Does Blocking GIP Cause Weight Loss? Tirzepatide Does the Opposite

Tirzepatide switches the GIP receptor on and produces more weight loss than any approved drug before it. A monthly injection that switches the same receptor off produced 16.3% to 19.9% weight loss in a 592-person phase 2.¹ Both directions work, and the field has not established why.

A 2025 review of the evidence behind both drug programs puts it plainly: "the fundamental question of whether to agonize or antagonize GIPR to treat obesity remains unanswered."²

What did the MariTide phase 2 show?

Weight loss "ranged from 16.3% to 19.9% with MariTide, compared to 2.6% with the placebo arm" in the cohort without type 2 diabetes, and 12.3% to 16.2% against 2.5% on a more conservative analysis.¹ Blocking the GIP receptor does appear to work on the human evidence available so far.

The drug carries a generic name too, maridebart cafraglutide, and it combines a GLP-1 receptor agonist with a GIP receptor antagonist in one molecule. The trial enrolled 592 adults across two cohorts for 52 weeks.¹ In the cohort with type 2 diabetes the figures were 12.1% to 17.0% against 1.4%, and 8.4% to 12.3% against 1.7%.¹ Amgen also reports that weight loss "had not plateaued by 52 weeks."¹

Glucose control moved too. In the participants with obesity and type 2 diabetes, Amgen reports a "robust and sustained reduction in hemoglobin A1c (HbA1c) of up to 2.2%."¹ Gastrointestinal events were the most frequent adverse events and mostly mild to moderate, with discontinuations for those events running "up to 7.8%" in the dose escalation arms.¹

Those are company-reported figures from a conference presentation, not numbers read out of a journal. The distinction matters for how much weight to put on them.

What does tirzepatide's label say about GIP?

Section 12.1 names both receptors and claims nothing about what each one contributes. The Mounjaro label opens it with "Tirzepatide is a GIP receptor and GLP-1 receptor agonist" and adds that tirzepatide "selectively binds to and activates both the GIP and GLP-1 receptors, the targets for native GIP and GLP-1."³

Nothing in that section separates the two effects. It does not say how much of the weight loss comes from the GIP half, or whether the GIP component is needed at all. Our comparison of semaglutide, tirzepatide and retatrutide covers what the clinical results look like side by side.

An approved drug can work without its own label explaining which half is doing the work. That situation is ordinary, and the gap it leaves is where the antagonist programs walked in.

What does GIP normally do?

GIP normally does two jobs, and one of them is handling fat. It is described in the review as "an intestinally derived, insulinotropic hormone commonly referred to as an incretin" and notes that "Endogenous GIP is secreted most potently in response to high lipid meals."²

Two effects follow. GIP helps "potentiate insulin secretion and improve glycemia in healthy volunteers", the incretin role it shares with GLP-1.² It also "serves a physiologic function to respond to lipid ingestion and promote lipid storage", and GIP infusion "enhances adipose tissue blood flow in lean patients."²

That second role is why blocking the receptor was ever a weight-loss idea. A hormone whose normal job includes promoting fat storage is an obvious thing to switch off.

One wrinkle sits in the diabetes data. In patients with type 2 diabetes, "Exogenous GIP-potentiated insulin secretion is nearly ablated", so the incretin half of GIP's effect is already blunted in the population both drug classes are tested in.²

What does the genetic evidence actually show?

Genetic evidence points in both directions. The paradox shows up one level down. Antag calls the receptor "a genetically-validated drug target" and Amgen says it "used its genetic expertise to identify GIP receptor inhibition as a key factor in reducing body mass."⁴¹ Neither phrase survives contact with the actual genetics.

Three missense variants in the receptor gene, "Arg190Gln, Glu288Gly, and Glu354Gln", "have been separately reported to associate with increased body mass index (BMI)" and also with "decreased adiposity."² The same variants, different reports, opposite directions.

Separately, variants associated with reduced type 2 diabetes risk, used as a stand-in for stronger receptor signaling, are "associated with reduced BMI and triglyceride levels."² That inference depends on assuming increased receptor action lowers diabetes risk, and the review states the assumption is "not fully established."²

Read on its own, the human genetics does not pick a side. Both companies describe a validated target and both are describing the same ambiguous evidence from opposite ends of it.

Has anyone shown which direction is right?

Nobody has shown which direction is right. Examining the evidence that launched both programs, the 2025 review reached no verdict, closing with the observation that "further insights into this unresolved paradox of GIPR agonism versus antagonism are undoubtedly on the horizon."²

Its summary of the two evidence bases is short on both sides. For agonism, "long-acting GIPR monoagonists, delivered alone or in combination with GLP-1R agonists, effectively lowered body weight", and a GIPR mono-agonist "has been reported to reduce body weight in healthy volunteers."² For antagonism, a potent antagonizing antibody abolishes almost all diet-induced weight gain in mice, and adding a GLP-1 agonist roughly doubles what that agonist achieves alone.²

Three explanations are on the table. Each gets examined in turn.

Does GIP agonism do anything on its own?

One reading says it does not. Possibility one is that "the GIPR agonism has no effect on its own."² That reading would make tirzepatide's weight loss a GLP-1 effect with a GIP component that neither helps nor hurts.

That reading is hard to square with the comparative results. Tirzepatide outperforms semaglutide on weight loss in head-to-head terms, and semaglutide is a GLP-1 agonist without the GIP half. If the GIP component contributed nothing, the difference would have to come from somewhere else in the molecule.

Beyond weight and glucose, the label records one more effect. Tirzepatide "increases insulin sensitivity, as demonstrated in a hyperinsulinemic euglycemic clamp study" at 28 weeks.³ Section 12.1 does not attribute that to either receptor, so the measurement sits in the same unallocated pile as the weight loss.

Does constant GIP stimulation end up acting like blockade?

Constant GIP stimulation may end up acting like blockade. Desensitization is the one most often repeated. Chronic GIP receptor agonism may induce what the review calls "functional antagonism", meaning a receptor stimulated continuously stops responding and the practical result resembles a block.² Our page on AT-7687 describes the same mechanism as constant stimulation wearing out the fat-cell receptor until it behaves as if blocked.

Mechanistically the appeal is obvious. If agonism becomes antagonism in practice, then the two drug programs are not opposites at all and the paradox dissolves, and a single mechanism explains two sets of clinical results.

Receptor biology makes it reasonable, not exotic. Continuous stimulation of a G protein-coupled receptor commonly reduces its surface expression and its signaling over time, so a long-acting agonist producing a functionally quiet receptor is not a strange proposal. Reviewers list it as one of three candidate explanations without endorsing it.²

Why does the dose comparison undercut that?

The dose comparison undercuts it because the doses do not line up. Here the objection is specific: "It is challenging to reconcile the functional antagonism hypothesis with the observation that an agonist delivered at much lower doses achieves a comparable degree of weight loss to an optimized antagonist."²

Read that slowly. A receptor that needs continuous heavy stimulation before it quits should require more drug, not less, to reach the same endpoint as a purpose-built blocker. The agonist gets there on less.

Dose comparisons across molecules are not clean, so the objection demotes desensitization without retiring it: potency, half-life, tissue distribution and receptor occupancy all differ between a peptide agonist and an antibody antagonist, and a dose in milligrams is not a measure of receptor engagement. What would settle it is a direct measurement of adipocyte and neuronal GIP receptor signaling under chronic agonist dosing in people, and no such measurement has been published.

Do the two approaches work through different routes?

The two approaches may work through different routes, and the review treats that account as the best supported. Its wording is that "there is compelling evidence that GIPR agonism and antagonism act via different neuronal mechanisms to drive weight loss."²

Different routes to the same outcome would explain everything the other two hypotheses strain to cover. Two drugs could move the same number on a scale through separate pathways, respond to different doses, and combine with GLP-1 agonism in different ways, without either one being a disguised version of the other.

It also sets up a question nobody has answered: if the routes differ, blocking and activating might not be interchangeable in their effects on anything other than weight.

One caveat belongs on that account. The primary papers behind it, including work on which brain regions mediate each effect, are paywalled and were not read for this page, so what appears here is the review's characterization of that evidence and not a reading of the experiments themselves.²

What did the preclinical work say about combining the two?

Blocking GIP alongside GLP-1 agonism beat either alone in animals. That finding is what both antagonist programs were built on. Amgen states that it "used its genetic expertise to identify GIP receptor inhibition as a key factor in reducing body mass", and that in preclinical studies "simultaneously activating GLP-1 and inhibiting GIP pathways had a stronger effect on weight loss" than targeting either receptor by itself.¹

Published work describes the same pattern. A potent GIP receptor antagonizing antibody "nearly abolishes HFD induced weight gain in mice", and combining it with the GLP-1 agonist liraglutide "nearly doubles the weight loss seen with liraglutide alone."²

The knockout animals carry the same ambiguity. Mice without the GIP receptor are "protected from diet-induced obesity (DIO) when placed on a high fat diet (HFD)", with "no body-weight phenotype on a chow diet", a shift toward lipid catabolism and "no change in food intake."² That reads as clean support for blocking the receptor until the comparison arrives: the review notes that "whole body GLP-1R KO mice demonstrate much the same phenotype as the GIPR KO animals."² That overlap, it says, "calls into question the translational relevance" of the models.²

Animal results motivated both programs, as usual. They are also the reason the human readouts matter more than the rationale behind them.

What is AT7687?

AT7687 is a peptide designed to block the GIP receptor, from the Danish company Antag Therapeutics. Its own description is a molecule "specifically designed to target and deactivate the GIP receptor."⁴ The company calls that receptor "a genetically-validated drug target contributing negatively to adiposity, insulin resistance, and cardiometabolic dysfunction."⁴

Our AT-7687 compound page carries the structural detail, the registry record and the pricing position. Nothing is for sale, because the molecule has never been marketed.

Phase 1 produced no efficacy figure. Antag reports a "favorable safety and tolerability profile" across single and multiple ascending dose cohorts, that "No severe or serious adverse events were observed in the study", and that "no gastrointestinal tolerability signals were identified."⁴ No weight figure from phase 1 appears in that announcement.

What is the AT7687 trial testing?

The AT7687 trial is testing weight loss on top of semaglutide, against semaglutide alone. The phase 2a is a "multi-center, randomized, double-blinded 13-week trial" in which "150 participants will be randomized", in "people living with overweight or obesity and type 2 diabetes who are not currently receiving anti-obesity medications."⁴

The control arm receives "placebo with semaglutide dose escalation", so both groups get a GLP-1 agonist and the question is what adding GIP blockade does on top.⁴ Change in body weight from randomization is the primary endpoint, with "change in HbA1c as the key secondary endpoint."⁴ Registration is NCT07724340, and Antag expects "Topline results expected in H1 2027."⁴

A 13-week readout answers a narrow question. Thirteen weeks is long enough to see a weight difference open up. It is far too short to say where it ends.

What will the AT7687 result not tell you?

The AT7687 result will not tell you which direction is correct. A positive result would show that GIP blockade adds something to semaglutide over 13 weeks. It would not show that blockade beats activation, because no trial has put a GIP antagonist against a GIP agonist.

Nothing in the AT7687 design compares it with tirzepatide. Antag's comparator is placebo on a semaglutide background, correct for the question it is asking and wrong for the question the paradox poses.

A negative result would be similarly bounded. One antagonist failing to add weight loss to semaglutide in 150 people over 13 weeks would not close the antagonist case, any more than one agonist succeeding has settled it.

Does any of this change what is on sale now?

None of this changes what is on sale now. AT7687 has never been approved or marketed anywhere, and maridebart cafraglutide is in development, not on pharmacy shelves. Tirzepatide is the approved drug in this story, and it is the agonist.

Anything sold under the AT7687 name did not come from the trial supply chain, since no such supply chain exists outside the study. Our guide to why most peptides have no human evidence covers the general version of that problem, and this compound is an unusually clean example: one registered trial, no posted results, no publication.

Practically, the reading is narrow. A GIP antagonist sold today is not the molecule in the Antag trial and not the molecule in the Amgen trial, because neither has a commercial form. Our MariTide page tracks what gets listed under that name. Anybody buying on the strength of the MariTide figures is buying on results produced by a different molecule, at a different dose, in a trial that enrolled 592 people.

Reading peptide marketing is where this lands, because the paradox gets used in both directions. A seller can point at tirzepatide and argue that GIP activation is the proven route. Another can point at the antagonist phase 2 and argue that blocking is the smarter one. Both can cite real human data, and neither is entitled to the conclusion.

Unresolved is where the review leaves it, with three live explanations and no head-to-head trial.² Anybody claiming the question is settled is selling something, and the molecule they are selling has no approved form.

Sponsor language repays close reading for the same reason. Amgen describes GIP receptor inhibition as "a key factor in reducing body mass" and Antag describes the receptor as a genetically validated target contributing negatively to adiposity and insulin resistance.¹⁴ Both statements are about a target, not about a result in people, and both companies are careful about that distinction in a way that resellers are not.

What has not been measured?

A head-to-head comparison has never been run. No trial has given one group a GIP agonist and another a GIP antagonist on the same GLP-1 background. That experiment would answer the question directly.

Long-term outcomes are the second gap. The longest antagonist data runs 52 weeks and had not plateaued, so the curve's shape past a year is unknown on that side.¹ Our guide to GLP-1 side effects covers what is established on the agonist side, where the evidence base is years deeper.

Tissue-level effects are the third. If agonism and antagonism act through different neuronal routes, then bone, insulin sensitivity and cardiovascular endpoints need measuring separately for each approach, and the published record treats weight as the primary question.² Nothing else has been measured on both sides.

Common questions

Does blocking the GIP receptor cause weight loss?

Why does tirzepatide activate GIP if blocking it also works?

What does the Mounjaro label say about GIP?

Is chronic GIP stimulation the same as blocking it?

What is the best-supported explanation?

What is AT7687?

What is the AT7687 phase 2a testing?

When will AT7687 results be available?

Did AT7687 show weight loss in phase 1?

Will the AT7687 trial settle the agonist versus antagonist question?

Is AT7687 the same as the drug named in the registry?

Can I buy AT7687?

Does MariTide have approval?

What does GIP do normally?

Is the GIP receptor a genetically validated target?

Do mice without the GIP receptor stay lean?

Which drug is the agonist and which is the antagonist?

Sources

Sources

1. Amgen. Results From Amgen's Phase 2 Obesity Study of Monthly MariTide Presented at the American Diabetes Association 85th Scientific Sessions. Company announcement, 23 June 2025. Read at source at investors.amgen.com. A sponsor announcement reporting results presented at a conference, cited for the bispecific GLP-1 agonist and GIPR antagonist mechanism, the 592 participants, the 52-week duration, the weight-loss ranges on both analyses in both cohorts, the absence of a plateau and the gastrointestinal adverse event characterization. The peer-reviewed report of this trial is paywalled and was not read, so no figure here comes from the journal version. ↩1 ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8 ↩9 ↩10 ↩11 ↩12 ↩13 ↩14
2. Douros JD, Mowery SA, Knerr PJ. The Premise of the Paradox: Examining the Evidence That Motivated GIPR Agonist and Antagonist Drug Development Programs. Journal of Clinical Medicine 2025;14(11):3812. DOI 10.3390/jcm14113812. Read at source, full text. Quoted for the unresolved-question statement, the summaries of the agonism and antagonism evidence bases, all three competing explanations, the dose objection to functional antagonism, the different-neuronal-mechanisms wording and the closing position. The article's figures are images and their underlying data are not in the text, so no figure data is quoted here. ↩1 ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8 ↩9 ↩10 ↩11 ↩12 ↩13 ↩14 ↩15 ↩16 ↩17 ↩18 ↩19 ↩20 ↩21 ↩22 ↩23 ↩24 ↩25 ↩26 ↩27 ↩28 ↩29 ↩30 ↩31
3. Mounjaro (tirzepatide) injection, US prescribing information, section 12.1. Read at source via the openFDA drug label API (https://api.fda.gov/drug/label.json), mechanism_of_action field. Quoted for the receptor agonist description and the binding statement. The section contains no statement separating the contributions of the two receptors, and the page says so rather than inferring one. ↩1 ↩2 ↩3 ↩4
4. Antag Therapeutics. Antag Therapeutics Announces Dosing of First Patient in Phase 2a Trial of First-in-Class GIPR Antagonist AT7687. Company announcement, Copenhagen, 28 July 2026. Read at source at antagtx.com. This is a sponsor announcement, not a peer-reviewed report or a regulatory document, and it is cited only for the trial design, the registration number, the company's own mechanism description and its characterization of phase 1 safety. No efficacy figure for AT7687 exists in it or anywhere else. ↩1 ↩2 ↩3 ↩4 ↩5 ↩6 ↩7 ↩8 ↩9 ↩10 ↩11 ↩12 ↩13 ↩14 ↩15