MariTide (AMG 133): Amgen's monthly obesity & diabetes injection
Amgen's experimental once-monthly injection that activates one hunger-controlling hormone signal and blocks another to reduce body weight and blood sugar, not yet an approved drug.
A researcher, an agent, or an algorithm wrote down the sequence and picked a target to hit.
An AI model like OpenFold3 or AlphaFold built a 3D structure and scored how well it fits the binding site.
A second contributor repeated the computation on their own hardware and the scores matched.
A chemistry service or a researcher ordered the sequence, it was manufactured, and mass spectrometry confirmed the right molecule was produced.
A binding or activity measurement confirmed that it actually does what the computer predicted — or didn't.
What this is
MariTide (maridebart cafraglutide, formerly AMG 133) is Amgen's investigational drug for obesity and type 2 diabetes — a once-monthly injection that does something no other approved obesity drug does: it uses an antibody as its backbone, with two GLP-1 activating peptide arms attached to it. The antibody component blocks the GIP receptor while those peptide arms activate the GLP-1 receptor simultaneously, making MariTide a bispecific antibody-peptide conjugate that hits both incretin receptors at once but in opposite directions. This mechanistic direction is the inverse of tirzepatide, which activates both receptors rather than blocking one; the GIP-antagonism approach was motivated by large-scale human genetics data showing that people who carry loss-of-function variants in the GIPR gene tend to have lower BMI. The antibody scaffold gives MariTide an approximately 21-day half-life, enabling once-monthly (Q4W) or every-eight-weeks (Q8W) subcutaneous dosing — roughly 12 injections per year compared with 52 for weekly GLP-1 agents. Because MariTide is a biologics-manufacturing product requiring mammalian cell culture, it cannot be reproduced by standard peptide synthesis; any product sold outside Amgen's registered trials claiming to be MariTide should be presumed to be something else.
History
MariTide — originally identified by its internal Amgen code AMG 133 and formally named maridebart cafraglutide — originated in Amgen's metabolic biologics program in the late 2010s and early 2020s. The mechanistic rationale drew on human genetics: large-scale sequencing had linked loss-of-function variants in the GIPR gene to lower BMI, suggesting that pharmacological GIPR antagonism might be weight-reducing. Amgen's approach was to test this directly by engineering a bispecific molecule combining a fully human GIPR-antagonist monoclonal antibody with GLP-1 agonist peptide arms attached via amino acid linkers — an unusual format in obesity research, where antibody scaffolds are well-established in oncology and autoimmune disease but had not previously been deployed in the incretin class. Phase 1 results were published in Nature Metabolism in early 2024 (Véniant and colleagues), demonstrating dose-dependent weight loss and a prolonged post-treatment effect, with weight loss maintained up to 150 days after the last dose. The Phase 2 trial results were published in the New England Journal of Medicine in September 2025 (n=592, 52 weeks), reporting up to approximately 21.6% weight loss at the highest dose (efficacy estimand), with weight-loss curves still declining at trial end. Amgen launched the Phase 3 MARITIME program in March 2025, with MARITIME-1 and MARITIME-2 targeting obesity and obesity-plus-T2D respectively, alongside planned cardiovascular, heart failure, sleep apnea, and chronic kidney disease programs.
What it does
MariTide's two active components work together: the GLP-1 agonist peptide arms suppress appetite, slow gastric emptying, and promote glucose-dependent insulin release — the same downstream effects as approved GLP-1 drugs like semaglutide. The antibody component simultaneously blocks the GIP receptor, adding a second weight-reducing signal motivated by the genetics data above. Because the antibody scaffold gives MariTide an approximately 21-day half-life (compared with roughly one week for semaglutide or five days for tirzepatide), appetite suppression and early satiety develop smoothly across the monthly dosing interval rather than oscillating. When treatment is stopped, the drug clears gradually — Phase 1 data showed weight loss maintained up to 150 days after the last dose, substantially longer than the four-to-six week washout typical of weekly GLP-1 agents (Véniant and colleagues, Nature Metabolism 2024).
Evidence
- Human: Moderate — Phase 1 (Véniant and colleagues, Nature Metabolism 2024) demonstrated dose-dependent weight loss and post-treatment weight maintenance up to 150 days after the last dose. The Phase 2 NEJM 2025 trial enrolled 592 participants across obesity and obesity-plus-T2D cohorts over 52 weeks, reporting weight loss of 12.3% to 16.2% under the treatment-policy estimand and up to approximately 21.6% under the efficacy estimand at the highest dose in the obesity cohort; HbA1c reductions of 1.2 to 1.6 percentage points were reported in the obesity-plus-T2D cohort. No weight-loss plateau was observed at 52 weeks. Phase 3 MARITIME-1 (approximately 3,500 participants, 72 weeks) and MARITIME-2 (approximately 999 participants, obesity plus T2D) launched March 2025 with primary readouts expected early 2027.
- Animal: Strong — preclinical studies in diet-induced obesity (DIO) mice and cynomolgus monkeys confirmed GIPR antagonism combined with GLP-1R agonism, demonstrating weight reduction and improved metabolic parameters (Véniant and colleagues, Nature Metabolism 2024). A separate head-to-head preclinical comparison of GIPR agonism versus GIPR antagonism in male mice was subsequently published, further characterizing the mechanistic distinction between the two approaches.
- Mechanistic: Both GIPR antagonism (MariTide) and chronic GIPR agonism (tirzepatide) have produced clinically meaningful weight loss in humans despite targeting the same receptor in opposite directions — a pharmacological question the field calls the "GIPR paradox." Characterization of GIPR genetic variants and β-arrestin signaling has been published (Kizilkaya and colleagues, Nature Metabolism 2024), supporting the genetics rationale for the antagonism approach. The full biology of why both directions produce weight loss remains an active research area, with receptor desensitization under chronic agonism as one working hypothesis (Madsbad and colleagues, Expert Opinion on Investigational Drugs 2025).
Known effects
- Weight loss in adults with obesity — Phase 2 human RCT; up to ~21.6% at 52 weeks (efficacy estimand); no plateau observed
- HbA1c reduction in adults with obesity + T2D — Phase 2 human RCT; 1.2–1.6 percentage points at 52 weeks
- Post-treatment weight maintenance — Phase 1 observation up to 150 days post-last-dose; mechanism is slow drug clearance given the ~21-day half-life
- Appetite suppression and slowed gastric emptying — mechanistic, consistent with GLP-1 class pharmacology
Safety signals
Adverse events reported in the Phase 2 trial (NEJM 2025) were predominantly gastrointestinal: nausea (most common, predominantly after the first dose), vomiting, and constipation. Extended dose-escalation schedules (4-week and 12-week ramps before reaching maintenance) substantially reduced GI adverse event rates compared with direct-to-maintenance starts. Discontinuation due to adverse events was 11% in dose-escalation arms per the Phase 2 report.
A pharmacological feature specific to monthly dosing is a tolerability asymmetry: if significant GI adverse events arise after a monthly dose, exposure cannot be quickly reduced — the drug persists at pharmacologically active levels for approximately three weeks. This distinguishes MariTide from weekly agents where dose reduction takes effect within days.
Immunogenicity (anti-drug antibody development) is an inherent consideration for antibody biologics and is being monitored across the MARITIME program; early data from Phase 2 were described as manageable, but long-term rates and clinical consequences require extended follow-up beyond 52 weeks.
By analogy to the GLP-1 class, medullary thyroid carcinoma history and Multiple Endocrine Neoplasia type 2 were applied as exclusion criteria in MARITIME trial protocols, consistent with standard GLP-1 class precautions. Cardiovascular outcomes for MariTide specifically are not established; MARITIME-CV is designed to test MACE reduction but has not yet reported. Long-term safety beyond 52 weeks is not available.
Regulatory status
- US (FDA): Not approved. Investigational; access limited to enrollment in Amgen's MARITIME Phase 3 program or related registered trials. No compounded equivalent is feasible given the antibody-biologic manufacturing requirements.
- EU and international: No regulatory authority has authorized MariTide for marketing. MARITIME trial sites span North America, Europe, and parts of Asia.
- WADA: WADA's S0 category prohibits substances not currently approved by any governmental regulatory health authority for human therapeutic use; MariTide meets that criterion. Athletes subject to WADA or equivalent anti-doping bodies should treat MariTide as prohibited.
Mechanism
MariTide is a bispecific antibody-peptide conjugate: the antibody component (the cafraglutide scaffold) is a fully human GIPR antagonist, while two GLP-1 agonist peptide arms (the maridebart moieties) are conjugated to the antibody via amino acid linkers. The stored 31-amino-acid sequence represents the GLP-1 peptide backbone; the antibody scaffold and linker chemistry that make MariTide a biologics-class molecule are not captured in that sequence.
GLP-1R agonism drives appetite suppression, slowed gastric emptying, and glucose-dependent insulin secretion — pathways shared with approved GLP-1 agents. GIPR antagonism embeds the pharmacological hypothesis motivated by human genetics: loss-of-function variants in GIPR associate with lower BMI in large-scale sequencing data, implying that blocking GIPR may reinforce weight-reducing signals. This is mechanistically opposite to tirzepatide's GIPR agonism, yet both approaches have produced clinically meaningful weight loss in human trials, giving rise to the "GIPR paradox." One working explanation involves receptor desensitization: chronic high-level agonism may functionally mimic antagonism through downregulation; the full biology remains an active research area (Madsbad and colleagues, Expert Opinion on Investigational Drugs 2025).
The antibody scaffold confers an approximately 21-day half-life — substantially longer than semaglutide (~1 week) or tirzepatide (~5 days) — enabling once-monthly or every-eight-weeks subcutaneous dosing and producing a prolonged post-treatment washout. Phase 1 data showed weight-loss maintenance up to 150 days after the last dose, interpreted as a consequence of gradual plasma-level decline below pharmacologically active thresholds rather than durable alteration of adipose biology (Véniant and colleagues, Nature Metabolism 2024).
Myths and misconceptions
"MariTide is available if you know where to look." MariTide is an investigational antibody-peptide conjugate in active Phase 3 development. Legitimate access is limited to enrollment in Amgen's MARITIME program or related registered trials. The molecule requires mammalian cell-culture biologics manufacturing and cannot be meaningfully reproduced by peptide synthesis alone — any product sold as "MariTide" in research-chemical channels should be presumed not to be MariTide.
"Monthly dosing makes MariTide strictly better than weekly GLP-1s." Monthly dosing is a genuine convenience advantage, but it also removes the ability to quickly adjust exposure if side effects arise. GI tolerability during titration is harder to manage when the last dose persists for three weeks. Amgen addressed this in Phase 2 and 3 with extended dose-escalation schedules, which substantially reduced GI adverse events, but the tolerability asymmetry is a real feature of the dosing paradigm.
"GIPR antagonism and GIPR agonism can't both work — one must be wrong." Both mechanisms have produced meaningful weight loss in human trials: tirzepatide (GIPR agonism plus GLP-1 agonism) and MariTide (GIPR antagonism plus GLP-1 agonism). The working explanation in the field involves receptor desensitization — chronic agonism may functionally mimic antagonism through downregulation — but the full biology is still being characterized. The "GIPR paradox" is an active research question, not a refutation of either program.
"The 150-day post-treatment effect means you only need occasional doses." Phase 1 post-treatment observation showed weight maintenance up to 150 days — a pharmacokinetic consequence of the ~21-day half-life, not evidence of durable disease modification. Obesity biology is expected to reassert adipose set-points once pharmacological levels fall below active thresholds; the 150-day window describes how long that withdrawal takes, not that the underlying condition has been altered.
Open questions
- Phase 3 efficacy ceiling: The Phase 2 weight-loss curves showed no plateau at 52 weeks. MARITIME-1 at 72 weeks is designed to determine whether the trajectory continues, flattens, or varies by dose arm.
- Head-to-head comparison with tirzepatide and semaglutide: No randomized head-to-head study has been reported. Cross-trial comparisons are confounded by differences in estimand methodology, population, and follow-up duration.
- Cardiovascular outcomes: MARITIME-CV is designed to test MACE reduction. Until it reports, cardiovascular benefit for MariTide is inferred from GLP-1 class evidence rather than established directly.
- Long-term immunogenicity: Anti-drug antibody development over two or more years of monthly dosing requires extended Phase 3 and post-marketing follow-up.
- Washout and weight-regain kinetics at 12–24 months: Phase 1 showed weight maintenance up to 150 days post-last-dose. Whether the slower washout translates into meaningfully different long-term weight-regain curves versus weekly GLP-1 cessation has not been directly tested.
- The GIPR paradox: Why both pharmacological GIPR antagonism and chronic GIPR agonism produce weight loss despite opposite receptor directions remains an active mechanistic question.
- Real-world tolerability of monthly dosing: Discontinuation patterns when GI adverse events cannot be quickly managed by dose reduction are untested outside trial settings.
Related peptides
- Tirzepatide — dual GIP/GLP-1 receptor agonist; activates GIPR rather than blocking it, making it the mechanistic foil to MariTide; approved for obesity and T2D
- Semaglutide — GLP-1 receptor agonist; once-weekly injection; the benchmark weekly GLP-1 agent against which MariTide's monthly convenience and efficacy trajectory are discussed
- Retatrutide — triple GIP/GLP-1/glucagon receptor agonist; another next-generation incretin in late-stage development
Research directions for this peptide, selected from the current sources — hypotheses you can explore and model. None of it is proven yet; tap any one to see the full thinking.
Because MariTide releases its active components slowly over a month rather than in a weekly pulse, could it avoid the nausea spike that happens after each weekly GLP-1 shot?
If flatter drug levels mean less nausea, patients could more reliably reach the doses needed for significant weight loss, and adherence to the treatment could improve substantially.
Does blocking the GIP receptor in fat tissue specifically reduce the deep abdominal fat that drives diabetes and heart disease, rather than just lowering body weight overall?
Losing visceral fat specifically improves insulin sensitivity and heart disease risk more than losing the same amount of subcutaneous fat; if MariTide achieves this, patients could see better health outcomes per kilogram lost compared to other obesity drugs.
Could the part of MariTide that blocks GIP receptors also protect bones from the thinning that often happens with rapid weight loss?
For older patients and postmenopausal women who are already at risk of fractures, an obesity drug that protects bones during weight loss would be much safer than current options, which can accelerate bone thinning.
Does MariTide's antibody block one receptor while its peptide arms activate another, without either arm interfering with the other?
If the two actions are truly independent, MariTide is unlikely to develop tolerance or resistance that undermines both effects at once, potentially making it a more durable long-term treatment than drugs targeting a single receptor.
▸full evidence table2 metrics
| metric | value | tool |
|---|---|---|
| ipTM | 0.7974955439567566 | openfold3-mlx |
| ranking score | 0.8522500991821289 | openfold3-mlx |
▸structural qualityopenfold3
| metric | value | note |
|---|---|---|
| gpde | 0.741 | global PDE — lower = better |
| disorder | 0.141 | fraction disordered |
| chain pair ipTM (A, B) | 0.797 | interface quality |
▸3-letter notation
▸recipeopenfold3-mlx 0.3.1
| parameter | value |
|---|---|
| model | openfold3-mlx 0.3.1 |
| weights | — |
| hardware | — |
| mlx version | — |
| python | — |
| random seed | — |
| msa strategy | — |
| diffusion samples | 1 |
| runtime | 431s |
| predicted by | mlx@peptide |
| predicted at | 2026-05-03 |
▸citationbibtex
@peptide{pep10902,
sequence = {HAEGTFTSDVSSYLEGQAAKEFIAWLVKGRX},
target = {gipr},
author = {peptidemodel},
year = {2026},
status = {computed}
}