Maridebart cafraglutide, known as MariTide and previously designated AMG 133, is an investigational antibody–peptide conjugate developed by Amgen for obesity and metabolic disease. It combines GLP-1 receptor activation with GIP receptor antagonism and has reduced weight in a phase 2 trial involving adults with and without type 2 diabetes. Its design also produces prolonged exposure that supports investigation of less frequent administration. The antibody, peptide attachments and two receptor activities are all necessary to understanding how this candidate differs from smaller metabolic peptides.
Overview
Véniant and colleagues (2024) describe a fully human monoclonal antibody against the GIP receptor carrying two GLP-1 analogue peptides through linkers. The antibody supplies receptor-blocking activity and a large protein scaffold; the attached peptides supply GLP-1 receptor agonism. The reported average molecular mass of the conjugate is 153,514 Da.
Molecular profile
- Development names
- MariTide; AMG 133 Source
- Molecular design
- Antibody–peptide conjugate Source
- Protein component
- Fully human monoclonal GIP-receptor antagonist antibody Source
- Peptide attachments
- Two GLP-1 analogue agonist peptides per antibody Source
- Reported average conjugate mass
- 153,514 Da Source
- Receptor activities
- GIP receptor antagonism; GLP-1 receptor agonism Source
- Clinical status
- Investigational; MARITIME phase 3 program Source
Forms and derivatives
AMG 133 / MariTide
Development and program names for maridebart cafraglutide.
Intact and total AMG 133 analytes
Laboratory definitions used to follow exposure. Intact assay requires at least one GLP-1 attachment; total assay also counts antibody that has lost all peptide attachments.
MariTide's molecular identity includes the antibody scaffold and its peptide attachments. Its incretin-receptor pharmacology comes from both components, so a reported sequence or mass needs to identify which part of the conjugate it describes.
Uses and research areas
Clinical research examines chronic weight management and glycemic effects in type 2 diabetes, with further studies addressing cardiovascular disease, heart failure and obstructive sleep apnea. The MARITIME development program reported by Amgen in 2026 includes phase 3 weight-management studies and trials aimed at specific complications. Those complication trials are important because a lower weight or HbA1c does not itself establish fewer cardiovascular events or improved sleep apnea.
Maintenance is another research area. A treatment that produces an initial response must also be evaluated for continued use, persistence of benefit and tolerability over time. MariTide's prolonged exposure makes less frequent maintenance schedules a development question, and the phase 2 extension selected people who had already responded to the first treatment year.
Mechanism of action
Receptor activation and receptor blockade in the same conjugate
GLP-1 and GIP are incretin hormones that participate in nutrient-responsive insulin secretion and metabolic regulation. An agonist activates its receptor; an antagonist interferes with activation by the receptor's usual ligand. MariTide's antibody component binds the GIP receptor and blocks GIP signaling, while its peptide components activate GLP-1 receptors. Investigators confirmed these activities in cell-based assays in the original translational program.
The discovery work by Wu and colleagues (2026) explains how the antibody, linker and peptide were optimized as a conjugate. The linker must permit the peptide to reach and activate its receptor while remaining attached to the protein scaffold. Changes to the peptide and attachment design can therefore alter both receptor response and how long the active conjugate persists.
Why GIP antagonism deserves its own explanation
Tirzepatide combines GLP-1 and GIP receptor agonism, whereas MariTide pairs GLP-1 agonism with GIP antagonism. Both are approaches to metabolic treatment, but a receptor can affect several tissues and cellular signaling processes. Its effect under sustained pharmacological stimulation is not fully described by whether an isolated assay records activation or blockade.
Kizilkaya and colleagues (2024) characterized human GIP receptor variants and compared carriers' metabolic traits. They distinguished Gs signaling, which drives production of the cellular messenger cAMP, from recruitment of beta-arrestin, a protein involved in receptor signaling and movement within cells. In the UK Biobank, variants impairing both processes were associated with lower BMI and other adiposity traits. Variants impairing Gs signaling while retaining beta-arrestin recruitment did not show those associations.
The cellular experiments helped explain that difference. Receptors retaining beta-arrestin recruitment continued Gs signaling after entering endosomes, internal membrane compartments; variants with impaired recruitment had little or no signaling from those compartments. A reduced response in one assay can therefore coexist with other preserved receptor functions. These observations make receptor location and sustained signaling relevant to GIP-directed strategies, while MariTide's clinical trials evaluate the effect of its complete conjugate.
Pharmacokinetics
What remains attached to the antibody
MariTide requires two exposure measurements. The phase 1 program's intact-conjugate assay counted antibody molecules with at least one GLP-1 analogue still attached. Its total-antibody assay also counted antibody material that had lost the peptide attachments. A longer-lasting total-antibody signal can therefore continue after part of the conjugate's GLP-1 component has been lost.
In Véniant and colleagues' human data (2024), the six-person, lowest single-dose cohort had a geometric mean terminal half-life of 14.3 days for intact AMG 133 and 21 days for total AMG 133. Median time to peak was five days for intact conjugate and 5.9 days for the total-antibody measurement. The distinction between analytes is essential to interpreting a claim about MariTide's half-life.
Population: Six adults with obesity, without diabetes, in the lowest single-dose cohort. Administration: Subcutaneous administration in the source study. Analyte: Intact AMG 133, defined as antibody with at least one attached GLP-1 analogue.
Véniant and colleagues (2024) reported a 14.3-day geometric-mean terminal half-life for intact AMG 133 in six adults with obesity after subcutaneous administration. The model begins at a normalized intact-conjugate concentration in the terminal phase and traces repeated halving at that rate. Absorption and early distribution are outside the illustration.
Véniant MM, Lu SC, Atangan L, Komorowski R, Stanislaus S, Cheng Y, et al. A GIPR antagonist conjugated to GLP-1 analogues promotes weight loss with improved metabolic parameters in preclinical and phase 1 settings. Nature Metabolism. 2024;6:290–303. doi:10.1038/s42255-023-00966-w. Véniant and colleagues (2024), Figure 4a, journal page 297 (PDF page 8), and clinical PK methods; single 21-mg subcutaneous cohort with six participants.
How the antibody scaffold extends exposure
Large antibody scaffolds reduce rapid renal filtration and can participate in FcRn-mediated recycling, as discussed in Wu and colleagues (2026). FcRn helps return some internalized antibody to circulation rather than directing all of it toward degradation. This supplies a biological explanation for persistence, while cleavage or loss of the peptide attachments contributes another route through which the conjugate's activity can change over time.
Research findings
The placebo-controlled phase 2 trial
Jastreboff and colleagues (2025) enrolled 592 participants in a dose-ranging trial with a 52-week primary weight endpoint. The cohort without diabetes included 465 people; the obesity–diabetes cohort included 127. Separating the cohorts allowed the researchers to evaluate weight responses in two clinically different populations and examine glycemic change in those with diabetes.
Using the treatment-policy estimand, mean weight reduction ranged from 12.3% to 16.2% across active groups without diabetes, compared with 2.5% with placebo. In the diabetes cohort, the corresponding range was 8.4–12.3%, compared with 1.7% with placebo. HbA1c fell by 1.2–1.6 percentage points in active diabetes groups, while the placebo mean increased by 0.1 points.
The treatment-policy analysis addresses the outcome of assignment to treatment, including the consequences of stopping it. Amgen's accompanying trial report (2025) also discussed larger estimates under an efficacy analysis. The two analyses answer different questions, which explains why summaries of this same trial can give different percentages without describing a different patient population.
Maintenance research in the second year
In its January 2026 presentation, Amgen reported an exploratory additional year in participants who had lost at least 15% of their weight in part 1. The sponsor stated that most participants maintained the earlier reduction with lower monthly or quarterly treatment and that cardiometabolic improvements persisted. It described the second-year tolerability favorably without providing a full numerical result table in that announcement.
The response-based selection is central to interpreting this extension. It follows people who had already obtained a substantial benefit, so its maintenance experience cannot be assumed for everyone who began the original trial. The extension is nevertheless useful for designing studies that examine a maintenance strategy after an initial response.
Safety and tolerability
MariTide remains investigational and is not FDA-approved as of October 13, 2026. Gastrointestinal adverse events were common in the phase 2 trial. Jastreboff and colleagues reported that lower starting amounts and escalation reduced their frequency. The result connects tolerability to how the treatment was introduced within the study, an issue particularly relevant to a candidate with prolonged systemic exposure.
Amgen's detailed 2025 report described nausea and vomiting as prominent events, with gastrointestinal-event discontinuation of up to 7.8% in the escalation groups. Investigators collected gastrointestinal symptoms through both standard adverse-event reporting and a daily patient questionnaire, which affects how completely those symptoms were captured. The published authors reported no unexpected safety signals, while larger and longer studies will characterize uncommon events more precisely.
The phase 3 program will also clarify safety and benefit in specific cardiometabolic populations. Current studies do not establish safety in pregnancy or childhood. The conjugate's manufacturing and attached-peptide structure are integral to its identity, so a product described only as a GLP-1 peptide would not establish that it reproduces the clinical material.
What the research means
MariTide provides human evidence for a combined GIP-antagonist and GLP-1-agonist strategy delivered through an antibody–peptide conjugate. The phase 2 weight and HbA1c findings support continued development, while its persistence supplies a reason to investigate less frequent treatment and maintenance. The latter goal requires direct study because a long half-life alone does not show that a particular interval preserves benefit or improves continuation.
The program's practical questions concern the whole conjugate: whether participants can sustain treatment, whether active peptide remains attached long enough for the intended regimen, and whether metabolic improvements translate into better clinical outcomes. Its receptor design is scientifically informative, but comparisons with tirzepatide or other candidates need matched clinical endpoints and methods for handling discontinuation.
FAQs
Is MariTide an amylin analogue?
MariTide's reported targets are the incretin receptors GIP and GLP-1. It is an antibody–peptide conjugate: the antibody blocks GIP receptors, and its two attached GLP-1 analogue peptides activate GLP-1 receptors. Véniant and colleagues (2024) describe this complete molecular design.
How does an antibody–peptide conjugate work here?
The antibody supplies GIP-receptor-blocking activity and a protein scaffold. Its attached peptides supply GLP-1-receptor activity. Wu and colleagues (2026) explain the development of the complete conjugate, including how linker and peptide choices affect its properties.
Why can MariTide have more than one reported half-life?
The measurements follow different analytes. An intact-conjugate assay requires a peptide attachment, while the total-antibody assay also follows antibody material after attachments are lost. The analyte must accompany any half-life value so readers can tell which molecular material persisted.
How does its GIP activity compare with tirzepatide?
MariTide's antibody blocks GIP receptors, while tirzepatide activates GIP receptors; both include GLP-1 receptor agonism. These differences make receptor signaling and tissue context useful research questions. Clinical benefit is assessed for the complete treatment in its own controlled trials.
Has quarterly maintenance been established for everyone starting treatment?
The reported extension selected participants who had already achieved substantial first-year weight reduction. Amgen's January 2026 presentation describes an exploratory maintenance experience in that selected group. It does not define an approved regimen.
Is MariTide approved?
MariTide remains investigational as of October 13, 2026. Amgen's 2026 program update describes ongoing phase 3 development for weight management and related clinical conditions.
Comparisons and related reading
Tirzepatide provides the clearest receptor comparison because it activates both GIP and GLP-1 receptors. Semaglutide focuses on GLP-1 receptor agonism. Zenagamtide, also called amycretin, combines GLP-1 and amylin-family activity within a peptide, offering another example of how molecular design can join metabolic pathways. MariTide's antibody scaffold adds a separate structural and pharmacokinetic dimension to these comparisons.
References
Defines the antibody–peptide structure and distinguishes intact-conjugate from total-antibody human exposure.
The conjugate-development paper explains selection of the antibody, peptide and linker and the exposure advantage of the protein scaffold.
Combined signaling defects were associated with lower adiposity; cellular experiments linked continued endosomal signaling to beta-arrestin recruitment.
The phase 2 trial reported weight loss in adults with and without diabetes and HbA1c changes in the diabetes cohort. Its treatment-policy analysis includes results after participants stopped assigned treatment.
Amgen’s report explains why estimates differ when analyses include treatment stopping versus assume continued treatment. It also describes daily symptom questionnaires and the number stopping because of gastrointestinal events during dose escalation.
Describes an exploratory second year in participants selected for a substantial first-year response.
Documents the ongoing MARITIME program and its weight-management and complication-focused studies.
Glossary
- Conjugate
- A molecular assembly formed by attaching distinct components through a defined linkage.
- Antagonist
- A substance that blocks or reduces activation of a receptor.
- Agonist
- A substance that activates a receptor.
- FcRn
- A receptor involved in recycling antibodies and albumin, helping prolong their persistence.
- Intact-conjugate analyte
- In the reported MariTide assay, antibody with at least one GLP-1 peptide attachment.
- Beta-arrestin
- A protein involved in receptor signaling and trafficking.
- Treatment-policy estimand
- A treatment-effect question that includes outcomes regardless of treatment discontinuation.