Mazdutide (IBI362 / LY3305677): Oxyntomodulin Backbone, GLP-1R/GCGR Co-Agonism, and the Lipidation Analytics Problem — A 2026 Research Guide

Most incretin compounds circulating in the research market are built on one of two scaffolds: the GLP-1 backbone (semaglutide, liraglutide) or the GIP backbone with GLP-1 activity engineered onto it (tirzepatide). Mazdutide is built on neither. It descends from oxyntomodulin — a 37-residue proglucagon cleavage product that mammals already make, and which has the unusual property of being a native, if weak, agonist at two different class B GPCRs simultaneously. That makes it a structurally distinct object in a market otherwise converging on the same handful of templates. It also creates an analytical problem: at 4,563 daltons with a C20 fatty diacid attached through a two-unit PEG linker to a specific lysine, this is a molecule where “correct mass” and “correct molecule” are not the same statement. This guide covers the molecular profile, the receptor-level mechanism, the published preclinical and in-vitro record, and what a certificate of analysis on a lipidated 33-mer actually has to demonstrate.

The information below describes in-vitro and preclinical laboratory research only. It is not medical guidance, and none of the findings described here establish safety or efficacy in humans.

Section 1 — Molecular Profile

Mazdutide is a long-acting synthetic analog of mammalian oxyntomodulin, developed by Innovent Biologics and licensed from Eli Lilly, and it appears in the literature under three names that all refer to the same molecule.

• Compound names: mazdutide; IBI362; LY3305677; OXM-3
• CAS number: 2259884-03-0
• Molecular formula: C210H322N46O67
• Molecular weight: approximately 4,563.1 Da (free base)
• Structure class: 33-residue linear synthetic peptide, C-terminally amidated, with a C20 fatty diacid (nonadecanoic acid) conjugated through a γGlu–AEEA–AEEA hydrophilic linker at the lysine in position 20
• Sequence: His-Aib-Gln-Gly-Thr-Phe-Thr-Ser-Asp-Tyr-Ser-Lys-Tyr-Leu-Asp-Glu-Lys-Lys-Ala-Lys(AEEA-AEEA-γGlu-nonadecanoic acid)-Glu-Phe-Val-Glu-Trp-Leu-Leu-Glu-Gly-Gly-Pro-Ser-Ser-Gly-NH2
• Shorthand: H-{Aib}-QGTFTSDYSKYLDEKKAK-{AEEA-AEEA-γGlu-C19 diacid}-EFVEWLLEGGPSSG-NH2

Three structural features carry the pharmacology and, downstream, the analytics.

The Aib substitution at position 2. Alpha-aminoisobutyric acid replaces the native residue at the position that dipeptidyl peptidase-4 cleaves in GLP-1 and glucagon-family peptides. This is the same protease-resistance strategy used in semaglutide and tirzepatide, and it is a non-proteinogenic residue — which means it will not appear in a standard proteomic database search and has to be accounted for deliberately in any identity determination.

The C20 diacid at Lys20. The fatty acid moiety, tethered through the γGlu and two AEEA (8-amino-3,6-dioxaoctanoic acid) spacer units, drives reversible albumin binding and is the basis for the extended circulating half-life. It also adds roughly 700 daltons of non-peptidic mass and a large hydrophobic domain that dominates reversed-phase retention behaviour — a point that matters in Section 6.

The C-terminal amide. The molecule terminates in a primary amide rather than a free carboxylate. That is a one-dalton difference from the free-acid form, and free-acid impurity is a recognised failure mode in amidated peptide synthesis.

Handling follows standard practice for a lipidated peptide of this size. Lyophilized material is typically stored sealed and desiccated at −20 °C or below, protected from light, with vendor stability data commonly citing two years at −80 °C and one year at −20 °C for powder. Aqueous solubility has been reported at approximately 22 mg/mL with sonication and pH adjustment; the acetate salt form is the common commercial presentation, so gross vial weight and net peptide content are distinct figures.

Section 2 — Mechanism

All activity described here is at the receptor and pathway level, in defined experimental systems.

The oxyntomodulin premise. Proglucagon is processed differently in different tissues. In intestinal L-cells, one of the products is oxyntomodulin — glucagon’s 29 residues plus an eight-residue C-terminal extension. That extension changes the receptor selectivity profile: native OXM is a low-potency agonist at both the glucagon receptor (GCGR) and the GLP-1 receptor (GLP-1R), where glucagon itself is strongly GCGR-selective and GLP-1 is strongly GLP-1R-selective. OXM is also cleared within minutes by DPP-4 and neprilysin, which is why the native peptide was never a viable pharmacological tool. Mazdutide is an attempt to keep the dual-receptor profile while engineering out the clearance problem: Aib2 for protease resistance, lipidation for albumin binding.

Binding profile. Published binding characterisation reports mazdutide interacting with both receptors at comparable affinity, with a modest bias toward GCGR. Reported inhibition constants are approximately 17.7 nM at the human glucagon receptor and 15.9 nM at the mouse glucagon receptor, and approximately 28.6 nM at the human GLP-1 receptor and 25.1 nM at the mouse GLP-1 receptor. The near-identity of the human and mouse figures is methodologically useful: it means rodent in-vitro systems are a reasonable proxy for the human receptor in this specific case, which is not true of every peptide in this class. Functional readout in isolated mouse islets has been reported with a half-maximal effective concentration of approximately 5.2 nM.

GLP-1R arm. GLP-1R is a class B secretin-family GPCR that couples principally to Gs. Agonist engagement activates adenylyl cyclase, raises intracellular cAMP, and activates protein kinase A and the Epac2 pathway. In pancreatic beta-cell systems this cascade is the canonical mechanism underlying glucose-dependent insulin secretion; in hypothalamic and hindbrain neuronal populations expressing GLP-1R, the same receptor has been studied for its role in central signalling relating to energy balance. Receptor-level β-arrestin recruitment and internalisation kinetics differ between GLP-1R agonists and are an active area of structural pharmacology.

GCGR arm. GCGR is likewise Gs-coupled and is expressed most densely on hepatocytes. cAMP/PKA signalling downstream of hepatic GCGR has been extensively characterised independently of any particular agonist: it drives CREB-dependent transcription, upregulates gluconeogenic and fatty-acid-oxidation programmes, and in preclinical models is associated with increased hepatic β-oxidation flux and altered lipid handling. The mechanistic rationale for co-agonism is that the GCGR arm contributes an energy-expenditure and hepatic-lipid component that pure GLP-1R agonism does not, while the GLP-1R arm is understood to offset the glycaemic consequence of unopposed glucagon receptor activation.

Throughout, these are pathways mazdutide has been studied for — GLP-1R and GCGR binding and Gs/cAMP/PKA pathway activation, islet insulin-secretion readouts, and hepatic lipid-handling programmes in preclinical models — not demonstrated clinical effects, and nothing here should be read as a claim about outcomes in humans.

Section 3 — Preclinical and In-Vitro Research Data

Receptor and islet work. The foundational in-vitro characterisation reports dual binding at human and mouse GCGR and GLP-1R at the affinities given above, with functional insulin-secretion activity in isolated mouse islet preparations. Chen and colleagues (2022, Journal of Medicinal Chemistry) situated this compound within a broader review of monomeric multitarget peptide design — useful context for understanding why the oxyntomodulin scaffold was chosen over a GLP-1 scaffold with glucagon activity grafted on.

Rodent models. Preclinical reporting in diet-induced obese mice, presented by Chen, Mezo, Coskun and colleagues at the American Diabetes Association Scientific Sessions in 2021, described changes in glucose control, body-composition endpoints, and energy expenditure following single subcutaneous administration. These are rodent findings in a controlled dietary model and should be read strictly as such.

Dual-agonism background literature. The mechanistic case for GLP-1R/GCGR co-agonism does not rest on mazdutide alone, and the background papers are the more informative reading. Pocai and colleagues (2009, Diabetes) reported that a PEGylated oxyntomodulin-derived co-agonist produced its metabolic phenotype in mice only when both receptors were engaged, establishing the two-receptor dependence genetically rather than by inference. Zimmermann and colleagues (2022, Molecular Metabolism) published the discovery and preclinical pharmacology of BI 456906, a separate GCGR/GLP-1R dual agonist, with detailed in-vitro potency characterisation and hepatic fatty-acid-oxidation data in DIO mice. A 2024 Nature Communications report on the oxyntomodulin analog G49 described a GCGR-dependent inter-organ sequence beginning with free fatty acid release from white adipose tissue and proceeding through adiponectin and FGF21 elevation to adipose beiging and increased energy expenditure. Taken together, this is the strongest published mechanistic scaffolding for what dual agonism does at the pathway level, and none of it is mazdutide-specific.

Multi-omics neurobiology. Dong, Bai, Yuan and colleagues (2025, eBioMedicine) reported a comparative study of mazdutide against dulaglutide in db/db mice, with cognitive-performance testing, histopathological assessment of neuronal structure, and integrated transcriptomic and metabolomic analysis. The reported molecular pathways clustered around neuroprotection, energy metabolism, and synaptic plasticity. This is the most detailed mechanism-oriented mazdutide paper currently in the literature and it is a mouse study.

The clinical record, and its boundary. A substantial clinical literature exists — phase 1b studies in Chinese adults with overweight or obesity (2021 and 2022, eClinicalMedicine), a phase 1b study in type 2 diabetes (2022, Nature Communications), a phase 2 trial (2024, Diabetes Care), and the phase 3 GLORY-1 study (2025, New England Journal of Medicine). Mazdutide injection was approved for marketing in China by the NMPA in June 2025. Those publications are cited below because they exist and because a researcher should know they exist. No efficacy figures from them are reproduced here, no therapeutic conclusion is drawn, and none of that record transfers to research-grade lyophilized material, which is a different product supplied under a different regulatory status for a different purpose.

Two limitations deserve plain statement. First, the mazdutide-specific in-vitro literature is thin relative to the clinical literature — the receptor characterisation rests largely on binding and islet data reported in supporting material rather than in a dedicated pharmacology paper, and much of the pathway reasoning is borrowed from the broader co-agonist field. Second, the entire clinical programme has been conducted in predominantly Chinese populations, which is a stated limitation in the meta-analytic literature and is worth carrying forward rather than eliding.

Section 4 — Published Literature

Real, published papers anchoring the mazdutide record. Researchers should consult primary sources directly.

• Ji L, Jiang H, An P, et al. (2021). “IBI362 (LY3305677), a weekly-dose GLP-1 and glucagon receptor dual agonist, in Chinese adults with overweight or obesity: a randomised, placebo-controlled, multiple ascending dose phase 1b study.” eClinicalMedicine 39:101088. PMID 34430840. PMC8374649. DOI 10.1016/j.eclinm.2021.101088.
• Jiang H, Pang S, Zhang Y, et al. (2022). “A phase 1b randomised controlled trial of a glucagon-like peptide-1 and glucagon receptor dual agonist IBI362 (LY3305677) in Chinese patients with type 2 diabetes.” Nature Communications 13(1):3613. PMID 35750681. PMC9232612. DOI 10.1038/s41467-022-31328-x.
• Ji L, Gao L, Jiang H, et al. (2022). “Safety and efficacy of a GLP-1 and glucagon receptor dual agonist mazdutide (IBI362) 9 mg and 10 mg in Chinese adults with overweight or obesity: a randomised, placebo-controlled, multiple-ascending-dose phase 1b trial.” eClinicalMedicine 54:101691. PMID 36247927. PMC9561728.
• Dong W, Bai J, Yuan Q, et al. (2025). “Mazdutide, a dual agonist targeting GLP-1R and GCGR, mitigates diabetes-associated cognitive dysfunction: mechanistic insights from multi-omics analysis.” eBioMedicine 117:105791. PMID 40479843. DOI 10.1016/j.ebiom.2025.105791. The db/db mouse multi-omics study.
• Ji L, Jiang H, Bi Y, et al. (2025). “Once-Weekly Mazdutide in Chinese Adults with Obesity or Overweight.” New England Journal of Medicine 392(22):2215–2225. DOI 10.1056/NEJMoa2411528. The phase 3 GLORY-1 report.
• Pocai A, Carrington PE, Adams JR, et al. (2009). “Glucagon-like peptide 1/glucagon receptor dual agonism reverses obesity in mice.” Diabetes 58(10):2258–2266. PMID 19602537. The genetic demonstration that both receptors are required.
• Zimmermann T, Thomas L, Baader-Pagler T, et al. (2022). “BI 456906: discovery and preclinical pharmacology of a novel GCGR/GLP-1R dual agonist with robust anti-obesity efficacy.” Molecular Metabolism 66:101633. PMC9679702. Detailed in-vitro potency and hepatic FAO characterisation for a comparator dual agonist.
• Chen T, et al. (2022). “The design and optimization of monomeric multitarget peptides for the treatment of multifactorial diseases.” Journal of Medicinal Chemistry 65(5):3685–3705. Design-space context for multitarget peptide scaffolds including this one.

Background reading on the receptors themselves — the class B GPCR structural literature on GLP-1R and GCGR, and the oxyntomodulin physiology reviews — provides the assay context that the mazdutide papers assume rather than restate.

Section 5 — Research Applications

In laboratory settings this compound appears principally in comparative incretin-receptor pharmacology. Reported in-vitro applications include competitive radioligand or fluorescence-polarisation binding assays against human and mouse GLP-1R and GCGR in transfected HEK293 or CHO systems; cAMP accumulation assays (HTRF or luminescent reporter) to derive functional EC50 at each receptor separately; β-arrestin recruitment and receptor internalisation assays to characterise signalling bias relative to GLP-1R-selective comparators; glucose-stimulated insulin secretion assays in isolated islets or INS-1/MIN6 beta-cell lines; hepatocyte assays for CREB phosphorylation, gluconeogenic gene expression, and fatty-acid-oxidation flux as GCGR-arm readouts; and albumin-binding and plasma-stability assays probing the contribution of the C20 diacid.

Assay-design variables worth attention. Because both arms converge on cAMP, running the two receptors in the same cell background is the only way to make the potency ratio meaningful — a GLP-1R EC50 from one cell line and a GCGR EC50 from another are not comparable numbers, and the co-agonist literature contains a good deal of this error. Because the lipidation drives albumin binding, serum or BSA content in the assay buffer materially shifts apparent potency by changing free compound concentration; free-fraction conditions must be stated, not assumed. Because the Aib2 residue exists specifically to resist DPP-4, any stability comparison against native oxyntomodulin or GLP-1 needs the protease source and activity specified. And because human and mouse receptor affinities are close for this compound but not for others in the class, cross-compound comparisons in rodent systems should not be generalised without checking the comparator’s own species profile.

Section 6 — How to Evaluate a Source

Mazdutide sits in the most heavily counterfeited segment of the research peptide market. The compounds adjacent to it on vendor menus — semaglutide, tirzepatide, retatrutide, survodutide, cagrilintide — are the same segment, and substitution between them is the documented failure mode. The analytical questions here are specific and answerable.

The lipidation is the identity question. The peptide backbone of mazdutide is 33 residues. The C20 diacid plus the γGlu and two AEEA linker units account for roughly 700 daltons of the 4,563 Da total. A synthesis that produces the correct backbone but fails, incompletely couples, or partially cleaves the lipid side chain yields material that is chromatographically distinguishable and mass-distinguishable — but only if someone looks. Des-lipidated peptide, single-AEEA truncation of the linker, and free-acid rather than amidated C-terminus are all realistic products of an imperfect synthesis, and none of them are visible in a purity percentage stated without a chromatogram.

Mass alone does not resolve it at this size. At 4,563 Da the molecule ionises as a multiply-charged envelope, and deconvoluted mass accuracy on a routine instrument is often only good to a few daltons. A one-dalton amidation difference is not reliably resolved by a low-resolution deconvoluted mass. What resolves the question is high-resolution LC-MS with reported charge states, or MS/MS fragmentation confirming the Lys20 modification site specifically, or comparison against a characterised reference standard on the same method.

Reversed-phase behaviour is dominated by the fatty acid. The hydrophobic C20 chain governs retention, which means backbone-related impurities — deletion sequences, D-amino acid epimers, Aib omission — can co-elute near the main peak on a poorly developed gradient. A shallow, well-resolved gradient with a stated column, mobile phase, and run time is the difference between a chromatogram that means something and a picture of one peak.

What the analytical package needs to show:

• Purity by HPLC, with a stated percentage and the trace. For a lipidated 33-mer the relevant impurity classes are deletion and truncation sequences, des-lipidated peptide, incomplete linker coupling, free-acid C-terminus, and aggregate. A bare “≥99%” identifies none of them.
• Identity by LC-MS, observed mass reported against approximately 4,563.1 Da for the free base, with sufficient method detail — ideally including fragmentation evidence for the site of lipidation — to establish that the modification was actually confirmed rather than assumed.
• Endotoxin by USP <85> LAL, with an actual figure, not a “meets specification” checkbox.
• Heavy metals by ICP-MS.
• Net peptide content and counterion, disclosed. Acetate and trifluoroacetate carry different mass burdens, and residual TFA is biologically active in some cell assays at concentrations that survive lyophilization — a real confound in cAMP and insulin-secretion work.
• Lot-specificity. The COA must correspond to the exact batch shipped, be dated, and name the accredited laboratory that performed the testing. A COA reused across lots is a marketing document.

Market pricing context: research-grade mazdutide at the 10 mg scale is commonly listed in roughly the $45–$90 range per vial at single-vial pricing, with multi-vial boxes bringing that down toward roughly $45–$55 per vial, and 20 mg presentations scaling accordingly. That spread is narrow enough that price carries almost no information about analytical quality. A 33-residue peptide with a non-proteinogenic residue, a C-terminal amide, and a site-specific lipidation through a two-unit PEG linker is a demanding synthesis — meaningfully harder than a short unmodified peptide — and a low price on a difficult synthesis is a question, not a bargain. The documentation is the argument.

Regulatory context worth knowing. Mazdutide is an approved product in China as of June 2025 and remains investigational elsewhere, which places it in the same category as the other novel incretin analogs in the research market: it is not on the FDA’s 503A bulk drug substances list, it was not among the seven peptides reviewed at the July 2026 Pharmacy Compounding Advisory Committee meeting, and it is not among the five scheduled for the PCAC review due before the end of February 2027. Approval in one jurisdiction does not change the status of material supplied for laboratory work, and it does not relax documentation or labelling discipline.

Section 7 — The PYXAX Verification Standard

PYXAX uses accredited independent laboratories in its verification network, including ILS Labs, Krause Analytical, and Janoshik. The current per-batch panel covers purity by HPLC, potency against label claim, and identity by LC-MS. A lot-specific COA is published in the COA Library before dispatch and names the laboratory that tested that batch. Endotoxin, heavy-metals, and final-vial sterility screening are not part of the current panel.

On a compound whose entire pharmacological identity depends on a fatty acid attached to one specific lysine in a 33-residue chain, the identity determination is not a formality appended to a purity number. It is the measurement that establishes whether the molecule in the vial is the dual agonist described in the literature or a peptide that merely resembles it. A published mass spectrum and chromatogram on the actual lot settles in one document what no amount of label copy can.

Founding batches are documented end to end, with full analytical data files available per lot. You can review current documentation in the COA library, read the underlying methodology on the verification standard page, see the mazdutide listing, or browse the full catalog in the shop.

All PYXAX compounds are supplied strictly for in-vitro and preclinical laboratory research use only. Not for human consumption. Not for veterinary use. Not for diagnostic procedures. These statements have not been evaluated by the FDA. Researchers are responsible for compliance with all applicable laws and regulations governing the use of research compounds in their jurisdiction.