Cagrilintide 5mg Peptide

Cagrilintide is a 37 residue analogue of human amylin carrying a C20 fatty diacid on the N-terminal lysine. Novo Nordisk developed it under the codes AM833 and NNC0174-0833, and Kruse and colleagues published the design work in the Journal of Medicinal Chemistry in 2021, where the molecule appears as analogue 23. It is the amylin half of the fixed combination with semaglutide that the sponsor calls CagriSema.

Amylin is co-secreted with insulin from pancreatic beta cells and acts on receptors in the area postrema and other circumventricular structures to end a meal. It is a poor drug on its own for a specific chemical reason: human amylin aggregates into amyloid fibrils, and the same beta-sheet propensity that builds islet amyloid in type 2 diabetes destroys a formulated solution. The only marketed amylin analogue before cagrilintide, pramlintide, solved aggregation but not clearance, and clears with a half-life of roughly 20 to 45 minutes.

Cagrilintide addresses both problems in one molecule. Six substitutions relative to human amylin suppress fibril formation and stabilise the central helix, and the acylation drives reversible albumin binding that pushed the reported human half-life to 159 to 195 hours. It is not selective for amylin receptors: cryo-electron microscopy structures published in 2025 show it engaging AMY1R, AMY2R, AMY3R and the bare calcitonin receptor, and that lack of selectivity is a deliberate design feature rather than an impurity of the pharmacology.

This 10mg vial is supplied as a lyophilised powder for in vitro and preclinical laboratory research. It is not the clinical fixed-dose formulation, and nothing on this page is guidance for use in a person.

Amylin receptors are not standalone proteins. They are heterodimers of the class B1 G protein-coupled calcitonin receptor (CTR) with one of three receptor activity-modifying proteins: CTR plus RAMP1 gives AMY1R, CTR plus RAMP2 gives AMY2R, and CTR plus RAMP3 gives AMY3R. CTR also reaches the cell surface without any RAMP, where it responds potently to calcitonin peptides and only weakly to amylin. Cagrilintide is an agonist at all four, and the 2025 cryo-electron microscopy work from Cao and colleagues resolved Gs-coupled complexes of cagrilintide with each one at 2.2 A for CTR, 2.2 A for AMY1R, 2.7 A for AMY2R and 3.0 A for AMY3R.

The structural basis for that breadth sits at the C-terminus. In rat amylin, the amidated Tyr37 contacts the RAMP extracellular domain, which is part of why amylin needs a RAMP to bind well. Cagrilintide replaces it with proline, mirroring Pro32 of salmon calcitonin. In the cagrilintide-bound AMY1R and AMY3R structures the small Pro37 side chain is buried entirely by calcitonin receptor extracellular domain residues, packing against Trp79 with no direct contact to RAMP1 or RAMP3 at all. That single change is why cagrilintide forms high-affinity interactions with the CTR protomer on its own.

The N-terminal region does the receptor activation. Cagrilintide forms the same bypass motif that gives amylin analogues their potency at AMY receptors, anchored by van der Waals packing between Phe23 of the peptide and Tyr146 and Tyr149 in transmembrane helix 1 of the receptor. The paired N14E and V17R substitutions form an intramolecular salt bridge across one face of the helix spanning residues 5 to 18, seen in a crystal structure of the non-lipidated backbone fused to maltose binding protein (PDB 7BG0), and that helix stabilisation is what suppresses the beta-sheet conversion that drives fibrillation.

Lipidation is not pharmacologically silent. Comparing cagrilintide with its non-lipidated backbone in a cAMP response element reporter assay, the two had similar potency and maximal response in cells co-expressing CTR and RAMP3, but the non-lipidated peptide lost potency at CTR expressed alone. The authors attribute this to the gamma-glutamyl linker forcing a different rotamer of His296 in extracellular loop 2 and a more open extracellular loop 3, so the fatty acid arm contributes to CTR engagement as well as to albumin binding and slow clearance.

Human amylin has one dominant liability as a drug candidate, and it is not potency. The peptide converts readily to beta-sheet-rich amyloid fibrils, the same process that deposits islet amyloid in type 2 diabetes pancreas, and a peptide that fibrillates in the vial is not a formulation. Kruse and colleagues opened their 2021 development paper by naming this as the central design constraint, and the six substitutions in cagrilintide split cleanly into two functional groups aimed at it.

The first group is A25P, S28P and S29P. These are the three proline substitutions that already exist in rat amylin, which does not fibrillate, and they are the same three that define pramlintide. Proline is a helix and sheet breaker; putting three of them into the C-terminal half of the peptide removes the segment most prone to beta-sheet stacking. The second group is the paired N14E and V17R, which are not in pramlintide. Positioned three residues apart on the same helical face, the glutamate and arginine side chains form an intramolecular salt bridge that stabilises the alpha-helix spanning residues 5 to 18. A crystal structure of the non-lipidated cagrilintide backbone fused to maltose binding protein (PDB 7BG0) shows the bridge, and every active-state receptor structure solved in 2025 shows it preserved with the two side-chain rotamers constrained.

The sixth substitution, Y37P, does something different. It is a potency and selectivity change rather than a stability one, and it is discussed in the receptor pharmacology section below. Separately from all six, the C20 alpha,omega-diacid attached through a gamma-glutamyl spacer to the alpha-amino group of Lys1 is the pharmacokinetic element: it is the same acylation logic used in semaglutide, exploiting reversible albumin binding to slow renal filtration and proteolysis.

The result is a molecule with a substantially different physical character from native amylin. Where the parent hormone aggregates, cagrilintide holds a stable helix. Where the parent hormone clears in minutes, cagrilintide showed a half-life of 159 to 195 hours in humans across a 28-fold dose range, with exposure proportional to dose and a median time to maximum concentration of 24 to 72 hours.

The amylin receptor family is unusual. There is no dedicated amylin receptor gene. Instead the calcitonin receptor, a class B1 GPCR, partners with receptor activity-modifying proteins to change its own pharmacology: with RAMP1 it becomes AMY1R, with RAMP2 AMY2R, with RAMP3 AMY3R, and without any RAMP it stays a calcitonin receptor that responds well to calcitonin peptides and poorly to amylin. A ligand's profile across those four states is therefore a design choice, not a fixed property of the target.

For years the working assumption in the field was that the amylin receptor phenotypes carried the weight effects, since a calcitonin-receptor-selective agonist does not control appetite or body weight in animals. The awkward observation is that dual CTR and AMY receptor agonists, salmon calcitonin among them, report higher efficacy than AMY-selective agonists. Cagrilintide was taken forward as a dual agonist on that reasoning.

In 2025 Cao and colleagues at Monash, working with Novo Nordisk, determined Gs-coupled cryo-electron microscopy structures of cagrilintide at all four receptor states: 2.2 A for CTR, 2.2 A for AMY1R, 2.7 A for AMY2R and 3.0 A for AMY3R. The receptors were co-expressed with a dominant-negative Gs and stabilised with nanobody 35. The comparison with rat amylin and salmon calcitonin complexes localises the non-selectivity to a single residue. Rat amylin's amidated Tyr37 reaches into the RAMP extracellular domain and forms the only direct peptide contact with the RAMP in previously reported complexes. Cagrilintide's Pro37 does not: it is buried by calcitonin receptor extracellular domain residues, forming extensive hydrophobic contacts with Trp79 and making no direct interaction with RAMP1 in AMY1R or RAMP3 in AMY3R. That is the structural reason cagrilintide can bind the CTR protomer alone with high affinity, and it is exactly the behaviour of Pro32 in salmon calcitonin.

The paper's own summary of the result is that cagrilintide adopts an amylin-like binding mode but induces distinct conformational dynamics at calcitonin-family receptors relative to rat amylin, salmon calcitonin and newer amylin-based peptides. For in vitro receptor work this is the practically useful point: a cagrilintide dose-response curve run only in an AMY3R line describes a fraction of the molecule's pharmacology, and comparing it against a selective amylin agonist requires including a RAMP-free CTR arm to see the difference at all.

The human dataset arrived in a specific order and each stage answered a different question. The phase 1b trial (NCT03600480) ran six sequential overlapping cohorts at a single US site between July 2018 and December 2019, randomising 96 participants with a body-mass index of 27.0 to 39.9 kg/m2 in a 3:1 ratio to cagrilintide at 0.16, 0.30, 0.60, 1.2, 2.4 or 4.5 mg once weekly or matched placebo, all on a background of semaglutide 2.4 mg once weekly. Its primary endpoint was treatment-emergent adverse events, and its most cited output is pharmacokinetic: cagrilintide exposure was dose-proportional, area under the curve over 168 hours ran from 926 to 24271 nmol x h/L, and cagrilintide did not alter semaglutide exposure or elimination.

The phase 2 dose-finding trial (NCT03856047) was the monotherapy readout: 706 participants on cagrilintide, 99 on liraglutide 3.0 mg and 101 on placebo, 26 weeks of treatment with an escalation period of up to six weeks, run across ten countries. Weight reduction was 6.0% to 10.8% on cagrilintide against 3.0% on placebo, with the 4.5 mg group beating daily liraglutide. Gastrointestinal events were the dominant tolerability signal, reported by 41% to 63% of cagrilintide participants against 32% on placebo, mostly nausea (20% to 47% versus 18%).

The 32 week phase 2 trial in type 2 diabetes (NCT04982575) is the only randomised comparison of cagrilintide alone, semaglutide alone and the combination in the same protocol, at 92 participants. It produced the cleanest separation of the two mechanisms in humans: on weight, cagrilintide alone outperformed semaglutide alone (-8.1% versus -5.1%); on HbA1c, semaglutide alone outperformed cagrilintide alone (-1.8 versus -0.9 percentage points). The combination led on both.

The phase 3a REDEFINE programme followed. REDEFINE 1 (NCT05567796) enrolled 3417 adults without diabetes for 68 weeks in a 21:3:3:7 randomisation, and REDEFINE 2 (NCT05394519) enrolled 1206 adults with type 2 diabetes and an HbA1c of 7% to 10% for the same duration. Both were published in the New England Journal of Medicine in 2025. REDEFINE 2 reported -13.7% body weight on the combination against -3.4% on placebo, with 73.5% of the combination group reaching an HbA1c at or below 6.5% against 15.9% on placebo. Gastrointestinal adverse events dominated the safety profile in both trials, affecting 79.6% of the REDEFINE 1 combination group and 72.5% of the REDEFINE 2 combination group.

The programme has also produced a negative result worth stating plainly, because it is routinely omitted. REDEFINE 4 compared the fixed combination against tirzepatide 15 mg once weekly over 84 weeks and did not meet its non-inferiority threshold for weight reduction: 23.0% versus 25.5% under the efficacy estimand, 20.2% versus 23.6% under the treatment-regimen estimand. The dedicated cardiovascular outcomes trial, REDEFINE 3, is event-driven in roughly 7000 adults with established cardiovascular disease and has not reported, so no statement about major adverse cardiovascular events can be made from this programme yet.

These compounds share a catalogue category and almost nothing else at the receptor. Semaglutide is a selective agonist at the GLP-1 receptor, a class B1 GPCR expressed in pancreatic beta cells, the hypothalamic arcuate nucleus and the nodose ganglion. Retatrutide is a triple agonist at the GIP, GLP-1 and glucagon receptors, all three in the same glucagon receptor subfamily. Cagrilintide targets the calcitonin receptor and its three RAMP heterodimers, a different subfamily of class B1 entirely, and reaches the brain principally through the area postrema rather than the arcuate.

The functional consequence shows up in the human data. In the head-to-head arms of the 32 week type 2 diabetes trial, cagrilintide alone produced more weight change than semaglutide alone (-8.1% versus -5.1%) and less HbA1c change (-0.9 versus -1.8 percentage points). Amylin agonism is therefore not a weaker version of incretin agonism; it moves body weight efficiently while moving glycaemia comparatively little, which is what you would expect from a satiation signal acting on meal termination rather than an insulinotropic hormone acting on beta cells.

Because the pathways are separate, their effects add rather than overlap. REDEFINE 1 measured all three arms in one protocol at 68 weeks: -11.5% on cagrilintide alone, -14.9% on semaglutide alone, -20.4% on the combination against -3.0% on placebo. A strictly additive model, subtracting the placebo effect once, would predict about 23.4%; the observed 20.4% falls short of that, so the two mechanisms combine sub-additively rather than synergistically. The combination still gained 5.5 percentage points over semaglutide alone, which is the number the programme was built on.

One asymmetry is worth noting for experimental design. Gastrointestinal adverse events were reported in 79.6% of the REDEFINE 1 combination group, and both classes drive them, so tolerability is not additive in the direction anyone wants. A phase 1 trial currently recruiting (NCT07607587) is specifically evaluating cagrilintide in participants who could not tolerate GLP-1 receptor agonists because of gastrointestinal adverse events, which is a direct test of whether the two mechanisms produce gastrointestinal effects through separable circuits.

Pramlintide is the closest structural relative and the comparison is precise: pramlintide is human amylin with A25P, S28P and S29P. Cagrilintide carries those same three prolines plus N14E, V17R and Y37P plus the Lys1 acylation. Three substitutions and one lipid chain separate a peptide that clears in roughly 20 to 45 minutes from one with a reported half-life of 159 to 195 hours, and separate a selective amylin receptor agonist from one that also engages the bare calcitonin receptor. Pramlintide was approved in the United States in 2005 as an adjunct to insulin, was never brought to the EU, Canada or Australia, and produces modest weight change in trials.

At the other end sits the selective amylin analogue class. Novo Nordisk's own NN1213, described by Dahl and colleagues in 2024, was optimised explicitly for selectivity as well as solubility, stability and clearance; in obese rats, daily administration over 21 days produced dose-dependent body-weight reduction that magnetic resonance imaging attributed primarily to fat mass loss. Eloralintide, from a separate programme and also stocked in this catalogue, belongs to the same selective lane.

The unresolved question the field is actually arguing about is whether calcitonin receptor co-agonism helps or is baggage. The structural paper states the tension directly: a calcitonin-receptor-selective agonist does not control appetite or body weight, which argues the amylin receptors carry the effect, yet dual CTR and AMY receptor agonists report higher efficacy than AMY-selective agonists, which argues co-activation contributes something. Because cagrilintide and a selective agonist differ at a small number of positions, the pair is a usable in vitro tool for that question: run both across CTR alone and CTR plus each RAMP and the selectivity difference is directly measurable in a cAMP readout.

Cagrilintide is more fibrillation-resistant than human amylin, not immune to it, and the difference matters in the laboratory. The entire substitution strategy in the molecule exists to raise the energy barrier to beta-sheet conversion, and that barrier is pH dependent for the same reason it is in the parent hormone: amylin-family peptides aggregate fastest near neutral pH, where the peptide sits closest to its isoelectric point and charge repulsion between molecules is weakest. Pharmaceutical formulation of this class accordingly sits in a mildly acidic window rather than at physiological pH.

In practice that makes the diluent choice a chemistry decision rather than a formality. A solution buffered near neutrality does not destroy the peptide immediately, but it puts it in the pH region where deamidation of the several asparagine and glutamine residues in the sequence proceeds fastest and where aggregation kinetics are least favourable. The sequence carries five asparagines, at positions 3, 21, 22, 31 and 35, plus a single glutamine at position 10. It contains no Asn-Gly, the fastest-deamidating motif in peptide chemistry, but it does carry Asn-Thr at 3 to 4 and again at 35 to 36, which sit one rank below Asn-Gly for succinimide formation and are the residues an accelerated stability study on this molecule should be watching.

Two failure modes are visible without instrumentation. Cloudiness or visible particulate in a solution that was clear indicates aggregation, and freeze-thaw cycling is the usual cause: ice-front cryoconcentration and the expanding ice-water interface both nucleate the beta-sheet conversion. Neither can be reversed. The lyophilised powder is far more stable than any solution, which is why the material ships lyophilised and why analytical work on stability should be dated from the point the powder is dissolved rather than from receipt.

One piece of received wisdom about handling this class is worth correcting, because it is repeated widely and it is wrong. Vigorous agitation is genuinely bad for an amylin analogue, but not because it breaks the disulfide bond. A carbon-sulfur or sulfur-sulfur covalent bond is not broken by shaking a vial; disulfides are cleaved by reducing agents, by thiol-disulfide exchange at alkaline pH, or by photolysis, none of which shaking supplies. What agitation actually does is generate air-water interface, and the hydrophobic face of an amphipathic peptide adsorbs to that interface, partially unfolds there and nucleates fibrils. The correct handling conclusion is the same, swirl rather than vortex, but the mechanism matters because it also explains why headspace, surfactant content and repeated pipetting change the outcome while bond chemistry has nothing to do with it.

There is one further consideration specific to acylated peptides. The C20 diacid gives cagrilintide surfactant-like behaviour, so it adsorbs to glass and to some plastics at low concentration. In binding assays and in dilution series run at nanomolar concentrations this is a real source of apparent potency loss, and it is the reason carrier protein is commonly included in assay buffers for lipidated peptides.

Identity for a molecule of this shape is established on mass first. The calculated average mass is 4409.01 g/mol for C194H312N54O59S2, with a monoisotopic mass of 4406.25 Da. On electrospray ionisation the peptide presents as a charge envelope rather than a single ion, typically in the 4+ to 6+ range, and deconvolution should return the intact mass within a few daltons.

Two specific mass results carry most of the diagnostic value, and both are consequences of how the molecule is built. A deconvoluted mass near 3955 Da rather than 4409 Da is the unacylated backbone: subtract the gamma-glutamyl spacer and the C20 diacid, together C25H43NO6, from the full formula and C169H269N53O53S2 at 3955.4 Da is what remains. A 453.6 Da deficit therefore means the acylation step failed, and since the acyl arm is the entire basis of the extended half-life, that material is pharmacologically a different compound. Separately, a result 2 Da high indicates a reduced open-chain species where the Cys2 to Cys7 disulfide has not closed, which is the other common structural defect in a synthesis of this class.

Purity is assessed by reversed-phase HPLC, and the acylation dominates the retention behaviour. The C20 diacid makes cagrilintide markedly more hydrophobic than an unmodified 37-mer of similar composition, so it elutes late in an acetonitrile gradient and the informative impurities are the ones that elute near it: the des-acyl backbone, which is far more polar and comes off substantially earlier, and truncation products from incomplete couplings, which often co-elute close to the main peak. Partial acylation is the realistic failure mode for this specific molecule, so the earlier-eluting shoulder is the peak worth integrating rather than ignoring.

Disulfide connectivity is not confirmed by mass alone and requires peptide mapping. Because the bridge is local, spanning only Cys2 to Cys7, an enzymatic digest under non-reducing conditions followed by LC-MS should show the intact bridged fragment, and a parallel reduced digest gives the assignment by mass shift. For work where aggregation state matters, thioflavin T fluorescence is the standard assay for beta-sheet amyloid content, and it is the assay the development paper's fibrillation work rests on; size exclusion chromatography gives the complementary readout of soluble oligomer content.