Eloralintide

How the molecule is built

Any amylin analogue has to solve two problems that native amylin does not: it aggregates into insoluble fibrils, and it is cleared within minutes. Pramlintide solved the first with proline substitutions borrowed from a non-fibrillating rodent sequence and never solved the second, which is why it is a mealtime injection. Eloralintide solves both, and the specific choices are unusual enough to be worth naming individually.

• Backbone: 37 residues, the same length as human amylin and as cagrilintide, with three non-coded amino acids at positions 11, 15 and 22. Non-coded residues are the standard lever for blocking proteolysis and for holding a conformation the natural sequence will not hold on its own. Position 22 sits at the leading edge of the 20–29 stretch that drives fibrillation in native human amylin — the region every analogue in this family has had to intervene in.
• C-terminus: amidated, as native amylin is. The free acid is substantially less active at the amylin receptors, so this is fidelity to the hormone rather than a modification of it.
• Cys2–Cys7: closed with a methylene thioacetal bridge instead of a disulfide. A disulfide is reducible; a thioacetal linkage is not redox-labile, so the N-terminal ring stays closed in environments that would open a conventional one. This is the single most distinctive piece of chemistry in the molecule, and it is a durability feature rather than a potency one.
• Lys26: acylated with a saturated linear C20 diacid through two γ-glutamic acid residues. The diacid binds albumin reversibly, shielding the peptide from renal filtration and enzymatic breakdown and releasing it slowly back into circulation. This is what supports once-weekly dosing.
• The comparison worth holding: semaglutide puts a C18 diacid on Lys26 through a γ-Glu/OEG spacer; cagrilintide uses a C20 diacid but attaches it at the N-terminus through a γ-glutamic acid linker; eloralintide uses a C20 diacid at a lysine side chain through two γ-Glu units. The albumin-binding strategy is shared across all three. The attachment chemistry is different in each, and it is the attachment point that determines which part of the molecule stays free to engage the receptor.
• The two sulfurs in the formula are Cys2 and Cys7 and nothing else — consistent with a single N-terminal ring and no free cysteine elsewhere in the chain.

Mechanism in detail

There is no standalone amylin receptor gene. Amylin receptors are heteromers: the calcitonin receptor complexed with one of three receptor activity-modifying proteins, giving AMY1 (RAMP1), AMY2 (RAMP2) and AMY3 (RAMP3). The RAMP is what converts a calcitonin receptor into an amylin receptor, which is exactly why selectivity between the two is a real medicinal-chemistry problem rather than a formality. Signalling is Gs-coupled and cAMP-driven, as it is at the GLP-1 receptor, but these are different genes with different tissue distributions and there is no cross-activation between the families.

• Area postrema: the primary central target. The area postrema is a circumventricular organ without a complete blood–brain barrier, so a circulating amylin analogue reaches it directly. Activation propagates to the nucleus tractus solitarius and the lateral parabrachial nucleus and terminates the meal in progress. The measurable effect is on meal size rather than on how often a meal is initiated.
• Stomach: gastric emptying slows, in the same direction as GLP-1 agonism but through a separate afferent route. This contributes to post-prandial glucose flattening and is part of the early gastrointestinal signal.
• Pancreatic α-cell: post-prandial glucagon secretion is suppressed, in a glucose-context-dependent rather than absolute way. This is the arm the first-generation analogue was approved on.
• Leptin signalling: amylin agonism restores leptin responsiveness in diet-induced obesity models. Few interventions do, and it is the mechanistic basis for describing amylin as acting on the defended body-weight set point rather than only on appetite in the moment.
• Energy expenditure: no documented arm. Amylin agonism is an intake-side mechanism, and reading it as a metabolic-rate intervention — the way retatrutide's glucagon component can be read — is a category error.
• Calcitonin receptor: the arm eloralintide is designed to leave out. Calcitonin receptors are expressed centrally and on osteoclasts, and in a non-selective agonist the central component plausibly adds to satiation while the skeletal component is the part with the least long-term data behind it. Selectivity is the attempt to keep the first without the second.

What selectivity buys — and what one trial can settle about it

Cagrilintide is non-selective by design: it agonises AMY1–3 and the bare calcitonin receptor, and the calcitonin arm is not thought to be inert. Eloralintide is the opposite design choice, and the 48-week monotherapy result is the first substantial evidence that a selective amylin agonist can carry a weight endpoint on its own. Roughly 20% mean reduction at 48 weeks on 9 mg is a magnitude that sits in dual-agonist territory rather than in the modest range the amylin class was once assumed to occupy.

What that does not settle is whether selectivity is the reason. Nobody has run the head-to-head. The two compounds have been studied in different trials, over different durations — 26 weeks for cagrilintide monotherapy against 48 for eloralintide — in differently selected populations, and a longer trial in this class flatters a compound because these curves are still descending at six months. The honest reading is that eloralintide has a strong 48-week monotherapy result and that the selectivity hypothesis remains a hypothesis with a good result attached to it.

One comparison is at least arithmetically clean. Eloralintide is about 4526 Da and cagrilintide about 4409 Da — within roughly 3% of each other, both 37-residue acylated amylin analogues. A milligram of one is very nearly the same molar quantity as a milligram of the other, so the dose ranges can be compared directly on a mass basis: cagrilintide was studied at 0.3–4.5 mg weekly, eloralintide at 1–9 mg. That is a genuine difference in the exposure the two programmes went after, not an artefact of unit conversion. It says nothing about relative potency at the receptor, which is a separate measurement.

What the phase 2 actually reported

• Design: multicentre, double-blind, randomised, placebo-controlled, 48 weeks, eloralintide as monotherapy. Registered as NCT06230523.
• Population: 263 adults with obesity, or overweight with at least one weight-related comorbidity. Subjects with type 2 diabetes were excluded, so this is a weight-endpoint trial in a non-diabetic population and nothing in it speaks to glycaemic management.
• Arms: fixed once-weekly subcutaneous doses of 1, 3, 6 and 9 mg, plus two escalation arms (3 → 6 → 9 mg, and 6 → 9 mg) that titrated up to 9 mg to test whether slower escalation improved tolerability.
• Primary outcome: every arm beat placebo. Mean weight reduction of 9.5% to 20.1% against 0.4% on placebo, increasing with dose — about 9 to 9.5% at 1 mg and about 20% at 9 mg.
• Safety: the commonest adverse events were mild-to-moderate gastrointestinal symptoms and fatigue, more frequent at higher doses. Incidence was lower with slower escalation, and in the 1 mg and 3 mg arms it was similar to placebo.
• A separate discovery-to-clinical-proof-of-concept paper covers the medicinal chemistry and the first-in-human work behind the compound.
• What is not in the record: no phase 3 readout, no approval anywhere, and no published monotherapy data beyond 48 weeks or above 9 mg weekly.

Reading the dose ladder

The escalation arms are the most useful part of this trial for anyone designing a protocol, because they answer a question the fixed arms cannot. Fixed arms tell you what a dose achieves; the escalation arms tell you that the route to the dose changes the cost of getting there. Reaching 9 mg by stepping up from a low start produced a lower incidence of the dose-related adverse events than the trial saw at the top of the range otherwise, which is the same tolerability logic the incretin programmes are built on and here it has direct evidence behind it rather than analogy.

The low arms deserve more attention than they usually get. 1 mg weekly is not a tolerance step in the way tirzepatide's 2.5 mg is — it produced roughly 9 to 9.5% mean weight reduction over 48 weeks with an adverse-event incidence similar to placebo. That is a real result at a dose that costs almost nothing in tolerability, and a protocol that treats 1 mg purely as a rung to be climbed past is discarding the arm with the best effect-to-burden ratio in the trial.

9 mg weekly is the ceiling with published data. Nothing above it has been reported at any duration, and 48 weeks is the longest observation window that exists for this molecule.

Supply becomes the binding constraint at the top of the ladder, and it is a hard constraint rather than an inconvenience. A single 5 mg vial cannot produce one 6 mg dose or one 9 mg dose — there is not enough peptide in the vial. Anyone planning either of the top two arms needs two vials in hand per dose, reconstituted to the same concentration, with the draw split across both. The arithmetic is worked through in the reconstitution section below.

Eloralintide and cagrilintide, side by side

This catalogue stocks both, both act on the amylin pathway, and the comparison gets asked constantly. The differences are specific rather than a matter of maturity.

• Receptor profile: eloralintide is selective for the amylin receptors; cagrilintide is non-selective across AMY1–3 and also agonises the bare calcitonin receptor. This is the substantive difference between them.
• Chemistry: both are 37-residue acylated amylin analogues of near-identical mass (~4526 versus ~4409 Da) carrying a C20 diacid. Eloralintide attaches it at Lys26 through two γ-Glu residues and closes its N-terminal ring with a methylene thioacetal; cagrilintide attaches its diacid at the N-terminus through a γ-glutamic acid linker.
• Monotherapy evidence: eloralintide has a 48-week phase 2 at 1–9 mg weekly with 9.5–20.1% mean weight reduction. Cagrilintide has a 26-week phase 2 at 0.3–4.5 mg weekly. Different durations and different populations — the numbers are not directly comparable even though the units are.
• Late-stage position: cagrilintide's phase 3 programme is a fixed-dose combination with semaglutide rather than a monotherapy, so its most mature evidence is about the pairing. Eloralintide's published evidence is monotherapy.
• The open question is the same one the field has been asking since the first selective agonist entered the clinic: whether amylin-receptor selectivity is better than dual amylin/calcitonin agonism. Selectivity may avoid calcitonin-arm effects nobody wants; the calcitonin arm may also be contributing to the effect. Neither compound has been tested against the other.