Larazotide

A synthetic eight-amino-acid peptide derived from a Vibrio cholerae toxin fragment, developed as an oral zonulin antagonist and tight-junction regulator for celiac disease, whose Phase 3 trial was stopped after a failed interim analysis in 2022.

Also referenced as: Larazotide acetate, AT-1001, INN-202

Also appears in: Immune

What is Larazotide?

Larazotide is a synthetic octapeptide (Gly-Gly-Val-Leu-Val-Gln-Pro-Gly) whose sequence was taken from a fragment of zonula occludens toxin, a protein secreted by Vibrio cholerae. It was designed as a locally acting, orally dosed antagonist of zonulin — the proposed human regulator of intestinal tight junctions — and became the most clinically advanced non-dietary drug candidate ever developed for celiac disease.

The program changed hands twice, which is why the compound carries several codes. Alba Therapeutics filed the original IND in 2005 and ran the early trials under the code AT-1001; the assets were licensed to Innovate Biopharmaceuticals in 2016, where it was designated INN-202; Innovate became 9 Meters Biopharma, which ran and then terminated the Phase 3 study.

It appears in the research-peptide market largely because of that clinical pedigree rather than because of a positive result. “Gut barrier” and “leaky gut” framing has made it a common ingredient in multi-peptide gut-health stacks, typically alongside other intestinal-repair compounds.

How it works

• Zonulin receptor antagonism (proposed). Larazotide was designed to competitively block zonulin signaling at the apical surface of enterocytes, preventing the tight-junction disassembly that follows gluten exposure in celiac disease (Slifer et al., American Journal of Physiology-Gastrointestinal and Liver Physiology, 2021;320(6):G983–G989). This is a design rationale rather than a confirmed molecular event: no zonulin receptor has been definitively identified, and larazotide’s direct binding target has never been established.
• Tight-junction assembly and actin rearrangement. In epithelial cell models, larazotide promoted tight-junction assembly and barrier function by promoting actin rearrangement and junctional distribution of ZO-1, occludin, claudins, and E-cadherin (Gopalakrishnan et al., Peptides, 2012;35(1):95–101). A companion paper from the same group describes the mirror-image result in a different context — larazotide inhibiting the ZO-1 and actin rearrangement provoked by gliadin and cytokines (Peptides, 2012;35(1):86–94) — so the two 2012 Peptides papers are easy to confuse when checking sources.
• Myosin light-chain kinase inhibition. A later mechanistic review attributed part of the effect to inhibition of myosin light-chain kinase, reducing tension on actin filaments and thereby facilitating junction closure (Slifer et al., 2021, above). Note that this review is the source for several mechanism claims that are often cited to the clinical trials instead.
• Barrier repair after injury. In ischemia-injured porcine jejunum, larazotide accelerated recovery of barrier function through tight-junction repair — the basis for interest beyond celiac disease (Slifer et al., PLoS One, 2021;16(4):e0250165). The effect was seen only at 1 µM, with neither 0.1 µM nor 10 µM working, echoing the narrow dose windows seen in the human trials.
• Cellular mechanism still unsettled. Papers published as recently as 2025 continue to describe larazotide’s underlying cellular mechanisms in the intestinal epithelium as incompletely understood.

Research status

Larazotide has an unusually deep human trial record for a peptide sold on the research market, and that record ends in failure at the decisive stage.

Early human work. The first proof-of-concept study in celiac subjects reported good tolerability and blunting of the gluten-induced rise in intestinal permeability seen in the placebo group (Paterson et al., Alimentary Pharmacology & Therapeutics, 2007;26(5):757–766). This single-dose study is the origin of the permeability claim, and it is the one place the claim held up.

Phase 2. In a randomized double-blind gluten-challenge study of 86 patients, the primary endpoint was the urinary lactulose/mannitol permeability ratio — and it produced no usable answer, with the authors reporting variability in the outpatient setting too high to assess permeability. Lower doses did appear to limit gluten-induced worsening of gastrointestinal symptoms, a secondary endpoint (Leffler et al., American Journal of Gastroenterology, 2012;107(10):1554–1562). A second placebo-controlled gluten-challenge trial in 184 patients reported reduced gluten-induced symptoms at 1 mg and lower anti-tissue-transglutaminase IgA, but again found no significant difference in intestinal permeability (Kelly et al., Alimentary Pharmacology & Therapeutics, 2013;37(2):252–262). The largest Phase 2b trial enrolled 342 adults with persistent symptoms despite a gluten-free diet and tested 0.5, 1, and 2 mg three times daily: the 0.5 mg arm met the primary symptom endpoint, while the 1 mg and 2 mg arms did not differ from placebo on any endpoint (Leffler et al., Gastroenterology, 2015;148(7):1311–1319.e6).

Phase 3 failure. The CeD-LA-3001 trial (ClinicalTrials.gov NCT03569007) was the first Phase 3 study of any drug for celiac disease, testing 0.25 mg and 0.5 mg three times daily against placebo in patients still symptomatic on a gluten-free diet. It was designed for 525 patients across three equal arms, but sponsor 9 Meters Biopharma announced in June 2022 that a pre-specified interim analysis had concluded the additional sample size required to demonstrate a significant difference from placebo was too large to justify continuing. Enrollment stopped at 307; the registry lists the trial as terminated, with the sponsor recorded as the reason.

Meta-analysis. A systematic review of four randomized trials covering 626 patients concluded larazotide was generally well tolerated and superior to placebo for gastrointestinal symptoms in patients undergoing gluten challenge — but found no benefit in patients following a gluten-free diet, which is the population both the Phase 2b and the Phase 3 trials were designed around (Hoilat et al., Clinics and Research in Hepatology and Gastroenterology, 2022;46(1):101782).

Work outside celiac disease. A Phase 2a trial in multisystem inflammatory syndrome in children was terminated early as case counts fell (NCT05022303) after enrolling just 12 children. Reported results suggested larazotide was tolerated in that group and pointed to faster spike-antigen clearance and symptom resolution as an adjuvant therapy (Yonker et al., Science Translational Medicine, 2025;17(809):eadu4284) — a signal from a dozen patients, not a controlled efficacy result. A Phase 2a study in long COVID in children and adults (NCT05747534, 107 participants, weight-based oral dosing four times daily for 21 days, ages 7–50) is listed as completed, with no results posted at the time of writing.

No regulatory authority has approved larazotide for any indication.

Common dosage forms

• Lyophilized powder in vials, commonly at around 5 mg, sold for reconstitution — a presentation inherited from research-peptide packaging conventions rather than from how the compound was studied.
• Oral capsules, which match the route used in every human trial.
• Multi-ingredient capsule blends, where larazotide is combined with other gut-oriented peptides in a single “gut” or “barrier” formula, so the larazotide content per capsule is often not stated separately.
• Modified variants such as N-acetyl larazotide, typically sold at 250 mcg per capsule. These are chemical analogs, not the peptide used in the clinical program, and no human trial has tested them.

This section describes formats only and is not dosing guidance.

Key considerations

• The pivotal trial failed. Larazotide is not an early-stage compound with an open question — it went all the way to Phase 3 and was stopped in 2022 when an interim analysis showed the effect was too small to chase. Marketing that cites its “Phase 3” status without that outcome is describing how far it got, not what it showed.
• The mechanism’s own biomarker never moved. Larazotide is sold on a tight-junction, gut-permeability story, yet the two controlled gluten-challenge trials that measured intestinal permeability directly failed to show a significant effect on it. Every positive human result in this program is a symptom score, not a barrier measurement.
• The positive signal was fragile and non-monotonic. The effective dose was not the same twice: lower doses in the 2012 trial, 1 mg in 2013, 0.5 mg in 2015 — with higher doses performing no better than placebo each time. A dose-response that runs backward and relocates between trials is a well-known warning sign, and none of it replicated at Phase 3 scale.
• The benefit, where it appeared, was in the wrong population. The meta-analysis found symptom benefit only under active gluten challenge, not in patients maintaining a gluten-free diet — but gluten-free patients with persistent symptoms were the commercial target and the Phase 3 population.
• Route mismatch in the research market. Larazotide was developed as a locally acting luminal agent working at the apical surface of intestinal cells, and every human trial dosed it orally. Injectable vial presentations have no counterpart in the clinical record.
• Serious naming collision. The code AT-1001 (also written AT1001) has been used independently for at least three unrelated compounds: larazotide; migalastat (marketed as Galafold) for Fabry disease; and a small-molecule α3β4 nicotinic acetylcholine receptor partial agonist studied for smoking cessation. Literature and registry searches on that code return a great deal of material about the other two, so a citation attached to “AT-1001” should always be checked against the actual compound. Larazotide also carried the sponsor code INN-202.
• The target itself is contested. Zonulin remains a debated construct in gastroenterology, no zonulin receptor has been definitively identified, and the Phase 3 result prompted reassessment of the zonulin-pathway hypothesis rather than just the molecule. Compound identity and content in multi-peptide blends are also difficult to verify independently.