NNZ-2591

A synthetic, orally active analog of cyclic glycine-proline, an endogenous IGF-1 metabolite, developed by Neuren Pharmaceuticals (proposed INN ercanetide) for Phelan-McDermid, Pitt Hopkins and Angelman syndromes, with a placebo-controlled phase 3 trial in Phelan-McDermid syndrome under way since late 2025.

Also referenced as: NNZ2591, NNZ 2591, Ercanetide, cyclo-L-glycyl-L-2-allylproline

Also appears in: Cognitive

COA Public product evidence No certificate records are bound to a tracked NNZ-2591 product page A record counts here only when a vendor publishes a certificate against one of its own tracked NNZ-2591 product pages. Other certificates naming this compound may still exist — search the wider certificate database. →

---

What is NNZ-2591?

NNZ-2591 is a synthetic analog of cyclic glycine-proline (cGP), a small ring-shaped metabolite that the brain derives from the N-terminal tripeptide of insulin-like growth factor 1. Chemically it is cyclo(L-glycyl-L-2-allylproline), a 2,5-diketopiperazine of two amino acids (C10H14N2O2, molecular weight 194.2, single (8aR) stereocenter) in which an allyl group has been added to the alpha-carbon of the proline to resist enzymatic opening of the ring; it is a cyclic dipeptide, so it qualifies as a peptide only in the most minimal sense and behaves pharmacologically like an oral small molecule, crossing the blood-brain barrier after peripheral dosing (Guan et al., Neuropharmacology, 2007;53(6):749–762). It originated at the Liggins Institute of the University of Auckland and is developed by Neuren Pharmaceuticals, the company behind trofinetide; it received the proposed international nonproprietary name ercanetide in 2025 (proposed INN list 133) and a United States Adopted Name in 2026.

The two Neuren compounds are siblings from the same IGF-1 fragment. Trofinetide (NNZ-2566) is a stabilized version of the linear tripeptide glycine-proline-glutamate (GPE); NNZ-2591 is a stabilized version of cGP, the two-residue ring that forms when the glycine-proline portion of GPE cyclizes and the glutamate is lost. NNZ-2591 reached the clinic for rare genetic neurodevelopmental disorders rather than for brain injury, with open-label phase 2 trials in Phelan-McDermid, Pitt Hopkins and Angelman syndromes reported in 2023 and 2024 and a placebo-controlled phase 3 trial in Phelan-McDermid syndrome enrolling since late 2025. It does not appear to be sold as a research-market peptide.

How it works

The mechanism proposed by its developers is normalization of IGF-1 signaling in the brain rather than direct agonism at any single receptor. The evidence is preclinical and comes largely from one academic group with ties to the sponsor.

• Modulation of IGF-1 bioavailability. Native cGP competes with IGF-1 for binding to IGF-binding protein 3; in cell culture it increased IGF-1 activity when IGF-1 was scarce and dampened it when IGF-1 was in excess, and mathematical modeling attributed the effect to a shift in the balance between bound and free IGF-1 (Guan et al., Scientific Reports, 2014;4:4388). The sponsor’s working model is that NNZ-2591 applies this auto-regulatory brake-and-accelerator to synaptic signaling that is disrupted in several genetically distinct neurodevelopmental disorders (reviewed in Guan et al., Molecules, 2023;28(3):1021).
• Neuroprotection after hypoxia-ischemia. Given intracerebroventricularly or subcutaneously two hours after hypoxic-ischemic injury in adult rats, NNZ-2591 reduced damage in cortex, hippocampus and striatum more potently than cGP itself, was stable in plasma, reached the cerebrospinal fluid, and over five days of subcutaneous treatment improved somatosensory-motor function while reducing caspase-3-mediated apoptosis and microglial activation and increasing astrocytic reactivity (Guan et al., Neuropharmacology, 2007;53(6):749–762).
• Functional recovery without dopamine rescue. In the 6-hydroxydopamine rat model of Parkinson’s disease, early central treatment limited the motor deficit, and delayed subcutaneous treatment begun two weeks after the lesion improved motor function from week seven onward and increased doublecortin-positive neuroblasts in the subventricular zone, without altering nigrostriatal dopamine depletion (Krishnamurthi et al., British Journal of Pharmacology, 2009;156(4):662–672).
• Cholinergic modulation and memory. Oral NNZ-2591 improved memory retention in scopolamine-treated rats in the Morris water maze, counteracted the scopolamine-induced rise in striatal choline acetyltransferase-positive neurons and hippocampal synaptophysin, and showed antagonism at muscarinic M2 receptors independent of scopolamine (Guan et al., Behavioural Brain Research, 2010;210(2):221–228).

The sponsor has also reported that NNZ-2591 corrected behavioral and synaptic deficits in mouse models of each target syndrome, and cites an unpublished 2025 mouse study for the neonatal brain-injury indication, but as of November 2026 those studies exist as company statements and conference material rather than as PubMed-indexed papers, and no human mechanistic data have been published.

Research status

Phase 1. A first-in-human study of single and multiple ascending oral doses in 28 healthy adults aged 18–55 ran from May 29, 2020 to February 11, 2021 (NCT04379869). No results have been posted to the registry or published.

Phelan-McDermid syndrome (phase 2 complete, phase 3 under way). PMS-001 (NCT05025241) treated 18 children aged 3–12 (mean 8.6 years) at four United States hospitals with twice-daily oral NNZ-2591 for 13 weeks. Primary endpoints were safety and pharmacokinetics; 14 efficacy measures were secondary. Top-line results were announced in December 2023, and the peer-reviewed report followed (Neumeyer et al., Neurology Genetics, 2026;12(1):e200338, published online December 23, 2025, with an erratum in 2026;12(4):e200405). Ten of 14 measures improved significantly from baseline by Wilcoxon signed-rank test: the PMS-specific Clinical Global Impression of Improvement averaged 2.4 with 16 of 18 participants rated improved (p < 0.0001), the Caregiver Impression of Change averaged 2.7 with 15 of 18 improved (p = 0.0003), and the Vineland-3 receptive communication raw score rose a mean 7.5 points from a baseline of 29.0 (p = 0.0001). Most adverse events were mild to moderate; one serious event (gastroenteritis, judged unrelated) occurred after treatment ended, and three participants discontinued, two after COVID-19 infection and one for seizures judged unrelated. A population pharmacokinetic analysis linked estimated steady-state 24-hour exposure above 300 µg·h/mL to CGI-I improvement.

After a Type C meeting in April 2025, the FDA agreed that a single pivotal trial with two co-primary endpoints, the Vineland-3 receptive communication raw score and the Phelan-McDermid Syndrome Assessment of Change, a syndrome-specific clinician global impression scale, could support a New Drug Application. Koala (NCT07281079) is that trial: a randomized, double-blind, placebo-controlled phase 3 study in about 160 children aged 3–12 over 13 weeks of treatment, registered with a start date of November 12, 2025, first participant dosed in February 2026, recruiting at sites in the United States and Canada, and estimated primary completion on October 31, 2027. A 52-week open-label extension (NCT07593391, about 180 participants, started June 8, 2026) accepts completers of Koala and eligible participants from the phase 2 trial. The program carries FDA Fast Track, Rare Pediatric Disease and orphan drug designations, plus European Union orphan designation.

Pitt Hopkins syndrome (phase 2 complete, next trial under FDA discussion). PTHS-001 (NCT05025332) enrolled 28 children aged 3–17 at five United States sites, of whom 16 passed screening and were dosed on the same 13-week design. Top-line results announced in May 2024 covered 11 completers: the PTHS-specific CGI-I averaged 2.6 with 9 of 11 improved (p = 0.0039), the Caregiver Impression of Change averaged 3.0 with 8 of 11 improved (p = 0.0234), and the CGI-Severity and caregiver top-three-concerns measures also reached significance, while the ten non-syndrome-specific measures did not. Four participants withdrew for adverse events, two of them (constipation, self-injury, abdominal distention and fatigue in one; sleep disorder and constipation in the other) judged drug-related, and one withdrew because of the protocol’s monitoring burden. No peer-reviewed publication had appeared as of November 2026. Written FDA feedback in the first quarter of 2026 indicated that a controlled trial modeled on the Phelan-McDermid phase 3 design would be acceptable, an approach the sponsor described as neither appropriate nor executable in so rare and more profoundly affected a population; a Type B End-of-Phase-2 meeting was scheduled for late October 2026, and its outcome had not been announced as of November 3, 2026.

Angelman syndrome (phase 2 complete, no next trial announced). NCT05011851 enrolled 17 children aged 3–17 at three Australian hospitals. The sponsor’s results, announced in August 2024, covered 16 dosed children (mean age 10 years) and 13 completers: the AS-specific CGI-I averaged 3.0 with 11 of 13 improved (p = 0.0010), the Caregiver Impression of Change averaged 3.2 with 8 of 12 improved (p = 0.0273), and all eight children aged 3–12 improved on both scales (mean 2.8, p = 0.0078). The most frequent adverse events were viral infection (31%), nasopharyngitis (25%), seizure (25%), upper respiratory infection, somnolence and constipation (19% each); there were no serious events. The indication holds FDA orphan, Fast Track and Rare Pediatric Disease designations and EU orphan designation, but no registration trial had been registered as of November 2026, and the sponsor’s 2026 presentations list it only among indications under consideration.

Prader-Willi syndrome (withdrawn). An open-label phase 2 trial (NCT05879614) was opened under an FDA-cleared IND with orphan designation, with a registered start date of September 2023, but enrollment was paused in 2024 and the record is now listed as withdrawn with zero participants enrolled, citing a revised development strategy. Prader-Willi syndrome still appears among the sponsor’s possible future indications.

Hypoxic-ischemic encephalopathy (preclinical). Neuren added a neonatal HIE program in 2025, returning the compound to the injury setting of its original rat studies. In February 2026 the FDA generally accepted the design and doses of an IND-opening pharmacokinetic and safety study but required an additional juvenile animal safety study first; the protocol for that study went to the FDA in May 2026, the sponsor now guides to an IND submission in the first half of 2027, and the FDA granted orphan drug designation for neonatal HIE on November 2, 2026.

Common dosage forms

• Oral solution. The only clinical presentation: the phase 2 trials used a 50 mg/mL oral solution, and the phase 3 Koala study and its extension use the same liquid format. No commercial product exists in any country.
• Preclinical formulations. The rat studies used oral gavage, subcutaneous injection and intracerebroventricular infusion.
• Reagent-grade powder. The compound is listed in analytical and chemical reagent catalogs under the names NNZ-2591 and ercanetide.

Key considerations

• Regulatory status. The Investigational status reflects an active, FDA-regulated phase 3 program in Phelan-McDermid syndrome with orphan, Fast Track and Rare Pediatric Disease designations, with no marketing application filed and no approval anywhere as of November 2026. Koala’s estimated primary completion is October 31, 2027.
• Open-label efficacy data. The 2023–2024 phase 2 signals come from single-arm trials of 16–18 children each, judged by clinician and caregiver impression scales designed for each syndrome, with sponsor employees among the authors of the only peer-reviewed report. Improvement from baseline without a placebo arm is vulnerable to expectation and regression to the mean, and in Pitt Hopkins syndrome the generic neurodevelopmental measures did not move. Koala is the first test against placebo.
• Short tolerability record. No serious drug-related events were reported across roughly 50 treated children in 13-week trials, but two Pitt Hopkins participants withdrew for drug-related gastrointestinal, sleep and behavioral events, and exposure beyond 13 weeks is only now being gathered in the open-label extension. The parent IGF-1 pathway’s mitogenic potential is a theoretical concern that the short trials cannot address.
• Program attrition. The Prader-Willi trial was withdrawn without enrolling anyone, no registration trial exists for Angelman syndrome, and the Pitt Hopkins path depends on an FDA meeting whose outcome is pending. The newest indication, neonatal HIE, has no human data at all.
• Naming and identity. NNZ-2591, ercanetide and cyclo(L-glycyl-L-2-allylproline) are the same molecule. It is not trofinetide (NNZ-2566, Daybue), which is a linear tripeptide analog of GPE, and it is not native cyclic glycine-proline (cGP), the unmodified two-residue ring from which it was derived.