Identity
Davunetide — also known as NAP, AL-108 and CP201 — is a synthetic eight-amino-acid peptide with the sequence NAPVSIPQ (Asn-Ala-Pro-Val-Ser-Ile-Pro-Gln). It is the active fragment of activity-dependent neuroprotective protein (ADNP), a protein released by glial cells in response to vasoactive intestinal peptide and essential for normal brain development. ADNP is one of the more frequently mutated genes in syndromic autism, and somatic ADNP mutations have been described in Alzheimer's disease — which is the thread linking davunetide's otherwise scattered list of attempted indications. It has been developed for intranasal administration throughout.
Development & history
- 1990s–2000s: ADNP and its active NAP fragment characterised by Illana Gozes and colleagues at Tel Aviv University. Preclinical work established microtubule stabilisation as the proposed mechanism, with neuroprotection across varied toxic insults in cell culture and reduced tau pathology in MAPT-mutant mouse models.
- 2007–2008: Developed clinically by Allon Therapeutics (Canada) as AL-108 (intranasal) and AL-208 (intravenous), positioned as a tau-directed neuroprotective agent.
- 2013: Phase 2 in amnestic mild cognitive impairment published (Morimoto et al.) — 144 subjects, 12 weeks intranasal, signals on attention and working memory, acceptable tolerability. This was the result that justified moving to a larger trial.
- 2010–2012: The pivotal phase 2/3 trial in progressive supranuclear palsy ran at 48 centres in six countries — 313 participants randomised 1:1 to 30 mg intranasal twice daily or placebo for 52 weeks.
- 2014: Results published in Lancet Neurol (Boxer et al.). The trial missed both co-primary endpoints (PSPRS and SEADL). The authors' conclusion was that davunetide is not an effective treatment for PSP, while noting that disease-modifying trials in PSP are feasible and should continue with other tau-directed candidates. Allon's programme did not recover; the compound passed through Paladin Labs.
- 2021: ExoNavis Therapeutics licensed davunetide from Tel Aviv University, refocusing it on ADNP syndrome — a rare monogenic disorder in which the rationale is replacement of a deficient gene product rather than modification of a complex late-onset disease.
- 2023: A post-hoc reanalysis of the PSP cohort (Gozes et al., Transl Psychiatry) reported that female participants deteriorated faster on placebo and that davunetide appeared to benefit women. Widely cited since as evidence the drug "works in women."
- 2024: ExoNavis states a phase 3 trial in 97 children with ADNP mutations began in October. ALZFORUM notes the trial does not appear in clinical trial registries.
- Designations: US FDA orphan drug (DRU-2017-6243) and rare pediatric disease designation; EMA orphan medicinal product designation — all for ADNP syndrome, all granted on the basis of rarity rather than demonstrated efficacy.
Mechanism (as proposed)
Davunetide's proposed mechanism is microtubule stabilisation. The peptide interacts with tubulin and with microtubule end-binding proteins, supporting the microtubule network that neurons depend on for axonal transport, cell morphology and synaptic function. Because tau's normal job is stabilising microtubules — and because tauopathies are diseases in which that stabilisation fails and hyperphosphorylated tau accumulates — a peptide that stabilises microtubules directly is a coherent therapeutic idea. In preclinical models this played out as expected: reduced tau hyperphosphorylation, less insoluble tau, protection against microtubule-disrupting insults, and improved behavioural performance in MAPT-mutant mice.
The honest reading is that this mechanism was tested in humans in the most favourable possible setting and did not deliver. Progressive supranuclear palsy is a "pure" tauopathy — the cleanest available human test of an anti-tau, microtubule-stabilising agent — and a well-powered, well-run, 52-week randomised trial found no benefit on either co-primary endpoint. A coherent mechanism with strong animal data that fails its direct human test is a common pattern in neurodegeneration drug development, and davunetide is one of its clearer examples.
Reading the evidence honestly
Three points are worth separating, because they are routinely blurred when davunetide is discussed outside the literature.
The negative trial is the main fact. It was large by neurodegeneration standards, international, double-blind, placebo-controlled, and 52 weeks long. It failed. Sources that lead with "clinical studies have demonstrated safety, favourable bioavailability and potential efficacy" are not describing that trial's outcome.
The "efficacy in women" claim is a post-hoc subgroup analysis of that failed trial. Post-hoc subgroup findings are the weakest tier of clinical evidence and the standard mechanism by which failed trials acquire a second life. It is a legitimate hypothesis worth prospective testing. It is not a demonstrated effect, it has not been replicated in a trial designed to test it, and its senior author is the compound's named inventor and an officer of the company that licensed it — a conflict that should be visible when weighing the claim, not hidden.
Orphan designation is not evidence. FDA orphan drug status, rare pediatric disease designation and EMA orphan medicinal product status are incentives to develop treatments for rare conditions. They are awarded on the basis of the disease's rarity and a plausible rationale, before efficacy is shown. Citing them as validation inverts what they mean.
None of this makes davunetide worthless. The ADNP syndrome programme has a more defensible rationale than the adult neurodegeneration attempts, the safety record across trials is unremarkable in the good sense, and the mechanism remains interesting. But the compound's grade reflects where the human evidence actually stands: one clean negative trial in the disease it was best suited to, one small unreplicated positive signal, and one live programme with nothing published and an unregistered trial.