Davunetide

What is Davunetide?

Davunetide is an 8-amino-acid peptide with the sequence NAPVSIPQ (Asn-Ala-Pro-Val-Ser-Ile-Pro-Gln), molecular formula C36H60N10O12 and molecular weight 824.9 daltons. Unlike several compounds on this site that are peptide-adjacent rather than genuinely peptides, davunetide is the real thing: a short linear peptide corresponding to a fragment of activity-dependent neuroprotective protein (ADNP), an 828-residue glial protein. It is commonly abbreviated NAP, after the first three residues. It has accumulated an unusual number of names across its development history, and knowing which is which makes the literature much easier to read. NAP is the original research designation from the laboratory that discovered it. AL-108 is the intranasal formulation and AL-208 the intravenous one, both from Allon Therapeutics. Davunetide is the international nonproprietary name, used for the intranasal product from around 2009. CP201 is the designation adopted later when the molecule was repositioned for ADNP syndrome. The honest summary of its status is that davunetide is one of the better-tested failed neuroprotective peptides. It went through a full development programme — five registered trials, roughly 520 participants in randomised placebo-controlled studies, including a 313-patient Phase 2/3 — and the pivotal trial produced a flat null result. Its sponsor entered insolvency proceedings months later. It holds a live FDA orphan drug designation for ADNP syndrome, but no trial in that indication appears in any public registry, and no human ADNP-syndrome data have been published. It is approved nowhere in the world for anything. Despite all of this it is widely sold through research-chemical and consumer peptide channels as a cognitive enhancer, which is the main reason it warrants a careful page rather than a footnote.

What Davunetide Is Investigated For

Davunetide's evidence base is unusual in a specific and instructive way: it is not thin, it is negative. Most compounds sold as research peptides have never been tested properly. Davunetide was, and it failed. The centrepiece is the AL-108-231 trial (Boxer et al., Lancet Neurology 2014): 313 patients with progressive supranuclear palsy randomised to 30 mg intranasal davunetide twice daily or placebo for 52 weeks across 48 centres. Both co-primary endpoints missed, and not narrowly — the change from baseline in the PSP Rating Scale was a median 11.8 in both arms with identical confidence intervals (p=0.41), and the Schwab and England activities-of-daily-living score was −0.20 in both (p=0.92). The authors' conclusion was one sentence: davunetide is not an effective treatment for PSP. This is a flat null, not an underpowered near-miss, and it is the single most important fact about this molecule. The cognitive record is no better, which matters because cognition is what davunetide is actually sold for. Three randomised placebo-controlled trials totalling roughly 520 participants measured cognitive outcomes and none met its prespecified cognitive endpoint. In amnestic mild cognitive impairment (n=144), analyses failed to detect a significant difference on the composite cognitive memory score; the only positives were unadjusted signals on two individual subtests. In schizophrenia (n=63), effects on MCCB-rated cognition were not significant versus placebo (p=0.45) — the one significant result was on UPSA, a functional-capacity rather than cognitive measure, and both measures showed an inverse dose-response with 5 mg outperforming 30 mg, which argues against a genuine pharmacological effect. The PSP trial measured cognition too and found nothing. What keeps the molecule alive is a series of post-hoc sex-stratified re-analyses published between 2023 and 2026, reporting that women in these trials benefited where the overall populations did not. These deserve neither dismissal nor credulity. They are unprespecified reanalyses of trials that missed their endpoints, published 10 to 14 years after the fact, overwhelmingly by the peptide's inventor — who is an officer of the company that now licenses it. No confirmatory trial has been registered to test the hypothesis they generate. Readers should treat the female-specific claim as an interesting question, not as a finding.

History & Discovery

Davunetide began as a fragment. In 1999, working in Illana Gozes' laboratory at the Sackler School of Medicine, Tel Aviv University, Merav Bassan and colleagues reported the complete 828-residue sequence of a glial protein they named activity-dependent neuroprotective protein, induced in astrocytes by vasoactive intestinal peptide. Within it sat an eight-residue stretch, NAPVSIPQ, that protected neurons in mixed culture at subfemtomolar concentrations — a potency figure so unusual that it defined the molecule's reputation for the next decade. The fragment was named NAP after its first three residues. Allon Therapeutics, a Vancouver company, licensed it and developed two formulations: AL-108 for intranasal delivery and AL-208 for intravenous. Between 2006 and 2010 the company ran a conventional development programme. An intravenous study in 234 patients looked at cognitive impairment after coronary artery bypass surgery; it has never been published and has never posted results. An intranasal Phase 2 in 144 patients with amnestic mild cognitive impairment reported in 2013 that it had failed to detect a significant difference on its composite cognitive memory score, while flagging signals on two individual subtests. An academic trial run through the TURNS network at UCLA tested it in 63 patients with schizophrenia and found no significant effect on cognition, though a functional-capacity measure improved. Davunetide received FDA orphan drug designation for progressive supranuclear palsy in December 2009 and an EU orphan designation in March 2010. The decisive trial ran from 2010 to 2012. AL-108-231 randomised 313 patients with progressive supranuclear palsy across 48 centres in six countries to 30 mg intranasal davunetide twice daily or placebo for 52 weeks, with the PSP Rating Scale and the Schwab and England activities-of-daily-living scale as co-primary endpoints. The result, published by Boxer and colleagues in Lancet Neurology in 2014, was as clean a negative as clinical research produces: median PSPRS change of 11.8 in both arms with identical confidence intervals, SEADL change of −0.20 in both arms, p-values of 0.41 and 0.92. Davunetide, the authors wrote, is not an effective treatment for PSP. An accompanying commentary asked when neuroprotective drugs should move from mice to men. Allon Therapeutics entered insolvency proceedings within months of the topline result and was acquired by Paladin Labs in 2013. The EU orphan designation was withdrawn at the sponsor's request in April 2013. Every registered davunetide trial had completed by then, and none has started since. The afterlife has two strands. The first is repositioning: as CP201, the molecule was proposed for ADNP syndrome — the neurodevelopmental disorder caused by mutations in the very protein it derives from, which is a genuinely compelling rationale. FDA granted orphan drug designation for ADNP syndrome in February 2018, and that designation is currently held on the FDA record by ExoNavis Therapeutics of Tel Aviv, which licensed the compound from Ramot at Tel Aviv University and lists Illana Gozes as a vice-president. The company has publicly described a Phase 3 trial in children and an intention to file for approval in progressive supranuclear palsy; neither appears in any public trial registry or regulatory record. The second strand is reanalysis. From 2023 onward, Gozes' group published a series of papers re-examining the old trial datasets by sex and reporting that women benefited where the overall populations had not — in progressive supranuclear palsy, in prodromal Alzheimer's disease, and in the cardiac-surgery study. These are unprespecified post-hoc analyses, published a decade or more after the trials, by the inventor of the compound and an officer of the company that licenses it. They have not been independently replicated and have not produced a registered confirmatory trial. They are best read as a hypothesis about why a null result might have concealed a real effect — an important question, and not the same thing as an answer.

How It Works

Nerve cells are held in shape by tiny internal scaffolding tubes called microtubules, which also act as the tracks along which the cell moves its cargo. In several brain diseases this scaffolding falls apart, partly because a protein called tau stops doing its stabilising job properly. Davunetide is a short fragment copied from a natural brain protein, and it appears to help hold the growing ends of those microtubules together by docking onto the proteins that cap them. The idea was that steadying the scaffolding would protect neurons in diseases where it disintegrates. The idea was reasonable, the cell biology is real, and when it was tested in a large trial in patients with a tau-driven brain disease, it did not help them.

Davunetide is the NAPVSIPQ fragment of activity-dependent neuroprotective protein (ADNP), an 828-residue glial protein identified in 1999 by Bassan and colleagues in Illana Gozes' laboratory at Tel Aviv University. ADNP expression is induced by vasoactive intestinal peptide in astrocytes, and the isolated NAP fragment showed neuroprotection in mixed neuron-astrocyte cultures at subfemtomolar concentrations — an extraordinary potency claim that framed the molecule's reputation from the outset. In the native protein the motif reads NAPVSIPQQ; the drug is the eight-residue form. The mechanism is microtubule stabilisation through the end-binding protein family, and this is the part of the davunetide story that has held up best. Microtubule plus-ends are regulated by EB1, EB2 and EB3, which recruit partner proteins through a short linear motif known as SxIP. Sequence analysis identified exactly such a motif within davunetide itself — the Ser-Ile-Pro at positions 5 to 7 — and demonstrated specific SxIP binding sites across all three EB family members. The functional test was clean: silencing EB1 or EB3, but not EB2, inhibited davunetide-mediated cell protection, and a sequence variant that preserved the SxIP motif retained activity while one that abolished it did not. Later work showed davunetide preventing microtubule degradation by recruiting both tau and end-binding proteins to microtubules, with expression of those proteins required for the effect. The tau connection is where the therapeutic hypothesis came from and where it broke down. Davunetide reduced the ratio of phosphorylated to total tau in subcortical regions of alpha-synuclein-overexpressing mice, and the preclinical rationale taken into trials was that it promoted microtubule stability and reduced tau phosphorylation. After the progressive supranuclear palsy failure, affinity chromatography from the originating laboratory showed that davunetide interacts preferentially with three-repeat tau — whereas progressive supranuclear palsy is characterised by four-repeat tau pathology. On that reading the pivotal trial tested the compound against the tau isoform it binds least well. This is a coherent explanation and it is also a post-hoc one from an interested party, so it should be read as a hypothesis the field has not tested rather than a settled account. Several further mechanisms have been proposed, almost all from the same laboratory: an ADNP-SIRT1 interaction at the EB1/EB3-tau level, an SH3-domain-binding site connecting ADNP to SHANK3 and hence to autism biology more broadly, nuclear as well as cytoplasmic distribution of the peptide, and entry into cells by dynamin-associated endocytosis. Independent work in 2026 reported davunetide promoting recovery after spinal cord injury through SIRT1-FOXO1-TFEB-mediated autophagy. The breadth is notable but so is the concentration of sourcing. The critical gap is translational. Nothing in this mechanistic account has been demonstrated in treated humans. No human biomarker study has shown davunetide stabilising microtubules or reducing tau phosphorylation in patients, and the only human central-nervous-system readout — a magnetic resonance spectroscopy substudy of N-acetylaspartate and choline in schizophrenia — enrolled 18 subjects and produced unchanged results.

Evidence Snapshot

Research Gaps & Open Questions

What the current literature has not yet settled about Davunetide:

• 01Whether the post-hoc female-specific efficacy signal is real is the central open question, and it is unresolved in a way that is nobody's fault but has gone unaddressed for over a decade. Four reanalyses between 2023 and 2026 report that women benefited in trials whose overall populations did not, across progressive supranuclear palsy, prodromal Alzheimer's disease and cardiac surgery. All are unprespecified, all come overwhelmingly from the inventor's laboratory, none has been independently replicated, and no confirmatory trial has been registered. A prospective sex-stratified study would settle it; none exists.
• 02The declared conflict of interest surrounding the favourable literature is itself a gap in the evidence base. Illana Gozes discovered the peptide, authored or senior-authored the great majority of the mechanistic work and all of the favourable reanalyses, and is an officer of ExoNavis Therapeutics, which licenses the compound. This does not make the work wrong, but it means the positive literature has had very little independent scrutiny, and readers should weight it accordingly.
• 03Whether the 3-repeat versus 4-repeat tau selectivity hypothesis explains the progressive supranuclear palsy failure has never been tested. If davunetide genuinely binds 3R tau preferentially, a trial in a 3R-predominant or mixed tauopathy would be the obvious experiment. Seven years after that hypothesis was published, no such trial has been run.
• 04No human target-engagement or biomarker data exist. Nothing shows that davunetide stabilises microtubules, reduces tau phosphorylation, or reaches meaningful brain concentrations in people. Without this, a null trial cannot distinguish between an ineffective drug and an inadequately delivered one — which is precisely the ambiguity the trial investigators identified as the field's central obstacle.
• 05The intravenous programme is essentially undocumented. The 234-patient AL-208 cardiac surgery trial completed in 2008, has never been published, and has posted no results. A 2025 reanalysis reported differential bioavailability between men and women with possible dose-dependent accumulation in men, caveated for small cohort size. An entire Phase 2 dataset in a surgical population sits unpublished.
• 06ADNP syndrome is the most mechanistically compelling indication davunetide has ever had — restoring function to a fragment of the very protein that is mutated — and there is no human data at all. The FDA orphan designation dates from 2018, the sponsor has publicly described a Phase 3 in children, and nothing appears in ClinicalTrials.gov or the EU Clinical Trials Register.
• 07Whether gray-market self-administration achieves any central exposure is unknown and unstudied. Community doses run at roughly a fifth of the lowest trial dose, in reconstituted research powder rather than a characterised intranasal formulation. No pharmacokinetic work addresses this scenario, which is nonetheless how most people who take davunetide actually take it.
• 08No long-term human safety data exist beyond 52 weeks, and none at all in children — a notable gap given that the only live designation targets a paediatric genetic syndrome.

Forms & Administration

Davunetide was developed principally as an intranasal spray (AL-108), with an intravenous formulation (AL-208) used in one cardiac-surgery trial. Trial doses were 5 mg once daily, 15 mg twice daily and 30 mg twice daily intranasally, and a single 300 mg intravenous dose. There is no approved product and no pharmaceutical formulation available anywhere. What circulates is research-grade lyophilised powder from laboratory chemical suppliers, labelled not for human or veterinary use, reconstituted by individuals for intranasal or subcutaneous use at roughly 2 to 5 mg per day — below every dose used in every trial, in a formulation whose nasal bioavailability bears no established relationship to the clinical product.

Dosing & Protocols

The ranges below reflect protocols commonly discussed in the literature and by clinicians — not a prescription. Actual dosing for any individual should be determined by a qualified healthcare provider who knows the patient.

Davunetide is not approved by any regulator in any country for any indication. The doses recorded here describe what was administered in completed clinical trials and exist for reference only. They are not a protocol, and the compound's largest and longest trial found no benefit at the highest dose tested.

Timeline of Effects

Common Questions

Who Davunetide Is NOT For

Drug & Supplement Interactions

No formal drug-interaction studies have been published for davunetide, and the trial reports contain no interaction analyses. As a small peptide cleared by proteolysis rather than hepatic metabolism, cytochrome P450-mediated interactions are not expected on mechanistic grounds, though this has not been demonstrated. The interactions worth considering are pharmacodynamic and largely theoretical. Davunetide's mechanism is microtubule stabilisation, which places it in nominal conceptual overlap with microtubule-targeting chemotherapeutics such as the taxanes and vinca alkaloids — agents that stabilise or destabilise microtubules systemically and with considerable toxicity. No interaction has been studied and the exposures are not comparable, but concurrent use would be unwise without oncological input. In the populations actually studied, concomitant medication burden was high and no signal emerged. A published post-hoc analysis of the progressive supranuclear palsy dataset specifically examined amantadine use and found no association with gait or cognitive measures, which is a useful if narrow piece of reassurance. For intranasal administration, concurrent intranasal corticosteroids, decongestants or other nasal sprays would be expected to alter deposition and absorption, and would compound the local irritation that was davunetide's most consistent adverse effect. None of this has been characterised in controlled studies. Anyone encountering davunetide in a clinical context should treat its interaction profile as unknown rather than as benign.

Safety Profile

Legal Status

Regulatory status changes over time. Verify current local rules with a qualified professional.

Myths & Misconceptions