Davunetide

Davunetide is the active core of ADNP, a protein essential for brain formation. It binds tubulin and enhances the association of end-binding proteins EB1 and EB3 with microtubule plus ends, which stabilises dynamic microtubules and supports axonal transport. In tauopathy models this translates into reduced tau hyperphosphorylation and preserved neuronal architecture. It is also neuroprotective against a range of insults in culture at femtomolar concentrations. Human development followed the mechanism logically - amnestic MCI, then schizophrenia, then progressive supranuclear palsy as a pure tauopathy - and the pivotal PSP trial was unambiguously negative on both primary endpoints, which is the most important fact about this compound.

Targets: Tubulin, EB1 and EB3 microtubule end-binding proteins, Tau phosphorylation

• · 30 mg per dose, 60 mg/day. This is the dose that failed in the PSP trial - it is included because it is the only properly characterised human regimen, not because it worked.
• · 5 mg and 15 mg twice-daily arms; the 15 mg arm showed improvement on some memory measures in an early phase 2 study.

Work out your exact syringe units →

• commonNasal discomfort, rhinorrhoea or epistaxis— The dominant adverse effect in trials, and unsurprising at 30 mg per nostril-dose.
• commonHeadache
• uncommonDysgeusia or taste disturbance
• uncommonDizziness

• Pregnancy and breastfeeding - no data.
• Nothing else is well established; the main argument against davunetide is that it did not work, not that it was dangerous.

• synergycerebrolysin — Both are pitched at neuroprotection in dementia; no combination data exists.
• synergyp21 — Both reduce tau pathology by different routes. Theoretical only.

• · No specific bloodwork.
• · If used for a cognitive indication, use a validated instrument at baseline and at 12 weeks. The phase 3 data say to expect nothing.

Not approved anywhere. Development was discontinued after the PSP failure; later revived in early trials for ADNP syndrome under the code CP201.

• Boxer et al. 2014, davunetide in patients with progressive supranuclear palsy: a randomised, double-blind, placebo-controlled phase 2/3 trial, Lancet Neurology (trial)
• Gozes et al., NAP (davunetide) and microtubule-associated neuroprotection - mechanism review (review)

Davunetide is worth studying carefully precisely because it is the compound in this class whose mechanism was most convincingly established and which then failed anyway. The molecular story is genuinely good: NAP is the active core of ADNP, a protein essential for brain formation; it binds tubulin and enhances the association of the end-binding proteins EB1 and EB3 with microtubule plus ends; that stabilises dynamic microtubules and supports axonal transport; in tauopathy models this reduces tau hyperphosphorylation and preserves neuronal architecture. It is neuroprotective in culture at femtomolar concentrations, which is a remarkable potency figure and was heavily cited during development. The clinical programme then followed the mechanism with unusual discipline - amnestic MCI first, where a phase 2a reported a positive effect on memory measures; then schizophrenia, where a 63-patient 12-week trial at 5 mg and 30 mg twice daily was unimpressive; then progressive supranuclear palsy, chosen deliberately as a pure tauopathy where a microtubule-stabilising tau-directed agent should show its clearest effect. That trial randomised 313 patients across 48 centres in six countries for 52 weeks at 30 mg twice daily intranasally, and the co-primary endpoints came out identical to placebo: median PSPRS change 11.8 in both arms (p=0.41), SEADL change -0.20 in both arms (p=0.92). The authors' conclusion was flat - davunetide is not an effective treatment for PSP. The lesson generalises well beyond this molecule. A femtomolar in-vitro potency, a clean mechanism, a rational indication and good tolerability produced nothing at all in the disease the mechanism predicted.

The primary thing that goes wrong is the decision to use it. Davunetide had the best-designed clinical programme of anything in this class and the answer came back clearly negative at the dose and duration that should have shown an effect. Using it anyway on the strength of the mechanism is choosing the preclinical story over the clinical result, which is exactly the error the trial existed to prevent. Beyond that: nasal adverse effects dominated in trials and are unsurprising at 30 mg per dose - epistaxis, rhinorrhoea and nasal discomfort were all meaningfully more common than placebo. Home reconstitution cannot reproduce the proprietary intranasal formulation's deposition characteristics, so even someone willing to buy 60 mg a day of research-grade peptide is not administering what was studied. And the cost arithmetic defeats the whole exercise before it starts - the clinical regimen is roughly 1.8 grams of peptide a month.

1. 5 mgFrom day 1 — 5 mg twice daily intranasally - the lower arm of the amnestic MCI and schizophrenia trials. No titration was used in any trial; dosing was fixed from the first day and tolerability was good across the entire range.
2. 15 mgFrom day 1, alternative — 15 mg twice daily - the arm of the early phase 2 MCI study that reported improvement on some memory measures. This is the only positive human signal davunetide ever produced and it was not replicated.
3. 30 mgFrom day 1, phase 3 dose — 30 mg twice daily, 60 mg/day, for 52 weeks - the regimen from the PSP trial. Listed because it is the best-characterised human regimen in this entire class, and it is the dose that definitively did not work. At 30 mg per administration, a 10 mg vial does not last a single dose.

• Tubulin — Direct binding established; neuroprotection in culture is reported at femtomolar concentrations, though that figure describes a cell-survival assay rather than a binding constant

  Microtubule stabilisation and support of axonal transport.

• EB1 and EB3 (microtubule end-binding proteins) — Enhances association with microtubule plus ends; not quantified as an affinity

  This is the specific molecular action that distinguishes davunetide from a generic tubulin ligand, and it is the mechanism the whole development programme was built on.

• Tau phosphorylation — Indirect consequence of microtubule stabilisation

  Reduced tau hyperphosphorylation and preserved neuronal architecture in tauopathy models. Did not translate into clinical benefit in progressive supranuclear palsy.

• Davunetide in patients with progressive supranuclear palsy (AL-108-231) Phase 2/3, randomised, double-blind, placebo-controlled · n=313 · 52 weeks · 2014

  Co-primary endpoints of change from baseline in the PSP Rating Scale and the Schwab and England Activities of Daily Living scale over 52 weeks, at 30 mg twice daily intranasally across 48 centres in six countries. Unambiguously negative: median PSPRS change 11.8 in both arms (p=0.41), SEADL change -0.20 in both arms (p=0.92). The authors concluded davunetide is not an effective treatment for PSP.

• Davunetide (AL-108) in schizophrenia Phase 2, randomised, double-blind, placebo-controlled, parallel group · n=63 · 12 weeks · 2012

  Cognition and functional capacity in schizophrenia, at 5 mg or 30 mg intranasally versus placebo over 12 weeks. Did not produce a convincing cognitive benefit. A companion MRS substudy in 18 of the participants examined N-acetylaspartate and choline in dorsolateral prefrontal cortex.

Nothing about this compound is acute. Trial endpoints were assessed at 12 weeks in schizophrenia and 52 weeks in PSP. No titration was used and none is needed; tolerability was good from the first dose across the full range. The honest timeline is that a 52-week trial at the maximum dose in the most mechanistically appropriate disease produced no separation from placebo at any timepoint.

There is very little to say here because there is very little reason to use the compound. The pairings people propose - with Cerebrolysin for general neuroprotection, or with P021 as two tau-directed agents by different routes - have no combination data and, more to the point, the monotherapy failed decisively in the disease its mechanism was designed for. Stacking a compound that did not work with another compound that has never been tested in humans is not a strategy. The one genuinely useful thing to know is arithmetic: the trial dose was 30 mg twice daily, so 60 mg a day, which means a 10 mg vial is a sixth of a single day's exposure. Anyone running a home protocol at 250-500 mcg is not approximating the clinical regimen in any sense - they are two orders of magnitude below it, and whatever they are testing, it is not what the trials tested.

Against everything else in this class, davunetide is the control experiment. It had the cleanest mechanism, femtomolar in-vitro potency, a rational disease target, excellent tolerability, real pharmaceutical funding and a properly powered blinded phase 2/3 trial - and it did not work. Read every confident preclinical claim about P021, FGL, Dihexa or the Semax analogues with that in mind, because none of them have been tested to anything like the same standard, and davunetide shows what happens when a compound from this space finally is. Against P021, the other tau-directed peptide here: P021 has intact rodent data and no human trials; davunetide has a definitive negative human trial. Which of those is the worse position depends on whether you think an untested hypothesis is preferable to a falsified one. Against Cerebrolysin: Cerebrolysin's trials were ambiguous, davunetide's was clear. Clarity is worth something even when the answer is no. The compound retains genuine interest for ADNP syndrome, a rare genetic disorder where it has been revived in early trials under the code CP201 and where the biological rationale is entirely different from the tauopathy story - that is a legitimate line of development and has nothing to do with nootropic use.

$200–$3000/month. The arithmetic is what kills this compound as a consumer product. The trial regimen of 60 mg/day needs about 1.8 grams a month, and research-grade peptide at typical grey-market pricing puts that in the low thousands of dollars monthly. A 10 mg vial at 40-80 USD is a sixth of one trial dose. People running 250-500 mcg daily spend little, but they are dosing two orders of magnitude below anything that has ever been studied - which is cheap in the same way that not taking it is cheap.

• ?No published human pharmacokinetic data despite a completed clinical development programme - no tmax, bioavailability, volume of distribution, clearance or protein binding.
• ?The phase 2a amnestic MCI result showing a positive memory effect is described in a developer-authored review rather than a primary trial publication I could resolve, and its participant numbers and duration I did not verify.
• ?How much of a home-reconstituted intranasal dose reaches the brain compared with the proprietary trial formulation is unknown and probably substantially less.
• ?The femtomolar neuroprotection figure comes from cell-culture survival assays and does not represent a binding affinity.
• ?Why a compound with clear target engagement in preclinical tauopathy models produced nothing in a pure human tauopathy has never been satisfactorily explained. Insufficient CNS exposure, wrong disease stage, and an incorrect mechanistic premise are all live possibilities.
• ?The status and results of the CP201 programme in ADNP syndrome I did not verify.
• ?No data exists on whether the sub-milligram doses used in grey-market protocols do anything at all; they are roughly 100-fold below the studied dose.

• Davunetide in patients with progressive supranuclear palsy: a randomised, double-blind, placebo-controlled phase 2/3 trial Boxer AL, Lang AE, Grossman M, Knopman DS, Miller BL, Schneider LS, Doody RS, Lees A, Golbe LI, Williams DR, Corvol JC, Ludolph A, Burn D, Lorenzl S, Litvan I, Roberson ED, Hoglinger GU, Koestler M, Jack CR, Van Deerlin V, Randolph C, Lobach IV, Heuer HW, Gozes I, Parker L, Whitaker S, Hirman J, Stewart AJ, Gold M, Morimoto BH; AL-108-231 Investigators, Lancet Neurology, 2014 · PMID 24873720

  The definitive negative trial. 313 patients, 52 weeks, both co-primary endpoints identical to placebo. The most rigorously conducted study of any compound in this class, and the most important single fact about davunetide.

• Effect of the neuroprotective peptide davunetide (AL-108) on cognition and functional capacity in schizophrenia Javitt DC, Buchanan RW, Keefe RS, Kern R, McMahon RP, Green MF, Lieberman JA, Goff DC, Csernansky JG, McEvoy JP, Jarskog F, Seidman LJ, Gold JM, Kimhy D, Nolan KS, Barch DS, Ball MP, Robinson J, Marder SR, Schizophrenia Research, 2012 · PMID 22169248

  The 63-patient, 12-week schizophrenia trial at 5 mg and 30 mg. The intermediate negative result between the promising MCI phase 2a and the definitive PSP failure.

• Effects of davunetide on N-acetylaspartate and choline in dorsolateral prefrontal cortex in patients with schizophrenia Jarskog LF, Dong Z, Kangarlu A, Colibazzi T, Girgis RR, Kegeles LS, Barch DM, Buchanan RW, Csernansky JG, Goff DC, Harms MP, Javitt DC, Keefe RS, McEvoy JP, McMahon RP, Marder SR, Peterson BS, Lieberman JA, Neuropsychopharmacology, 2013 · PMID 23325325

  The magnetic resonance spectroscopy substudy of the schizophrenia trial. Notable as a rare attempt to demonstrate target engagement in living human brain rather than inferring it.

• Addressing Alzheimer's disease tangles: from NAP to AL-108 Gozes I, Stewart A, Morimoto B, Fox A, Sutherland K, Schmeche D, Current Alzheimer Research, 2009 · PMID 19874271

  The development-stage review describing the phase 2a amnestic MCI result that showed a positive impact on memory function. This is the source of the one encouraging human signal davunetide ever produced, written by the developing group before the later failures.