P-21

P021 (Cntf-derived neurogenic peptide)

P-21 (P021 (Cntf-derived neurogenic peptide)) is a cognitive & mood research compound. Small neurogenic peptide derived from CNTF — enhances BDNF signaling and adult hippocampal neurogenesis.

How P-21 works

Small neurogenic peptide derived from CNTF — enhances BDNF signaling and adult hippocampal neurogenesis.

Proposed benefits

Researched for focus, memory, neuroprotection, mood and stress resilience.

Where to get P-21

The evidence for P-21

Graded by what exists behind each claim.

Human clinical evidence

• No human trials: development stopped at preclinical, so the ceiling on any claim here is a rodent model — and these transfer poorly.

📊 Correlative data

• Limited community use. Cam supplied the dosing in this Vault directly, which is worth stating — the published record has none, so what exists here comes from practice rather than a paper.
• Beyond that the record is self-reported: community dosing logs are real information about tolerability and almost none about efficacy.

🧪 How the mechanism reads

• A peptidergic analog derived from ciliary neurotrophic factor, developed at the NIH. Rodent work reports increased neurogenesis in the hippocampus and improved cognition in Alzheimer's models, without CNTF's weight-loss side effect.
• Removing the appetite effect while keeping the neurotrophic action is the design goal, and it is why P-21 exists at all — CNTF itself failed clinically partly because of that side effect.

Why an empty tier is not a verdict → · What community dosing logs are worth →

How to read these tiers: they say how much human evidence exists, not how well something works — and ✗ flags harm, never a disappointing trial. How the evidence tiers work →

What P-21 actually does

Write the structure out, because almost nobody does. The primary literature gives this compound as Ac-DGGL(A)G-NH2 Kazim 2014; the paper that introduced the series writes it as Ac-DGGLAG-NH2 and calls it P21 Li 2010. Read that as a four-residue peptide — Asp-Gly-Gly-Leu — carrying an acetyl group on the N-terminus, an amide on the C-terminus, and an adamantylated glycine at the C-terminal end. The reason for the adamantane is stated explicitly by the group that built it: it was added to the C-terminal to increase blood-brain barrier permeability and to decrease degradation by exopeptidases, in order to make the peptide druggable Baazaoui 2017.

That is three separate protease defenses on a four-residue molecule, and each blocks a different enzyme class. The N-terminal acetyl removes the free amine that aminopeptidases require. The C-terminal amide removes the free carboxylate that carboxypeptidases require. The adamantane — a rigid tricyclic C10H16 cage of 136.2 Da, the same cage carried by amantadine and memantine — adds steric bulk that exopeptidases cannot get around, and adds lipophilicity to a peptide that would otherwise be far too polar to cross a membrane. This is a designed molecule pretending to be a fragment, and the design is the interesting part.

Where the four residues came from, and why four. The sequence was obtained by epitope mapping the most active region of ciliary neurotrophic factor Baazaoui 2017. CNTF itself failed clinically, and it failed for reasons that are the entire design brief: anorexia, skeletal muscle loss, hyperalgesia, cramps and muscle pain Blanchard 2010. The sibling peptide in the same program, Peptide 6c, is GDDL — CNTF residues 147–150 — and it reproduced the neurogenic effect without weight loss or any other apparent side effect Blanchard 2010. So the family logic is: keep the four residues that carry neurogenesis, discard the protein that carries the cachexia.

What it does downstream, and what it does not have. In the triple transgenic Alzheimer mouse, chronic treatment reduced abnormal hyperphosphorylation of tau at the major neurofibrillary-pathology sites, lowered soluble amyloid-beta with a trend toward fewer plaques in CA1, raised BDNF expression, and reduced GSK-3beta activity, which the authors propose as the link between the two Kazim 2014. Note their own careful reading of the amyloid result: it was consistent with reduced generation, not increased clearance. And note what is absent from this entire literature — no receptor. Nothing published names a receptor this peptide binds. It is a neurotrophic-factor mimetic without a demonstrated molecular target Kazim 2016, which is a materially weaker position than ‘CNTF analog’ makes it sound.

Cell, rodent, human — and where it stops

The cell step is missing, and that is itself worth noticing. This series was characterized in living animals from the start. There is no cell-culture demonstration of a target engagement, no binding curve and no dose-response in neurons in the papers that define it Li 2010 Kazim 2014 Baazaoui 2017. Most compounds on this site have cell data and no animal data; this one has the reverse, and the missing layer is the one that would name a receptor.

In rodents, and the exposures are extraordinary. Normal adult C57Bl/6 mice given the peptide peripherally showed improved learning and both short-term and spatial reference memory, with increased neurogenesis and maturation of newborn neurons in the granular cell layer and subgranular zone of the dentate gyrus Li 2010. The sibling tetrapeptide was delivered by subcutaneous extended-release depot pellet for 30 days of continuous dosing, with stereological counts of doublecortin and NeuN cells and BrdU uptake, higher MAP2a,b and synaptophysin, higher basal c-fos, better spatial memory and no weight loss Blanchard 2010. In the flagship experiment, 3xTg-AD and wild-type female mice were treated in the diet for 12 months, starting at 9–10 months of age Kazim 2014; in the prevention study the diet ran from 3 months to 21 months, beginning 6–9 months before any overt pathology Baazaoui 2017. Across neurogenic regions in AbetaPP transgenic mice, the CNTF-derived peptides raised doublecortin-positive and BrdU-positive cell counts and PCNA-positive cells, while cerebrolysin in the same comparison reduced TUNEL staining without altering PCNA — the authors' point being that the two work by different modes, proliferation against apoptosis Rockenstein 2011.

In humans: nothing. No trial, no pharmacokinetics, no tolerability series, no case report.

The obstacles, and they are unusually concrete.

• Twelve and eighteen months of continuous dietary dosing is not a four-week course. The two results that carry this compound's reputation are lifetime-scale oral exposures in mice Kazim 2014 Baazaoui 2017. A 4–8 week subcutaneous or intranasal protocol is not a shortened version of that experiment; it is a different one.
• The strongest result is a prevention result. Dosing began 6–9 months before pathology appeared Baazaoui 2017. Reading a prevention study in a genetically loaded mouse as a treatment for an adult who already has symptoms is a category change, not an extrapolation.
• The endpoints are unmeasurable outside a mouse. Tau phosphorylation at specific epitopes, soluble Abeta, dentate doublecortin counts, synaptophysin density. There is no consumer version of any of these.
• No mg/kg dose exists to scale from. Diet dosing means intake tracks food intake, and the abstracts state the design rather than a milligram-per-kilogram figure. The 200 µg–4 mg range used in practice is therefore not derived from the published work at all — it cannot be, because the published work does not contain the number.
• Identity is harder here than for any other peptide on this site. An adamantylated, doubly-capped tetrapeptide is a more demanding synthesis than a plain peptide, and a molecule sold as P-21 without the adamantane cage would, by its designers' own stated reasoning, have worse brain penetration and faster exopeptidase degradation Baazaoui 2017. No vendor publishes an analysis showing the cage is present.

P-21 pharmacokinetics — how much of it actually gets in

What degrades it, and what was done about it. Exopeptidases — aminopeptidases at one end, carboxypeptidases at the other — and the molecule is built specifically against them: acetyl cap, C-terminal amide, and an adamantylated glycine added to reduce exopeptidase degradation and raise blood-brain barrier permeability Baazaoui 2017. Whether that is enough is a separate question from whether it was attempted, and no measured stability figure — not a serum incubation half-life, not a brain-to-plasma ratio — has been published for it.

The oral evidence is real and the oral number is missing. This is one of very few peptides on this site with a positive oral result: 12 months of dietary administration produced central effects in mice Kazim 2014, repeated over 18 months in a second study Baazaoui 2017. That is stronger evidence of oral activity than almost anything else in this catalog has. What does not exist is a percentage: no oral bioavailability figure, no plasma concentration, no CSF measurement, in any species. The compound is known to work by mouth and not known to reach any particular concentration by mouth.

The shape of the exposure is the opposite of the usual story. For most peptides the interesting fact is a short plasma residence against a longer effect. Here every published effect came from continuous exposure over months — food, or a 30-day subcutaneous depot pellet Blanchard 2010 — and no result has ever been produced by intermittent bolus dosing. A daily subcutaneous injection is a peak-and-trough pattern that no successful experiment with this molecule has used.

The injectable and intranasal comparators. The protocols in circulation are subcutaneous or intranasal. The published chronic work is oral, and the one subcutaneous route used in the family was a slow-release implant rather than an injection. Nobody has compared routes for this molecule, so the case for injecting it rests on convention rather than on any measurement.

What would have to be true, and how you would know it was not

Four predictions. The first comes straight out of the reason the molecule exists.

1. Body weight and grip strength must not fall — and if they do, the design has failed in exactly the way its authors were trying to avoid. CNTF's clinical failure was anorexia and skeletal muscle loss Blanchard 2010, and the whole claim for the tetrapeptide is that it kept the neurotrophic action and dropped that liability. Weigh weekly. Measure grip strength with a dynamometer at baseline and at 8 weeks, three attempts per hand, best of three. Stable weight and stable grip is the mechanism working as designed; a quiet 2–3 kg loss with reduced grip is the parent protein's side-effect profile reappearing, and nobody is looking for it.

2. MoCA at baseline and 12 weeks, with the caveat stated up front. If a person is taking this for cognition, the endpoint should be a scored instrument rather than a feeling. MoCA is free and takes ten minutes. It also has a practice effect and a ceiling in healthy adults, so a rise of one point is noise and should be read as such; use different versions and the same time of day.

3. Nothing in blood will move, and that is the prediction against the product. No published mechanism connects this molecule to any standard panel. Predict an unchanged CBC and unchanged morning cortisol after 8 weeks. That is not a reason to skip the draw — it is the control that tells you a reported effect is not coming from something else that changed at the same time.

4. The timescale prediction, which most courses will fail. Every positive published result used months of continuous exposure Kazim 2014 Baazaoui 2017 Blanchard 2010. A four-week course is shorter than any experiment that has ever shown an effect with this compound. So a strong subjective effect at week 2 is not the published mechanism arriving early; it is a different phenomenon, and the honest reading is expectancy until a scored instrument says otherwise.

What nobody has tested yet

Six experiments. The first one is the reason this page exists.

1. Nobody has published an analysis of a commercial vial confirming the adamantane is there. The cage is the compound's entire brain-penetration and stability strategy Baazaoui 2017, it adds 136 Da that a mass spectrometer sees instantly, and a plain Ac-DGGLAG-NH2 without it would look superficially similar on a certificate of analysis that only reports purity. One LC-MS run per batch would settle whether buyers are getting the designed molecule or its undecorated backbone. Nobody does it.

2. Nobody has measured a concentration of this peptide in plasma or CSF in any species. Not once, in fifteen years.

3. Nobody has compared oral with subcutaneous or intranasal delivery. Every efficacy result is oral; every retail protocol is not. This is the single largest unexamined gap between what was studied and what is sold.

4. Nobody has run it in a normally ageing human, even though the closest published analog is exactly that. The first result in the series was in normal adult mice, not a disease model Li 2010. That is the population buying it. A 12-week study in healthy adults over 50, with a scored cognitive battery, would be both the most obvious study to run and the first one in humans.

5. Nobody has checked whether the muscle-sparing claim holds outside a mouse. Absence of weight loss was demonstrated for the sibling peptide over 30 days in mice Blanchard 2010. Whether an adamantylated CNTF fragment spares muscle in a person over months has never been looked at, and it is measurable with a scale and a dynamometer.

6. There is no dose-response of any kind. No minimum effective dose, no ceiling, no toxicity threshold, in any species, for P021 specifically. Every number in circulation for this compound is somebody's guess dressed as a protocol.

P-21 — its own safety story, not its class's

The class block on this page warns about cholinergics, dopaminergics and glutamatergics. This compound is none of those, and it has four risks of its own that the block cannot raise.

1. The parent protein's side-effect profile is the map of what to watch. CNTF produced anorexia, skeletal muscle loss, hyperalgesia, cramps and muscle pain in clinical use Blanchard 2010. The tetrapeptide was built to shed those, and in mice over 30 days it appears to have Blanchard 2010. Nobody has checked in a person. Appetite, weight and muscle cramp are therefore the specific things to log on this compound — not because they are known side effects, but because they are the known side effects of the thing it was derived from.

2. Neurogenesis is a growth signal, and the tumor question has never been asked. The published effects include increased BrdU uptake and increased PCNA-positive cells — markers of proliferation — sustained across months of dosing Rockenstein 2011 Kazim 2014. No study in this series carried a tumor endpoint. This is not an alarm: proliferation of neural progenitors in the dentate gyrus is not carcinogenesis, and there is no report of a tumor with any of these peptides. It is the specific unstudied question the mechanism raises, and it does not appear on any page selling this compound.

3. Part of this molecule is not a peptide, and that part has relatives with their own pharmacology. Adamantane is the shared core of amantadine and memantine, both CNS-active drugs. This is extrapolation and should be read as such: nothing published says what the adamantylated glycine does once the peptide bond in front of it is eventually cleaved, and no metabolite of P021 has ever been identified. A peptide that leaves a lipophilic cage behind is not the same disposal problem as a peptide that resolves into amino acids.

4. What zero human exposure means. There is no published human of any age who has taken this. Every dose, every cycle length and every tolerability claim in circulation is unsourced, and the compound's own designers describe it as a candidate rather than a treatment Kazim 2016.

Sources read for this page

• Li B, Wanka L, et al. Neurotrophic peptides incorporating adamantane improve learning and memory, promote neurogenesis and synaptic plasticity in mice. FEBS Letters 2010 · PMID 20600002
• Blanchard J, et al. Beneficial effect of a CNTF tetrapeptide on adult hippocampal neurogenesis, neuronal plasticity, and spatial memory in mice. Journal of Alzheimers Disease 2010 · PMID 20952820
• Rockenstein E, et al. Regional comparison of the neurogenic effects of CNTF-derived peptides and cerebrolysin in AbetaPP transgenic mice. Journal of Alzheimers Disease 2011 · PMID 21860085
• Kazim SF, et al. Disease modifying effect of chronic oral treatment with a neurotrophic peptidergic compound in a triple transgenic mouse model of Alzheimer's disease. Neurobiology of Disease 2014 · PMID 25046994
• Baazaoui N, Iqbal K. Prevention of dendritic and synaptic deficits and cognitive impairment with a neurotrophic compound. Alzheimers Research and Therapy 2017 · PMID 28655344
• Kazim SF, Iqbal K. Neurotrophic factor small-molecule mimetics mediated neuroregeneration and synaptic repair: emerging therapeutic modality for Alzheimer's disease. Molecular Neurodegeneration 2016 · PMID 27400746

P-21 — safety, predicted from mechanism

P-21 — interference & stacking

Bloodwork to run alongside P-21

Baseline first, then again at 8–12 weeks.

The Brain Fog & Cognition panel covers these in one order — 12 markers, $233.06 with the discount applied.

Check results you already have → · All 103 markers A–Z

P-21 — frequently asked questions

What P-21 is used for

P-21 appears under 1 goal in the goal router.

Related Cognitive & Mood compounds

← Explore the full Protocol Vault

↑ Back to on this page