F tier · safety concern
No human being has ever been given P021: a ClinicalTrials.gov search returns no study of the compound under any name, and the Alzheimer's Drug Discovery Foundation's own 2025 assessment states flatly that safety in humans has not been assessed. Essentially every in vivo paper is co-authored by the originating laboratory, and the one attempt in a disease model chosen by an outside group (Cdkl5 knockout mice, 2024) failed to restore memory, social behaviour or motor function and failed to raise BDNF at all.
p21
P021 is a synthetic tetrapeptide, Ac-Asp-Gly-Gly-Leu-, capped at the C-terminus with the unnatural, rigid, highly lipophilic amino acid 3-aminoadamantane-1-carboxylic acid, designed to resist exopeptidases and cross the blood-brain barrier. It was designed at the New York State Institute for Basic Research in Developmental Disabilities by Khalid Iqbal, Inge Grundke-Iqbal and Herbert Mösler as a small-molecule mimetic of ciliary neurotrophic factor, and in rodents it increases dentate-gyrus neurogenesis and BDNF expression and lowers tau hyperphosphorylation. It has never been administered to a human being, has no registered clinical trial, and has no marketing authorisation anywhere.
// we supply this one
available as a research reagent
≥ 98% HPLC · lyophilised powder · batch certificate published. Grade F above is our own and is not adjusted because we stock it.
the explanation
P21 is a very short lab-made peptide based on a fragment of a natural nerve-growth protein, with a bulky cage-shaped chemical group stuck on the end so it survives longer in the body and gets into the brain. In mice and rats, especially genetically engineered Alzheimer's mice, it increased the birth of new brain cells and improved memory tests. No human has ever taken it in a study, so nobody knows whether it works or is safe in people.
regulatory status
Unapproved worldwide; preclinical research chemical only
P021 holds no marketing authorisation in the UK, EU, US or any other jurisdiction, is not a prescription-only medicine because it is not a licensed medicine at all, and is not a controlled drug. There is no MHRA-authorised product containing it, and a ClinicalTrials.gov search returns no registered study, so it has never been an investigational medicinal product in a UK trial; anything sold here is an unlicensed substance supplied strictly as a laboratory research reagent for in vitro and non-human use. The composition of matter is claimed in Research Foundation for Mental Hygiene patent US8592374B2 (inventors Herbert Mösler, Khalid Iqbal and Inge Grundke-Iqbal; priority 16 March 2007, granted 26 November 2013), and the Alzheimer's Drug Discovery Foundation records Phanes Biotech as the developer. It is not a lawful dietary-supplement ingredient anywhere — no New Dietary Ingredient notification exists and it appears on no compounding bulks list — and because no government health authority has approved it for human therapeutic use it falls squarely within section S0, Non-Approved Substances, of the WADA Prohibited List, which prohibits any such substance at all times.
how it works · proposed mechanism
P021 is described as a CNTF mimetic, but its proposed mechanism is not agonism at the CNTF receptor complex. The Alzheimer's Drug Discovery Foundation's assessment describes it as competitively inhibiting leukaemia inhibitory factor signalling, which normally restrains the formation of neural progenitors from stem cells, thereby releasing a brake on neurogenesis rather than pushing a growth-factor accelerator. Downstream, the reported signature is increased BDNF with activation of JAK/STAT3, MAPK and Akt pathways.
Derived from CNTF but not a CNTF receptor agonist
The pharmacophore corresponds to CNTF residues 148-151 within Peptide 6 (CNTF 146-156, VGDGGLFEKKL), the 11-mer from which P021 was cut down. Because the proposed action is competition with leukaemia inhibitory factor signalling rather than engagement of the CNTFR-alpha/LIFR-beta/gp130 complex, the clinical history of recombinant CNTF cannot be read across to P021 in either direction. That is important, because recombinant CNTF failed in amyotrophic lateral sclerosis trials with anorexia, weight and muscle loss, hyperalgesia and cramps — effects P021 is reported not to reproduce in rodents.
The adamantyl cap is pharmacokinetics, not pharmacology
3-aminoadamantane-1-carboxylic acid was attached C-terminally specifically because it is rigid, bulky and highly lipophilic, blocking exopeptidase attack and raising membrane permeability. The reported consequences are stability above 90% in artificial gastric juice, complete stability in artificial intestinal fluid, and a murine plasma half-life above 3 hours. Notably, no study has directly assayed how much P021 actually reaches the brain in any species, so blood-brain barrier penetration remains an inference from design intent rather than a measurement.
BDNF-TrkB-CREB induction and neurogenesis
In 3xTg-AD mice, treated animals showed roughly four-fold higher numbers of DCX-positive and Ki-67-positive cells in the dentate gyrus than vehicle controls, alongside restored BDNF protein and increased GluN2A, GluA1 and GluA2/3 subunits. In aged non-transgenic rats, BDNF, TrkB and the pCREB/CREB ratio rose and Morris water maze latency improved. The independent-model exception is instructive: in Cdkl5 knockout mice, BDNF was not increased at all, and the behavioural benefit was limited to reduced muscular rigidity.
Downstream effects on tau and amyloid are secondary readouts
Reported reductions in tau hyperphosphorylation of roughly 50% after prolonged treatment, and lower soluble Abeta40 and Abeta42, are framed as consequences of restored plasticity rather than direct anti-tau or anti-amyloid action; P021 is not an aggregation inhibitor or a secretase modulator. Total tau also fell in hippocampus and cerebrospinal fluid of aged rats, though cerebrospinal fluid Abeta did not change. All of these are endpoints in rodent brains, and none has ever been measured in a human given the compound.
what’s reported
evidence shape
7 sourcesThe composition of the cited literature by study type. This describes the shape of the evidence, not its quality — and it does not by itself set the grade.
⚠ the catch
P021 is sold for human self-administration on the strength of a rodent literature in which the human column is empty. ClinicalTrials.gov returns no registered study under P021, P-021 or any variant, and the Alzheimer's Drug Discovery Foundation's May 2025 assessment states in terms that no studies have tested P021 in humans and that safety in humans has not been assessed. The evidence base is also not independent in the way the marketing implies: the productive papers run from Li 2010 through Kazim, Khatoon, Bolognin, Baazaoui and Wei, all with Khalid Iqbal as senior or corresponding author, and even the 2024 study from Elisabetta Ciani's group at the University of Bologna — the closest thing to an outside test — lists Iqbal as a co-author. That study is the most informative single result in the file, and it is negative: in Cdkl5 knockout mice, chronic P021 did not significantly restore memory, social behaviour or motor function, and BDNF protein, the compound's proposed proximate mediator, was not increased. Finally, most of the headline data come from a triple-transgenic Alzheimer's mouse or from 19-24-month-old rats, and even there the aged rats treated with P021 did not reach the performance of young animals — so the strongest claim the literature supports is partial rescue of an artificial deficit in a rodent, not enhancement of a healthy adult human brain.
key published findings
- The molecule is chemically well defined: Ac-DGGL-AG-NH2 as claimed in patent US8592374B2, where AG is 3-aminoadamantane-1-carboxylic acid; C27H42N6O8, 578.66 Da, CAS 1246751-68-7, consistent with the 578.3 monoisotopic mass quoted by the Alzheimer's Drug Discovery Foundation.
- There are zero registered clinical trials and zero humans ever dosed; the only human-relevant statement in the independent 2025 assessment is that safety in humans has not been assessed.
- In 3xTg-AD mice, prolonged treatment produced roughly four-fold higher DCX+ and Ki-67+ dentate-gyrus cell counts, roughly 50% lower tau hyperphosphorylation, and 87% survival at week 71 versus 41% on vehicle.
- In aged non-transgenic rats (19-24 months), treatment reduced hippocampal and cerebrospinal-fluid total tau, raised BDNF, TrkB and pCREB/CREB, and improved Morris water maze latency — but treated aged rats did not reach young-animal performance, and cerebrospinal-fluid Abeta did not change.
- The one study run primarily by an outside group (Mottolese et al., Journal of Neurodevelopmental Disorders 2024, University of Bologna, with Iqbal as co-author) was largely negative in Cdkl5 knockout mice: no significant restoration of memory, social behaviour or motor function, and no increase in BDNF.
- Rodent tolerability is reassuring on its own terms — up to 18 months of treatment without weight loss, tumours or signs of pain, and no anorectic effect unlike recombinant CNTF — but blood-brain-barrier penetration has never been directly measured in any species.
limitations of the evidence
- Zero human exposure: no phase 1, no pharmacokinetics, no bioavailability, no immunogenicity and no toxicology package in humans.
- The in vivo literature is not independent — Khalid Iqbal is an author on effectively every in vivo P021 paper, including the negative Cdkl5 study, so no laboratory unconnected to the originator has replicated the efficacy findings.
- The strongest efficacy data come from transgenic disease models (3xTg-AD, Ts65Dn) or from aged rats, not from healthy adult animals; the single normal-mouse study was in 8-10-month-old C57Bl/6 mice using implanted depot pellets, a route with no relevance to how the compound is sold.
- No molecular target has been identified and no binding affinity published; the proposed mechanism, competition with leukaemia inhibitory factor signalling, is inferred from downstream markers rather than measured receptor pharmacology.
- Brain exposure has never been assayed directly, so the entire rationale for the adamantyl cap rests on design intent plus indirect behavioural readouts.
- The name is dangerously ambiguous: 'P21' also denotes the cyclin-dependent kinase inhibitor protein p21/CDKN1A and unrelated PCNA-blocking p21 peptides, and at least one chemical catalogue attaches that description to this CAS number while listing a molecular formula one CH2 short of the patented structure.
documented safety signals
- No human safety data exist at all — no adverse-event profile, no tolerated exposure range, no monitoring precedent, and no published genotoxicity, carcinogenicity or reproductive-toxicity study in any species.
- The rodent record is reassuring but limited to the endpoints looked at: no weight loss, tumours, pain signs, grooming or posture changes over up to 18 months, and no anxiety effect; body weight increased in both 3xTg-AD and wild-type mice.
- Class-level context cuts the other way: full-length recombinant CNTF caused anorexia, skeletal muscle loss, hyperalgesia, severe cramps and muscle pain in amyotrophic lateral sclerosis trials. P021 is a different molecule with a different proposed mechanism and is reported not to be anorectic, but CNTF is the protein whose name is used to sell it.
- Chronically stimulating adult neurogenesis and neurotrophic signalling in a healthy human brain has no characterised long-term safety profile, and the theoretical proliferative concern has never been formally studied for this compound.
- Absence of reported harm in humans reflects absence of human study, not evidence of safety.
identity
| full name | P021 — N-acetyl-L-aspartyl-glycyl-glycyl-L-leucyl-N-[3-(aminocarbonyl)adamantan-1-yl]amide, an adamantylated tetrapeptide derived from ciliary neurotrophic factor residues 148-151 |
| category | Cognitive |
| modality | peptide |
| formula | C27H42N6O8 |
| molar mass | 578.66 g/mol |
| cas | 1246751-68-7 |
| half-life | Plasma half-life reported as over 3 hours in mice; no human pharmacokinetic data of any kind has been published, and no study has directly measured how much P021 reaches the brain. |
| sequence | Ac-Asp-Gly-Gly-Leu-NH-[3-carbamoyladamantan-1-yl] (written in the originating patent as Ac-DGGL-AG-NH2, SEQ ID NO:18, where the C-terminal residue is the unnatural amino acid 3-aminoadamantane-1-carboxylic acid in its amide form). The parent peptide is Peptide 6, CNTF 146-156, VGDGGLFEKKL. |
laboratory handling
| storage | Lyophilised powder is stored desiccated at -20 °C and protected from light. Once reconstituted, laboratory practice is refrigeration at 2-8 °C for short-term use, or aliquoting and freezing to avoid repeated freeze-thaw cycles. Handling information only; this material has no approved human use. |
| solubility | Short hydrophilic peptide, readily soluble in sterile or bacteriostatic water and in dilute aqueous buffer. Lot-specific solubility is stated on the certificate of analysis. |
| co-studied with | There are no human combination data of any kind, and no animal study has tested P021 alongside another nootropic, cholinesterase inhibitor or anti-amyloid agent. Any stacking claim is vendor invention with no experimental basis. |
Handling information describes laboratory practice for a research reagent. It is not a preparation guide for use in humans.
the receipts
7 citedUnited States Patent and Trademark Office; assignee Research Foundation for Mental Hygiene Inc · 2013 · official
U.S. National Library of Medicine, ClinicalTrials.gov · 2026 · official
World Anti-Doping Agency · 2025 · official
Alzheimer's Drug Discovery Foundation (Cognitive Vitality) · 2025 · review
Journal of Neurodevelopmental Disorders 16:65 · 2024 · preclinical
Neurobiology of Disease (PMID 25046994) · 2014 · preclinical
ChemicalBook chemical registry record · 2026 · reference
others in Cognitive