Region — US. Neutral scientific reference — not offered for sale here.
Cart · 0
Set region
Vallydia
ExploreShopToolsJournalTrust
Explore  /  P21 (P021)
Research reference — not for sale

P21 (P021)

F
lead outcome
Human evidence for any indication
grades vary by outcome ↓
Peptide
also called — P021 · Peptide 21 · P21 Adamantane · GLXC-21260 · Ac-DGGL-adamantylglycine-NH2 · CNTF-derived peptide mimetic
neurogenesis (preclinical)BDNF signalling (proposed)tau pathology (preclinical)CNTF mimetic

Status: reference entry — not for sale. No dosing, reconstitution, or administration is published (intentional). Neutral scientific reference only. P21 has never been studied in humans; any dose figure encountered online is extrapolated from rodent feeding studies, not a validated protocol.

In brief

P21 (also written P021) is a small synthetic peptide reverse-engineered from the most active region of ciliary neurotrophic factor (CNTF) — residues 148-151, the tetrapeptide DGGL — with an adamantylated glycine added for lipophilicity, protease resistance and blood-brain barrier penetration. The design goal was to capture CNTF's pro-neurogenic effects while avoiding what killed CNTF clinically: hypothalamic STAT3-driven anorexia and antibody formation. In rodents that goal appears met — plasma half-life over three hours versus about three minutes for recombinant CNTF, oral bioavailability, and a consistent record across Alzheimer's, ageing and Down syndrome models of increased hippocampal neurogenesis, raised BDNF, reduced tau phosphorylation and improved maze learning. Two caveats dominate. The evidence is almost entirely from one laboratory (Khalid Iqbal's group), and the single prominent independent test — in CDKL5 deficiency disorder, 2024 — failed to reproduce the BDNF increase and found limited behavioural benefit. And there are no human studies of any kind: no trial, no pharmacokinetics, no safety data, nothing registered. Clean preclinical work, unusually concentrated, entirely untested in people. Not to be confused with p21/CIP1/WAF1, the unrelated cell-cycle inhibitor protein.

Legal standing, by region
European Union
Not approved

No marketing authorisation; not an authorised food supplement or cosmetic ingredient. Sold only as a non-medicinal research chemical, which does not authorise human use.

United Kingdom
Not approved

No marketing authorisation; research chemical status only.

United States · your region
Not FDA-approved

Not approved for human or veterinary use, not a dietary supplement, and not intended for human consumption — a status the research-chemical suppliers themselves state explicitly. No clinical trials are registered.

International
Not approved anywhere

Approved in no jurisdiction for any indication.

Evidence, by outcome
How we grade →

An honest grade per outcome — drawn from the evidence, not any catalogue. Hype and undemonstrated marketing claims grade low.

OutcomeEvidence base · effectGrade
Human evidence for any indication
Any "research dose" circulating for P21 is extrapolated from rodent oral feeding protocols (typically 60 nmol/g of feed in the originating lab's work). There is no validated human protocol, because there has never been a human study.
None. No clinical trial of P21 of any phase has been published or registered. There is no human pharmacokinetic, dose-response, efficacy or safety data. The Alzheimer's Drug Discovery Foundation's Cognitive Vitality review states directly that no studies have tested P021 in humans. · Nothing demonstrated in humans.
F
Cognition, neurogenesis and tau pathology in mouse models
Two structural limitations. First, **the literature is concentrated almost entirely in the originating laboratory** (Iqbal, Kazim, Baazaoui, Bolognin and colleagues at the New York State Institute for Basic Research) — consistent direction from one group is weaker evidence than convergent findings from many. Second, Alzheimer's mouse models have historically translated to human benefit extremely poorly; a compound that rescues 3xTg-AD mice has cleared a bar that dozens of failed drugs also cleared.
A substantial and directionally consistent rodent record. Kazim SF, et al. Neurobiol Dis. 2014 (PMID 25046994) — oral P021 in 3xTg-AD mice produced Morris water maze performance comparable to wild-type, significantly reduced phosphorylated tau at multiple epitopes, increased BDNF, and restored Ki-67+ and DCX+ dentate gyrus cells to wild-type levels; treatment begun at 12 months, after pathology onset, still rescued deficits. Bolognin S, et al. Neurobiol Aging. 2014 (PMID 24702821) — rescue of cognitive-ageing decline in aged rats. Baazaoui N, Iqbal K. Alzheimers Res Ther. 2017;9:45 — preventive treatment from 3 months for 18 months increased DCX+ and Ki-67+ cells roughly fourfold with no tumours or weight loss. Khatoon S, et al. J Alzheimers Dis. 2015 (PMID 26401692) — BBB permeability and reduced total tau in aged rats. Kazim et al. (2017) — efficacy in the Ts65Dn Down syndrome model. · Consistent improvements in neurogenesis markers, BDNF expression, tau phosphorylation and maze learning across several rodent models.
B
Independent replication
One failed replication is not a refutation, and a different disease model may legitimately behave differently. But it is the only independent data point available, and it is negative — which matters a great deal for a compound whose entire case rests on one lab's work. Worth stating alongside: unlike the neighbouring compound Dihexa, whose foundational mechanism papers were retracted in 2025, no retractions or expressions of concern have been identified in the P21 literature. The problem here is not integrity; it is concentration and absence of human data.
Sparse. The one prominent independent test identified — Bhattacharyya et al. (2024), in CDKL5 deficiency disorder — **failed to replicate the BDNF increase in vivo** and reported limited behavioural benefit. · The central mechanistic claim did not reproduce outside the originating programme in the one published attempt.
D
Mechanism: LIF inhibition and BDNF/TrkB signalling
Elegant on paper, but demonstrated within the same programme that demonstrated the efficacy, and the BDNF step specifically failed to replicate in the one independent study. The mechanism is a hypothesis with supporting data, not an established fact.
Proposed mechanism is **competitive inhibition of leukaemia inhibitory factor (LIF) signalling**, which normally suppresses neurogenesis — so P21 acts by removing an inhibitory constraint rather than by directly agonising the CNTF receptor. Downstream, increased BDNF and TrkB signalling through PI3K/Akt leads to GSK-3β inactivation, the proposed route to reduced tau hyperphosphorylation. · Coherent, specifically articulated mechanism that explains the observed effects and why P21 avoids CNTF's systemic problems.
C
Design advantage over full-length CNTF
A better-designed molecule is not a proven one. And the specific liability that ended CNTF development — **antibody formation** — has been flagged as a possible adverse event for P21 too (Lozupone et al., 2023), though none has been reported to date in animals. With no human exposure, immunogenicity in people is untested.
Full-length CNTF failed clinically — a trial in amyotrophic lateral sclerosis showed no benefit on pulmonary function, motility or survival, with substantial weight loss in some patients; Axokine (recombinant CNTF for obesity) was derailed by anorexia and antibody formation. P21's design deliberately isolates the neurogenic activity (CNTF residues 148-151) and adds an adamantyl group for lipophilicity and exopeptidase resistance. Reported result: plasma half-life over 3 hours in mice versus roughly 2.9 minutes for recombinant CNTF, ~90% gastric and >95% intestinal stability, oral bioavailability, BBB penetration, and no hypothalamic STAT3-driven anorexia. · The engineering objective appears to have been met in rodents.
B
Human safety
Reassuring rodent safety across long treatment is genuinely favourable, but it is rodent safety. Potential immunogenicity is specifically flagged in the review literature. Material sold online carries the usual separate risks of unverified identity and purity — relevant here because the adamantylated structure is not trivial to synthesise correctly.
No human data. In rodents, chronic treatment over 18 months produced no tumours, weight loss or observed adverse effects, and mortality in 3xTg-AD mice was reduced. · Unknown in humans.
—
Disclosure

Vallydia sells its own cosmetic serums, and some ingredients graded here belong to the same categories as those products. Grades are drawn from the published evidence by the method we publish, and applied to our own ingredients on the same terms — our copper-peptide serum is graded no more kindly than the peptides it competes with. We disclose the interest so you can weigh it.

Trade names shown here (for example Matrixyl, Argireline, Syn-Ake) are the property of their respective owners and are used only to identify the ingredient under discussion. Their appearance implies no affiliation with, or endorsement by, the proprietor.

Evidence changes

Get notified if this changes

We track new research and update these grades as the evidence moves. Leave your email and we'll tell you if the evidence for P21 (P021) changes - nothing else.

One ingredient, one purpose. No marketing, no sharing. Unsubscribe anytime. See our privacy policy.

Identity

P21 (also written P021, and sometimes sold as "P21 Adamantane") is a small synthetic peptide derived from ciliary neurotrophic factor (CNTF). Its core is the tetrapeptide DGGL, corresponding to residues 148-151 of human CNTF — identified through epitope mapping with neutralising antibodies as the biologically active region — with the C-terminal glycine bearing an adamantane group. That bulky tricyclic modification raises lipophilicity and blocks exopeptidase degradation, which is what gives the molecule oral bioavailability and blood-brain barrier penetration it would otherwise lack. The parent compound in the same programme was an 11-residue peptide (Peptide 6, Ac-VGDGGLFEKKL-NH₂); P21 is its compact, stabilised successor.

Important disambiguation: P21 the CNTF-derived peptide is entirely unrelated to p21^Cip1/Waf1 (gene CDKN1A), the cyclin-dependent kinase inhibitor studied extensively in cancer biology. They share nothing but a name, and conflating them is the most common error in material written about this compound.

Development & history

  • Background: Full-length CNTF looked promising as a neurotrophic therapy and failed in humans. A clinical trial in amyotrophic lateral sclerosis showed no benefit on pulmonary function, motility or survival, with substantial weight loss in some patients; Axokine, recombinant CNTF developed for obesity, was derailed by anorexia and by antibody formation against the protein. The anorectic effect traced to hypothalamic STAT3 activation.
  • 2010: The laboratory of Khalid Iqbal at the New York State Institute for Basic Research in Developmental Disabilities, with Inge Grundke-Iqbal and colleagues, published the founding work: adamantane-incorporating neurotrophic peptides improving learning, memory, neurogenesis and synaptic plasticity in mice (FEBS Letters). The design intent was explicit — keep CNTF's neurogenic activity, discard its systemic liabilities.
  • 2014: Two key rodent papers. Kazim et al. (Neurobiol Dis) showed chronic oral P021 in 3xTg-AD mice normalised Morris water maze performance, reduced phospho-tau at multiple epitopes, raised BDNF and restored dentate gyrus neurogenesis markers — including when treatment began after pathology onset. Bolognin et al. (Neurobiol Aging) showed rescue of cognitive-ageing decline in aged rats.
  • 2015: Khatoon et al. characterised BBB permeability and reduced total tau in aged rats (J Alzheimers Dis).
  • 2017: Baazaoui & Iqbal (Alzheimers Res Ther) reported preventive treatment from 3 to 21 months increasing neurogenesis markers roughly fourfold, with no tumours or weight loss, and reduced mortality in treated 3xTg-AD mice. Separate work extended to the Ts65Dn Down syndrome model.
  • 2024: Bhattacharyya et al., studying CDKL5 deficiency disorder, reported failure to replicate the in-vivo BDNF increase and limited behavioural benefit — the only prominent independent test.
  • To date: No clinical trial has been registered or published. No human pharmacokinetic, safety or efficacy data exist.

Mechanism (as proposed)

P21's proposed mechanism is a disinhibition rather than a direct agonism, which is the conceptually interesting part. Leukaemia inhibitory factor (LIF) signalling normally suppresses neurogenesis; P21 is reported to competitively inhibit it, lifting that brake on neural progenitor proliferation and differentiation in the dentate gyrus. Downstream, this is associated with increased BDNF expression and TrkB signalling through PI3K/Akt, culminating in GSK-3β inactivation — GSK-3β being a principal kinase driving abnormal tau phosphorylation, which is the proposed link to reduced tau pathology.

Crucially, because P21 does not reproduce full CNTF's activation of hypothalamic STAT3, it is reported not to cause the anorexia that ended CNTF's clinical development. That selectivity is the design's central claim.

The honest qualifier: this mechanism was worked out largely within the same research programme that reported the efficacy, and its BDNF step is precisely what the one independent study failed to reproduce.

Reading the evidence honestly

P21 is a case where the preclinical work looks genuinely good and the evidence structure is the problem.

The rodent record is real and reasonably deep. Multiple models — 3xTg-AD, aged rats, Ts65Dn — with consistent direction across neurogenesis markers, BDNF, tau phosphorylation and behaviour, including therapeutic (post-onset) as well as preventive dosing, plus favourable pharmacokinetics and long-term rodent safety. This is a better preclinical package than many compounds in this register can show.

But it is concentrated in one laboratory, and the one independent test was negative. Consistency within a single research programme is not the same as replication. The 2024 CDKL5 study failed to reproduce the BDNF effect in vivo and found limited behavioural benefit. That is one study in a different disease model, and it does not overturn the body of work — but when it is the only outside data point, it deserves to sit in the summary rather than the footnotes.

And there are no human data at all. Not a phase 1, not a pharmacokinetic study, not a registered trial. Alzheimer's mouse models have an especially poor translation record, so "rescues 3xTg-AD mice" is a bar that many subsequently failed drugs also cleared. The specific liability that ended CNTF — antibody formation — has been flagged as a theoretical risk for P21 and cannot be assessed without human exposure.

One point in P21's favour worth stating, since it distinguishes it from a neighbour: the nearby synaptogenesis compound Dihexa had its foundational mechanism papers retracted in 2025. No retractions or expressions of concern have been identified in the P21 literature. The issue here is not research integrity — it is a clean, narrow, single-source evidence base with nothing on the human side of the ledger.

Chemical identifiers

Cross-reference identifiers for the authoritative external databases — not a recipe, and nothing about how to use it.

PubChem CID56589645
InChIKeyLUJZBZPLIRWWGJ-NBONATBJSA-N

via PubChem CID 56589645 ("Peptide 021"), molecular formula C27H42N6O8 — matches the reported Ac-Asp-Gly-Gly-Leu-adamantylglycine-NH2 structure and is verified by exact name/formula match. This is NOT p21^Cip1/Waf1 (CDKN1A), the unrelated cell-cycle inhibitor protein. No clean CAS resolved; SMILES omitted (not independently verified). · high confidence

Sources — 8 cited
01Li B, Wanka L, Blanchard J, Liu F, Chohan MO, Iqbal K, Grundke-Iqbal I. Neurotrophic peptides incorporating adamantane improve learning and memory, promote neurogenesis and synaptic plasticity in mice. FEBS Lett. 2010;584(15):3359-3365.
02Kazim SF, Blanchard J, Dai CL, et al. Disease-modifying effect of chronic oral treatment with a neurotrophic peptidergic compound in a triple transgenic mouse model of Alzheimer's disease. Neurobiol Dis. 2014 (PMID 25046994).
03Bolognin S, Buffelli M, Puoliväli J, Iqbal K. Rescue of cognitive-aging decline in aged rats by oral P021. Neurobiol Aging. 2014 (PMID 24702821).
04Khatoon S, Chalbot S, Bolognin S, et al. P021 is BBB-permeable and reduces total tau in aged rats. J Alzheimers Dis. 2015 (PMID 26401692).
05Baazaoui N, Iqbal K. Prevention of dendritic and synaptic deficits and cognitive impairment with a neurotrophic compound. Alzheimers Res Ther. 2017;9(1):45.
06Bhattacharyya et al. (2024) — independent study in CDKL5 deficiency disorder reporting failure to replicate in-vivo BDNF increases and limited behavioural benefit.
07Lozupone M, et al. (2023) — review of peptide and oligonucleotide-based drugs targeting abnormal tau, noting possible antibody formation against P021 as a theoretical adverse event.
08Alzheimer's Drug Discovery Foundation, Cognitive Vitality report: P021 (last updated May 2025) — notes absence of any human studies.
Updated 2026-07-22 (no clinical trial is registered — re-check whether any human study or further independent replication appears)

Grades reflect the published evidence, not our interest. No dosing, reconstitution, or administration is published for research compounds — that restraint is deliberate.

Related compounds
CerebrolysinC
Peptide
Dihexa (PNB-0408)F
Peptide
Davunetide (NAP)F
Peptide
SemaxC
Peptide
Explore by goal
Neuro & cognitive →
Vallydia

A neutral reference and a lawful-lane shop. Registered in Spain. Information for those who seek it — never promotion.

Region — United States
ExploreRegisterThe Register — full indexCategoriesTrust & COAHow we gradeOpen data
ShopCosmetic peptidesJournalQuizzes
TermsPrivacyCookiesReturnsShippingImprint

This site provides neutral scientific reference and sells only products lawful in your region. Nothing here is medical advice, a recommendation, or an offer to supply unapproved medicines. No dosing or administration is published for research compounds. Cosmetic peptides per Regulation (EC) 1223/2009. Unapproved injectable peptides are neither sold nor advertised in the EU (Directive 2001/83/EC, Title VIII). © 2026 Vallydia SL — Registered in Spain.

P21 (P021) — evidence, uses & regulatory status · Vallydia