P21 5mg Peptide

P21, also written P021, is a four-residue peptidomimetic derived from ciliary neurotrophic factor and capped at its C-terminus with an adamantane cage. The literature shorthand is Ac-DGGLAG-NH2: an acetylated Asp-Gly-Gly-Leu core taken from the CNTF active region, with an adamantylated glycine bolted on. It came out of Khalid Iqbal's laboratory at the New York State Institute for Basic Research in Developmental Disabilities on Staten Island, which has published essentially the entire preclinical file on it since the first report in 2010.

The first thing to settle is what this molecule is not. A search for "p21" returns overwhelmingly the cyclin-dependent kinase inhibitor p21, gene symbol CDKN1A, also called p21 Waf1/Cip1. That is a 164-residue human protein of about 18.1 kDa, a transcriptional target of p53 that binds cyclin-CDK complexes and arrests the cell cycle, and it is one of the most studied proteins in cancer biology. It has no sequence, no structure and no mechanism in common with the CNTF-derived peptide sold under the name P21. The two share four characters of nomenclature and nothing else, and any statement about cell cycle arrest, senescence or tumour suppression belongs to CDKN1A, not to this compound.

The design problem P021 was built to solve is a specific one. Recombinant CNTF is a genuinely potent neurotrophic protein, but its plasma half-life in humans is about 2.9 minutes, it does not cross the blood-brain barrier well, and the amyotrophic lateral sclerosis trials of the 1990s reported anorexia, substantial weight loss, hyperalgesia, severe cramps and muscle pain. Epitope mapping with neutralising antibodies narrowed the active region to residues 146 to 156, giving the 11-mer Peptide 6 (Ac-VGDGGLFEKKL-NH2). Truncating that to its minimal active core, Ac-DGGL-NH2, and then adding the adamantyl group produced P021. Every result described below comes from rodents or cultured cells. There is no human data of any kind.

The proposed mechanism has two arms that meet downstream. The first is competitive inhibition of leukemia inhibitory factor signalling. LIF, like CNTF, signals through gp130 and holds neural stem cells in an undifferentiated state, suppressing their commitment to a neuronal lineage. In cell culture, P021 inhibited LIF-induced phosphorylation of STAT3 by roughly 30 percent (Li and colleagues, FEBS Letters, 2010). Relieving that brake is the proposed route to the increase in neural progenitor cell formation seen in the dentate gyrus.

The second arm is induction of brain-derived neurotrophic factor. Across the rodent studies, chronic P021 raised BDNF protein and phosphorylated CREB in hippocampus and cortex, without moving pro-BDNF, TrkB or CNTF itself. In Ts65Dn mice the BDNF increase was about two-fold relative to vehicle-treated trisomic animals, which brought it back to the wild-type level (Kazim and colleagues, Scientific Reports, 2017). BDNF acting at TrkB engages PI3K and Akt, and Akt phosphorylates glycogen synthase kinase-3 beta at Ser9, which is the inhibitory site.

That last step is what connects a neurotrophic compound to a tau result. GSK-3 beta is a major tau kinase, and inhibiting it lowers phosphorylation at the epitopes that define Alzheimer neurofibrillary pathology. In 3xTg-AD mice, P021 in diet at 60 nmol per gram of feed raised Ser9-phosphorylated GSK-3 beta and reduced tau labelling with AT8 (pSer202, pThr205), PHF-1 (pSer396, pSer404) and 12E8 (pSer262, pSer356), while leaving total tau unchanged (Kazim and colleagues, Neurobiology of Disease, 2014). The compound is therefore not a tau aggregation inhibitor and not an immunotherapy; the tau effect is downstream of a kinase, which is why total tau does not move.

The adamantane is a delivery decision, not a pharmacophore. The cage is lipophilic and bulky, and it was added to raise membrane permeability and to block exopeptidase attack on the C-terminus, both of which the unmodified tetrapeptide lacked. It works for stability: P021 retains over 90 percent integrity in artificial gastric juice after 30 minutes and essentially 100 percent in artificial intestinal fluid after 2 hours at 37 degrees Celsius, and plasma half-life in mice is over 3 hours against 2.9 minutes for recombinant CNTF. Whether it works for brain entry is less settled, and that gap is covered below.

This is the single most common error in material written about the compound, and it is worth being blunt about. p21, gene symbol CDKN1A, is a human protein of 164 amino acids and about 18,119 Da, catalogued in UniProt as P38936 and known variously as p21 Waf1, p21 Cip1 and CDK-interacting protein 1. It is transcribed in response to p53 activation, binds cyclin-CDK complexes, blocks phosphorylation of retinoblastoma protein and thereby arrests the cell cycle at G1 and at the DNA damage induced G2 checkpoint. It also competes with POLD3 for PCNA binding, which is how it inhibits DNA polymerase delta.

P021 is a synthetic construct of four amino acids and an adamantane cage, total mass 578.66 Da, with no biological relationship to CDKN1A whatsoever. It does not bind cyclins, does not respond to p53, and has never been studied in a cell cycle context. The overlap is purely typographic: the Iqbal group numbered its compounds sequentially, and peptide number 21 in that series happened to collide with one of the most cited protein names in oncology.

The practical consequence is that literature searches contaminate quickly. A PubMed query for the bare string p21 returns tens of thousands of cell cycle papers. The queries that actually retrieve this compound are P021, Peptide 021, DGGLAG, or the combination of CNTF with peptide mimetic. Anyone reading about "p21 and senescence", "p21 in tumour suppression" or "p21 and PCNA" is reading about a different molecule.

CNTF is a four-helix bundle cytokine of the IL-6 family that signals through a tripartite receptor of CNTF receptor alpha, LIF receptor beta and gp130, activating JAK-STAT3, MAPK-ERK and PI3K-Akt. It supports motor neuron survival, drives adult hippocampal and subventricular zone neurogenesis, and has real effects on oligodendrocytes, retina, muscle and adipocytes. Those same pleiotropic actions are why it failed as a systemic drug: the ALS CNTF Treatment Study Group reported anorexia, skeletal muscle loss, hyperalgesia, severe cramps, muscle pain, nausea, cough and mouth ulcers in 1996.

The Iqbal group's route around this was to keep the neurogenic epitope and discard the receptor engagement that produces the systemic effects. Epitope mapping with neutralising antibodies identified residues 146 to 156 as a biologically active region, and the corresponding 11-mer, Ac-VGDGGLFEKKL-NH2 (Peptide 6), enhanced dentate gyrus neurogenesis and Morris water maze reference memory in adult C57BL/6 mice after 30 days from a subcutaneous slow-release bolus. Peptide 6 has a plasma half-life over 6 hours against roughly 3 minutes for recombinant CNTF, and it acts by competitively inhibiting LIF signalling.

Peptide 6 was then truncated to its minimal active core, the tetrapeptide Ac-DGGL-NH2 (Peptide 6c), which retained neurogenic and neurotrophic activity. Adding adamantylated glycine to that C-terminus produced P021, first reported in FEBS Letters in 2010. In that study, adult C57BL/6 mice aged 8 to 10 months received 25 nmol per day from subcutaneous extended-release depot pellets for 35 days and improved on one-trial object recognition and spatial reference memory, with increased BrdU-positive cells in the granule cell layer and subgranular zone and increased synaptophysin and synapsin I immunoreactivity in the dentate gyrus.

One nomenclature detail is worth carrying: because P021 was built by truncation and then chemical capping, it is described as a tetrapeptide in some papers and a pentamer in others, depending on whether the adamantylated glycine is counted as a residue. Both descriptions refer to the same molecule.

The triple transgenic 3xTg-AD mouse carries APP Swedish, PS1 M146V and tau P301L, and develops intraneuronal amyloid beta by around 4 months and tau pathology around 12 months. Four separate treatment windows have been tested, and the pattern of what responds is informative.

Starting at 9 to 10 months, after amyloid onset but before tau pathology, 60 nmol per gram of feed for 6 or 12 months reduced phospho-tau at all three epitope sets, raised Ser9-phosphorylated GSK-3 beta, restored dentate gyrus neurogenesis to wild-type counts, restored MAP2 and synaptophysin in CA1, CA3 and dentate gyrus, and rescued novel object recognition to wild-type discrimination levels. The amyloid effect was much narrower: soluble Abeta fell at 15 to 16 months but not at 21 to 22 months, insoluble Abeta did not move, and plaque load showed only a trend in CA1. The authors read this as reduced Abeta generation rather than clearance.

Starting at 3 months, during the synaptic compensation period and before any overt pathology, and continuing to 21 or 22 months, the effects were larger. Neurodegeneration measured by Fluoro-Jade C fell in CA1, CA3, dentate gyrus and parietal cortex, plaque load fell by roughly 40 percent in the subiculum and 20 percent in CA1, tau hyperphosphorylation fell by about half, and episodic memory was rescued. The most striking number in the whole file is survival: at week 71, 87 percent of P021-treated 3xTg-AD females were alive against 41 percent of vehicle-treated ones, and hindlimb paralysis occurred in 1 of 32 treated animals against 5 of 34 controls.

Two prenatal and early postnatal studies pushed the window earlier still. Dosing mothers from prenatal day 8 to postnatal day 21 raised GluN1, GluN2A, GluA1, pCREB and PSD-95 and lowered GFAP in the adult offspring, and reduced tau hyperphosphorylation and plaque load out to 22 months, while leaving synapsin, synaptophysin, MAP2 and Iba1 unchanged. A companion study dosing from birth to postnatal day 120 restored hippocampal BDNF and raised GluN2A, GluA1 and GluA2/3 along with synapsin 1, PSD-95 and MAP2 transcripts. In 2026 a diffusion MRI study reported that these early treatment effects are detectable as brain microstructure changes in vivo, which is the first non-histological readout in the series.

The Ts65Dn mouse is trisomic for a segment of chromosome 16 orthologous to human chromosome 21 and shows both developmental delay and Alzheimer-like memory impairment. Treating mothers at 200 nmol per gram of formulated diet from gestational day 8 through postnatal day 21 rescued the delayed appearance of surface righting, cliff aversion and ear twitch in the pups, prevented the presynaptic protein deficit, lowered GSK-3 beta activity, and roughly doubled BDNF and raised phosphorylated CREB at both 3 weeks and 7 months of age. Adult spatial reference memory in the Morris water maze was attenuated toward wild type. PSD-95, TrkB, CNTF and total CREB did not change, and body weight, anxiety-like behaviour and activity level were unaffected.

The CDKL5 deficiency disorder study published in 2024 is the honest counterweight and deserves as much attention as the positive results. In SH-SY5Y cells with CDKL5 knocked out, P021 restored proliferation, survival and neuronal maturation in a dose-dependent way, and corrected GSK-3 beta signalling. In Cdkl5 knockout mice given 60 nmol per gram of diet from postnatal day 21 to 90, it did not work: BDNF did not rise, doublecortin-positive cell survival did not recover, CA1 pyramidal neuron survival and dendritic spine maturation were unchanged, and behaviour on Morris water maze, passive avoidance, marble burying, nest building and rotarod was not restored. The only in vivo signal was reduced muscular rigidity on the catalepsy bar test. The authors' interpretation is that CNTF-dependent responses may themselves be broken in a CDKL5-null brain.

Read together, these two results suggest the compound needs an intact BDNF induction pathway to do anything, and that a genotype which disrupts that pathway is not rescued by pushing on it. That is a more useful framing than treating every model as a win.

The CNTF receptor complex is structurally related to the leptin receptor, and the two are distributed similarly across the hypothalamic nuclei that regulate feeding. When CNTF caused unexpected weight loss in motor neuron disease trials, the question was whether it was acting as a cachectic cytokine or as a leptin mimic. Work at Regeneron settled it toward the latter: CNTF activated hypothalamic leptin-like pathways in diet-induced obesity models that no longer respond to leptin, suppressed food intake without triggering the hunger and stress signals that follow food deprivation, and produced no binge rebound on withdrawal (Lambert and colleagues, PNAS, 2001). A modified CNTF derivative later improved glucose, insulin, triglyceride and hepatic steatosis measures in db/db mice and suppressed stearoyl-CoA desaturase 1, the same gene leptin suppresses in ob/ob mice (Sleeman and colleagues, PNAS, 2003). The clinical version reached a Phase 2 obesity trial with a significant weight reduction, and then failed in Phase 3 on peripheral adverse effects and neutralising antibodies.

This matters for reading P021 because it means the anorectic action of CNTF is a receptor-level property of the intact cytokine and its leptin-like hypothalamic signalling, not an unavoidable feature of the neurogenic epitope. P021 does not reproduce it. Across rodent studies running as long as 18 months, no weight loss was observed; body weight rose modestly in both 3xTg-AD and wild-type mice at 60 nmol per gram of feed, with food consumption unchanged. In the Down syndrome and CDKL5 studies, body weight was likewise unaffected.

The honest limit here is that no published study has measured leptin, leptin receptor signalling, neuropeptide Y, or hypothalamic STAT3 in a P021-treated animal. The absence of an anorectic effect is a robust and repeatedly reported observation. The mechanistic explanation for that absence, that the truncated fragment does not engage the hypothalamic arm of CNTF signalling, is an inference from the structure and from the phenotype, not a measured result. Anyone describing P021 as a leptin sensitiser is going well beyond what has been published.

The molecular data on this compound is unusually badly reported, so it is worth setting out what the registries actually hold. PubChem CID 56599151, CAS 1246751-68-7, UNII VV8CZC8PAS: molecular formula C27H42N6O8, average molecular weight 578.66 g/mol, monoisotopic mass 578.3064 Da, InChIKey LUJZBZPLIRWWGJ-ZKHAZHQOSA-N. The commonly circulated value of 578.3 is the monoisotopic mass presented as though it were the average molecular weight. The commonly circulated formula C30H54N6O5 does not correspond to any structure with an aspartate residue and is not what the registries hold.

The deposited structure resolves the C-terminal cap as 3-aminoadamantane-1-carboxamide amide-bonded directly to the leucine carbonyl. That is one methylene short of a literal adamantylglycine residue, which is why the formula carries 27 carbons rather than 28. The papers' shorthand of an added adamantylated glycine and the registry structure are describing the same commercial entity at different levels of resolution, and both are in circulation. The one reading that is simply wrong is treating the A in Ac-DGGLAG-NH2 as alanine: it denotes the adamantane, and a genuine Ac-Asp-Gly-Gly-Leu-Ala-Gly-NH2 hexapeptide would be a different compound with a mass near 502 Da and none of the lipophilicity the design depends on.

Stability data comes from the developing laboratory. P021 retains over 90 percent integrity in artificial gastric juice after 30 minutes and effectively 100 percent in artificial intestinal fluid after 2 hours, both at 37 degrees Celsius, which is why the animal work is done through diet and oral gavage rather than injection. Plasma half-life in mice is over 3 hours. The lyophilised powder is handled at -20 degrees Celsius and protected from light and moisture; once in solution the compound is held at 2 to 8 degrees Celsius, and repeated freeze-thaw cycles are avoided in the same way as for any small peptide. Identity is normally confirmed by reversed-phase HPLC with mass spectrometric confirmation of the 578.3 monoisotopic ion, and the adamantane conjugation step is the one where impurity profiles diverge, since incomplete and over-conjugated species co-elute closely with the target.

Cerebrolysin is the closest functional comparator and the one buyers ask about most. It is a porcine brain-derived hydrolysate, a complex undefined mixture of low molecular weight peptides and free amino acids with neurotrophic activity attributed partly to CNTF-like components. P021 is the opposite design choice: one defined molecule of known formula and mass, synthesised rather than extracted, with a single proposed mechanism. A retracted 2011 comparison in AbetaPP transgenic mice had reported that both increased neuroblast counts but through different routes, the hydrolysate by reducing apoptosis and the CNTF peptides by increasing proliferation. That paper was retracted in 2026, so the comparison should be treated as unsupported rather than as established.

Adamax shares only the chemistry, not the parentage. It is an adamantane-conjugated analogue of the ACTH(4-10) melanocortin fragment, structurally in the Semax family, and the adamantyl group is doing the same lipophilicity job in both molecules. That shared modification is why the two are often shelved together, and it is not evidence of a shared mechanism: melanocortin fragments and CNTF fragments act on unrelated receptor systems.

Semax and Selank are Russian-developed regulatory peptides with their own literatures and their own BDNF findings, and Dihexa is an angiotensin IV analogue acting on hepatocyte growth factor and c-Met. All four are studied for hippocampal plasticity, which is what puts them in the same catalogue category, and none of them shares a receptor with P021. The distinguishing feature of P021 within this group is that its entire published record comes from a single institutional cluster, and that the record includes long-duration dosing studies of 12 to 18 months, which is unusual for a research peptide at this stage.

There is no human data. A query of ClinicalTrials.gov returns no registered study of P021 in any indication, at any phase, and none has ever been run. Safety in humans has not been assessed. Everything in the record is rodent or cell culture, and Phanes Biotech, the company holding the license, describes the compound as in preclinical development for Alzheimer's disease and related neurodegenerative conditions.

Blood-brain barrier penetration is asserted more confidently than the data supports. The adamantane was added to increase it, and behavioural and biochemical effects in brain after peripheral dosing are consistent with brain exposure. But no study has directly assayed the concentration of P021 reaching the brain in mice, a limitation the authors themselves note. The computed properties do not obviously favour it either: topological polar surface area of 226 square angstroms and XLogP of -1.1 are not the numbers of a molecule that crosses membranes freely, which makes the absence of a direct measurement more conspicuous rather than less.

Almost the entire file comes from one laboratory. Khalid Iqbal is co-founder and Chief Scientific Officer of the company developing the compound and holds patents on its composition and use, and that conflict is disclosed on the papers. Independent replication is thin: the 2024 CDKL5 work from the University of Bologna is the clearest external test, and its in vivo arm was negative. The 2026 diffusion MRI collaboration with the Medical University of South Carolina is the other outside contribution.

Finally, no study has examined P021 in healthy young animals as a cognitive enhancer in the sense the term is usually used. The wild-type work in the 2010 report used mice aged 8 to 10 months, and the rest of the record is disease models, aged animals, or developmental rescue. Immunogenicity is an open question rather than a resolved one: no immune reaction to P021 has been detected in rats to date, but the possibility of antibody formation has been raised in review, and the neutralising antibodies that ended the clinical CNTF derivative programme are the reason it is worth asking about.