PE-22-28

Spadin analog

PE-22-28 (Spadin analog) is a cognitive & mood research compound. TREK-1 potassium-channel blocker derived from spadin — promotes neurogenesis with fast-acting antidepressant effects in models.

How PE-22-28 works

TREK-1 potassium-channel blocker derived from spadin — promotes neurogenesis with fast-acting antidepressant effects in models.

Proposed benefits

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

Where to get PE-22-28

The evidence for PE-22-28

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

• Very limited community use — it is a recent research compound with almost no practical record, and what exists is a handful of self-reports about mood.
• 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 spadin analog that blocks the TREK-1 potassium channel. TREK-1 knockout mice show a depression-resistant phenotype, and blocking it produces antidepressant-like effects in rodents within days rather than the weeks SSRIs require.
• The speed is the interesting part mechanistically — a channel-level intervention does not need the receptor adaptation that makes conventional antidepressants slow. Whether that transfers to humans is entirely untested.

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 PE-22-28 actually does

Start with where the molecule comes from, because it is not what the name suggests. Sortilin, also called neurotensin receptor-3, is made as a precursor and cleaved by the protease furin in the late Golgi. That cleavage releases a 44-amino-acid propeptide, abbreviated PE Mazella 2018. Spadin is PE 12–28 — 17 residues of that propeptide Djillani 2019. PE 22–28 is the last 7 residues of spadin. So this compound is a fragment of a fragment of a propeptide, and the arithmetic checks: 28 minus 12 plus 1 is 17, the stated length of spadin, and 28 minus 22 plus 1 is 7, the stated length of this one.

It was not designed. It was salvaged from what survived. The group started from spadin's own blood degradation products and built the 7-residue peptide out of the piece that was still there after the parent had been chewed up Djillani 2017. That is an unusual and honest origin story for a research compound: the drug is the metabolite.

The target is a background potassium channel. TREK-1 is a two-pore-domain (K2P) potassium channel that leaks K+ continuously and so holds neurons near a hyperpolarized resting state. Sortilin physically interacts with TREK-1 — shown by co-immunoprecipitation from COS-7 cells and cortical neurons co-expressing both, and by colocalization in mouse cortical neurons — and the propeptide it releases blocks the channel Mazella 2010 Mazella 2018. Block a background K+ leak and the cell sits closer to firing threshold. Mice with the TREK-1 gene deleted show a depression-resistant phenotype, which is what made the channel a target in the first place Mazella 2010.

The potency numbers are the most striking thing on this page. Spadin binds TREK-1 with an affinity near 10 nM and blocks it with an IC50 of 40–60 nM Mazella 2010 Djillani 2017. On human TREK-1 expressed in HEK cells and measured by patch clamp, PE 22–28 has an IC50 of 0.1 nM, with the G/A analog at 0.12 nM and the biotinylated version at 1.2 nM Djillani 2017. Removing ten residues made the peptide roughly 400 times more potent at the channel. That is not trimming for stability that happened to keep activity; it is a genuine structure-activity result.

The subtlety nobody repeats: binding and blocking come apart. Modifications of the N- or C-terminal ends of PE 22–28 either maintained or abolished the effect on channel activity without affecting the peptide's affinity Djillani 2017. A molecule can therefore occupy the channel and do nothing. For a buyer that has a hard consequence: a batch made with the wrong terminal chemistry could pass an identity test on mass alone and be pharmacologically inert.

And the residues themselves are not public. Vendors sell this as ‘PE-22-28’ and the primary papers identify it by position rather than by sequence in their abstracts and body text; the sequences appear only in a figure. This page will not print a sequence it cannot verify from a source, so the honest statement is that a 7-residue peptide is being sold and nobody in the retail chain publishes what the seven residues are. That is a solvable problem and nobody solves it.

Cell, rodent, human — and where it stops

In cells. Human TREK-1 expressed in HEK cells, patch clamp, IC50 0.1 nM for PE 22–28 Djillani 2017. For the parent: block confirmed in COS-7 cells, in cultured hippocampal pyramidal neurons and in CA3 neurons in brain slices Mazella 2010. On mouse cortical neurons, PE 22–28 and its derivatives raised PSD-95 expression, which is a synaptogenesis readout Djillani 2017.

In rodents, with the doses. Spadin at 100 µg/kg and PE 22–28 and analogs at 3.0–4.0 µg/kg intraperitoneally, or 1 mg/kg by gavage, in mice Djillani 2017. Forced swim: reduced immobility. Novelty-suppressed feeding after a 4-day sub-chronic course: reduced latency to eat. Neurogenesis after only 4 days of treatment, most prominently with the G/A analog. For spadin, an intravenous 4-day course raised hippocampal CREB phosphorylation and neurogenesis and increased 5-HT neuron firing in the dorsal raphe Mazella 2010.

And then the finding that should govern how anybody uses this. In a mouse model of focal ischemia and post-stroke depression, the short spadin analog was given at 0.03 µg/kg intraperitoneally once daily for 7 days to activate TREK-1 and produce neuroprotection, and then at 3 µg/kg once daily to inhibit it and produce the antidepressant effect. Electrophysiology confirmed the reason: the peptide is biphasic — at low doses channel activity went up, at higher doses it was inhibited Pietri 2019. A hundred-fold dose window separates two opposite actions on the same channel.

In humans: nobody. Not one person has been given PE 22–28 in any published setting. The only human data in this field is on the other side of the equation: soluble sortilin and the sortilin-derived propeptide are detectable in human serum and their levels differ in major depressive disorder, which makes them candidate biomarkers rather than treatments Mazella 2018.

The obstacles.

• The biphasic dose-response is the obstacle, and it is not generic. A person guessing at a microgram dose has no way to know which side of the line they are on, and the two sides predict opposite channel effects Pietri 2019.
• The dose arithmetic puts the retail protocol above both animal regimens. The antidepressant mouse dose is 3–4 µg/kg. Body-surface-area scaling divides a mouse dose by about 12, giving roughly 0.25–0.33 µg/kg, or about 18–23 µg for a 70 kg adult. Protocols in circulation run 50–200 µg, two to eleven times that, and hundreds of times the scaled equivalent of the 0.03 µg/kg neuroprotective dose. Nobody has established which regime a human at 100 µg is in.
• The oral datum exists and is ignored. 1 mg/kg by gavage reproduced what 3–4 µg/kg did intraperitoneally Djillani 2017 — a ratio of roughly 250 to 1, which is a measured statement about what surviving the gut costs this molecule.
• Species and channel. The original work was explicitly on rodent TREK-1 Mazella 2010; the 0.1 nM potency figure is against human TREK-1 in a cell line Djillani 2017, which is a genuine improvement and worth stating precisely rather than blurring.
• The behavioral endpoints are screening assays. Forced swim and novelty-suppressed feeding have a long history of predicting antidepressants that later failed in people. Rapid onset in a swim test is not rapid onset in depression.

PE-22-28 pharmacokinetics — how much of it actually gets in

The numbers everyone quotes are duration-of-effect numbers, not plasma half-lives, and the distinction matters. Spadin's in vivo activity disappears beyond about 7 hours; PE 22–28 and its analogs remain active for up to 23 hours, with half-effect times around 14–23 hours against about 6 hours for spadin Djillani 2017 Djillani 2019. Those are behavioral windows. No plasma concentration-time curve has been published for PE 22–28 in any species, so every ‘half-life’ figure circulating for this compound is a borrowed pharmacodynamic number wearing a pharmacokinetic label.

What degrades it. Serum peptidases — and the evidence is unusually direct, because the compound was designed out of the parent's blood degradation products Djillani 2017. A 7-residue peptide with a free N-terminus is an aminopeptidase substrate; the reason this fragment lasts three times longer than its parent is that the parent's vulnerable stretch has already been removed.

The oral barrier, with a real number attached. 1 mg/kg by gavage against 3–4 µg/kg intraperitoneally for a comparable behavioral effect Djillani 2017. That ratio — about 250-fold — is the cost of the stomach, the pancreatic proteases, the brush border and hepatic first pass for this molecule. It is also the reason nobody sells an oral form: the peptide survives the gut just well enough to work at a dose 250 times larger, which is a scientific result and a commercial dead end.

The injectable comparator. Every efficacy experiment above used intraperitoneal injection, which is not subcutaneous: intraperitoneal absorption goes substantially through the portal circulation and the liver, while a subcutaneous dose does not. The protocols in circulation are subcutaneous. Nobody has published a comparison of the two routes for this peptide, so the equivalence everyone assumes between an animal intraperitoneal dose and a human subcutaneous one is an assumption.

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

Four predictions. The third is the one that would embarrass the product, and the fourth is the one that should make anyone careful.

1. Sleep changes before mood does. TREK-1 is a background potassium channel and blocking it raises excitability broadly, not selectively in a mood circuit. Run actigraphy for two weeks before starting and through a four-week course. The mechanism predicts a measurable change in sleep onset latency or fragmentation earlier than any change in a mood score. If sleep is completely unaffected while mood improves markedly within days, the effect is more likely expectancy than channel block.

2. Morning cortisol will not move. Nothing in the TREK-1 story touches the HPA axis. Baseline and 4 weeks; a flat line is the expected result and a large move points somewhere else.

3. The rapid-onset claim is falsifiable in a week, and against the product. The entire pitch for this mechanism is speed: spadin produced antidepressant-like effects within 4 days against the 3–4 weeks a conventional antidepressant needs Mazella 2010 Djillani 2019. So if someone is three weeks into a course telling themselves it is still building, the mechanism being sold is not the mechanism operating. Either it works in the first week or the rodent pharmacology does not transfer at this dose.

4. Below some dose, the prediction reverses. The same short analog activated TREK-1 at 0.03 µg/kg and inhibited it at 3 µg/kg Pietri 2019. TREK-1 activation is hyperpolarizing and, in the ischemia model, neuroprotective — the opposite behavioral direction from block. Nobody can currently say where that crossover sits in a human, which means an under-dose is not a weaker version of the intended effect but potentially the opposite one. That is a prediction, it is testable with a dose-response in mice in a fortnight, and it has not been tested for this peptide.

What nobody has tested yet

Six experiments. The first two are the ones a buyer should care about most.

1. Nobody has published the sequence in a form a buyer can check against a vial. Not the vendors, not the abstracts. Until the seven residues are public and a batch has been run against them by mass spectrometry, identity for this compound is taken on trust.

2. Nobody has mapped the biphasic curve for PE 22–28 itself. The activation-at-low-dose, inhibition-at-high-dose result was established for the short analog used in the stroke study Pietri 2019. Whether PE 22–28 flips sign at low dose, and at what concentration, is unpublished — and it is the single most decision-relevant fact about the compound for anyone taking it.

3. Nobody has measured a plasma concentration in any species. One pharmacokinetic study in rats, five timepoints, would replace every borrowed half-life figure on the internet with a real one.

4. Nobody has run the terminal-chemistry check on commercial material. Djillani's own result says N- and C-terminal modifications can abolish channel block while leaving affinity intact Djillani 2017. A functional assay — does this batch block hTREK-1 in a patch clamp — is the only test that distinguishes active from inert material, and no vendor performs one.

5. Nobody has used the human biomarker that already exists. Serum soluble sortilin and sortilin-derived propeptide are measurable and differ in major depressive disorder Mazella 2018. A self-experiment group could measure the endogenous ligand of the target they are blocking, before and after — the closest thing to a mechanistic human readout this compound could have.

6. Nobody has tested the sleep prediction, and wearables make it nearly free. Continuous sleep staging across a four-week course, twenty people, no laboratory required. It would be the first human physiological dataset of any kind for this molecule.

PE-22-28 — its own safety story, not its class's

The class block on this page is written for pharmacological CNS drugs — phenibut, trazodone, propranolol, cabergoline. This compound is nothing like any of them, and its own risks are specific.

1. A knockout phenotype is not a toxicology package. The argument that blocking TREK-1 is safe rests on TREK-1-deleted mice being viable and depression-resistant Mazella 2010. A mouse that developed without the channel is not the same as an adult brain in which the channel is blocked intermittently. No repeat-dose toxicity study, no organ histopathology and no reproductive study for PE 22–28 appears in the published record.

2. The biphasic finding is the safety headline. Getting the dose wrong on this compound does not simply do less — at the low end the same class of molecule activates the channel it is supposed to block Pietri 2019. Compounds whose dose-response changes sign are the ones where self-titration is most dangerous, and this is the only compound in this cohort with a published sign change.

3. The context of use is the real hazard. This is bought as a fast antidepressant by people who are depressed, with zero human data of any kind. The failure mode in depression is not a lab abnormality; it is deterioration and suicidal risk, and the specific danger is somebody stopping or delaying a treatment that has randomized evidence in order to try one that has none. Worsening mood, new hopelessness or any thought of self-harm during a course is an immediate reason to contact a clinician, not to adjust a dose.

4. What zero reported adverse events means here. It means zero humans have been dosed in any published setting. There is no denominator at all, so nothing about tolerability in people is known in either direction.

Sources read for this page

• Djillani A, Pietri M, Moreno S, Heurteaux C, Mazella J, Borsotto M. Shortened Spadin Analogs Display Better TREK-1 Inhibition, In Vivo Stability and Antidepressant Activity. Frontiers in Pharmacology 2017 · PMID 28955242
• Mazella J, Petrault O, et al. Spadin, a sortilin-derived peptide, targeting rodent TREK-1 channels: a new concept in the antidepressant drug design. PLoS Biology 2010 · PMID 20405001
• Djillani A, Mazella J, Heurteaux C, Borsotto M. Fighting against depression with TREK-1 blockers: Past and future. A focus on spadin. Pharmacology and Therapeutics 2019 · PMID 30291907
• Pietri M, et al. First evidence of protective effects on stroke recovery and post-stroke depression induced by sortilin-derived peptides. Neuropharmacology 2019 · PMID 31325429
• Mazella J, Borsotto M, Heurteaux C. The Involvement of Sortilin/NTSR3 in Depression as the Progenitor of Spadin and Its Role in the Membrane Expression of TREK-1. Frontiers in Pharmacology 2018 · PMID 30670975

PE-22-28 — safety, predicted from mechanism

PE-22-28 — interference & stacking

Bloodwork to run alongside PE-22-28

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

PE-22-28 — frequently asked questions

PE-22-28 inside a finished plan

One arm of 1 Protocol Blueprint, free to read in full.

What PE-22-28 is used for

PE-22-28 appears under 2 goals in the goal router.

Related Cognitive & Mood compounds

← Explore the full Protocol Vault

↑ Back to on this page