PE-22-28: The Antidepressant Claim Is Ten Mice in a Water Cylinder

01What it is

PE-22-28 is seven amino acids, GVSWGLR, and the name is a set of coordinates. Sortilin is a receptor protein on cell surfaces, built with a 44 amino acid leader piece at the front, called the propeptide, that is cut off when the receptor matures. Residues 22 through 28 of that propeptide are GVSWGLR, which I counted off UniProt myself. The parent is spadin, seventeen residues ending in those same seven; the short one is also called mini-spadin.

The sequence was not designed. Djillani and colleagues incubated spadin in mouse serum, watched it fall apart, and built short peptides from the fragments. (PMID 28955242) It was found in the wreckage of its parent.

TREK-1, the target, is a potassium channel, a pore that lets potassium leak out of a nerve cell and so makes it harder to fire. The antidepressant idea starts from a gene deletion rather than a drug. Heurteaux and colleagues bred mice without the TREK-1 gene and reported they resisted depression in five behavioural models. (PMID 16906152) Everything since is an attempt to reproduce a lifelong genetic absence with an injection into an adult.

02What the trials found

None that I could find, and being exact about that matters. A PubMed search for PE 22-28 and its spellings returns one record, the 2017 mouse paper. ClinicalTrials.gov returns zero for both names, and the EU Clinical Trials Register replies "Query did not match any clinical trials." I could find no published or registered study giving PE-22-28, spadin or any analogue of either to a person.

One document points the other way. The inventors wrote, in a review published on 8 January 2019, that "the further use of spadin and/or spadin analogs in clinical trials is currently in progress." (PMID 30670975) Seven and a half years later I could find no registration of it, and no result.

So the record is mice: male C57Bl/6J, seven to nine weeks old, one laboratory in Valbonne, France. The swim test is the antidepressant claim. A mouse is put in a cylinder half filled with water at 22 degrees for six minutes and someone times how long it stops swimming during the last four. Less floating, less despair. Thirty minutes after one injection into the abdominal cavity at 3.0 to 4.0 micrograms per kilogram, ten mice per group, PE 22-28 cut floating time to 91.80 seconds against 161.7 on saline.

Read what that is: how long a mouse stops swimming, scored by eye, half an hour after a needle in the belly. The words blind, blinded, randomised and randomly do not appear in the body of that paper, and the immobility time, in its own words, "was manually measured."

03The dose that opens the channel and the dose that closes it

In 2019 the same laboratory published a mouse stroke study using mini-spadin and reported that the peptide "had a biphasic action on TREK-1. At low doses, the channel activity was increased whereas at higher doses it was inhibited." (PMID 31325429) They used it deliberately: 0.03 micrograms per kilogram into the abdominal cavity daily for seven days after the stroke, to open the channel and protect the brain, then 3 micrograms per kilogram to close it for the antidepressant effect. Two doses a hundredfold apart, opposite effects on the same target, in mice. That paper is not open access, so I read the abstract.

So PE-22-28 is not a blocker in the way a product page implies. Two settings tried on a dial with a hundredfold span, doing opposite things at either end, nothing turned in between.

04How it compares

The potency claim does not survive checking. These numbers come from cells in a dish rather than animals, and not all from the same dish: the headline figure was read from human TREK-1 channels put into a kidney cell line and recorded one cell at a time, and the comparison figure from a different cell line. IC50 is the concentration producing half the maximum effect, and the 2017 paper prints its own four times without agreeing with itself: 0.12 nanomolar twice and 0.10 twice, swapped between PE 22-28 and its alanine substituted cousin. The denominator is worse. It cites spadin at 40 to 60 nanomolar, crediting the 2010 paper that measured it, in a different cell line, at 70.7. (PMID 20405001)

Then the parent's block did not replicate. Ma and Lewis at Portsmouth put mouse and human TREK-1 into frog eggs and could not significantly inhibit the activated channel with spadin at 1 micromolar, fourteen times its IC50. (PMID 32317978) They tested spadin, not PE-22-28, and they name the limit themselves: results from frog eggs may not hold in other cell types, and peptides made by different methods may not behave the same way. But the potency claim is a ratio, and its denominator is what failed.

I found no drug approved anywhere whose approval rests on blocking TREK-1, and no selective TREK-1 blocker approved for any indication. The pharmacology is not unprecedented though: fluoxetine inhibits TREK-1 at about 6 micromolar (PMID 28955242), at least a hundred times weaker than spadin's 40 to 60 nanomolar, and the group that made spadin calls fluoxetine a TREK-1 inhibitor with proven clinical efficacy in depression.

05Where the dose came from

Not from the paper. Every published dose of PE-22-28 is in a mouse: 3.0 to 4.0 micrograms per kilogram into the abdominal cavity for the swim tests, and 1 milligram per kilogram down a tube into the stomach in an oral panel whose legend contradicts its own caption. The market runs on two others, up the nose and under the skin, and my PubMed search pairing either name with the nose, the mouth or the skin returned nothing.

Four pages, guides and vendor pages, give per dose figures from 100 to 2,000 micrograms, mostly up the nose. A twentyfold spread for one molecule, convention rather than evidence. One page says so: its dosing "comes from community practice, not from the published research."

The arithmetic here is mine, done to check the method rather than to propose anything. The standard body surface area conversion divides a mouse dose by 12.3, so 3 micrograms per kilogram in a mouse becomes about 17 micrograms for a 70 kilogram adult and 4 becomes about 23. The low end of the convention is roughly four to six times that, the high end ninety to a hundred and twenty times.

I found no study in any species measuring how much PE-22-28 gets into the blood, how far it spreads or how fast it clears, which is what a pharmacokinetic study does. The 23 hours belongs elsewhere: that experiment injected two modified analogues into the abdominal cavity at 3.2 to 40 micrograms per kilogram, ten mice per time point, and PE 22-28 was not in it. Every once daily on a protocol page is a schedule with no measured half life behind it.

06What could go wrong

I could find no published human safety data on PE-22-28. No tolerability record, no dose ceiling and no adverse event count came back from my searches, and I found no formal toxicology study in any species. Empty is not the same as clean.

The primary paper's safety argument is a prediction and says so in its own words: "we could expect that spadin analogs would be devoid of side effects." On vendor pages that becomes "no side effects." The test underneath it was real: in kidney cells carrying the channels, PE 22-28 left four related potassium channels untouched at 100 nanomolar, and all three analogues left hERG, the cardiac channel, untouched at 10 micromolar, a hundred times higher.

Against that, mice born without TREK-1 are more easily damaged by stroke and seizures, and resist being put under with inhaled anaesthetics. (PMID 15175651) The spadin group raised that hazard themselves, tested it, and reported no worsening of pain, seizures, stroke damage, heart function or blood sugar in rodent models, which is as far as the abstract goes; I could not open the full text. (PMID 21807005) That was spadin, and the anaesthetic finding is not addressed anywhere I could find.

07What nobody knows

How long PE-22-28 lasts in an animal, and whether it reaches the brain. The only brain measurement I found belongs to the parent.

Which side of the biphasic curve a human dose lands on.

Whether the compound does anything in a female animal. Every behavioural experiment I found used males.

Whether it is named on the 2026 WADA Prohibited List; that list did not load for me. The compounding side has an answer: PE 22-28 and spadin appear nowhere in FDA's 14 May 2026 list of nominated substances.

08What the selling rests on

Thin evidence is the normal condition of everything here. What is unusual is how little of the selling belongs to the molecule in the vial. The 23 hours was measured on a modified analogue at ten times the dose, and the paper cannot decide which of its two modified analogues it belongs to, while the potency ratio is built on a denominator its own source contradicts. Even the premise underneath both is a mouse born without the gene rather than a mouse given the drug.

If you have seen PE-22-28 sold anywhere, I want to know whether the page quoted 23 hours, and whether it named the molecule that figure was measured on.

Frequently asked

Has PE-22-28 been given to people?+

Not in anything I could find. Searches of PubMed, ClinicalTrials.gov and the EU Clinical Trials Register in August 2026 returned no published study and no registration in which PE-22-28, spadin or any analogue of either was given to a human being by any route. The inventors wrote in a review published in January 2019 that clinical trials were in progress, and seven and a half years later I could find no registration of one and no published result.

What did the animal studies actually measure?+

Mostly how long a mouse floats. The antidepressant result is a forced swim test: a mouse is placed in a water filled cylinder for six minutes and the time it spends motionless during the last four is measured by hand. Thirty minutes after a single injection into the abdominal cavity at 3.0 to 4.0 micrograms per kilogram, ten male mice per group spent about 91.8 seconds motionless against 161.7 seconds in the saline group. Every behavioural experiment in that paper used male mice of one strain, from the laboratory that designed the peptide.

Where does the 300 times more potent than spadin figure come from?+

From a comparison whose sources do not agree with each other. The 2017 paper prints its own half maximum concentration for PE 22-28 four times, all of them measured on channels put into a kidney cell line rather than in any animal, giving 0.12 nanomolar twice and 0.10 twice, and swapping the value with a modified analogue. It then cites spadin at 40 to 60 nanomolar, crediting a 2010 paper that actually reports 70.7 from a different cell line. Three denominators against two numerators give six different answers, from 333 to 707 fold, and the same group's own review states it as a range of 40 to 350 times.

Does PE-22-28 last 23 hours?+

The 23 hour figure was measured on a different molecule. In the 2017 mouse experiment that produced it, the injections were of two modified analogues, one alanine substituted and one biotin tagged, at 3.2 to 40 micrograms per kilogram into the abdominal cavity, ten mice per time point, and PE-22-28 itself was not in the experiment. I could find no measurement of how long PE-22-28 lasts in any species by any route.

Where do the doses that circulate come from?+

From community practice rather than from research, and they are reported here rather than recommended. Four pages read in August 2026, guides and vendor pages, give per dose figures running from 100 to 2,000 micrograms, mostly up the nose, which is a twentyfold spread for one molecule. One of those pages says outright that its dosing information comes from community practice and not from the published research. The only published doses I could find are in mice, 3.0 to 4.0 micrograms per kilogram injected into the abdominal cavity, and scaling those by the standard body surface area conversion gives about 17 to 23 micrograms for a 70 kilogram adult.

Is PE-22-28 safe?+

I could find no published human safety data on PE-22-28. No tolerability record, no dose ceiling and no adverse event count turned up in my searches, and I found no formal toxicology study in any species. The primary paper's own safety argument is a prediction rather than a measurement, in its authors' words that they could expect the analogues to be devoid of side effects. The same laboratory later reported that the peptide opens its target channel at low doses in mice and closes it at higher ones, so the relationship between amount and direction of effect is not established. An empty safety record is not the same as a clean one.

---

This content is for educational and informational purposes only and is not medical advice. Peptides discussed are research compounds and may not be approved for human use. Nothing here should be used to diagnose, treat, cure, or prevent any disease. Always consult a qualified healthcare professional before starting any peptide, supplement, or protocol. Individual responses vary. Do not self-administer compounds without proper medical supervision.