PE 22-28 10mg Peptide

PE 22-28 is a seven-residue peptide, Gly-Val-Ser-Trp-Gly-Leu-Arg, taken from the middle of a 44-amino-acid propeptide that is released when the sorting receptor sortilin, also called neurotensin receptor 3, is matured by the protease furin in the Golgi. The name is literal. The group that studies this family abbreviates that propeptide as PE, and this fragment spans its residues 22 to 28. A longer fragment of the same propeptide, residues 12 to 28, is spadin.

The compound exists because of a potassium channel. TREK-1, the product of the KCNK2 gene and also written K2P2.1, is a two-pore-domain background channel that sits open at rest and holds neurons near their resting potential. In 2006 a laboratory at the Institut de Pharmacologie Moleculaire et Cellulaire in Valbonne, outside Nice, reported that mice lacking Kcnk2 were resistant to depression-like behaviour in five separate models, showed more efficient serotonergic neurotransmission, and mounted a substantially smaller corticosterone rise under stress. Untreated knockout animals behaved like normal animals given fluoxetine. That result turned a background channel into a drug target, and the search for a blocker produced spadin in 2010 and PE 22-28 in 2017.

What makes the target unusual is that it has nothing to do with monoamine reuptake. Fluoxetine does block TREK-1, but with a half-maximal inhibitory concentration near 6 micromolar; PE 22-28 blocks the same channel at 0.10 nanomolar. The literature on PE 22-28 is small, entirely preclinical, and comes almost exclusively from the Valbonne laboratory. There is no registered clinical trial for PE 22-28 or for spadin, and no published pharmacokinetic study in any species.

PE 22-28 blocks TREK-1, and the block is state dependent. In whole-cell patch clamp on a HEK293 line stably expressing human TREK-1, the current is first opened with 10 micromolar arachidonic acid, then the peptide is applied. Under those conditions 100 nanomolar PE 22-28 reduced current density at 0 mV by 55.5 plus or minus 4.6 percent (n = 13, p < 0.0001), against 44.4 plus or minus 8.8 percent for spadin (n = 7). Full concentration-response curves gave an IC50 of 0.10 nanomolar for PE 22-28 and 0.12 nanomolar for its alanine-substituted analogue, against 40 to 60 nanomolar for spadin: a gain of more than 300-fold from a peptide less than half the length. Inhibition is much weaker when the channel has not been opened first, which the authors read as the peptide needing an open-state conformation to reach the selectivity filter.

Losing a background potassium conductance depolarises the cell. The downstream consequences were mapped mostly on spadin, and they connect the channel to serotonin without touching a transporter. In vivo electrophysiology showed spadin raising the firing rate of dorsal raphe serotonergic neurons by 113 percent, an increase abolished by electrolytic lesion of the medial prefrontal cortex and occluded by the mGluR2/3 antagonist LY 341495. The effect therefore depends on prefrontal TREK-1 channels coupled to mGluR2/3 receptors, and it is additive with, rather than downstream of, 5-HT4 receptor stimulation.

The plasticity arm runs on a faster clock than monoaminergic drugs. A four-day course of PE 22-28 in mice roughly doubled BrdU-positive cell counts in the hippocampus, and in cultured mouse cortical neurons 0.1 micromolar PE 22-28 raised the postsynaptic scaffold protein PSD-95 about twofold between 5 and 36 hours. Work on spadin traced this to MAPK and PI3K signalling and to a rapid rise in hippocampal BDNF, explicitly not through mTOR, which is the route ketamine uses. That distinction matters: two fast-acting mechanisms, two different intracellular pathways.

The starting problem was durability, not potency. Spadin worked in mice but its behavioural effect was gone beyond seven hours after a single injection, and an earlier attempt to fix that with retro-inverso analogues produced two peptides that extended the effect to about 16 hours but showed in vitro toxicity at higher concentrations and cost too much to synthesise. So the group went the other way and asked what the body was already doing to spadin.

Incubating spadin with mouse serum at 37 degrees Celsius and running the products on reversed-phase HPLC showed almost all of the parent peptide gone by 30 minutes, replaced by two peaks that mass spectrometry identified as PE 14-25 and PE 12-27. Those two, plus three peptides designed around them, were screened at 100 nanomolar against arachidonic-acid-activated human TREK-1. The ladder is informative. PE 14-25 did nothing at all, at 114.7 plus or minus 10.6 percent of the arachidonic acid current (n = 8, p = 0.09). PE 12-27 blocked 28.4 plus or minus 9.9 percent. PE 22-27 gave 25.7 plus or minus 20.0 percent and did not reach significance, PE 22-25 gave 36.0 plus or minus 17.5 percent at p = 0.06, and only PE 22-28 blocked cleanly, at 55.5 plus or minus 4.6 percent.

Two structure-activity conclusions fall out of that screen. Removing the single C-terminal arginine, which is the only difference between PE 22-28 and PE 22-27, destroys the activity. And the seven-residue window is genuinely minimal: shorten it at either end and the block goes. The group then built 16 analogues on PE 22-28 as a core. Most N-terminal additions killed it, including stearic acid at 5.4 percent block, the Fmoc group at 5.8 percent and 5-FAM at 6.6 percent. Only three peptides survived the combined electrophysiology and behavioural filter: PE 22-28 itself, its glycine-to-alanine variant at position 22, and the biotinylated form of that variant.

The 2006 knockout paper is the reason this compound class exists. Deleting Kcnk2 in mice produced resistance to depression-like behaviour across five separate models, increased the efficacy of serotonergic neurotransmission, and substantially blunted the corticosterone rise under stress. The knockouts behaved like naive animals treated with fluoxetine, without any drug. That is an unusual result for a background potassium channel, which had until then been studied mostly in the contexts of anaesthesia, pain and ischaemia.

The human evidence for the target is genetic rather than pharmacological, and it is more interesting than most summaries suggest. In the STAR*D effectiveness cohort, variants in KCNK2 showed no association with remission on first-step citalopram among 1554 genotyped participants, but among the 751 who went on to next-step treatment they were associated with response at a Bonferroni-corrected gene-based empirical p below 0.001, with a similar effect size again at the third step. The signal is specific to treatment resistance, not to depression itself. Separately, an fMRI study in 31 healthy volunteers found that the three TREK1 genotypes previously linked to positive antidepressant response were associated with potentiated basal ganglia responses to monetary gains, with no effect on responses to losses.

None of that is evidence for PE 22-28. It is evidence that the channel it blocks is doing something in humans, which is a different and weaker claim, and the two are worth separating when reading anything written about this peptide.

Spadin is 17 residues, APLPRWSGPIGVSWGLR, and weighs about 1849 g/mol. PE 22-28 is its C-terminal seven residues, GVSWGLR, at 773.89 g/mol. On the channel, the shorter peptide is more than 300 times more potent: 0.10 nanomolar against 40 to 60 nanomolar. In behaviour the ratio holds up. A comparable reduction in forced swim immobility needed 100 micrograms per kilogram of spadin against 3.0 micrograms per kilogram of PE 22-28, and in the corticosterone model PE 22-28 at 3.0 micrograms per kilogram slightly outperformed spadin at 100 micrograms per kilogram, at 98.1 plus or minus 8.8 seconds against 117.4 plus or minus 6.9 seconds, with a control value of 164.9 plus or minus 6.0 seconds.

Duration is where the record needs a caveat that is almost never stated. Spadin has a measured half-effect time of about 6 hours in the forced swim test. The 14 to 23 hour figures widely attributed to PE 22-28 come from a time-course experiment run on the two derivatives, not on the plain peptide. G/A-PE 22-28 gave 14 hours at 3.2 micrograms per kilogram and 21 hours at 32 micrograms per kilogram, and the biotinylated variant gave 17 and 23 hours at 4.0 and 40 micrograms per kilogram. PE 22-28 itself was not in that time course, and its own action duration has not been published.

The two peptides also carry different safety evidence. The wider panel showing no change in cardiac rate or systolic pressure, no worsening of kainate-induced seizures, no increase in infarct size after focal ischaemia and no change in glycaemia was run on spadin in 2012. For PE 22-28 the record is narrower: four other two-pore-domain channels and hERG.

Selectivity is the strongest part of this compound's in vitro file. Native HEK293 cells were transfected with human TREK-2, TRAAK, TRESK and TASK-1. TREK-2 and TRAAK were opened with 10 micromolar arachidonic acid, then 100 nanomolar PE 22-28 was applied at peak current. No change in amplitude was seen in any of the four. TREK-1 and TREK-2 share 63 percent identity and TREK-1 and TRAAK 45 percent, so discriminating between them at a hundredfold above the TREK-1 IC50 is a meaningful result rather than a formality.

The hERG test was run at 10 micromolar, five orders of magnitude above the working concentration at TREK-1, and left both the end-of-pulse and tail current densities unchanged. That matters because drug-induced block of IKr is the standard mechanism behind QT prolongation and torsades de pointes, and it is a recurring liability in psychoactive drug development.

TREK-1 itself, however, is expressed well outside the brain: in cardiac tissue, vascular smooth muscle, the gastrointestinal tract, adipocytes, chondrocytes and pancreatic beta cells. A clean hERG result says nothing about what blocking TREK-1 in those tissues would do. Recent work has implicated TREK-1 in adipocyte differentiation, in NLRP3 inflammasome activity in macrophages, in alveolar epithelial injury and in astrocytic potassium handling, so this is not a brain-specific target, and the consequences of blocking it elsewhere have not been characterised for this peptide.

Four paradigms carry the behavioural case, all in male C57Bl/6J mice aged 7 to 9 weeks, with 10 animals per group. In the Porsolt forced swim test, a single intraperitoneal injection of 3.0 micrograms per kilogram cut immobility over the final 4 minutes of a 6-minute session to 91.8 plus or minus 6.1 seconds from a saline value of 161.7 plus or minus 6.5 seconds. In the learned helplessness test, four days of the same amount shortened escape latencies across 30 trials following a four-day inescapable shock training session. In the novelty-suppressed feeding test, four days of treatment cut the latency to start eating in a brightly lit open arena to 153.2 plus or minus 5.4 seconds from 226.1 plus or minus 35.0 seconds.

The fourth is a chemically induced model rather than an acute screen. Seven weeks of corticosterone produces a depressive-like state in mice, and both acute and four-day PE 22-28 reversed the forced swim and novelty-suppressed feeding readouts in those animals. That is a more demanding test than the acute paradigms, because the animals start from an altered baseline.

The honest framing is that these are screening assays. Reduced immobility in a forced swim test is produced by many compounds, including psychostimulants that nobody would call antidepressants, and a single behavioural readout cannot separate an antidepressant-like effect from increased locomotion. What raises the interest here is convergence: the same 3.0 micrograms per kilogram that shifts these behaviours also produces the neurogenesis and synaptic markers on the same four-day clock, and the compound has a defined molecular target with an IC50 measured on cloned human channels.

The neurogenesis experiment used BrdU incorporation across the whole hippocampus, with eight 40-micrometre vibratome sections from bregma 3.3 to 5.3 counted per animal. After four daily intraperitoneal injections at 3.0 to 4.0 micrograms per kilogram, counts were 1736 plus or minus 126 for PE 22-28, 2110 plus or minus 132 for its glycine-to-alanine variant and 1809 plus or minus 267 for the biotinylated variant, against 899 plus or minus 109 for saline, with n = 5 per group. The four-day timeframe is the point of the experiment. Chronic monoaminergic treatment raises this marker only after two to four weeks, and that lag is one of the standing explanations for why those drugs take weeks to work.

Synaptogenesis was read as PSD-95 abundance by Western blot in primary mouse cortical neurons at 0.1 micromolar peptide. Levels dipped slightly over the first 5 hours, then rose continuously, ending roughly twice as high at 36 hours. Taken with the neurogenesis counts, the interpretation offered is that most of the newly born cells go on to mature rather than being pruned.

The mechanistic detail behind this came from the parent peptide. Spadin activates MAPK and PI3K in cultured neurons in a time and concentration dependent way, and the PI3K arm accounts for its protection against staurosporine-induced apoptosis. It raises both mRNA and protein for PSD-95 and synapsin, increases the proportion of mature dendritic spines in cortical neurons, and produces a rapid rise in hippocampal BDNF after injection. Critically for the comparison people want to make, spadin does not go through mTOR, which is the pathway that carries ketamine's rapid effect. Two fast-acting mechanisms, two different routes.

The 2019 stroke study is the most complicated result in this literature and the one that most changes how the peptide should be read. Working in a mouse model of focal ischaemia with post-stroke depression, the group found that the peptide, which they call mini-spadin in that paper, acts biphasically on TREK-1: at low concentrations it increases channel activity, and only at higher concentrations does it inhibit. They exploited that directly. A single low amount, 0.03 micrograms per kilogram given intraperitoneally 30 minutes after ischaemia onset and then once daily for seven days, was used to activate the channel and produce neuroprotection. The peptide was then given at 3 micrograms per kilogram, once daily four days a week, to block it and produce the antidepressant-like effect. Across that protocol the animals limited the loss of body weight, showed less delayed dopaminergic degeneration in the substantia nigra, improved on motor and cognitive tests, and showed reduced forced swim and novelty-suppressed feeding signals, with enhanced neurogenesis and synaptogenesis alongside.

A biphasic concentration response means the direction of the effect at this target depends on concentration, which is not a property that survives casual translation between species, routes or timeframes. It is the single most important qualification on everything else written about the compound.

A second line of work sits outside the nervous system. TREK-1 is expressed in pancreatic beta cells, and the sortilin propeptide stimulates insulin secretion only in the presence of glucose. In 2021 the same laboratory reported that the propeptide, spadin and mini-spadin all protected beta cells against interleukin-1-beta-induced death by raising intracellular calcium through membrane depolarisation, working through the calmodulin kinase pathway and phosphorylation of CREB, with mini-spadin additionally promoting beta-cell proliferation. The unifying mechanism, depolarisation through loss of a background potassium current, is the same one that operates in neurons.

The composition dictates the handling. GVSWGLR contains no cysteine, so disulfide scrambling is not a failure mode, and no methionine, so the classic oxidation liability is absent. It does contain a single tryptophan, which is the residue most vulnerable to photo-oxidation in a short chain and the reason amber vials and dark storage are not decorative. That same tryptophan is analytically useful. With no tyrosine in the sequence, absorbance at 280 nanometres reports on one residue only and gives a clean concentration check independent of gravimetric mass.

Proteolytic stability was the original design problem. The parent 17-mer was almost entirely gone from mouse serum within 30 minutes at 37 degrees Celsius, and the degradation was traced to metalloproteases using 1,10-phenanthroline as an inhibitor. Shortening the peptide was the response to that, though the shortened form still has a short in vitro serum half-life. What improved was the duration of the functional effect, not the resistance of the molecule in a tube.

Two termini decide whether a batch is the right molecule. Removing the C-terminal arginine, giving PE 22-27, abolishes channel block. Adding bulk to the N-terminal glycine abolishes it in most cases: stearic acid, the Fmoc group, 5-FAM, myristoyl and palmitoyl derivatives all fell below significance, while biotin, dansyl and a glycine-to-alanine substitution were tolerated. A certificate of analysis that reports purity by area percent without confirming the intact mass therefore leaves the pharmacologically decisive part of the molecule unverified. Reversed-phase C18 HPLC with an acetonitrile gradient from 20 to 70 percent is the separation used in the source work, and mass spectrometry is what distinguished the degradation fragments in the first place.

One practical note on labelling. Peptides of this class are supplied as an acetate or trifluoroacetate salt, so the gravimetric contents of a vial exceed the free-base peptide content. Net peptide content belongs on the analysis, not inferred from the label.

There is no human data of any kind. A search of ClinicalTrials.gov for spadin and for PE 22-28 returns no registered studies, and neither the parent peptide nor any shortened analogue has entered a Phase 1 trial in the two decades since the target was proposed.

There is also no pharmacokinetics. No plasma concentration, no measured half-life, no clearance figure and no brain penetration measurement has been published for PE 22-28 in any species. The 14 to 23 hour numbers in circulation are behavioural half-effect times for two derivatives, and a behavioural duration is not a pharmacokinetic parameter. Whether the peptide crosses the blood-brain barrier intact, or acts through a peripheral route, is unresolved. The group synthesised biotinylated variants precisely because biotinylation has been reported to improve peptide brain uptake, which implies uptake was a live question.

The evidence base is narrow in the way that matters most. TREK-1 as a channel and spadin as a tool compound have been taken up by other laboratories, including a 2020 Korean study that used spadin to establish that TWIK-1/TREK-1 heterodimers carry astrocytic passive conductance. PE 22-28 specifically appears almost only in publications from the Valbonne group and its collaborators. The key neurogenesis result rests on five animals per group and the behavioural work on ten. There is no published toxicology, no repeat-exposure study, no genotoxicity or reproductive data, and the off-target screen covered four related potassium channels plus hERG rather than a broad panel.

One further gap is specific to how this compound gets described in product catalogues. Because it is short and potent, and because the parent peptide carries a much larger body of work, findings from spadin are routinely attributed to PE 22-28 without the substitution being flagged. The serotonergic electrophysiology, the BDNF and mTOR mechanism, the cardiac and seizure panel, and the effect on 5-HT firing rate were all measured on spadin. Whether they hold for the seven-residue fragment is a reasonable inference and not a finding.