Semax vs Selank

Published August 2026 · Patriot Labs Research Team · ~11 min read

Why these two get lumped together

Semax and Selank arrive in the same paragraph so often that the pairing feels like a fact about the molecules. It is really a fact about their history. Both were developed in Russia, both came out of work associated with the Institute of Molecular Genetics of the Russian Academy of Sciences, both are seven amino acids long, both are given intranasally in most of the published animal work, and both have literatures dominated by the same handful of research groups.

That shared provenance is real. What it is not is evidence that they do similar things. Set the two sequences next to each other and the resemblance is confined to the last three residues:

Semax — Met–Glu–His–Phe–Pro–Gly–Pro
Selank — Thr–Lys–Pro–Arg–Pro–Gly–Pro

Four residues of business end, three residues of shared tail. The tail is the same idea in both cases; the business ends come from molecules that have nothing to do with each other. The peptide basics guide covers the naming vocabulary this one assumes.

The shared design: a fragment plus a tail

Both molecules are the product of the same design problem. Short endogenous peptides are frequently potent and almost always fragile. Blood and tissue are full of aminopeptidases, carboxypeptidases and endopeptidases whose job is to dismantle exactly this kind of molecule, and a free tetrapeptide put into that environment tends not to last. A researcher who wants to study what a particular four-residue sequence does has a practical problem before they have a biological one.

The strategy applied to both is to leave the bioactive fragment alone and defend it from one end. In each case the C-terminus was extended with the tripeptide Pro–Gly–Pro. Published descriptions of Semax state that the PGP tripeptide “was included to ensure the resistance of Semax to peptidases,” and frame it as conferring greater metabolic stability than natural melanocortins. Published descriptions of Selank state that the peptide was elongated at the C-terminus by three natural L-amino acids — Pro-Gly-Pro — to improve metabolic stability and yield a relatively longer duration of action. Proline is conformationally awkward for proteases, its ring locks the backbone, and many peptidases will not cleave adjacent to it. Two unrelated fragments, one identical piece of armour.

The armour is partial, not absolute, and this is worth stating plainly. A study using isotopically labelled Semax incubated with rat basal forebrain cell cultures and plasma membranes reported that the splitting away of Met-Glu from one end and Gly-Pro from the other, with formation of pentapeptides, were the predominant degradation processes, and that glial and neuronal preparations produced different degradation patterns. So the tail slows breakdown; it does not prevent it, and it is itself cleaved. Anyone reading a claim that PGP makes these peptides “protease-resistant” should read that as relative, not categorical. The general chemistry of how peptide bonds come apart is covered in the degradation guide.

There is a further complication that makes the tail more than inert scaffolding. In a rat focal-ischaemia experiment, Pro-Gly-Pro administered on its own altered transcription of neurotrophin and neurotrophin-receptor genes, enhancing Bdnf and TrkC transcription at three hours after occlusion and Ngf, TrkB, TrkC and TrkA at twenty-four hours. The authors described Semax's effect as more selective for the ischaemic tissue and PGP's influence as largely non-specific. That means the shared tail is not a neutral handle: it is a fragment with its own reported activity, present in both molecules.

Two very different parents

Now the part that gets flattened in most side-by-side write-ups.

Semax carries a fragment of ACTH. Adrenocorticotropic hormone is the pituitary output of the stress axis — the signal that drives adrenal corticosteroid release. Semax's N-terminal four residues, Met-Glu-His-Phe, correspond to positions 4–7 of that hormone. Crucially, this is a fragment chosen because it is not the hormonal part. Published accounts of the molecule state directly that Semax does not exhibit hormonal activity, and describe it as operating without the hormonal and toxic effects associated with the parent compound. The same Met-Glu-His-Phe stretch is also present in the α-MSH sequence, which is why the current literature more often classifies Semax as a melanocortin derivative than as an ACTH analogue.

A naming quirk follows from this and causes real confusion. Older papers — including much of the 1990s and 2000s work — call Semax an “ACTH(4–10) analogue,” because it is a seven-residue molecule occupying the same positional territory. More recent papers from the same groups call it ACTH(4–7)PGP, which is the more accurate description: only four ACTH-derived residues are retained, and positions 8–10 have been replaced outright by the Pro-Gly-Pro tail. Both labels refer to the same compound.

Selank carries tuftsin. Tuftsin is not a hormone and has nothing to do with the stress axis. It is a tetrapeptide, Thr-Lys-Pro-Arg, that sits inside an antibody: it corresponds to residues 289–292 of the heavy chain of γ-globulin, released enzymatically from that larger protein. Its classical description is as a natural immune activator — a stimulator of phagocytosis in macrophages and polymorphonuclear granulocytes, with reported antibacterial and antineoplastic activity and effects on leukocyte chemotaxis. Selank is that tetrapeptide with the same PGP tail appended, and is routinely described in the literature as a synthetic analogue of the endogenous tuftsin molecule.

The two starting materials are about as far apart as two short human peptides can be: one is a piece of an endocrine signalling hormone with the endocrine part deliberately left out, the other is a piece of an antibody whose native job is telling phagocytes to eat things. Whatever these molecules turn out to do, there is no structural reason to expect them to do the same thing.

Head to head

What has actually been studied for Semax

The Semax literature clusters around three themes.

Neurotrophins. The most-cited mechanistic thread concerns brain-derived neurotrophic factor and nerve growth factor. In rats subjected to permanent middle cerebral artery occlusion, Semax enhanced transcription of Bdnf, TrkC and TrkA at three hours after occlusion, Nt-3 and Ngf at twenty-four hours, and Ngf again at seventy-two hours. The reported effect was selective for ischaemic tissue in a way that Pro-Gly-Pro alone was not. This is transcript-level work in a lesion model — it describes changes in gene expression, not a demonstrated functional outcome in an intact organism.

Ischaemia and reperfusion. A transient middle cerebral artery occlusion model in Wistar rats — ninety-minute occlusion, then reperfusion — was used for RNA-sequencing at twenty-four hours. Semax was associated with 394 differentially expressed genes, suppressing inflammation-associated transcripts and activating neurotransmission-associated ones. A companion protein-level study in the same model reported upregulation of active CREB in subcortical structures, downregulation of MMP-9 and c-Fos in adjacent cortex, and suppression of active JNK in both. The RNA-Seq groups were three animals each — normal for sequencing work, and a real constraint on inference.

Monoamines. In rodents, Semax was reported to raise striatal tissue content of the serotonin metabolite 5-HIAA by roughly 25% two hours after administration, with a larger rise in extracellular 5-HIAA over one to four hours. The dopaminergic finding is more conditional: Semax alone did not alter baseline dopamine or its metabolites, but given before d-amphetamine it markedly amplified amphetamine's effect on extracellular dopamine and on locomotor activity. That is a modulatory result, not a direct agonist result, and it only appears under challenge.

Beyond these, a study in transgenic APPswe/PS1dE9 mice reported reductions in amyloid plaque burden and changes on behavioural testing with Semax and a derivative. That paper also states that Semax appears on the Russian List of Vital and Essential Drugs for Medical Application, citing Decree No. 2406-r of 12 October 2019. We were not able to independently verify the text of that decree, so treat the registration detail as reported in the peer-reviewed literature rather than as something we confirmed at source.

What has actually been studied for Selank

The Selank literature clusters differently.

GABAergic signalling. This is the dominant mechanistic hypothesis. A binding-oriented study concluded that one of Selank's anti-anxiety molecular mechanisms may be associated with subtype-selective, concentration-dependent allosteric modulation of GABA receptors, describing the peptide as a positive allosteric modulator and reporting non-cumulative interactions with diazepam and olanzapine. Separately, intranasal Selank in male Wistar rats altered expression of 45 genes in frontal cortex at one hour and 22 at three hours, with Gabre, Gabrq and Hcrt showing changes of up to 128-fold at the three-hour timepoint.

The cell-culture picture is more equivocal, and the equivocation is instructive. In human IMR-32 neuroblastoma cells, which express functional GABA_A receptors, Selank alone produced no change in the mRNA levels of any of the 84 genes examined. Its effect appeared only in combination: it nearly completely suppressed the expression changes GABA alone produced, and it amplified the effects of olanzapine. A molecule that does nothing alone in a dish and something clear in combination is a modulator, and modulators are harder to characterise than agonists.

Immune and inflammatory signalling. This is the inheritance from tuftsin, and it is the part of Selank's profile with no counterpart in Semax. A single administration in mice changed expression of genes for chemokines, cytokines and their receptors in spleen at six and twenty-four hours, and the effect was not confined to the full heptapeptide — the Gly-Pro fragment produced changes across most of the genes studied. Work in rats also reported effects of both tuftsin and Selank on haemostatic parameters: fibrin-depolymerisation activity in vitro, and combined antiplatelet, anticoagulant and fibrinolytic effects in plasma after intranasal administration, with Selank exceeding native tuftsin.

Selank's anxiolytic-pathway framing is why it is often discussed alongside compounds such as DSIP, though those act through unrelated proposed mechanisms and the two literatures do not connect.

Where they do overlap

Two genuine overlaps exist beyond the shared tail, and both are worth knowing because they are the most-abused points in vendor copy.

Enkephalin-degrading enzymes. An in-vitro study using human serum reported that both Semax and Selank inhibit enkephalin-degrading enzymes in a concentration-dependent manner, with potency exceeding reference inhibitors, and that pentapeptide fragments retained inhibitory activity while tri-, tetra- and hexapeptide fragments did not. The authors proposed this as a possible common basis for the activity of both peptides. It is a plausible shared mechanism, and it is also a serum biochemistry finding — not a demonstration that this pathway accounts for anything observed in an animal.

Direct comparisons are rare and messy. Two studies have looked at both molecules in one design. A behavioural study in BALB/c and C57BL/6 mice compared intraperitoneal and intranasal routes for Selank, Semax and Noopept and found strain- and route-dependence rather than a clean profile: in BALB/c mice all three peptides improved exploratory activity and reduced anxiety-related measures by both routes, with intranasal administration favouring cognitive endpoints and intraperitoneal favouring anxiolytic ones — while C57BL/6 mice showed almost no behavioural response, and intraperitoneal Semax in that strain produced an anxiety-increasing effect. A resting-state fMRI study in 52 healthy human participants scanned at three timepoints reported differences in functional connectivity between the right amygdala and a temporal-lobe region, and described both general and specific effects of the two peptides.

The honest reading of both is that the peptides are neither equivalent nor cleanly separable, and that the direction of an effect can flip with the strain of mouse and the route of administration. For context on where these two sit relative to the rest of the class, see the cognition-focused peptide overview.

Evidence at a glance

What the evidence does not establish

The single most important limitation applies to both, equally. The published literature on Semax and Selank is heavily concentrated in one research network and one language. A large share of the primary work carries authors associated with the Institute of Molecular Genetics of the Russian Academy of Sciences, and appears in Russian-published journals or their English translations. Several of the clinical claims that circulate about both peptides trace back to Russian-language reports we could not retrieve and read in full. Independent replication by unaffiliated groups outside that network is, in the sources we located, sparse to absent. That is not an accusation; it is a statement about the structure of the evidence. Findings not reproduced by an independent laboratory carry less weight than findings that have been, and that discount applies to nearly everything below.

Registered status in Russia is not verified evidence of anything. A peer-reviewed paper reports Semax's inclusion on a Russian essential-medicines list by decree; we cite that report but did not verify the decree text at source, and we could not locate an equivalent verified document for Selank. National registration is also a regulatory fact, not a scientific one — it says a body reviewed a dossier under one country's rules. It is not a claim we are entitled to make about these materials, and neither compound is approved for human use in the United States.

Most of the mechanistic work is transcript-level. A very large fraction of the Semax literature, and a substantial part of the Selank literature, consists of measuring which genes change expression after administration. Transcript changes are a real observation and a poor substitute for a functional one: mRNA does not always become protein, protein does not always become activity, and 394 differentially expressed genes at one timepoint in one lesion model is a description of a system in flux, not a mechanism. RNA-Seq group sizes in the published designs are small — three animals per group in the ischaemia work.

The receptor targets are not settled. Published discussion of Semax refers to suggestions of allosteric interaction with a large number of different receptors, without a single established primary target. For Selank, GABA_A allosteric modulation is the leading hypothesis, but the cell-culture result — no effect on any of 84 genes when applied alone — sits awkwardly with a simple receptor story. Neither peptide has a clean, independently confirmed molecular target in the sources reviewed here.

The Pro-Gly-Pro tail confounds both. Free Pro-Gly-Pro has its own reported activity on neurotrophin gene transcription, and a Gly-Pro fragment alone changed immune gene expression in mouse spleen. Any effect seen with either heptapeptide could therefore be partly a tail effect, a fragment effect, or a metabolite effect, and both molecules are degraded to smaller pieces in tissue. Attributing an observation to the intact heptapeptide requires controls that are frequently absent.

Direction of effect is not stable across models. The BALB/c versus C57BL/6 comparison is the clearest warning available: the same peptide, in the same experiment, produced anxiety-reducing measures in one strain and an anxiety-increasing effect in the other by one route. Anyone treating rodent behavioural findings for these peptides as a fixed property of the molecule is over-reading them.

Human data located here is minimal. The EEG study we retrieved involved nine volunteers against six controls and dates from 1996. The imaging study involved 52 participants and reports connectivity differences — a physiological observation, not an outcome. Neither is a controlled clinical trial, and neither is evidence of benefit for anything.

Nothing transfers between the two. Sharing a C-terminal tripeptide does not make Selank a source of evidence about Semax or the reverse. The two direct comparisons that exist found differences, not equivalence. Reading across from one peptide's literature to the other is the most common error in the material written about this pair.

Frequently asked questions

Are Semax and Selank the same kind of molecule? They are built the same way but from different parts. Both are synthetic seven-residue peptides in which a short fragment of a larger endogenous molecule is extended at the C-terminus with Pro-Gly-Pro to improve resistance to peptidases. The fragments are unrelated: Semax carries Met-Glu-His-Phe from ACTH, Selank carries Thr-Lys-Pro-Arg from tuftsin.

What does the Pro-Gly-Pro tail actually do? Published descriptions of both peptides state it was added to confer peptidase resistance and extend duration of action. It is not absolute protection — a radiotracer study of Semax against rat basal forebrain cell cultures and plasma membranes found that loss of the N-terminal Met-Glu and the C-terminal Gly-Pro, with pentapeptide formation, were the predominant degradation routes.

Is Semax derived from a stress hormone? It carries ACTH(4–7), four residues of adrenocorticotropic hormone. Published accounts state that Semax itself does not exhibit hormonal activity, and the same four residues occur in the α-MSH sequence, which is why current papers usually call it a melanocortin derivative rather than an ACTH analogue.

Why is Selank described as an immune peptide as well as a brain peptide? Because its parent is one. Tuftsin is a fragment of an immunoglobulin heavy chain, classically characterised as a stimulator of phagocytosis. Selank has been reported to alter chemokine and cytokine gene expression in mouse spleen and to affect haemostatic parameters in rats, which is a profile Semax does not share. Comparative details are on the Selank product page.

Which one has better evidence? Neither, in the sense the question usually means. Semax has a larger and more mechanistically detailed body of rodent lesion-model and transcriptomic work. Selank has more direct receptor-level work and a distinct immunological literature. Both share the same core weakness — concentration in a single research network with little independent replication — so a larger volume of publications does not translate into greater confidence.

Can findings for one be applied to the other? No. A shared C-terminal tripeptide is a formulation-stability feature, not a shared mechanism. The two studies that examined both peptides in a single design reported differences between them.

Is it approved for human use? No. Neither Semax nor Selank is an approved drug in the United States, and neither is approved for human use. All Patriot Labs products, including Semax and Selank, are sold strictly for in-vitro research and laboratory use only, and are not for human or veterinary consumption.

All Patriot Labs products are sold strictly for in-vitro research and laboratory use only. Not for human or veterinary consumption. This guide is educational and describes peptide chemistry and published research in general terms; it is not medical advice, does not describe how to use any product, and the references cited do not constitute a product claim.