If you want a clean illustration of how far ahead of the evidence the peptide conversation runs, watch what happens to a single animal study. In December 2025, a Russian group published the first head-on test of Semax in a transgenic Alzheimer's mouse model — 80 animals, three behavioural tests, and brain histology. Within weeks the finding was circulating as though Semax had been shown to treat Alzheimer's disease. It hasn't. What is Semax research actually showing here is narrower and more interesting than that: a 1980s ACTH fragment analogue, still unapproved outside Russia, producing a plaque signal in one of neuroscience's least predictive models.

What is Semax, and why is an Alzheimer's model new for it?

Semax is a synthetic heptapeptide — Met-Glu-His-Phe-Pro-Gly-Pro — built from the ACTH(4–7) fragment with a proline-glycine-proline tail bolted onto the C-terminus. The tail slows proteolysis; the truncation strips out the steroidogenic activity of full-length ACTH. It came out of the Institute of Molecular Genetics in Moscow in the early 1980s, and the published literature on Semax has overwhelmingly been about two things: ischaemic stroke and cognition/learning in rodents.

Amyloid is a different problem. Most of Semax's rodent work has involved acute insults — occlude an artery, then measure what the peptide does to gene expression, infarct volume or behaviour. Chronic, slowly accumulating amyloid pathology is a much harder target, and until this paper Semax had essentially not been characterised against it in a transgenic model.

The 80-mouse study, described plainly

The work was published by Radchenko and colleagues in Acta Naturae (Vol. 17, No. 4, 2025). The design:

  • Animals: 60 male APPswe/PS1dE9/Blg (APP/PS1) mice on a mixed C57Bl6 background, plus 20 male wild-type C57Bl6 controls — 80 in total.
  • Compounds: Semax itself, and a derivative heptapeptide, Met-Glu-Asp-Arg-Pro-Gly-Pro, in which the His-Phe pair of the ACTH(4–7) core is substituted while the Pro-Gly-Pro tail is retained.
  • Readouts: open field, novel object recognition and Barnes maze for behaviour; histological counting of amyloid inclusions in cortex and hippocampus.

The reported result: both peptides improved cognitive performance relative to untreated transgenic animals, and histology showed a reduced number of amyloid inclusions in both cortex and hippocampus. The authors' own framing is that this demonstrates "potential" for developing corrective strategies — not efficacy.

Two details matter more than the headline. First, the derivative is not Semax. Swapping His-Phe for Asp-Arg produces a different molecule with a different pharmacology, no registration anywhere, and no human exposure history. Any reader who sees "Semax and its derivative worked" and buys Semax has already lost the thread. Second, some of the behavioural effects described are locomotor as much as cognitive — Semax was reported to increase velocity and reduce distance travelled in the maze. In rodent work, changes in movement and exploration can inflate or confound apparent learning scores, which is exactly why a single lab's behavioural battery is a starting point rather than a conclusion.

Why plaque reduction in APP/PS1 mice is weak evidence

This is the part usually left out. APP/PS1 mice carry human familial Alzheimer's mutations driven to overexpression; they develop plaques on a compressed timeline that sporadic, late-onset human disease does not follow. The model is useful for mechanism. Its record as a predictor of clinical benefit is poor, and it is not even internally consistent — the same intervention can move plaque load in one APP/PS1 line and not another, as reported for long-term liraglutide across two transgenic models.

So the honest summary is: one lab, one paper, 80 animals, unreplicated, no human data. It is a legitimate reason to run the next experiment. It is not a reason to change anything you do.

Semax in Australia: the regulatory position hasn't moved

Here is the Australian angle, and it is unchanged by this study. Semax is unscheduled in the Poisons Standard — and that is routinely misread as permission. It isn't. Semax is an unapproved therapeutic good in Australia: it holds no ARTG registration, which means importing or supplying it outside a TGA pathway (Special Access Scheme, Authorised Prescriber, or an approved clinical trial) is unlawful. "Not scheduled" and "legal to buy" are separate questions, and only the second one matters practically.

Elsewhere: Semax is a registered medicine in Russia as nasal drops and appears on Russia's Vital & Essential Drugs list — which is why Russian-language clinical literature exists at all. But there are no internationally registered trials of Semax, which is the single most important fact about its evidence base and the reason no Western regulator has evaluated it. For athletes, Semax does not appear on the WADA 2026 Prohibited List, though anti-doping specialists have described its status as unclear rather than cleared — a distinction worth respecting if you are tested.

One caution on sourcing, since it applies across this whole space: clinical trial registries now contain fictional demonstration records for peptides, complete with plausible-looking "recruiting" statuses and invented sponsors. If you see a Semax trial listed, check that the sponsor is a real, verifiable organisation before you believe it.

What would actually move the needle

A short list, in order of value:

  1. Independent replication of the amyloid histology in a different lab and a different APP/PS1 line.
  2. Pharmacokinetics — route, brain exposure, dose-response — which remain thinly characterised for Semax relative to how confidently it is discussed.
  3. A registered human trial anywhere outside Russia. Until one exists, every Semax claim rests on rodent work plus a regional regulatory approval that no other agency has independently reviewed.

For a wider view of how other compounds sit against the same evidence bar, see our peptides index. And if you want new primary literature summarised as it lands, rather than three months later as a marketing claim, join the free updates list.

FAQ

What is Semax?

Semax is a synthetic heptapeptide (Met-Glu-His-Phe-Pro-Gly-Pro) derived from the ACTH(4–7) fragment with a Pro-Gly-Pro tail added to resist enzymatic breakdown. It was developed in Moscow in the early 1980s and has mostly been studied in rodent models of ischaemic stroke and learning.

Is Semax legal in Australia?

Semax is unscheduled in the Poisons Standard but is an unapproved therapeutic good with no ARTG registration, so importing or supplying it outside a TGA pathway such as the Special Access Scheme or an approved clinical trial is unlawful.

Does Semax work for Alzheimer's disease?

There is no human evidence. A December 2025 study in 80 mice reported improved cognitive test performance and fewer amyloid inclusions with Semax and a derivative peptide, but it is a single unreplicated animal study in a model with a poor record of predicting clinical benefit.

Has Semax been tested in humans outside Russia?

No. Semax is registered in Russia as nasal drops and sits on Russia's Vital & Essential Drugs list, but there are no internationally registered clinical trials, so no Western regulator has evaluated it.

Sources

This article is independent information, not medical advice. Semax is an unapproved therapeutic good in Australia with no ARTG registration and no internationally registered clinical trials; nothing here is a recommendation to obtain or use it.