Brain Peptides Explained: What the Research Actually Shows

Neuropeptides for brain health are having a moment — Cerebrolysin, Semax, Selank, Dihexa, P21, and FGL are the six names that keep coming up. Before getting excited about any of them, three questions matter, in order: does the molecule actually reach the brain, is there documented evidence it does something useful once it's there, and what's its current regulatory status.

The real bottleneck: the blood-brain barrier

The blood-brain barrier is built from tightly sealed endothelial cells lining the brain's capillaries, and it's extremely good at blocking most therapeutic molecules by design. Only small, fat-soluble molecules under roughly 400-500 daltons cross it passively — a threshold that rules out more than 98% of small-molecule drugs and essentially all biologics in their native form.

That's why intranasal delivery shows up so often with these compounds — it reaches the CNS via the olfactory and trigeminal nerves, largely bypassing the bloodstream and the barrier. It's genuinely useful, but technique-sensitive: incorrect head angle, spray placement, or a hard inhale sends the dose down the throat instead of into the nasal cavity where it can be absorbed.

Cerebrolysin

A mixture of low-molecular-weight peptide fragments extracted from porcine brain protein, in clinical use since the 1970s and approved in roughly 45 countries (not the US). Given by IV or IM, not intranasally. The evidence is genuinely mixed by indication:

• Acute ischemic stroke: a 2023 Cochrane review of 7 trials, 1,773 patients, found no survival benefit and a signal toward more non-fatal serious adverse events.¹
• Stroke rehabilitation: the CARS trial found a large effect on upper-limb motor recovery at 90 days — flagged by its own authors as exploratory.²
• TBI: the CAPTAIN trial series missed its primary endpoint individually, but a pooled analysis (185 patients) found benefit at 30 and 90 days.³
• Vascular dementia: improved cognition with no safety signal, but insufficient evidence to recommend routine use per Cochrane.⁴

Semax

A synthetic seven-amino-acid ACTH analog developed in Russia in the 1980s for stroke recovery, since expanded to cognitive support, anxiety, and neuroprotection in Russian clinical practice. Animal studies show a single dose measurably raising hippocampal BDNF and TrkB, with improved learning task performance.⁵ Anecdotally, effects are often noticed within days rather than weeks — unusual for this category.

Selank

Semax's sibling compound, built from tuftsin rather than ACTH, and fully approved in Russia as a prescription intranasal product. Its niche is anxiety, mild depression, and asthenia rather than stroke recovery. The most cited comparison pitted Selank against the benzodiazepine medazepam for generalized anxiety and found comparable relief without sedation, dependency risk, or cognitive dulling.⁶

Dihexa (and successors ATH-1017, ATH-1105)

Dihexa was engineered at Washington State University in the 2010s as a stabilized angiotensin IV analog designed to cross the blood-brain barrier, but never progressed past mouse studies due to poor pharmacokinetics. Its redesign, ATH-1017 (fosgonimeton), reached human Alzheimer's Phase 3 trials with clean Phase 1 safety data, but the program was discontinued after Phase 3 failed to slow cognitive decline. A second redesign, ATH-1105, is in earlier development with positive Phase 1 results, now targeting ALS instead.⁷

Dihexa is scheduled for review at the FDA's next PCAC meeting, expected before the end of February 2027, alongside several other peptides under consideration for the 503A list.

P21

A small peptide derived from ciliary neurotrophic factor (CNTF), designed for oral dosing and blood-brain barrier permeability, with a mechanism broadly overlapping Cerebrolysin's multi-pathway approach. Published evidence is currently limited to mouse models — no human clinical data yet.⁸

FGL

A 15-amino-acid peptide from the University of Copenhagen, isolating the FGFR1-activating region of NCAM (neural cell adhesion molecule). In animal studies, FGL given after learning tasks produced durable memory improvement, alongside anti-inflammatory effects — calming microglial activation in aged rats and promoting remyelination after stroke in other models.⁹ Human Phase 1 safety data was published in 2007 and looked clean, but no Phase 2 or 3 efficacy trials have followed since.

What this means practically

• Delivery method matters as much as the molecule — which is why so many of these are dosed intranasally, and why technique affects outcomes
• Animal and cell-model evidence isn't the same as human clinical evidence — P21 and FGL in particular have compelling mechanisms but little to no human trial data
• Regulatory status is moving, but a favorable advisory vote isn't an approval, and none of these six are FDA-approved drugs for the uses discussed here
• "Neurotrophic" and "neuroprotective" cover genuinely different mechanisms — BDNF upregulation, anti-inflammatory signaling, synaptic growth factors — worth distinguishing rather than treating as interchangeable