What is Dihexa? What it does, benefits and what the evidence shows

Dihexa is a synthetic, lab-made molecule derived from angiotensin IV, a fragment of a blood-pressure hormone, and it is sold as a memory and cognition booster. It has only ever been studied in cells and animals, mostly rodents. We found no human trial, no approval, and no registered clinical study. There is a further problem buyers should know about: two of the papers that established how dihexa is supposed to work were retracted in April 2025, after journal notices of concern in 2021.

What Dihexa is and how it is supposed to work

Dihexa has the full chemical name N-hexanoic-Tyr-Ile-(6) aminohexanoic amide. It was engineered from a shorter angiotensin IV analogue, norleucine-1-angiotensin IV, whose active core the developers traced to three amino acids. They then modified that core so it would resist breakdown and cross into the brain, and the result was dihexa. It was reported to reverse memory deficits in a rat model of scopolamine-induced impairment and in aged rats, and to promote the growth of new synapses in cells (McCoy et al., Journal of Pharmacology and Experimental Therapeutics, 2013). A later paper gave it another development name, PNB-0408 (Wells et al., Journal of Huntington's Disease, 2024).

The proposed mechanism is hepatocyte growth factor (HGF) signalling. HGF is a natural growth factor that acts through a receptor called c-Met, and the claim was that dihexa binds HGF and boosts its signalling, which in turn drives the formation of new synapses (Benoist et al., 2014). That claim is where the trouble is.

The retractions

The 2014 paper that reported dihexa binding HGF and acting through c-Met, and the 2012 paper from the same group that developed the analogue family as HGF/Met modifiers, both carry retraction notices published in April 2025 (retraction of Benoist et al., 2014; retraction of Kawas et al., 2012). Both had been the subject of notices of concern in September 2021 (notice on Benoist et al.; notice on Kawas et al.). The journal also issued a notice of concern for the 2013 paper that first described dihexa; we found notices of concern for it but no retraction notice, so it stands as published with that flag on it (Europe PMC record of the notice).

The retraction notice for the 2014 paper states that, following an investigation by Washington State University, Figures 1B and 2A/C and the data in a later erratum submission were found to contain falsified and/or fabricated data, and that Leen H. Kawas and Joseph W. Harding were found to be solely responsible (retraction notice, Journal of Pharmacology and Experimental Therapeutics, 2025). We could not open the notice for the 2012 paper, so we can't state its reasons. The HGF/c-Met explanation has lost its two main supporting papers, and the original 2013 description carries an expression of concern. Everything below that mentions HGF or c-Met as dihexa's mechanism should be read with that in mind.

Approval status

We found no approval for dihexa. We searched DailyMed (the U.S. label database) for dihexa and found no label, and a ClinicalTrials.gov search for dihexa and its chemical name returned no registered study. We did not check the European Medicines Agency or other regulators' databases.

FDA's page on bulk drug substances that may present significant safety risks for compounding lists dihexa acetate in a table of substances that were previously in "category 2" and were withdrawn by their nominators. The entry says FDA "has not identified any human exposure data on drug products containing dihexa acetate administered via any route of administration" and that it lacks important information about safety, including whether it would cause harm if given to humans (FDA, page current as of April 22, 2026). In plain terms, there is no human safety record for regulators to look at.

What people use it for

What the research actually shows

Memory in animals

The original 2013 paper reported that dihexa reversed memory deficits in the rats described above (McCoy et al., 2013), but, as noted, that paper carries an expression of concern. A review of the brain HGF/c-Met system that Wright and Harding, two of the developers, co-wrote describes dihexa as showing promise in animal models of Alzheimer's disease (Wright and Harding, Journal of Alzheimer's Disease, 2015). It is a review by the compound's own developers, so it is not independent confirmation.

A more independent data point is a 2021 study in APP/PS1 mice, a common Alzheimer's model. The authors reported that dihexa restored performance in a spatial water-maze test, reduced markers of inflammation in the brain, and activated a signalling pathway called PI3K/AKT (Sun et al., Brain Sciences, 2021). That study attributes the effect to PI3K/AKT rather than testing HGF, which is worth noticing given the retractions. It is a single mouse study.

A 2018 systematic review of 32 animal studies of angiotensin IV and related peptides found that in models of cognitive deficit, 8 of 9 studies of angiotensin IV and its analogues, dihexa among them, found improved performance on some memory tasks (Ho and Nation, Neuroscience & Biobehavioral Reviews, 2018). That is a pattern of mostly positive rodent reports, not evidence about people, and the review was written before the retractions.

A negative result

In a 2024 study, 40 rats given a mitochondrial toxin that mimics Huntington's disease were divided into three groups, one of which also received PNB-0408 (dihexa). The compound did not protect the rats from the weight loss, learning and memory problems or movement problems the toxin caused, and the authors concluded it may not be an effective strategy in that model (Wells et al., Journal of Huntington's Disease, 2024). One study does not settle anything, but it is a reminder that the compound has not worked everywhere it was tried.

Nerve repair and hearing cells

In rats with a cut and repaired sciatic nerve, motor function improved at 8 to 16 weeks when dihexa was combined with mesenchymal stem cells and placed in the muscle; the paper reports the same for stem cells combined with G-CSF, and the abstract does not describe dihexa alone doing anything (Weiss et al., Annals of Medicine and Surgery, 2021). In zebrafish larvae, dihexa at one concentration protected inner-ear-type hair cells from two antibiotics known to damage them (Uribe et al., Frontiers in Cellular Neuroscience, 2015). Two authors of that study, Kawas and Harding, also authored the retracted papers, and it leans on the HGF explanation.

A related drug idea that reached humans, and failed

Dihexa itself has never been given to people in a trial we could find. A different molecule, fosgonimeton, was designed as a positive modulator of the same HGF system. It was studied in a randomized placebo-controlled phase 2/3 trial of 287 people with mild-to-moderate Alzheimer's disease in the primary analysis. It did not meet its primary or secondary endpoints, and more people on the drug stopped treatment because of side effects, mostly injection-site reactions (Porsteinsson et al., Journal of Alzheimer's Disease Reports, 2025). Secondary sources describe a close relationship between fosgonimeton and dihexa, but we could not confirm it from a primary source, so this is context for the HGF idea, not a result for dihexa.

Everything else

We found no controlled human study of dihexa for memory, focus, learning, mood or aging in healthy people. Those uses rest on rodent findings and anecdote.

The theoretical cancer concern

HGF and its receptor c-Met are not just about brain synapses. Reviews describe the HGF/c-Met pathway as involved in tumor cell growth, invasion, migration and blood-vessel formation, and several drugs that block c-Met are already on the market for cancer (Zhu et al., Signal Transduction and Targeted Therapy, 2026). Another review says the pathway is critical in development and tissue repair, that it promotes malignant tumor progression, and that either too much or too little activity is linked to disease (Sun et al., Journal of Cell Communication and Signaling, 2026).

The worry that follows is that a compound designed to switch that pathway on could, in principle, encourage tumors. We found no study that tested dihexa for that, in either direction, and no case report. So this is a question the literature has not answered, not a demonstrated harm. It is also complicated by the retractions: if dihexa does not work through HGF/c-Met as first claimed, the concern loses its stated basis, but nobody has shown what it does work through.

Side effects and risks

There is no human safety data for dihexa. What the record shows:

• No human exposure data. FDA's own statement is that it has identified none, by any route.
• No safety study of any kind that we found, including no long-term animal toxicology paper among the sources we opened.
• A shaky evidence base. The key mechanism papers are retracted and the discovery paper carries an expression of concern.
• The cancer question is open, as above.

What the lab tests show

As of September 29, 2026, VendorVerify has no lab results we can count for dihexa. That doesn't mean vials are good or bad; it means nobody has submitted a report we can use. If you have had a dihexa vial tested, submitting the report is the fastest way to change that.

Because we don't have enough lab data on this compound to rank vendors for it, here are the highest-ranked vendors across all compounds right now. A vendor's overall record is a signal, not a guarantee for a compound they have never had tested.

Buying smarter

With no compound-specific data, lean on the general checks. Look at each vendor's full record and how many reports sit behind the score, and treat a vendor-posted COA as a starting point (how to read a peptide COA, fake peptide COAs). For a compound with no track record in our data, an independent identity, purity and quantity test on your own vial is worth paying for; how to get a peptide tested walks through it, and endotoxin in research peptides covers when to add that test.

If a report comes back with a fail, stop using that batch, tell anyone who bought from the same run, bring the report to the vendor, and submit it to VendorVerify. People researching dihexa often also read about Semax and Selank, two other cognition-marketed peptides. Our peptide hub lists every compound we track, and what are grey market peptides explains the market itself.

Database figures are from VendorVerify as of September 29, 2026. This isn't medical advice; dihexa is sold for research use only, and we found no approval for it as a drug.