Dihexa: What It Is, Claimed Cognitive Effects, Risks, Why It's Research-Only

Key takeaways

• Dihexa is a chemically capped fragment of angiotensin IV, designed at Washington State University to survive digestion and reach the brain. It is sold online as a nootropic "peptide" in capsules, creams and vials.
• Every result behind the cognitive claims comes from rats, mice and lab dishes: reversal of drug-induced memory loss, faster maze learning in old rats, better memory in an Alzheimer's mouse model. No human has taken dihexa in a registered trial.
• The research record is damaged. The paper that tied dihexa to its proposed mechanism, hepatocyte growth factor (HGF) and its receptor c-Met, was retracted in April 2025, and the paper that introduced dihexa carries an expression of concern. The company that licensed it dropped it in 2015.
• There is no human dose, and the risks are unknown rather than known to be small: no published toxicology, a plasma half-life of roughly 9 to 13 days in rats, and a target pathway that drives growth and spread in several cancers.
• No US pharmacy may compound dihexa. It came off FDA's Category 2 list on April 15, 2026 only because its nomination was withdrawn; it is not on the permitted list, and an FDA advisory committee review is due by February 2027.

Dihexa at a glance

Also called

PNB-0408, N-hexanoic-Tyr-Ile-(6) aminohexanoic amide, angiotensin IV analog

Class

Angiotensin IV analog: a capped tyrosine-isoleucine dipeptide (a peptidomimetic) proposed to amplify hepatocyte growth factor (HGF)/c-Met signaling; developed at Washington State University

Route

Oral gavage, intraperitoneal injection and infusion into the brain in rats and mice; sold as capsules, creams and injectable vials with no human pharmacokinetic data

FDA status

Not FDA-approved and never tested in a registered human trial. Removed from FDA's Category 2 list on April 15, 2026 after its nomination was withdrawn; not on the 503A bulks list, so it cannot be legally compounded; advisory committee review due by February 2027.

FDA Category 2 (compounding restricted)

No

Available by prescription (US)

No: there is no approved product and no US pharmacy may compound it

Best-studied use

Rat and mouse models of memory impairment (scopolamine, old age, an Alzheimer's transgenic mouse, repeated mild brain injury) and neuron cultures; no human studies

Typical monthly cost (telehealth, 2026)

No legal product exists to price; research capsules and vials are unapproved drugs and we do not list their prices

What dihexa is

Dihexa's chemical name describes it: N-hexanoic-Tyr-Ile-(6) aminohexanoic amide. At its core are two amino acids, tyrosine and isoleucine, joined by a peptide bond. One end is capped with a six-carbon fatty acid, the other with a six-carbon amino acid amide. That makes it less a peptide than a peptidomimetic, a small molecule assembled from peptide parts, although FDA's compounding lists and every online seller group it with the peptides. It also appears in the literature under the code name PNB-0408 (Wells 2024).

It comes from a long research program in the laboratories of Joseph Harding and Jay Wright at Washington State University. Angiotensin IV, a six-amino-acid fragment of the blood-pressure hormone angiotensin, improved learning in rodents when it was infused into the brain, but it falls apart in minutes and cannot cross the gut or the blood-brain barrier. In the paper that introduced dihexa, the group reported that its parent compound, the angiotensin IV analog Nle1-angiotensin IV, lasted 1.4 minutes in rat serum, while dihexa lasted about 335 minutes (McCoy 2013). In live rats the plasma half-life was 12.7 days after an intravenous dose and 8.8 days after an intraperitoneal one; blood was sampled only for 5 days, so those figures are extrapolations. Radiolabeled dihexa infused into a rat's carotid artery was concentrated in the brain within 30 minutes.

The paper disclosed that Wright and Harding were founders and shareholders of M3 Biotechnology, a company developing drugs based on the technology (McCoy 2013). That company is now Athira Pharma (Retraction Watch 2021). In a 2021 filing with the SEC, Athira said it held an exclusive license to dihexa, tried to develop it from 2011 to 2015, and stopped because its studies did not show the properties a successful drug needs. It then designed fosgonimeton (ATH-1017), a different molecule that the filing describes as a prodrug converted in the blood into dihexa (Athira 2021). The dihexa sold online today is not Athira's product; it is synthesized by research-chemical suppliers.

How it is supposed to work

The original theory was that angiotensin IV acts on its own receptor in the brain. The Washington State group later proposed something different: that dihexa binds hepatocyte growth factor (HGF), a growth factor that despite its name acts throughout the body, and helps it activate its receptor, the tyrosine kinase c-Met (also written MET). In the brain, HGF and MET support neuron survival and the formation of synapses (Martino 2025). The 2014 paper that made this case reported that dihexa bound HGF with high affinity, switched on c-Met when only a trace of HGF was present, increased dendritic spines in rat neurons, and lost its memory effect in rats when an HGF blocker was infused into the brain (Benoist 2014).

That paper was retracted in April 2025 (Benoist 2014). The mechanism is still plausible, and one independent rat study published in 2025 found that MET-blocking drugs cancelled dihexa's effect on working memory (Martino 2025), which fits the idea without showing that dihexa binds HGF. As of October 2026 the HGF/c-Met explanation rests on a retracted paper, the developers' reviews, and a small amount of independent support.

A research record with retractions

This is the part of the dihexa story most sellers leave out.

• September 2021. The Journal of Pharmacology and Experimental Therapeutics placed expressions of concern on four papers from the dihexa laboratory, published between 2011 and 2014, over possible image manipulation (Retraction Watch 2021). One of them is McCoy 2013, the paper that introduced dihexa.
• October 2021. Athira's board reported that a special committee had found that its chief executive, Leen Kawas, a co-author of the papers from her graduate work, had altered images in her 2011 doctoral dissertation and in at least four research papers published from 2011 to 2014. She resigned on October 18, 2021. The committee also found that Washington State's dihexa patent incorporates images from papers she co-authored (Athira 2021).
• April 2025. The journal retracted three of the four papers: Benoist 2014 (the HGF/c-Met mechanism), Kawas 2012 (the development of angiotensin IV analogs as HGF/Met modifiers) and Kawas 2011, a paper on related compounds designed to block MET (JPET 2025).
• October 2026. McCoy 2013, which holds nearly all of dihexa's rat data, is not retracted but still carries its expression of concern on PubMed (McCoy 2013).

A retraction does not prove a compound inert, and independent laboratories have since reported effects in mice and rats. It does mean that the mechanism paper no longer counts as evidence, and that the most-cited dihexa data come with a warning attached by the journal that published them.

What the evidence actually shows

Dihexa claims, graded by the evidence behind them

Claim	Grade	What the evidence shows	Sources
Reverses drug-induced memory loss in rats	D	In rats given scopolamine into the brain, dihexa restored water-maze learning when infused into the brain (0.1 or 1 nmol), injected into the abdomen (0.25 and 0.5 mg/kg; 0.05 mg/kg did nothing) or given by stomach tube (1.25 and 2 mg/kg). 2 mg/kg by mouth fully reversed the deficit. 8 to 10 rats per group, 8 days of testing. The paper carries a 2021 expression of concern.	McCoy 2013
Improves learning in old rats	D	24-month-old rats, 3 male and 3 female per group, given 2 mg/kg daily by stomach tube learned the water maze faster than untreated old rats on most test days. The untreated group was highly variable.	McCoy 2013
Improves memory in an Alzheimer's mouse model	D	An independent group treated male APP/PS1 mice from 6 to 9 months of age with 1.44 or 2.88 mg/kg; water-maze performance, neuron counts and a synapse protein improved and inflammatory markers fell (12 mice per group for behavior). The paper gives the route as both intragastric and intraperitoneal.	Sun 2021
Rescues working memory after repeated mild brain injury	D	In a study of 52 male and female rats, those given three mild closed-head injuries in a week regained T-maze working memory when dihexa was infused into the brain, best at 1 nmol; MET-blocking drugs abolished the effect. Independent laboratory.	Martino 2025
Protects against Huntington's-like brain damage	D	Negative. In 40 male rats exposed to a mitochondrial toxin that mimics Huntington's disease, dihexa did not prevent the weight, motor or memory deficits over five weeks.	Wells 2024
Works by amplifying HGF/c-Met signaling	D	The paper reporting that dihexa binds HGF and that its memory effect depends on c-Met was retracted in April 2025. An independent 2025 rat study found MET blockers cancelled dihexa's effect, consistent with the idea but not proof of binding.	Benoist 2014; Martino 2025
Builds new synapses, 10 million times more potently than BDNF	D	In rat neuron cultures, dihexa at 1 picomolar increased dendritic spines about threefold. The BDNF comparison comes from the university's 2012 news release, not from a head-to-head study in animals or people.	McCoy 2013; WSU 2012
Improves memory, focus or intelligence in healthy people	D	No human study of dihexa exists. ClinicalTrials.gov lists no trial under its name, its code name or its chemical name.	ClinicalTrials.gov 2026
Treats Alzheimer's disease in people	D	No trial of dihexa. Fosgonimeton, a separate injectable drug that its developer says the body converts into dihexa, did not beat placebo on any primary or secondary endpoint in a 26-week trial of 287 people with mild-to-moderate Alzheimer's.	Porsteinsson 2025; Athira 2021

Three things stand out.

Every grade is D. The record is a handful of rodent experiments, a few cell-culture studies and no human data. The experiments used roughly 6 to 13 animals per group.

The routes do not match the products. Several of the positive results came from infusing dihexa directly into the brain's fluid spaces, a route no consumer uses. The oral results are rat results with a stomach tube; the injected ones went into the abdominal cavity, not under the skin.

Independent replication is thin and mixed. Outside the original laboratory, one group reported benefits in Alzheimer's mice, one in brain-injured rats, and one found no protection in a Huntington's model. That is a reason to study dihexa further, not a reason to take it.

What you are promised versus what is documented

Seller pages and forum posts make a consistent set of claims. Here is how they line up with the record.

• "Regrows brain connections." Documented as more dendritic spines in rat neuron cultures and inferred from rat behavior. Synapse counts in a living human brain have never been measured on dihexa.
• "10 million times more potent than BDNF." This comes from the 2012 news release about the McCoy paper, which said dihexa was "seven orders of magnitude" more powerful than brain-derived neurotrophic factor at forming new synapses (WSU 2012). We did not find that comparison in the published paper. Potency in a dish says nothing about benefit in a brain.
• "Works orally." True in rats at 2 mg/kg (McCoy 2013). In people, oral absorption has never been measured; the paper's figure for absorption from the human gut is a software prediction.
• "A nootropic for focus and productivity." We found no study, in animals or people, designed to test dihexa on attention, focus or performance in healthy subjects. The animal models were of induced or age-related impairment.
• "An Alzheimer's or Parkinson's treatment." Not supported in humans. The company that licensed dihexa abandoned it, and the trial of its successor molecule failed (Porsteinsson 2025).
• "Safe because it comes from a natural peptide." Dihexa is an engineered molecule designed to persist in the body. No toxicology study has been published.

Dihexa versus the realistic alternatives

Sources: (Hua 2022), (Porsteinsson 2025), (van Dyck 2023).

The table is not a ranking of three similar products. Lecanemab is for people with diagnosed early Alzheimer's disease, its benefit is modest, and its risks need monitoring; it is here because it shows what a tested memory drug looks like. For someone with no memory disorder who wants an edge, none of these is an answer, and dihexa is the one with the least known about it.

The other peptides sold for the brain do not change that picture. Semax and Selank, registered as nasal drops in Russia, have small Russian studies in stroke and anxiety but no controlled trial for focus or memory in healthy people, and neither can be prescribed or compounded in the US in October 2026; our Semax and Selank profile and Semax versus Selank comparison cover them. Our ranking of peptides for cognitive function puts dihexa alongside every candidate, including the large trials of intranasal insulin and semaglutide that found no benefit for memory.

Dosage: what was studied versus what sellers claim

Scaled by body surface area, the standard first step in estimating a human-equivalent dose, the 2 mg/kg oral rat dose corresponds to about 23 mg a day for a 70 kg adult; scaled by weight alone it would be 140 mg. That is arithmetic, not a dose. Drug developers use it only to choose a cautious starting point for a supervised first-in-human study, and the rat half-life of 9 to 13 days does not tell anyone how long dihexa lasts in a person. Our dihexa dosage page works through every study, the conversions and the seller charts in detail.

We do not publish a reconstitution example for dihexa. The reconstitution calculator exists for medications a pharmacy has dispensed with a labeled concentration; a research vial of dihexa has neither.

Side effects and who should not use it

What is known in humans about dihexa itself: nothing. There is no trial, no case series and no published toxicology. FDA says it has not identified any human exposure data for dihexa by any route of administration and lacks the information to know whether it would harm people (FDA Category 2). The absence of reported side effects is an absence of data.

What is known about the related drug. In fosgonimeton's phase 1 study, 88 people received daily injections for up to 9 days; adverse events were similar on drug and placebo, mostly mild injection-site pain and itching, with one case each of moderate neutropenia and an allergic skin reaction (Hua 2022). In the 26-week LIFT-AD trial, serious adverse events were 4.2% on fosgonimeton and 6.9% on placebo, but 14.2% stopped because of adverse events, mostly injection-site reactions, against 4.6% on placebo (Porsteinsson 2025). Those are results for a pharmaceutical-grade injection of a different molecule. They do not describe a capsule or vial of uncertain content.

The cancer question. What is established: HGF and MET play a causal role in the survival, growth, blood-vessel formation and spread of cancer cells, which is why drugs that block MET have been developed (Gherardi 2012). What is extrapolation: that a compound designed to amplify that signal could help an existing tumor grow. We found no study testing dihexa in any tumor model, and there is no evidence that dihexa causes cancer. The honest position is that nobody has looked, and that the pathway is one where looking matters. The dihexa safety page sets out what is established and what is not.

Product quality. Research capsules, creams and vials are not tested for identity, purity, dose or sterility by anyone accountable, and dihexa has no pharmacopeial standard.

Do not use dihexa if you:

• have cancer, a history of cancer or a high inherited cancer risk;
• are pregnant, trying to conceive or breastfeeding (no reproductive data exist);
• are a child or teenager;
• are looking for treatment for memory loss, dementia, ADHD or another diagnosed condition, which has tested options a clinician can offer.

For everyone else, the practical answer in October 2026 is the same: there is no legal, labeled product to use.

Legal status in the United States

Consistent with our legal tracker:

• Not approved. Dihexa has never been approved by FDA or, as far as we can find, by any other regulator, and it has never been tested in a registered human trial (ClinicalTrials.gov 2026).
• Off Category 2, but not permitted. Dihexa was one of the twelve peptides FDA announced on April 15, 2026 it was removing from Category 2, its list of bulk substances that may pose significant safety risks, because their nominations had been withdrawn. FDA's page now lists dihexa acetate among the substances "nominated but withdrawn", with the note that it found no human exposure data by any route (FDA Category 2). Removal ended the safety flag; it did not make dihexa eligible for compounding.
• Not compoundable. A 503A pharmacy may compound a substance only if it has a USP monograph, is a component of an FDA-approved drug or is on FDA's bulks list. Dihexa passes none of the three tests, so no licensed pharmacy may make it, with or without a prescription.
• Review pending. FDA's Pharmacy Compounding Advisory Committee is due to review dihexa, with LL-37, injectable GHK-Cu, PEG-MGF and melanotan II, by February 2027. A favorable vote would only be advice; FDA would still have to complete rulemaking, which it has not done for the six peptides the committee recommended in July 2026. Our LL-37 profile follows the same review.
• Research labels are not a legal route. Possessing dihexa is not a crime. Selling it for human use is selling an unapproved drug, whatever the label says; in 2026 FDA's warning letters to research-peptide sellers cited dosing guides, calculators and bundled bacteriostatic water as evidence of intended human use. Dihexa was not among the products named in the letters our tracker lists, but the reasoning applies to it.

Cost and how to get it prescribed

There is no legal way to get dihexa prescribed, and so no legitimate price to report. A telehealth clinician cannot make a non-compoundable substance compoundable, as our guide to what telehealth can prescribe explains. In our October 2026 clinic directory, some providers listed dihexa among their products; we flag it, because such a product either is not what it claims to be or is not coming from a pharmacy filling a lawful prescription.

If you are comparing what legal peptide therapy costs, our cost guide covers the products a licensed pharmacy can dispense, and the peptide therapy guide walks through the evaluation, labs and monitoring a real program includes.

If memory is the reason you are looking

Older adults are the people most often searching for dihexa, usually because they or a parent are noticing memory changes. That deserves a clinician's evaluation, not a research chemical. Memory problems have many causes, some of them treatable, and early Alzheimer's disease now has approved treatments with measured, if modest, effects: in its 18-month phase 3 trial, lecanemab slowed decline by 0.45 points on an 18-point dementia rating scale, with infusion reactions in about a quarter of patients and brain swelling or effusions on MRI in 12.6% (van Dyck 2023). Whether that trade-off is worth it is a conversation to have with a neurologist or memory clinic. Our cognitive-function ranking describes what such an evaluation looks for.

There are no dihexa data for women specifically, for menopause, or for pregnancy. For attention problems or ADHD, effective medications have decades of trials behind them; the dihexa safety page points to that evidence.

Bottom line

Dihexa is a real piece of neuroscience: an attempt to turn a fragment of angiotensin into a stable, brain-penetrant drug that amplifies a growth-factor system involved in building synapses. In rats and mice it improved memory in several models. It has also never been tested in a human being, its mechanism paper has been retracted, the paper that introduced it carries an expression of concern, the company that licensed it walked away in 2015, and the successor molecule that the body converts into dihexa failed its Alzheimer's trial.

What remains is a compound with a long half-life in rats, an unstudied relationship to a pathway that drives several cancers, no human dose and no legal source in the United States. A reasonable position is to watch the February 2027 committee review. An unreasonable one is to swallow a research capsule because a rat learned a maze faster.

Frequently asked questions

+−What is dihexa?

Dihexa is a small synthetic molecule built from a fragment of angiotensin IV, a breakdown product of the blood-pressure hormone angiotensin. Researchers at Washington State University capped two of its amino acids, tyrosine and isoleucine, with fatty chains so that it would resist digestion and reach the brain. It was developed as a possible Alzheimer's drug, was dropped by the company that licensed it, and is now sold online as a nootropic without any human trial behind it.

+−What does dihexa do?

In rats and mice it improved maze learning: it reversed memory loss caused by the drug scopolamine, helped 24-month-old rats learn faster, improved memory in an Alzheimer's mouse model and restored working memory after repeated mild brain injury. In rat neuron cultures it increased the number of dendritic spines, the small structures where synapses form. None of these effects has been measured in a human being.

+−Is dihexa a peptide?

Only partly. It contains two amino acids joined by a peptide bond, but both ends are chemically capped, one with a six-carbon fatty acid and the other with a six-carbon amino acid amide, which is what lets it survive in the blood. Chemists would call it a peptidomimetic. FDA's compounding lists and online sellers group it with the peptides.

+−Is dihexa legal?

It is not FDA-approved, and as of October 2026 no licensed US pharmacy may compound it, because it is not on FDA's list of permitted bulk substances. It came off FDA's Category 2 safety-risk list on April 15, 2026 only because its nomination was withdrawn, and an FDA advisory committee review is due by February 2027. Possessing it is not a crime, but selling it for human use, including with a research-use-only label, is selling an unapproved drug.

+−Has dihexa been tested in humans?

No. ClinicalTrials.gov lists no study of dihexa under its name, its code name PNB-0408 or its chemical name (searched October 10, 2026), we found no published human study, and FDA says it has not identified any human exposure data for dihexa by any route. The only human trials in the family are of fosgonimeton, a different injectable drug that its developer says the body converts into dihexa; its main Alzheimer's trial missed every primary and secondary goal.

+−Is dihexa the same as fosgonimeton?

No. Fosgonimeton (ATH-1017) is a separate drug, dihexa with a phosphate group attached, developed by Athira Pharma for injection under the skin. Athira's 2021 filing with the SEC says it is converted in the blood into dihexa. Its trial results describe a pharmaceutical-grade injection at specific doses, not the capsules, creams or vials sold online, so they cannot be used to justify those products.

+−Why were the dihexa studies retracted?

In September 2021 the Journal of Pharmacology and Experimental Therapeutics placed expressions of concern on four papers from the dihexa laboratory, published from 2011 to 2014, over possible image manipulation. Athira's board found that its then chief executive, a co-author, had altered images in her 2011 dissertation and in at least four papers. In April 2025 the journal retracted three of the four, including the paper that linked dihexa to HGF and c-Met. The paper that introduced dihexa still carries its expression of concern.

+−Does dihexa cause cancer?

There is no evidence that it does, and also no study that looked. What is established is that the pathway dihexa is said to amplify, HGF and its receptor MET, drives the survival, growth and spread of several cancers, and FDA has approved two MET-blocking drugs for a form of lung cancer. Whether dihexa feeds existing tumors is unknown; that uncertainty is the reason anyone with cancer or a high cancer risk should avoid it.

+−Does dihexa work orally?

In rats, yes: 2 mg per kilogram by stomach tube fully reversed scopolamine-induced memory loss, and 1.25 mg per kilogram partly did. In people, nobody has measured how much of a swallowed dose is absorbed or reaches the brain. The oral-absorption figure in the original paper is a computer prediction, not a measurement.

Sources

1. StudyMcCoy AT et al. Evaluation of metabolically stabilized angiotensin IV analogs as procognitive/antidementia agents. J Pharmacol Exp Ther 2013;344:141-54. Expression of concern published 2021 (J Pharmacol Exp Ther 378:313). (2013)
2. StudyBenoist CC et al. The procognitive and synaptogenic effects of angiotensin IV-derived peptides are dependent on activation of the hepatocyte growth factor/c-met system. J Pharmacol Exp Ther 2014;351:390-402. RETRACTED April 2025 (J Pharmacol Exp Ther 392:103567). (2014)
3. SourceJ Pharmacol Exp Ther 2025;392:103567. Retraction notice to 'The procognitive and synaptogenic effects of angiotensin IV-derived peptides are dependent on activation of the hepatocyte growth factor/c-Met system' (Benoist 2014). Companion notices of April 2025 retract Kawas 2012 (PMID 40312092) and Kawas 2011 (PMID 40312094). (2025)
4. StudySun X et al. AngIV-analog dihexa rescues cognitive impairment and recovers memory in the APP/PS1 mouse via the PI3K/AKT signaling pathway. Brain Sci 2021;11:1487. (2021)
5. StudyMartino KA et al. Hepatocyte growth factor/MET activator rescues working memory deficits after repeated mild traumatic brain injury. Neurotrauma Rep 2025;6:991-8. (2025)
6. StudyWells RG et al. Effects of an angiotensin IV analog on 3-nitropropionic acid-induced Huntington's disease-like symptoms in rats. J Huntingtons Dis 2024;13:55-66. (2024)
7. NewsWashington State University. Prospective Alzheimer's drug builds new brain cell connections, improves cognitive function of rats. News release, October 11, 2012 (reprinted by ScienceDaily). (2012)
8. NewsRetraction Watch. Four papers by Athira CEO earn expressions of concern. September 24, 2021. (2021)
9. SourceAthira Pharma, Inc. Form 8-K filed October 21, 2021: special committee findings, dihexa development 2011 to 2015 and its discontinuation, and ATH-1017 (fosgonimeton) as a prodrug of dihexa. (2021)
10. StudyHua X et al. Safety, tolerability, pharmacokinetics, and pharmacodynamics of the positive modulator of HGF/MET, fosgonimeton, in healthy volunteers and subjects with Alzheimer's disease: randomized, placebo-controlled, double-blind, phase I clinical trial. J Alzheimers Dis 2022;86:1399-413. (2022)
11. StudyPorsteinsson AP et al. Fosgonimeton in mild-to-moderate Alzheimer's disease (LIFT-AD). J Alzheimers Dis Rep 2025;9:25424823251405817. (2025)
12. ReviewGherardi E, Birchmeier W, Birchmeier C, Vande Woude G. Targeting MET in cancer: rationale and progress. Nat Rev Cancer 2012;12:89-103. (2012)
13. Studyvan Dyck CH et al. Lecanemab in early Alzheimer's disease. N Engl J Med 2023;388:9-21. (2023)
14. SourceClinicalTrials.gov. Searches for 'dihexa', 'PNB-0408' and 'N-hexanoic-Tyr-Ile': no registered studies (searched October 10, 2026). (2026)
15. RegulatoryFDA. Certain bulk drug substances for use in compounding that may present significant safety risks: Category 2 list and bulk drug substances nominated but withdrawn (dihexa acetate). Content current as of April 22, 2026. (2026)