Dihexa (PNB-0408)

Dihexa, also catalogued as PNB-0408, is a small synthetic molecule whose full chemical name is N-hexanoic-Tyr-Ile-(6)-aminohexanoic amide. Its molecular formula is C₂₇H₄₄N₄O₅, its molecular weight is 504.66 g/mol, and it is indexed under CAS number 1401708-83-5 and PubChem CID 129010512. Despite being marketed in some circles as a peptide, it contains only two canonical amino acid residues, tyrosine and isoleucine, capped at the N-terminus with hexanoic acid and at the C-terminus with a 6-aminohexanoic amide group. It is therefore more accurately described as a peptidomimetic, a small drug-like molecule built on a peptide scaffold.

Dihexa emerged from a research program at Washington State University led by Joseph W. Harding and John W. Wright, which studied the procognitive properties of angiotensin IV (AngIV), a six amino acid fragment of the renin-angiotensin system with the sequence Val-Tyr-Ile-His-Pro-Phe. AngIV and its metabolically stabilized relative Nle1-AngIV had been reported to improve learning in rodent memory tasks, but they were poor drug candidates: rapidly degraded, poorly absorbed and essentially unable to reach the brain after peripheral administration.

The development path was one of stepwise simplification. Work published in 2011 reported that the procognitive information in Nle1-AngIV was concentrated in its three N-terminal residues, and that tripeptide and tetrapeptide fragments could still reverse memory deficits in rats. A follow-up paper in 2013 described the addition of the hexanoyl and aminohexanoic amide caps, producing the compound named dihexa for its two six-carbon extensions. The stated design goals were metabolic stability, oral availability and passage across the blood-brain barrier, three properties the parent peptides lacked.

What dihexa never acquired is a clinical file. It was patented, discussed in reviews as a possible Alzheimer's disease candidate, and then stalled at the preclinical stage. Everything a reader encounters today about dihexa's effects comes from rodent experiments, cell culture work, patent documents or uncontrolled personal reports, not from controlled human research.

This content is for educational purposes only. Dihexa is a research compound, not an approved medicine. Consult a healthcare professional for any question about cognition, memory or neurological health.

The mechanism most often attributed to dihexa involves the hepatocyte growth factor (HGF) / c-Met system. HGF is a pleiotropic growth factor; its receptor c-Met (encoded by the MET gene) is a receptor tyrosine kinase expressed in many tissues, including neurons. In the brain, HGF/c-Met signaling has been associated with neurite outgrowth, dendritic arborization, synapse formation and neuroprotection, which is why a 2015 review by Wright and Harding in the Journal of Alzheimer's Disease argued that this system was an overlooked therapeutic target in Alzheimer's disease.

According to the original proposal, dihexa does not activate c-Met directly. It was reported to bind HGF with high affinity and to promote HGF dimerization, which is a prerequisite step for productive receptor engagement, thereby allowing c-Met phosphorylation to occur in the presence of HGF concentrations that would otherwise be subthreshold. In this model dihexa behaves as a positive modulator of an endogenous growth factor rather than as a growth factor substitute. The proposed downstream consequence was spinogenesis and synaptogenesis: new dendritic spines, enlarged spine heads and additional functional synaptic connections in hippocampal neurons, which was offered as the explanation for improved performance in rodent memory tasks.

This is not the only mechanism ever proposed for the angiotensin IV family. An older and still active line of research treats AngIV analogs primarily as inhibitors of insulin-regulated aminopeptidase (IRAP, historically called the AT4 receptor). A 2020 review in Frontiers in Pharmacology by Hallberg and Larhed traces the medicinal chemistry effort to convert AngIV into small IRAP-inhibiting peptidomimetics, a program that has produced its own series of cognition-related candidates. The HGF/c-Met account and the IRAP account are distinct explanations for overlapping observations, and the choice between them was never settled by independent evidence for dihexa specifically.

A third strand comes from independent downstream work. A 2021 study in Brain Sciences using APP/PS1 transgenic mice reported that dihexa's behavioral and anti-inflammatory effects depended on the PI3K/AKT pathway, since the PI3K inhibitor wortmannin reversed them. PI3K/AKT lies downstream of many growth factor receptors, including c-Met, so this result is compatible with the HGF hypothesis without demonstrating it.

The critical caveat is that the papers that established the HGF/c-Met mechanism for this compound class have since been withdrawn from the literature, which is the subject of a dedicated section below. The mechanism should currently be read as a plausible hypothesis whose primary supporting evidence has been retracted, not as an established pharmacology.

The oldest surviving piece of the puzzle is the 2011 paper by Benoist and colleagues in the Journal of Pharmacology and Experimental Therapeutics. It reported that Nle1-AngIV derived peptides as short as tetrapeptides and tripeptides reversed scopolamine-induced deficits in the Morris water maze, a rodent spatial memory task, and that a peptide's procognitive activity correlated with its capacity to increase dendritic spine number and spine head size. This paper concerns the precursor fragments rather than dihexa itself, and it is a pharmacological reversal of a drug-induced deficit, not a model of a human neurodegenerative disease.

Dihexa was introduced in a 2013 paper from the same group, which described it as metabolically stable, sufficiently blood-brain barrier permeant, orally active and capable of inducing spinogenesis at picomolar concentrations in cultured neurons. The picomolar potency figure is the origin of the claims about extraordinary potency that still circulate. That paper carries a Notice of Concern published in 2021, so these specific figures should not be treated as reliable.

The most informative independent study is the 2021 report in Brain Sciences by Sun and colleagues in APP/PS1 mice, a transgenic model of amyloid pathology. Orally administered dihexa restored spatial learning and memory performance in the Morris water maze. The authors also reported reduced activation of astrocytes and microglia, lower levels of the pro-inflammatory cytokines IL-1β and TNF-α, higher levels of the anti-inflammatory cytokine IL-10, increased expression of the presynaptic marker synaptophysin, and loss of these effects when PI3K was inhibited. This is a positive, mechanistically coherent result from a group with no stake in the original patents.

Negative data also exist, and they matter just as much. A 2024 study in the Journal of Huntington's Disease by Wells and colleagues tested PNB-0408 in forty male Wistar rats randomized to vehicle, the mitochondrial toxin 3-nitropropionic acid, or 3-nitropropionic acid plus PNB-0408. The compound did not protect against the toxin-induced deficits, and the authors concluded that it did not appear to be an efficacious strategy in that model. Dihexa therefore has at least one published failure in an independent laboratory and a different disease model.

Taken together, the preclinical file is thin for a compound with this much consumer visibility: a handful of rodent studies, small group sizes, short durations, behavioral endpoints that are notoriously sensitive to experimental conditions, no published chronic toxicology, no published carcinogenicity assessment and no published reproductive safety work. For context on how to read this kind of evidence base, see our overview of peptide safety and the limits of preclinical data.

Any honest guide to dihexa has to address its publication record, because the integrity problem is unusually central. In September 2021, the Journal of Pharmacology and Experimental Therapeutics published Notices of Concern on four papers from the Washington State University group: Kawas et al. 2011 on mimics of the HGF dimerization domain, Kawas et al. 2012 on the development of AngIV analogs as HGF/Met modifiers, McCoy et al. 2013 on metabolically stabilized AngIV analogs (the paper that introduced dihexa), and Benoist et al. 2014 on the dependence of procognitive and synaptogenic effects on the HGF/c-Met system.

In April 2025, the journal went further and published formal retraction notices for three of those articles: the 2011 dimerization domain paper, the 2012 HGF/Met modifier paper and the 2014 paper that is the primary published basis for the HGF/c-Met mechanism. The notices state that, following an institutional investigation at Washington State University, specified figures and data were found to contain falsified or fabricated material. As of this writing the 2013 paper that first characterized dihexa has not been retracted; it remains flagged by its 2021 Notice of Concern.

It is important to be precise about what this does and does not mean. A retraction for image manipulation does not demonstrate that dihexa is biologically inert, and it says nothing about the independent 2021 APP/PS1 study or the independent 2024 Huntington's model study. What it does is remove the evidentiary foundation for the compound's most striking claims: the picomolar potency figures, the reported binding affinity for HGF, the dimerization model and the comparisons that produced the widely repeated assertion that dihexa is orders of magnitude more potent than BDNF. Those numbers now rest on papers the publishing journal has withdrawn.

There is a second-order consequence that is rarely mentioned. One of the retracted papers reported anti-Met and anticancer activity for molecules in this family. That paper is sometimes cited to reassure readers that AngIV analogs do not promote tumor growth. Since it has been retracted, it cannot be used as reassurance in either direction, and the oncological question remains genuinely open.

For readers, the practical rule is simple: when a vendor page, a forum post or an article quotes a dramatic potency statistic for dihexa, check whether the underlying citation is one of the withdrawn papers. In most cases it is.

No. As of October 2026 there is no published human clinical trial of dihexa, no published human pharmacokinetic study, no published human safety or tolerability data, and no established human dosing. Searches of the trial registries return no interventional record for dihexa or PNB-0408. The compound never completed the regulatory and toxicological work that precedes first-in-human administration, which is why no responsible source can state what an effective or safe human exposure would be.

The closest the HGF/c-Met hypothesis has come to a human test is a different molecule. Fosgonimeton (ATH-1017), developed by Athira Pharma, is a small-molecule prodrug that is converted in plasma to the active metabolite ATH-1001 and acts as a positive modulator of the HGF/MET system. Its preclinical characterization was published in Neurotherapeutics in 2023, reporting neurotrophic, anti-inflammatory and procognitive effects in cell and animal models of dementia. Fosgonimeton is not dihexa: it is a distinct chemical entity, given by subcutaneous injection, from a different sponsor and with its own data package.

Fosgonimeton was taken into LIFT-AD, a randomized, placebo-controlled Phase 2/3 trial (NCT04488419) in which 287 participants with mild-to-moderate Alzheimer's disease who were not taking acetylcholinesterase inhibitors received either daily subcutaneous fosgonimeton 40 mg or placebo. The published results report that the trial did not meet its primary or secondary endpoints: fosgonimeton did not significantly improve the ADAS-Cog11 cognitive scale or the ADCS-ADL23 function scale compared with placebo. The authors noted consistent non-significant numerical differences favoring the drug and argued that this may indicate biological activity worth further study, which is a reasonable scientific position and also a clear statement that efficacy was not demonstrated.

Two inferences follow. First, the broader idea of pharmacologically enhancing HGF/MET signaling to improve cognition has been tested in humans once, at scale, and it failed its prespecified endpoints. Second, that failure is not evidence against dihexa specifically, and a different outcome would not have been evidence for it, because the two compounds differ in structure, route and pharmacokinetics. The honest summary is that dihexa's human effects are unknown, and the one human readout available for its proposed pathway was negative.

Dihexa's visibility in the nootropic world is out of proportion to its evidence, and the reasons are mostly narrative rather than scientific. The first is the potency claim. A compound described in the literature as inducing synapse formation at picomolar concentrations, and compared favorably with BDNF, is a compelling story in a field where most options are modest stimulants or poorly characterized herbals. That the comparison traces back to papers now under Notice of Concern or retracted rarely accompanies the claim.

The second is the oral and brain-penetrant design. Most research peptides discussed for cognition require injection or intranasal administration, which limits their appeal. Dihexa was engineered to survive the gut and cross the blood-brain barrier, and that engineering goal is frequently repeated online as if it were a verified human property. It was demonstrated in rodents, which is not the same claim.

The third is availability. Dihexa is a small molecule that is relatively straightforward to synthesize, so it appears in grey-market catalogues labelled for laboratory use. Buyers have no practical way to verify identity, purity, residual solvents or the accuracy of a stated quantity, and certificates of analysis supplied by sellers are not independent. Dihexa is also lipophilic by design and poorly water soluble, which pushes suppliers and users toward organic solvent or oil vehicles whose safety for human exposure has never been evaluated for this compound.

The fourth is the absence of a null result that feels satisfying. Because no human trial exists, enthusiasts cannot be contradicted by trial data, and the discussion is dominated by self-reports. Uncontrolled self-reports in this space are a weak signal: expectation effects are large for subjective cognition, people typically combine several compounds at once, and reports of adverse outcomes are systematically underreported in communities invested in a substance. Our comparative overview of peptides studied for brain and cognitive function sets dihexa alongside better-documented options, and the companion guide on whether peptides are safe explains why research-grade material is a distinct risk category.

None of this is an argument that dihexa is uninteresting. HGF/c-Met is a legitimate neurobiological target, and the 2021 APP/PS1 data are worth replicating. It is an argument that enthusiasm has run far ahead of evidence, and that the gap is being filled by marketing rather than data.

The first and most important safety statement is an absence: there is no published human safety data for dihexa, and the peer-reviewed record contains no chronic toxicology study, no carcinogenicity bioassay, no reproductive or developmental toxicity assessment and no drug interaction work. Short rodent experiments designed to measure behavior are not safety studies. When a source states that dihexa is well tolerated, it is extrapolating from the absence of reported problems in small animal cohorts, which is not the same as evidence of safety.

The mechanistic concern that deserves careful, non-alarmist explanation is oncological. The c-Met receptor is one of the better-characterized oncogenic drivers in human cancer. A 2024 review in the International Journal of Molecular Sciences describes MET as a principal driver of invasive growth, a developmental program that tumors co-opt to sustain malignancy, and summarizes the mechanisms through which MET becomes aberrantly activated, including gene amplification, MET exon 14 skipping mutations and protein overexpression. An entire class of oncology drugs, including MET tyrosine kinase inhibitors and antibodies against MET or HGF, exists specifically to block this pathway in patients.

Dihexa was designed to do the opposite: to amplify HGF signaling at c-Met. That is a coherent reason for caution, particularly for repeated or long-term exposure, and particularly in anyone with a personal or family history of cancer or with undiagnosed pathology. It is equally important to state the limits of this reasoning. HGF/c-Met also serves normal regenerative functions in liver, kidney, lung and nervous tissue; positive modulation of an endogenous ligand is pharmacologically different from constitutive receptor activation by a mutation; and no study has ever tested whether dihexa promotes tumor growth. The concern is biologically grounded and unresolved, not demonstrated. The one paper that reported anticancer activity for molecules in this family has been retracted, so it offers no counterweight.

Beyond oncology, several unknowns follow from the same pharmacology. c-Met is widely expressed outside the brain, and effects on fibrotic processes, angiogenesis or epithelial proliferation have not been characterized for this compound. Nothing is known about interactions with prescription medicines, about behavior in people with hepatic or renal impairment, or about the consequences of the solvents used to dissolve a poorly water-soluble molecule.

Medical disclaimer: this section is educational and is not advice. Dihexa is not approved for human use in any jurisdiction we are aware of, and self-administration of research chemicals carries risks that have not been quantified. Discuss any cognitive symptom or concern with a qualified healthcare professional.

Dihexa is not an approved drug. It has no marketing authorization from the US Food and Drug Administration, the European Medicines Agency or, to our knowledge, any other major regulator, for any indication. It has no approved product label, no official prescribing information and no pharmacopeial monograph, which means there is no recognized standard defining what an acceptable grade of the material even is.

It is also not a dietary supplement ingredient. A synthetic peptidomimetic developed as a drug candidate does not qualify as a supplement in the United States or as a food supplement ingredient in the European Union, and regulators on both sides of the Atlantic have taken enforcement action against companies selling unapproved peptide products with health claims. Material that reaches consumers is typically labelled for research use only, a designation that signals the opposite of a quality guarantee for human use: it tells the buyer the product was never manufactured, tested or released under pharmaceutical standards.

Legality varies by jurisdiction and is frequently misunderstood. In many countries the sale of a research chemical for laboratory purposes occupies a grey zone, while importing it for human consumption, administering it to another person or advertising therapeutic effects can breach medicines legislation. Readers should assume that national rules differ substantially and that the absence of a specific prohibition is not permission.

For competitive athletes the situation is clearer and stricter. The World Anti-Doping Agency prohibits, at all times, pharmacological substances that are not currently approved for human therapeutic use by any governmental health authority under its non-approved substances category. A compound in dihexa's regulatory position falls squarely within that description, so athletes subject to anti-doping rules should treat it as prohibited regardless of whether it appears by name on any list.

Finally, a point of professional practice: because dihexa has no approved use and no established human dose, there is no legitimate clinical protocol to publish. This guide deliberately contains no dosage, no reconstitution instruction and no administration schedule. Sources that supply such numbers are inventing them or reproducing invented figures, and the absence of human pharmacokinetic data means the numbers cannot be sanity-checked even in principle.

Dihexa is usually encountered in lists alongside other compounds discussed for cognition, which obscures how different their evidence bases are. The useful comparison is not which molecule sounds most potent, but how much human data exists and how trustworthy the underlying literature is.

Read down the first column and dihexa is an outlier in a specific way. Semax and Selank have thin but real human literature and an approval history in one jurisdiction. Cerebrolysin has been through many controlled trials, with results that reviewers continue to argue about, which is what a normal contested evidence base looks like. Dihexa has rodent data, a retracted mechanistic foundation and zero human trials, yet it is often presented as the most powerful option of the group.

There is also a structural difference worth noting. Semax and Selank are short peptides administered intranasally or parenterally, and Cerebrolysin is a complex porcine brain-derived peptide preparation given parenterally under medical supervision in the countries where it is registered. Dihexa is a small lipophilic peptidomimetic designed for oral use, which makes casual self-experimentation easier and therefore the lack of safety data more consequential.

If the goal is to understand which cognition-related peptides have anything resembling a clinical file, our comparative article on peptides for brain and cognitive function covers each of these in more depth, including where the published evidence stops. Dihexa belongs in that conversation as a scientifically interesting but unvalidated candidate, not as a usable option.