Dihexa research still has 8 open questions, and most of them trace to one event. In April 2025, the Journal of Pharmacology and Experimental Therapeutics retracted 3 papers from the Washington State University group that first characterized this material, after an institutional investigation found falsified or fabricated figure data. The retractions did not erase every finding. They removed the published basis for one specific claim: that Dihexa acts through the hepatocyte growth factor and c-Met receptor system. What is left is a molecule with a replicated behavioral signal in 1 independent rodent model, a published null result in another, no human testing, and no long-term toxicology on record.
Disclaimer. Dihexa is supplied strictly as a research chemical. It is not for human consumption, not for therapeutic use, and not for diagnostic use. Nothing below is guidance on administration to people. This article covers the published scientific record, the open mechanistic questions in it, and what a research plan should control for.
The 8 questions, in short:
-
Does Dihexa act through hepatocyte growth factor and c-Met at all?
-
Does it activate that pathway or block it, and at which concentration does the direction flip?
-
Where did the claim that it outperforms BDNF by 7 orders of magnitude actually come from?
-
Will the behavioral signal replicate outside 1 laboratory and 1 model?
-
Does it do anything measurable in a system that is not already impaired?
-
Does modulating this pathway translate at all, given what happened to the related clinical program?
-
What happens under chronic exposure, given that c-Met is an oncogenic pathway?
-
What are the pharmacokinetics in any species other than the rat?
Research Use Only, And Why That Framing Matters Here
Every material listed by Nordic Chems is sold for laboratory research and development only, to be handled by trained and qualified personnel. That is not a boilerplate line in this case. It is the condition that makes the rest of this article coherent.
A material with no human safety file, no published long-term toxicology, and a contested mechanism is a legitimate object of laboratory study and nothing else. The statements here have not been evaluated by the US Food and Drug Administration. Nothing on this page is intended to diagnose, treat, cure, or prevent any disease. The terms of service govern every order, and they require that purchasers confirm research intent and professional handling capability.
What Is Dihexa In Chemical Terms?
Dihexa (PNB-0408) is a synthetic angiotensin IV analog with the chemical name N-hexanoic-Tyr-Ile-(6) aminohexanoic amide. Its CAS number is 1401708-83-5, its formula is C27H44N4O5, and its molecular weight is 504.66. It was developed at Washington State University and is studied in neuropharmacology research on growth factor signaling and synaptic structure.
It also carries the identifiers ATH-1001 and UNII 9WYX65A5C2 in the NIH substance registry. The parent molecule, angiotensin IV, is degraded quickly and does not cross the blood-brain barrier. The N-terminal and C-terminal modifications that produced Dihexa were designed to raise hydrophobicity and metabolic stability, which took molecular weight from roughly 774 down to 505.
That chemistry is settled. Almost nothing downstream of it is.
The 3 Retractions That Reset The Record
In April 2025 the Journal of Pharmacology and Experimental Therapeutics retracted 3 papers from the originating laboratory. The retraction notice for the 2014 paper, Benoist et al., states that following an investigation by Washington State University, Figures 1B and 2A/C and data in a later erratum submission were found to contain falsified or fabricated data, and names Leen H. Kawas and Joseph W. Harding as solely responsible.
Each of the 3 carried a different load:
-
Kawas et al. 2011, J Pharmacol Exp Ther 339:509-518. Characterized HGF dimerization-domain mimics as anti-Met and anticancer agents.
-
Kawas et al. 2012, J Pharmacol Exp Ther 340:539-548. Developed the angiotensin IV analogs as HGF and Met modifiers.
-
Benoist et al. 2014, J Pharmacol Exp Ther 351:390-402. Supplied the claim that the behavioral and synaptogenic effects depend on HGF and c-Met activation.
A fourth paper, McCoy et al. 2013 in the same journal, carries a Notice of Concern rather than a retraction. That matters more than it looks. McCoy 2013 is the source of the rodent behavioral data, the spine-density counts, the half-life figures, and the blood-brain barrier evidence that most secondary writing on this material still cites. It has not been withdrawn, but it is flagged.
The damage has a specific pattern: the papers that explained the mechanism are gone, and the paper that characterized the molecule is under a caution notice.
The Mechanism Itself Is Unproven
Dihexa's proposed mechanism is no longer supported by a live publication. The retraction of Benoist 2014 removed the direct evidence that the observed effects depend on hepatocyte growth factor and c-Met, because that paper carried the antagonist-blockade and receptor-knockdown experiments that made the causal argument.
What survives is weaker and indirect. Other groups have used Dihexa in HGF-responsive systems and observed activity, but those studies assumed the HGF binding mechanism rather than testing it independently. Assuming a mechanism and confirming one are different experiments.
The study that would close this question is not complicated. A direct binding assay plus c-Met phosphorylation readout plus selective c-Met knockdown, run by a group with no connection to the original laboratory, published with raw figure data. That experiment has not appeared.
Activator Or Blocker: The Direction Problem
Reported activity reverses with concentration, and no one has resolved which direction matters.
At picomolar and low nanomolar concentrations, Dihexa has been described as potentiating HGF-dependent c-Met activation. At micromolar concentrations, it has been described as inhibiting HGF dimerization, which would reduce c-Met signaling. The same molecular family was patented and published under both readings: one of the now-retracted 2011 papers characterized these dimerization-domain mimics as anti-Met and anticancer agents.
That is not a small inconsistency. It means the sign of the effect depends on exposure level, and the exposure level at which the behavior switches has not been mapped. Any in vitro design using this material has to treat concentration as a primary variable rather than a secondary one, because the same reagent may do opposite things at 10 picomolar and at 1 micromolar.
Where Did The "7 Orders Of Magnitude" Number Come From?
This is the claim worth checking, because it is everywhere.
Nearly every page written about this material repeats that Dihexa is 7 orders of magnitude more potent than brain-derived neurotrophic factor. The Alzheimer's Drug Discovery Foundation reviewed the primary literature and reported something different: the comparison appeared in a press release from the authors of the 2013 publication, and the effects of BDNF and Dihexa were not directly compared in that publication.
A 10-million-fold potency ratio is a dramatic number. It is also the kind of number that should be traceable to a head-to-head assay with both agents run under the same conditions. As far as the published record goes, it is not.
Anyone citing that figure in a grant application, a protocol rationale, or a literature review should pull the source first. The defensible version of the claim is narrower: in cell-culture assays of new neuronal connections, the authors characterized activity at very low concentrations. That is a real observation. The ratio is a comparison that was asserted rather than measured.
Has Any Independent Lab Replicated The Effect?
Once, partially, and through a different pathway.
Sun et al., published in Brain Sciences in 2021, is the one independent replication on record. A group at Hebei Medical University tested Dihexa in APP/PS1 transgenic mice at 1.44 and 2.88 mg/kg amounts administered, from 6 to 9 months of age. They reported improved Morris water maze performance, increased synaptophysin expression, more surviving neurons on Nissl staining, and reduced inflammatory markers. They attributed the effect to PI3K and AKT signaling, and the PI3K inhibitor wortmannin at 0.5 mg/kg reversed it.
Read that carefully. It is corroboration of the behavioral signal from outside the original laboratory. It is not corroboration of the HGF and c-Met mechanism. The replicating group identified a different pathway.
There is also a reporting wrinkle in that paper worth flagging: the methods describe both intragastric and intraperitoneal administration, and the results section references both injection and intragastric procedures. Until that is clarified, the study cannot be cleanly cited as either an oral or a parenteral result.
One independent replication is a data point, not a body of evidence.
What Does Dihexa Do In A System That Is Not Impaired?
Very little, based on what has been published, and this is the question most secondary writing skips.
The reported behavioral effects come from impaired models: scopolamine-induced deficits, aged animals, transgenic amyloid models, lesion models. In animals with normal cognition, the Alzheimer's Drug Discovery Foundation review notes that Dihexa did not improve performance. The reading offered at the time was that the HGF and c-Met system is not heavily engaged during normal learning and responds instead to injury.
If that holds, it reshapes the research question. The material would be a conditional modulator of a repair-response pathway rather than a general enhancer of synaptic function. Those are different biological claims, they predict different experiments, and only one of them is consistent with what has actually been measured.
It also means a study design using healthy animals as the test population, rather than as a control arm, is probably measuring nothing.
Then there is the result nobody cites. Wells et al. 2024, in the Journal of Huntington's Disease, ran 40 male Wistar rats in 3 groups across 5 weeks, testing PNB-0408 against 3-nitropropionic acid neurotoxicity. The authors, working at Whitworth University with collaborators at Oregon Health and Science University, reported that PNB-0408 did not protect the animals from the induced deficits. A published null result from an independent group belongs in the evidence base just as much as a positive one.
What Does The Fosgonimeton Result Mean For The Pathway In 2026?
It is the closest thing to a clinical test the underlying hypothesis has received, and the readout was negative on the endpoints that mattered.
First, the correction that most writing on this topic gets wrong: fosgonimeton is not Dihexa. It is a separate molecule, developed by Athira Pharma, formerly M3 Biotechnology, a Washington State University spinout that originated from the same laboratory and the same HGF and MET premise. Dihexa itself never entered clinical development.
In September 2024, Athira reported that its Phase 2/3 LIFT-AD trial, which randomized 312 participants to once-daily subcutaneous fosgonimeton 40 mg or placebo over 26 weeks, missed its primary endpoint, a Global Statistical Test combining cognition and function scores. It also missed both key secondary endpoints. A biomarker signal was reported alongside the miss.
What that does and does not tell you:
-
It does not show that Dihexa would behave the same way. Different molecule, different exposure, different population.
-
It does suggest that modulating this pathway did not translate to clinical endpoints in the one adequately powered human study the hypothesis has had.
For anyone designing preclinical work in this space in 2026, that is a reason to focus on mechanism and target engagement rather than on behavioral endpoints alone. The behavioral endpoint has already been tested at the clinical level for a related agent and did not hold.
Long-Term Safety And The c-Met Cancer Signal
No long-term toxicology has been published in any species. That is the plainest statement in this entire article, and it has not changed since the Alzheimer's Drug Discovery Foundation reviewed the material in August 2021.
The specific unresolved concern is theoretical but not speculative. The HGF and c-Met axis is a well-characterized oncogenic pathway. Aberrant activation of it, through receptor mutation, overexpression, or amplification, is documented across multiple human tumor types and correlates with progression and poorer outcomes. A molecule proposed to potentiate that axis therefore carries an obvious question: what happens under chronic exposure?
No carcinogenicity study, no chronic tumorigenesis study, no repeat-exposure toxicology. The review states this directly and names tumorigenesis and cancer progression as the untested concern.
Worth noting: the clinical program for the related molecule excluded participants with a history of or newly diagnosed malignant tumor. The people designing human studies on this pathway treated the oncology question as live.
One caveat on that source. The report predates the retractions by almost 4 years, and its mechanism section cites papers that have since been withdrawn. Its safety and trial-status conclusions do not depend on those papers and still stand. Its mechanistic description should be read with the retraction in mind.
Pharmacokinetics Outside The Rat
Nearly all pharmacokinetic data on this material comes from one species, and the half-life figures are unusual enough to deserve independent confirmation.
In rats, reported half-life was 12.68 days after intravenous administration and 8.83 days after intraperitoneal administration, with a serum half-life of 335.5 minutes measured in rat serum. Those are the numbers circulating in the literature. They come from McCoy 2013, the paper carrying the Notice of Concern.
A multi-day half-life changes experimental design in ways that are easy to get wrong. Washout intervals, accumulation across repeated exposure, and the interval between final administration and tissue collection all depend on it. If the figure is off, a study that assumes clean washout may be measuring residual exposure.
There is also the vehicle problem. The original rodent work used DMSO at high concentrations as the carrier, reported at 75% DMSO for intravenous and intraperitoneal routes. DMSO has biological activity of its own. Vehicle-only control arms are not optional in this work.
The oral bioavailability figure that circulates on secondary sites, usually given as roughly 38%, does not trace cleanly to a primary source in the material reviewed here. Treat it as unverified.
Established, Contested, And Unknown: Where Each Claim Sits
Every claim about this material sorts into 1 of 6 tiers, and knowing which tier a claim sits in decides whether it belongs in a protocol rationale.
Established. The chemical identity. CAS 1401708-83-5, formula C27H44N4O5, molecular weight 504.66, all confirmed in the NIH substance registry.
Supported but qualified. A behavioral signal in impaired rodent models, resting on 1 independent replication with the earlier data under a Notice of Concern. PI3K and AKT involvement, from that same single study in a single model. Blood-brain barrier penetration in rodents and the multi-day rat half-life, both resting on the flagged 2013 paper.
Contested. HGF and c-Met as the operative mechanism. The primary evidence was retracted in April 2025 and nothing independent has replaced it.
Unsupported as stated. The potency ratio against BDNF, which traces to a press release rather than a head-to-head experiment.
Negative. No improvement reported in animals with normal cognition. No protection in the 3-nitropropionic acid rat model.
Absent. Human pharmacology of any kind. Long-term toxicology and carcinogenicity in any species.
That is where this material stands in October 2026.
4 Checks Worth Building Into A Dihexa Protocol
A contested evidence base does not make a material unusable in the laboratory. It changes what a protocol has to control for. 4 checks:
1. Cite around the retractions, not through them. Benoist 2014 and the 2011 and 2012 Kawas papers should not appear as supporting citations in a rationale section. If a secondary review you are drawing on cites them, that review inherits the problem. Check the publication status of every reference before it goes into a protocol.
2. Treat concentration as a primary variable. Given the reported reversal between picomolar and micromolar activity, a single-concentration design cannot distinguish potentiation from inhibition. Run a range.
3. Build vehicle controls that match the original work. If you are comparing against published rodent results, the DMSO concentrations in those studies were high enough to require matched vehicle arms.
4. Verify the material before you verify the hypothesis. A null result means nothing if the contents of the vial were never confirmed. This is the one variable you control, and it starts with where you are sourcing Dihexa for laboratory work.
What The Supplier Record Can And Cannot Settle
A contested literature is not something a supplier can repair. The identity of the material in the vial is, and that is worth separating from everything else in this article.
Nordic Chems states that every batch is independently tested by accredited laboratories for identity, purity, and concentration, so that the label matches the contents. It works with ISO-certified and cGMP-compliant laboratories, and every product carries labelling with the chemical name, concentration, and structure. Testing is carried out by those third-party laboratories rather than in-house, a point the terms of service set out directly: the purchaser remains responsible for verifying a material against their own standard methods before it goes into a protocol.
That division is worth stating plainly. A supplier that names the standard its testing partners work to is making a stronger claim than one that verifies its own material. The published detail behind that claim sits on the quality control page and the reasons this testing standard exists.
That same standard governs Dihexa in research quantities and the rest of the research chemical catalog, including materials whose literature is far more settled than this one.
Dihexa research will stay unsettled until an independent group runs the mechanism experiment and someone publishes a chronic toxicology study. Neither has happened. Until they do, the discipline available to any research plan is in the controls, the citations, and the certificate that says what is actually in the vial. Anyone ordering research material online should ask for that record before the work starts, not after.
Disclaimer. All materials are for laboratory research and development use only. Not for human consumption. Not for veterinary use. Not for diagnostic or therapeutic application. These statements have not been evaluated by the US Food and Drug Administration and are not intended to diagnose, treat, cure, or prevent any disease. Handling requires appropriate PPE and standard laboratory safety practice, in compliance with all applicable regulations.
FAQs
Was the main Dihexa research paper retracted?
Yes. In April 2025 the Journal of Pharmacology and Experimental Therapeutics retracted 3 papers from the Washington State University group, including Benoist et al. 2014, which supplied the evidence that the effects depend on the hepatocyte growth factor and c-Met system. The retraction notice states that an institutional investigation found falsified or fabricated data in specific figures. A fourth paper, McCoy et al. 2013, carries a Notice of Concern rather than a retraction.
Has Dihexa been tested in humans?
No. There is no completed human study of Dihexa in the published record, and the Alzheimer's Drug Discovery Foundation reported finding no registered trials and no NIH-funded programs testing it. All available data comes from rodent models and cell culture. Dihexa is supplied for laboratory research only and is not for human consumption.
Is Dihexa the same material as fosgonimeton?
No. They are separate molecules. Fosgonimeton, also called ATH-1017, was developed by Athira Pharma, a Washington State University spinout that originated from the same laboratory and the same HGF and MET premise. Fosgonimeton reached Phase 2/3 testing. Dihexa itself never entered clinical development.
What does Dihexa research still need to prove about the HGF and c-Met mechanism?
It needs an independent group to run the mechanism experiment that the retracted paper claimed to have run: direct binding, c-Met phosphorylation, and selective receptor knockdown, published with raw figure data. The only independent replication to date, Sun et al. 2021 in APP/PS1 mice, attributed the effect to PI3K and AKT signaling rather than confirming HGF and c-Met.
Has any independent group replicated Dihexa's effects?
Partially, and the record includes a negative result. Sun et al. 2021, at Hebei Medical University, reported improved maze performance and increased synaptophysin in APP/PS1 mice at 1.44 and 2.88 mg/kg amounts administered. Wells et al. 2024, an independent group publishing in the Journal of Huntington's Disease, tested 40 Wistar rats over 5 weeks and reported no protection against 3-nitropropionic acid neurotoxicity.
What is the main safety question in Dihexa research?
No long-term toxicology has been published in any species. The specific unresolved concern is that the HGF and c-Met axis is a well-documented oncogenic pathway, and aberrant activation of it is associated with tumor progression across multiple human cancer types. No carcinogenicity or chronic exposure study on this material exists.
Is Dihexa approved for human use?
No. Dihexa is not approved by any regulatory authority anywhere, for any indication. It is supplied strictly as a research chemical for laboratory use by trained and qualified personnel, and it is not for human consumption, veterinary use, diagnostic use, or therapeutic use.
