Dihexa occupies an unusual place in the research-peptide literature. It was designed as a metabolically stabilised, orally available analogue of angiotensin IV and was widely described, for more than a decade, as a hepatocyte growth factor (HGF)/c-Met mimetic several orders of magnitude more potent than brain-derived neurotrophic factor in synaptogenesis assays. That mechanistic story does not survive the published record. In April 2025 the Journal of Pharmacology and Experimental Therapeutics retracted the two papers that established the HGF/c-Met framing. What remains is a small, mixed preclinical column, a 2013 characterisation that still carries an editorial Expression of Concern, and a complete absence of human data.

This article grades that remainder. It is not a protocol, not a recommendation, and not a claim that Dihexa treats, prevents, or reverses any human condition.

1. What Dihexa is

Dihexa is N-hexanoic-Tyr-Ile-(6) aminohexanoic amide (also designated PNB-0408). Chemically it is a capped dipeptide: a tyrosine–isoleucine core with an N-terminal hexanoyl group and a C-terminal 6-aminohexanoic amide. It was developed at Washington State University from the norleucine¹-angiotensin IV (Nle¹-AngIV) lineage, itself a metabolically more stable analogue of the endogenous hexapeptide angiotensin IV (Val-Tyr-Ile-His-Pro-Phe).[1][7]

The design intent, as stated in the originating characterisation, was to retain the putative procognitive core of Nle¹-AngIV — reported to reside in the three N-terminal residues — while improving metabolic stability and barrier permeability relative to the parent peptides.[1] That is a chemical-optimisation claim. It is not evidence of a human effect.

Dihexa is not a licensed medicine in the United Kingdom. It has no MHRA marketing authorisation, no Summary of Product Characteristics, and no Yellow Card safety file as a medicinal product. Under the Human Medicines Regulations 2012 it is an unlicensed substance whose legal status turns on how it is presented and for what purpose it is supplied.[13][14]

2. How the claim was built

Angiotensin IV and related AT₄-lineage peptides have a longer experimental history than Dihexa itself. A 2018 systematic review by Ho and Nation identified 32 experimental (non-human) studies of angiotensin IV, angiotensin-(1-7) and their receptors; in models of cognitive deficit, eight of nine included studies reported that Ang IV or named analogues (Nle¹-AngIV, Dihexa, LVV-hemorphin-7) improved performance on spatial working-memory or passive-avoidance tasks.[10] Almost all of that literature used intracerebroventricular or other near-learning administration. The review is useful as a map of the class. It is not a validation of oral Dihexa, and it predates the 2025 retractions.

Two papers from the originating laboratory sit underneath the Dihexa-specific claim.

Benoist et al., 2011 reported that C-terminal truncated Nle¹-AngIV analogues as small as tri- and tetrapeptides reversed scopolamine-induced Morris water maze deficits in rats and increased hippocampal dendritic-spine number and head size.[7] That paper has not been retracted. It did not name or test Dihexa. It is precursor evidence for the chemical lineage, not direct evidence for the marketed compound.

McCoy et al., 2013 is the characterisation paper. The authors described sequential N- and C-terminal modifications that improved stability while retaining activity against scopolamine-induced maze deficits and hippocampal synaptogenesis, culminating in N-hexanoic-Tyr-Ile-(6) aminohexanoic amide. They reported oral activity, blood–brain barrier permeability, reversal of scopolamine-induced deficits, and improved learning in aged (24-month) rats.[1] In September 2021 the same journal issued an Expression of Concern for this article.[2] The paper has not been formally retracted. An unretracted paper under an Expression of Concern is not a clean primary source. It remains the main published description of Dihexa’s in-life rodent pharmacology from the originating group.

A subsequent 2015 review by Wright and Harding in the Journal of Alzheimer’s Disease synthesised the HGF/c-Met system as an Alzheimer’s target and placed Dihexa inside that frame.[11] That review now sits on mechanism papers that have since been withdrawn. It should be read as a contemporaneous position piece, not as independent confirmation.

3. The HGF/c-Met story, and its retraction

The claim that Dihexa acts by potentiating HGF at c-Met was not a side comment. It was the proposed molecular mechanism.

Kawas et al., 2012 (J Pharmacol Exp Ther 340:539–548; PMID 22129598) framed angiotensin IV analogues as hepatocyte growth factor/Met modifiers.[3] It was retracted in April 2025.[4]

Benoist et al., 2014 (J Pharmacol Exp Ther 351:390–402; PMID 25187433) argued that the procognitive and synaptogenic effects of angiotensin IV-derived peptides, including Dihexa, depend on activation of the HGF/c-Met system.[5] It was retracted in April 2025.[6] The retraction notice records that, following a Washington State University investigation, specified figures and data in a subsequent erratum submission were found to contain falsified and/or fabricated elements, and that Leen H. Kawas and Joseph W. Harding were found to be solely responsible.[6]

The mechanism should therefore be read as proposed — in papers later retracted — not as established fact. Citing either paper without the retraction is a citation error.

The circulating claim that Dihexa is “seven orders of magnitude more potent than BDNF” is an assay-level comparison from this programme, not a clinical potency, and it is not a finding that can be treated as settled once the mechanism papers have been withdrawn. Even on its own terms it compared synaptogenic activity in cell-culture assays, not cognitive outcomes in any species, and not any human endpoint.

A related 2011 JPET paper from the same group on HGF dimerization-domain mimics (not Dihexa-specific) was retracted on the same date.[4] The pattern is a laboratory-level integrity failure in the HGF/c-Met biochemical file, not a single isolated figure.

4. What remains after the retractions

After removing the withdrawn papers, the Dihexa-specific column is short.

Independent positive preclinical study. Sun et al. (2021), from China Pharmaceutical University and Nanjing Medical University, administered Dihexa to APP/PS1 transgenic mice — a standard amyloid-model of Alzheimer’s disease — and reported restored Morris water maze performance, increased neuronal-cell and synaptophysin (SYP) signals, reduced astrocyte and microglia activation, lower IL-1β and TNF-α, higher IL-10, and dependence on PI3K/AKT signalling (the PI3K inhibitor wortmannin reversed the anti-inflammatory and anti-apoptotic readouts).[8] This is the cleanest unretracted support that Dihexa, as a named compound, does something in a mammalian cognitive model. It is still a mouse study. It does not restore the retracted HGF/c-Met biochemistry; the authors attribute the effect to PI3K/AKT. It does not generate a human evidence grade.

Independent negative preclinical study. Wells et al. (2024) tested PNB-0408 (Dihexa) against chronic 3-nitropropionic acid in male Wistar rats, a mitochondrial-toxin model used to mimic Huntington’s-like motor and cognitive deficits. PNB-0408 did not protect against the weight, motor, or cognitive deficits induced by 3-NP.[9] A negative result in a different model does not falsify Sun 2021. It does bound any claim of general neuroprotection. The Dihexa literature is not a one-direction column.

The McCoy 2013 characterisation remains published, under the 2021 Expression of Concern noted above.[1][2] It is single-laboratory, and the journal has already told readers to treat it with caution.

No human file. As of this writing there is no published peer-reviewed human trial of Dihexa, no registered interventional study on the public clinical-trial registers that we can identify, and no published human pharmacokinetic or safety dataset. Absence of a trial is not a finding of harm. It is a finding of no human evidence.

5. The c-Met caveat, stated carefully

Even if the withdrawn papers had stood, the proposed target is not a casual wellness pathway. MET, which encodes the HGF receptor c-Met, is a proto-oncogene. Dysregulated HGF/MET signalling is implicated in tumour growth, invasion and metastasis, and MET is a validated oncology drug target.[12] That is a reason any development programme aiming at HGF/c-Met potentiation would need a serious oncology risk assessment before first-in-human exposure. It is not a demonstrated human harm from Dihexa, because Dihexa has not been studied in humans. The honest position is narrower: the proposed mechanism, now withdrawn, pointed at a pathway that oncology actively inhibits, and there is no human safety file against which to bound that theoretical risk.

6. UK legal status

The Human Medicines Regulations 2012, regulation 2, define a medicinal product as any substance presented as having properties for preventing or treating disease in human beings, or any substance that may be used by or administered to human beings with a view to restoring, correcting or modifying a physiological function by pharmacological, immunological or metabolic action.[13] The MHRA applies that definition by presentation and by function.[14]

Dihexa has no UK marketing authorisation. It is not a prescription-only medicine in the POM sense, because it is not an authorised medicine; supplying or advertising it as a medicine is the regulated act. A research-use-only supply, honestly labelled and not presented for human administration, sits outside the medicines definition. Presenting the same substance as a cognitive enhancer, an Alzheimer’s candidate for personal use, or a nootropic for human consumption would engage the presentation limb, regardless of a “research” sticker on the vial.[13][14]

Dihexa is not listed as a controlled drug under the Misuse of Drugs Act 1971. Controlled-drug status and medicines-law status are different questions. Unlicensed is not a synonym for illegal, and it is not a synonym for safe to supply as a medicine.

This page describes published research. It does not authorise, recommend, or describe human administration.

7. Evidence grade

Peptide Data grades on a four-point scale: Strong, Moderate, Limited, Anecdotal.

Human cognitive or neurodegenerative claims for Dihexa: anecdotal. There is no human trial. There is no human safety database. Vendor and forum reports are not an evidence grade.

Unretracted preclinical remainder: limited, and mixed. One independent positive mouse study (Sun 2021) and one independent negative rat study (Wells 2024) sit beside a single-lab 2013 characterisation under an Expression of Concern and a precursor 2011 Nle¹-AngIV paper that did not test Dihexa. The HGF/c-Met mechanism papers are withdrawn and do not count.

This is not in the same evidence class as licensed GLP-1 or dual GIP/GLP-1 receptor agonists, which rest on multi-thousand-participant Phase 3 programmes and MHRA authorisations. It is closer to other research peptides whose human column is empty and whose preclinical column is either thin, conflicted, or both.

8. What this page is not

This is not a finding that Dihexa “does nothing”. Sun 2021 is a real paper, and Ho and Nation’s class-level review is a real map of angiotensin-IV experimental pharmacology. It is a finding that the mechanism most often attached to the compound has been formally withdrawn, that the remaining Dihexa-specific animal data are limited and not unidirectional, and that the step from those data to any human use has not been taken in the published record.

No consumption, self-administration, or dosing guidance is offered. Experimental doses used in rats and mice are properties of those protocols; they are not instructions. Research use only is the absolute line.

This article is AI-researched and editorially reviewed. It is provided for research and educational purposes only and is not medical advice. Research peptides are not licensed for human consumption in the UK.