Dihexa is a small synthetic molecule derived from angiotensin IV, a peptide investigated in brain signaling. Researchers changed the parent structure to make a more stable peptidomimetic: a molecule that retains selected peptide features while changing its chemistry. Rodent experiments have asked whether it helps animals perform memory tasks after injury or disease-like changes. The evidence includes favorable maze findings, an unfavorable Huntington-like experiment and serious publication problems in the early work. No reviewed human trial establishes cognitive benefit or clinical safety.
Overview
The chemical name N-hexanoic-Tyr-Ile-(6)-aminohexanoic amide identifies a defined parent compound, called PNB-0408 by Wells and colleagues (2024). Its structure was developed from an angiotensin IV analogue. FDA’s safety entry separately names Dihexa acetate and reports finding no human exposure data for that form by any route.
Molecular profile
- Tested parent form
- N-hexanoic-Tyr-Ile-(6)-aminohexanoic amide; a defined synthetic peptidomimetic Source
- Other research name
- PNB-0408 Source
- Chemical design
- Angiotensin-IV-derived molecule with modified ends; not the intact angiotensin IV peptide Source
- Acetate distinction
- Dihexa acetate is the substance named in FDA’s compounding safety entry; the entry supplies no human exposure data Source
Forms and derivatives
Dihexa / PNB-0408
These names identify the parent material in the rat Huntington-like experiment. The chemical name describes its hexanoyl end, Tyr-Ile segment and aminohexanamide end.
Dihexa acetate
FDA evaluates the acetate-named bulk substance separately. Human exposure and safety for that acetate form remain uncharacterized.
McCoy and colleagues’ 2013 report introduced the chemistry and rat exposure measurements and carries a 2021 Notice of Concern. The later 2014 HGF/c-Met mechanism paper was retracted in 2025. The concern limits confidence in the early pharmacology, and the withdrawn mechanism measurements cannot establish the proposed HGF/c-Met explanation.
Uses and research areas
The research targets learning and memory in specific experimental settings. Early work used rats whose maze learning was disrupted by scopolamine, a drug that blocks part of acetylcholine signaling, and older rats with learning difficulties. Later studies used APP/PS1 mice with inherited Alzheimer-related changes, rats exposed to a toxin that produces Huntington-like deficits, and rats with repeated mild head injuries. A swimming maze tests learning the location of a hidden escape platform; a delayed alternation maze tests remembering the previous choice long enough to choose the other arm. Neither task directly measures a person’s independence, work performance or progression of dementia.
The models also differ biologically. Genetic changes, temporary blockade of neurotransmission, disturbed cellular energy production and mechanical head injury do not create the same brain problem. The favorable results in APP/PS1 mice and head-injured rats therefore need to be read beside the unfavorable toxin experiment. The later experiments therefore provide both favorable and unfavorable findings in distinct forms of experimental brain injury.
Mechanism of action
Hepatocyte growth factor, or HGF, is a protein signal that activates the MET receptor on cells. MET signaling can influence cell survival, movement and growth. Dihexa has been investigated as a compound that strengthens this signaling. However, the widely cited 2014 claim that HGF/c-Met activation explains its learning and synapse findings came from a paper that has been withdrawn. The retraction notice says a Washington State University investigation found falsified or fabricated data in Figures 1B and 2A/C and in a subsequent erratum submission. Those data cannot establish Dihexa’s binding strength, receptor activation or mechanism.
Later work supplies narrower observations. In Sun and colleagues’ mouse experiment, Dihexa treatment accompanied changes in PI3K/AKT, an intracellular signaling route involved in cell survival. Adding wortmannin, a PI3K inhibitor, reversed several inflammatory and cell-death-related effects. This supports involvement of that signaling route in that experiment; it does not independently identify every molecular target of Dihexa.
The 2025 head-injury study also tested a MET antagonist and cultured kidney-cell colonies. Dihexa shifted E-cadherin, a cell-adhesion protein, from the cell membrane toward the cell interior, and the antagonist blocked the behavioral benefit in injured rats. The antagonist experiment and cell-adhesion assay support further investigation of HGF/MET signaling. They do not directly measure Dihexa’s binding to its proposed target.
Pharmacokinetics
The 2013 report with a Notice of Concern measured Dihexa in rat plasma after intravenous and intraperitoneal administration. Its calculated terminal half-life was 12.68 days after intravenous administration and 8.83 ± 2.41 days (mean ± standard error) after intraperitoneal administration, with four rats in the latter analysis. The intravenous profile rose again after an initial decline; the authors discussed recirculation through the gut and liver as a possible explanation. This is more complicated than one smooth exponential decline, and the formal concern further limits confidence in quoting the estimates as settled properties.
In a separate part of that report, investigators measured brain-associated radioactivity after labeled Dihexa and a vascular tracer were infused into rats. Radioactivity measured at one time point does not supply a human concentration curve, and distinguishing intact drug from labeled breakdown products matters when interpreting delivery. Oral maze activity also does not by itself quantify absorption into blood.
A human half-life has not been established. The long rat estimates come from a small exploratory study under a formal concern. The newer head-injury experiment delivered Dihexa directly into a fluid-filled cavity in the brain (ventricle), bypassing the usual question of movement from blood into brain. It cannot establish oral or systemic brain delivery in people. FDA’s Dihexa acetate assessment likewise supplies no human exposure measurements.
Research findings
Later mouse findings in an Alzheimer-related model
Sun et al. (2021) studied male APP/PS1 mice beginning at six months of age. After three months of treatment, Dihexa-treated mice found a hidden platform faster during training and crossed its former location more often during the later probe trial than APP/PS1 comparison mice given preparation without Dihexa. The maze analyses used 12 mice per group. Swimming speed did not show a group difference, which helps assess whether faster movement alone explained the result. Brain measurements also found more Nissl-stained neuronal cells and more synaptophysin, a protein associated with nerve terminals, alongside lower inflammatory measurements.
The paper describes administration both as intragastric and as injection. Because those routes differ, it is an uncertain source for a firm oral-delivery claim. Its mouse behavior and tissue findings are useful observations, but they do not show that people with Alzheimer’s disease retain memory or daily function with Dihexa.
An unfavorable result in toxin-exposed rats
Wells et al. (2024) randomized 40 male Wistar rats to a preparation-only control, the mitochondrial toxin 3-nitropropionic acid, or the toxin plus PNB-0408. Across five weeks, toxin-exposed rats gained less weight and developed movement and learning problems. Adding PNB-0408 did not protect the measured outcomes. The result is an unfavorable finding in this five-week toxin model; the size and precision of any remaining treatment effect are unclear.
Direct brain administration after repeated mild injury
In Martino et al. (2025), rats underwent three mild closed-head injuries over one week or sham procedures. Investigators tested whether the animals could alternate their T-maze choices after a delay. At the higher tested amount, Dihexa infused into a fluid-filled cavity in the brain (ventricle) increased correct choices to 81.5%, compared with 67.1% in injured rats given the buffered salt solution DPBS (P<0.0001). Testing took less time numerically, but that comparison was too uncertain to establish a shortening of the session (reported P=0.9777). A MET antagonist blocked the improvement. The authors combined male and female behavioral results. The Methods identify the control infusion as DPBS, while the figure labels it artificial cerebrospinal fluid; that inconsistency leaves the exact control composition uncertain.
The experiment followed behavior for days after injury using direct brain administration. Human concussion recovery and delivery from an oral or systemic preparation remain untested by this design.
Safety and tolerability
No human benefit or safety profile for parent Dihexa was established in the reviewed research. FDA reports that it found no human exposure data for Dihexa acetate and lacks information needed to know whether administration would cause harm. The animal studies are not designed to estimate uncommon human adverse events, long-term cancer risk, pregnancy risk or interactions with medicines. No reviewed source establishes an FDA-approved therapeutic use for Dihexa.
The proposed growth-factor mechanism makes unwanted effects on other tissues a subject for proper toxicology. It does not justify either a numerical human cancer-risk claim or a claim of safety. Likewise, a long estimated rat half-life cannot be used as a human administration schedule. Chemical identity, manufacturing quality and the biological behavior of any supplied material would each require evidence beyond a vendor label.
What the research means
Dihexa remains a preclinical topic with unresolved reliability and translation problems. Later studies report biological activity and improved behavior in selected models; the toxin study supplies a clear counterexample to universal protection. The 2013 pharmacology report’s formal concern and the 2014 mechanism paper’s retraction narrow what can responsibly be treated as established. Independent reproduction, dependable exposure measurements and suitable toxicology would be necessary before a human efficacy trial could be interpreted confidently.
For a reader comparing cognitive research compounds, the useful distinction is the experiment actually performed. A rat choosing a maze arm correctly, a protein moving within cultured cells and a person maintaining daily activities are different outcomes. The current literature does not connect those steps into a demonstrated human treatment effect.
FAQs
Has Dihexa improved memory in people?
The studies reviewed here tested cells or rodents. Improved animal maze performance cannot establish better everyday memory, prevention of dementia or treatment of Alzheimer’s disease in people. FDA also reports finding no human exposure data for Dihexa acetate by any route. Sun et al. (2021), Martino et al. (2025), FDA assessment
Why does the retraction matter?
The 2014 paper was the prominent experimental argument linking Dihexa to HGF/c-Met signaling. Its retraction means those results cannot establish the claimed mechanism. Later experiments remain assessable on their own methods; their existence does not repair the withdrawn data. Retraction notice (2025), Martino et al. (2025)
Is the original Dihexa report also retracted?
The 2013 McCoy report carries a 2021 Notice of Concern in the journal record. That is a different editorial action from the 2025 retraction of the 2014 paper. This profile describes the early measurements as reported observations under a formal concern. McCoy et al. (2013), Journal notice (2021), Retraction notice (2025)
Does Dihexa have a proven human half-life?
No human half-life was established in the reviewed sources. The much longer values quoted online come from small rat experiments in the 2013 paper with a Notice of Concern. They cannot determine a human dosing interval or duration of a cognitive effect. McCoy et al. (2013), Journal notice (2021), FDA assessment
Did every animal experiment find a benefit?
In the 40-rat experiment using 3-nitropropionic acid to injure brain energy metabolism, Dihexa did not protect the measured movement, learning and weight-gain outcomes over five weeks. This result shows why favorable findings in one model cannot stand for every neurological condition. Wells et al. (2024)
Does the growth-factor hypothesis prove a cancer risk?
No human Dihexa safety study in the reviewed material can quantify cancer risk. HGF/MET is a growth-signaling system, which makes sustained and off-target effects legitimate subjects for toxicology. The retracted mechanism paper cannot establish either a human cancer hazard or reassurance about it. Martino et al. (2025), Retraction notice (2025), FDA assessment
Comparisons and related reading
References
The original Dihexa chemistry, rat exposure and learning report carries a 2021 Notice of Concern. Its measurements are described with that warning and are not validated human pharmacology.
The journal’s formal concern applies to the 2013 report that supplies the original Dihexa chemistry, rat learning and exposure measurements.
The editor retracted the 2014 mechanism paper after an institutional investigation identified falsified or fabricated data in specified figures and a subsequent erratum submission. The withdrawn paper is not affirmative mechanism evidence here.
Genetically modified mice given Dihexa performed better in a swimming maze and had changes in inflammatory and cell-survival measurements. The paper describes administration inconsistently, limiting oral-delivery claims.
In a toxin-induced rat model, Dihexa did not prevent the measured movement, learning and weight-gain deficits. This supplies a consequential unfavorable result from a different disease model.
Rats with repeated mild head injuries made more correct delayed maze choices after direct brain-ventricle Dihexa than after buffered salt solution. The session-time comparison remained uncertain. Separate cell experiments examined movement of a cell-adhesion protein.
FDA reports finding no human exposure data for Dihexa acetate by any route and describes major gaps in its safety information.
Glossary
- Peptidomimetic
- A molecule designed to resemble selected features of a peptide while changing its chemical structure.
- HGF and MET
- Hepatocyte growth factor and its cell-surface receptor, a signaling pair studied in tissue responses and cell growth. c-Met is another name for MET.
- APP/PS1 mouse
- A genetically modified mouse model used to study selected Alzheimer-related changes; it does not reproduce the whole human disease.
- Working memory
- Holding information briefly to guide the next action, such as choosing the opposite arm of a maze after a delay.
- Terminal half-life
- The estimated time for the concentration in the late phase of a measured concentration curve to halve; it is not a direct measurement of clinical-effect duration.