what is the difference
Dihexa and fosgonimeton
The same proposed mechanism, characterised to two completely different depths, with every fact on both sides carrying a date and a document.
Dihexa is a small analog of angiotensin IV, a fragment of a hormone in the blood-pressure system, with fatty acid groups added at both ends. Those groups let it survive in blood and reach the brain, and they make it behave more like a small drug molecule than like a peptide. It also carries the code name PNB-0408, and it came out of a single laboratory at Washington State University.
Fosgonimeton, developed under the code ATH-1017, is a separate molecule aimed at the same place: the hepatocyte growth factor system and its receptor c-Met. It went through a full clinical programme, including a randomised, placebo-controlled phase 2 and 3 trial in mild to moderate Alzheimer's disease, with results posted on the registry and a published report.
Not medical advice
This page reports what has been documented about a compound. It is not medical advice, not a diagnosis, not a dosing protocol, and not a recommendation to obtain or use anything described here. Talk to a licensed clinician about anything concerning your health. Read the full disclaimer.
Reviewed against editorial standards Updated
The short answer
Fosgonimeton is the one that has been tested in people. Its phase 2 and 3 trial, LIFT-AD, enrolled 554 participants and did not achieve its primary or secondary endpoints: the between-group difference on the combined primary measure was reported as minus 0.08 with a p value of 0.70.
Dihexa has never been given to a person in any published study, and ClinicalTrials.gov returned no registration for it on 2 August 2026. The 2014 paper that established its proposed mechanism was retracted in 2025, after an expression of concern published in 2021.
Side by side
| Property | Dihexa | Fosgonimeton |
|---|---|---|
| What it is | A small angiotensin IV analog with hexanoic groups at both ends, also known as PNB-0408 | A small-molecule positive modulator of the hepatocyte growth factor system, developed under the code ATH-1017 |
| Proposed mechanism | Hepatocyte growth factor signalling through the c-Met receptor | The same system, described in its trial report as positive modulation of hepatocyte growth factor signalling |
| Where it came from | A single laboratory at Washington State University | A clinical-stage company programme, with the trial report listing company employees among its authors |
| Published human studies | None, of any design | A phase 2 and 3 trial published in 2025, plus a completed phase 2 and a phase 1 study of absorption and excretion |
| Registrations on ClinicalTrials.gov | 0 | 5 |
| The trial that matters | None exists | LIFT-AD, NCT04488419, randomised, placebo-controlled, quadruple-masked, 554 enrolled, started September 2020, completed 15 July 2024, results first posted 4 April 2025 |
| Primary endpoint | Not applicable | A Global Statistical Test combining two standard Alzheimer's scales. The trial did not achieve its primary or secondary endpoints |
| The headline number | Not applicable | A between-group difference of minus 0.08 on that combined measure, with a p value of 0.70, in the primary analysis population of 287 |
| What happened next | Nothing registered. The compound remains preclinical | The open-label extension study, NCT04886063, was terminated on 23 October 2024, and the registry records the reason as the parent trial not meeting its primary endpoint |
| Status of the key mechanism paper | Retracted. PubMed flags the 2014 paper as a retracted publication, and carries an expression of concern from 2021 and a retraction notice from 2025 | No retraction on the record |
| Independent animal work | A 2021 study in APP/PS1 mice reporting memory and signalling differences, and a 2024 rat study of a Huntington's-like model in which the compound did not protect against the deficits | Not surveyed here. The human trial is the record that decides the question |
| Corporate record | No sponsor | The registry lists the lead sponsor as LeonaBio. SEC EDGAR records the same entity's former names as Athira Pharma, Inc. through 8 January 2026 and M3 Biotechnology, Inc. before that |
| About the target | c-Met is a proto-oncogene | c-Met is a proto-oncogene |
Why the two get mixed up
The two get conflated because the mechanism sentence is identical. Product copy for dihexa describes hepatocyte growth factor and c-Met, and so does the trial report for fosgonimeton. A reader who searches that mechanism finds a large randomised trial in a peer-reviewed journal, with nothing on the surface to indicate that the trial tested a different molecule.
The lineage encourages it further. Both molecules were pursued on the same idea about the same growth factor system, and the corporate record behind the clinical programme reaches back to the period the academic work was published. What follows from this site's rules is unambiguous: a trial of one molecule sets a tier for that molecule, two compounds acting on one target are two evidence bases, and a shared mechanism moves nothing between them.
The direction of the result does the rest. A trial that misses its endpoints is easy to leave out of a summary, so a reader can meet the mechanism through the enthusiastic literature and never meet the trial that tested it. That result is on this page for the same reason the retraction is: the documents exist, they are dated, and naming only the encouraging half of them would be misleading by omission.
What the published record covers for each
Start with what the record no longer contains. The 2014 paper in the Journal of Pharmacology and Experimental Therapeutics is the one usually cited for dihexa's mechanism, and PubMed flags it with the publication type Retracted Publication [1]. Its record carries an expression of concern published in 2021 and a retraction notice published in 2025, and that notice is a separate PubMed record [2]. Both are cited here, and no evidence tier is assigned to the retracted paper's findings: a retraction takes a paper off the record, and tiering its results would put them back.
What survives on dihexa is animal work, and it points in two directions. A 2021 study in Brain Science reported that dihexa was associated with differences in memory testing and in PI3K and AKT signalling in APP/PS1 mice, a strain engineered to develop Alzheimer's-like pathology [Animal] [4]. A 2024 study in the Journal of Huntington's Disease randomised 40 male Wistar rats into vehicle, toxin, and toxin plus compound groups in a model of Huntington's-like symptoms produced by a mitochondrial toxin, and reported that the compound did not protect the animals from the deficits the toxin produced [Animal] [5]. The originating laboratory's 2013 characterisation of this class of analogs is also on the record and is not retracted [3].
What this evidence can and cannot show. These results come from mice and rats in engineered and chemically induced models of neurodegeneration, not from people. An animal model is a deliberate simplification: the pathology is created on purpose, the animal is young, the amount given is scaled to body weight in a way that does not translate directly, and the outcome is measured at a fixed point rather than lived with. Mechanism is a reason to run a trial. It is not a substitute for one, and compounds that looked mechanistically convincing have failed in people many times.
The same mechanism has now had that trial, on the other molecule. LIFT-AD was a randomised, placebo-controlled, quadruple-masked phase 2 and 3 study in mild to moderate Alzheimer's disease, with 554 participants enrolled and a primary analysis population of 287. Its primary endpoint was a Global Statistical Test combining the ADAS-Cog11 cognitive scale and the ADCS-ADL23 daily-function scale. The published report states that the trial did not achieve its primary or secondary endpoints: a between-group difference on the combined measure of minus 0.08 with a p value of 0.70, and differences on the two component scales that were also not statistically significant [Human RCT] [6] [7].
The same report records that more participants in the treatment group than in the placebo group stopped taking part because of adverse events, 14.2 percent against 4.6 percent, mostly on account of injection site reactions, and that serious adverse events were balanced between the groups [Human RCT] [6]. The open-label extension, NCT04886063, was terminated on 23 October 2024, and the registry entry states the reason in its own words: the parent trial did not meet its primary endpoint [8].
Two record-keeping facts sit behind all of this, both public and both dated. The registry lists the lead sponsor of the fosgonimeton programme as LeonaBio, and the SEC EDGAR submissions record for that corporate entity lists its former names as Athira Pharma, Inc., through 8 January 2026, and M3 Biotechnology, Inc. before that [9]. Separately, c-Met is a proto-oncogene, a gene whose normal job is to drive cell growth and division and whose overactive forms are implicated in cancer. That is a fact about the target rather than a finding about either compound, and it is worth knowing about any molecule designed to turn it on.
Our takeThe identical mechanism, characterised to two different depths. One molecule has zero human studies and a retracted mechanism paper. The other took the same idea into 554 people, missed, and published the numbers.
Frequently asked questions
Is dihexa the same as fosgonimeton?
No. They are two different molecules aimed at the same hepatocyte growth factor and c-Met system. Fosgonimeton, developed under the code ATH-1017, has a clinical programme behind it. Dihexa has no registration and no published human study of any design, and a trial of one molecule is not evidence about the other.
Was the dihexa research retracted?
The 2014 paper that established its proposed mechanism was, in 2025. PubMed flags the original with the publication type Retracted Publication, carries an expression of concern published in 2021, and holds the retraction notice as a separate record [1] [2]. Other work on the compound has not been retracted [3] [4] [5]. No human study of dihexa exists to retract.
What happened in the fosgonimeton trial?
LIFT-AD enrolled 554 participants with mild to moderate Alzheimer's disease against placebo and measured a Global Statistical Test combining a cognitive scale and a daily-function scale. The published report states that the trial did not achieve its primary or secondary endpoints, with a between-group difference of minus 0.08 and a p value of 0.70 [6]. Results are also posted on the registry entry [7].
Does the failed trial mean the mechanism is wrong?
It means one molecule aimed at that mechanism did not separate from placebo on the endpoints that trial measured, in that population, over that period. Trials fail for reasons other than a wrong mechanism: the wrong molecule, the wrong exposure, the wrong population, the wrong measure. What the result does establish is that the mechanism has been tested at scale in people once, and what came back is on the public record with its numbers.
Why does a retracted paper matter if other studies exist?
Because of what it was the source of. That paper is where the hepatocyte growth factor and c-Met explanation for dihexa came from, and almost every later description of the compound rests on it. A retraction does not establish that the mechanism is wrong. It establishes that the evidence which put it on the record is no longer part of the record, and the independent rodent work that remains reports a different signalling pathway [4].
References
- Benoist CC, Kawas LH, Zhu M, Tyson KA, Stillmaker L, Appleyard SM, Wright JW, Wayman GA, Harding JW. 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. PMID 25187433 DOI 10.1124/jpet.114.218735 Retracted in 2025, after an expression of concern published in 2021.
- Benoist CC, Kawas LH, Zhu M, Tyson KA, Stillmaker L, Appleyard SM, Wright JW, Wayman GA, Harding JW. 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" [J Pharmacol Exp Ther 351 (2014) 390-402]. J Pharmacol Exp Ther. 2025. PMID 40312093 DOI 10.1016/j.jpet.2025.103567
- McCoy AT, Benoist CC, Wright JW, Kawas LH, Bule-Ghogare JM, Zhu M, Appleyard SM, Wayman GA, Harding JW. Evaluation of metabolically stabilized angiotensin IV analogs as procognitive/antidementia agents. J Pharmacol Exp Ther. 2013. PMID 23055539 DOI 10.1124/jpet.112.199497
- Sun X, Deng Y, Fu X, Wang S, Duan R, Zhang Y. AngIV-Analog Dihexa Rescues Cognitive Impairment and Recovers Memory in the APP/PS1 Mouse via the PI3K/AKT Signaling Pathway. Brain Sci. 2021. PMID 34827486 DOI 10.3390/brainsci11111487
- Wells RG, Azzam AF, Hiller AL, Sardinia MF. Effects of an Angiotensin IV Analog on 3-Nitropropionic Acid-Induced Huntington's Disease-Like Symptoms in Rats. J Huntingtons Dis. 2024. PMID 38489193 DOI 10.3233/JHD-231507
- Porsteinsson AP, Sabbagh M, Tariot PN, Church KJ, San Martin J, Ooi KC, Daggett S, Hale MD, Holub R, Moebius HJ. Fosgonimeton in mild-to-moderate Alzheimer's disease. J Alzheimers Dis Rep. 2025. PMID 41393340 DOI 10.1177/25424823251405817
- US National Library of Medicine, ClinicalTrials.gov. LIFT-AD, a randomized, placebo-controlled, double-blind study of ATH-1017 in subjects with mild to moderate Alzheimer's disease (NCT04488419). Phase 2 and 3, 554 enrolled, started 28 September 2020, completed 15 July 2024, results first posted 4 April 2025. The same registry returned no study of dihexa when it was searched on 2 August 2026. ClinicalTrials.gov. 2026. Registry record
- US National Library of Medicine, ClinicalTrials.gov. Open label study of ATH-1017 for treatment of mild to moderate Alzheimer's disease (NCT04886063). Terminated 23 October 2024, with the registry recording the reason as the parent phase 2 and 3 trial not meeting its primary endpoint. ClinicalTrials.gov. 2024. Registry record
- US Securities and Exchange Commission, EDGAR. Company submissions record for CIK 0001620463, read 2 August 2026. Current name LeonaBio, Inc.; former names recorded as Athira Pharma, Inc. through 8 January 2026 and M3 Biotechnology, Inc. before that. SEC EDGAR. 2026. EDGAR company record