Does Dihexa Actually Work? The 2025 Retractions and What Evidence Still Stands

Two separate questions get tangled together whenever people ask whether dihexa "works." The first is whether dihexa is a real, synthesizable compound with defined chemistry. It is. The second is whether the striking cognitive-enhancement claims attached to it are supported by sound data. After 2025, that second answer is substantially weaker than it was a decade ago.

The compound itself, N-hexanoic-Tyr-Ile-(6)-aminohexanoic amide, was designed at Washington State University (WSU) as a metabolically stabilized, orally bioavailable analog of angiotensin IV (AT IV, a hexapeptide fragment of angiotensin II that published animal research had linked to memory effects). That chemistry work is not what came apart. What came apart is the mechanistic and potency evidence layered on top of it: the imaging and blot data that were used to argue dihexa drives synapse formation through the HGF/c-Met pathway at extraordinary potency.

So the honest framing is this. Dihexa is legitimate as a molecule. Its headline story as a uniquely potent synaptogenic compound rests on a body of work that now includes one full retraction and one Notice of Concern, authored in part by a researcher found to have manipulated images. That does not prove dihexa is inert. It means the strongest published support for the exciting claims has been withdrawn, and independent replication to fill the gap does not exist in the published record.

The dihexa story unfolded over roughly 15 years, from initial publication to formal retraction. The sequence matters because it shows the claims circulated widely for years before the integrity problems surfaced.

The practical takeaway: much of the enthusiasm around dihexa in nootropic and research communities was seeded during the 2013 to 2019 window, before any of the corrective notices appeared. Content that predates 2021 generally repeats the original claims without knowledge of what followed.

Leen Kawas earned her PhD at WSU in the laboratory that developed dihexa, then co-founded and led Athira Pharma, a biotech built around a related HGF-pathway compound. In 2021, an internal Athira investigation, prompted by outside allegations, examined images in her doctoral dissertation and in papers she co-authored between 2011 and 2014.

According to reporting on that investigation and subsequent institutional actions, the review concluded that images had been altered, and Kawas left the company. WSU subsequently revoked her doctoral degree. Public commentary on the flagged figures, including analysis posted on PubPeer, focused on apparent duplication and manipulation within western blot and microscopy panels, which are precisely the data used to support the pathway and synaptogenesis claims.

Two points deserve emphasis for a research audience. First, image integrity findings speak to whether the reported data can be trusted as presented, not necessarily to whether the underlying biological hypothesis is true or false. A manipulated blot does not prove the opposite conclusion; it removes the evidence. Second, the misconduct findings were serious enough to draw a federal settlement of roughly $4 million from Athira over how the situation intersected with grant applications, and serious enough for a journal to retract a paper more than a decade after publication. Journals retract reluctantly. That threshold being crossed is a strong signal.

The single most repeated dihexa claim is that it promotes synaptogenesis at concentrations roughly seven orders of magnitude, about 10 million times, lower than BDNF (brain-derived neurotrophic factor, a growth factor that supports synapse formation and is a common benchmark for pro-plasticity compounds). This is the number that made dihexa a sensation.

That comparison traces directly to the WSU publications, including the paper now retracted (PMID 25187433) and the paper now under a Notice of Concern (PMID 23055539). The femtomolar-range potency was inferred from the same class of imaging assays that were later flagged. When the supporting data for a quantitative claim is retracted, the claim does not automatically become false, but it loses its published basis. There is no independent laboratory replication of the ~10^7 potency figure in the peer-reviewed literature.

For a researcher, the responsible read is to treat the potency headline as a hypothesis that was never independently confirmed, rather than as an established property of the compound.

Stripping away the withdrawn material, a modest amount of genuinely useful information remains.

The chemistry survives. Dihexa's structure, synthesis route, metabolic stabilization strategy, and design rationale as an AT IV analog are described in the 2013 paper and are not the subject of the image-manipulation findings. The compound is real and can be made and characterized, which is why it exists as a research chemical at all.

The broader angiotensin IV literature survives and predates the controversy. Multiple groups, independent of the Kawas figures, reported that AT IV and related fragments produce memory effects in animal models through mechanisms distinct from classical angiotensin receptors. The AT IV cognitive hypothesis is older and wider than dihexa itself.

The general HGF/c-Met synaptogenesis biology survives as a field. HGF signaling through the c-Met receptor is an established regulator of neuronal growth and survival in work from many laboratories. What is compromised is the specific, dihexa-linked demonstration that this compound drives that pathway at the reported potency.

What does not survive is arguably the most important part for anyone asking whether dihexa works: independent behavioral replication and any human efficacy data. The clinical successor compound developed from this same HGF-pathway program, fosgonimeton, was reported in 2024 to have not met the primary endpoint in its pivotal Alzheimer's disease trial. That outcome is a distinct compound and does not directly test dihexa, but it removes the one adjacent line of human evidence that optimists had pointed to.

Dihexa's proposed mechanism is potentiation of signaling through c-Met (the receptor tyrosine kinase encoded by the MET proto-oncogene). That same pathway is one of the better-characterized drivers of cancer biology. HGF/c-Met signaling promotes cell proliferation, migration, invasion, and angiogenesis, and MET amplification or activation is implicated across multiple tumor types. A substantial oncology drug-development effort is aimed at inhibiting c-Met, which is the opposite of what a c-Met potentiator would do.

This creates a legitimate theoretical concern that is independent of the retractions: a compound designed to chronically enhance c-Met signaling engages a pathway that, in other contexts, supports tumor growth. It is a mechanism-level question, not a claim that dihexa causes cancer.

The honest status is that no published carcinogenicity study, no long-term rodent tumor-incidence data, and no human safety follow-up exists for dihexa. There is therefore no evidence that it is carcinogenic and no evidence that it is safe on this axis. For a proliferation-linked pathway, the absence of long-term safety data is itself a meaningful limitation that researchers should weigh explicitly.

Even setting aside the integrity problems, the dihexa literature has hard boundaries that no amount of enthusiasm can cross.

It cannot tell you that dihexa improves cognition in humans. There is no published human efficacy trial. Every claim of human benefit is extrapolation from compromised or unreplicated animal data, not a finding.

It cannot tell you the true potency of the compound. The signature ~10^7 potency comparison rests on retracted and flagged imaging. Without independent replication, the real dose-response relationship in any living system is unestablished.

It cannot tell you the long-term safety profile. No chronic-exposure toxicology, no carcinogenicity assessment given the c-Met/MET oncogene link, and no human pharmacovigilance data exist in the published record.

It cannot tell you that the mechanism is real as reported. The specific evidence that dihexa acts through HGF/c-Met to drive synaptogenesis is the exact material that was retracted. The general pathway biology from other labs does not fill this gap for this compound.

What the record can support is narrow: dihexa is a defined AT IV-derived molecule with characterized chemistry, developed in a program whose central efficacy and mechanism papers now carry a retraction and a Notice of Concern. That is the ceiling of what an evidence-first reading allows.

Dihexa is a useful case study in why provenance matters. The compound became widely discussed on the strength of a few dramatic numbers, those numbers came disproportionately from one research program, and that program later failed integrity review. The claims outran the evidence, then the evidence was withdrawn, and the claims kept circulating anyway.

A disciplined approach is to weight independent replication heavily and single-source spectacular findings lightly, especially when the single source has a known integrity problem. A potency claim of 10 million-fold over a gold-standard growth factor is exactly the kind of extraordinary result that demands independent confirmation before it enters your working model. For dihexa, that confirmation is absent.

For researchers who still wish to study the compound in appropriate preclinical models, the value now lies in treating the earlier claims as open questions to be tested rather than as established background facts. The mechanism, the potency, and the behavioral effects are all live hypotheses that the current literature does not settle. Framing dihexa work that way is both more honest and more scientifically productive than repeating the retracted headline.

For a general primer on separating strong from weak preclinical claims, see How to Evaluate Peptide Research.

Blackwell BioLabs offers Dihexa strictly for research purposes only (RUO). All batches are third-party tested with HPLC purity confirmation and mass spectrometry identity verification. Nothing in this article should be read as a claim of human cognitive benefit, a safety assurance, or a recommendation for use in humans; the published efficacy and mechanism evidence for dihexa includes a formal retraction and a Notice of Concern, and no human efficacy trial exists.

For the surrounding research context, see Dihexa Explained, the mechanism background in Dihexa and HGF/c-Met Research, and the methodology primer How to Evaluate Peptide Research.