Dihexa: Retracted Foundations, Independent Replication and Unresolved Risk

Dihexa (N-hexanoic-Tyr-Ile-(6) aminohexanoic amide, developmental code PNB-0408, PubChem CID 11527385) is a synthetic peptidomimetic derived from angiotensin IV. It was developed at Washington State University in the laboratory of Joseph W. Harding through systematic structure-activity relationship studies aimed at creating a metabolically stable, orally bioavailable, blood-brain-barrier-penetrant compound that would replicate the cognitive-enhancing effects of angiotensin IV without requiring central administration.

Dihexa is technically a hexapeptide derivative, not a full peptide, though it retains peptide-like properties. Its hexanoic acid modifications at both termini were engineered specifically for metabolic stability and oral absorption, properties that are rare among peptide compounds and contributed to the early excitement surrounding the molecule. The compound gained widespread attention for a claimed potency comparison: in specific cell culture assays measuring synaptogenesis, dihexa was reported to be active at picomolar concentrations where BDNF required nanomolar concentrations, a ratio frequently described as “10 million times more potent than BDNF.” This comparison refers specifically to molar potency for synapse formation in a particular assay, not to overall cognitive effect magnitude.

Property	Detail
Full chemical name	N-hexanoic-Tyr-Ile-(6) aminohexanoic amide
PubChem CID	11527385
Origin	Washington State University, Joseph Harding laboratory
Parent compound	Angiotensin IV (Val-Tyr-Ile-His-Pro-Phe)
Proposed mechanism	HGF/c-Met potentiation (foundational paper retracted April 2025)
Alternative pathway	PI3K/AKT signalling (Sun 2021, independent replication)
Human clinical trials	None (all data preclinical as of May 2026)
Human safety data	None published
Spinoff company	Athira Pharma (formerly M3 Biotechnology). IPO 2020, $622M peak valuation

Dihexa is part of a broader cluster of nootropic and neuropeptide compounds covered on this site. For the cross-category overview, see the nootropic and neuropeptide compounds hub.

The bottom line: Dihexa is a synthetic angiotensin IV derivative with striking preclinical synaptogenic data, but its evidence base is uniquely compromised by confirmed research misconduct, retracted foundational papers, and a complete absence of human safety or efficacy data.

Dihexa belongs to a family of angiotensin IV analogs originally designed to target the AT4 receptor, later identified as insulin-regulated aminopeptidase (IRAP). The original hypothesis held that inhibiting IRAP in the brain enhanced memory by preserving neuropeptide substrates. However, the Harding group’s trajectory shifted after a 2010 study by Albiston, Fernando, and colleagues found that IRAP knockout mice showed accelerated age-related spatial memory decline rather than the improvement predicted by the inhibition hypothesis. This paradoxical result pushed the research programme toward an alternative explanation: that dihexa-class compounds produced their cognitive effects by activating the HGF/c-Met receptor system.

The resulting HGF/c-Met hypothesis proposed that dihexa acts as an allosteric modulator facilitating HGF dimerisation and binding to the c-Met receptor tyrosine kinase, amplifying downstream signalling cascades involved in neuronal growth, survival, and synaptic connectivity.

However, the primary evidence for this mechanism – the Benoist et al. 2014 paper in the Journal of Pharmacology and Experimental Therapeutics that characterised dihexa’s HGF binding affinity (Kd = 65 pM) and c-Met phosphorylation – was formally retracted in April 2025 after Washington State University’s investigation confirmed the figures contained falsified and fabricated data. The earlier Kawas et al. 2012 paper in the same journal, which originally framed dihexa-class molecules as HGF-dimerisation mimetics, was also retracted in April 2025. This means the foundational biochemical characterisation and primary mechanistic framing for dihexa have both been withdrawn from the published record.

Independent replication: the PI3K/AKT pathway

An independent 2021 study by Sun and colleagues at China Pharmaceutical University and Nanjing Medical University (published in Brain Sciences) used APP/PS1 transgenic Alzheimer’s mice – a different and more disease-relevant model than the scopolamine-treated rats used in the original WSU work. The Chinese group found that dihexa restored spatial learning and cognitive function in Morris water maze testing, increased neuronal cells and synaptophysin protein expression, decreased astrocyte and microglia activation, reduced pro-inflammatory cytokines (IL-1beta, TNF-alpha), and increased the anti-inflammatory cytokine IL-10. Crucially, the study identified the PI3K/AKT signalling pathway rather than HGF/c-Met as the mechanism, with the PI3K inhibitor wortmannin significantly reversing dihexa’s anti-inflammatory and anti-apoptotic effects.

Additional indirect evidence comes from stem cell biology: multiple independent laboratories unrelated to WSU have used dihexa as a functional HGF receptor agonist in hepatocyte differentiation protocols since 2015 (Siller et al. 2015, Mathapati et al. 2016), suggesting genuine HGF interaction despite the compromised binding data. This use case is covered in detail in the stem cell biology section below.

The bottom line: Dihexa’s proposed HGF/c-Met mechanism rests on retracted foundational papers, but independent replication from a Chinese research group confirms cognitive effects in Alzheimer’s mice through a PI3K/AKT pathway, and functional evidence from stem cell research suggests genuine HGF interaction despite the compromised binding data.

Understanding dihexa requires understanding the commercial entity built on its research. Leen Kawas, the PhD student who co-developed dihexa in Harding’s laboratory, co-founded M3 Biotechnology in 2013 to commercialise dihexa and a related prodrug molecule, ATH-1017 (fosgonimeton). M3 rebranded as Athira Pharma in 2019. By September 2020, Athira went public at $17 per share, raising $238 million and reaching a peak valuation of $622 million. Kawas controlled 6.5 million shares worth over $110 million at peak.

In 2021, Athira’s board placed Kawas on leave after allegations emerged that she had manipulated images in her doctoral dissertation and in published research papers. A special committee investigation confirmed that Kawas altered images in her 2011 doctoral dissertation and at least four research papers co-authored between 2011 and 2014. The manipulations included copying and pasting data between experiments, digitally altering western blot band intensities, and reusing identical images to represent different experimental conditions.

The consequences cascaded over the following years. On January 6, 2025, the US Department of Justice announced that Athira Pharma agreed to pay $4,068,698 to settle False Claims Act allegations that it failed to report research misconduct when applying for NIH grants between January 2016 and June 2021. The whistleblower who forced accountability was Andrew Mallon. Separately, a securities class action settlement cost an additional $10 million. Washington State University revoked Kawas’s PhD.

Paper	Status	Significance
Kawas et al. 2012, JPET	Retracted April 2025	Original HGF-dimerisation mimetic framing
Benoist et al. 2014, JPET	Retracted April 2025	Primary HGF/c-Met binding characterisation
McCoy et al. 2013, JPET	Expression of concern (not retracted as of May 2026)	Behavioural and synaptogenic effects, oral cognitive restoration in rats
Sun et al. 2021, Brain Sciences	Published (independent, no integrity concerns)	Independent replication of cognitive effects via PI3K/AKT pathway

Athira’s lead clinical compound, fosgonimeton (ATH-1017), a prodrug of dihexa, entered clinical trials for Alzheimer’s disease. The Phase 2 ACT-AD trial failed its primary endpoint, and the larger LIFT-AD trial failed both primary and key secondary endpoints. The company’s special committee noted that Athira had conducted alternative preclinical studies to support ATH-1017’s activity independently of the challenged papers, and that the issued US patent claiming ATH-1017 did not cite any paper found to contain altered images. Nevertheless, fosgonimeton’s clinical failure raises questions about whether dihexa’s preclinical promise translates to human efficacy through any pathway.

The bottom line: Dihexa’s commercial development involved confirmed data fabrication, a $4 million False Claims Act settlement, $10 million in securities litigation, a revoked PhD, and the failure of its prodrug in human Alzheimer’s trials – the most extensively documented research integrity case in the research peptide space.

All published evidence for dihexa is preclinical. No completed or ongoing human efficacy trial appears in PubMed or ClinicalTrials.gov as of May 2026. The evidence base, such as it exists after the retractions, can be assessed as follows:

Finding	Source	Integrity status
HGF dimerisation and c-Met activation	Kawas 2012	Retracted – cannot be relied upon
HGF binding Kd = 65 pM, c-Met phosphorylation, synaptogenesis	Benoist 2014	Retracted – cannot be relied upon
Oral cognitive restoration in scopolamine and aged rats	McCoy 2013	Expression of concern – interpret with caution
Cognitive rescue in APP/PS1 AD mice via PI3K/AKT	Sun 2021 (independent)	Clean – no integrity concerns
Sciatic nerve repair with stem cells	Weiss 2021	Clean – no integrity concerns
Functional HGF agonism in stem cell hepatocyte differentiation	Multiple independent groups	Clean – functional use confirms interaction

The bottom line: The science behind dihexa is compromised but not necessarily wrong – an independent Chinese study confirms cognitive effects through a different pathway, and functional stem cell evidence suggests genuine HGF interaction, but the quantitative binding data and the specific mechanistic framing that drove commercial development are no longer reliable.

Dihexa in stem cell biology

One of dihexa’s least discussed but most consequential real-world uses has nothing to do with nootropics. Since 2015, dihexa has been used as a standard small-molecule tool in stem cell differentiation protocols for generating hepatocyte-like cells from human pluripotent stem cells. The Siller et al. 2015 (Stem Cell Reports) and Mathapati et al. 2016 (Current Protocols in Stem Cell Biology) publications established a protocol in which 100 nM dihexa replaces recombinant HGF protein during the terminal hepatoblast-to-hepatocyte maturation phase, producing cells with polygonal morphology, distinct nuclei, and functional hepatocyte characteristics within three days of treatment.

This protocol remains in active use. A May 2025 Frontiers in Cell and Developmental Biology paper (Asumda et al.) directly compared the small-molecule (dihexa-based) and growth-factor-based hepatocyte differentiation approaches, confirming both produce functional hepatocyte-like cells with the dihexa protocol offering advantages in cost and reproducibility. The key significance for dihexa’s mechanistic credibility is that these stem cell laboratories have nothing to do with the Harding/Kawas research group and use dihexa purely as a functional HGF receptor agonist. The fact that dihexa reliably drives HGF-dependent hepatocyte maturation across multiple independent laboratories provides indirect but genuine evidence that the compound does interact with HGF-dependent pathways, even though the specific binding affinity data from the retracted papers cannot be relied upon.

The bottom line: Dihexa’s continued use as a standard reagent in stem cell hepatocyte differentiation protocols across independent laboratories worldwide provides the strongest surviving functional evidence that the compound genuinely engages HGF-dependent signalling – a finding entirely separate from the nootropic claims and uncontaminated by the Athira Pharma integrity crisis.

Dihexa acetate was placed on FDA Category 2 of the interim 503A bulks list in September 2023, meaning the FDA had identified significant safety concerns and compounding was not permitted. On April 15, 2026, the FDA announced removal of 12 peptides from Category 2 due to withdrawal of nominations, with dihexa acetate among them. The removal became effective approximately April 22, 2026.

Removal from Category 2 does not place dihexa on the positive 503A bulks list or into Category 1. It exists in a regulatory grey area until the PCAC meets to review it, which is scheduled before February 2027. For the distinction between these categories, see the Category 1 vs Category 2 explainer.

Jurisdiction	Status
FDA (USA)	Not approved. Category 2 removed April 22, 2026. PCAC review before February 2027. Not on 503A bulks list
MHRA (UK)	Not licensed
TGA (Australia)	Not approved
WADA	Not individually listed

For a broader view of how different countries approach peptide regulation, see the guides for the US, the UK and Australia.

The bottom line: Dihexa’s regulatory path is further behind than other compounds in this cluster – removed from Category 2 but not yet reviewed by PCAC, with retracted foundational research that will complicate any future evaluation.

No human safety data exists for dihexa as of May 2026. No Phase I trial has been conducted for dihexa itself (as distinct from the prodrug fosgonimeton/ATH-1017). This means the complete safety profile in humans is unknown.

The most significant theoretical safety concern is oncological. The HGF/c-Met pathway is frequently dysregulated in cancers, and c-Met is an established oncogene. A compound that potentiates HGF signalling could theoretically promote tumour growth, angiogenesis, or metastasis. This concern remains theoretical because no long-term safety study has been conducted to assess or rule it out, but it represents a mechanistic red flag that would likely feature prominently in any PCAC review.

Fosgonimeton (the dihexa prodrug ATH-1017), which did enter human clinical trials under Athira Pharma, reported adverse events in its Phase 2 programmes. However, fosgonimeton is a distinct compound from dihexa with different pharmacokinetics, and its adverse event profile cannot be directly applied to the parent molecule.

For information on evaluating the purity and identity of any research peptide, see the certificate of analysis guide.

The bottom line: Dihexa has zero published human safety data, a theoretical oncological concern via c-Met pathway potentiation, and its prodrug’s clinical failures raise questions about whether preclinical efficacy translates to humans.

Within the cognitive and neuropeptide cluster on this site, dihexa occupies a fundamentally different position from Semax or Selank (which have approved clinical use in Russia and decades of published human data) or even DSIP (which has limited but existent human exposure data). Dihexa is the only compound in the cluster with retracted foundational research and a complete absence of human data. Its angiotensin IV lineage distinguishes it mechanistically from every other compound in this cluster.