How Dihexa Modulates HGF in the Brain

Dihexa

1 Key Paper Retracted

The central study linking dihexa to HGF/c-Met synaptogenesis was retracted in 2025 after Washington State University confirmed fabricated data.

J Pharmacol Exp Ther, Retraction 2025

J Pharmacol Exp Ther, Retraction 2025

View as image

Dihexa (N-hexanoic-Tyr-Ile-(6) aminohexanoic amide) is an angiotensin IV-derived peptide proposed to enhance cognition by modulating the hepatocyte growth factor (HGF) / c-Met receptor system in the brain. The hypothesis was compelling: HGF drives synaptogenesis and neuronal survival, and dihexa would amplify that signal at sub-therapeutic HGF concentrations. But the central study establishing this mechanism was retracted in April 2025 after Washington State University's investigation found fabricated data. That retraction does not erase all evidence for the HGF/c-Met connection, but it removes the foundation on which the strongest mechanistic claims were built. For a broader look at the potency claims surrounding this peptide, see dihexa's potency claims: what the original research actually says.

The Proposed Mechanism: HGF/c-Met Signaling

Hepatocyte growth factor (HGF) is a pleiotropic growth factor that binds the c-Met receptor tyrosine kinase. While originally characterized in liver biology, HGF/c-Met signaling plays critical roles in the central nervous system. Wright et al. (2015) reviewed the evidence for the brain HGF/c-Met system as a target for Alzheimer's and Parkinson's treatment, documenting that HGF promotes neuronal survival, drives dendritic arborization, and stimulates the formation of new synaptic connections (spinogenesis and synaptogenesis).^[4]

The proposed mechanism for dihexa works in three steps:

1. HGF binding: Dihexa binds directly to HGF with high affinity
2. Facilitated dimerization: This binding promotes HGF dimerization, which is required for c-Met receptor activation
3. c-Met phosphorylation: The dimerized HGF activates c-Met even at HGF concentrations too low to trigger signaling on their own

The downstream effects of c-Met activation include:

• PI3K/AKT pathway: neuronal survival and anti-apoptotic signaling
• MAPK/ERK pathway: synaptic plasticity and gene expression
• NMDA receptor potentiation: enhanced long-term potentiation, the cellular basis of memory

Wright et al. (2015) also proposed that the brain HGF/c-Met system represents an alternative neurotrophic target to the better-known BDNF/TrkB system, with the advantage that HGF can be modulated by small molecules (like dihexa) that cross the blood-brain barrier.^[5]

Why HGF Matters for Cognition

HGF is not a fringe target. The c-Met receptor is expressed on neurons in the hippocampus, cortex, and cerebellum. HGF levels decline with age, and reduced HGF/c-Met signaling has been documented in Alzheimer's disease brain tissue. In cell culture, HGF promotes the growth of new dendritic spines (the tiny protrusions on neurons that form synaptic connections) and increases synapse density.

The appeal of modulating HGF rather than delivering it directly is practical. HGF is a large protein (82 kDa) that cannot cross the blood-brain barrier. Delivering it to the brain would require surgical infusion, as has been attempted (with mixed success) for other neurotrophic factors like GDNF in Parkinson's disease. A small molecule that enhances existing HGF activity could theoretically achieve the same neurotrophic benefit from a peripheral injection or oral dose. This was the core rationale for developing dihexa: not as a replacement for HGF, but as an amplifier of whatever HGF signaling already exists in the brain.

In Alzheimer's disease, where HGF levels in the hippocampus are reduced compared to age-matched healthy brains, an HGF amplifier could theoretically restore a declining neurotrophic signal. In traumatic brain injury, where HGF is transiently elevated as part of the repair response, amplification could boost the brain's endogenous recovery mechanisms. Both applications depend on the compound actually reaching the brain and specifically engaging the HGF/c-Met system, which is exactly what the retracted data claimed to demonstrate.

The Retraction: What Happened

The foundational mechanistic study was Benoist et al. (2014), published in the Journal of Pharmacology and Experimental Therapeutics. This paper presented evidence that dihexa and its parent compound Nle(1)-AngIV induced c-Met phosphorylation in the presence of subthreshold HGF concentrations, augmented HGF-dependent cell scattering, and induced hippocampal spinogenesis and synaptogenesis.

In 2021, the journal's editors issued an expression of concern after receiving reports of possible image manipulation. Washington State University conducted a formal investigation that concluded key figures in the paper contained fabricated and falsified data. Researchers Leen H. Kawas and Joseph W. Harding were found solely responsible. The paper was formally retracted in April 2025.

The consequences extended beyond the journal:

• WSU revoked Kawas's PhD
• Athira Pharma, the biotech company Kawas co-founded and led as CEO, had developed ATH-1017 (fosgonimeton), a prodrug of dihexa, as its lead Alzheimer's candidate
• In January 2025, Athira agreed to pay over $4 million to settle False Claims Act allegations related to NIH grants referencing the compromised research
• At least one additional paper from the same lab (on angiotensin IV analogs as HGF/Met modifiers) was also retracted

The retraction does not prove that dihexa has no biological activity. It means the specific evidence showing that dihexa's cognitive effects depend on HGF/c-Met activation contained fabricated data. The mechanistic link between dihexa and HGF/c-Met synaptogenesis has lost its primary evidential support.

This distinction matters. A compound can have real biological effects through a different mechanism than what was claimed. Aspirin worked for decades before its prostaglandin inhibition mechanism was understood. But the specific narrative that dihexa amplifies HGF signaling to create new synapses cannot be treated as established science when the key supporting figures were fabricated.

What Independent Evidence Remains

Animal Cognition Studies

Several studies from independent groups have tested dihexa in animal models without relying on the retracted data.

Sun et al. (2021) tested dihexa in APP/PS1 transgenic mice (a model of Alzheimer's disease) and found that the peptide rescued cognitive impairment as measured by the Morris water maze.^[2] The study identified PI3K/AKT signaling as the active pathway but did not specifically confirm that HGF/c-Met was the upstream trigger. The cognitive rescue was real, but the mechanism could operate through pathways other than HGF.

Wells et al. (2024) tested an angiotensin IV analog (structurally related to dihexa) in a rat model of Huntington's disease induced by 3-nitropropionic acid.^[6] The results showed neuroprotective effects, including reduced striatal lesion size and improved motor performance. This extends the evidence for angiotensin IV-class peptides beyond Alzheimer's models into other neurodegenerative conditions. The study does not depend on any retracted data, but it tests a related compound rather than dihexa itself, and cannot confirm the HGF/c-Met mechanism specifically.

Systematic Review

Ho et al. (2018) conducted a systematic review of 37 experimental studies examining cognitive benefits of angiotensin IV and angiotensin-(1-7).^[3] The review found consistent procognitive effects across multiple animal models, behavioral paradigms, and laboratories. Improvements were observed in spatial memory (Morris water maze), passive avoidance, and object recognition tasks. The review included studies predating the retracted work and studies from groups with no connection to the Harding laboratory, providing evidence that the angiotensin IV class of peptides has genuine cognitive activity independent of any single research group's claims. The critical question is whether that activity operates through HGF/c-Met, through insulin-regulated aminopeptidase (IRAP) inhibition, or through mechanisms not yet characterized.

The Original McCoy 2013 Study

McCoy et al. (2013) evaluated metabolically stabilized angiotensin IV analogs, including dihexa, as procognitive and antidementia agents.^[1] This study predates the retracted work and was authored partly by the same group (Harding lab at WSU), but it has not been retracted. The procognitive effects it reports (improved performance in scopolamine-challenged rats) have been partially replicated by independent groups. Any data from this lab should be interpreted with caution given the broader integrity issues. For more context on dihexa and Alzheimer's research, see the dedicated article.

Alternative Mechanistic Hypotheses

The retraction of the HGF/c-Met evidence reopens the question of how angiotensin IV-class peptides produce cognitive effects. Two alternative mechanisms have been proposed:

IRAP Inhibition

Angiotensin IV was originally identified as an inhibitor of insulin-regulated aminopeptidase (IRAP), also known as the AT4 receptor. IRAP degrades several neuropeptides involved in memory and cognition, including vasopressin, oxytocin, somatostatin, and cholecystokinin. Inhibiting IRAP would increase local concentrations of these neuropeptides in synaptic spaces, potentially explaining the memory-enhancing effects without requiring any HGF/c-Met engagement.

Hanzevacki et al. (2025) used quantum mechanics/molecular mechanics simulations to distinguish how IRAP handles its substrate oxytocin versus its inhibitor angiotensin IV, providing molecular-level detail about the binding interaction.^[7] Their analysis showed that angiotensin IV occupies the IRAP active site in a way that prevents substrate access, confirming the enzyme inhibition mechanism at the atomic level. This work supports IRAP as a legitimate and well-characterized target of angiotensin IV-class compounds.

The IRAP hypothesis has an important advantage over the HGF/c-Met hypothesis: it has been independently validated by multiple research groups using different techniques, including radioligand binding studies, enzyme inhibition assays, and computational modeling. The evidence for IRAP inhibition does not depend on any retracted data.

Direct Receptor Activity

Beyond IRAP and HGF/c-Met, dihexa may interact with other aminopeptidases that regulate neuropeptide signaling in the brain. The pharmacology of dihexa at targets beyond HGF has not been thoroughly characterized, in part because the HGF/c-Met hypothesis dominated the research direction for over a decade. With that hypothesis now weakened, a broader pharmacological profile of dihexa's binding targets could clarify which mechanism actually drives the procognitive effects observed in animal studies.

The Gap Between Mechanism and Marketing

Dihexa has been described in peptide vendor marketing as "millions of times more potent than BDNF." That claim derives from comparing the effective concentration of dihexa for inducing neurite outgrowth (picomolar) with that of BDNF (nanomolar to micromolar), but compares two entirely different assay systems measuring different endpoints. The claim was always misleading on its own terms, and the retraction of the key mechanistic paper further undermines the precision of the underlying data.

No human clinical trials of dihexa have been completed. Athira Pharma's fosgonimeton (ATH-1017) was in clinical development but the program was disrupted by the data integrity investigation. The compound's clinical status remains uncertain.

The HGF/c-Met system remains a legitimate neurotrophic target in the brain. The question is whether dihexa specifically engages this system, and the primary evidence supporting that specific claim has been retracted.

Where the Science Stands in 2026

The dihexa-HGF story illustrates a broader problem in peptide research: much of the excitement around certain compounds rests on a small number of publications from a single laboratory. When that research is compromised, the entire evidence base collapses to whatever independent replication exists.

For dihexa specifically:

• The HGF/c-Met mechanism has lost its foundational evidence
• The IRAP inhibition mechanism is independently validated but explains angiotensin IV-class activity broadly, not dihexa specifically
• Independent animal studies show procognitive effects, but no study has definitively identified the responsible receptor or pathway
• No human data exists for dihexa in any indication
• The peptide is sold by research chemical vendors and gray market sources, but its use in humans is based entirely on preclinical animal data from a research program partially tainted by fabrication
• The gap between the marketing claims surrounding dihexa and the actual evidence base is among the largest in the peptide space

The amyloid-beta hypothesis in Alzheimer's faced similar challenges when key studies were questioned, and the broader lesson applies here: extraordinary claims about peptide potency demand extraordinary evidence, and a single retracted paper can shift the entire risk-benefit calculation.

The Bottom Line

Dihexa was proposed to enhance cognition by facilitating HGF dimerization and c-Met receptor activation in the brain. The central study supporting this mechanism was retracted in 2025 after fabricated data was confirmed. Independent animal studies show angiotensin IV-class peptides have genuine procognitive effects, but whether those effects operate through HGF/c-Met or through alternative pathways like IRAP inhibition remains unresolved. No human data exists for dihexa in any indication.

Frequently Asked Questions

How does dihexa work in the brain?

Dihexa was proposed to work by binding hepatocyte growth factor (HGF) and facilitating its dimerization, which activates the c-Met receptor on neurons. This would trigger PI3K/AKT and MAPK/ERK signaling cascades that promote synaptogenesis and neuronal survival. The key 2014 paper supporting this specific mechanism was retracted after data fabrication was confirmed. Alternative mechanisms including IRAP inhibition may explain its observed cognitive effects in animal models.

Was the dihexa research retracted?

Yes. The key 2014 study by Benoist et al. linking dihexa's cognitive effects to HGF/c-Met activation was retracted in April 2025. Washington State University's investigation found that figures contained fabricated and falsified data, with researchers Leen Kawas and Joseph Harding found solely responsible. At least one additional paper from the same lab was also retracted. WSU revoked Kawas's doctoral degree.

Does dihexa have any legitimate research behind it?

Independent studies do show cognitive effects. Sun et al. (2021) found dihexa rescued memory deficits in Alzheimer's model mice via PI3K/AKT signaling. Ho et al. (2018) reviewed 37 studies showing consistent procognitive effects for angiotensin IV-class peptides. These independent results suggest biological activity, but the specific HGF/c-Met mechanism remains unconfirmed after the retraction.

What is the HGF c-Met system in the brain?

The HGF/c-Met system is a growth factor signaling pathway where hepatocyte growth factor (HGF) binds the c-Met receptor tyrosine kinase. In the brain, this system promotes neuronal survival, drives formation of new dendrites, and stimulates synapse creation. Wright et al. (2015) proposed it as an alternative neurotrophic target to BDNF for treating Alzheimer's and Parkinson's disease.

Has dihexa been tested in humans?

No. No human clinical trials of dihexa have been completed. Athira Pharma developed a prodrug of dihexa called fosgonimeton (ATH-1017) for Alzheimer's disease, but the clinical program was disrupted after the data fabrication investigation involving the company's co-founder. The compound's clinical development status remains uncertain.

Is dihexa really millions of times more potent than BDNF?

That claim compares effective concentrations across entirely different assay systems measuring different endpoints. Dihexa induced neurite outgrowth at picomolar concentrations while BDNF required higher concentrations in a different assay, but these numbers cannot be directly compared. The claim was always scientifically misleading, and the retraction of the key mechanistic paper further undermines the data behind it.