For research and educational purposes only · Not medical advice · Consult a qualified physician before any human use
Dihexa is a synthetic angiotensin IV derivative proposed to amplify HGF/c-Met signaling. Its foundational mechanistic paper was formally retracted in April 2025 for research misconduct. Its pharmaceutical successor compound (fosgonimeton) failed all endpoints in a Phase 2/3 Alzheimer's trial (n=312, September 2024). Zero human clinical trials of dihexa itself exist. The c-Met proto-oncogene safety concern is completely uncharacterized.
Dihexa has zero published human clinical trials. It has never been formally investigated in humans under an IND. No FDA, EMA, or any other regulatory approval exists anywhere in the world.
The foundational mechanistic paper (Benoist et al., 2014, Journal of Pharmacology and Experimental Therapeutics) — the paper establishing HGF/c-Met as the molecular target — was formally retracted in April 2025 due to research misconduct involving image manipulation. The 2013 cognitive effects paper (McCoy et al.) received an Expression of Concern for the same reason and remains published but compromised.
The pharmaceutical derivative of dihexa, fosgonimeton (ATH-1017), developed by Athira Pharma specifically to recapitulate dihexa's HGF/c-Met mechanism, failed all primary and key secondary endpoints in the Phase 2/3 LIFT-AD trial (n=312, September 2024). GST p=0.70, ADAS-Cog11 p=0.35. These p-values are not near-miss results; they indicate essentially no detectable signal above placebo.
In January 2025, Athira agreed to pay over $4 million to settle False Claims Act allegations related to NIH grants referencing the compromised research. The "ten million times more potent than BDNF" claim widely circulated in wellness communities originates from a paper that carries an Expression of Concern and uses a comparison that is pharmacologically misleading.
Dihexa (N-hexanoic-Tyr-Ile-(6)-aminohexanoic amide) is a synthetic peptide originally developed by researchers at Washington State University as part of a program to create therapies for Alzheimer's disease and cognitive decline. It is a derivative of angiotensin IV, a peptide in the renin-angiotensin system, chemically modified to improve potency, stability, and blood-brain barrier penetration.
The scientific lineage begins with the endogenous peptide angiotensin IV (AngIV), a hexapeptide fragment of the renin-angiotensin system that had been observed to enhance learning and memory in rodent models. Joseph Harding's laboratory at Washington State University conducted structure-activity relationship studies to identify the minimum active sequence and then chemically stabilize it. Each modification was selected to improve pharmacokinetic performance: shortening the peptide reduced enzymatic cleavage sites, while increased hydrophobicity facilitated passage across lipid membranes including the blood-brain barrier.
The key pharmacological hypothesis at discovery was that AngIV and its analogs including dihexa did not work through the classical AT4 receptor as previously assumed, but instead through allosteric potentiation of the hepatocyte growth factor (HGF) / c-Met receptor system. However, the paper establishing this mechanistic target (Benoist et al., 2014) was formally retracted in April 2025 due to research misconduct.
Dihexa's proposed mechanism, as originally described in the now-retracted Benoist et al. 2014 paper, is allosteric potentiation of the HGF/c-Met signaling pathway. Rather than acting as a direct agonist at c-Met, dihexa was proposed to bind HGF itself with high affinity, stabilizing an active HGF conformation that lowers the threshold for c-Met receptor activation. At subthreshold HGF concentrations, dihexa would have no effect; only in the presence of some endogenous HGF would it amplify the signal.
The downstream pathway proposed: c-Met activation via PI3K/AKT and MAPK/ERK signaling, leading to dendritic arborization, increased spine density, synaptogenesis, and the formation of functional memory-encoding synapses. This pathway dependency was partially confirmed by the independent Yun/Sun 2021 APP/PS1 mouse study using PI3K inhibitor blockade.
Critical caveat: the paper that established HGF/c-Met as the mechanistic target has been formally retracted. While the behavioral findings in rodent models were not the direct locus of the image manipulation, the mechanistic basis for dihexa's cognitive effects is now scientifically unsupported in its original form. Independent work has provided partial replication of behavioral effects, but the mechanism underlying those effects cannot be confidently stated.
By day 7, oral dihexa (2 mg/kg) was able to reverse the cognitive deficit seen with scopolamine, with performance no different from controls (McCoy 2013, Expression of Concern). Dihexa restored spatial learning and cognitive functions in the Morris water maze test in APP/PS1 mice (Yun 2021, independent replication, in good standing). Dihexa increased neuronal cells and the expression of SYP protein in APP/PS1 mice. Furthermore, dihexa decreased the activation of astrocytes and microglia, markedly reduced levels of the pro-inflammatory cytokines IL-1β and TNF-α, and increased the levels of the anti-inflammatory cytokine IL-10 (Yun 2021).
The McCoy 2013 paper characterized dihexa's potency at inducing synaptogenesis as approximately seven orders of magnitude (ten million times) more potent than BDNF in cell culture assays. This figure, widely circulated in wellness communities, requires critical context: it measures molar concentration efficiency in a specific cell culture assay, not overall cognitive effect magnitude. BDNF and dihexa act through entirely different receptors (TrkB vs. HGF/c-Met), making the comparison pharmacologically misleading. The paper from which this data came carries an Expression of Concern for image manipulation.
Very Low. The cognitive improvement data in scopolamine-impaired and APP/PS1 mice is biologically interesting and partially replicated independently (Yun 2021). However, the mechanistic paper establishing HGF/c-Met as the target has been retracted, and the foundational 2013 paper carries an Expression of Concern with manipulated western blot images. The behavioral findings remain hypothesis-generating but cannot be taken as mechanistically validated in their original form. The compound may have some biological activity, but the scientific foundation has been substantially undermined.
The Phase 2/3 LIFT-AD trial enrolled 312 participants with mild-to-moderate Alzheimer's disease and tested once-daily subcutaneous injections of fosgonimeton 40 mg over 26 weeks. Fosgonimeton failed to meet its primary endpoint of change on the Global Statistical Test. Treatment led to a -0.08 change in GST favoring active drug, which did not reach statistical significance (P=0.70).
Secondary endpoints were similarly negative: ADAS-Cog11 change of -1.09 vs. -0.39 for placebo (difference -0.70, P=0.35); ADCS-ADL23 P=0.61. These p-values are not near-miss results; they indicate essentially no detectable signal above placebo noise.
In pre-specified subgroups characterized by more rapid disease progression (moderate AD and APOE4 carriers), cognition and function improved or stabilized in the fosgonimeton treated group. Biomarkers showed changes consistent with HGF modulation. However, subgroup analyses from a failed primary endpoint trial cannot support regulatory approval or clinical confidence.
Low. The most consequential evidence for evaluating dihexa's clinical prospects, and it is unambiguously negative for the primary indication. Fosgonimeton was developed specifically to recapitulate dihexa's HGF/c-Met mechanism with improved pharmacokinetics. Its definitive LIFT-AD failure raises serious questions about whether HGF/c-Met modulation produces meaningful cognitive benefits in humans at all. The LIFT-AD failure is not a case of a drug that worked but in the wrong patient population or at the wrong dose: p=0.70 on the primary endpoint reflects essentially no biological signal. All Athira clinical programs have been discontinued.
In the SHAPE Phase 2 trial in Parkinson's disease dementia and Lewy body dementia (n=28), the primary endpoint was not met. In a very small post-hoc subgroup of 5 patients on the 40 mg dose, statistically significant improvement in ADAS-Cog13 was observed (-7.2 points, P=.0321). This is a post-hoc analysis of 5 patients from a failed primary endpoint study and cannot support clinical conclusions.
For traumatic brain injury, the theoretical rationale exists through HGF/c-Met's known roles in neuroprotection and neural repair following acute injury. No dedicated dihexa TBI studies have been published.
Very Low. The SHAPE n=5 post-hoc signal is interesting but entirely insufficient to draw conclusions from a failed primary endpoint trial. Traumatic brain injury remains theoretical only. Neither indication has dedicated evidence of sufficient quality to characterize dihexa's effects.
The McCoy 2013 study (Expression of Concern) examined dihexa in normal-cognition rodents as part of the study design. Dihexa did not improve performance in normal-cognition animals at the doses studied. Improvement was only observed in cognitively impaired animals (scopolamine-treated or aged rats). This is consistent with dihexa's proposed allosteric mechanism: if dihexa amplifies HGF signaling only above a pathological deficit threshold, enhancement in healthy individuals with normal HGF/c-Met function would not be expected.
The wellness and nootropic community use case for dihexa (as a cognitive enhancer for healthy individuals) is therefore unsupported even by the compromised preclinical data, not just by human trials.
Not established and explicitly negative in normal-cognition rodent studies. The theoretical mechanism (allosteric potentiation of HGF signaling) predicts no benefit in healthy individuals with normal baseline signaling. Zero human evidence. The nootropic use case for dihexa is not only unvalidated in humans but is explicitly contradicted by the preclinical data that constitutes dihexa's entire evidence base.
Preclinical only. Animal studies used 0.01-1 mg/kg.
Dihexa has no validated human dosing protocol. All dosing information in the research community is extrapolated from rodent studies or self-reported anecdotally, with no pharmacokinetic or safety validation. The absence of a Phase 1 human trial means that basic parameters (Tmax, Cmax, half-life, bioavailability, maximum tolerated dose) are completely unknown in humans. The c-Met proto-oncogene safety concern is unaddressed by any carcinogenicity study. Research-grade quality standards cannot meaningfully be applied when no validated human use exists. This compound should not be self-administered based on current evidence.
Dihexa vs. the rest of this series: Dihexa is the most evidentially compromised compound reviewed. Every other peptide in this series has at least one of the following: independent replication from multiple research groups, human trial data (even if limited), or a mechanistic paper that has not been retracted. Dihexa has none of these after the April 2025 retraction and the LIFT-AD Phase 2/3 failure.
Dihexa vs. BPC-157 (comparison of 'exciting preclinical, no human trials'): Both have substantial preclinical literature and zero validated human clinical data. But BPC-157's research base, while concentrated in one Croatian group, has not been subject to fraud findings. Dihexa's foundational paper has been retracted for misconduct. BPC-157's FDA Category 2 compounding prohibition reflects regulatory concern about evidence quality; dihexa's absence from that list reflects only that it has not received regulatory attention, not that it is safer or better validated.
Dihexa vs. Epithalon (single-source research concern): Both raise source concentration concerns, but for different reasons. Epithalon's literature comes from one Russian group with no fraud findings. Dihexa's literature originates in manipulated research that led to formal retraction, a CEO resignation, and a federal False Claims Act settlement.
Dihexa has no established safety profile in humans. It has never been administered to humans in any published study. All safety information is theoretical and based on the compound's proposed mechanism of action.
The primary theoretical safety concern is that c-Met is a proto-oncogene. HGF/c-Met pathway dysregulation is documented in gastric cancer, lung cancer, hepatocellular carcinoma, and several other malignancies. Pharmaceutical companies have developed c-Met inhibitors (capmatinib, tepotinib, crizotinib) as anticancer drugs, precisely the opposite pharmacological direction from dihexa. Whether dihexa's allosteric potentiation of HGF/c-Met carries oncogenic risk is entirely unknown because no carcinogenicity studies have been conducted. This is the most significant uncharacterized safety concern of any compound in this research series.
The fosgonimeton (ATH-1017) Phase 1 safety data (n=88, 9-day multi-dose escalation) found no treatment-related serious adverse events. However, this covers 9 days of exposure in 88 subjects for a chemically distinct molecule, and cannot be extrapolated to dihexa itself or to longer-term use.
No adverse event database exists for dihexa. Self-experimenter anecdotal reports circulate online but have no clinical validity and no systematic collection.
Dihexa is the most evidentially compromised compound in this research series. It sits at the intersection of three compounding negative factors not seen together in any other compound reviewed: (1) zero human clinical trials of the compound itself; (2) research fraud disclosures that resulted in formal retraction of the foundational mechanistic paper; and (3) failed Phase 2/3 clinical trial of the class mechanism via the direct pharmaceutical successor compound (fosgonimeton, LIFT-AD, September 2024).
The compound may still have biological activity. The Yun 2021 independent replication in APP/PS1 mice provides some basis for this view, and the behavioral water maze data was not the direct locus of the image manipulation. But the current state of its evidence base does not support human use for any indication. The c-Met oncological safety concern remains completely unaddressed. The "ten million times more potent than BDNF" claim, which has driven enormous wellness community interest, originates from compromised research and is pharmacologically misleading regardless.
Honest characterization requires stating clearly: the scientific and clinical trajectory of dihexa's entire development program has been negative, and the recent events (retraction, LIFT-AD failure, False Claims Act settlement) represent genuine scientific closure on the original mechanistic hypothesis rather than temporary setbacks.
For research and educational purposes only · Not medical advice · Consult a qualified physician before any human use