Dihexa

Dihexa

Dihexa is am extremly potent synthetic peptide that promotes synaptogenesis (formation of new synaptic connections) through modulation of hepatocyte growth factor. It came out of Washington State University in the early 2010s as a candidate Alzheimer's drug, and the original lab data made it sound almost mythical. Headlines claimed it was "ten million times more potent than BDNF", the brain's own master growth signal. That number is real but specific to a cell-culture assay, and it doesn't translate cleanly to a living brain.
The real appeal is that Dihexa is one of the few compounds where the proposed mechanism is structural, meaning it appears to help the brain build new synaptic connections rather than just temporarily boosting neurotransmitters the way caffeine or modafinil do. The catch is that no human trials have ever been completed on Dihexa itself, the corporate prodrug version (fosgonimeton) failed its Phase 2/3 Alzheimer's trial in 2024, and one of the foundational mechanism papers was retracted in 2025 over data fabrication. Anyone using it is running a personal experiment with limited safety data.
There is a theoretical cancer risk due to promoting cell growth, a major reason why it hasn't cleared FDA trials. if you have any cancer history or risk factors, avoid this completely

Deep-dive

Dihexa is a hexapeptide analog of angiotensin IV, formally N-hexanoic-Tyr-Ile-(6) aminohexanoic amide, also known by its development code PNB-0408. The parent peptide angiotensin IV has cognitive effects in animals but breaks down rapidly in plasma and doesn't cross the blood-brain barrier. Dihexa was engineered around those two problems: an N-hexanoyl group at one end and an aminohexanoic amide at the other resist peptidases, and the increased hydrophobicity allows passive diffusion across the BBB. The original synthesis paper from McCoy and colleagues in 2013 reported a circulating half-life of around 9 to 12 days in rats after IV or IP administration, which is unusually long for a peptide and reflects how aggressively the molecule was stabilised.
Mechanism. Dihexa binds with high affinity to hepatocyte growth factor (HGF), a protein the body produces to drive tissue repair and cellular survival. The Dihexa-HGF complex activates the c-Met receptor on neurons, which triggers downstream PI3K/AKT and MAPK signalling, promotes dendritic spine formation, and ultimately increases synaptic density. Sun and colleagues in 2021 showed Dihexa restored cognitive function in APP/PS1 mice (an Alzheimer's model) specifically through PI3K/AKT activation. This is currently the strongest independent (non-Washington State University) supportive data and has not been challenged.
The "10 million times BDNF" claim. This comes from in vitro synaptogenesis assays where Dihexa was active at picomolar concentrations while BDNF required nanomolar concentrations to produce equivalent dendritic spine growth. The dose-response ratio is genuine in those specific assays. It does not translate to a 10-million-fold effect in a living brain, where pharmacokinetics, tissue distribution, and receptor density all change the picture. Treat the headline number as a marketing line, not a clinical prediction.
Healthy brains may not respond. This is the single most important practical finding from the animal literature. In the WSU studies, Dihexa improved cognition in animals with induced cognitive deficits but had no measurable effect on rats with normal cognition. The HGF/c-Met system seems to be recruited primarily during repair, not during everyday learning. If you have a healthy, well-functioning brain, the realistic expectation is that benefits may be subtle or absent. Dihexa looks much more promising for someone recovering from concussion, dealing with age-related cognitive shift, or experiencing post-illness brain fog than for a 25-year-old optimiser.
The data integrity problem. Anyone considering Dihexa should know the full story. Leen Kawas, a PhD student in the WSU lab that developed Dihexa, co-founded M3 Biotechnology (later renamed Athira Pharma) in 2013 to commercialise the compound. Athira pivoted to fosgonimeton (ATH-1017), a prodrug that converts to Dihexa in the body, and went public in 2020 raising over $200 million. In 2021 an internal Athira investigation found that Kawas had altered images in her doctoral dissertation and at least four research papers published between 2011 and 2014, including foundational papers on Dihexa's mechanism. Kawas resigned. WSU later revoked her PhD. The 2014 Benoist paper that established the HGF/c-Met mechanism was formally retracted in 2025. Athira settled a False Claims Act case with the Department of Justice for $4 million in early 2025 tied to NIH grant applications that cited the manipulated research. The retraction does not mean Dihexa "doesn't work", it means the specific paper that proved the HGF/c-Met mechanism contained fabricated images. The mechanism has been partially rebuilt by independent work like the Sun 2021 paper above.
The human trial that failed. Athira's prodrug fosgonimeton has been the only version of this molecule actually tested in humans. The Phase 2/3 LIFT-AD trial in 315 patients with mild-to-moderate Alzheimer's, reading out in September 2024, missed both its primary cognitive endpoint and its key secondary endpoints. The change in the Global Statistical Test favoured fosgonimeton over placebo by 0.08 points with a p-value of 0.70, meaning the result was not distinguishable from chance. The earlier ACT-AD Phase 2 trial had also missed its primary endpoint. The drug was safe and well-tolerated but did not show clinical benefit. This isn't the result you'd expect from a true 10-million-fold synaptogenesis amplifier.
Women. Almost all preclinical Dihexa research used male rodents, the standard but unhelpful default in neuroscience. There is no female-specific Dihexa data in humans. The mechanism, however, raises sex-specific concerns men don't carry to the same degree. The HGF/c-Met pathway is particularly active in breast tissue. High HGF and c-Met expression in breast tissue is associated with invasive breast cancer and is causally linked to metastasis. HGF/c-Met signalling is especially active in basal-like and triple-negative breast cancer, where it drives migration and treatment resistance. Normal breast epithelial cells suppress HGF expression; cancerous and pre-cancerous cells lose that suppression. A compound that potentiates HGF/c-Met signalling systemically is theoretically pushing in the wrong direction for breast tissue, and women have meaningfully higher baseline breast cancer risk than men do. This doesn't mean Dihexa causes breast cancer, the studies haven't been done, but it means the cancer-risk calculation isn't symmetrical between sexes. Women with any personal or family history of breast cancer should treat Dihexa as contraindicated rather than cautionary, and the same applies to women on hormone therapy where HGF responsiveness in breast tissue is altered. The pathway is also active in placental development and uterine remodelling, which is why pregnancy and breastfeeding are absolute contraindications.
Older adults. The HGF/c-Met system declines with age, and this is part of why Dihexa's preclinical data looked strongest in dementia models rather than healthy young animals. The proposed use case where the cost-benefit aligns best is age-related cognitive shift, post-stroke recovery, or post-concussion symptoms, where the system being targeted is genuinely underactive. Older adults also have higher baseline cancer screening importance, so the monitoring threshold should be lower, not higher.
Limitations of the evidence. To summarise honestly: the human evidence is one failed Phase 2/3 trial of a prodrug. The animal evidence is real and has been independently replicated, but it shows benefit specifically in cognitive-deficit models, not in healthy animals. The mechanism evidence had its keystone paper retracted but has been partly rebuilt. The long-term safety evidence is essentially zero. The cancer-risk concern is theoretical but mechanistically well-grounded. Dihexa is a compound where the practical question isn't "does it work" but "does the potential benefit, in your specific situation, outweigh a poorly-quantified but plausible long-term risk."

Dosage:

  • Oral: 8-25 mg once daily, taken in the morning. Most people land at 15-20 mg. Dihexa is unusual among peptides in being orally active, capsules and tablets work without losing meaningful potency
  • Subcutaneous (alternative): 2-5 mg once daily, also in the morning. No clear advantage over oral for most people, the molecule was designed for oral use
  • Take with a fatty meal. The molecule is hydrophobic and absorption improves significantly with dietary fat. Avoid taking it on a completely empty stomach
  • Cycle: 4-6 weeks on, 2-4 weeks off. The reasoning is to avoid sustained activation of a growth pathway that, pushed continuously, has theoretical cancer-promoting potential. There is no good data on what cycling looks like in humans, this is a community-derived structure that errs on the cautious side
  • Start low. 8-10 mg for the first week to assess tolerance, then titrate up if needed. Most reported side effects are dose-dependent
  • Don't take it in the evening. Insomnia and overstimulation are the most common side effects and are worse later in the day. Morning dosing only
  • Women of reproductive age should not use Dihexa during pregnancy, breastfeeding, or while trying to conceive. The HGF/c-Met pathway is involved in placental development and there is no reproductive safety data
  • Hard contraindications: personal history of cancer, strong family history of breast, lung, gastric, liver, or kidney cancer, known precancerous findings, hormone therapy in women, active hepatitis or chronic liver disease. The pathway Dihexa activates is the one oncology drugs are designed to block
  • Source matters. Dihexa is sold as a research chemical, not a regulated pharmaceutical, and purity varies enormously between vendors. Always require a third-party certificate of analysis with HPLC purity and mass spectrometry confirmation. Cheap, unverified vendors are not worth the risk on a compound with this mechanism

Here's what you can expect:

Dihexa is not a stimulant. There's no acute kick the way you get from coffee or a racetam. The proposed effects are structural, meaning new synapses, and they build gradually. People who report benefit usually notice clearer recall, better word retrieval, and easier learning of new material around weeks two to three, with subjective effect peaking around weeks four to six.
The realistic expectation depends heavily on baseline. If you're using Dihexa to recover from concussion, address age-related cognitive shift, or push through post-illness brain fog, you have a reasonable mechanistic case for benefit. If you're a healthy 30-year-old looking for an edge, expect very little, the animal data suggests the system Dihexa targets isn't recruited during normal cognition.
Onset is slow. Crash on stopping is generally not reported, which is consistent with structural rather than neurochemical effects. Some people report partial persistence of benefits after a cycle ends, again consistent with new synapses being formed rather than transient signalling changes. Whether that translates to long-term human benefit is unknown.

Side effects & risks:

  • Headaches in the first week are the most common side effect, usually mild and transient, often resolving with dose reduction
  • Insomnia and overstimulation if dosed too late in the day. The compound has a long half-life, dosing past noon is a recipe for poor sleep
  • Mild irritability or restlessness, also typically dose-dependent and front-loaded in a cycle
  • GI discomfort with oral dosing, usually mild, helped by taking with food
  • No human safety dataset for Dihexa itself. The closest data is from fosgonimeton, the injectable prodrug, which was generally well-tolerated in Alzheimer's trials at a fixed dose for six months in older patients. That doesn't generalise to multi-year use for cognitive enhancement in healthy adults
  • Liver effects are theoretically possible. HGF was named for its activity on hepatocytes. There's no human evidence of liver toxicity from Dihexa, but baseline and follow-up liver enzymes are sensible
  • Theoretical risk of aberrant tissue remodelling. HGF is involved in wound healing, fibrosis, and angiogenesis. Sustained pathway activation could in principle drive unwanted tissue changes. No human data, but a reason cycling is the default
Sold as a research chemical, not approved for human use by any regulatory agency.