Dihexa: The Complete Guide

StackTrax doesn't recommend Dihexa use. This guide exists because the compound is in circulation and users deserve to know the actual record. Three load-bearing facts.

The foundational mechanism paper has been retracted. Benoist et al. J Pharmacol Exp Ther 2014 (PMID 25187433) — the paper that established dihexa's "K_d ≈ 65 pM HGF binding" claim and the "c-Met knockdown blocks dihexa effect" mechanism — was retracted in April 2025 (PMID 40312093) for image manipulation. The K_d figure that appears across vendor copy and forum posts is a number from a retracted paper. Two additional foundational papers from the same lab carry Expressions of Concern: Benoist 2011 (PMID 21719467, the Nle1-AngIV-analog series from which dihexa was derived) and McCoy 2013 (PMID 23055539, the founding dihexa paper).

The dihexa-derived prodrug failed Phase 2/3. Fosgonimeton (ATH-1017), Athira Pharma's lead asset, was a dihexa-derived prodrug for Alzheimer's disease. It failed primary endpoint in the Phase 2/3 LIFT-AD trial in September 2024. Earlier Phase 1 was published as Hua 2022 (PMID 35180125). The molecule that was supposed to be dihexa's clinical proof-of-concept didn't work in mild-to-moderate Alzheimer's.

The mechanism is direct activation of an oncogene. Dihexa is reported to act by enhancing HGF / c-Met signaling. c-Met is one of the most-studied oncogenic receptor tyrosine kinases in solid-tumor biology — central to invasion and metastasis (Comoglio, Trusolino, Boccaccio, Nat Rev Cancer 2018, PMID 29674709). Vendor framing of dihexa as a "safe nootropic" is methodologically incoherent given the target.

Commercial-conflict context: the WSU lab that developed dihexa (Harding & Wright) co-founded Athira Pharma. Athira's CEO 2014–2021, Dr. Leen Kawas, was a co-author on multiple dihexa papers and resigned in October 2021 after an internal investigation found altered Western blot images. In January 2025, Athira agreed to a $4.07M False Claims Act settlement with DOJ for failing to disclose the misconduct in NIH grant applications 2016–2021.

If after all of this you still want a Dihexa guide, the rest of this page is what the actual published record says — including the parts that aren't retracted.

Dihexa (PNB-0408; N-hexanoic-Tyr-Ile-(6) aminohexanoic amide) is a peptidomimetic small molecule, not a true peptide. It's a stabilized derivative of angiotensin IV (AngIV), an endogenous hexapeptide fragment of the renin-angiotensin system. Some vendor and encyclopedia content describes Dihexa as a "synthetic peptide" or "six-amino-acid peptide" — that's wrong. Dihexa is a capped dipeptidomimetic with only two proteinogenic amino acids (Tyr and Ile) plus a hexanoic-acid cap and aminohexanoic-acid amide tail.

It was designed in the labs of Joseph W. Harding and John W. Wright at Washington State University. The lab spent ~20 years studying AngIV procognitive effects before designing Dihexa as a metabolically stable, orally-bioavailable derivative. The minimum pharmacophore was a Tyr-Ile dipeptide core (Benoist 2011, PMID 21719467 — Expression of Concern).

The "10-million-times-more-potent-than-BDNF" claim circulates widely in nootropic communities. It originates in WSU press materials from ~2012, not from a head-to-head primary-literature comparison. There's no peer-reviewed paper that directly compares Dihexa and BDNF on the same synaptogenesis assay. The number is real in the sense that someone said it; it's not real in the sense of being an experimentally validated potency ratio.

Regulatory status

• FDA: not approved. No NDA, no IND for Dihexa itself publicly listed. Not on the 503A bulks list. Sold gray-market as "research chemical — not for human use." The dihexa-derived prodrug fosgonimeton had an active IND under Athira (NCT03298672 Phase 1, NCT04491006 ACT-AD Phase 2, NCT04488419 LIFT-AD Phase 2/3) — those are studies of fosgonimeton, not dihexa.
• EMA: not approved.
• WADA 2026: not specifically named, but at minimum S0 (non-approved substances), and plausibly S2.4 (HGF mimetic / growth-factor mimetic) by structure-activity argument. The harder framing is S2.4, not S0.

Per the now-retracted Benoist 2014 paper (PMID 25187433), dihexa binds endogenous HGF (hepatocyte growth factor) with high affinity (K_d ≈ 65 pM — a number from the retracted paper). The dihexa–HGF complex sensitizes the c-Met receptor tyrosine kinase to subthreshold HGF concentrations. Activated c-Met then triggers downstream PI3K/AKT and MAPK/ERK cascades, producing dendritic spinogenesis, increased synapsin / VGLUT1 / PSD-95 expression, and behaviorally measured procognitive effects in scopolamine-deficit and aged-rat models.

The K_d = 65 pM HGF-binding number, the "HGF antagonist blocks Dihexa effect" experiment, and the c-Met knockdown experiment all originate in the retracted Benoist 2014 paper. Any source that uncritically reproduces "K_d = 65 pM" is reproducing a number from a retracted paper.

Independent third-party preclinical confirmation comes from Sun et al. 2021 (PMID 34827486) in Brain Sci — APP/PS1 mouse model (Alzheimer's), independent group at Nanjing — confirming a procognitive effect of dihexa via the PI3K/AKT axis. Important caveat: this paper doesn't independently re-prove the HGF-binding affinity or the c-Met-knockdown experiment. So the headline mechanism story rests on a retracted paper plus a vendor-citation chain plus one independent procognitive-effect confirmation.

The c-Met oncology context matters: Comoglio, Trusolino, Boccaccio. Nat Rev Cancer 2018 (PMID 29674709) is the canonical review of MET as an oncogene. c-Met activation drives invasion, metastasis, and tumor survival across multiple solid-tumor lineages (lung adenocarcinoma, gastric, hepatocellular, glioblastoma, renal, ovarian, breast). MET-amplified tumors are aggressive; c-Met inhibitors (capmatinib, tepotinib, crizotinib in part) are FDA-approved oncology drugs. Dihexa's proposed mechanism is the opposite of what the oncology field is trying to do.

Fosgonimeton (ATH-1017) is the dihexa-derived prodrug developed by Athira Pharma. It's a different molecule from dihexa — designed for in vivo conversion — but its development was the closest thing to a clinical proof-of-concept for the HGF/c-Met-positive-modulator class.

The trial program: Phase 1 NCT03298672 (safety, tolerability, PK in healthy volunteers and Alzheimer's patients; published as Hua 2022 PMID 35180125, generally safe / well-tolerated); Phase 2 ACT-AD NCT04491006 (mild-to-moderate Alzheimer's); Phase 2/3 LIFT-AD NCT04488419 — failed primary endpoint September 2024, no peer-reviewed publication of the failed result yet. Reda et al. 2024 (PMID 38599894, Neurotherapeutics) characterized fosgonimeton preclinically.

The LIFT-AD failure means the most well-funded, most-rigorous test of the dihexa/HGF/c-Met-positive-modulator hypothesis in humans didn't succeed in mild-to-moderate Alzheimer's — the indication that motivated the entire program. This is the most important calibration point for anyone considering Dihexa for cognitive benefit.

Claim	Source	Status
Procognitive effect in scopolamine-deficit rats	McCoy 2013 PMID 23055539	Expression of Concern Sept 2021
Kd ≈ 65 pM HGF binding	Benoist 2014 PMID 25187433	RETRACTED April 2025
HGF antagonist + c-Met knockdown both block Dihexa effect	Benoist 2014	RETRACTED
APP/PS1 procognitive effect via PI3K/AKT	Sun 2021 PMID 34827486	Independent third-party; preclinical
"10-million-times more potent than BDNF"	WSU press release ~2012	Not anchored in primary literature
Human cognitive benefit (via fosgonimeton)	LIFT-AD Phase 2/3, Sept 2024	FAILED primary endpoint
Human PK / safety of dihexa itself	None	No human study of dihexa exists
Long-term safety	None	No long-term study at all

Direct activation of the HGF / c-Met pathway is mechanistically aligned with multiple oncology phenotypes. c-Met overexpression is a hallmark of many solid tumors (lung adenocarcinoma, gastric, hepatocellular, glioblastoma, renal, ovarian, breast). HGF-driven invasion and metastasis is a well-characterized solid-tumor mechanism. MET amplification defines an aggressive subset of NSCLC; c-Met inhibitors (capmatinib, tepotinib) are FDA-approved targeted oncology drugs. The pharmaceutical industry spends substantial resources blocking c-Met. Dihexa's proposed mechanism is the opposite.

For an AD target population (older adults, baseline cancer risk elevated), this concern is particularly load-bearing. Even if dihexa worked as advertised, the long-term cancer-latency question would dominate. The Sun 2021 APP/PS1 mouse study (PMID 34827486) was 14 weeks — below tumorigenesis timescales. No long-term carcinogenicity study of Dihexa has been published.

StackTrax doesn't endorse Dihexa use. The dosing below is community / forum convention. There's no validated human protocol for Dihexa itself.

Parameter	Common Range
Dose	8–45 mg/day oral
Cycle length	2–8 weeks (community); no validated cycle structure
Route	Oral (designed for oral bioavailability and BBB penetration); SC or IM very rare

If used despite the warnings: shortest possible exposure (cancer-axis risk has long latency); don't use during cancer screening lapses; don't stack with other growth-factor or IGF-axis modulators; recognize the asymmetry that short-term subjective effect is benign while long-latency oncology risk is not.

Acute community reports: headache (most commonly), fatigue or brain fog (paradoxical, given the procognitive framing), mild GI effects, mood changes (mixed reports). None of these is anchored in a human safety study — there's no human PK / safety paper on dihexa itself.

Theoretical and class concerns: cancer (the dominant concern — c-Met activation is a solid-tumor oncogene mechanism, and no long-term carcinogenicity study has been done); tissue overgrowth or fibrosis (HGF/c-Met drives tissue remodeling; chronic activation could plausibly produce off-target tissue effects); cardiovascular (c-Met has cardiovascular roles; chronic agonism uncharacterized); off-target receptor effects (the dihexa molecule has not been profiled against a broad receptor panel).

• Active or prior cancer of any kind — absolute. c-Met activation is mechanistically aligned with tumor invasion / metastasis.
• Family history of cancer — relative caution.
• Lapsed cancer screening — don't start until current.
• Pregnancy, breastfeeding, planning pregnancy — absolute.
• Concurrent IGF-axis modulators / HGH / SARMs — uncharacterized stacking; theoretical additive growth-factor load.
• Athletes subject to anti-doping testing — don't use. WADA S0 minimum, plausibly S2.4 (HGF mimetic).
• Anyone unwilling to accept long-latency oncology risk for short-term cognitive benefit that the published clinical program (fosgonimeton LIFT-AD) failed to demonstrate in humans.

Fosgonimeton (ATH-1017) — the dihexa-derived prodrug. Different molecule. Failed Phase 2/3 LIFT-AD in September 2024.

Angiotensin IV (AngIV) — the endogenous hexapeptide that dihexa was derived from. Less stable; not in research-chemical commerce.

Cerebrolysin — porcine brain extract, multi-component. Different family entirely.

Semax — ACTH(4–7) analog nootropic; different mechanism (BDNF / NGF upregulation, not c-Met).

NSI-189 — small-molecule neurogenesis enhancer; different mechanism.

Capmatinib / tepotinib — FDA-approved c-Met inhibitors (oncology). They do the opposite of what dihexa is reported to do. Their existence as approved cancer drugs is the calibration for why c-Met activation is a serious concern.