A 24-amino-acid mitochondrial-derived peptide encoded by the 16S rRNA mitochondrial gene — first described in 2001 by the Hashimoto / Nishimoto group at Keio from a cDNA library of an Alzheimer brain region that was unusually spared from disease.
Humanin is a 24-amino-acid mitochondrial-derived peptide (MDP) — sequence MAPRGFSCLLLLTSEIDLPVKRRA — encoded within the 16S rRNA region of the mitochondrial genome. It was the first member of the MDP family identified.
The discovery paper is Hashimoto et al. PNAS 2001 (PMID 11371646), "A rescue factor abolishing neuronal cell death by a wide spectrum of familial Alzheimer's disease genes and Abeta." It came out of a functional cDNA-library screen for clones that protected neuronal cells from familial-AD-gene-induced and Aβ-induced death; the source library was the occipital lobe of an AD patient — a region relatively spared from the disease. (Some references date humanin's discovery to 2003. That's the year the IGFBP-3 binding partner was identified, Ikonen 2003 PMID 14561895. The peptide itself was first described in 2001.)
The discovery lab was Ikuo Nishimoto's group at Keio University in Tokyo. Nishimoto died in 2003; subsequent Japanese work continued under Yuichi Hashimoto and Masaaki Matsuoka. The major Western lab is Pinchas Cohen (now at USC Leonard Davis School of Gerontology), who later led the discovery of MOTS-c in 2015 and the broader MDP family (SHLPs and others). Humanin is the founding member of that family — MOTS-c, SHLP1–6, and other MDPs were discovered later on humanin's discovery template.
Humanin acts through multiple intersecting cellular survival pathways. The mechanism story is unusually well-mapped for a research peptide because two decades of work have come from both the Japanese discovery group and the Cohen lab.
The most-cited mechanism is anti-apoptotic action via Bax inhibition. Humanin binds Bax (a pro-apoptotic Bcl-2-family protein) and prevents its translocation from cytosol to mitochondria, blocking the mitochondrial apoptosis pathway. Source: Guo et al. Nature 2003 (PMID 12732850).
A second mechanism is IGFBP-3 partnership. Humanin binds insulin-like growth factor binding protein-3 (IGFBP-3) and modulates IGFBP-3-dependent cell survival. This was the bridge to Cohen-lab humanin work (Ikonen 2003, PMID 14561895). Third, humanin activates a heterotrimeric cytokine-receptor complex composed of CNTFR, WSX-1, and gp130, driving intracellular STAT3 phosphorylation (Hashimoto 2009, PMID 19386761). Downstream of that receptor humanin also engages ERK1/2 and AKT, with signalling that differs by age in the hippocampus (Kim 2016, PMID 27384491). Fourth, hypothalamic / intracerebroventricular humanin in rodents improves peripheral insulin sensitivity and glucose homeostasis (Muzumdar 2009, PMID 19623253).
One claim worth ignoring: the "TRIM5α inhibition" mechanism that shows up in some encyclopedia entries and vendor copy has zero PubMed support. Skip it.
Most preclinical humanin literature uses analogs, not the wild-type peptide. Vendor copy that says "humanin protects against X in mouse models" is technically correct but leans on analog data more than wild-type data.
| Analog | Modification | Notes |
|---|---|---|
| HNG (S14G-Humanin) | Ser → Gly at position 14 | Most-studied. ~1000-fold more potent than wild-type in some neuroprotection assays. Workhorse of Alzheimer's mouse studies (Tajima 2005 PMID 15678515; Zhang 2012 PMID 21993310). |
| HNGF6A | S14G + F6A | Cohen-lab work; further-stabilized analog with improved metabolic activity. |
| Colivelin (AGA-(C8R)-HNG17) | Hybrid of HNG and ADNF | Enhanced potency and stability (PMID 26964005). |
| P3S (longevity variant) | Pro³ → Ser³ | Enriched in centenarians, especially APOE4 carriers (PMID 38520065). |
Build your protocol, log every dose, monitor your body's response, and get reminders so you never miss a dose.
Start Tracking FreeNo trial has ever administered humanin to humans. Clinical trials indexed on PubMed do involve humanin, but they measure it as an outcome — circulating humanin rises after acute endurance exercise and after resistance training, for instance — rather than giving it as a drug. So all human evidence here is observational or biomarker work. Worth knowing for context: the sister MDP MOTS-c did reach humans — CohBar's analog CB4211 completed a randomised, placebo-controlled Phase 1a/1b in 88 subjects (NCT03998514, completed April 2021), reported as the first mitochondrial-derived-peptide candidate to enter clinical testing. No results were ever posted to the registry or published. Humanin itself has had no equivalent.
| Domain | Evidence | PMIDs |
|---|---|---|
| Neuroprotection (Alzheimer's) | Wild-type humanin rescues neuronal cells in vitro | 11371646, 11717357, 11327724 |
| HNG cognition (mouse) | S14G improves memory in Aβ-injection and APPswe/PS1dE9 models | 15678515, 21993310 |
| Anti-apoptosis / Bax | Bax binding blocks mitochondrial translocation | 12732850 |
| IGFBP-3 partnership | Regulates cell survival | 14561895 |
| Cytokine-receptor signaling | CNTFR/WSX-1/gp130 → STAT3 | 19386761 |
| Insulin action (rodent) | ICV humanin improves peripheral insulin sensitivity | 19623253 |
| Cardio-protection (rodent) | Reduces myocardial IR injury | 27434747, 34896254 (review) |
| Coronary endothelial function (human observational) | Higher circulating humanin = preserved function (n=102) | 23220334 |
| Cognitive aging | Mouse intervention + human observational | 30242290 |
| Aging-related MDP decline | Endogenous humanin and SHLPs drop with age | 27070352 |
| Retinal protection (in vitro) | RPE protection from oxidant injury | 32768357 |
| T2DM / AD biomarker | Plasma humanin differs by disease state | 33131010 |
| P3S longevity variant | Enriched in APOE4-carrying centenarians | 38520065 |
| Lifespan (worm) / healthspan (mouse) | C. elegans overexpressing humanin lived longer — 19.0 vs 17.7 days mean, about +7%, and the gain disappeared in daf-16/FOXO mutants. In middle-aged mice, twice-weekly HNG did not extend lifespan; it reduced visceral fat and improved metabolic markers without changing food intake. Humanin declines with age in mice, monkeys and humans, but stays flat across two decades in the naked mole-rat. Offspring of centenarians (n=18) had significantly higher circulating humanin than age-matched controls (n=19) | 32575074 |
| Cardioprotection — large animal | HNG 2 mg/kg cut infarct size in a pig model, the closest thing to human-relevant cardiac data here. The effect vanished when ischemic time was lengthened, which the authors flag as the open problem for translation | 32760857 |
| Cardioprotection — dose-response (rodent) | Dose-dependent infarct reduction, maximal at 2 mg/kg IP (mice); in rats, HNG given during ischemia worked at 252 mcg/kg but not lower doses | 20651283, 28726291 |
| Beta-cell insulin secretion | A potent humanin analog increased glucose-stimulated insulin secretion from pancreatic islets | 23995290 |
| Muscle protection (human cells) | Humanin partially preserved myotube area against dexamethasone-induced atrophy — but MOTS-c outperformed it in the same experiment, fully preserving area and fusion | 41732124 |
| Inflammation resolution | Human macrophages make humanin during efferocytosis, and that production promotes resolution of inflammation | 40877234 |
| Tumor promotion (not a benefit) | Listed here so it is not read past: in triple-negative breast cancer models, exogenous humanin protected tumor cells from apoptosis, accelerated tumor growth and spontaneous lung metastasis, and impaired chemotherapy's antitumor and antimetastatic effect. Silencing humanin did the reverse. See Side Effects | 32444831 |
Not medical advice. These figures describe what is reported in the literature and what practitioners and communities do — not a recommendation, and for most compounds here no human dose-finding study exists. Talk to a qualified healthcare provider.
No human-trial-derived dose exists. Same caveat as MOTS-c — anything specific in vendor or forum copy is community practice, not clinically validated.
| Parameter | Common Range |
|---|---|
| Dose per injection | 500–1000 mcg SC |
| Frequency | Daily, EOD, or twice weekly |
| Cycle length | 4–8 weeks on / similar off |
| Route | SC standard; intranasal occasionally reported |
Humanin's circulating half-life is roughly 30 minutes (Yen 2020, PMID 32575074). Stabilized analogs — HNG, HNGF6A, colivelin — were designed partly to overcome that clearance. The short half-life is not a footnote: the Yen authors offer it as one explanation for why twice-weekly HNG improved healthspan markers in middle-aged mice yet did not extend their lifespan.
Sources disagree on cycle length. Community write-ups variously describe 10–15 day cycles repeated several times a year, and 4–8 week cycles with similar time off. Neither is anchored to a dose-finding study, so treat the spread as a measure of how little is known rather than a menu.
For scale, published animal work used routes and doses that do not convert to human subcutaneous dosing: 0.16 mcg/kg/min by intracerebroventricular infusion (rat), 2 mg/kg intraperitoneal (mouse cardioprotection), 84–252 mcg/kg intravenous (rat), 0.4 mg/kg/day for 16 weeks (mouse atherosclerosis). None of these is a human dose, and no human dose-finding study exists.
Pre-filled with a typical Humanin setup. Edit any field — the draw updates live.
Insulin syringe — 100 units = 1 mL
Free account. Saves your reconstitution + schedules doses + tracks every vial.
Dosing cheat sheet, reconstitution reference, and cycle planning — delivered to your inbox.
Not medical advice. These figures describe what is reported in the literature and what practitioners and communities do — not a recommendation, and for most compounds here no human dose-finding study exists. Talk to a qualified healthcare provider.
For a 5 mg vial in 2 mL bacteriostatic water: 2.5 mg/mL, so 1 unit on a U-100 syringe = 25 mcg. A 500 mcg dose is 20 units; a 1000 mcg dose is 40 units.
Lyophilized vials are stable at room temperature short-term; refrigerate for longer. Reconstituted vials should be refrigerated and used within 28 days. For the full protocol, see the Bacteriostatic Water guide.
No trial has administered humanin to humans, so there is no clinical safety dataset. Published animal work reports no significant toxicity, including at 2 mg/kg HNG in mice and pigs.
Guides often describe humanin's cancer risk as a mechanistic guess from its anti-apoptotic action. It is more than that. In Scientific Reports (Moreno Ayala 2020, PMID 32444831), exogenous humanin given to mice bearing triple-negative breast cancer reduced the tumors' apoptotic rate, accelerated tumor growth, increased spontaneous lung metastasis, and blunted chemotherapy's antitumor and antimetastatic effect. Silencing humanin had the opposite effect — reduced cancer-cell viability and better chemosensitivity. The same paper found humanin and its receptors expressed in human breast-cancer specimens (TCGA) and humanin upregulated in TNBC biopsies versus healthy tissue. The authors' own conclusion is that caution is warranted when giving exogenous humanin for degenerative disease. This is one tumor type in one model, and other work reports humanin shielding healthy tissue from chemotherapy without shielding tumors — but it is a real result pointing the wrong way, and it is unresolved.
A growth and reproduction tradeoff. In the Yen 2020 work, both humanin-overexpressing worms and transgenic mice showed reduced body size, reduced body fat, and reduced brood size — the authors read this as the familiar reproduction-for-longevity tradeoff seen in other longevity models, and link it to lowered IGF-I. Sustained elevation is not a free lunch.
Other mechanism-derived concerns. Humanin binds IGFBP-3, so chronic dosing could plausibly alter IGF-1 bioavailability. Humanin is endogenous, so theoretical immunogenicity is low — though no immunogenicity data has been published.
A search of the FDA's FAERS database returns exactly one report naming humanin: a 56-year-old, September 2024, coded serious, reconstituted with distilled water rather than bacteriostatic water. The reported reactions were dizziness, weakness, blurred vision, facial paresis, paralysis, respiratory failure and botulism. FAERS reports are self-submitted, unverified, and establish no causality, and that cluster is the textbook presentation of a contaminated injection rather than anything in humanin's pharmacology. Read it as a warning about sterile technique and vial provenance — the reason the bacteriostatic water protocol specifies a preserved diluent — not as evidence about the peptide. No EudraVigilance or Yellow Card review was performed for this page.
Community reports describe mild injection-site irritation and occasional GI discomfort as the most common issues. The user base is very small and none of it is anchored in a peer-reviewed safety paper.
MOTS-c — sister mitochondrial-derived peptide. Different sequence, different gene origin (12S rRNA vs 16S), different mechanism (AMPK activation vs Bax/IGFBP-3/STAT3). Both Cohen-lab MDP family but mechanistically distinct.
SHLP1–6 — small humanin-like peptides; further Cohen-lab MDPs. Less studied; not commonly available as research chemicals.
Cerebrolysin — porcine brain extract, multi-component. Different family entirely.
HNG vs wild-type humanin — if the goal is the published preclinical effect, HNG (S14G) is the molecule the strongest mouse studies actually used.
Humanin is a research peptide not approved by the FDA for human use. It is sold only as a research chemical, and StackTrax does not endorse or facilitate personal use.
Quality varies enormously among research-chemical suppliers. At minimum, look for:
Lyvn does not carry Humanin. They are the vendor we recommend elsewhere on this site, and they publish the full independent panel per batch — which makes their catalogue a useful yardstick for the criteria above even on a compound they do not stock. We are not recommending a specific supplier for Humanin itself, because we have no basis to.
For research use only. Not for human consumption. 21+.
Build your protocol, log every dose, monitor your body's response, and get reminders so you never miss a dose.
Start Tracking FreeNo. Humanin is not approved by the FDA or EMA. There is no NDA on file and no IND. It is sold as a research chemical only. Humanin is not explicitly named on the WADA 2026 list, but is likely covered by the S2 catch-all (peptide hormones, growth factors, related substances and mimetics) since humanin signals via cytokine-receptor complexes.
Humanin is a 24-amino-acid mitochondrial-derived peptide (MDP) encoded within the 16S rRNA region of the mitochondrial genome, rather than the nuclear genome. It was the founding member of the MDP family identified in 2001. Sister MDPs include MOTS-c (12S rRNA, AMPK activation) and SHLP1-6 (small humanin-like peptides). The Cohen lab at USC discovered MOTS-c and SHLPs later, building on humanin's discovery template.
Humanin was first described in Hashimoto et al. PNAS 2001 (PMID 11371646) by Ikuo Nishimoto's group at Keio University in Tokyo. It came out of a functional cDNA-library screen for clones that protected neuronal cells from familial Alzheimer's-gene and amyloid-beta-induced death. The source library was the occipital lobe of an AD patient, a region relatively spared from the disease. Some references date the discovery to 2003, but that is when the IGFBP-3 binding partner was identified (Ikonen 2003), not the peptide itself.
No trial has administered humanin to humans. Some indexed clinical trials do involve humanin, but they measure it as an outcome — circulating humanin rises after endurance and resistance exercise — rather than giving it as a drug. The human evidence is therefore observational: a coronary endothelial function study (n=102, PMID 23220334) where higher circulating humanin correlated with preserved function, plasma humanin differences by T2DM/AD disease state (PMID 33131010), the P3S longevity variant enriched in APOE4-carrying centenarians (PMID 38520065), and higher circulating humanin in offspring of centenarians (n=18 vs 19 controls, PMID 32575074). Efficacy data is all in vitro or in animals. The sister peptide MOTS-c did reach humans as CohBar's analog CB4211 (Phase 1a/1b, 88 subjects, NCT03998514, completed 2021) — but results were never posted or published, and humanin itself has had no equivalent trial.
Humanin acts through multiple intersecting pathways. The most-cited is anti-apoptotic action via Bax inhibition (Guo 2003, PMID 12732850), where humanin binds Bax and prevents its translocation to mitochondria. It also binds IGFBP-3 to modulate IGFBP-3-dependent cell survival, activates a heterotrimeric cytokine-receptor complex (CNTFR/WSX-1/gp130) driving STAT3 phosphorylation, and in rodents improves peripheral insulin sensitivity. The TRIM5-alpha mechanism that shows up in some vendor copy has zero PubMed support.
HNG (S14G-Humanin) is a humanin analog with a Ser-to-Gly substitution at position 14. It is roughly 1000-fold more potent than wild-type humanin in some neuroprotection assays and is the workhorse of Alzheimer's mouse studies (Tajima 2005, Zhang 2012). Most preclinical humanin literature uses analogs like HNG, HNGF6A, or colivelin rather than wild-type, partly because wild-type humanin has a minutes-scale half-life in rodents. Vendor copy that says "humanin protects against X in mouse models" is technically correct but leans on analog data more than wild-type data.
No human-trial-derived dose exists. Community practice ranges are 500 to 1000 mcg SC per injection, dosed daily, every other day, or twice weekly, in 4 to 8 week cycles with similar off time. SC is standard; intranasal is occasionally reported. For a 5 mg vial in 2 mL bacteriostatic water, the concentration is 2.5 mg/mL, so 1 unit on a U-100 syringe equals 25 mcg (500 mcg = 20 units; 1000 mcg = 40 units).
There is no clinical safety dataset, because no trial has given humanin to humans. The most important published signal is not a side effect in the usual sense: in a triple-negative breast cancer mouse model, exogenous humanin accelerated tumor growth and spontaneous lung metastasis and impaired chemotherapy (PMID 32444831), which makes active cancer, prior cancer, and current chemotherapy firm contraindications rather than theoretical ones. Animal work also shows a growth and reproduction tradeoff with sustained elevation (PMID 32575074). FAERS holds exactly one report naming humanin — a 2024 serious case reconstituted with distilled water whose reaction pattern reads as botulism from a contaminated injection, not humanin pharmacology; it is a lesson about sterile technique, not about the peptide. Community reports describe mild injection-site irritation and occasional GI discomfort. Avoid in pregnancy or breastfeeding (no data), use caution in autoimmune conditions, and monitor glucose if on insulin or diabetes medication.
Disclaimer: This guide is for educational and informational purposes only and is not intended as medical advice, diagnosis, or treatment. The compounds discussed are not FDA approved for human use. Always consult a qualified healthcare provider before starting any new supplement or peptide protocol. StackTrax does not sell peptides or supplements directly — purchase links go to third-party vendors. StackTrax is not responsible for the products, quality, or business practices of any third-party vendor.
© 2026 StackTrax, LLC. All rights reserved.
Privacy · Terms · Do Not Sell or Share My Personal Information
StackTrax guides cover peptides and compounds that are not FDA-approved for the uses discussed. The dosing, reconstitution, and safety information is compiled from published research and community protocols for educational purposes only.
Before using any compound mentioned here, consult a qualified healthcare provider. StackTrax does not sell, prescribe, or recommend these substances for personal use.