Humanin

01 · What it is

Humanin’s evidence base is twenty-five years deep, published in Nature and PNAS, and built almost entirely on a synthetic mutant that is not the molecule sold under the name.

Humanin is a 24-residue peptide encoded by a short open reading frame within the mitochondrial 16S ribosomal RNA gene. It was found in 2001 by a group at Keio University in Tokyo, using functional expression screening of a cDNA library made from the occipital lobe of an Alzheimer’s brain — a region that had been comparatively spared. The screen asked which cDNA could rescue neuronal cells from death caused by familial Alzheimer’s disease mutations, and one clone did: it abolished death caused by a wide spectrum of familial Alzheimer’s genes and by amyloid-β, without affecting other forms of cell death ². Two companion papers characterised the effect against Swedish mutant amyloid precursor protein ¹ and across a battery of Alzheimer’s-relevant insults ³, and a fourth established that the peptide is produced in vivo, confirming the 24-residue sequence and noting that the open reading frame is 99% identical to a stretch of mitochondrial DNA ⁴.

It was the first mitochondrial-derived peptide. Its existence is what prompted the search that produced MOTS-c [see that guide].

The molecular biology went further than most of this field ever gets. Humanin binds Bax and prevents its translocation from cytosol to mitochondria, blocking the committed step of intrinsic apoptosis — reported in Nature in 2003 ⁷. Independently, a yeast two-hybrid screen for IGFBP-3-interacting proteins pulled out humanin, establishing a second binding partner and a link to the insulin-like growth factor system ⁹. Structure-function work mapped the residues that matter ⁶, showed the peptide must dimerise to work and that dimerisation-blocking mutants act as antagonists ¹¹, and identified TRIM11 as a protein that binds and destabilises it ⁵.

And then almost everything therapeutic was done with something else. Two serine residues, at positions 7 and 14, regulate the rescue function, and substituting glycine for the serine at position 14 produces S14G-humanin — written throughout the literature as HNG, and described in paper after paper as a potent or highly potent analogue ⁸. Search the last decade of humanin research and the overwhelming majority of the intervention studies use HNG: platelets ¹⁸, hair follicles ¹⁹, cochlear hair cells ²⁸, cardiac tissue ²⁰, epidermal stem cells ¹⁴, sperm ¹⁵, germ cells ¹⁶, gouty joints ²², the endometrium ²⁶, the ovary ²⁹, the retina ^12,23,27, neurotransmitter release ²⁴.

That is not a trivial distinction. HNG is a designed mutant with a single substitution at a residue the structure-function work identified as functionally decisive ^6,8. A guide that reported HNG’s results under the heading “humanin” would be describing a different molecule, and this one does not.

No human has been given either. The human record consists of measurements: humanin in seminal plasma correlated with sperm quality ¹⁷, plasma humanin in patients with age-related macular degeneration ²³, humanin and MOTS-c as candidate markers in a prostate cohort of 375 men ³⁰, humanin in neuron-derived extracellular vesicles as a secondary readout inside an exercise trial in Alzheimer’s disease ²⁵, and humanin and MOTS-c rising in human skeletal muscle and plasma after acute high-intensity exercise ²¹.

One review argues the opposite of the received account and belongs in the first 200 words: humanin plasma levels are inversely correlated with growth hormone and IGF-1 expression, which the authors suggest may promote accelerated ageing, and they title the paper “Mitochondrial-Derived Peptides Exacerbate Senescence” ¹³.

02 · Evidence at a glance

Evidence grade

Preclinical — no human administration study of humanin or of HNG is indexed

Structure

24 residues, encoded by a short open reading frame within the mitochondrial 16S rRNA gene ^2,4

Discovery

2001, by functional expression screening of cDNA from a spared region of an Alzheimer’s brain ²

Status in the field

The first mitochondrial-derived peptide identified

What most intervention studies actually use

S14G-humanin (HNG), a synthetic single-substitution mutant ⁸

Binding partners identified

Bax ⁷; IGFBP-3 ⁹; TRIM11, which destabilises it ⁵

Structural requirement

Must dimerise; dimerisation-blocking mutants are antagonists ¹¹

Human administration studies

None

Human subjects given humanin or HNG

0

Human observational data

Seminal plasma ¹⁷; AMD plasma ²³; prostate cohort, n=375 ³⁰; exercise ^21,25

A contrarian finding

Plasma humanin inversely correlated with GH and IGF-1; argued as potentially pro-ageing ¹³

Approval status

None, any jurisdiction

03 · Mechanism of action

Bax, and the committed step of apoptosis

The best-established molecular action is inhibition of Bax. Bax is a pro-apoptotic Bcl-2 family protein that sits inactive in the cytosol until a death signal converts it to an active conformation, whereupon it translocates to mitochondria and permeabilises the outer membrane. That permeabilisation is the point of no return in intrinsic apoptosis.

Humanin interacts with Bax and prevents its translocation from cytosol to mitochondria ⁷. The finding was published in Nature, and it is the reason humanin appears in cell-death literature far outside neuroscience.

A peptide encoded in the mitochondrial genome that blocks a protein from reaching the mitochondrion is an elegant piece of biology, and it is the single most robust thing in this file.

IGFBP-3, and a link the field has not resolved

A yeast two-hybrid screen designed to find IGFBP-3-interacting proteins returned humanin, and the interaction regulates cell survival and apoptosis ⁹. IGFBP-3 has both IGF-dependent and IGF-independent pro-apoptotic activity, so a peptide that binds it sits at a junction between the mitochondrial death machinery and the growth factor axis.

This connection is also where the contrarian reading enters. If humanin rises when GH and IGF-1 fall ¹³, the peptide may be a marker of a low-growth-signalling state rather than a cause of anything beneficial — and low growth signalling has both pro-longevity and pro-frailty readings depending on context.

The receptor question

Early work found specific binding sites for humanin on neuronal cells, with rescue blocked by a tyrosine kinase inhibitor but not by a PI3K inhibitor, implying action through a cell-surface receptor linked to tyrosine kinases [as summarised in 10]. A 2004 review describes the peptide acting through multiple mechanisms, including a putative cell-surface receptor and direct interaction with intracellular partners ¹⁰.

The extracellular receptor account has never reached the certainty of the Bax result, and this guide does not present it as settled.

Structure: two serines, dimerisation, and the origin of HNG

The 24-residue peptide contains serines at positions 7 and 14, and both regulate its rescue function; isomerisation and dimerisation potentiate activity ⁸. Mutational scanning identified the residues essential for neuroprotection ⁶, and mutants that interfere with dimerisation inhibit neuroprotection by wild-type humanin — that is, they behave as antagonists ¹¹.

S14G-humanin, HNG, substitutes glycine at position 14. The literature describes it consistently as a potent or highly potent variant ^15,20,22, and it is the form used in nearly all of the therapeutic experiments in Section 4.

Degradation

TRIM11, a tripartite motif protein, binds humanin and destabilises it ⁵. That is one of very few papers in this project describing how a peptide of this kind is cleared at the protein level, and it implies that endogenous humanin concentration is actively regulated rather than passively determined.

04 · Key research findings

Neuroprotection — the founding result. Humanin abolished neuronal cell death caused by multiple familial Alzheimer’s disease genes and by amyloid-β, without effect on other death stimuli ², characterised in detail against V642I and Swedish APP mutants and presenilin 1 and 2 mutants ^1,3. In vivo production of the peptide was subsequently demonstrated ⁴. In a 2022 study, intraperitoneal HNG raised hippocampal acetylcholine in mice, did not affect physical activity, and modestly improved object memory ²⁴.

Specificity — rescuing one class of insult and not others — is a stronger result than broad cytoprotection, and it is what made the original papers persuasive. Twenty-five years later, no clinical programme has followed.

Retina. Humanin G protected transmitochondrial ARPE-19 cybrids carrying AMD patient mitochondria from mitochondrial and cellular damage ¹². In later work, inflammation-associated marker proteins were elevated in AMD and HNG treatment reduced them, with humanin protein levels also measured in the plasma of AMD patients ²³. A third study examined HNG’s effect on angiogenesis and neurodegeneration markers in AMD retinal pigment epithelium ²⁷.

Three studies, one group, one cell system, plus a plasma measurement. Consistent within itself.

Reproductive tissue. HNG supplementation of freezing media protected human sperm from freeze-thaw damage, with protection increasing with concentration ¹⁶. HNG prevented temozolomide-induced male germ cell apoptosis in SCID mice bearing human medulloblastoma, following earlier work by the same group in a metastatic melanoma allograft in which HNG protected germ cells but not cancer cells ¹⁶. HNG alleviated intrauterine adhesions in a rat model by inhibiting endometrial epithelial cell ferroptosis ²⁶, and ameliorated ovarian dysfunction in a cyclophosphamide-induced premature ovarian insufficiency mouse model ²⁹. Separately, humanin was detected in human seminal plasma and spermatozoa for the first time, with levels related to sperm quality ¹⁷.

The claim that HNG protects normal germ cells from chemotherapy without protecting the tumour is the most therapeutically interesting result in this file, and it rests on two studies from one group ¹⁶.

Cardiovascular and platelets. S14G-humanin protected against streptozotocin-induced cardiac dysfunction in a diabetic cardiomyopathy model ²⁰. HNG inhibited platelet activation and thrombus formation by stabilising platelet microtubules ¹⁸.

An antithrombotic effect is a safety-relevant finding as much as a therapeutic one, and it has not been examined in any species that bleeds under observation.

Inflammation and injury. HNG protected against monosodium urate crystal-induced gouty arthritis, acting on oxidative stress and NLRP3 inflammasome-mediated inflammation ²². HNG and SHLP3 — another mitochondrial-derived peptide — reduced gentamicin-induced hair cell loss in organ of Corti explants ²⁸. HNG was cytoprotective in ultraviolet-B-induced epidermal stem cell injury ¹⁴. HNG promoted hair growth by inhibiting the anagen-to-catagen transition ¹⁹. In 2025, human efferocytic macrophages were shown to produce humanin, and the peptide promoted resolution of inflammation ³¹.

The macrophage finding is the most important recent addition, because it identifies an endogenous human cell type that makes the peptide and a physiological role for it — which is the kind of result that would justify a development programme.

Exercise and human measurement. Acute high-intensity exercise increases humanin and MOTS-c concentrations in human skeletal muscle and plasma ²¹. In the ADEX randomised controlled trial of exercise in mild to moderate Alzheimer’s disease, humanin was measured alongside proBDNF and BDNF in plasma neuron-derived extracellular vesicles ²⁵.

The ADEX study is the closest humanin comes to appearing in a randomised trial, and the intervention was exercise. Humanin was an outcome measure, not a treatment.

Cancer biomarker work. In 375 men suspected of prostate cancer, plasma and exosomal levels of humanin, MOTS-c, GAS5, miR-21 and miR-103 were measured for risk stratification ³⁰.

A cohort of that size, with humanin as a candidate diagnostic marker, is the largest human dataset involving this peptide — and it involves measuring it, not giving it.

The dissenting review. A 2018 review argues that mitochondrial-derived peptides exacerbate senescence, noting that humanin plasma levels are inversely correlated with growth hormone and IGF-1 expression, which may promote accelerated ageing, while acknowledging that both humanin and MOTS-c improve insulin sensitivity in mouse models of type 2 diabetes ¹³.

A published argument that the compound’s associations point the opposite way to the popular framing. It is a review rather than an experiment, and it deserves to be read alongside the promotional account rather than after it.

05 · Evidence overview

06 · Safety profile

Animal data. No formal toxicology exists. The rodent studies report efficacy endpoints over days to weeks and do not systematically record adverse events ^{16,19,20,22,26,29}. One study noted explicitly that intraperitoneal HNG did not affect the physical activity of treated mice ²⁴, which is the closest thing to a tolerability observation in the file.

Human data. No human has been given humanin or HNG in any indexed study. Every human dataset is a measurement of endogenous peptide ^{17,21,23,25,30}, and none carries information about administering it.

What is genuinely unknown. Pharmacokinetics in any species — a 24-residue unmodified peptide, actively destabilised by at least one identified protein ⁵, would be expected to clear quickly, and no study reports a concentration-time curve. Whether the anti-apoptotic action is desirable in a person: blocking Bax translocation ⁷ is protective for a neuron under stress and is also the mechanism by which cells that should die do not, and the germ-cell-versus-tumour selectivity reported in two studies ¹⁶ is the only work addressing this and comes from one group. The antithrombotic effect ¹⁸, entirely unexamined for bleeding risk. Whether the inverse relationship between humanin and the growth hormone–IGF-1 axis ¹³ means that raising humanin is beneficial, harmful, or simply downstream of something else. Repeat-dose toxicology, genotoxicity, reproductive toxicology and carcinogenicity: none performed. Immunogenicity, unassessed. And the identity of commercial material: no published analysis has examined any product sold as humanin, which matters unusually here, since the difference between humanin and HNG is a single residue that cannot be detected without sequencing.

07 · US regulatory status

Current as of 6 September 2026. Humanin is not approved as a drug in the United States or any other jurisdiction and is not a controlled substance. Neither is S14G-humanin.

No marketing authorisation has been sought or granted anywhere for either molecule. No investigational new drug programme appears in the peer-reviewed record. The compound was identified in 2001 ² and has generated a substantial academic literature since without entering clinical development — a twenty-five-year gap between a Nature mechanism paper ⁷ and the absence of a first-in-human study.

Under the World Anti-Doping Code, peptide hormones and growth factors are prohibited at all times. Competitors should consult the current Prohibited List directly rather than rely on secondary summaries, including this one.

08 · Limitations of the evidence

1. Most of the therapeutic literature is about a different molecule. S14G-humanin is a designed mutant at a residue that structure-function work identified as functionally decisive ^6,8, and it is what nearly every intervention study since 2017 has used ^{12,14,15,16,18,19,20,22,23,24,26,27,28,29}. Results obtained with HNG are not evidence about humanin, and the two are routinely presented interchangeably.

1. No human has been given either peptide. The entire human record is observational measurement ^{17,21,23,25,30}.

1. The most-cited human association points the wrong way for a longevity claim. Plasma humanin correlates inversely with growth hormone and IGF-1, and a published review argues this may promote accelerated ageing ¹³. That argument has not been rebutted in the indexed literature.

1. Correlational human data cannot establish direction. Humanin rising after exercise ²¹, being related to sperm quality ¹⁷, or differing across a prostate cohort ³⁰ is compatible with the peptide causing the difference, responding to it, or both tracking a third factor.

1. The anti-apoptotic mechanism is a double-edged one. Preventing Bax translocation ⁷ is the basis of every protective result in this file and is also, in a cell with accumulated damage, exactly what should not happen. The two studies reporting selectivity between germ cells and cancer cells ¹⁶ come from a single group and have not been independently replicated.

1. Animal-to-human translation is untested, and the models are acute. Streptozotocin diabetes, crystal-induced arthritis, gentamicin ototoxicity, cyclophosphamide ovarian injury, UV-B skin injury ^{14,20,22,28,29}. Each isolates one insult over days. None resembles a chronic human condition.

1. Research concentration was extreme early and is diffuse now, which cuts both ways. The 2001–2004 foundational work came almost entirely from one Tokyo group ^{1,2,3,4,5,6,8,10,11}, which is a replication weakness at the origin; the recent cytoprotection work comes from many unrelated groups ^{12,18,22,26,28,31}, which is a strength for the phenotype and does nothing for the translational case.

1. No randomised controlled trial of the peptide exists in any species for any endpoint. The one randomised trial in which humanin appears tested exercise ²⁵.

1. Publication bias cannot be estimated. A literature this large and this uniformly positive about a cytoprotective peptide, with one dissenting review in twenty-five years ¹³, is not obviously a complete record of what has been tried.

1. Material identity is unestablished and unusually hard to establish. No published analysis has examined a commercial humanin preparation, and the distinction that matters most here — humanin versus HNG — is one amino acid.

09 · References

1. Hashimoto Y, Ito Y, Niikura T, Shao Z, Hata M, Oyama F, Nishimoto I. Mechanisms of neuroprotection by a novel rescue factor humanin from Swedish mutant amyloid precursor protein. Biochem Biophys Res Commun. 2001 May 4;283(2):460–468.

   PMID 11327724 ↗

2. Hashimoto Y, Niikura T, Tajima H, et al. A rescue factor abolishing neuronal cell death by a wide spectrum of familial Alzheimer’s disease genes and Abeta. Proc Natl Acad Sci U S A. 2001 May 22;98(11):6336–6341.

   PMID 11371646 ↗

3. Hashimoto Y, Niikura T, Ito Y, et al. Detailed characterization of neuroprotection by a rescue factor humanin against various Alzheimer’s disease-relevant insults. J Neurosci. 2001 Dec 1;21(23):9235–9245.

   PMID 11717357 ↗

4. Tajima H, Niikura T, Hashimoto Y, et al. Evidence for in vivo production of Humanin peptide, a neuroprotective factor against Alzheimer’s disease-related insults. Neurosci Lett. 2002 May 24;324(3):227–231.

   PMID 12009529 ↗

5. Niikura T, Hashimoto Y, Tajima H, et al. A tripartite motif protein TRIM11 binds and destabilizes Humanin, a neuroprotective peptide against Alzheimer’s disease-relevant insults. Eur J Neurosci. 2003 Mar;17(6):1150–1158.

   PMID 12670303 ↗

6. Yamagishi Y, Hashimoto Y, Niikura T, Nishimoto I. Identification of essential amino acids in Humanin, a neuroprotective factor against Alzheimer’s disease-relevant insults. Peptides. 2003 Apr;24(4):585–595.

   PMID 12860203 ↗

7. Guo B, Zhai D, Cabezas E, Welsh K, Nouraini S, Satterthwait AC, Reed JC. Humanin peptide suppresses apoptosis by interfering with Bax activation. Nature. 2003 May 22;423(6938):456–461.

   PMID 12732850 ↗

8. Terashita K, Hashimoto Y, Niikura T, et al. Two serine residues distinctly regulate the rescue function of Humanin, an inhibiting factor of Alzheimer’s disease-related neurotoxicity: functional potentiation by isomerization and dimerization. J Neurochem. 2003 Jun;85(6):1521–1538.

   PMID 12787071 ↗

9. Ikonen M, Liu B, Hashimoto Y, Ma L, Lee KW, Niikura T, Nishimoto I, Cohen P. Interaction between the Alzheimer’s survival peptide humanin and insulin-like growth factor-binding protein 3 regulates cell survival and apoptosis. Proc Natl Acad Sci U S A. 2003 Oct 28;100(22):13042–13047.

   PMID 14561895 ↗

10. Nishimoto I, Matsuoka M, Niikura T. Unravelling the role of Humanin. Trends Mol Med. 2004 Mar;10(3):102–105.

    PMID 15106598 ↗

11. Hashimoto Y, Terashita K, Niikura T, et al. Humanin antagonists: mutants that interfere with dimerization inhibit neuroprotection by Humanin. Eur J Neurosci. 2004 May;19(9):2356–2364.

    PMID 15128389 ↗

12. Nashine S, Cohen P, Chwa M, Lu S, Nesburn AB, Kuppermann BD, Kenney MC. Humanin G (HNG) protects age-related macular degeneration (AMD) transmitochondrial ARPE-19 cybrids from mitochondrial and cellular damage. Cell Death Dis. 2017 Jul 20;8(7):e2951.

    PMID 28726777 ↗

13. Mendelsohn AR, Larrick JW. Mitochondrial-derived peptides exacerbate senescence. Rejuvenation Res. 2018 Aug;21(4):369–373.

    PMID 30058454 ↗

14. Wang X, Liu X, Zhao Y, Sun H, Wang Y. Cytoprotective role of S14G-humanin (HNG) in ultraviolet-B induced epidermal stem cells injury. Biomed Pharmacother. 2019 Feb;110:248–253.

    PMID 30508736 ↗

15. Yang C, Xu L, Cui Y, Wu B, Liao Z. Potent humanin analogue (HNG) protects human sperm from freeze-thaw-induced damage. Cryobiology. 2019 Jun;88:47–53.

    PMID 30959025 ↗

16. Jia Y, Lue Y, Swerdloff RS, Lasky JL, Panosyan EH, Dai-Ju J, Wang C. The humanin analogue (HNG) prevents temozolomide-induced male germ cell apoptosis and other adverse effects in severe combined immuno-deficiency (SCID) mice bearing human medulloblastoma. Exp Mol Pathol. 2019 Aug;109:42–50.

    PMID 31085184 ↗

17. Rao M, Wu Z, Wen Y, Wang R, Zhao S, Tang L. Humanin levels in human seminal plasma and spermatozoa are related to sperm quality. Andrology. 2019 Nov;7(6):859–866.

    PMID 30920769 ↗

18. Ren L, Li Q, You T, Zhao X, Xu X, Tang C, Zhu L. Humanin analogue, HNG, inhibits platelet activation and thrombus formation by stabilizing platelet microtubules. J Cell Mol Med. 2020 Apr;24(8):4773–4783.

    PMID 32174022 ↗

19. Kim SM, Kang JI, Yoon HS, et al. HNG, a humanin analogue, promotes hair growth by inhibiting anagen-to-catagen transition. Int J Mol Sci. 2020 Jun 26;21(12):4553.

    PMID 32604799 ↗

20. Chen X, Yun C, Zheng H, et al. The protective effects of S14G-humanin (HNG) against streptozotocin (STZ)-induced cardiac dysfunction. Bioengineered. 2021 Dec;12(1):5491–5503.

    PMID 34506248 ↗

21. Woodhead JST, Merry TL. Mitochondrial-derived peptides and exercise. Biochim Biophys Acta Gen Subj. 2021 Dec;1865(12):130011.

    PMID 34520826 ↗

22. Zhang J, Lei H, Li X. The protective effects of S14G-humanin (HNG) against mono-sodium urate (MSU) crystals-induced gouty arthritis. Bioengineered. 2022 Jan;13(1):345–356.

    PMID 34965184 ↗

23. Nashine S, Cohen P, Wan J, Kenney MC. Effect of Humanin G (HNG) on inflammation in age-related macular degeneration (AMD). Aging (Albany NY). 2022 May 16;14(10):4247–4269.

    PMID 35576057 ↗

24. Ikegawa N, Kozuka A, Morita N, Murakami M, Sasakawa N, Niikura T. Humanin derivative, HNG, enhances neurotransmitter release. Biochim Biophys Acta Gen Subj. 2022 Oct;1866(10):130204.

    PMID 35843407 ↗

25. Delgado-Peraza F, Nogueras-Ortiz C, Simonsen AH, et al. Neuron-derived extracellular vesicles in blood reveal effects of exercise in Alzheimer’s disease. Alzheimers Res Ther. 2023 Sep 20;15(1):156.

    PMID 37730689 ↗

26. Zou L, Huang J, Zhang Q, Mo H, Xia W, Zhu C, Rao M. The humanin analogue (HNG) alleviates intrauterine adhesions by inhibiting endometrial epithelial cells ferroptosis: a rat model-based study. Hum Reprod. 2023 Dec 4;38(12):2422–2432.

    PMID 37814907 ↗

27. Nashine S, Kenney MC. Effects of Humanin G (HNG) on angiogenesis and neurodegeneration markers in age-related macular degeneration (AMD). Mitochondrion. 2024 Jan;74:101818.

    PMID 38029849 ↗

28. Lu Y, Bartoszek EM, Cortada M, Bodmer D, Levano Huaman S. Mitochondrial-derived peptides, HNG and SHLP3, protect cochlear hair cells against gentamicin. Cell Death Discov. 2024 Oct 21;10(1):445.

    PMID 39433756 ↗

29. Huang J, Zhang D, Zou L, Liu H, Xia W, Zhu C, Rao M. S14G-Humanin ameliorates ovarian dysfunction in a cyclophosphamide-induced premature ovarian insufficiency mouse model. Mol Hum Reprod. 2025 Jul 3;31(3):gaaf042.

    PMID 40811024 ↗

30. Coradduzza D, Cruciani S, Sibono L, et al. Diagnostic relevance of Humanin, GAS5 and miR-21/miR-103 in prostate disease risk stratification. Clin Exp Med. 2025 Aug 6;25(1):279.

    PMID 40768089 ↗

31. Maraux M, Vetter M, Zuffo LD, et al. HUMANIN produced by human efferocytic macrophages promotes the resolution of inflammation. Cell Death Dis. 2025 Aug 28;16(1):656.

    PMID 40877234 ↗