Peptide LexiconAll compounds

Humanin: Seven Registry Records, None of Them a Dose

Humanin is a peptide your own mitochondria make, with 542 published papers and 312 carrying a human tag. Every one of its seven trial-registry records measures it rather than gives it — including the two filed as interventional.

PeptideMade by the human bodyNo human administration on record

Caroline S · Published 2026-09-13

Length

24 amino acids

Sequence

A 24-amino-acid peptide encoded within a short open reading frame in the mitochondrial 16S ribosomal RNA gene region, MT-RNR2 — not in the cell's nuclear DNA

Origin

Discovered in 2001 in a screen for factors that protect nerve cells against Alzheimer's-related death. It was the first of the mitochondrial-derived peptides to be described; MOTS-c, from the 12S region, came later.

Last reviewed

2026-09-13

What is it good for?

It is sold for longevity, metabolic health and mitochondrial function. In the research literature it is mostly studied the other way round — as a substance the body's own mitochondria release under stress, and therefore as a possible marker of that stress.

Illustration: A white multi-well laboratory plate on a pale bench with a copper block at one end and a dark green wall behind.
Illustration

What it is

Humanin is a peptide of 24 amino acids that the human body makes. What sets it apart is where the instructions live: inside the mitochondrial 16S ribosomal RNA gene region — the mitochondrion's own DNA, rather than the chromosomes in the cell nucleus.

It was found in 2001, in a screen looking for factors that protect nerve cells from the kind of death seen in Alzheimer's disease, and it was the first peptide shown to be encoded that way. A whole category followed it: the mitochondrial-derived peptides. The other member of that family with an entry in this library is MOTS-c, a 16-residue peptide from the neighbouring 12S region, and the two are usually sold together.

It is marketed for longevity, metabolic health and mitochondrial function. The research literature largely runs in the opposite direction — treating humanin as something the body releases under stress, and therefore as a possible signal of that stress.

A large literature, and what the large numbers measure

542 PubMed records name it in title or abstract as of 2026-09-13. That is a bigger file than most compounds here have, and it reflects genuine scientific interest in the molecule.

The breakdown matters more than the total:

Query Records
humanin[tiab] 542
with PubMed's human tag 312
mentioning mice, rats or mouse models 235
randomized-controlled-trial publication type 5
paired with centenarian 6
concerning HNG (S14G-humanin) 103

The 312 is the number to be careful with. PubMed's human tag marks research involving people or human material. For a peptide the body produces, that overwhelmingly means studies that took a sample and measured how much of it was there. The tag is accurate; it simply answers a different question from the one a reader asking "has this been tested in people?" has in mind. This library has hit the same trap on MOTS-c, LL-37 and VIP — every time the compound is one the body already makes.

And 103 of the records are about a different molecule. HNG, written in papers as S14G-humanin, is humanin with a single amino acid substituted, and the literature describes it as a potent humanin analogue. Much of the protective animal work uses it for exactly that reason — it produces a clearer effect. A result obtained with HNG is evidence about HNG. The titles often say only "humanin analogue", so the distinction survives only if someone reads the methods.

The registry, read record by record

This is where the compound's file is genuinely unusual, and it takes one minute to check.

ClinicalTrials.gov held 7 records on 2026-09-13 — more than most compounds in this library manage. Five are observational. Two are filed as interventional, which ordinarily means somebody is being given something.

Here is what the two interventional studies intervene with:

  • NCT07438002Analysis of Effects of High-intensity Physical Exercise in Subjects With Dialyzed Chronic Kidney Disease (Casa di Cura Dott. Pederzoli, n=50, recruiting). The intervention is exercise.
  • NCT07678073Comparison of the Effects of General Anesthesia and Combined Spinal-Epidural Anesthesia (University of Gaziantep, n=68, recruiting). The intervention is the choice of anaesthetic.

In both, humanin is among the substances measured afterwards.

The five observational records are openly what they are: plasma humanin in acute kidney injury (Guangdong Provincial People's Hospital, n=60), humanin's value in diagnosis and prognosis of acute kidney injury after heart transplant (n=60), muscle biopsies taken during orthopaedic surgery in cerebral palsy (Eva Pontén, n=150), IVF outcomes in polycystic ovary syndrome (Assiut University, n=100), and the largest completed one — humanin isoforms in cardiac muscle and blood plasma against major complications after cardiac operations (University of Tartu, n=106, completed).

Across all seven records, not one administers humanin to anybody. The registry contains humanin studies and contains no humanin dosing. A reader who counts seven registrations and concludes the compound is in human trials has been misled by an accurate number — and the two that say interventional are what makes the count so convincing.

This is a distinct shape from the reasons this library has already catalogued for an empty registry. Here the registry is not empty. It is full of the wrong thing, and the label on two of the records points the wrong way.

What the biomarker question actually is

The studies above are asking whether the amount of humanin in someone's blood or tissue predicts how they do — after heart surgery, after a transplant, in kidney injury, in fertility treatment.

That question does not have an obvious answer, and the ambiguity is the reason it is being asked. Mitochondria appear to release these peptides under stress. So a high reading is consistent with a strong protective response, and equally consistent with a great deal of stress requiring one. Until those are told apart, a rising number cannot be read as a benefit — including the rise after exercise, which is the finding most often quoted in marketing.

Where it stands

There is no approved product and no application. Drugs@FDA and DailyMed both returned nothing on 2026-09-13.

There is also no dose, no route, no dosing interval and no safety profile for giving it to a person, anywhere in the published or registered record. That it is a natural human peptide is not a substitute: insulin, thyroid hormone and adrenaline are all natural, and all are dangerous in the wrong amount.

Only 6 of the 542 records involve centenarians, which is a thin foundation for the longevity claim that sells it.

The honest summary: humanin is a real and scientifically interesting molecule with a substantial laboratory literature, a growing role as a possible marker of mitochondrial stress, and — after twenty-five years — not one study in which it was given to a human being.

For the other mitochondrial peptide sold beside it, see the MOTS-c entry; for the mitochondrial compound in this library that does have registered human trials, SS-31; and for the cluster these claims belong to, peptides for longevity.

What the research shows

Exists in the human body and is measurable in blood and tissue

This is the best-established fact about it and the subject of most of its human literature

Protects cells and tissue in laboratory and animal models

A large preclinical literature — 235 records mention mice, rats or mouse models

Has been given to a person in a registered study

Zero. All seven registry records measure the peptide; none administers it

The animal work tests the natural peptide

Substantially not — 103 of the 542 records concern HNG, the S14G variant described as a potent analogue, a different molecule from the one sold

Has an established dose, route or safety profile in people

None exists in the published or registered record

Bars show how much of the evidence is in humans, not how well anything works.

Where it stands

United States

No approved product and no application. Drugs@FDA and DailyMed returned no record on 2026-09-13.

Published record

542 PubMed records name it in title or abstract as of 2026-09-13. 312 carry PubMed's human tag, 235 mention mice, rats or mouse models, and 5 carry the randomized-controlled-trial publication type.

Trial registry

7 records on ClinicalTrials.gov (2026-09-13): 5 observational and 2 interventional. In the two interventional studies the interventions are physical exercise and a choice of anaesthetic; humanin is an outcome being measured in both.

What the registry studies are about

Acute kidney injury after cardiac surgery and after heart transplant, muscle biopsies in cerebral palsy, IVF outcomes in polycystic ovary syndrome, high-intensity exercise in dialysis patients, and anaesthetic technique. Sponsors include the University of Tartu (n=106, completed), Guangdong Provincial People's Hospital and Assiut University.

How it is sold

As a research powder for injection, usually alongside MOTS-c and the other longevity peptides, with reference to the mitochondrial literature rather than to any human administration study.

Frequently asked questions

What is humanin?

A peptide of 24 amino acids that your own mitochondria make. Unusually, the instructions for it sit inside mitochondrial DNA rather than in the cell's nucleus — it was the first peptide found to be encoded that way, in 2001, during a search for factors that protect nerve cells from Alzheimer's-related damage.

What does the research actually show?

Two quite different things that are easy to blur. In cells and in animals, humanin and its variants protect tissue against several kinds of stress, and that literature is substantial. In people, the research almost entirely measures the peptide the body already produces — in plasma, in cardiac muscle, in follicular fluid — to see whether the amount tracks with illness or recovery. The first line is about what the molecule can do. The second is about what its level indicates.

There are 312 human papers. Doesn't that mean it has been studied in people?

It means it has been measured in people. PubMed's human tag marks research involving humans or human material, which includes every study that draws blood and assays a peptide. It does not mean anybody was given anything. This library has now hit the same trap on several compounds where the molecule is one the body makes — the count is real, and it answers a different question from the one a reader is asking.

What about the seven trial registrations?

They are the clearest illustration of the point, and they are worth checking individually. Five are openly observational — plasma humanin in acute kidney injury, humanin isoforms after cardiac surgery, muscle biopsies in cerebral palsy, IVF outcomes in polycystic ovary syndrome. The two marked interventional look more promising until you read what is being done: in one the intervention is a high-intensity exercise programme in dialysis patients, in the other it is the choice between general and spinal-epidural anaesthesia. In both, humanin is one of the things measured afterwards. Not one of the seven gives humanin to anyone.

Is the animal evidence at least about the compound being sold?

Partly, and less than it looks. 103 of the 542 records concern HNG — S14G-humanin — a variant with one amino acid changed, described in the literature as a potent analogue and used in much of the protective work precisely because it is stronger. A result obtained with HNG is evidence about HNG. Anyone reading the preclinical literature as support for injecting humanin has to separate the two, and the paper titles frequently do not make that easy.

Does a higher humanin level mean better health?

The biomarker studies are asking that question and have not settled it, and the direction is not obvious either way. Mitochondria appear to release these peptides under stress, so a high level can plausibly mark either a robust protective response or a lot of stress to respond to. That ambiguity is why the registry studies exist — and it is also why a number that rises after exercise cannot be read straight off as a benefit.

Is it safe to take?

Unknown, with no human administration study of any size to draw on. The fact that the body produces it is not a safety argument: the body produces insulin, thyroid hormone and adrenaline, all of which are dangerous in the wrong amount at the wrong time, and none of them arrives by injection from a research powder at a concentration nobody has established.

What would change the picture?

A phase 1 study — any study that administers the natural peptide to healthy volunteers and reports what happened. After 542 papers and twenty-five years, that study does not exist. The gap between a well-studied molecule and a tested treatment is the gap this entry describes, and here it is unusually wide.