Humanin protected cultured neurons and improved memory in aged mice

What is humanin?

Humanin is a 24-amino-acid signaling peptide encoded within the 16S rRNA region of mitochondrial DNA, not a designed molecule. The body translates and secretes it. Researchers discovered humanin in 2001 while screening brain tissue from Alzheimer's patients for whatever let certain neurons survive insults that should have killed them ^[1].

That screen found this small mitochondrial peptide. Humanin abolished neuronal death caused by a wide range of familial Alzheimer's disease mutations and by amyloid-beta itself, and it left the neurons' response to unrelated toxic insults unchanged ^[1]. That specificity made humanin scientifically interesting from the start.

How does humanin work?

Humanin engages the formyl peptide receptor-like 1 (FPRL1) and a receptor complex involving CNTFR and gp130. Downstream, it activates pro-survival signaling, most notably through the STAT3 and Akt pathways, and interferes with pro-apoptotic proteins such as BAX.

The practical effect in cell and animal models is cytoprotection: cells exposed to stress that would normally trigger programmed death survive instead. The same broad mechanism is proposed for humanin's cardiac and metabolic effects, not a separate one.

What is HNG, and how does it differ from humanin?

HNG, also written S14G-humanin, is the synthetic analog most people encounter, and it swaps one amino acid for glycine at position 14. That single substitution makes HNG substantially more potent than native humanin in the lab, and HNG is the version used in essentially all of the animal dosing studies ^[2].

Structural work has mapped which parts of the molecule matter. The cysteine at position 8, the serine-to-glycine swap at position 14, and the domain spanning roughly proline 3 to proline 19 are all required for the rescue activity ^[2]. That mapping explains why HNG behaves differently from native humanin in the lab. It says nothing about how either performs once injected into a human.

What does humanin do in cultured neurons?

Humanin and HNG blocked cell death in cultured neurons triggered by multiple familial Alzheimer's disease gene mutations and by amyloid-beta fragments ^[1]⁠^[2]. The effect was specific to those Alzheimer's-relevant insults rather than a general anti-death effect ^[1]⁠^[2].

In vitro neuroprotection is where humanin research started and remains its most internally consistent finding. It is solid preclinical pharmacology. It is not evidence that the peptide does anything comparable in a living human brain.

What does humanin do in mice?

Humanin improved spatial memory and cognitive performance on standard rodent tests in 18-month-old mice, which were aged rather than young ^[3]. A separate line of work injected aged mice with the HNG analog twice weekly for fourteen months at 4 mg/kg ^[4].

The aged mice given HNG showed less myocardial fibrosis, less cardiac fibroblast proliferation and less cardiac apoptosis than untreated aged mice, with the effect tracked to activation of the Akt/GSK-3beta signaling pathway ^[4].

Both are legitimate rodent pharmacology with clear mechanistic follow-through, run over long timeframes. Fourteen months is a serious chronic-dosing study by animal-research standards ^[4]. Both stop at mice. Separate evidence reviews cover the mitochondrial peptide stack, SHLP2 and MOTS-c.

What does the human genetic data on humanin show?

In a large, nationally representative human cohort, the mitochondrial SNP rs2854128, which reduces circulating humanin levels, was associated with faster cognitive aging ^[3]. The finding is real and non-trivial: it links naturally occurring variation in humanin production to a meaningful human outcome.

It is an association between genetically determined endogenous levels and an outcome measured across a population. It is not a trial in which humans received humanin and were followed for a cognitive endpoint. Nothing in that dataset says what happens if a person with normal humanin genetics receives exogenous HNG by injection.

Why doesn't the humanin correlation justify injecting it?

Observational data linking higher circulating humanin to markers of healthy aging establish humanin as a biomarker worth understanding, not as a cause ^[3]. The correlation does not establish that pushing levels higher with an injected analog would produce the same downstream biology.

People with genetically higher lifetime humanin exposure have had that exposure since birth, in a specific tissue and temporal pattern, alongside whatever other genetic and physiological factors travel with that SNP. A bolus of injected HNG in adulthood is a different intervention, and no study has tested it in humans.

The rodent injection data are the closest thing to an interventional test that exists. Mice are not small humans, particularly for a peptide acting through mitochondrial and receptor-mediated pathways with known cross-species differences.

Do the humanin dosing protocols online come from any study?

No published study supports the human dosing protocols, injection frequencies or subjective benefit reports for HNG that circulate on forums. None of the four studies cited below addresses them, so they are forum anecdote rather than evidence.

The only dosing figure with a citation is the mouse protocol: 4 mg/kg of the HNG analog, twice weekly, for fourteen months ^[4]. Extrapolating a mouse mg/kg dose to a human is not standard practice without pharmacokinetic bridging data, and none exists for humanin.

What is still unknown about humanin?

The practical questions are open from exposure through to risk:

• Pharmacokinetics. No human study establishes absorption, half-life or how quickly levels return to baseline for humanin or HNG, injected subcutaneously or otherwise.
• Human dose. An effective or safe human dose is unknown.
• Safety. Long-term human safety is undocumented. The reported side effects in research-use contexts, injection-site reactions, mild fatigue, uncommon headache and transient nausea at higher doses, come with no clinical safety study behind them and are anecdotal.
• Malignant cells. A pro-survival, anti-apoptotic peptide could in principle help stressed or damaged cells, including malignant ones, resist death signaling. That concern is mechanistically coherent given the BAX-interacting pathway, and no study has quantified or tested it as a risk in humans.