Humanin — the mitochondrial peptide that protects neurons

A small sequence of RNA, expressed from an unusual location, encoded a short peptide that, when introduced to neurons under amyloid-beta attack, protected them from dying. They called it Humanin.

The unusual location was the key thing. Humanin was encoded not by the nuclear genome, where most human proteins are encoded, but by a short open reading frame nested inside the mitochondrial genome — specifically in the 16S ribosomal RNA gene. The mitochondrial genome is small. It encodes just thirteen proteins involved in oxidative phosphorylation, plus the ribosomal and transfer RNAs needed to make them. The idea that it also encodes short signaling peptides with neuroprotective function wasn't something the field was looking for. Humanin landed in the literature as a genuine surprise.

It is a twenty-four amino acid peptide. Twenty-four amino acids is short — smaller than most biologically active peptides, closer to the size of a regulatory signal than a structural protein. Its circulating levels can be measured in blood. It can be detected in tissues. It declines with age. It is inversely correlated with markers of Alzheimer's risk in observational data. And it does something that most peptides encoded in mitochondrial DNA do not do: it leaves the cell. It signals.

Understanding why that matters requires a brief detour into what the mitochondrial genome is and what it was designed — through billions of years of evolution — to do. The mitochondrial genome is a remnant. Mitochondria were once free-living bacteria, and they still carry bacterial DNA, bacterial ribosomes, and bacterial gene expression machinery. Over evolutionary time, most mitochondrial genes migrated to the nuclear genome. The thirteen proteins that stayed behind are the ones most tightly integrated into the inner membrane — the core subunits of the electron transport chain that are apparently too hydrophobic, or too positionally critical, to be expressed anywhere else and imported. The mitochondrial genome was thought to be a stripped-down maintenance manual: minimal, functional, specific.

What Humanin and the peptides discovered after it suggest is that the mitochondrial genome carries a second layer of encoded function — one that operates as a communication system between the mitochondrion and the rest of the cell, and between the cell and other tissues. The mitochondria, it turns out, have something to say. When they're under stress, when conditions are dangerous, when cellular survival is threatened, they secrete peptides that carry that information outward. Humanin is the first such peptide that was clearly characterized.

The family it belongs to is now called mitochondrial-derived peptides, or MDPs. The list has grown since Humanin was discovered to include MOTS-c, which regulates metabolic homeostasis and has been studied for its effects on insulin sensitivity and exercise adaptation; a series of peptides called SHLPs, short humanin-like peptides 1 through 6, each with distinct receptor interactions and biological activities; and possibly others not yet characterized. The MDP field is young, active, and increasingly convinced that the mitochondrial genome's signaling function is real and important — that these are not curiosities but a class of retrograde signals through which mitochondria coordinate cellular stress responses.

Humanin's biology, as it has been worked out over the two decades since its discovery, covers several distinct mechanisms. The most studied is its anti-apoptotic effect. Apoptosis is programmed cell death — a controlled process by which a cell dismantles itself when it receives the right signals. In neurons under stress, including the stress of amyloid-beta toxicity, the apoptotic pathway can be inappropriately activated: the neuron receives a death signal that should have been blocked, and it dies when it shouldn't have. Humanin appears to interfere with this inappropriate activation at multiple points. It binds IGFBP-3, an insulin-like growth factor binding protein that has its own pro-apoptotic signaling role, and sequesters it away from the apoptotic cascade. It interacts with Bax, a protein that triggers the mitochondrial phase of apoptosis, and prevents it from inserting into the outer mitochondrial membrane. And it interacts with the tripartite receptor complex composed of CNTFR, WSX-1, and gp130 on the cell surface, activating STAT3 signaling in ways that promote cell survival.

These aren't redundant overlapping effects on a single pathway. They're distinct points of intervention at different stages of the apoptotic process, which is part of why Humanin was an interesting research target from the beginning: a peptide that works at multiple nodes in a death pathway is a peptide with unusual leverage.

The metabolic biology is a second distinct story. Humanin circulates in blood at measurable concentrations and appears to act as a systemic hormone-like signal. It has been researched for effects on insulin sensitivity, with some animal studies suggesting it improves glucose handling and reduces fat accumulation. It appears to have anti-inflammatory properties, reducing cytokine production in some inflammatory contexts. In the vascular system, preclinical work has looked at its effects on endothelial function and atherosclerotic plaque development — findings that, if they translate, would connect Humanin biology to cardiovascular aging as well as neurodegeneration. These metabolic effects are less well characterized than the neuroprotective ones, and they haven't been tested in human clinical trials. They're preclinical signals. They point toward a molecule with broader metabolic relevance than the original Alzheimer's research context suggested.

The age-related decline in circulating Humanin is one of the more striking findings in the observational literature. Humanin levels in blood appear to fall with age in a pattern that roughly parallels the increase in neurodegeneration risk, metabolic dysfunction risk, and overall cellular stress load. This inverse correlation has been observed in humans across multiple populations. It's also been observed in long-lived human families — centenarians and their offspring appear to have higher circulating Humanin levels than age-matched controls without exceptional longevity. This doesn't establish Humanin as a cause of longevity. It's consistent with Humanin being part of a physiological system that works better in people who age well, and that declines as that system breaks down. Correlation, not causation. But the correlation is specific enough to take seriously.

The mechanism behind the age-related decline isn't fully established, but the leading hypothesis connects it to the overall decline in mitochondrial function with aging. If Humanin is produced by mitochondria under stress, and if mitochondria produce it as a survival signal, then the decline in Humanin may reflect declining mitochondrial mass, declining mitochondrial stress-response capacity, or a combination. The mitochondria of an aged cell are fewer, less functional, and less responsive to the same stresses that would have prompted a robust Humanin signal in a younger cell. Less signal means less protection means more vulnerability to the apoptotic and inflammatory triggers that accumulate with age.

It's worth being clear about where the evidence sits, because the difference between "this is an interesting and well-characterized biology" and "this is something you should be taking" is significant and the field is still firmly in the first category.

Humanin is not FDA-approved. It is not an approved drug in any jurisdiction as of mid-2026. It exists as a research compound — synthesized, studied in cell culture and animal models, measured in human observational studies, and increasingly investigated in the context of aging biology. There are no completed phase 2 or phase 3 human clinical trials in any indication. The pharmacokinetic challenges are real: Humanin has a short half-life in circulation, crossing the blood-brain barrier in therapeutically meaningful concentrations is not trivial, and dose optimization for human use hasn't been established through the standard trial process. These aren't reasons to dismiss the biology. They're accurate descriptions of where the science is.

The MDP story as a whole is at a stage that feels like the early peptide hormone era in some ways — a class of molecules with real biological activity and interesting mechanisms, waiting for the technical and clinical infrastructure to catch up with the discovery biology. Humanin was found by accident while looking for something else, which is how biology often reveals its depth. The question of what it means for human health — whether you can meaningfully raise it, whether doing so would alter disease trajectories, whether it can be administered in a form that reaches the tissues where it matters — is a genuinely open one, being worked on now.

What the discovery has already changed is the conceptual picture. The mitochondrial genome was understood as a minimal operating manual for electron transport. It appears to also be a signaling archive — a set of sequences that encodes the mitochondria's own vocabulary for communicating cellular stress, survival priorities, and possibly longevity signals to the rest of the body. Humanin was the first word read from that vocabulary. It is unlikely to be the last, and the sentences may turn out to be important.