Asking the cells that didn't die

Most peptides in this catalog were found by asking what a signal does. Humanin was found by asking a stranger question — not what kills a neuron, but what the neurons that didn’t die were holding on to. In 2001 a Keio University lab took the occipital cortex of a patient who had died with Alzheimer’s and screened the few surviving neurons for any gene that could rescue cells from the death triggered by familial-Alzheimer’s mutations and amyloid-β. One clone did it almost completely. It encoded a peptide just twenty-four residues long, and they named it humanin.

Then came the part that made it strange. The sequence wasn’t in the nuclear genome at all. It was written inside mitochondrial DNA — tucked within the 16S ribosomal RNA gene, the region that otherwise codes for part of the mitochondrion’s own protein-building machinery. A survival message, hiding in the powerhouse’s rRNA, read out of the one genome nobody expected to be dictating peptide hormones.

That origin puts humanin in a small, recently opened club: the mitochondrial-derived peptides (MDPs). Its better-known sibling MOTS-c is encoded a little further along the same mitochondrial genome and tilts toward metabolism and exercise; humanin is the family’s cytoprotectant, the one whose single obsession is keeping cells alive under stress. The shared lesson is the one the field is still absorbing: the mitochondrion is not only a factory that answers to the nucleus. It writes back — short peptides that travel out to the rest of the cell, and into the blood, carrying news about the state of the engine room.

What makes humanin worth a long look is that it is read in two different places, and does something different in each. It is at once an intracellular brake bolted directly onto the machinery of cell death and a secreted hormone that works a receptor from the outside. Neither reading depends on the other; the peptide simply means “survive” wherever it happens to be.

Apoptosis — a cell’s orderly suicide — usually runs through Bax, a protein that, when activated, moves from the cytosol to the mitochondrial outer membrane, punches it full of pores, and lets the death signals (cytochrome c and the rest) spill out. Once Bax reaches the membrane and oligomerises, the decision is effectively made.

Humanin steps in before that. Working from inside the cell, it binds Bax directly and prevents its translocation to the mitochondrion — the pore never forms, the membrane stays sealed, the cascade never starts. The seminal demonstration was Guo and colleagues in Nature in 2003, showing humanin interferes with Bax activation itself. It reaches related triggers the same way — sequestering Bid and Bim, the proteins that would otherwise hand Bax its cue. It is less a drug acting on a pathway than a hand held over the trigger.

A parallel intracellular partner sharpens the picture. Humanin binds IGFBP-3, an insulin-like-growth-factor binding protein with its own pro-apoptotic streak, and neutralises it — tying humanin’s survival signal into the IGF axis that runs through so much of this catalog’s growth and metabolic biology.

Secreted, the identical peptide behaves like a proper hormone. It docks a three-part receptor assembled from CNTFRα, WSX-1, and gp130 — a borrowed committee of cytokine-receptor subunits — and through it fires the JAK/STAT3 pathway, the same intracellular relay a dozen survival and anti-inflammatory signals converge on. A second route, through the formyl-peptide receptor FPR2/FPRL1, links it to inflammatory and vascular signalling. The outcome each time rhymes with the intracellular one: transcriptional programs that favour survival and dampen stress.

So the peptide has, in effect, two independent ways to say the same thing. Inside, it is mechanical — grab Bax, hold the line. Outside, it is a message on a receptor — and, uniquely for this family, a message the cell can send, a genuine mitochondrial hormone circulating between tissues rather than a private note to self.

Because humanin travels in the circulation, it can be measured — and the measurements are the most human thing about the story so far. Circulating humanin falls with age, declining across the decades in ways that track other markers of mitochondrial decline; it tends to run higher in the exceptionally long-lived and in the offspring of centenarians, and it moves with insulin sensitivity and metabolic health. The peptide reads, tantalisingly, like a dial on the engine room that someone might one day want to turn.

The tool most preclinical work reaches for is not native humanin but HNG (S14G-humanin) — a one-residue swap reported to be roughly a thousand times more potent in neuroprotection assays, because the native peptide is short-lived and comparatively weak. It is the version behind most of the striking animal results, and it is worth keeping straight: much of what gets attributed to “humanin” is really the behaviour of an engineered analog.

This is where the creed — bullish on the science, sceptical on the page — has to earn its keep, because the biology is easy to fall for. The mechanism is real and unusually well characterised for so young a molecule: the Bax brake, the IGFBP-3 tie-in, and the trimeric receptor are each grounded in primary work, and an effect that shows up through two independent routes is more likely a true property than an artefact of one assay.

• Established. Humanin is encoded in mitochondrial DNA, is cytoprotective in cell and rodent models, binds Bax to block apoptosis, and signals through a CNTFRα/WSX-1/gp130 receptor to STAT3. This is the well-supported core.
• Associative in humans. Blood levels fall with age and run higher in the long-lived and their offspring, and correlate with insulin sensitivity — real observations, but correlations, not proof that raising humanin changes outcomes.
• Not yet shown. No controlled human trials establish that administering humanin (or HNG) treats or prevents Alzheimer's, diabetes, cardiac injury, or ageing itself. The clinical case is absent, not merely early.

The gap between those three lines is the whole story. Humanin is one of the more remarkable ideas in the catalog — a survival hormone the mitochondrion writes in its own genome and speaks two ways at once — and one of the least tested in people. Both halves of that sentence are true, and the marketing that has begun to attach itself to the peptide tends to quote only the first. The honest summary is that we found humanin by asking the cells that survived what they were holding on to, and we are still, twenty-odd years later, working out whether we can hand it to the ones that didn’t.