Nine Hours in the Killifish Room

The killifish room is kept at twenty-six degrees and smells faintly of pond. Four hundred tanks are stacked along three walls in a rack system that hums continuously, and in each tank a small, absurdly beautiful fish is dying on schedule.

Nothobranchius furzeri comes from ephemeral pools on the savannah of south-eastern Africa — puddles, essentially, that fill during the rains and are gone within months. An animal that breeds in a puddle has no time for a leisurely life history. The eggs survive in dried mud in a state of suspended development; the rains arrive; the fish hatch, mature in under three weeks, spawn furiously, and die. In captivity, with abundant food and no predators, they last four to six months. They are the shortest-lived vertebrate that can be bred in a laboratory, and that single fact is why this room exists.

"A mouse study takes three years and costs a fortune," says the postdoctoral researcher who runs the facility, tapping a tank where an older male hangs motionless near the surface, his tail frayed and his colours gone muddy. "He's fourteen weeks. In mouse terms he's about seventy-five. I can run a full lifespan experiment twice in a year. That's the entire reason anybody puts up with the smell."

She is thirty-one, has been doing this for five years, and has the particular dryness of someone who has spent a long time being asked at parties whether her work will let people live forever.

---

The question the room is designed to address is deceptively simple: is ageing one thing or many?

The intuitive view — the one most people hold without examining it — treats ageing as accumulated wear, like the gradual failure of a machine that has been running too long. Under this account there is nothing to study and certainly nothing to treat. Things break. Eventually enough of them have broken.

The field's organising framework, set out in an influential 2013 paper and revised a decade later, proposes something different and considerably more actionable. It identifies around a dozen distinct biological processes — the hallmarks of ageing — each of which can be measured, each of which worsens with age, and each of which, when experimentally accelerated, produces an animal that ages faster. Genomic instability. Telomere attrition. Epigenetic drift. Loss of protein quality control. Mitochondrial dysfunction. Chronic low-grade inflammation. Among them, and currently the most tractable, is cellular senescence.

A senescent cell is one that has stopped dividing but has declined to die. Cells reach this state for good reasons: a cell with damaged DNA that continues to divide is the first step towards a tumour, so arrest is a protective response, and in a young body such cells are cleared away by the immune system within days. The trouble is that clearance becomes less efficient with age while production continues. The cells accumulate. And they are not inert. They secrete a cocktail of inflammatory signals, growth factors and enzymes that degrade surrounding tissue — a phenomenon abbreviated to SASP, the senescence-associated secretory phenotype. A senescent cell does not merely fail to contribute. It actively degrades its neighbourhood, and it recruits nearby healthy cells into the same state.

"That's what makes it interesting therapeutically," she says. "It's not diffuse damage everywhere. It's a specific population of identifiable cells doing identifiable harm. In principle you can go and remove them."

Which is precisely what has been attempted. Compounds that selectively kill senescent cells while sparing healthy ones — senolytics — have produced results in mice that would be dismissed as fantasy if they had not been independently replicated. Aged animals given intermittent courses show improved cardiac function, better exercise capacity, restored fur, delayed onset of multiple age-related diseases, and modestly extended median lifespan. Crucially, treatment beginning late in life still works.

She anticipates the obvious question before it is asked.

"Mice are not people, and mouse ageing is not human ageing. Everybody in this building knows the graveyard of interventions that looked spectacular in rodents and did nothing in humans. The human trials that exist are small, they're mostly in specific diseases rather than ageing itself, and the honest summary is: promising, unproven, don't buy anything."

---

By afternoon the room has settled into a rhythm. Tanks are checked, dead fish logged and removed, a cohort of eight-week-olds photographed under anaesthesia for automated scoring of pigmentation and posture. The work is repetitive to a degree that outsiders find surprising, and it is punctuated by small losses. Individual fish are known by their tank codes. Some are known by more than that.

"You're not supposed to get attached," she says, entering a number. "You do anyway. The ones that live unusually long — you notice them. You start hoping."

The deeper reason for studying short-lived animals is not merely convenience. It is that ageing rates across species vary by orders of magnitude in ways that pure wear-and-tear cannot explain. A mouse and a naked mole-rat are similar in size and metabolic profile; one lives four years and the other more than thirty, with almost no increase in mortality risk as it ages. Some clams live for centuries. Certain whales appear to resist cancer despite carrying a thousand times more cells than a human. If ageing were simply entropy, such variation would be impossible. It is regulated. Regulation implies levers.

Finding the levers is what the room is for. Killifish from different wild populations — some from wetter regions, with longer rainy seasons — live measurably longer, and the genetic differences between them can be mapped. Diet, temperature and a small set of drug candidates are tested against those baselines. The results accumulate slowly, one four-month generation at a time.

At six in the evening she performs the last check, adjusts a feeder, and writes the day's mortality on a whiteboard beside the door: 4. Underneath, in older handwriting that has not been erased in months, someone has written the lab's unofficial motto — We are not trying to make you immortal. We are trying to make the last decade less terrible.

She switches off the overhead lights. The pumps continue. In four hundred tanks, four hundred small lives run down at a rate that is, for once, being watched closely enough to learn something from.