A Dozen Body Signals Could Reveal a Slow-Aging Drug in Months

A MOUSE given a promising anti-aging drug at four months old will not tell you whether it worked until it is old and, often, dead. That’s the bottleneck that has throttled the entire field: a full lifespan study takes three to four years and can run US$300,000, and until the animals finish dying, nobody knows if the compound did anything.

A new paper in Frontiers in Science proposes a shortcut. Twelve measurable biological signals, drawn from a decade of mouse experiments, appear to shift the same way in every strain of mouse known to age slowly, whatever caused the slow aging in the first place, and the authors argue that reading those signals at a single timepoint could show whether a drug is working in months rather than years.

Biogerontology has stopped being purely observational. Rapamycin, acarbose, several single-gene mutations and a calorie-restricted diet all reliably extend mouse lifespan substantially, and the field’s central problem has shifted from finding an anti-aging intervention at all to testing candidates fast enough to keep up.

Fourteen agents or combinations have significantly extended lifespan in mice tested under the National Institute on Aging‘s Interventions Testing Program (ITP), a standardized protocol run at three independent sites with enough mice, both sexes included, to reliably detect a modest effect. Eight of those fourteen produced increases of 12% or more. Rapamycin alone lifted median lifespan by 23% in males and 26% in females, the largest effect of any single drug tested. Combined with acarbose, it pushed male lifespan up 29%, the biggest gain the program has recorded for any pharmacological intervention.

None of that solves the waiting problem. Researchers still had to watch each cohort age out to know whether a candidate worked, until they started comparing the mice at the molecular level instead of just at the finish line.

The Same Signal Keeps Showing Up

Comparing calorie-restricted mice, five long-lived single-gene mutants and four kinds of drug-treated mice against their own well-fed, undrugged littermates, the team behind the new paper kept finding the same handful of changes regardless of which intervention produced the slow-aging state. Levels of a fat-browning protein, uncoupling protein 1 (UCP1), rose. So did a liver enzyme, GPLD1 (short for a mouthful: glycosylphosphatidylinositol-specific phospholipase D1), and two proteins linked to brain health, brain-derived neurotrophic factor (BDNF) and doublecortin. Activity in the growth-signalling mTOR pathway dropped. Inflammatory signalling through the regulator protein NF-kB quieted down. Twelve such changes recur often enough, the authors argue, to serve as aging rate indicators (ARIs): a slow-aging fingerprint that shows up whether the animal got a drug, a diet or a mutation.

“We hope aging rate indicators (ARIs) will advance the search for anti-aging drugs in three ways,” says Richard Miller, a geneticist at the University of Michigan, and senior author on the paper. New candidates could be screened in mice more cheaply, changes in treated people could be compared against the mouse signature, and the shared indicators themselves might point toward the handful of upstream control pathways that many different anti-aging interventions apparently converge on.

An Odometer Versus a Speedometer

The distinction the authors draw is between a car’s odometer and its speedometer. A conventional biomarker of aging, the kind derived from methylation patterns or physical decline, tells you how much aging has already accumulated, and doing that reliably takes two measurements separated by enough time for a difference to show up: an odometer reading. An ARI, in principle, does not need the wait. It is meant to shift, once, into a new steady state as soon as an animal enters a slow-aging condition, whether that shift happens early or late in life, and the direction of the shift, up or down, does not matter, only that it is consistently different from an untreated animal’s. A protocol tested in the paper exposed mice to a drug or a calorie-restricted diet at four months old and measured the indicators eight months later, at just 10 to 20 mice of each sex per group, far fewer animals and far less time than a lifespan study demands.

A machine-learning study folded into the roadmap put a number on how well this might generalize. Researchers trained a model on plasma samples from mice given four different interventions, rapamycin, acarbose, an estrogen compound and a diabetes drug called canagliflozin, alongside calorie restriction, using over a thousand identified metabolites, then asked it to predict the lifespan benefit of the one intervention the model had never seen. It succeeded for every one of the five, at a level unlikely to arise by chance, and a model trained only on male mice still predicted benefits correctly in females, even for drugs that only extend lifespan in males.

None of the twelve candidate indicators has been validated in a species other than the mouse, and only two of them, irisin and GPLD1, can currently be measured from a simple blood sample rather than a dissected organ, which matters enormously if the goal is eventually testing living, breathing people rather than mice. Most of the underlying cross-model comparisons also come from a single laboratory group working with one genetically heterogeneous mouse stock, so independent replication in other colonies is still owed.

If the indicators hold up, the payoff reaches well past mice. Several of the same drugs that shift ARIs also delay markers linked to diabetes, several cancers, heart disease and neurodegeneration in the animals that receive them, a pattern geroscience researchers elsewhere have argued reflects shared aging pathways sitting upstream of many distinct chronic diseases, and a screening method that runs on a handful of blood or tissue samples rather than a three-year survival curve could let researchers chase those downstream connections at a pace the field has never had. It could also let a clinician compare a person’s blood chemistry after a candidate drug to the pattern already established in slow-aging mice, well before anyone knows whether that person will actually live longer, the same gap that has slowed proposals for large human anti-aging trials such as the long-discussed metformin trial known as TAME (Targeting Aging with Metformin), proposed nearly a decade ago and still unfunded.

What the mouse work has already overturned is the assumption that timing is everything: several of the fourteen ITP-validated interventions, rapamycin included, still worked when started in late middle age rather than from birth, so a drug that shifts an aging rate indicator in a middle-aged volunteer would not automatically be dismissed as arriving too late.

FAQ