A one-week stress signal in fly brains made the flies live longer

Cells have an emergency system for hard times. When amino acids run short, or when the mitochondria that make a cell's energy get into trouble, the cell slows protein production and switches on a set of rescue genes. Biologists call this the integrated stress response, and the protein that does most of the switching is called ATF4.

Whether that response is good or bad for aging has been an awkward question. Worms and yeast live longer when their version of ATF4 is turned up, and several long-lived mouse strains have more ATF4 activity — but removing the stress response also extends life in worms, and a recent study found that damping ATF4 down lengthens life. A new preprint argues that both camps can be right, because what matters is not whether the response is on, but for how long and in which tissue.

What they did

The researchers worked in fruit flies, using a genetic switch that only turns on when the flies eat a particular drug, so ATF4 activity could be raised and then lowered again at will. Mostly female flies were used, because the lifespan effects turned out to be limited to females. They compared ATF4 turned up throughout adult life; turned up body-wide for only the first one or three weeks of adult life, or in midlife, then switched off; and turned up in one tissue at a time — neurons, the fat body (the fly's combined liver and fat tissue), the gut, or flight muscle.

Alongside lifespan they measured body weight, food intake, egg laying, climbing ability (a standard test of neuromuscular health in flies) and survival of starvation, and read out which genes were active in head, gut, muscle and fat body at several ages. Each lifespan trial included control flies carrying only the switch, so that effects of the inducing drug itself could be separated out.

What they found

• Duration decided the sign of the effect. Permanent, body-wide ATF4 activation in adults shortened life. Confining it to the first one or three weeks extended life instead, three weeks working better than one. Pooling four trials (1,166 deaths), the transient treatment roughly halved the risk of death, and the advantage persisted into old age. Midlife activation also worked, less strongly.
• Body-wide activation came with costs. During the induction weeks the flies were lighter and laid far fewer eggs. Both recovered afterwards, but their body composition stayed shifted: less total protein, with fat unchanged. Food intake did not drop, so they were not simply eating less.
• One week in neurons alone was enough — and cost nothing. ATF4 in the fat body and gut, in muscle, or in neurons for three weeks did nothing for lifespan. One week in neurons extended it, with a pooled risk of death 0.55 times that of controls across three trials, and slowed the age-related decline in climbing, while leaving body weight and egg laying normal.
• The strongest trace was left in a tissue that was never touched. In 50-day-old flies — 41 days after the brain signal had been switched off — the fat body had over 3,000 genes running at a different level, against fewer than 110 in head, gut or muscle. Genes that pack DNA away into silent, tightly wound heterochromatin were turned up, 39 of the 99 genes assigned to that job. "Jumping genes" — parasitic DNA sequences that normally escape silencing as animals get old — shifted downward in 87 of 103 families, though the total was not statistically significant (0.51-fold of controls, p = 0.064).
• Old fat-body mitochondria worked better. Respiratory capacity in fat bodies of 50-day-old flies was almost twice as high at complex I, the entry point of the respiratory chain, with no increase in the amount of respiratory protein measured — better-working machinery rather than more of it. The same flies survived starvation longer.
• The drug version did not work. A compound called BTM-3528, which switches on the same response in human cells, produced the expected changes in cultured fly cells. Fed to flies it did not extend life, and the target gene rose only slightly and only in the head — so it barely reached the tissues: a delivery failure rather than a test of the idea.

What to keep in mind

The authors list their own limits. The study shows that neuronal ATF4 is sufficient to extend lifespan, not that it is needed for the benefit of body-wide activation. The brain-derived signal that reaches the fat body has not been identified. The chromatin and jumping-gene changes are correlations, not demonstrated causes of the longer life, and gene activity in the one-week-neuron flies was measured only at old age, so what happens during the induction week itself is unknown. And, as with many longevity findings in flies, the lifespan effect appeared only in females, for reasons nobody can yet explain.

Two further cautions. This is a preprint: it has not been through peer review, and details may change. And the intervention is a genetic switch in a fly — there is no treatment here, and the one compound tried as a stand-in did not work.

Why it matters

Most work on aging asks which pathways to push. This one asks when, for how long, and where — and finds that the same push can add life or subtract it depending on the answer. That fits the old idea that a mild, brief stress can leave an organism better off while a sustained one wears it down, and gives the idea a measurable shape: one week, in neurons, in early adult life.

The other striking part is the reach. A signal confined to the nervous system for seven days left a different tissue in an altered state six weeks later, long after the signal itself had faded. Whatever carries that memory — the authors suspect a relay in which other transcription factors take over from ATF4 — it is the kind of mechanism that would have to exist for a brief intervention early in life to matter much later. Whether anything similar operates in mammals is an open question.