The Onset of Breathing at Birth Switches Off Regenerative Capacity

TL;DR: The authors make a bold, simple claim — the first breath after birth acts as a body-wide "off switch" for growth signaling. Breathing blows off CO₂, which alkalinizes the inside of cells within a minute, which activates the kinase SNRK, which shuts down mTOR. Because mTOR is the master correlate of axon-regrowth competence, this single physiological event may be why mammals stop being able to regenerate shortly after birth. The striking part: it's reversible. Making adult mice breathe extra CO₂, or deleting Snrk, turns mTOR back on and regrows injured optic-nerve axons.

Why this is a strange and interesting question

We've known for a long time that regenerative capacity collapses in early postnatal life. Neonatal mouse spinal cord heals scar-free; a week later it doesn't. Embryonic retinal ganglion cells (RGCs) sprout axons vigorously; postnatal ones barely do. And mTOR activity — the signaling hub that licenses anabolic growth — drops sharply between embryo and adult. The standard story treats this as a gradual, programmatic maturation: transcriptional changes, chromatin closing down, growth genes getting switched off over days.

The authors ask a sharper question: what if it isn't gradual at all? What if a discrete event flips the switch? Birth is the most violent physiological transition in a mammal's life — within minutes the animal goes from placental life support to breathing, feeding, and thermoregulating on its own. Could one of those transitions be the trigger?

The first measurement already reframes everything. Phospho-S6 (p-S6), the standard readout of mTOR activity, is present in most RGCs before birth. Within five minutes of a C-section delivery it is nearly gone, and it stays gone for at least 12 hours. The same collapse happens in brain, heart, and skeletal muscle, and it hits p-S6, p-S6K, and p-4EBP1 — three downstream mTOR targets — while total protein is unchanged. That last detail matters: if total protein is constant and only phosphorylation drops, this is not transcription or translation. It is an acute signaling event, fast enough to be a switch, not a program.

Finding the trigger by subtraction

Birth bundles several changes together: loss of umbilical nutrients, onset of breathing, temperature drop. The elegant part of the paper is how cleanly they pull these apart, using the fact that a C-section lets you independently control the cord (nutrients) and the amniotic sac (breathing).

Keeping pups at 37 °C didn't preserve mTOR — temperature is out. Keeping the umbilical cord intact didn't preserve it either, and blood glucose wasn't even reduced in the first five minutes — nutrients are out. But pups left inside the intact amniotic sac, unable to breathe air, kept their mTOR activity high for up to 30 minutes, regardless of whether the cord was cut. The moment a pup breathed air, p-S6 vanished — the decline starting ~4 minutes after the first breath and essentially complete by 5.

So it's breathing. And then one more subtraction nails the chemistry. Breathing does two things at once: raises blood O₂ and dumps CO₂. Deliver the pup into 10% O₂ and mTOR still collapses. Deliver it into 10% CO₂ and mTOR is preserved. It's the loss of CO₂, not the gain of oxygen, that throws the switch.

The first breath as a chemical off-switch

In utero high CO₂ mTOR ON → grow breathe → CO₂ exhaled pH inside cell ↑ SNRK ON mTOR OFF regeneration lost adult: inhale CO₂ → pH ↓ → SNRK off → mTOR ON again

Womb is a high-CO₂ world; growth programs are tuned to it. Blowing off CO₂ is the signal that growth-time is over. Same chemistry runs backward in adults — and across heart, brain, and muscle, not just retina.

The core logic. Losing CO₂ raises intracellular pH (the CO₂/HCO₃⁻/H⁺ equilibrium shifts, consuming protons), SNRK reads the alkalinization, and mTOR shuts off. Adding CO₂ back acidifies and reverses it.

The kinase, and watching it fire in a newborn's eye

Why would losing CO₂ suppress mTOR? The authors had an independent clue: a CRISPR kinase screen across more than a thousand kinase/phosphatase targets after optic-nerve crush flagged SNRK — an unusual AMPK-family kinase — as the single strongest negative regulator of axon regeneration (32 hits met criteria; SNRK was the standout). AMPK-family kinases are classic mTOR brakes, often acting through the TSC complex. And indeed, knocking out Tsc1 in RGCs preserved mTOR across birth, placing TSC in the path.

Delete Snrk in embryonic RGCs and they keep their p-S6 at 5 minutes after birth, at embryonic levels. The switch depends on SNRK.

The technical centerpiece is showing this live. They adapted transpupillary two-photon imaging to newborn mice, delivered by C-section and imaged in a custom holder while tactile stimulation triggered the first breaths (breathing frequency climbing from 7 to 45 breaths/min over five minutes). Using the ratiometric reporter ExRai-AMPKAR, they watched AMPK-family activity in individual RGCs rise 2–3 minutes after the first breath and keep climbing until ~8 minutes — temporally preceding, and then overlapping, the mTOR collapse. Pups that never breathed showed no change. In Snrk-knockout RGCs the rise was largely abolished. Delivering into 10% CO₂ blocked it; 10% O₂ did not.

Then they chased the mechanism one layer deeper with more biosensors. ATP stayed flat — so this is not energy stress, consistent with SNRK lacking the γ subunit that makes canonical AMPK an AMP/ATP sensor. Calcium stayed flat for the first 4 minutes, then rose. But intracellular pH, read with the FRET sensor pHLemon-rev, rose within ~1 minute of breathing and plateaued by 5 — and 10% CO₂ abolished it. The ordering is the argument: alkalinization comes first (the early trigger), calcium rises later (a likely amplifier for the sustained phase).

The pH→SNRK→mTOR chain plays out in minutes at birth but hours in the adult — the authors attribute the slowdown to postnatal maturation of ion/pH homeostasis. The ordering (pH first, kinase next, mTOR last) is preserved, which is their main causal argument.

The payoff: it's reversible in adults

If the switch is just chemistry, you should be able to run it backward. Put adult mice in 10% CO₂ and intracellular pH in RGCs drops within ~2 minutes; SNRK/AMPK activity falls after ~1 hour; p-S6 climbs by 4 hours and reaches ~70% p-S6-positive RGCs by 48 hours. The same slow, stepwise ordering, just dilated in time. Deleting Snrk in adult RGCs likewise reactivates mTOR and, after optic-nerve crush, produces robust axon regeneration.

And the headline functional result: adult mice kept in 5% CO₂ from a week before injury through two weeks after showed significantly more CTB-labeled regenerating axons past the crush site than air controls (N = 7 air vs 10 CO₂). CO₂ also reactivated mTOR in brain, heart, and skeletal muscle — hinting this is a body-wide knob, not an RGC quirk.

The authors can't resist the evolutionary speculation, and it's a fair one: amniotes (mammals, birds) are lung-breathers with poor regeneration, while fish and salamanders regenerate beautifully. Maybe committing to air-breathing, CO₂-dumping extrauterine life was a trade — metabolic independence bought at the cost of regenerative competence.

What to believe, and what to push on

The subtraction logic is genuinely strong. The amniotic-sac experiment and the CO₂-vs-O₂ swap are clean, causal, and hard to explain away — breathing, specifically CO₂ loss, is the trigger for the mTOR collapse, and p-S6 is read by an antibody that has nothing to do with the biosensors. The Snrk and Tsc1 genetic experiments anchor the pathway independently. That core is convincing.

Where I'd apply pressure is the middle of the chain, the live-imaging mechanism. The authors are admirably upfront that ExRai-AMPKAR is itself pH-sensitive — its signal rises as pH goes from 6.8 to 7.4, which is exactly the direction and range of the alkalinization they're measuring. So the "AMPK-F activity" trace is partly entangled with the pH change it's supposed to be downstream of. Their defenses are the phospho-dead mutant (flat) and the Snrk knockout (signal largely gone), and those genuinely help — but they concede the phosphorylated sensor could have stronger pH dependence than the mutant. The precise ordering of pH-then-kinase, which carries the causal argument, rests partly on sensors measured in the same cells under the same pH swing. Worth watching.

Other caveats: how alkalinization actually activates SNRK is inferred (from yeast Snf1) rather than shown biochemically; the TSC and SNRK contributions aren't fully reconciled; and the regeneration from CO₂ alone, while significant, isn't benchmarked against the strong Pten/mTOR manipulations this same lab pioneered, so we don't know if it's a comparably large effect or a modest one. There's no functional vision recovery, chronic 5% CO₂ is a whole-body perturbation with its own confounds, and this is all mouse. The "general across all tissues" claim rests on Western blots and a few extended-data panels.

Still — the shape of the idea is what makes this worth your time. A single, timed, chemical event that flips a conserved growth switch across the body, running on the same CO₂/pH physiology that also gates seizures and sperm capacitation, and that stays reversible for life. If even the RGC portion holds up, "modulate CO₂/pH to reopen a regenerative window" becomes a testable, almost embarrassingly simple therapeutic hypothesis.

The sections to read closely: the CO₂/O₂ dissociation (Fig. 4) for the cleanest causal evidence, the newborn two-photon imaging and the two-phase pH/calcium model (Figs. 3 and 5), and in Methods, the biosensor pH-sensitivity controls — that's where the mechanism is most load-bearing and most contestable.