The Sleeping Switch

A Glove of Living Cells

The axolotl, Ambystoma mexicanum, is a creature that seems to have misread the rules. Most animals, when they lose a limb, replace it with a scar. The axolotl replaces it with the limb. It can do this not just once but repeatedly throughout its life, though the regenerated tissue becomes slightly smaller with each iteration. Beyond limbs, it can regenerate its tail, its heart, and portions of its brain. It does not heal so much as rebuild from memory.

That word—memory—is not a metaphor here. The cells that flood the wound site retain what developmental biologists call positional identity: each one knows where in the body it is, what it's supposed to become, and how much of it is missing. This positional memory is encoded in the cell's molecular machinery, in chemical gradients and gene expression patterns that were laid down during embryonic development and never fully erased. When the limb comes off, those cells don't forget. They pick up exactly where the blueprint left off.

For more than a century, scientists have asked a single question about this capability: which genes are running the program? Because if you can identify the switch, you can study what it does. And if you understand what it does, you can ask whether it exists anywhere else.

In April 2026, a research group led by Wake Forest University biologist Josh Currie published the answer—or at least the first, most important part of it—in the Proceedings of the National Academy of Sciences. The genes are called SP6 and SP8. They are transcription factors: proteins that switch other genes on and off. And they are not unique to the axolotl. They appear in zebrafish. They appear in mice. They appear in every mammal whose genome has ever been sequenced, which is to say they appear in us, dormant, waiting in the dark of every cell in the human body.

The discovery is not a cure. No human will regrow a hand next year because of this paper. But the paper does something subtler and more significant than a cure: it identifies the conserved genetic architecture that makes regeneration possible, and it demonstrates, for the first time in a mammal, that you can partially restore regenerative capacity by delivering one of regeneration's downstream signals via viral gene therapy. The question has shifted from can mammals regenerate? to why did they stop, and can we nudge them to start again?

What SP8 Does in the Dark

Transcription factors are biology's master switches. They don't perform any cellular function themselves. What they do is bind to specific sequences of DNA and either activate or repress the genes sitting downstream. SP8 is one of the Sp/KLF family of zinc-finger transcription factors, a group with deep evolutionary roots. What makes SP8 and its partner SP6 so interesting is where they turn on: in the wound epidermis.

Currie's lab spent years mapping gene expression in the regenerating axolotl. When they looked at which genes lit up in the wound epidermis in the hours after amputation, SP8 stood out. It wasn't just expressed; it was expressed specifically in the thin epidermal sheet that forms over the wound, the structure called the apical epithelial cap, which every biologist who has ever studied limb regeneration knows is absolutely essential for what comes next. Without that cap, the blastema fails to form. Without the blastema, there's no regeneration.

To test SP8's function directly, Currie's team used CRISPR gene editing to delete it from living axolotls. The result was clean and decisive: without SP8, the animals could not properly regenerate their limb bones. The wound closed, but the structural rebuilding failed. The architectural instruction was missing.

The team then looked at mice. Mice can, it turns out, regenerate the very tips of their digits under the right conditions—a limited capability that has been known since the 1970s, largely overlooked, rarely studied. When researchers looked at whether SP6 and SP8 were expressed in regenerating mouse digit tips, they were. The same genes, doing something at least partially analogous, in a mammal. When both genes were knocked out in mice using conditional genetics, digit regeneration was significantly impaired.

This is where the story turns from characterisation to intervention.

The third collaborating lab in the study, working on zebrafish, had previously identified an enhancer—a short piece of non-coding DNA that acts as a volume control for a nearby gene—that responds to tissue injury in fish fins and cranks up the expression of a growth-promoting signal called FGF8. FGF8 is one of the molecules that SP8 normally activates. If SP8 is the master switch, FGF8 is one of the lights it turns on.

The researchers packaged that zebrafish enhancer driving FGF8 expression into an adeno-associated viral vector (AAV) and injected it directly into the digit stumps of mice shortly after amputation. The treated digits, even in mice where SP6 and SP8 had been genetically deleted, showed 32% greater total bone volume and 24% greater regenerated bone length at 28 days compared to untreated controls. The viral vector was, in effect, bypassing the broken switch and delivering part of the downstream signal directly.

Partial recovery. Not full regeneration. But partial recovery in an animal that had, minutes before treatment, no regenerative capacity at all in those digits. That is a meaningful number.

What the Salamander Has Always Known

Regeneration biology has had a complicated century. The field has been revived, dismissed, and revived again—a cycle that tracks, roughly, with the tools available to study it. For most of the twentieth century, researchers knew that salamanders could regenerate and that mammals largely couldn't, but the molecular architecture beneath that difference was opaque. The genes involved were unknown. The signals were guessed at. The few scientists who pursued the question often found themselves marginalised from mainstream developmental biology.

The genomics revolution changed this. When it became possible to sequence transcriptomes, to see in fine detail which genes are expressed in which cells at which moments, the axolotl went from an interesting curiosity to one of the most molecularly legible organisms in experimental biology. Researchers sequenced its enormous genome (ten times larger than the human genome), built atlases of gene expression in regenerating tissue, and began to find conserved elements: sequences that exist, largely unchanged, across hundreds of millions of years of evolution.

Jessica Whited's lab at Harvard has been one of the most productive contributors to this modern wave. Her group has traced how amputation triggers body-wide molecular changes in the axolotl, how the sympathetic nervous system activates stem cells far from the injury site, how the organism coordinates a whole-body response to the loss of a single limb. When asked about the implications of this work for other organisms, Whited is careful but unambiguous: "I think it's going to inspire a lot of future work to try to figure out not just how this works in an axolotl, but also how it works in other systems."

The Currie study fits squarely into that project. The key word in his team's findings is conserved. Conservation in evolutionary biology means that a gene or a sequence has been maintained across species through selection pressure: it works, and losing it costs too much to be tolerated. The fact that SP8 and SP6 are doing something analogous in axolotls, zebrafish, and mice argues that they are part of a very old wound-response program, one that predates the divergence of fish and tetrapods. Mammals didn't lose these genes. They kept them.

Ken Muneoka's lab at Texas A&M has been working a parallel line of evidence. His team has published extensively on mammalian digit regeneration—the same limited capability that Currie's lab exploited—and has been pushing back for years against the conventional wisdom that mammalian regeneration is simply impossible. His work has shown that what mammals lack is not the genetic machinery for regeneration but the regulatory environment that activates it. The mechanical and biochemical context of a mammalian wound suppresses signals that would, in a salamander, tell the tissue to rebuild. Muneoka's recent work shows that regenerative failure in mammals can be rescued—that the capacity exists and can be coaxed forward.

What Currie's paper adds to this picture is the upstream identity of the relevant signals. Before, researchers knew that FGF8 and related growth factors were involved in regeneration. Now they know more precisely what turns those growth factors on, and they know that the genetic sequence that produces that switch is shared across species. That is a road map. It doesn't tell you how to get there, but it tells you, for the first time, where you're going.

The AAV gene therapy approach is itself significant. Adeno-associated viruses are among the most clinically advanced gene delivery vehicles in medicine. FDA-approved AAV therapies already exist for conditions ranging from spinal muscular atrophy to a form of hereditary blindness. The infrastructure for using AAVs in human tissue exists, is tested, is understood. When researchers at Wake Forest inject an AAV carrying a zebrafish FGF8 enhancer into a mouse's amputated digit and measure 32% more bone volume a month later, they are not working in a preclinical vacuum. They are working with tools that are already on a path toward human application.

The Ancient Instructions We Carry

Go back to the wound. Go back to that translucent sheet of cells sliding over the axolotl's stump in the first hours after the cut. What we now understand about that sheet—the wound epidermis, the apical epithelial cap—is that it is an organ of communication as much as of cover. It is the tissue that activates SP8. SP8 is what begins the cascade: the blastema, the progenitor cells, the rebuild.

In a mouse, in a human, that same sheet forms after a wound. It has to; without epidermal closure, the body dies. The sheet forms, the wound closes, and then the regenerative cascade does not follow. The cells that cover a human wound are doing something structurally similar to what the axolotl's cells are doing, but they're running a different program. The orchestra plays the same instruments; it just reads from a different score.

The score may be changeable. That is what Currie's paper argues, carefully, with evidence.

Human SP8 is not broken. Human SP6 is not broken. They are present, intact, embedded in the genome. What's missing is the regulatory context that would turn them on at the right level, in the right place, at the right time after injury. The zebrafish enhancer that drives FGF8 expression—the one that the team packaged into their AAV vector—is essentially a piece of regulatory DNA borrowed from another organism, one that still knows how to activate the cascade. It's not an alien instruction; it's a version of an instruction we once shared and then stopped listening to.

Josh Currie was precise about what the result means and doesn't mean. "We can use this as a kind of proof of principle," he said after publication, "that we might be able to deliver therapies to substitute for this regenerative style of epidermis in regrowing tissue in humans." Proof of principle. Not a promise. The path from mouse digit tip to human arm passes through many years of work, many safety studies, enormous complexity. The gap between regrowing bone volume in a two-millimetre mouse digit and regrowing a human forearm is not trivial. Nobody in this field is pretending it is.

But something has changed in what the question even is. For most of the last century, the question was whether mammalian regeneration was possible at the molecular level. That question is effectively closed. It is possible. The genes are present. The signals exist. The question now is one of context, regulation, delivery—and those are engineering problems as much as biological ones. Hard engineering problems, but the kind that yield to sustained attention and good tools.

The axolotl does not know it is magical. It loses its arm, the cells do what the genes tell them to do, and three to eight weeks later it walks away. The extraordinary thing is not that the axolotl can do this. The extraordinary thing is that we are realising, piece by piece, that the instructions for doing it are written in our genome too. They've been there since before there were salamanders, since before there were mammals, since whatever shared ancestor first learned to tell a wound: this is not the end, rebuild.

The axolotl never forgot. We just stopped listening. For the first time, there is reason to think we can learn to listen again.