When a salamander loses a leg, it grows back. When a zebrafish loses a fin, it grows back. When a mouse loses a fingertip, it grows back, but only the very tip, and then the process stops cold. Three independent research teams have now discovered why all three animals begin the same repair sequence: a shared pair of genes, called Sp6 and Sp8, switches on in the skin covering a fresh wound within hours of amputation. The finding, published in the Proceedings of the National Academy of Sciences, suggests that mammals never lost the genetic instructions for regeneration. They just lost the ability to finish the job.
Three labs, three species, one answer
The study was a coordinated effort split across three organisms. Kenneth Poss’s lab at Duke handled zebrafish. Peter Currie’s group at the Australian Regenerative Medicine Institute contributed axolotl work, though Currie is primarily recognized for zebrafish and muscle biology research, and his precise role in the axolotl arm of the study has not been detailed in public statements. Kenneth Brown’s team at Wake Forest University focused on mouse digit tips. Each lab ran its experiments independently, then compared results.
The convergence was striking. In all three species, the transcription factors Sp6 and Sp8 activated rapidly in the wound epidermis, the thin layer of skin that seals over an injury site. In axolotls and zebrafish, that activation set off a cascade that rebuilt entire limbs and fins. In mice, the same initial program fired but produced only a partial digit tip before stalling.
The proof came from knockout experiments. When researchers used CRISPR to delete Sp8 in axolotls, regeneration failed entirely. In mice engineered to lack both Sp6 and Sp8 in their epidermal cells through conditional gene knockouts, digit tip regrowth also collapsed. The genes were not just correlated with regeneration. They were required for it.
Because axolotls, zebrafish, and mice sit on distant branches of the vertebrate family tree, the shared program almost certainly traces back to a common ancestor hundreds of millions of years ago. That means humans, who share the vast majority of their protein-coding genome with mice, likely carry the same dormant instructions.
A toolkit that already works across species
The Sp6/Sp8 discovery did not emerge in isolation. It fits into a growing body of work showing that regeneration is controlled less by unique “regeneration genes” and more by regulatory switches that determine when and where ordinary genes turn on.
A foundational 2016 study in Nature identified what researchers called tissue regeneration enhancer elements, or TREEs: stretches of non-coding DNA that activate only after injury and amplify pro-regenerative gene expression in zebrafish hearts and fins. Those enhancers acted like volume knobs, and critically, they could be engineered to drive either regenerative or anti-regenerative programs in targeted tissues.
A follow-up line of research pushed the concept further. Scientists showed that a zebrafish regeneration-associated enhancer controlling a growth factor called hb-egfa during spinal cord repair could be transplanted into neonatal mouse tissue, where it successfully delivered pro-regenerative signals. That was early proof that fish-derived genetic switches are not locked to fish biology. They can function in mammalian cells.
Other pieces of the puzzle have fallen into place in parallel. Comparative single-cell atlases spanning axolotl, human, mouse, chicken, and frog tissues have allowed researchers to map which cell types express regeneration-associated genes across species. Separate work on the mTORC1 signaling pathway revealed that both axolotl limb cells and mouse digit cells activate mTORC1 after amputation, but the downstream protein translation diverges sharply, helping explain why the same starting signal produces full regrowth in one animal and a stump in another. And earlier mouse studies established that a specific population of Lgr6-positive nail stem cells, responsive to Wnt signaling, is required for digit tip regrowth, defining the cell population that the Sp6/Sp8 program appears to coordinate.
Together, these findings sketch a coherent model. Injury triggers Sp6 and Sp8 in the wound epidermis. TREE-like enhancers amplify regenerative gene expression. Local stem and progenitor cells respond by forming a blastema, the mass of reprogrammed cells that rebuilds complex tissue. In zebrafish and axolotls, the full sequence runs to completion. In mice, it starts but breaks down partway through.
Why mammals stall
The central mystery is no longer whether mammals have regeneration genes. They do. The question is what blocks the later stages.
One major obstacle is structural. Axolotls form blastemas: organized clusters of dedifferentiated cells that can become bone, muscle, nerve, and skin. Mice do not form blastemas after digit amputation, and no laboratory intervention has successfully induced one. Without that intermediate structure, even a correctly triggered Sp6/Sp8 program may have no cellular architecture to work with. The genes can issue instructions, but there is no workforce to carry them out.
Scarring is another barrier. Mammalian wound healing prioritizes rapid closure over reconstruction. Fibroblasts flood the injury site, deposit collagen, and seal the wound with scar tissue, a process that physically walls off the kind of cellular reorganization regeneration requires. Immune cell behavior at the wound site, extracellular matrix composition, and mechanical forces all influence whether tissue remodels or simply scars over.
Oxygen may also play a role. Analysis published in Nature Lab Animal has highlighted oxygen-sensing capacity as a factor distinguishing regenerative from non-regenerative limb contexts. Low oxygen, or hypoxia, can promote stem-like cellular states, but sustained hypoxia damages tissue. How Sp6 and Sp8 interact with oxygen-sensitive pathways like HIF signaling has not been explored.
Developmental timing adds yet another layer. Some regenerative responses in mammals, including partial heart repair and limited spinal cord regrowth, occur only in newborns and vanish within days or weeks after birth. The neonatal mouse experiments with zebrafish enhancers suggest young tissues are far more permissive to pro-regenerative signals. Whether adult human tissues retain any comparable window of plasticity, or whether they would need extensive reprogramming to respond, remains unknown.
What this does not mean yet
No one has tested whether activating Sp6 and Sp8 in human epidermal cells triggers any regenerative response. The experimental evidence stops at mouse digit tips, the only structure mammals are known to regrow after amputation under normal conditions. Whether the same transcription factors could initiate regrowth of a full finger, let alone an arm, has no direct experimental support in any mammalian system beyond neonatal mice.
None of the three collaborating laboratories have announced clinical trial timelines, funding commitments, or human application plans. Institutional statements from Wake Forest University and Duke University frame the findings in terms of long-term research potential, not near-term therapy. Readers should understand human limb regrowth as a research direction that just gained a critical new clue, not a medical prospect with a timeline attached.
It is also worth noting what this discovery does not address. Other promising approaches to regenerative medicine, including bioelectric patterning research led by Michael Levin at Tufts University, 3D bioprinting of tissue scaffolds, and stem cell therapies, operate on different principles. The Sp6/Sp8 work identifies a genetic starting switch but does not resolve the downstream engineering challenges those other fields are tackling. A complete solution, if one ever arrives, will likely require insights from multiple approaches working in concert.
Why Sp6/Sp8 narrows the search but does not close it
The strongest evidence in this story is experimental and direct: removing Sp8 in axolotls stopped regeneration, and removing Sp6 and Sp8 in mouse epidermis stopped digit tip regrowth. Those are loss-of-function results published in a peer-reviewed journal with named methods and organisms. The cross-species enhancer work provides a second layer of primary evidence showing regeneration can be controlled through specific DNA switches that function across species boundaries.
The claims about human relevance rest on inference, not experiment. Because humans share most of their genome with mice, and because mice share the Sp6/Sp8 wound response with axolotls and zebrafish, researchers reason the same dormant program likely exists in human tissue. That logic is sound as a hypothesis. It has not been tested.
The mTORC1 findings offer a cautionary note even within that optimism. When mammals and regenerative animals share the same signaling pathway, the way cells interpret the signal can differ enough to produce opposite outcomes. Simply switching on the right genes may not be sufficient. The cellular context, including stem cell populations, protein translation machinery, scarring responses, and oxygen levels, will likely need to be addressed in parallel.
For now, the Sp6/Sp8 discovery does something specific and valuable: it narrows the search. Instead of asking whether mammals have any regeneration capacity at all, researchers can now ask what prevents a verified, conserved genetic program from running to completion. That is a more tractable question, and as of May 2026, it is the question driving the next round of experiments in all three labs.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
