Cut off an axolotl’s leg, and within weeks a knob of cells called a blastema will sprout from the wound and rebuild the missing limb, bone by bone, muscle by muscle, right down to the fingertips. Scientists have long known that these cells somehow remember which body part they are supposed to become. What they could not explain was how.
A team at Northeastern University, led by biologist James Monaghan, now points to a surprisingly simple answer: the speed at which cells chew up a single molecule. In a study published in Nature Communications in spring 2025, the researchers show that an enzyme called CYP26B1 breaks down retinoic acid, a derivative of vitamin A, inside the blastema. That breakdown creates a chemical gradient that tells cells near the wound tip to build fingers rather than upper arms. Block the enzyme, and those distal cells behave as though they have been flooded with retinoic acid, reprogramming themselves to produce shoulder-like structures instead.
The finding lands during a fast-moving stretch of axolotl research that is reshaping how scientists understand positional memory, the biological system that ensures a regrown limb comes back in the right shape. As of May 2026, at least three other studies have filled in neighboring pieces of the puzzle, together sketching a layered molecular map that guides regeneration along every axis of the limb.
One enzyme, one gradient, one switch
The core experiment is elegant in its directness. Monaghan’s group amputated axolotl limbs at different points along the shoulder-to-fingertip axis, then chemically inhibited CYP26B1 in the resulting blastemas. Without the enzyme to clear retinoic acid, cells that should have known they were “distal” (near the hand) instead adopted “proximal” identities (near the shoulder), effectively mimicking what happens when researchers drench tissue in excess retinoic acid.
The team also identified a gene called Shox that responds to shifts in the retinoic acid gradient, acting as a molecular translator between a chemical concentration and a specific limb segment. The U.S. National Science Foundation, which funded the work, highlighted the results as a step forward in understanding how retinoic acid breakdown controls limb patterning.
Put plainly: the difference between regrowing a hand and regrowing a shoulder may come down to how fast cells destroy a single vitamin A byproduct.
A second axis, a self-sustaining loop
Retinoic acid degradation addresses the proximal-to-distal question (shoulder versus fingers), but a limb also has a front and a back. A separate study, published in Nature, tackled that anterior-to-posterior axis by mapping a feedback circuit between two signaling molecules, Hand2 and Sonic hedgehog (Shh), in axolotl connective tissue cells.
Using transgenic reporters and lineage tracing, the researchers demonstrated that an initial burst of Shh activity triggers Hand2 expression, which in turn sustains Shh, locking posterior identity into place long after the original signal fades. The role of Sonic hedgehog in limb development has been recognized for decades, but this work reveals how a self-reinforcing loop preserves that positional information as a durable molecular memory, not just a fleeting cue.
Fine-tuning at the fingertip
A third study, published in npj Regenerative Medicine in 2026, zoomed in further to examine hedgehog signaling during axolotl digit regeneration. By perturbing the pathway at specific developmental stages, the authors showed that hedgehog activity is required for correct digit number and spacing, tying a well-known developmental signal to the fine-grained architecture of individual fingers and toes. The paper also cataloged which patterning mechanisms are conserved across amphibian species and which diverge at the digit scale.
Taken together, the three studies outline a tiered system: retinoic acid degradation sets the broad proximal-to-distal address, the Hand2-Shh loop maintains front-to-back identity, and hedgehog signaling sculpts the details of each digit.
Regeneration is not just development on replay
One of the more striking recent findings comes from a separate line of work in newts. Researchers at Caltech studied Pleurodeles waltl newts carrying mutations in FGF10, a growth factor gene critical for building limbs during embryonic development. The mutant animals grew up with severely malformed legs. Yet after amputation, their hindlimbs regenerated normally. The results were reported in PNAS, though a direct link to the paper is not available at this time.
That result matters because it suggests the regeneration program is not simply a rerun of fetal limb construction. The instructions for building a limb from scratch and for rebuilding one after injury only partially overlap, leaving room for regeneration-specific pathways like the retinoic acid and hedgehog circuits described above. If regeneration can override developmental errors, the minimal set of molecular signals truly required for regrowth may be different, and possibly smaller, than scientists assumed.
The gaps that remain
The biggest unanswered question is whether any of these triggers can be activated in mammals. Humans carry genes for retinoic acid signaling and Sonic hedgehog, and mammalian embryos use both during limb development. But no published data so far show that manipulating CYP26B1 or the Hand2-Shh circuit produces regenerative outcomes in mammalian tissue. Every experiment to date has been performed in salamanders and newts, animals with an unusually high capacity for regrowth. Claims about future human therapies remain speculative.
A related gap involves the nervous system. A study published in Communications Biology introduced a simplified assay called CALM and linked nerve-derived signals to changes in a chromatin mark (H3K27me3) during the earliest stages of patterning competence in axolotl limb cells. That raises the possibility that nerves prime cells to respond to retinoic acid and hedgehog cues by opening regions of the genome that contain key patterning genes. But no published work yet integrates the neural step with the retinoic acid degradation mechanism or the Hand2-Shh loop in a single experiment. Whether these systems operate in sequence, in parallel, or through some more complex arrangement has not been resolved.
There is also the question of memory between injuries. Adult salamanders can regenerate limbs repeatedly, which implies that cells either maintain a stable record of their coordinates or can rapidly reconstruct it after each amputation. The Hand2-Shh loop offers one candidate for long-term storage along the front-to-back axis, while differential retinoic acid degradation could, in principle, be re-established from scratch based on tissue geometry. Direct evidence for how positional maps are archived in uninjured limbs, and how they are recalled when a blastema forms, is still thin.
Where the molecular logic of regrowth stands in mid-2026
The strongest claims in this body of work rest on peer-reviewed experiments that use genetic perturbation and lineage tracing, methods designed to test whether a gene is necessary and sufficient for an outcome rather than merely correlated with it. By selectively blocking CYP26B1 or altering Hand2 and Shh activity, researchers watched limbs regenerate with specific segments duplicated, missing, or transformed. That kind of causal evidence is unusually clear for a process as complex as limb regrowth.
Contextual sources, including an NSF summary of the Monaghan lab’s work and news coverage from outlets like Nature and Phys.org, provide accessible framing but do not add independent experimental data. They are useful for confirming researcher identities and funding, not for predicting clinical timelines.
The most defensible takeaway, as of May 2026, is this: limb regeneration in amphibians depends on a coordinated set of chemical gradients and feedback circuits that assign positional identities along multiple axes. Retinoic acid degradation via CYP26B1 specifies the shoulder-to-fingertip fate. The Hand2-Shh loop stabilizes front-to-back orientation. Hedgehog signaling shapes individual digits. Nerves and growth factors modulate the overall readiness of cells to respond. How much of this framework will carry over to species with limited regenerative abilities remains an open and pressing question. But the current data represent a genuine shift: from viewing regeneration as an inscrutable biological trick to seeing it as a problem built from interpretable, and potentially engineerable, molecular logic.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
