How Salamanders Regrow Limbs: The Strange Biology of Vertebrate Regeneration

Among vertebrate animals, the salamander is the regeneration champion. An adult axolotl can regrow a severed limb with full bone, muscle, nerves, and skin in a few weeks, and the replacement is anatomically identical to the original. Newts can regrow eyes, including the lens and retina. Some species can regenerate sections of heart muscle, parts of the spinal cord, jaw, and even the lens of the eye after multiple removals. No mammal can do any of these things to any meaningful degree. A child who loses a fingertip past the last joint will regrow some tissue at the very tip; an adult who loses any larger piece of any limb does not regrow it. The salamander stands out as an exception that has fascinated developmental biologists for two centuries, both for its own sake and for what it might teach about why the rest of us cannot do the same.

The blastema

The mechanism that makes salamander regeneration possible is a structure called the blastema. When a limb is severed, the cells immediately around the wound do something that adult mammalian cells do not. They dedifferentiate. The mature muscle cells, bone cells, cartilage cells, and connective-tissue cells in the stump undergo a partial reversal of their developmental commitment, becoming a population of multipotent progenitor cells that resemble the cells of the embryonic limb bud. This population accumulates at the wound site as the blastema, and over the following weeks the blastema cells re-execute the developmental program that built the original limb in the embryo, producing all the tissue types of the new limb in the correct anatomical arrangement.

The blastema is the central object of regeneration biology. The molecular signaling that creates it, maintains it, and drives the differentiation of its cells back into mature limb tissues is the subject of decades of careful research. The work of Jeremy Brockes, David Gardiner, Susan Bryant, James Monaghan, and many others has identified key signaling molecules — fibroblast growth factors, bone morphogenetic proteins, retinoic acid, Wnt pathway components — that regulate the process. The 2018 axolotl genome sequencing by Nowoshilow and colleagues, and the 2021 single-cell transcriptomic atlases of regenerating limbs, have begun to identify which cells in the blastema correspond to which cells in the original limb development program.

Why mammals cannot do this

The standard mammalian response to a deep wound is scarring. Connective-tissue cells called fibroblasts proliferate at the wound site, deposit large amounts of collagen, and produce a dense fibrous tissue that closes the wound but does not reproduce the original tissue structure. The scar is functional in the sense that it stops bleeding, prevents infection, and restores some mechanical integrity, but it is not a continuation of the original tissue program. A scarred section of skin does not regrow hair follicles, sweat glands, or full sensory innervation; a scarred section of heart does not regain contractile function.

The contrast with the salamander is striking. After a salamander limb is severed, the wound closes within hours but the cells underneath do not scar. They form an apical epithelial cap, a structure that is functionally equivalent to the apical ectodermal ridge that drives limb development in vertebrate embryos. The cap signals to the underlying mesenchymal cells, which dedifferentiate to form the blastema, and the regeneration program proceeds.

The molecular question of why mammals scar where salamanders regenerate is partially answered. Mammalian fibroblasts respond to wound signals by producing collagen and inflammatory cytokines that recruit immune cells; the immune response and the fibroblast response together produce the scar. Salamander fibroblasts respond to the same wound signals by dedifferentiating and reentering a developmental program. Why the cellular response differs is the subject of ongoing research; the answer involves both the signaling molecules present at the wound site and the response of the cells to those signals, and probably also the immune-system context — salamanders have a more limited inflammatory response than mammals do.

The fetal regeneration window

One of the more interesting observations in the field is that mammalian fetuses can regenerate. A fetal mouse or fetal sheep with a limb amputated in utero can regrow it. The capacity is lost over the course of fetal development; by birth it is essentially gone, and the response shifts to scarring. The genetic and developmental machinery for regeneration is present in the mammalian genome — it has to be, for limb development to work in the first place — but adult mammalian tissues have lost access to it.

This finding shifts the framing of the research question. The question is not "what does the salamander have that mammals lack?" — both have the same fundamental developmental machinery. The question is "what blocks the developmental machinery in adult mammalian tissues, and can the block be released?"

The MRL mouse and the human fingertip

The MRL/MpJ mouse strain, identified at the Wistar Institute in the 1990s, has an unusual ability to heal ear punches without scarring and to regenerate ear tissue with relatively complete anatomy including cartilage and hair follicles. The strain was originally bred for autoimmune research, and the regeneration capacity was discovered accidentally when researchers noticed that ear-tagged mice from that strain had healed cleanly while mice from other strains had not. The mechanism is not fully understood, but it provides one of the clearest indications that adult mammals can have substantial regenerative capacity if the right combination of factors is present.

Human fingertip regeneration is the other indication. Children younger than about 11 who lose the tip of a finger past the last joint can regrow it, including the nail and bone, if the wound is left to heal without surgical closure. The capacity is real but limited — it does not extend to larger amputations or to adults — and it suggests that the developmental machinery for regeneration is not entirely absent from mammals; it is conditionally accessible.

The applied research surface

Research on salamander regeneration has obvious applied interest. If the molecular block that prevents adult mammalian limb regeneration could be identified and removed, the medical implications would be substantial. The current state of the field is that the cellular and molecular components are increasingly well-characterized, but the integration into a working therapeutic intervention is many years away. The closest analogues in current clinical practice are tissue engineering approaches that use scaffolds and stem cells to encourage limited regeneration of specific tissues, and these are far short of the salamander capability.

The 2024 publication by Lin and colleagues in Nature on cross-species regenerative-program reactivation, the work of Kenneth Poss at Duke on heart regeneration in zebrafish, and the parallel research programs on lizard tail regeneration, planarian regeneration, and Hydra regeneration are all part of a broader research effort to map the diversity of regenerative capacities across the animal kingdom and identify the conserved and divergent mechanisms.

The deeper observation

The salamander regeneration story is in part a story about how much of biology is contingent rather than necessary. There is no fundamental physical or chemical reason that mammalian tissues cannot regenerate; the capacity is in the genome and is exercised in the embryo. The adult mammalian inability to regenerate is a developmental and evolutionary outcome rather than a biological law. Mammalian evolution selected for fast wound closure and infection prevention over slow but anatomically perfect regeneration, probably because the trade-off favored survival in the mammalian niche. The salamander niche apparently favored the opposite trade-off, and the salamander retained the capacity that the mammalian lineage gave up.

What the research program is slowly converging on is the recognition that regeneration is not an exotic capacity belonging to a few unusual species but a latent capacity present in most vertebrates that has been suppressed in particular lineages for particular reasons. Releasing the suppression in a controlled and safe way is among the harder problems in current biology, but it is a problem with a path forward rather than a problem with no path at all.