Axolotls: Re-accessing Embryonic Programming Mechanisms for Regeneration

Introduction to Ambystoma Mexicanum, The Axolotl

Photograph by Stephen Dalton/Animals Animals—Earth Scenes, National Geographic

The Ambystoma mexicanum commonly referred to as Axolotl is an amphibian part of the order Caudata/Urodela known for its unique characteristics in development and regeneration. Axolotls are native to Lake Xochimilco and Lake Chalco in Mexico. Both Lake Xochimilco and Lake Chalco have been drained due to expansion, however, axolotls do exists in small numbers in these areas. Axolotls have a condition called neoteny, characterized by the retention of larval features throughout adult life. Axolotls retain their tadpole-like dorsal fin and gills and never produce lungs, and are therefore restrained to water their whole lives. Neoteny has been found in other amphibians besides axolotls and is usually accompanied with an iron deficiency in the organism. Iron is an essential element in the production of thyroxine hormones necessary for growth, development, and random genetic mutations. However, juvenile axolotls are still able to reach sexual maturity in their larval stage, which is a phenomena called perennibranchiate.  However, metamorphosis can be induced artificially by injecting hormones into axolotls.   In rare cases, BBC news and other studies have reported this metamorphosis can occur spontaneously [1]. In metamorphosis axolotls lose their gills while growing lungs and eyelids, which gives them an evolutionary land dwelling potential. Due to their intriguing scientific characteristics, axolotls have been interesting models for development due to neoteny, their unique phases of embryogenesis, and their ability to regenerate organs of the body. Here we examine axolotls ability to re-access embryonic processes during regeneration.


Endo et al., 2004

The regeneration of limb buds and the spinal cord in developing embryos occurs in all vertebrates during embryogenesis.As the embryo develops, the organism fails to maintain this regenerative potential, and by birth this regenerative capability is lost. (Gardiner et al., 2005) However, axolotls maintain their regenerative potential beyond their embryonic/larval period. Therefore, making axolotls the only species known to regenerate complete postnatal limbs and organs.  Injured sites of axolotls return to an embryonic-like developmental state and follow the Accessory Limb Model, recently developed in 2004 by an analysis of dermal cell migration and blastema formation (Endo, 2004), to regenerate the lost limb.

Figure 1: Endo et al., 2004

The Accessory Limb Model explains the steps in limb regeneration. The first event is re-epithelialization to cover the wound surface for wound healing without scar formation (Figure 1A). Next, only if a nerve is deviated to the site of the wound, additional signals cause dermal fibroblasts at the wound periphery to dedifferentiate and migrate toward the nerve to form the early blastema (Figure 2A). Limb development from the blastema is dependent on dermal cells from the opposite side of the limb grafted to the wound site (Figure 1B & 2B) (Kragle et al.,2009).

Figure 2: Endo et al., 2004

Then a limb is regenerated from the blastema (Figure 3). This process shows axolotls capability to re-access the regenerative program of the earlier developed embryo. Therefore it is hypothesized that the axolotls regenerating tissues are embryonic-like and give rise to postembryonic tissues that are equivalent to tissues developed at the end of embryogenesis. Promoters and inhibitors of the Accessory Limb Model are being studied to create therapeutic interventions to enhance regenerative abilities in other species such as humans. (Kragl et al., 2009)

Figure 3: Gardiner et al., 2004

Howard Hughes Medical Institute (HHMI) lecture on the Accessory Limb Model

Exploring the Blastema in Regeneration in Axolotls

A controversial issue in axolotl regeneration lies within the origin of the limb: the blastema. Histological results have shown the blastema as a homogeneous group of pluripotent cells capable of differentiating into multiple lineages. In fact, many studies have suggested cells that occur in the blastema are somehow reprogrammed to be multipotent or pluripotent like pre-embryogenesis cells in the limb blastema. (Ulmanski et al., 1938; Echeverri et al., 2002)

Other studies have proposed axolotl muscle and Schwann cells have the capability to dedifferentiate into multipotent and pluripotent states (Lo et al., 1993; Morrison et al., 2006) and (Wallace et al., 1972). For this to occur, cell memory of dedifferentiated cells must be erased, which would allow the cell to redifferentiate into different cell lineages from which they originated. In 1973, Wallace and Wallace implanted nerve pieces into irradiated host and showed that Schwann cells and fibroblasts could differentiate into dermis, blood vessels, cartilage and muscle cells. Therefore, implying an evolutionary ability for the dedifferentiation of cells from different lineages into blastemal cells followed by a redifferntiation into tissues unlike those from which they originated (Wallace et al., 1972). Other studies have implied dedifferentiation of specifically muscle into a multipotent progenitor. These muscle experiments tracked dextra-injected newt myotubes implanted into the limb blastema finding the label in cartilage (Lo et al., 1993). BrdU was also used to track cultured newt muscle satellite cells in vivo where the labeled cells were tracked in cartilage, muscle, and epidermis of the fully regenerated limb (Morrison et al., 2006). These findings were recently tested and challenged by using integrated green fluorescent protein trans-gene in the axolotl in order to elucidate the mechanism of dedifferentiation to a pluripotent or multipotent state. However, the results showed no indication of cells being reprogrammed to a complete pluripotent state, but instead the results showed the cells have memory of their origin and return to their origin cell’s fate. Here, we review this study that indicates that cells of the limb blastema do not become pluripotent, but retain a memory of their tissue origin. Therefore, implying a heterozygous blastema in limb regeneration of axolotls.

Determining Redifferentiated Potential of Axolots

The study “Cells Keep a Memory of Their Tissue Origin during Axolotl Limb Regeneration” determines the pluripotency potential of axolotl tissue (Kragl et. al, 2009). First, this study explores the lineage of dermis after dedifferentiation. Each major limb is tagged with grafting the embryonic region that produces that limb tissue from GFP+ transgenic donors into GFP- host embryos or by direct grafting of GFP+ limb tissue into a GFP- unlabeled host in order to follow the dedifferentiated cells through the healing and regeneration of the limb. In order to control the findings of this experiment, no chimaeric expression in the regenerate to indicate signaling was found.

Redifferentiation Potential of Dermal Cells

As previously described in the Accessory Limb model (Kragl et al., 2009), epidermal cells migrate over the stump immediately after amputation. To determine pluripotent potential and actual fate of dermal cells, skin grafts of GFP+ donors were transplanted into GFP- hosts at the amputation site (Fig 4A). The skin grafts generated a large amount of GFP+ cells in the regenerated limb (Fig 4B). To determine if dermal cells were redifferentiating into skeletal muscle during regeneration, GFP+ skin transplant cells from the newly formed limb were analyzed for a skeletal muscle marker, PAX7. PAX7 immunofluorescence showed no overlap between PAX7+ and GFP+ cells indicating dermal cells were not differentiating into muscle cells as previously shown. Further more, single cell polymerase chain reaction of GFP+ blastema cells showed no expression of the skeletal muscle Myf5, another skeletal muscle marker, whereas 31% of the control skeletal muscle-derived blastema cells expressed Myf5 (Fig 6d). Unlike muscle, cartilage cells formed from cleanly labeled dermis limbs(Fig 4g).  However, GFP+ Schwann cells were not found in the regenerated limb. These results conclude dermal cells have the ability to differentiated into cartilage, but not Schwann or muscle cells.

Figure 4. Dermis makes cartilage and tendons only. A) Schematic of experiment. Full thickness skin from GFP+ transgenic 8-cm-long donors was grafted onto GFP- hosts. B) Time course immunostanined for MHCI(muscle-specific myosin heavy chain). Skin grafts generated a large amount of GFP+ cells in the regenerated limb. C) Immunostained for PAX7,longitudinal section of 12-day blastema., no overlap between GFP+ and PAX7+. Showing dermal cells did not enter the myogenic lineage. D and E) Cross-section of regenerated limbs. GFP+ cells were negative for muscle markers Myf5, PAX7, and GFP. F-H Longitudinal cross sectionsimmunostained for anti-MHCI. GFP cells contributed to connective tissue (F), tendons (G), and cartilage (H). Blue is DAPI and scale bars b,0.5mm;c-h 50μm. (Modified) Kragl et al. Nature 460, 60-65 (2009)

Redifferentiation Potential of Cartilage

Pieces of GFP+ upper-arm cartilage was used to replace a section of cartilage in GFP- before amputation. GFP+ cells contributed to the blastema and the regenerated cartilage (Fig 5b).  Similarly to dermal cells results, cartilage progenitors lacked myogenic markers (PAX7, GFP, and Myf5) showing no contribution to the regeneration of muscle (Fig5c and 6d).  The GFP+ cells were found in tendons, perichondrium and dermis (Fig. 5d). Suggesting that cartilage cells only have the potential to redifferentiate back to cartilage which further suggests these cells have memory of their origin which challenges earlier findings described above.

Figure 5. Cartilage does not make muscle; M Kragl et al. Nature 460, 60-65 (2009) doi:10.1038/nature08152; Cartilage does not make muscle. (A) Schematic of experiment. To following limb cartilage, pieces of GFP+ upper-arm cartilage replaced cartilage of GFP- hosts before amputation. B) Time course immunostanined for MHCI(muscle-specific myosin heavy chain). C) Immunostained for PAX7, longitudinal section of 12-day blastema., no overlap between GFP+ and PAX7+. D) Longitudinal section of regenerated limb 30 days post amputation immunostained for MHCI(red). Majority cells found in regenerated skeleton; no signal was found in muscle. Blue is DAPI (b-d) and scale bars b,0.5mm;c, 50μm; d, 100 μm. (Modified) Kragl et al. Nature 460, 60-65 (2009)

Redifferentiation Potential of Muscle

Muscle cells also displayed a memory, differentiating back into muscle cells. No markers for cartilage or epidermis were found in longitudinal sections representing labeling of 25,000 nuclei in GFP+ muscle fibers (Fig 6e). These results indicate that even after dedifferentiation muscle cells are restricted to redifferntiation in the skeletal muscle lineage.

Figure 6. Muscle does not make cartilage or epidermis; M Kragl et al. Nature 460, 60-65 (2009) doi:10.1038/nature08152; Muscle doesn’t make cartilage or epidermis. A) Schematic of experiment. GFP+ embryonic presomitic mesoderm to label limb muscle fibers and satellite cells. B) Time course of GFP+ cells in regenerated limb. C) Immunostained for PAX7, longitudinal section of 12-day blastema, GFP+ were positive for PAX7. D) Single cell PCR GFP+ muscle-derived blastema cells expressed Myf5, but GFP+ cells from others tissues did not express Myf5. mRNAs for these genes were normalized to expression of RP4. Numbers of cells/blastemas/animals/experiments analyzed were as follows for each tissue. Skeleton: 152/8/8/4, Schwann cells: 402/6/6/6, dermis: 230/12/12/6, muscle: 184/6/6/3. D) Longitudinal section of regenerated limb 30 days post amputation. No GFP+ cells were found in cartilage or epidermis (above dotted line). Blue is DAPI (merged figures) and scale bars b,0.5mm;c, 50μm; d, 100 μm. (Modified) Kragl et al. Nature 460, 60-65 (2009)

Redifferentiation Potential of Nerves

Dedifferentiated Schwann cell fate was also explored with the GFP labeling transplant experiments. GFP+ neural fold was grafted using white mutant embryos as hosts and donors. In these grafting experiments, restriction of Schwann cells was also shown. GFP expression was only found in nerve tracts,and no expression of cartilage or muscle was found (Fig 7d and e).

Figure 7. Schwann cells only give rise to Schwann cells; M Kragl et al. Nature 460, 60-65 (2009) doi:10.1038/nature08152; Schwann cells only make Schwann cells. A) Schematic of experiment. GFP+ neural fold was grafted using white mutant embryos as host and donors. B) Confocal fluorescence image of GFP+ Schwann cells in hand to ensure specific labeling. Close association of the GFP signal with neuronal-specific BIII-tubulin and glia-specific myelin basic protein(MBP). C) Time course of GFP+ cells in regenerated limb. Schwann cells enter distal tip at day 25 post amputation. D) Longitudinal section of GFP+ Schwann in regenerated hand closely associated to BIII- tubulin and MBP. E) Part D shown with DAPI F-J) Section of regenerating limbs that had been X-rayed and rescued by a non-irradiated nerve implant. F-H) Regenerate rescued by nerve with GFP-labeled Schwann cells. Cartilage was negative for both GFP and nuclear cherry. I and J) A regenerate rescued by nerve tissue with GFP+ expression. Cartilage cells where positive for GFP showing they were derived from non-Schwann cells. G and J) Higher magnification of regions in Fand I. H) Higher magnification of G) showing MBP and BIII-tubulin staining. Scale bars b,0.5mm;c, 50μm; d, 100 μm. (Modified) Kragl et al. Nature 460, 60-65 (2009)


The results discussed above challenge and disprove the previous idea of a homogeneous blastema filled with dedifferentiated pluripotent cells. This paper provides a method of labeling cells with GFP, which provides a much higher resolution in the examination of cell plasticity than the previous methods used to analyze the fate of dedifferentiated cell types. Due to the reliability of this high-resolution method, the previous results can be attributed to induced plasticity created by contamination of implanted cells.  However, the new results revealing cell’s memory of their tissue origin leave even more questions. The ability of axolotls to re-accesses the embryonic programs for tissue and organ formation without dedifferentiation to pluripotent states needs further exploration. Although, the Accessory limb model of axolotls does not reprogram dedifferentiated cells to a complete totipotent state, the results above show the cells involved at the amputation plane are undergoing reprogramming events and potentially reprograming the cells involved to a dedifferentiated state and maybe even a multipotent state. These reprogramming characteristics are analogous to induced pluripotent stem cells. Currently, induced pluripotent stem cells are criticized due to the remaining memory of tissue origin the cells still possess which drives them into specific lineages limiting their regenerative capabilities. However, the axolotl is a model of successful ‘induced pluripotent-like stem cells’ in tissue and organ regeneration. Therefore, further knowledge about axolotl cell dedifferentiation and redifferentiation without ever entering a complete totipotent state, but still re-accessing embryonic programs and successfully regenerate tissues and organs is needed. Axolotls provide us with an excellent model for understanding the limitations and potentially eliminating the current barriers of induced pluripotency, and axolotls are a great model for how an organism can re-enter embryonic processes at the site of injury to not only heal but also regenerate lost tissues and organs which is a crucial topic for in regenerative medicine.

Axolotls in Academia

University of Florida

University of Nottingham

University of California, Irvine


Echeverri, K., and E. Tanaka. “Mechanisms of Muscle Dedifferentiation during Regeneration.” Seminars in Cell & Developmental Biology 13.5 (2002): 353-60.

Endo, Tetsuya, Susan V. Bryant, and David M. Gardiner. “A Stepwise Model System for Limb Regeneration.” Developmental Biology 270.1 (2004): 135-45. Print.

Gardiner, David M. “Ontogenetic Decline of Regenerative Ability and the Stimulation of Human Regeneration.” Rejuvenation Research 8.3 (2005): 141-53.

Kragl, Martin, Dunja Knapp, Eugen Nacu, Shahryar Khattak, Malcolm Maden, Hans Henning Epperlein, and Elly M. Tanaka. “Cells Keep a Memory of Their Tissue Origin during Axolotl Limb Regeneration.” Nature 460.7251 (2009): 60-65.

Lo, D. C. “Reversal of Muscle Differentiation During Urodele Limb Regeneration.” Proceedings of the National Academy of Sciences 90.15 (1993): 7230-234. Print.

McCusker, Catherine, and David M. Gardiner. “The Axolotl Model for Regeneration and Aging Research: A Mini-Review.” Gerontology 57 (2011): 565-71.

Morrison, J. I. “Salamander Limb Regeneration Involves the Activation of a Multipotent Skeletal Muscle Satellite Cell Population.” The Journal of Cell Biology 172.3 (2006): 433-40. Print

Umanski, E. E. “The Regeneration Potencies of Axolotl Skin Studied by Means of Exclusion of the Regeneration Capacity of Tissues through Exposure to X-rays.” Bull. Biol. Med. Exp. USSR 6 (1938): 141-45. Print.

Wallace, H. “The Components of Regrowing Nerves Which Support the Regeneration of Irradiated Salamander Limbs.” Journal of Embryology and Experimental Morphology 28 (1972): 419-35.

Leave a Reply

Your email address will not be published. Required fields are marked *

8 − seven =