Position-specific induction of ectopic limbs in non-regenerating blastemas on axolotl forelimbs

Abstract

Ectopic retinoic acid (RA) has been hypothesized to reprogram the positional identity of cells in developing and regenerating limbs to a single positional value corresponding to the posterior-ventral-proximal (PVPr) position on the limb. We tested this hypothesis by using RA to reprogram the information of blastema cells that were induced to form at different positions around the limb circumference. We observed that RA treatment of blastemas in anterior and dorsal locations, but not posterior and ventral locations, resulted in the induction of complete ectopic limbs. These position-specific differences in limb induction are probably due to differences in the positional disparity between the RA-reprogrammed blastema cells and the cells at the periphery of the wound. These observations are consistent with the hypothesis that RA treatment reprograms the information in blastema cells to the PVPr position on the limb, since anterior and dorsal positions have the largest disparity and posterior and ventral have the smallest disparity from the PVPr identity.

Keywords: Axolotl; limb regeneration; positional information; reprogramming; retinoic acid.

Figure 1

Treatment of a “non‐regenerative” anterior wound with retinoic acid (RA) induces the formation of ectopic limbs. (A) Diagram of an axolotl forelimb indicating the positions at which wounds were made around the limb circumference. (B) Diagram illustrating the experimental procedures performed on anterior wound sites. An anterior wound site with a deviated nerve forms an ectopic blastema within 5−10 days. Without a posterior skin graft, the ectopic blastema integrates into the limb and the wound heals over without forming an ectopic limb (Endo et al. 2004). If a skin graft from the posterior side of the limb is grafted into the wound site, an ectopic limb is generated (Endo et al. 2004). Ectopic blastemas without a graft (non‐regenerating) were treated with RA to determine whether RA can replicate the positional confrontation normally induced by the posterior tissue graft. (C) Images of the same ectopic limb induced by RA treatment of an anterior ectopic blastema taken over a period of 7 weeks. The first image is the blastema on the day of RA treatment. The blastema forms an amorphous mass that eventually forms two well patterned ectopic limbs. These images are representative of what was observed in 10 of 20 ectopic structures that eventually formed paired ectopic limbs from an RA‐treated anterior wound (see Table 1). Blue scale bars are 1 mm in length.

Figure 2

Histology of RA‐induced paired limbs from an anterior wound site. Histological analyses were performed on sections that transected each of the two ectopic limbs that grew from an RA‐treated anterior wound site, harvested 10 weeks post‐treatment. Two complete limbs, including the zeugopod, stylopod, and autopod, formed from this experimental manipulation. A large mass of tissue formed proximal to the limb structures. (A) Fluorescent images were obtained of the cryosectioned limbs stained with DAPI (blue) and phalloidin‐rhodamine (red). Muscle tissue, rich in F‐actin (stained with phalloidin‐rhodamine), was distributed throughout the more distal limb structures but was not observed in the mass of tissue located proximally. (B) Sections were stained with eosin Y, hematoxylin, and alcian blue for histological analysis. The proximal mass of tissue predominantly differentiated into connective tissue and cartilaginous elements (dark blue) that could not be identified as corresponding to skeletal elements that are part of the normal limb pattern. Blue scale bars are 1 mm in length.

Figure 3

RA‐induced regeneration phenotypes are position‐specific. Quantification of the regeneration phenotypes for RA‐induced blastemas. Upon the completion of differentiation, the ectopic outgrowths were collected and stained for bone and cartilage in whole mount preparations in order to analyze the presence of skeletal elements in the regenerate. From our observations of these skeletal preparations, we divided the regeneration phenotypes into five categories: (1) no ectopic structures; (2) a single skeletal element; (3) multiple (jointed) symmetrical skeletal elements; (4) a single limb; or (5) two paired limbs (also see Table 1). (A) Examples of blastemas that exhibited the five different regeneration phenotypes that were quantified in this study. Images were taken of the blastemas starting the day of RA injection and ending when the tissue in the wound site had completely differentiated (determined by the formation of mature skin). (B) The histogram represents the percentage of ectopic blastemas located on the anterior, dorsal, posterior, or ventral axis that differentiated into each phenotype. Only the wounds located on the anterior or dorsal axis generated one or two paired limbs when treated with RA (also see Table 1). Blue scale bars are 1 mm in length.

Figure 4

Model of RA‐induced supernumerary limbs from ectopic blastemas. (A) Diagram of an amputated limb showing the distribution of positional information around the limb circumference as described in Bryant et al. (1981) and Bryant and Iten (1976). The experiment was performed on wound sites that were made on the most anterior (“9”), dorsal (“12/0”), posterior (“2/3”), or ventral (“5”) locations. (B) Models representing the positional interactions between the RA‐reprogrammed blastema (gray circle with the “4” coordinate) and the cells in the surrounding wound margin. The boundary of the square wound is indicated by heavy‐weighted lines. The thin lines and corresponding numbers on the outside of the square correspond to the distribution of circumferential positional information in the host limb as illustrated in (A). The thin lines and corresponding numbers inside the square wound are the predicated positional values that would be intercalated as a result of the interactions between the reprogrammed blastema cells (“4”) and the surrounding host cells. When the RA‐treated blastema is located in anterior and dorsal wound sites, the positional disparity is sufficient to induce the formation of two complete limb axes. In contrast, the interaction between the reprogrammed blastema cells with posterior or ventral host cells does not generate enough positional diversity to induce ectopic limb axes.