Identification of regenerative roadblocks via repeat deployment of limb regeneration in axolotls

Abstract

Axolotl salamanders are powerful models for understanding how regeneration of complex body parts can be achieved, whereas mammals are severely limited in this ability. Factors that promote normal axolotl regeneration can be examined in mammals to determine if they exhibit altered activity in this context. Furthermore, factors prohibiting axolotl regeneration can offer key insight into the mechanisms present in regeneration-incompetent species. We sought to determine if we could experimentally compromise the axolotl's ability to regenerate limbs and, if so, discover the molecular changes that might underlie their inability to regenerate. We found that repeated limb amputation severely compromised axolotls' ability to initiate limb regeneration. Using RNA-seq, we observed that a majority of differentially expressed transcripts were hyperactivated in limbs compromised by repeated amputation, suggesting that mis-regulation of these genes antagonizes regeneration. To confirm our findings, we additionally assayed the role of amphiregulin, an EGF-like ligand, which is aberrantly upregulated in compromised animals. During normal limb regeneration, amphiregulin is expressed by the early wound epidermis, and mis-expressing this factor lead to thickened wound epithelium, delayed initiation of regeneration, and severe regenerative defects. Collectively, our results suggest that repeatedly amputated limbs may undergo a persistent wound healing response, which interferes with their ability to initiate the regenerative program. These findings have important implications for human regenerative medicine.

Fig. 1

Regenerative decline after repeated amputation. Both forelimbs were amputated and allowed to fully regenerate. a Experimental overview. b Skeletal preparations of limbs following successive rounds of amputation. The limb in bottom right panel failed to regenerate beyond the plane of amputation (dashed line). c Representative examples of sibling control limb (top left panel) and limb that failed to regenerate after repeated amputation (bottom left panel). d Cumulative distribution plot of loss in ability to regenerate beyond the plane of amputation (right graph). Scale bars in b are 1 mm

Fig. 2

Regenerative decline after repeated amputation is less severe if the plane of amputation is shifted distally along the axis of the limb. Both forelimbs were amputated for each animal and allowed to fully regenerate. a Experimental overview. b Representative images of axolotl limbs after 5 rounds of amputation at the same plane (left panels) or with serially distal planes of amputation (right panels). c Quantification of the number of limbs able to regenerate digits after each amputation round following repeated amputation in the same plane or repeated, progressively distal amputations. Asterisk (*) indicates p < 0.05 (Fisher’s Exact Test). All scale bars equal to 1 mm

Fig. 3

Aborted limb stumps exhibit persistent collagen deposition. A Masson’s trichrome stain was performed on limbs that failed to regenerate (“stumps”) following repeated amputation and on intact control limbs. a Column a (leftmost) depicts an intact specimen (no amputations). b Column b (middle) depicts a specimen following same-plane, repeated amputation. c Column c (right) depicts a specimen following serial-plane, repeated amputation. Top panels show representative images of intact control a and failed regenerates from limbs amputated in the same plane b and limbs amputated serially distally c. Middle and lower horizontal panels are higher magnification views of images in the top panels. Brackets delineate epidermis, arrowheads indicate substratum compactum (dermis), double arrowheads indicate connection between substratum compactum and epidermis, and arrows indicate epidermal tongues. Scale bars in top rows are 500 µm, and scale bars in middle and bottom rows are 100 µm

Fig. 4

Transcriptomic analyses suggest an antagonistic expression pattern for amphiregulin. Failed regenerates that had undergone 5 amputations and sibling control limbs that had never been injured were amputated proximally and harvested at 3 days post-amputation. a Overview of RNA-sequencing strategy. b Heatmap showing the results of k-means clustering (k = 2) of significantly differentially expressed genes between repeatedly amputated limbs and control limbs. Several genes with known roles in regenerative processes are shown. c Gene Ontology analyses of genes that are downregulated in the repeated amputation condition. Panel on the left shows the top 10 most significantly enriched Biological Processes. Panel on the right is a treemap of all significantly enriched Biological Processes. d Gene Ontology analyses of genes that are upregulated in the repeated amputation condition. Panel on the left shows the top 10 most significantly enriched Biological Processes. Panel on the right is a treemap of all significantly enriched Biological Processes. “A”: forward locomotion; “B”: multicellular organismal process; “C”: directional locomotion; “D”: developmental process. A full list of significantly enriched Biological Processes can be found in Supplementary Table 2. e qRT-PCR showing normalized areg expression at 3 dpa in normally-regenerating control limbs (first amputation) and in limbs compromised by repeat amputation (sixth amputation). Error bars are SEM. N = 8 limbs per condition. Asterisks (**) denotes p < 0.01. f qRT-PCR showing normalized areg expression at 3 dpa in normally-regenerating control limbs (first amputation) and in limbs amputated twice and thrice. Error bars are SEM. N = 4-5 limbs per condition. Asterisks (**) denotes p < 0.01; Asterisk (*) denotes p < 0.05; n.s. denotes not significant. g Plot showing the 11 genes with the highest initiation ratio (ratio of early expression to late expression). Expression data for panel g were obtained from [32]

Fig. 5

Amphiregulin is expressed during the wound healing stage of limb regeneration. In situ hybridization analyses of areg expression in wild-type juvenile axolotls; a–f are regenerating limb samples; g–h are flank skin wound samples. a Expression of areg at 3 h post-amputation. Arrowheads indicate expression of areg in the leading edge of the wound epidermis. Epi epidermis. b Higher magnification view of panel a. c Section stained with sense control probe at 3 h post-amputation. d Expression of amphiregulin at 12 h post-amputation. Arrowheads indicate expression of areg in the wound epidermis (we). e Tissue section stained with areg anti-sense in situ probe at 72 h post-amputation. f Tissue section stained with areg anti-sense in situ probe at 14 days post-amputation (medium bud blastema). g In situ hybridization showing areg expression in leading edge of wound epidermis at 6 h post-biopsy. h Higher magnification view of panel g. We wound epidermis, bl blastema. Images are representative of four biological replicates per time point. Scale bars are 100 µm

Fig. 6

Overexpression of amphiregulin disrupts limb regeneration. Axolotl limbs with no prior injuries were electroporated with either plasmid encoding GFP (control) or plasmids encoding GFP plus AREG. a Overview of experimental strategy. b-b”) Representative images of EGFP control limbs from 8 to 16 days post-amputation (dpa). Arrowheads denote amputation planes. c-c”) Representative images of areg mis-expressing limbs from 8 to 16 days post-amputation (dpa). Arrowheads denote amputation planes. d Quantification of blastema lengths in b–b” and c–c”. N = 24 animals for control and N = 23 animals for areg overexpression. e-g Representative images of control limbs and limbs exhibiting severe regenerative defects or no regeneration beyond the stylopodium following areg overexpression. e’–g’) Representative skeletal preparations of control limbs and limbs exhibiting severe regenerative defects or no regeneration beyond the stylopodium following areg overexpression. h Quantification of defects after control egfp and areg overexpression. The two groups exhibit significantly different morphologies (p < 0.01, Fisher’s exact test). N = 48 limbs for control and N = 46 limbs for areg overexpression

Fig. 7

Overexpression of amphiregulin results in abnormally thick wound epidermis, alterations in cellular proliferation, and increased mTOR signaling in the wound epidermis during limb regeneration. Axolotl limbs with no prior injuries were electroporated with either plasmid encoding GFP or plasmids encoding GFP and AREG. a Multi-timepoint Masson’s trichrome staining of tissue sections from egfp control or areg overexpressing limbs. For 0 dpa, 12 hpa, 1 dpa, and 3 dpa, N = 5 limbs per group per timepoint. For 8 dpa, N = 5 limbs for control and N = 6 limbs for areg overexpression. b Representative immunofluorescent staining of phospho-Histone H3 (pH3) on tissue sections from egfp control limbs at 8 dpa. c Representative immunofluorescent staining of phospho-Histone H3 (pH3) on tissue sections from areg overexpressing limbs at 8 dpa. d Quantification of the percentage of pH3-positive nuclei in the wound epidermis. Asterisk (*) indicates p < 0.05. N = 5 limbs for control, and N = 6 limbs for areg overexpression. e Quantification of the percentage of pH3-positive nuclei in non-wound epidermal tissues. f Representative immunofluorescent staining of phospho-rpS6 (pS6) on tissue sections from egfp control limbs at 8 dpa. g Representative immunofluorescent staining of pS6 on tissue sections from areg overexpressing limbs at 8 dpa. h Quantification of the percentage of pS6 positive cells in non-wound epidermal tissues. n.s. not significant. N = 5 limbs for control, and N = 6 limbs for areg overexpression. i Quantification of the percentage of pS6 positive cells in the wound epidermis. Asterisk (*) indicates p < 0.05. N = 5 limbs for control, and N = 6 limbs for areg overexpression. Scale bars are 100 µm