In salamanders, grafting of a left limb blastema onto a right limb stump yields regeneration of three limbs, the normal limb and two ‘supernumerary’ limbs1,2,3,4. This experiment and other research have shown that the juxtaposition of anterior and posterior limb tissue plus innervation are necessary and sufficient to induce complete limb regeneration in salamanders5,6,7,8,9,10. However, the cellular and molecular basis of the requirement for anterior–posterior tissue interactions were unknown. Here we have clarified the molecular basis of the requirement for both anterior and posterior tissue during limb regeneration and supernumerary limb formation in axolotls (Ambystoma mexicanum). We show that the two tissues provide complementary cross-inductive signals that are required for limb outgrowth. A blastema composed solely of anterior tissue normally regresses rather than forming a limb, but activation of hedgehog (HH) signalling was sufficient to drive regeneration of an anterior blastema to completion owing to its ability to maintain fibroblast growth factor (FGF) expression, the key signalling activity responsible for blastema outgrowth. In blastemas composed solely of posterior tissue, HH signalling was not sufficient to drive regeneration; however, ectopic expression of FGF8 together with endogenous HH signalling was sufficient. In axolotls, FGF8 is expressed only in the anterior mesenchyme and maintenance of its expression depends on sonic hedgehog (SHH) signalling from posterior tissue. Together, our findings identify key anteriorly and posteriorly localized signals that promote limb regeneration and show that these single factors are sufficient to drive non-regenerating blastemas to complete regeneration with full elaboration of skeletal elements.
This is a preview of subscription content, access via your institution
Subscribe to this journal
Receive 51 print issues and online access
$199.00 per year
only $3.90 per issue
Rent or buy this article
Prices vary by article type
Prices may be subject to local taxes which are calculated during checkout
Similar content being viewed by others
NCBI Reference Sequence
The following sequences have been deposited in the NCBI GenBank database with the corresponding accession numbers: fgf10, KU882013 ; fgf17, KU882014; grem1, KU882015; hand2, KU882016; gli3, KU882017; hoxd13, KU882018; fgf8, KU882019; fgf9, KU882020.
Iten, L. E. & Bryant, S. V. The interaction between the blastema and stump in the establishment of the anterior–posterior and proximal–distal organization of the limb regenerate. Dev. Biol. 44, 119–147 (1975)
Maden, M. Structure of supernumerary limbs. Nature 286, 803–805 (1980)
Maden, M. Supernumerary limbs in amphibians. Integr. Comp. Biol. 22, 131–142 (1982)
Stocum, D. L. Determination of axial polarity in the urodele limb regeneration blastema. J. Embryol. Exp. Morphol. 71, 193–214 (1982)
Bryant, S. V. Regenerative failure of double half limbs in Notophthalmus viridescens. Nature 263, 676–679 (1976)
Bryant, S. V. & Baca, B. A. Regenerative ability of double-half and half upper arms in the newt, Notophthalmus viridescens. J. Exp. Zool. 204, 307–323 (1978)
Stocum, D. L. Regeneration of symmetrical hindlimbs in larval salamanders. Science 200, 790–793 (1978)
Endo, T., Bryant, S. V. & Gardiner, D. M. A stepwise model system for limb regeneration. Dev. Biol. 270, 135–145 (2004)
Satoh, A., Gardiner, D. M., Bryant, S. V. & Endo, T. Nerve-induced ectopic limb blastemas in the axolotl are equivalent to amputation-induced blastemas. Dev. Biol. 312, 231–244 (2007)
Nacu, E. & Tanaka, E. M. Limb regeneration: a new development? Annu. Rev. Cell Dev. Biol. 27, 409–440 (2011)
Bryant, S. V., French, V. & Bryant, P. J. Distal regeneration and symmetry. Science 212, 993–1002 (1981)
Meinhardt, H. A boundary model for pattern formation in vertebrate limbs. J. Embryol. Exp. Morphol. 76, 115–137 (1983)
Laufer, E., Nelson, C. E., Johnson, R. L., Morgan, B. A. & Tabin, C. Sonic hedgehog and Fgf-4 act through a signaling cascade and feedback loop to integrate growth and patterning of the developing limb bud. Cell 79, 993–1003 (1994)
Chiang, C. et al. Manifestation of the limb prepattern: limb development in the absence of sonic hedgehog function. Dev. Biol. 236, 421–435 (2001)
Harfe, B. D. et al. Evidence for an expansion-based temporal Shh gradient in specifying vertebrate digit identities. Cell 118, 517–528 (2004)
Towers, M., Mahood, R., Yin, Y. & Tickle, C. Integration of growth and specification in chick wing digit-patterning. Nature 452, 882–886 (2008)
Zhu, J. et al. Uncoupling Sonic hedgehog control of pattern and expansion of the developing limb bud. Dev. Cell 14, 624–632 (2008)
Niswander, L., Jeffrey, S., Martin, G. R. & Tickle, C. A positive feedback loop coordinates growth and patterning in the vertebrate limb. Nature 371, 609–612 (1994)
Zúñiga, A., Haramis, A. P., McMahon, A. P. & Zeller, R. Signal relay by BMP antagonism controls the SHH/FGF4 feedback loop in vertebrate limb buds. Nature 401, 598–602 (1999)
Lewandoski, M., Sun, X. & Martin, G. R. Fgf8 signalling from the AER is essential for normal limb development. Nature Genet. 26, 460–463 (2000)
Moon, A. M. & Capecchi, M. R. Fgf8 is required for outgrowth and patterning of the limbs. Nature Genet. 26, 455–459 (2000)
Khokha, M. K., Hsu, D., Brunet, L. J., Dionne, M. S. & Harland, R. M. Gremlin is the BMP antagonist required for maintenance of Shh and Fgf signals during limb patterning. Nature Genet. 34, 303–307 (2003)
Panman, L. et al. Differential regulation of gene expression in the digit forming area of the mouse limb bud by SHH and gremlin 1/FGF-mediated epithelial-mesenchymal signalling. Development 133, 3419–3428 (2006)
Nissim, S., Hasso, S. M., Fallon, J. F. & Tabin, C. J. Regulation of Gremlin expression in the posterior limb bud. Dev. Biol. 299, 12–21 (2006)
Mariani, F. V., Ahn, C. P. & Martin, G. R. Genetic evidence that FGFs have an instructive role in limb proximal-distal patterning. Nature 453, 401–405 (2008)
Sobkow, L., Epperlein, H.-H., Herklotz, S., Straube, W. L. & Tanaka, E. M. A germline GFP transgenic axolotl and its use to track cell fate: dual origin of the fin mesenchyme during development and the fate of blood cells during regeneration. Dev. Biol. 290, 386–397 (2006)
Boyce, F. M. & Bucher, N. L. Baculovirus-mediated gene transfer into mammalian cells. Proc. Natl Acad. Sci. USA 93, 2348–2352 (1996)
Kost, T. A. & Condreay, J. P. Recombinant baculoviruses as mammalian cell gene-delivery vectors. Trends Biotechnol. 20, 173–180 (2002)
Kaikkonen, M. U. et al. Truncated vesicular stomatitis virus G protein improves baculovirus transduction efficiency in vitro and in vivo. Gene Ther. 13, 304–312 (2006)
Hopkins, R. & Esposito, D. A rapid method for titrating baculovirus stocks using the Sf-9 Easy Titer cell line. BioTechniques 47, 785–788 (2009)
Kragl, M. et al. Cells keep a memory of their tissue origin during axolotl limb regeneration. Nature 460, 60–65 (2009)
Nacu, E. et al. Connective tissue cells, but not muscle cells, are involved in establishing the proximo-distal outcome of limb regeneration in the axolotl. Development 140, 513–518 (2013)
Schnapp, E., Kragl, M., Rubin, L. & Tanaka, E. M. Hedgehog signaling controls dorsoventral patterning, blastema cell proliferation and cartilage induction during axolotl tail regeneration. Development 132, 3243–3253 (2005)
We thank B. Gruhl, S. Moegel, M. Armstead, H. Goers, A. Wagner and S. Kaudel for animal care, H. Q. Le and K. Goehler for assistance, C. Cannistraci for advice on statistical analysis, C. Antos, D. Knapp, W. Masselink, T. Sugiura and Y. Taniguchi for advice on the manuscript, K. Crawford for discussion, S. Eaton for providing us with the plasmid with human SHH, K. Airenne, S. Ylä-Herttuala, and M. Kaikonnen for the plasmid with VSV-GED, members of the MPI-CBG protein facility for assistance in baculovirus preparation, and members of the MPI-CBG antibody facility for the anti-MHC antibody. This work was supported by central funds from the CRTD and MPI-CBG and an ERC Advanced Investigator grant to E.M.T. C.R.O. was supported by the Portuguese Foundation for Science and Technology (FCT). Schematic illustrations of axolotl limbs were redrawn based on illustrations from ref. 10 under copyright transfer agreement to E.N. and E.M.T. as authors.
The authors declare no competing financial interests.
Extended data figures and tables
Extended Data Figure 1 SAG-induced accessory limbs contain muscles, axons, and primarily lower arm and hand elements.
a, Immunostaining for myosin heavy chain (MHC) reveals muscle fibres in the accessory limb (AL). b, Immunostaining for β3-tubulin reveals axons in the accessory limb. c, Immunostaining for MEIS, an upper arm identity marker, reveals that the accessory limb contains primarily lower arm and hand elements, as evidenced by the absence of MEIS in the accessory limb. Expression of MEIS is seen in the host upper arm on which the accessory limb was induced. MEIS staining is also observed in the epidermis, however epidermis does not play a role in positional identity. d, Overlay of anti-MHC (green), anti-β3-tubulin (white), anti-MEIS (red) and Hoechst (marks nuclei, blue) staining. Number of accessory limbs analysed, n = 5. Scale bar, 1 mm.
Extended Data Figure 2 HH signalling is sufficient for accessory limb outgrowth from ABs and is necessary for outgrowth from ABs complemented with posterior skin.
a, Treatment of ABs with lower concentrations of SAG still results in accessory limb outgrowth, but with a decreased number of digits. Data from a single experiment. b, SAG (40 nM) also induces accessory limb outgrowth from ABs in large animals (13–14 cm snout to tail tip), indicating that this effect is not size-dependent. Data from a single experiment. c, Blocking HH signalling with cyclopamine (8 μM) in ABs complemented with posterior skin blocks regeneration. Data from a single experiment. d, Percentage of accessory limb outgrowths from ABs treated with different concentrations of SAG. See Supplementary Table 1 for details. Short blue horizontal lines represent average percentages for each condition. The dotted line is for visual aid and intersects the y axis at 0%. Numbers in upper right corners of photomicrographs indicate the number of accessory limb outgrowths out of the total number of ABs. dpw, days post-wounding. Scale bars, 2 mm.
Extended Data Figure 3 SAG treatment does not make the determination state of accessory limbs more posterior.
a, SAG-induced accessory limbs (SAG-AL) regenerate at most a spike after amputation, indicating a lack of functional positional discontinuity. SAG-ALs were amputated and imaged at 14 and 21 days post-amputation. No further growth of the spike was observed 21 days post-amputation. Red or black lines, amputation plane. No SAG-ALs regenerated a patterned accessory limb after amputation (n = 0 of 5). Experimental data from a single experiment. Alcian blue/alizarin red staining: cartilage is blue, ossifications are red. b–d, Testing determination state of SAG-ALs by skin transplantation to ABs or PBs. b, Schema illustrating the experimental setup and potential outcomes. Accessory limbs were generated by SAG treatment of ABs from transgenic animals constitutively expressing GFP. The SAG-ALs were divided into a medial and lateral side, with skin from one side transplanted to an anterior host site and skin from the opposite side transplanted to a posterior host site. Two possible outcomes were expected, depending on whether SAG posteriorized cells during initial accessory limb outgrowth. c, A host AB with a skin transplant from a SAG-AL shows no accessory limb outgrowth, indicating no determination of posterior identity during SAG-AL outgrowth (n = 0 of 22). d, A host PB with a skin transplant from the same SAG-AL as in c results in new accessory limb outgrowth, revealing the anterior identity of the transplanted skin (n = 15 of 21). Data from a single experiment. See Supplementary Table 2 for details. DF, dark field images. GFP panels show the transplanted skin. dpt, days post transplantation of skin. Scale bars, 2 mm.
a, ABs express blastema markers prrx1, msx1, msx2, and twist1, and anterior marker gli3. The y axis indicates the number of normalized counts obtained by Nanostring nCounter technology. Each data point is the mean normalized count number from biological replicates (n = 3 or 4; for individual n values and complete data set, see Supplementary Table 7a; data from one experiment). Error bars represent s.d. For data points where s.d. was small, no error bars are shown. b, gli3 is expressed at 8 and 12 days in SAG-treated or untreated ABs. c–h, Expression of FGF–SHH loop components in SAG-treated versus untreated ABs. c, d, f, In ABs, fgf8, fgf9, and grem1 maintenance, but not induction, requires SAG treatment. c, d, fgf8 and grem1 are upregulated in ABs in the presence or absence of SAG at 8 days, but are downregulated at 12 days unless SAG is present. f, fgf9 is expressed at the tip of the bumps at 8 days and maintains this expression at the tip in 12-day SAG-treated samples. In untreated 12-day samples there is no more fgf9 expression at the tip, although there seems to be expression at the base of the bump. This could represent a secondary source of fgf9 that is not part of the FGF–SHH circuitry or it could be an unspecific signal; see Methods. g, fgf10 was expressed in all conditions. h, Unlike other fgf genes, fgf17 expression required SAG at all time points. fgf17 was not expressed in untreated samples at either 8 or 12 days, but was expressed at both time points when treated with SAG. e, SAG is not sufficient to upregulate shh in ABs. Shh is not expressed in ABs at 8 or 12 days in the presence or absence of SAG. For detailed information on the number of samples analysed and results, see Supplementary Table 3. Inserts show magnification of positive cells. Images of 12-day SAG-treated blastemas in panels f–h show different sections from the same limb that were hybridized separately for each of the indicated genes. Scale bars: all 8-day images, 100 μm; all 12-day images, 200 μm.
Extended Data Figure 5 FGF signalling is required for SAG-induced accessory limb outgrowth from ABs, and is sufficient to induce single-digit accessory limb outgrowth from ABs.
a, FGF signalling is required for SAG-induced accessory limb outgrowth. ABs were treated with 15 nM SAG or water starting on the day of wounding. At 10 days post-wounding, concomitant treatment with DMSO or 15 µM PD173074 (inhibitor of FGFR1 and FGFR3) was initiated. Treatment was terminated at 22 days post-wounding. Images of ABs were taken at 10, 15 and 30 days post-wounding. ABs treated with water and DMSO fail to develop accessory limbs (n = 1 of 16). ABs treated with SAG and DMSO form accessory limbs (n = 14 of 14). ABs treated with SAG and PD173074 grow until initiation of PD173074 treatment, after which growth halts and ABs fail to develop accessory limbs (n = 0 of 16), eventually regressing. This indicates that FGF signalling acts downstream of SAG and is required for accessory limb outgrowth. Dashed yellow lines demarcate ABs. The experimental data come from two experiments. dpw, days post-wounding. Scale bars, 1 mm. b, Ectopic baculovirus-induced expression of FGF8 in ABs results in single digit accessory limb outgrowths. Percentage of substantial outgrowths from ABs transduced with fgf8 or mCherry (negative control). See Supplementary Table 4 for details. Short blue horizontal lines represent average percentage for each condition. The dotted line is for visual aid and intersects the y axis at of 0%.
Extended Data Figure 6 FGF8, but not HH signalling, is sufficient to drive accessory limb outgrowth from PBs.
a–c, f, HH signalling is not sufficient to drive accessory limb outgrowth from PBs (for details see Supplementary Table 1). a, SAG is sufficient to drive accessory limb outgrowth from ABs (n = 23 of 32) but not PBs (n = 1 of 32) on the contralateral limb of the same animal. Data from one experiment. The images are the same as in Fig. 2a. b, Untreated PBs also produce no accessory limb outgrowth (n = 1 of 37; data from two experiments). c, Positive control: PBs complemented with anterior skin transplants from animals that constitutively express GFP grew into accessory limbs without SAG treatment (n = 10 of 19, data from two experiments). f, Percentage of accessory limb outgrowths from PBs without treatment (H2O), treated with SAG, or complemented with anterior skin. d, e, g, FGF8 is sufficient to drive accessory limb outgrowth from PBs (data from three experiments; for details see Supplementary Table 5). d, PBs expressing baculovirus-induced FGF8 grow into accessory limbs (n = 27 of 50). The images are the same as in Fig. 2b. e, Negative control: PBs expressing baculovirus-induced mCherry do not grow into accessory limbs (n = 3 of 49). g, Percentage of accessory limb outgrowths from PBs transduced with fgf8 or mCherry (negative control). Short blue horizontal lines in all graphs represent average percentage for each condition. Dotted lines are for visual aid and intersect the y axes at 0%. Dashed yellow lines demarcate blastemas. DF, dark field images; dpw, days post-wounding. Scale bars, 2 mm.
Extended Data Figure 7 PBs do not express fgf8 or fgf17 even though intact HH signalling is present, but do express grem1, fgf10 and fgf9.
a, g, PBs do not express fgf8 and fgf17 even upon SAG treatment. Fgf8 and fgf17 are not expressed in PBs at 8 or 12 days in the presence or absence of SAG, but are expressed in positive control PBs containing an anterior skin transplant. b–d, f, Early induction of grem1, shh, ptc1 (a HH signalling target) and fgf10 in posterior wounds is independent of SAG, and SAG is not sufficient for maintenance of grem1 and shh expression. b, c, Grem1 and shh are present in PBs at 8 days and are downregulated by 12 days independent of SAG treatment. This observation differs slightly from the Nanostring nCounter data (Fig. 2d) where grem1 is downregulated at 15 days. The discrepancy is probably due to the variation in regeneration timelines across different experiments. Grem1 and shh are present in control PBs containing an anterior skin transplant at both 8 and 12 days. d, Ptc1 is expressed at 8 days in PBs that were untreated or treated with SAG or complemented with an anterior skin graft. f, Fgf10 is present in all conditions at 8 and 12 days. e, Fgf9 is expressed in a few cells at the tip of 8-day untreated PBs, but is absent from 12-day untreated PBs and 8- or 12-day SAG-treated PBs. Fgf9 is present in control PBs complemented with anterior skin grafts. Detailed information on the number of samples analysed and results can be found in Supplementary Table 3. Inserts show magnification of positive cells. Images of 12-day PBs complemented with anterior skin (e–g) show different sections from the same limb that were hybridized separately for each of the indicated genes. Scale bars: all 8-day samples, 100 μm; all 12-day samples, 200 μm.
Extended Data Figure 8 Schematic illustration of the expression pattern of fgf8, fgf9, fgf10, fgf17, grem1 and shh in ABs and PBs treated with SAG or untreated and PBs complemented with anterior skin transplant.
ABs upregulate fgf8, fgf9, fgf10 and grem1 but not shh in the absence of SAG at 8 days post-wounding, indicating that initial upregulation of fgf8, fgf9 and fgf10 (yellow) and grem1 (blue) is independent of SAG. ABs downregulate expression of fgf8 and grem1 by 12 days, and eventually regress. ABs treated with SAG maintain fgf8 and grem1 expression at 12 days, indicating that HH signalling is necessary for their maintenance. SAG-treated ABs also express fgf17. ABs treated with SAG eventually develop accessory limbs (ALs). Untreated PBs upregulate fgf9, fgf10, grem1 and shh (red) but not fgf8 nor fgf17. SAG-treated PBs also do not upregulate fgf8 or fgf17 and do not produce accessory limb outgrowths, indicating that SHH signalling is not sufficient for accessory limb outgrowth from PBs. PBs complemented with an anterior skin transplant express fgf8, fgf9, fgf10, fgf17, grem1 and shh and form accessory limbs.
Extended Data Figure 9 Expression patterns of gli3, hand2, fgf8, fgf9, fgf10, fgf17, grem1, shh and hoxd13 in normal limb regeneration.
a, Expression of Gli3, hand2, fgf8 and grem1 was analysed on adjacent sections of a 10-day blastema. Sections are arranged from the distal tip of the blastema (top, plane A) to the proximal base of the blastema (bottom, plane E). Gli3 is expressed throughout the blastema, with stronger expression in the anterior half that lacks hand2 expression and weaker expression in the hand2-expressing zone. Fgf8 is expressed in the anterior half that lacks hand2 expression. Grem1 is expressed more proximally than the other genes (starting from plane B). It is primarily expressed in the anterior half of the blastema, but is also expressed in the posterior half, particularly in the proximal part (planes D and E). b, Unlike fgf8; fgf9, fgf10 and fgf17 are not excluded from the hand2-expressing, posterior half of a blastema (11 days). c, Shh is expressed in a small posterior domain in the blastema (12 days). The planes are sections of the same blastema along the proximo-distal axis. Scale bars, 200 μm.
Extended Data Figure 10 HH signalling is required for FGF8-induced accessory limb outgrowth from PBs.
a, Schematic outline of the experiment. Each replicate experiment consisted of transducing limbs with fgf8 or mCherry (negative control), followed by injury with nerve deviation at 10 or 14 days after transduction. Treatment of FGF8-expressing animals with 8 μM cyclopamine or an equivalent amount of ethanol (cyclopamine solvent) was started at 6 days post-wounding. Additionally a group of non-transduced PBs were complemented with anterior skin and treated with 8 μM cyclopamine (as controls for blocking HH signalling) or ethanol. b, Examples of typical outcomes in each condition. Images from the condition PB+FGF8+cyclopamine are the same as in Fig. 2c. Scale bars, 2 mm. dpw, days post-wounding. c, Percentage of substantial outgrowths in each condition. Data from four experiments. For detailed description of the number of limbs in each condition and the way they were scored see Supplementary Table 6. The dotted line is for visual aid and intersects the y axis at 0%. d, The effect of cyclopamine is specific to HH signalling. Limbs were amputated and treatment was started 6 days post-amputation with 8 μM cyclopamine alone or in combination with 600 nM SAG. Treatment with cyclopamine blocked regeneration, which was rescued by co-treatment with SAG. Numbers indicate regenerated limbs out of total amputated limbs. Data from two experiments. dpa, days post-amputation.
This file contains Supplementary Tables 1-6. (PDF 284 kb)
This file contains Supplementary Table 7. (XLSX 102 kb)
This file contains Supplementary Table 8. (XLSX 60 kb)
This file contains Supplementary Table 9. (XLSX 49 kb)
This file contains Supplementary Table 10. (XLSX 68 kb)
About this article
Cite this article
Nacu, E., Gromberg, E., Oliveira, C. et al. FGF8 and SHH substitute for anterior–posterior tissue interactions to induce limb regeneration. Nature 533, 407–410 (2016). https://doi.org/10.1038/nature17972
This article is cited by
Akt/mTOR integrate energy metabolism with Wnt signal to influence wound epithelium growth in Gekko Japonicus
Communications Biology (2022)
Cell Regeneration (2021)
Nature Cell Biology (2021)
Single-cell RNA-seq reveals novel mitochondria-related musculoskeletal cell populations during adult axolotl limb regeneration process
Cell Death & Differentiation (2021)
Cellular and Molecular Life Sciences (2021)