A Role for Sonic Hedgehog in Axon Guidance

During development of the nervous system, axons at times navigate considerable distances before finally reaching their appropriate targets. The highly motile growth cone at the tip of an extending axon is exquisitely sensitive to attractive and repellent cues in the extracellular environment. These cues might be substrate-associated or diffusible, working at short range or over long distances, respectively. Axons often travel in distinct steps, reaching intermediate targets before continuing on to their ultimate destinations. These intermediate targets can consist of groups of specialized cells that express guidance molecules attractive to approaching axons. One particularly good model system for studying the factors that regulate axon guidance in this fashion has been the ventral midline of the developing vertebrate spinal cord.

Commissural neurons of the vertebrate embryo differentiate in the dorsolateral regions of the spinal cord. From here they send axons that navigate along stereotypical trajectories, traveling ventrally along the lateral margin of the spinal cord before turning at the level of the motor column and taking a more direct route to the ventral midline floor plate. After crossing the floor plate in the ventral commissure, these axons then turn abruptly and travel in the longitudinal axis on the contralateral side (Figure 1A). It has become evident that guidance molecules expressed by specialized ependymal cells that make up the floor plate are critical for attracting commissural axons as they travel toward the midline.1-3 Netrins make up a small family of secreted proteins with evolutionarily conserved roles in axon guidance affecting pathfinding at the midline from worms to mammals.3,4 Netrin-1 is produced by the vertebrate floor plate and stimulates both outgrowth and attraction of commissural axons in vitro, making it a good candidate for a role in long-range axon guidance.5 Knocking out either Netrin-1 or its receptor, Deleted in Colorectal Cancer (DCC), results in pathfinding defects that include stalling and abnormal trajectories of most commissural axons as they project to the floor plate.3,6 However, a small percentage of axons do still reach the midline, and the observation that Netrin-1-/- floor plate tissues still attract commissural axons in vitro, suggests the presence of additional attractive cues.3

The morphogen, Sonic Hedgehog (Shh), is expressed by the notochord and later by cells of the floor plate where it plays a well-described role in determining cell fate in the ventral neural tube.7 The ventral to dorsal gradient of Shh also places it in position to affect commissural axons approaching the midline floor plate.8 A recent study by Tessier-Levine and colleagues has shown that Shh is a likely mediator of Netrin-1-independent attraction at the vertebrate midline.9

Transfected COS cells expressing diffusible Shh in vitro are attractive to commissural axons from E11 rat spinal cord explants.9 The classical Shh pathway involves high-affinity binding to the receptor Patched (Ptc), which relieves inhibition of the signaling component of the receptor Smoothened (Smo).10 Suppressing Smo activity with the specific inhibitor cyclopamine blocks Shh-mediated attraction in vitro. To further implicate Shh as a midline attractant, studies were also carried out in vivo. Transcription factors of the Gli family (Gli1-3) are important players in Shh-mediated signal transduction pathways in the developing vertebrate nervous system.11 The neural tube of mice lacking Gli2 function, and therefore lacking normal Shh responses, fails to develop a functional midline floor plate.12 Although Shh expression is undetectable in the spinal cord of Gli2-/- mice, some Netrin-1 expression remains in the periventricular region and extending axons are still capable of reaching the midline.9,13 However, axons do appear defasciculated and often project abnormally into the lateral motor columns (Figure 1B). Knockouts of Shh, Ptc-1, or Smo each result in mice with defects in spinal cord patterning and are therefore not suitable for use in studying Shh-effects on commissural pathfinding.14-16 However, a Cre recombinase system using the Wnt-1 promoter was designed to more specifically target Smo function in commissural neurons. Commissural axons in mutant embryos lacking Smo activity exhibit a phenotype similar to Gli2-/- mice. Although many axons arrive at the ventral midline, prominent defasciculation and misrouted projections into the motor column are observed (Figure 1C). The most profound defects were observed in Gli2/Netrin-1 double knockout embryos where virtually all commissural axons stall without reaching the ventral spinal cord, resulting in elimination of the ventral commissure (Figure 1D).

These results suggest that the morphogen Shh acts in concert with Netrin-1 to regulate axon guidance at the vertebrate midline. Shh may not be the only morphogen with the ability to regulate axon guidance. For instance, Bone Morphogenetic Proteins (BMPs), members of the TGF-beta superfamily, may act as guidance cues as well.17 The roof plate at the dorsal midline is known to release factors that repulse commissural axons in vitro, an activity that is inhibited by function blocking antibodies to BMP-7 and is absent in BMP-7-/- roof plate explants.17 Members of the Wnt family have also been implicated as regulators of axon guidance. Targeted deletion of the Wnt receptor Frizzled-3, results in severe defects in embryonic axon tracts of the vertebrate forebrain.18 In addition, Wnt-5 binding to the receptor tyrosine kinase Derailed, is necessary for proper axon guidance at the Drosophila midline.19 Morphogen gradients in the developing embryo play classical roles in determining cell fate, but their expression patterns also place them in a position to affect axon guidance. Future studies will determine whether other morphogens act as guidance cues, while also leading to a better understanding of the mechanisms by which they mediate their effects on growing axons.

Figure 1.
  1. Wild-type commissural axons in the developing vertebrate embryo originate in the dorsal spinal cord and travel ventrally toward the midline floor plate where they cross in the ventral commissure before turning orthogonally on the contralateral side.
  2. The Gli2 knockout leads to loss of the floor plate and concomitant loss of Sonic Hedgehog (Shh) expression and most of the Netrin-1 expression. Most axons do reach ventral commissure but many defasciculate and send wandering projections into the motor column.
  3. Inactivation of the signaling portion of the Shh receptor, Smoothened (Smo), results in commissural axon projections reminiscent of those in Gli2-/- spinal cords. Many axons reach the midline but exhibit marked defasciculation and wandering projections.
  4. Commissural axons from Gli2 and Netrin-1 double knockouts stall and most fail to enter the ventral regions of the spinal cord.

    [Note: figure adapted from Charron, F. et al. (2003) Cell 113:11.]


  1. Placzek, M. et al. (1990) Development 110:19.
  2. Serafini, T. et al. (1994) Cell 78:409.
  3. Serafini, T. et al. (1996) Cell 87:1001.
  4. Ishii, N. et al. (1992) Neuron 9:873.
  5. Kennedy, T.E. et al. (1994) Cell 78:425.
  6. Fazeli, A. et al. (1997) Nature 386:796.
  7. Marti, E. & P. Bovolenta (2002) Trends Neurosci. 25:89.
  8. Gritli-Linde, A. et al. (2001) Dev. Biol. 236:364.
  9. Charron, F. et al. (2003) Cell 113:11.
  10. Murone, M. et. al. (1999) Curr. Biol. 9:76.
  11. Ruiz i Altaba, A. (1998) Development 125:2203.
  12. Ding, Q. et al. (1998) Development 125:2533.
  13. Matise, M.P. et al. (1999) Development 126:3649.
  14. Chiang, C. et al. (1996) Nature 383:407.
  15. Goodrich, L.V. et al. (1997) Science 277:1109.
  16. Zhang, X.M. et al. (2001) Cell 106:781.
  17. Augsburger, A. et al. (1999) Neuron 24:127.
  18. Wang, Y. et al. (2002) J. Neurosci. 22:8563.
  19. Yoshikawa, S. et al. (2003) Nature 422:583.