by Marc Tessier-Lavigne, Ph.D.,
Department of Anatomy, University of California, San Francisco.
Tessier-Lavigne presented this report as part of NINDS's
Neuroscience Lecture Series on Dec. 11, 1995.

Abstract

The functioning of the nervous system is dependent on the network of connections among neurons that arise during development. This network forms when each neuron sends out an axon to its target cells during embryogenesis. One mechanism that contributes to guiding axons to their targets is long-range chemotropism: axons can be attracted by diffusible attractants, secreted by target cells and repelled by long-range chemorepellent substances, secreted by non-target cells, that create exclusion zones that the axons avoid. Our laboratory has been interested in identifying these long-range chemoattractants and repellents in order to determine their contribution to guidance in vivo and their mechanisms of action. We have focused in particular on axon guidance in the developing spinal cord, where spinal commissural axons are attracted to an intermediate target, the floor plate of the spinal cord, by a floor plate-derived attractant.

Through biochemical purification, we were able to identify a good candidate for the attractant, a novel 78-kDa protein we call netrin-1, as well as a closely related molecule, netrin-2. Netrin-1 is expressed by floor plate cells and can mimic the chemoattractant activity of floor plate cells in an in vitro assay. Direct evidence for the involvement of netrin-1 in guiding commissural axons along a dorsal-to-ventral circumferential trajectory in vivo was obtained when we found that these axons become misrouted in mice that have a mutation in the netrin-1 gene.

Remarkably, the netrins are vertebrate homologs of the UNC-6 gene product in the nematode Caenorhabditis elegans. Mutations in unc-6 impair circumferential migrations of axons and cells in the nematode. The finding that unc-6 is required for migrations in both a ventral and a dorsal direction has led to the hypothesis that UNC-6 may attract some axons while it repells others. We have found that this is true of netrin-1, because a population of axons that grow away from the floor plate, trochlear motor axons, are repelled by netrin-1. Thus, UNC-6 and the netrins define a highly conserved family of bifunctional axon-guidance molecules.

What was your starting point, and how have your questions evolved?

The starting point was in studies I performed as a postdoctoral fellow with Tom Jessell and Jane Dodd at Columbia University in New York. In collaborative work with another fellow, Marysia Placzek, we investigated the cellular interactions that are responsible for directing commissural axons and discovered and characterized a chemoattractant activity in floor plate cells. After joining the faculty at UCSF, I continued efforts to identify the active factor. My lab's initial efforts were aimed simply at identifying the factor; our focus has now shifted to determining whether the netrins account for all of the bioactivity of floor plate cells and the precise roles played by netrins in axon guidance. We also wish to determine the mechanisms through which the netrins produce their effects, and whether chemotropism is a widespread mechanism of axon guidance.

Which findings have been most surprising to you or to other scientists?

The first surprise was that a relatively large protein (-80 kDa), related to extracellular matrix molecules, could function as a long-range chemoattractant. We had expected that a long-range attractant would be a smaller molecule (<10 kDa), as is the case for chemoattractants for cells of the immune system.

The second surprise was finding a close kinship between the netrins and UNC-6. Nowadays, it is perhaps not surprising that these vertebrate molecules should have a relative in C. elegans, nor is it particularly surprising that they are all involved in axon guidance. The surprise, however, is the fact that these molecules are involved in very similar guidance events in vertebrates and nematodes and - as shown by our collaborators in Corey Goodman's laboratory at the University of California, Berkeley - in fruit flies as well. In each organism, an UNC-6-netrin family member is expressed at the midline of the developing nervous system where the protein appears to play a role in attracting some axons while simultaneously repelling others. This degree of conservation - not just of structure but also of precise function - is still astonishing to me and prompted Goodman to quip, "The spinal cord is the worm within us."

The final surprise is the finding that these molecules are bifunctional guidance cues - simulataneously attracting some axons and repelling others. This indicates parsimony in the elaboration of guidance mechanisms. It also suggests that we should perhaps think of guidance cues as being present not specifically to attract or repel, but rather - much like signposts on a freeway - to provide directional information that axons can act upon in different ways depending on the guidance machinery present in their growth cones.

What were the greatest stumbling blocks, and what new observations, techniques, reagents, or insights helped you get past them?

The greatest stumbling block was the small size of floor plate tissue, which was the original source of chemoattractant activity. This precluded direct purification of the chemoattractant. To identify the attractant, I therefore took three approaches: expression cloning, screening known factors, and searching other tissues for a more abundant source of activity. Expression cloning involves screening a floor plate expression library for plasmids that could confer the chemoattractant activity. This approach was unsuccessful because of the low specific activity of the factor. Screening known factors - especially factors known to be chemoattractants in non-neural systems - on the assumption that the floor plate factor might be already identified also proved unsuccessful. Searching other tissues for a more abundant source of activity, which ultimately proved successful, was motivated by the hope that we could discover a similar activity in another tissue that would point us toward the correct molecule in floor plate cells. Extracts of brain tissue from defined embryonic stages turned out to possess activity similar to that in floor plate, and embryonic brain proved to be sufficiently abundant for a purification (though, in the end, we still needed 25,000 brains).

The activity from the brain tissue turned out to be due to two proteins, netrin-1 and netrin-2. While I think it was not luck that we isolated proteins related to the floor plate factor, we were lucky that netrin-1 turned out to be expressed in floor plate. I had only expected that the active component in brain extracts would be a distinct relative of the floor plate factor.

How can clinical scientists capitalize on this research?

Factors that promote the growth of axons may be useful for promoting regrowth of axons (regeneration) in adults following trauma or injury to the nervous system. We are starting to collaborate with several groups to determine whether the netrins could be useful for stimulating nerve repair.

How are you following up this work, and what questions would you ultimately like to answer?

Our current efforts are aimed at understanding how the netrins mediate their attractive and repulsive actions. We are thus trying to identify the receptors on growth cones that mediate the netrins' effects, as well as looking at downstream events that ultimately lead to growth cone reorientation. In addition, we have undertaken a large effort to identify other chemoattractants and repellents that guide developing axons in order to see whether chemotropism is a widespread mechanism of guidance and whether it is a unifed mechanism functioning in all cases through similar types of receptors and second-messenger systems.