The time was just right. For decades people had been looking for the molecules involved in the exquisitely precise wiring of the nervous system. How is it that neuronal axons, the extensions of neurons, can navigate to targets unerringly? What information are they detecting? What molecules are serving as guideposts? There had been a number of false starts in the 1980s, a number of attempts to identify these molecules. A number of candidates had been tested and then discarded; it was discovered that they either weren’t involved or were only side players, not the central players. However, by the early to mid-1990s, over the course of a few years, a number of discoveries of the key axon guidance molecules were made. At the time, three of what I like to call the "big four" families of axon guidance were identified. These were the netrins, semaphorins, and ephrins. The fourth family of key axon guidance molecules, the slits, wasn’t identified until 1999.

If you look at the article, you’ll see that we knew there was a midline repellant that hadn’t yet been identified. In principle, this could have been a member of one of the three families we’d already discovered, but there were indications that it wasn’t. That turned out to be slit. So, already in 1996 there was the sense of something lurking out there. It was also clear at the time that there was another chemotractant missing. It wasn’t until 2003 that we showed that the other chemotractant is sonic hedgehog.

No. Although the "big four" are key regulators of guidance, other types of molecules have emerged as regulators as well. For example, one of the pleasing discoveries that was made recently is that molecules we think of as morphogens also function as axon guidance molecules. It is an open question as to how many more types of guidance cues remain to be identified.

What Corey Goodman and I set out to do in that article was really to step back, and rather than take in the welter of detail emerging, we wanted to organize it and systematize it and pull out the major themes and principles that were emerging. And I think that’s why the review is referenced so frequently. It does lay out the principles in a way that is rather timeless, and as new discoveries have been made, they add to and elaborate on the principles we enunciated there. At least so far, they haven’t added whole new dimensions or invalidated the general conclusions we drew at the time.

It’s interesting. By coincidence, Corey and I were both asked to write chapters for a textbook edited by Max Cowan. We asked for permission to pool our efforts. That joint chapter was published in 1997, but it was written before we wrote our Science review. That chapter is 71 pages long with 14 pages of references. So in it we really took a broad historical perspective, because we wanted to be comprehensive. But in writing the chapter, we were able to organize our thoughts. When we were then asked to write the review for Science, we were able to distill our thoughts since we’d already gone through the field in a comprehensive way. It was fortuitous that we’d done that, and it worked out very well. We had an opportunity to make it better, and to make it more concise.

It’s gone in a few different directions. First of all, a number of other examples of these guideposts, or signposts, have been identified—the slits, first and foremost. As another example, the receptors for the semaphorins weren’t known at the time. They turned out to be receptor complexes. So there has been some fleshing out of the ligands and the receptors. And then, as I mentioned, there have been a lot of studies aimed at determining specific contributions of individual members of these families, large multi-genic families. There are about 20 semaphorins, for example. Identifying what each of these molecules does has required a lot of work, and it has helped us refine our understanding of how they contribute to axon guidance. As we’ve gone along, evidence has been accumulating that other factors contribute to axon guidance as well—the morphogens, for instance. So the first advance has been to flesh out the identity of signals and receptors, and determining specific contributions of these signals and receptors to axon guidance events.

Part two is the signal transduction mechanisms downstream of the receptors. The field has been interested in the question: once the ligand binds the receptor, how does it convey its information? That information might be attraction, repulsion, stopping, promoting extension, and so forth. So obviously the receptor has to engage a signal transduction pathway downstream. There’s been a lot of progress there, identifying the key molecular players in these pathways. Perhaps not surprisingly, the signal transduction pathways have a lot to do with cell motility pathways in general. Today, in 2005, many of the players have been identified, but how the whole fits together to make the machine work as effectively as it does is, however, still poorly understood. It has been said that we are still in the "Baroque" period in this field: we have a lot of the players identified, but we don’t know how they all fit together.

A third direction in which the field has expanded is in applying insights from the development of axonal connections to the problem of regeneration. When axons are severed in the adult, the cut end, the part that is still attached to the cell body, will reform a growth cone and try to re-grow. It does this more or less successfully in the peripheral nervous system but not in the central nervous system—in the spinal cord and the brain. And so there are mounting efforts from a number of laboratories, including ours, to determine to what extent the principles and molecules that are used to wire the nervous system are also reused to rewire of the nervous system.

And then there’s a fourth area, which is understanding the full morphogenesis of neurons. Neurons will send out axons that are guided, but these axons will also branch. A number of groups, including ours, have sought to identify the molecules that regulate branching. It turns out that the molecules that guide axons are also used to regulate branching as well. These molecules are also involved in regulating dendrite development—the other end of the neuron. To make the story complete, all these molecules tend to be involved in regulating the morphogenesis of non-neural tissues as well. A dramatic example is the morphogenesis of blood vessels. It turns out that all the major families of axon guidance molecules also regulate vascular morphogenesis—the patterning of blood vessels, arteries and veins and capillaries. For example, last October we and our collaborators showed that the netrins are involved in regulating vascular morphogenesis.

In short, there are many ways that the field is both staying its original course, focused on specific issues of guidance of axons to targets, but also intersecting wonderfully with a whole range of other studies on growth, guidance, migration, and morphogenesis of other tissues and organs, and also not just in development but also in regeneration.

I was quite pleased with the review when we finished it. I think if you read it today, the principles we enunciate there have held true. And the way we organized it, the way we conceptualized the field, is certainly the way that I continue to view the field, the prism through which both I and other workers in the field look at it. So there isn’t much I would change. I think perhaps the one additional thing we could have done would have been to have a section on regeneration. At the time, there was very little data. It would have been mostly speculation, but we could have perhaps tried to create a framework for thinking about regeneration.

  What is the biggest challenge in the field today? Or the biggest challenges?

It’s still the case that there isn’t a single neuron in the body where we can say that, when that neuron sends out its axon and the axon grows all the way to its target, that we know all the molecules guiding it to the target. In no case have we been able to tell the story of a guidance event in full. Until we can do that, we won’t know just how close we are to having a full explanation of guidance, as opposed to many robust partial explanations of the phenomena. So getting a full view of the process is still the big challenge. We don’t know how close or far we are, because each investigator in the field focuses on a particular neuron, a particular portion of its trajectory. I think we have to make a big push to understand all aspects of the trajectory of at last one class of neuron. That would be a step forward, but it’s a challenge. If you want to extend it further, then understanding at a mechanistic level in the growth cone, the signal transduction mechanisms that integrate and transduce the guidance signals into direct motion and motility. That is a big challenge, and we are still a far way off from having that view. Those are the two challenges I would single out as perhaps the most significant.

That’s a good question. You take away the pieces you think are necessary and you ask yourself when you’re done if you have the behavior of an axon that is unguided, or whether some guidance still remains. That is the best operational way of addressing that issue.

Marc Tessier-Lavigne, Ph.D.
Genentech, Inc.
San Francisco, CA, USA