Life on the road

As nerve axons migrate during wiring of the nervous system, molecular signposts at intermediate targets show them the way. Such a target has now bee n found to keep axons alive — provided they're on the right track.

To form the connections of the nervous system, axons must find their targets by setting out on a journey that may be long and complicated. How do they select the right path? What happens to those that stray off in the wrong direction? Intermediate targets are known to act as signposts along the way, sending out guidance molecules that tell the axons which way to go. On page 765 of this issue, Wang and Tessier-Lavigne¹ show that intermediate targets may have another function — they can also support the survival of migrating axons. Such a mechanism, by keeping axons alive only if they follow the right route, could ensure that the correct neural connections are formed.

The concept that nerves derive support from the targets they ultimately innervate can be traced back more than a century². The best-known way in which this support — now known as a neurotrophic effect — occurs is by prevention of programmed cell death. The idea is that, during development, an excess of neurons is produced initially. Their axons then compete for a limiting amount of neurotrophic factors in the target. The result is a match in the sizes of the nerve and target, as well as elimination of mistargeted axons. A molecular basis for this phenomenon has been established through studies on nerve growth factor, the related neurotrophins, and other molecules that act as neurotrophic factors in final targets³.

Wang and Tessier-Lavigne¹ set out to investigate whether similar principles might apply to an intermediate target — an idea they refer to as en passant (in passing) neurotrophic activity. As their model system they chose spinal commissural axons (those that cross from one side of the spinal cord to the other), and their intermediate target was a wedge-shaped group of cells called the spinal-cord floor plate⁴ (Fig. 1a). Commissural axons are guided towards the floor plate, at least in part, by diffusible chemoattractants belonging to the netrin family. After crossing the midline, these axons turn and grow alongside the floor plate towards the brain, eventually heading off to reach their final targets.

Figure 1: Commissural neurons and their intermediate target, the floor plate.

a, Commissural neurons in the rat dorsal spinal cord send out their axons, which soon turn (1) towards the floor plate, guided by chemoattractants. They then cross the floor plate (2) and turn again to grow alongside it (3) in the direction of the brain. Later they leave (4) to reach their final targets. b, Wang and Tessier-Lavigne¹ grew commissural axons from spinal-cord explants in vitro. They found that the explants degenerated at about the time they should normally have reached the floor plate. However, floor plate-conditioned culture medium (FPCM) contains a neurotrophic activity that can rescue them. Later, FPCM is not sufficient, but the axons can be rescued by FPCM if molecules in the neurotrophin family (NT) are also added.

As a first step, Wang and Tessier-Lavigne developed an in vitro assay. In the presence of netrin-1, rat dorsal spinal-cord explants send out commissural axons. But within days the axons degenerate and the cell bodies die. Could this degeneration be rescued by the intermediate target? The authors found that the axon degeneration and cell death were indeed dramatically rescued with floor plate-conditioned culture medium (FPCM), revealing a powerful neurotrophic activity.

Is the activity localized to the intermediate target? The molecule responsible has not yet been identified, and, although it seems to be a polypeptide, tests of many candidates failed to identify any with appropriate activity. So, Wang and Tessier-Lavigne tested the localization by assaying slices taken from different parts of the spinal cord. They found that the activity is highly concentrated in the floor plate, with some also in more dorsal regions. These results indicate either lower production of the activity in dorsal regions, or diffusion of a signal produced in the floor plate.

If this activity is to fit the en passant neurotrophic model, a key issue is whether it can promote survival of cell bodies by acting at a distance on their axons. This was shown for nerve growth factor in classic experiments using the ‘Campenot chamber’, where axons were separated from their cell bodies by growth through a sealed barrier. Spinal commissural axons grow too slowly for this approach to be used, but Wang and Tessier-Lavigne developed an innovative and persuasive alternative. They placed a dorsal spinal-cord explant next to a floor plate explant in a collagen gel. By examining the axons and staining for dying cell bodies, they showed that, independent of the location of the cell bodies, the cells survived only if the axons had grown near the floor plate. This result is consistent with the idea that survival of commissural neurons in the embryo could depend on the distant journey of their axons.

So spatial aspects seem to fit the model — but what about time? In vivo, commissural axons grow out between embryonic day 11 (E11) and E13, and take about a day to reach the floor plate. Wang and Tessier-Lavigne found that, in the absence of FPCM, all of the dorsal spinal-cord explants looked healthy at the equivalent of E14. Yet, just one day later, all had begun to degenerate (Fig. 1b). This was true irrespective of whether the explants were placed in culture at E11 or E13, suggesting that a clock runs in the dorsal spinal cord in vivo or in vitro. This clock sets off a requirement for neurotrophic support just as the commissural axons are supposed to be reaching the floor plate. Axons that don't get there by the time the clock strikes presumably suffer a fate worse than Cinderella's.

Later, after the developing axons grow away from the floor plate, one might expect them to come to depend on neurotrophic activity from their final targets. Consistent with this idea, Wang and Tessier-Lavigne found that, at the equivalent of E18, axons can no longer be supported by FPCM. Instead, they need a combination of FPCM and neurotrophins (Fig. 1b). This could fit with a two-step model of development, in which the growing axons first depend on an en passant signal from the intermediate target, and later need an additional signal from the final target.

The en passant neurotrophic activity identified in these elegant studies could help solve the neural wiring problem in several ways. First, if axons go astray, this mechanism might help to eliminate them quickly, before they interfere with the orderly pioneering and assortment of axon tracts. Second, it could prevent axons reaching the wrong final target, where they might otherwise be incorporated in aberrant neural circuits¹. Third, it opens the possibility of a combinatorial mechanism, where a limited number of factors derived from the intermediate or final targets could be used in different combinations to specify many distinct connections.

Wang and Tessier-Lavigne's work adds a fascinating new dimension to the increasing recognition that neurotrophic effects may go beyond the simple model of support by final targets. During development, the responsiveness of some neurons to different neurotrophins switches. Although this has not been tied unequivocally to intermediate targets, neurotrophins may contribute support at cell bodies or along pathways, while axons are still on the way to their targets⁵. There is also some analogy in later events, when motor neurons require support both from their muscle target and from the Schwann cells that wrap around their axons after reaching the target³. We also have several new questions. What is the precise developmental significance of the floor plate activity detected in vitro? Could the work have therapeutic implications in spinal-cord regeneration? What molecules are responsible for the activity? We don't have all the answers yet — but after all, life's a journey.