Fig. 1. Do growth cones have a map?

Fig. 2. Is it a nose...?

Fig. 3. ...or a hand...?

Axon guidance

Think about wiggling your big toe. How did a nerve grow out from your spinal cord, all the way down your leg and find the correct muscle in your foot? Why didn’t it stop at one of the dozens of other muscles en route? How did it find its way into the leg or even ‘know’ that it should go there? (Fig. 1).

As the axon grows out towards its target, it is guided by a specialised structure at its tip called a growth cone. This is sometimes described as being like a nose (Fig. 2) sniffing out chemical signals that direct it. As you can see in the title bar, it has long finger-like projections which it is thought to use to feel its way around (Fig. 3) so perhaps it is more like Mr. Tickle than Mr. Nosey. Maybe a better analogy for this combination of sensitivity to touch and smell would be a mouse's whiskers or a beetle's antennae.

As far as we know the axon doesn’t have an innate map of where it is, or should be, going. It just has a drive to extend which is then moulded by its environment to shape the final pattern of nerve growth. This makes sense when you consider that this is taking place in a developing embryo so its surroundings are constantly changing and expanding.
The environmental cues that it picks up can be broadly thought of as either positive, promoting growth towards it, or as negative, causing the axon to shrink back and turn away (Fig. 4). Serial combinations of these signals can then steer the axon to its goal. This behaviour is straightforward to reproduce by growing neurons in a dish, on a surface patterned with alternating areas of attractive and repulsive proteins.

The growth cone carries receptor proteins on its surface. When these bind specific partner proteins, called ligands, it triggers cascades of signals inside the growth cone and axon which will change their behaviour. So, for instance, when an attractive ligand binds to its receptor it stimulates the growth cone to extend the axon. If the receptor or the ligand are asymmetrically distributed then the growth cone will turn towards it. The reverse happens with a repulsive ligand to steer an axon away. In principle this neatly explains how a nerve will grow towards a certain muscle and away from another one. This theory is supported by experiments: the same pairings of receptors and ligands that attract or repel growth cones in a dish in the lab are found on nerves and muscles in developing embryos. This process operates in flies, fish, frogs, mice and men. If the genes that code for the receptors or ligands are experimentally knocked out in any of these animals then nerves no longer find their correct muscle. In the case of humans there are congenital disorders in which this occurs, see our research for more information about these.

Fig. 4. Growth cones turn towards positive cues (green) and are repelled by negative ones (red).

It’s not just nerves and muscles that are paired up by this system, the same mechanism is used by neurons to navigate to other areas of the brain or nervous system. However, spotting even subtle changes in the registration between nerves and muscles much simpler than finding differences among the millions of interlinked pathways within the brain. There are plenty of well-documented examples of brain tracts that are disrupted when certain guidance proteins are lost.
For many guidance signals it may be that the changes are too subtle to be identified or simply that no-one has looked in the right place. There is an ambitious and exciting project under way to map all the brain’s connections which, if successful, will provide a valuable reference to help understand the overall wiring pattern.
Let us imagine that we had a complete description of all the neuronal connections, we still need to unravel how they formed. There are millions more pathways than there are genes in the entire human genome so how does each neuron get a specific label to reach its destination? We know that one gene can produce several versions of a protein, indeed one gene known to be important for brain development can produce 38000 different protein isoforms! So it is possible that each neuron could display a unique combination of receptors on its surface but this raises another, equally difficult, problem in its place – how is this labelling co-ordinated to make sure that each neuron wears a unique tag?

See also: