Think about flexing your toe.
For a second, visualize the impulse, just reaching the back of the frontal lobe - the primary motor cortex.
Descend. Go down the hemispheres, through the front of the brainstem. Cross to the other side. Go down the spinal cord — almost all of it. Once you’ve reached the lumbar region, connect. Leave the spinal canal. Find similar branches and converge. Exit pelvis. Go down the back of the thigh, nestled in between thick muscle bodies. Branch out. Continue down the back of the leg, then the foot, all the way to the big toe. Move.
An electrical impulse just traversed the whole length of your body. For this, it used the most efficient highways you possess - myelinated axons, propagating impulses at up to 120 m/s (or, in race-car terms, 430 km/h).
But how do these highways come to be? How does an axon, essentially an elongation of a cell, know how to reach and connect to a particular muscle, or portion of skin, or another neuron?
Today, we’ll take a look at this embedded GPS axons possess, which has fascinated neuroscientists for quite a while. In short, we’ll cover:
As the nervous system develops, neurons migrate to what is to become their final home. Be it the temporal lobe, the spine, or a ganglion, they reach their destination and afterwards begin connecting. They will emit one or more dendrites and most often just one axon.
The axon extends through the so-called growth cone, a structure that permanently emits finger-like projections, exploring the environment and guiding the movement towards a pre-determined target.
Watch below how axons from the eye of the Xenopus frog extend during development to reach the brain:
Inside the growth cone, cytoskeletal structures - like microtubules- self-assemble to sustain the axon’s elongation, much in the way a scaffold holds a building under construction.
On the left, you can see a close-up of a growth cone, with its cytoskeleton marked - the microtubules in green, and actin filaments in red.
Many other elements are necessary to sustain this impressive growth - axonal membrane needs to be produced, as well as different components of the cytoplasm.
Keep in mind that most of the protein synthesis actually takes place in the neuron’s cell body, so appropriate building materials need to be produced and then accurately sent to the growth cone area.
A growth cone does more than just ‘grow’. Far from being a structure that simply elongates with no direction or purpose, it moves in a way that allows the axon to explore the environment and effectively decide which way to go.
Watch below as axons decide to pursue or, alternatively, avoid different substances added in their way:
A variety of receptors sit on the outside of the growth cone membrane, simply waiting for the right signal. Once they’ve detected the molecule they were designed for, they can trigger a whole downstream signaling cascade that determines the axon to either grow towards or away from it. Taking this information in, the growth cone then effectively reassembles its cytoskeleton - re-thinking the way it will emit its next filopodia.
Guided by a multitude of these receptors at the same time, the axon’s path becomes somewhat restricted. By avoiding several locations while intently moving towards others, the growth cone is steering towards its correct target.
Axons are surprisingly efficient in finding their goal. They do not stray from their stereotyped path, nor do they go by trial and error. The inescapable truth that axons are, in fact, guided, has been proven many times by studying the early development of the nervous system.
The very first axons that extend out of the nervous system (sometimes referred to as pioneers) do so in a highly stereotyped and conserved manner. Ultimately, more and more axons are added onto a path, creating the nervous tracts present in the adult.
Observing that the first axons must be guided, given that they always extend across the same paths, led to a clear conclusion: Specifically localized guided information must be available in the developing brain, and growth cones must have a way of detecting and responding to these guidance cues.
This then begs the question: What are the guidance cues, and how do axons sense them?
Axons are clearly guided by something, and that something may be certain chemicals in their environment - some attracting them, some pushing them away.
This hypothesis was pioneered by the very person that discovered growth cones in the first place, Ramon y Cajal:
“If one admits that neuroblasts are endowed with chemotactic properties, then one might also imagine that they are capable of ameboid movements, initiated by factors secreted from epithelial, neural, or mesodermal elements. As a result, their processes may be oriented in the direction of chemical gradients, and thus guided to the secreting cells”
Ramon y Cajal 1892
This surprisingly modern outlook matches one of the best explanations we have for axon guidance so far: gradients.
As far back as the ‘80s, researchers have shown that the spatial distribution of guidance cues may determine an axon’s final path. To do this, they often used neurons in the retina as a model - given the fairly convoluted way they take in order to reach the brain, involving several twists and turns.
Over time, many of the cues retinal axons use in their path have been discovered. It has been found that the visual pathway can be divided in several molecularly distinct segments, from the retina all the way up to the central nervous system. Proteins such as netrins, slits, semaphorins or ephrins mark these domains, effectively generating a map for the developing axon to follow.
Not long after that, a new hypothesis arised, one that still stands today: axon guidance doesn’t only rely on the presence of a cue substance, but potentially also on its distribution. Thus stemmed the idea of gradient guidance. This idea presumes that gradients of the cue substances are formed very early on in the embryo, and then cautiously followed by developing axons. The growth cone is then expected to detect a favourable change in the concentration of a cue substance, rather than just its presence, allowing for a lot more precision and room for re-routing if needed.
As a matter of fact, more recent studies attest that, during their complex orientation, growing axons may not only detect concentration gradients, but also temperature gradients. And the influence of physical factors doesn’t seem to stop here.
Interestingly, it appears that signal molecules are not the only factors that can guide an axon’s movement. Mechanical stimulation may also play a role.
In a fascinating experiment performed more than 2 decades ago, D. Bray set out to physically interact with axonal growth cones. He devised a system in which he attached microelectrodes to the very tip of a growth cone, and pulled at it in a controlled manner.
Impressively, this led to the axons elongating in a significant manner. Not simply stretching, but properly elongating by means of protein synthesis and appropriate signaling pathways. After a few hours under this permanent tension, some axons had more than doubled in length compared to their control counterparts.
This led scientists to expect mechanosensation - and, in that sense, mechanosensitive molecules, to play a role in axon extension and perhaps even pathfinding.
On one hand, the growth cone creates a certain degree of tension on the elongating axon. Perhaps maintaining that tension is what sustains elongation, as D. Bray noticed in his towing experiments.
On the other, the axon may find and wish to avoid physical obstacles in its path. Further studies seem to attest to this necessary equilibrium.
A 2016 paper explored how the stiffness of the environment affects axon growth. Indeed, axons seem to prefer softer substrates even in the absence of other signaling molecules to guide their paths. Furthermore, getting rid of the molecules that are supposed to mediate mechanosensitivity led to aberrant growth and pathfinding errors.
One such molecule, which is of particular interest in development, is Piezo 1 - an ion channel named after its apparent ability to transform mechanical stimulation (stretch) into electrical signaling.
In a 2019 paper, Song et al. have found that Piezo 1 plays a role in axon regeneration following injury. Supposedly, when an axon is no longer inhibited by mechanical stimuli from the outside, it can grow unhindered, effectively regenerating at a higher rate post-injury.
Although mechanosensitivity certainly complicates the issue of axon guidance, it most certainly does not complete it. Many other variables need to be taken into account - such as the role glial cells play, how some axons prefer dendrites and others prefer muscles, how the growth cone of the first developed neuron differs from that of the last, and so on.
Pinpointing precisely how axons grow and orient themselves during development remains a challenge. This process only employs a minute amount of trial and error, compared to the way some pathfinding algorithms in computers may work. Axon guidance works within a highly engineered and very finely tuned system. Growth cones don’t get lost on the way and don’t have the option to start from the beginning. Instead, they constantly explore and rely on a myriad of inputs to guide their next movement, and the next one after that, until they reach their final target.
Further insights into this system may allow us to understand why is it that axons are so adept at growing and guiding themselves during embryonic life, yet seem almost helpless after an injury in the adult brain. Ultimately, we may even find ways to help them along in the latter case.
Further reading: How does an axon grow? Goldberg, J. Genes & Dev. 2003. 17: 941-958