Hands and feet—and the digits that extend from them—emerged about 400 million years ago, endowing the first land animals with fine-motor control, a major turning point in evolution.

A study published last week in Neuron looks at the genes that are switched on and off in an embryo to wire up fingers and toes to the central nervous system. Front and center in this process is the motor neuron, the nerve cell that controls movement.

Motor neurons in the spinal cord extend long, wirelike tendrils all the way out to fingers and toes. Signals that pass along these “wires” control movements for grasping, and enable the dexterity required for making tools, writing with a pen, playing a musical instrument or typing tweets and texts on a smartphone.

The molecular-scale view provided in this latest study suggests how the wiring process may have evolved. It may also provide insight into diseases, such as amyotrophic lateral sclerosis (ALS), which begins with muscle weakness in the extremities. Study co-author Thomas Jessell, who co-directs Columbia University’s Zuckerman Institute, talked to Scientific American about the findings and their implications.

[An edited transcript of the interview follows]

What experiments did you do for this study, and what were your main findings?
The big question is: How do you control the innervation [or the connection to the central nervous system] of digit muscles? To try to answer that, we looked for molecular differences that could reveal aspects of motor neuron diversity—and the study started with a molecular screen to try and find genetic differences between different subtypes of motor neurons. Alana Mendelsohn [an MD/PhD student in Jessell’s lab] did the screen using a genetic-screening method (RNA sequencing), and she found that motor neurons innervating the shoulder, elbow and wrist are similar to each other—but those innervating the digits are quite different from the rest.

We found 20 genes that distinguished motor neurons that innervate digit muscles from the others, and there’s a strange gene code involved in the cells’ development. The motor neurons that innervate the digits switch on master developmental control genes called Hoxc8 and Hoxc9—which is strange, because these two genes are never co-expressed by any other groups of motor neurons. Digit-innervating motor neurons also lack the RALDH2 gene, which encodes an enzyme that synthesizes a key developmental biochemical called retinoic acid, and is critical to the development of all other types of motor neurons.

We also forced motor neurons to express other Hox genes and exposed them to retinoic acid, both of which had a dramatically adverse effect on their differentiation. That provided additional support for the idea that motor neurons need to avoid retinoic acid to acquire this particular fate. So if you’re trying to make neurons different from one another, you can do that by adding something to one set of neurons, or by subtracting it. We weren’t looking for this, so it was a big surprise.

What does this tell us about how fine-motor control evolved?
What I find most interesting is the idea that the embryonic tissue that goes on to form limb and the motor neurons is regulated by coordinated molecular mechanisms—under the guidance of a genetic program that has been conserved over the course of evolution. The same genes—RALDH2 and the Hox genes—control motor neuron development, and one intriguing theoretical postulate is that RALDH2 may also be involved in limb development.

Prehistoric bony fish were known to have motor neurons that innervate their appendages, such as fins. The unique Hox gene coding that we observed in this study is likely to have evolved in a transitional organism, in which fins began to develop a more limblike structure.

Whales today don’t have digit muscles, but some fishes gradually acquired embryonic digit-forming tissue in their fin muscles. The zebra fish, for example, has some semblance of digit-patterning tissue, and so the big question is: How do these differences emerge? How do digit-innervating motor neurons get to be so different from the others? And how do they generate long-distance fibers that stretch out to the digit muscles as opposed to short ones that are going to innervate, say, a shoulder muscle? I wouldn’t say that we have provided any great insight into that, other than to say that neurons acquire the ability to grow that extensive distance by virtue of their genetic profiles.

The limbs pattern [structure] themselves and the motor neurons that innervate them pattern themselves in parallel. All motor neurons work in the same way. But how do they choose to project to the extremities, where you find the digit muscles? Our findings tell us that two things have to happen: One is that Hox genes must be altered by evolution to get this strange variant code. The other is that you have to avoid retinoic acid if you want to have a chance of being a motor neuron that innervates digit muscles. Maybe high retinoic acid levels in the extremities would make the motor neuron axons freak out when they get there.

Do your findings have any therapeutic potential?
They’re relevant in the context of treating diseases like amyotrophic lateral sclerosis, a motor neuron degenerative disease that starts with weakness in the extremities and eventually leads to complete loss of all motor function. Until now researchers trying to develop treatments wouldn’t have known how to generate the motor neurons that innervate the limb digit muscles, but our findings give them an opportunity to do so.

How are you following up this research?
We know that Hox genes encode transcription factors that control the activity of other genes, but we found that digit-innervating motor neurons also express other genes, such as FIGN and CNEP4], and we still don’t know the function of any of these other genes. If we knew what those functions were, then maybe we’d have a better way of manipulating these cells, so I’d be keen on performing some experiments to explore what these other genes do. We also found that digit-innervating motor neurons express signaling molecules called bone morphogenetic proteins and other proteins called SMADs, and so we need to explore what they’re doing in some detail, in order to resolve whether there’s anything interesting going on there.

Motor neurons are vastly complicated, and we’re trying to understand them in relation to their function, so there all sorts of other questions we could choose to examine. What would happen, for example, if we prevented digit-innervating motor neurons from being generated during embryonic development? Those digit muscles will just sit there without receiving a nerve supply. But what would they do then? Would they waste away without a nerve supply?