During development, motor neurons in the spinal cord become progressively specialized until they're destined to make specific connections with one muscle target. Without receiving specialized signals during embryogenesis, a naive motor neuron wouldn't know where to send its axons or which muscles to connect to. But what turns one cell into a motor neuron that reaches down into fingers and another into a cell that connects with more nearby shoulder muscles is still largely unknown.

"Understanding how different types of neurons are generated is a big question in developmental biology and also for clinical researchers," says Dr. Jeremy Dasen, a postdoctoral research fellow with Dr. Tom Jessell, professor of biochemistry and molecular biophysics in the Center for Neurobiology and Behavior. "If we're going to use cells as therapies, we need to understand the steps that direct motor neurons to their particular fates."

Now the two researchers have identified a set of signals in the developing spinal cord that start to turn generic motor neurons into more specialized cells. The research was published in the Oct. 30 issue of Nature.

The new research reveals that Hox-c proteins control one of the first events in the specialization process, the grouping of motor neurons into different columns. Each motor column spans several segments of the spinal cord, and once grouped into a column, the neurons send axons to common targets in restricted regions of the body. For example, in the chick, neurons in the lateral motor column (LMC) only connect with upper limb muscles, while the neurons in the Column of Terni (CT) just below, only connect with sympathetic ganglia (which then connect to organs like the kidneys and pancreas).

Drs. Dasen and Jessell first saw that, before columns are formed, two Hox-c proteins appear in the spinal cord in areas that correspond to two future columns. Hoxc6 was expressed only in cells fated to become the LMC, while Hoxc9 was expressed only in cells fated for the CT, suggesting that the two proteins could be involved in directing cell fate.

To see if the Hox-c proteins controlled the cells' identities, Dr. Dasen then switched Hox-c expression patterns in chick embryos and looked at protein markers to determine the identity of each neuron. In one set of embryos, when he made all the motor neurons express Hoxc6, the cells changed fates and became LMC neurons. The opposite happened in a second set of embryos, where he made all the motor neurons express Hoxc9. In these embryos, nearly all the motor neurons became CT cells.

The neurons not only looked like they had changed identity, but they also acted like it. When Dr. Dasen changed neurons originally slated for the LMC into CT neurons, the cells acted like true CT cells and directed their axons to sympathetic ganglia, not LMC targets in the limb muscles.

"Hox-c proteins really tell motor neurons what to do," Dr. Dasen says.

But what tells Hox-c proteins what to do and where to be turned on? Drs. Jessell and Dasen also answered these questions by finding that a gradient of fibroblast growth factor signaling along the length of the spinal cord determines the initial Hox-c pattern. A low concentration of fibroblast growth factor near the top of the spinal cord sets up Hoxc6 expression, while increasing amounts toward the middle set up Hoxc9. The researchers did not look at motor columns near the base of the spinal cord, but they think that fibroblast growth factor signals may also turn on another Hox protein, which then organizes a column of neurons that send connections to leg muscles.

Though column identity tells a motor neuron where to go in general, the neurons only become destined for a specific muscle after they are grouped into smaller motor pools within each column. All the neurons in a given motor pool connect to one specific muscle target.

"Nobody knows exactly how motor neurons are constrained to a particular pool identity," Dr. Dasen says, "but because different pools express different Hox proteins, it's likely that Hox proteins are involved and control which specific muscle is targeted."

He adds, "A role for Hox proteins in motor pool identity would be very exciting, as it would provide a general mechanism for how specific connectivity patterns emerge in the central nervous system and would provide a starting point for exploring more refined aspects of locomotor function."

Tom Jessell is a Howard Hughes Medical Institute investigator.