Linking Motor Neuron Identity and Muscle Target Connectivity
The vertebrate central nervous system contains hundreds of different classes of neurons, each forming synaptic connections with a selective set of target neurons. Insight into the way in which neurons establish precise patterns of connectivity has begun to emerge from the study of spinal motor circuits. One fundamental aspect of motor circuit assembly is the formation of precise connections with target muscles in the developing limb. Mammalian limbs typically contain more than 50 different muscle groups, demanding the generation of an equivalent number of motor neuron subclasses, and posing a formidable challenge for growing axons as they seek out their particular muscle target.
How do motor neurons acquire specialized identities that direct their axons to specific muscle targets? We have found that three core properties of the motor neurons that innervate limb muscles—their diversity, their stereotyped position, and their connectivity—are established by a Hox transcriptional regulatory network. These Hox interactions constrain motor pools to specific rostrocaudal levels of the spinal cord and drive the diversification of motor neurons at a single segmental level. This Hox regulatory network also directs motor neuron connectivity with limb muscles—altering motor neuron Hox profiles changes target muscle connectivity. In turn, Hox genes determine the specificity of target muscle connections by regulating downstream transcription factors that exhibit more restricted profiles of expression. LIM homeodomain proteins establish motor axon trajectories by controlling the axonal expression of Eph kinase receptors. Other Hox-regulated transcription factors, notably ETS proteins, control the expression of cadherins—cell surface recognition molecules that promote the clustering of motor neurons into coherent pools. Thus, the challenge of specifying a hundred or more distinct subclasses of motor neurons, each projecting to a specific target cell group, is achieved through the regulatory interactions of a structurally related and chromosomally clustered set of Hox transcription factors.
We have also analyzed the contribution of the entire cohort of 21 motor neuron Hox proteins to topographic motor mapping through inactivation of the gene encoding an essential Hox cofactor, FoxP1. In mouse FoxP1 mutants Hox activity is lost and all molecular features of motor neuron subtype are erased, yet the overall pattern of motor nerve branching and limb muscle innervation is preserved. These findings indicate that motor nerve branching within the limb is constrained by pre-established permissive and inhibitory domains, and that the role of the FoxP1/Hox program is to provide motor neurons with identities that enable their axons to respond to local guidance cues that direct them to select just one of many available axonal conduits. Our findings imply that during evolution, the FoxP1/Hox program served a key role in expanding the range of motor neuron subtypes to ensure effective innervation of new peripheral motor targets.
The information content of the combinatorial FoxP1/Hox program far exceeds the requirements for motor neuron diversification, suggesting that Hox proteins have additional roles in the assembly of spinal circuits. Indeed, complex patterns of Hox protein expression also define spinal interneurons and sensory neurons, consistent with the idea that Hox proteins contribute more extensively to the formation of locomotor circuits. More generally, the self-organizing features inherent in this Hox transcriptional regulatory network may help to endow developing spinal motor circuits with their apparent high degree of genetic determination.
Sensory Feedback Control of Motor Output
The coordination of motor output depends critically on sensory feedback information provided by proprioceptive sensory neurons. The selectivity of proprioceptive afferent–motor neuron connectivity is thought to have its basis in the formation of distinct afferent termination zones in the spinal cord, as well as in the recognition of specific motor neuron targets.
We have found that the specificity of proprioceptive axonal inputs to motor neurons is controlled by two main classes of transcription factors, Runx and ETS proteins. The level of Runx3 expression by sensory neurons appears to be a primary determinant of the projection pattern of sensory axons within the spinal cord. In turn, the expression and activity of Runx3 in sensory neurons is gated by neurotrophin signals derived from the periphery, and by expression of the ETS transcription factor Er81. Elimination of Er81 function prevents the formation of direct connections between proprioceptive sensory afferents and motor neurons, abolishing coordinated motor output. These studies reveal that common sets of transcription factors control sensory and motor projection patterns in the developing spinal cord.
In studies to define the recognition systems that govern the specificity of synaptic connections formed between sensory and motor neurons, we have worked with Silvia Arber (Biozentrum, University of Basel) on the role of semaphorin-plexin signaling. We have found that a recognition system involving expression of sema3e by motor neuron pools, and its high-affinity receptor plexinD1 by proprioceptive sensory neurons, is a critical determinant of synaptic specificity in sensory-motor circuits. Changing the profile of sema3e-plexinD1 signaling in sensory and motor neurons leads to a functional and anatomical rewiring of monosynaptic connections. Thus, the specificity of sensory-motor connections in the mammalian spinal cord depends on a recognition system in which the matching expression of a sema ligand and its plexin receptor prevents synapse formation. Whether the intricate patterns of sensory-motor connectivity in the mammalian spinal cord emerge solely through layers of repellent filtering or also involve adhesive recognition remains to be determined.
Interneuron Circuitry and Motor Coordination
Local interneuron circuits have a major role in the coordination of motor behaviors. We have explored how information transfer at synaptic connections between proprioceptive sensory neurons and motor neurons is subject to two GABAergic inhibitory constraints: presynaptic inhibition of sensory transmitter release and postsynaptic inhibition of motor output. We have found that GABAergic interneurons mediating pre- and postsynaptic inhibition possess distinct transcriptional identities, differ in their GABA synthetic pathways, and form target connections with high precision. The assembly of these inhibitory circuits has its origins in the selectivity with which the two GABAergic interneuron classes respond to inductive and recognition signals provided by their sensory and motor targets. Molecular genetic manipulations show that construction of the presynaptic inhibitory circuit involves a stringent program of recognition specificity and synaptic differentiation that is directed by signals, including brain-derived neurotrophic factor, supplied selectively by proprioceptive sensory terminals.
Finally, the simple repetitive movements that underlie locomotion are also regulated by localized neural networks known as central pattern generators. Little is known, however, about the organization of the locomotor central pattern generator circuit in walking mammals, in part because of the difficulty in identifying and manipulating its intrinsic interneuronal components. We have used profiles of transcription factor expression to define distinct sets of ventral interneurons, each with a different intraspinal projection pattern and target connectivity. Genetic and physiological studies performed with the labs of Martyn Goulding (Salk Institute for Biological Studies), Kamal Sharma (University of Chicago), and Ole Kiehn (Karolinska Institute, Stockholm) have defined classes of interneurons with key roles in establishing left-right alternation in motor activity, and in the interneuronal circuitry that patterns locomotor behavior.