Dorsal root ganglion (DRG) neurons specifically project axons to central and peripheral targets according to their sensory modality. The Runt-related genes Runx1 and Runx3 are expressed in DRG neuronal subpopulations, suggesting that they may regulate the trajectories of specific axons. Here we report that Runx3-deficient (Runx3(-/-)) mice displayed severe motor uncoordination and that few DRG neurons synthesized the proprioceptive neuronal marker parvalbumin. Proprioceptive afferent axons failed to project to their targets in the spinal cord as well as those in the muscle. NT-3-responsive Runx3(-/-) DRG neurons showed less neurite outgrowth in vitro. However, we found no changes in the fate specification of Runx3(-/-) DRG neurons or in the number of DRG neurons that expressed trkC. Our data demonstrate that Runx3 is critical in regulating the axonal projections of a specific subpopulation of DRG neurons.

The formation of laminar-specific projections is a key event in the development of appropriate neuronal connections in many regions of the central nervous system. In order to provide a framework for defining functions of molecules related to spinal laminar targeting of dorsal root ganglion neurons in mice, we have characterized the initial trajectories of sensory axons in relation to the maturation of their target laminae in the spinal cord. We show that morphological and biochemical differentiation of distinct clusters of neurons in the dorsal region of the spinal cord precedes initial collateral branching from sensory axons. Between embryonic day (E) 12.5 and E13.5, sensory axons develop swelling ("nodes") along their entire intraspinal extent and elaborate interstitial collateral branches from these nodes. Collaterals from the different classes of sensory axons then penetrate the gray matter of the spinal cord sequentially. Each class of sensory axons projects directly to its target lamina, never branching into inappropriate laminae en route. Some cutaneous afferents traverse the entire width of the spinal cord to reach superficial laminae on the contralateral side, strictly avoiding both the ventral spinal cord and inappropriate laminae of the deep dorsal horn. The pathways taken by developing sensory afferents are compatible with the idea that cells in inappropriate laminae exert inhibitory influences on sensory axons which regulate their laminar specificity.

Runx1-deficient mice die around embryonic day 11.5 due to impaired hematopoiesis. This early death prevents the analysis of the role of Runx1 in the development of sensory ganglia. To overcome the early embryonic lethality, we adopted a new approach to utilize transgenic Runx1-deficient mice in which hematopoietic cells are selectively rescued by Runx1 expression under the control of GATA-1 promoter. In Runx1-deficient mice, the total number of dorsal root ganglion (DRG) neurons was increased, probably because of an increased proliferative activity of DRG progenitor cells and decreased apoptosis. In the mutant DRG, TrkA-positive neurons and peptidergic neurons were increased, while c-ret-positive neurons were decreased. Axonal projections were also altered, in that both central and peripheral projections of CGRP-positive axons were increased. In the dorsal horn of the spinal cord, projections of CGRP-positive axons expanded to the deeper layer, IIi, from the normal terminal area, I/IIo. Our results suggest that Runx1 is involved in the cell fate specification of cutaneous neurons, as well as their projections to central and peripheral targets.

Experiments over the past decade in which NGF/TrkA signaling has been abolished by antibodies or targeted gene mutations have shown that 70-85% of dorsal root ganglion (DRG) neurons require NGF for survival during development. There is consensus that many of the NGF-dependent neurons are small-diameter, peptidergic neurons subserving nociception. These neurons express the signaling receptor for NGF, TrkA. There is a major discrepancy, however, between the percentage of DRG neurons which require NGF for survival (70-85%) and percentage of DRG neurons expressing TrkA receptors (40-50%). The identity of these non-TrkA expressing, NGF-dependent neurons has not been established. A candidate group is a population of small DRG neurons with unmyelinated axons which bind BSI isolectins from the plant, Bandeiraea simplicifolia. We show here that most of these BSI-binding DRG neurons do not express TrkA in adult mice. However, in mutant mice in which NGF/TrkA signaling has been abolished by inactivation of the trkA gene, BSI-staining in the DRG and dorsal horn is completely eliminated. BSI-binding DRG cells are thus the first identified neuronal population in which cells do not express TrkA in maturity, but require NGF/TrkA signaling for survival during embryonic development. These neurons must either depend on NGF via a novel, indirect mechanism or alternatively, downregulate TrkA expression during development.

1. Spontaneous motoneuronal activity was monitored from the lumbar ventral roots in an isolated spinal cord preparation from rat fetuses at embryonic days (E) 13.5-18.5. 2. Spontaneous bursts that were synchronized in both left and right ventral roots were observed periodically (mean interval, 1.5-2.6 min) from E14.5 to 17.5. This activity was abolished in Ca(2+)-free saline or by application of tetrodotoxin (1 microM), indicating that it was synaptically mediated. 3. The glutamate receptor blocker kynurenate (4 mM) failed to block spontaneous bursts at E14.5-15.5, though it completely abolished them at E17.5. The glycine receptor antagonist strychnine (10 microM) completely blocked spontaneous bursts at E14.5-15.5. Bicuculline, a GABAA receptor antagonist, reduced the amplitude of the spontaneous bursts. 4. At E15.5, a brief application of glycine (250 microM to 2 mM) evoked excitatory responses resembling the spontaneous bursts in both time course and amplitude. Such glycine-induced responses were not observed under Ca(2+)-free conditions, suggesting that they were synaptically evoked. These synaptic responses were not blocked by kynurenate (4 mM), but they were abolished by strychnine (10 microM). 5. It is concluded that glycine and GABA generate the earliest spontaneous motor activity of the fetus and function transiently as excitatory transmitters in the embryonic spinal cord.