Roles of Hedgehog pathway components and retinoic acid signalling in specifying zebrafish ventral spinal cord neurons.

In mouse, Hedgehog (Hh) signalling is required for most ventral spinal neurons to form. Here, we analyse the spinal cord phenotype of zebrafish maternal-zygotic smoothened (MZsmo) mutants that completely lack Hh signalling. We find that most V3 domain cells and motoneurons are lost, whereas medial floorplate still develops normally and V2, V1 and V0v cells form in normal numbers. This phenotype resembles that of mice that lack both Hh signalling and Gli repressor activity. Ventral spinal cord progenitor domain transcription factors are not expressed at 24 hpf in zebrafish MZsmo mutants. However, pMN, p2 and p1 domain markers are expressed at early somitogenesis stages in these mutants. This suggests that Gli repressor activity does not extend into zebrafish ventral spinal cord at these stages, even in the absence of Hh signalling. Consistent with this, ectopic expression of Gli3R represses ventral progenitor domain expression at these early stages and knocking down Gli repressor activity rescues later expression. We investigated whether retinoic acid (RA) signalling specifies ventral spinal neurons in the absence of Hh signalling. The results suggest that RA is required for the correct number of many different spinal neurons to form. This is probably mediated, in part, by an effect on cell proliferation. However, V0v, V1 and V2 cells are still present, even in the absence of both Hh and RA signalling. We demonstrate that Gli1 has a Hh-independent role in specifying most of the remaining motoneurons and V3 domain cells in embryos that lack Hh signalling, but removal of Gli1 activity does not affect more dorsal neurons.

Zebrafish V2 cells develop into excitatory CiD and Notch signalling dependent inhibitory VeLD interneurons.

The vertebrate spinal cord contains distinct classes of cells that form at precise dorsal-ventral locations and express specific combinations of transcription factors. In amniotes, V2 cells develop in the ventral spinal cord, just dorsal to motoneurons. All V2 cells develop from the same progenitor domain and hence are initially molecularly identical. However, as they start to become post-mitotic and differentiate they subdivide into two intermingled molecularly-distinct subpopulations of cells, V2a and V2b cells. Here we show that the molecular identities of V2a and V2b cells are conserved between zebrafish and amniotes. In zebrafish, these two cell types both develop into interneurons with very similar morphologies, but while V2a cells become excitatory Circumferential Descending (CiD) interneurons, V2b cells become inhibitory Ventral Lateral Descending (VeLD) interneurons. In addition, we demonstrate that Notch signalling is required for V2 cells to develop into V2b cells. In the absence of Notch signalling, all V2b cells develop as V2a cells.

Pax2/8 act redundantly to specify glycinergic and GABAergic fates of multiple spinal interneurons.

The spinal cord contains several distinct classes of neurons but it is still unclear how many of the functional characteristics of these cells are specified. One of the most crucial functional characteristics of a neuron is its neurotransmitter fate. In this paper, we show that in zebrafish most glycinergic and many GABAergic spinal interneurons express Pax2a, Pax2b and Pax8 and that these transcription factors are redundantly required for the neurotransmitter fates of many of these cells. We also demonstrate that the function of these Pax2/8 transcription factors is very specific: in embryos in which Pax2a, Pax2b and Pax8 are simultaneously knocked-down, many neurons lose their glycinergic and/or GABAergic characteristics, but they do not become glutamatergic or cholinergic and their soma morphologies and axon trajectories are unchanged. In mouse, Pax2 is required for correct specification of GABAergic interneurons in the dorsal horn, but it is not required for the neurotransmitter fates of other Pax2-expressing spinal neurons. Our results suggest that this is probably due to redundancy with Pax8 and that the function of Pax2/8 in specifying GABAergic and glycinergic neuronal fates is much broader than was previously appreciated and is highly conserved between different vertebrates.