Robert R. Bernhardt

Increased expression of specific recognition molecules by retinal ganglion cells and by optic pathway glia accompanies the successful regeneration of retinal axons in adult zebrafish.

Retinal ganglion cells (RGCs) in adult zebrafish can regenerate their axons. We show that successful axonal regeneration is accompanied by the re‐expression by RGCs of mRNAs encoding specific recognition molecules that are expressed at high levels in the larval retina but are down‐regulated in the adult. Message levels for l1.1 and l1.2 (two homologs of mammalian L1), n‐cam (homologous to mammalian N‐CAM), beta3 (related to the beta3 and beta2 subunits of mammalian Na, K‐ATPase), and tn‐c (homologous to mammalian tenascin‐C) were high in larval RGCs undergoing axonogenesis and low in adult RGCs. After an optic nerve crush, axotomized adult RGCs showed increased levels of l1.1, l1.2, and n‐cam mRNA expression, whereas the levels of beta3 and tn‐cmRNA remained unchanged. The optic nerve crush also induced the expression of some of these mRNAs in the optic nerve and tract where they are not normally detectable. This lesion induced up‐regulation by presumptive glia was observed for l1.1, l1.2, n‐cam and beta3 but not for tn‐c. The combination of a neuronal (intrinsic) response to axotomy with an environmental (extrinsic) response may be an important determinant allowing for the successful axonal regeneration.

Pathfinding by identified growth cones in the spinal cord of zebrafish embryos

The spinal cord of early (18–20 hr) zebrafish embryos consists of a small number of neurons per hemisegment. The earliest neurons are identified and project growth cones that follow stereotyped, cell- specific pathways to reach their termination sites. We have studied the pathways taken by 4 of the early neurons in order to delineate the cells and structures their growth cones encounter during pathfinding. These neurons are 3 classes of commissural neurons (CoPA, CoSA, and CoB), which have contralateral longitudinal axons, and the VeLD neuron, which has an ipsilateral longitudinal axon. These growth cones encounter a defined set of cells and structures. Commissural growth cones appear to bypass the longitudinal axons of several identified neurons, including those from contralateral commissural neurons they encounter immediately following projection from the cell bodies. In contrast, these growth cones appear to extend in association with the longitudinal axons of commissural cells after crossing the ventral midline. Another set of cells of interest are the floor plate cells, a row of cells that constitute the ventral floor of the cord. At the floor plate growth cones exhibit cell-specific behaviors which may be influenced by the floor plate. (1) The floor plate may attract specific growth cones. The CoPA, CoSA, CoB, and VeLD growth cones all extend to the floor plate while other identified growth cones do not. (2) The floor plate may mediate cell-specific turns and induce some growth cones to cross the midline while inhibiting others from doing so. The commissural growth cones extend directly under the floor plate to cross the midline and turn anterior (CoPA and CoSA) or bifurcate (CoB); the VeLD growth cone turns away from the midline and extends posteriorly. (3) The floor plate may mediate changes in the substrate affinities of growth cones. Commissural growth cones bypass longitudinal pathways before they have encountered the floor plate, but not after. The description of pathfinding by these growth cones suggests that some elements in their environment are ignored while others are not. Most interestingly, a single structure (the floor plate) may mediate multiple, cell-specific effects on spinal growth cones.

Growth cone guidance by floor plate cells in the spinal cord of zebrafish embryos

The spinal cord of early zebrafish embryos contains a small number of neuronal classes whose growth cones all follow stereotyped, cell-specific pathways to their targets. Two classes of spinal neurons make cell-specific turns at or near the ventral midline of the spinal cord, which is occupied by a single row of midline floor plate cells. We tested whether these cells guide the growth cones by examining embryos missing the midline floor plate cells due either to laser ablation of the cells or to a mutation (cyc-1). In these embryos the growth cones followed both normal and aberrant pathways once near the ventral midline. This suggests that normally the midline floor plate cells do provide guidance cues, but that these cues are not obligatory.

Zebrafish Tenascin-W, a New Member of the Tenascin Family

A cDNA clone encoding tenascin-W, a novel member of the tenascin family, was isolated from a 20- to 28-h postfertilization (hpf) zebrafish cDNA library on the basis of the conserved epidermal growth factor-like domains represented in all tenascin molecules. An open reading frame of 2796 base pairs encodes a mature protein consisting of heptad repeats, a cysteine-rich amino terminal region, 3.5 epidermal growth factor-like repeats, five fibronectin type III homologous repeats, and a domain homologous to fibrinogen. These domains are the typical modular elements of molecules of the tenascin family. Sequence comparison demonstrated that TN-W shares homologies with the members of the tenascin family but is not a species homolog of any identified tenascin. The expression pattern of tn-w was analyzed by in situ hybridization in 1-day-old embryos, in 3-day-old larvae, and in juvenile zebrafish. At 24-25 hpf, tn-w mRNA was expressed in the lateral plate mesoderm, most conspicuously in the presumptive sclerotome. Migrating cells of sclerotomal and neural crest origins also showed high levels of expression. At 3 days, expression by sclerotomal and neural crest cells continued to be observed while expression in the somitic mesoderm was decreased. In juvenile fish, tn-w was expressed weakly by cells in the myosepta and, more strongly, by presumably nonneuronal cells in the dorsal root ganglia. In these tissues and at the same developmental stages, the expression of tn-w partially overlapped with the distribution of tn-c mRNA. In addition, tn-c was expressed in the central nervous system (CNS) and in the axial mesoderm, neither of which expressed tn-w at any of the age stages examined. The expression pattern of tn-w suggests an involvement in neural crest and sclerotome cell migration and in the formation of the skeleton. Similar and possibly overlapping functions could also be performed by tn-c, which appears to have additional functions during the development of the CNS.

Characterization of peripheral myelin protein 22 in zebrafish (zPMP22) suggests an early role in the development of the peripheral nervous system

Peripheral myelin protein 22 (PMP22) is a component of compact myelin of the peripheral nervous system (PNS). Mutations affecting PMP22 are associated with hereditary neuropathies in humans and rodents. Although mammalian PMP22 is expressed in several tissues, the disease pathology is restricted to the PNS. We describe the characterization of a PMP22‐related cDNA from zebrafish and the distribution of its cognate mRNA. Phylogenetic considerations and mRNA expression in cranial nerves are consistent with the interpretation that the encoded protein is the orthologue of mammalian PMP22. In situ hybridization analysis during development showed zebrafish PMP22 expression in embryonic sclerotome cells, in neural crest cells, and in migratory derivatives of both populations. Based on this specific expression pattern prior to the onset of myelination, we hypothesize that zebrafish PMP22 may play a role in early PNS development and that disturbance of such functions may contribute to the PNS‐restricted defects caused by mutations in the mammalian PMP22 gene. J. Neurosci. Res. 57:467–478, 1999.

Expression of a Na,K-ATPase Beta3 Subunit During Development of the Zebrafish Central Nervous System

Zebrafish beta3, a full length cDNA clone encoding a zebrafish Na,K‐ATPase beta subunit, was isolated. The protein shares highest homology with the beta3 subunits of amphibians and mammals, slightly less homology with the beta2 subunits, and is distinct from the betal subunits. The fish beta subunit co‐assembled with alpha subunits to form Na,K‐ATPase enzymes when expressed in Xenopus oocytes. Embryonic expression was first detected by whole‐mount in situ hybridization between 8–12 hr post‐fertilization (hpf) in the head mesoderm. Subsequently, and up to 24 hpf, the mRNA was confined to four dorsal domains in the anterior neural tube. After a transient downregulation during the second day, expression was again conspicuous in the nervous system of 3‐day‐old larvae. Based on its distribution pattern, the fish beta subunit could be involved in setting up regional identities in the developing fish CNS and in the differentiation of distinct cell types.

Axonal outgrowth by identified neurons in the spinal cord of zebrafish embryos

The spinal cord of early zebrafish embryos contains a small number of neurons per hemisegment. The earliest neurons are identifiable as individual neurons or small groups of homogeneous neurons and project growth cones that follow stereotyped, cell-specific pathways to their targets. These growth cones appear to bypass some axons but follow others during pathfinding, suggesting that they can distinguish among the different axons they normally encounter. Furthermore, identified growth cones exhibit cell-specific behaviors in apparent contact with the floor plate cells, which are found at the ventral midline of the early cord. These observations suggest the testable hypothesis that the floor plate may mediate multiple, cell-specific actions on identified growth cones in the zebrafish cord. One hypothesized action is inhibition of specific growth cones to prevent them from crossing the ventral midline.

Cellular and molecular bases of axonal pathfinding during embryogenesis of the fish central nervous system.

The accessibility of the zebrafish embryo offers unique possibilities to study the mechanisms that guide growing axons in the developing vertebrate central nervous system. This review examines the current understanding of the pathfinding decisions by the growing axons, their substrates, and the recognition molecules that mediate axon-substrate interactions. The detailed analysis of pathfinding at the level of individual axons demonstrates that growing axons chose their paths unerringly. To do so, they rely on cues presented by their environment, in particular by neuroepithelial cells. Our understanding of the molecular bases of axon-substrate interactions is increasing. Members of most classes of recognition molecules have been identified in fish. Experimental evidence for the functions of these molecules in the zebrafish nervous system is accumulating. In the future, this analysis is expected to profit greatly from genetic screens that have recently been initiated.