Marilyn J. Anderson

Glial fibrillary acidic protein in regenerating teleost spinal cord.

Immunohistological and ultrastructural studies were carried out on normal and regenerating spinal cord of the gymnotid Sternarchus albifrons, and in the brain and spinal cord of the goldfish Carassius auratus, to examine the distribution of glial fibrillary acidic protein (GFAP) in these tissues. Sections of normal goldfish brain and spinal cord exhibited positive staining for GFAP. In normal Sternarchus spinal cord, electron microscopy has revealed filament-filled astrocytic processes; however, such astrocytic profiles were more numerous in regenerated cord. Likewise, while normal Sternarchus spinal cord showed only a small amount of GFAP staining, regenerated cords were strongly positive for GFAP. Positive staining with anti-GFAP was observed along the entire length of the regenerated cord in Sternarchus, and was especially strong in the transition zone between regenerated and unregenerated cord. Both regeneration of neurites and production of new neuronal cell bodies occur readily in such regenerating Sternarchus spinal cords (Anderson MJ, Waxman SG: J Hirnforsch 24: 371, 1983). These results demonstrate that the presence of GFAP and reactive astrocytes in Sternarchus spinal cord does not prevent neuronal regeneration in this species.

Neurogenesis in tissue cultures of adult teleost spinal cord.

[3H]Thymidine autoradiography of explant cultures from the spinal cord of an adult teleost, Sternarchus albifrons, reveals the presence of thymidine-labeled cells with neuronal morphology. These cells have also been identified as neurons by positive staining with two monoclonal antibodies against neurofilaments. Thymidine labeling occurs in cultured neurons derived from both normal (histologically and functionally mature) and regenerating spinal cord of adult Sternarchus albifrons. These results provide evidence that some cells in spinal cord of adult Sternarchus retain the ability to incorporate thymidine and undergo neuronal differentiation in vitro.

Morphology of regenerated spinal cord in Sternarchus albifrons

SummaryThe tail of the gymnotid Sternarchus albifrons, including the spinal cord, regenerates following amputation. Regenerated spinal cord shows a rostro-caudal gradient of differentiation. Cross sections of the most distal regenerated cord show radially enlarged ependymal cells, relatively undifferentiated cells, and numerous blood vessels. More anterior sections contain well differentiated electromotor neurons, glial cells, and myelinated axons. The number of electromotor-neuron cell bodies in cross sections of regenerated spinal cord is three to six times the number in nonregenerated cord. Distinct tracts of axons, easily identifiable in normal cord, are not distinguishable in cross sections of regenerated cord. Some reorganization of the spinal cord also appears to take place anterior to the site of transection.Individual electromotor neurons in the regenerated spinal cord have morphologies largely similar to those of normal electrocytes, i.e., cell bodies are rounded, lack dendrites, have synapses characterized by gap junctions with presynaptic axons, and lack an unmyelinated initial segment. The presence of electromotor neurons with normal morphology in regenerated spinal cord correlates with the re-establishment of relatively normal electrocyte axonSchwann cell relationships in the regenerating electric organ of this sternarchid.

Generation of electromotor neurons in Sternarchus albifrons: Differences between normally growing and regenerating spinal cord

This study examines the regulation of the number of electromotor neurons during postnatal growth of the spinal cord in the gymnotiform teleost Sternarchus albifrons. It specifically asks whether a large overproduction of electromotor neurons and a wave of cell death, similar to those occurring during spinal cord regeneration in this species, play a role in the on-going growth at the caudal tip of the normal spinal cord. Neurons are produced from ependymal precursors at the caudal end of the spinal cord during both normal growth in the adult and regeneration of the spinal cord in this species. Previous studies have demonstrated that during spinal cord regeneration after amputation of the tail in Sternarchus, there is an initial massive (up to fivefold) overproduction of electromotor neurons, followed by a wave of cell death which reduces the number of these neurons to the normal level. In the present study, transverse sections through the caudalmost spinal segment of normal adult Sternarchus were examined. Proceeding rostrally from the caudal tip of the cord, the number of electromotor neurons increases monotonically to reach the normal number at a site 4-5 mm rostral to the caudal tip. Neither a massive overproduction of electromotor neurons nor a wave of neuronal death are observed during on-going growth of the normal spinal cord. The mechanisms by which the neuronal number is modulated are therefore different in the on-going normal growth of spinal cord versus regeneration of spinal cord in this species.

Cell death of asynaptic neurons in regenerating spinal cord

The weakly electric fish Sternarchus albifrons possesses a unique class of asynaptic neurons, the electromotor neurons, whose axons constitute the electric organ. The cell bodies of origin of the electrocyte axons are located in the spinal cord. Both spinal cord and electromotor neurons ( electrocytes ) regenerate after amputation of the tail. Sternarchus spinal cords which have regenerated for 1 or more years show a progression in number of perikarya of electromotor neurons along the rostro-caudal axis. The most recently regenerated region of the cord is at the caudal end, which consists of a tube of ependyma. Progressing rostrally along regenerated spinal cord from the caudal end, numerous cells are generated and large numbers of electromotor neurons differentiate. The maximum number of electromotor neurons per transverse section of regenerated cord is five times higher than in normal mature cord. Rostral to this, the number of electromotor neurons decreases gradually to the normal number near the transition zone (the border with unregenerated cord). As the more rostral regenerated cord has presumably had a longer period of regeneration, we conclude that excess numbers of electromotor neurons are generated initially, and that subsequently the number of these neurons is decreased by cell death. This conclusion is supported by the fact that younger regenerates (2-4 months) have larger-than-normal numbers of perikarya of electromotor neurons extending up to the transition zone (Anderson and Waxman , 1981). No evidence of migration or depletion of electromotor neurons from unregenerated cord rostral to the amputation has been observed. Since the axons of the electromotor neurons in Sternarchus do not normally form any synapses, this study provides evidence that factors other than synaptic competition must be responsible for determining cell death during regeneration of these spinal neurons.

Regeneration of spinal electrocyte fibers in Sternarchus albifrons: development of axon-Schwann cell relationships and nodes of Ranvier.

SummaryThe electrocyte fibers in the gymnotid Sternarchus albifrons are highly differentiated myelinated axons which exhibit several types of nodes of Ranvier and characteristically short internode lengths. In the present study, regeneration of the electrocyte fibers following removal of the tail was examined by electron microscopy. By 36 days following extirpation, the regenerating electrocyte axons exhibit Type I nodes of Ranvier, with a normal morphology, and Type II nodes of Ranvier with a large nonmyelinated gap and a polypoid elaboration of the axon surface. Moreover, in the regenerating axons the internode length ∶ diameter ratios are quite small. Thus, relatively normal axon-Schwann cell relations and a relatively normal differentiation of the axon surface are achieved during regeneration of the Sternarchus electrocyte fibers.

Explant cultures of teleost spinal cord: source of neurite outgrowth

The source of neurite outgrowth in explant cultures of normal adult Apteronotus spinal cord was examined. Explants which contained the central region of spinal cord, including ependyma, showed neurite outgrowth in culture. Explants which did not contain ependyma showed no neurite outgrowth. It is concluded that the ependymal region is necessary for neurite outgrowth in these cultures of adult teleost spinal cord. In addition, our failure to observe axon outgrowth clearly attributable to fluorescently back-labeled electromotor neurons in these cultures suggests that the exuberant neurite outgrowth in vitro is most probably due to cells other than the electromotor neurons. This explant culture system provides a unique opportunity to study neuronal differentiation, regeneration, and neurogenesis in vitro.

Retrograde axon reaction following section of asynaptic nerve fibers

SummaryThe asynaptic spinal neurons of the gymnotid teleost Sternarchus albifrons show several distinct characteristics of the retrograde reaction of the perikaryon (which corresponds to chromatolysis in mammals) following axotomy. Nuclei of affected cells are characteristically eccentric. Large bundles of neurofilaments, never seen in normal perikarya of these cells, become prominent following axotomy. There is a marked increase in the number and size of dense bodies in the affected perikarya. Large arrays of parallel rough endoplasmic reticulum, never seen in normal cells, are frequent in the axotomized neurons. These results demonstrate that disconnection from synaptic terminals is not a necessary condition for the retrograde reaction of the perikaryon following axotomy.

Retrograde labeling of regenerated electromotor neurons with HRP in a teleost fish, Sternarchus albifrons: relation to cell death.

SummaryBack-labeling of regenerated electromotor neurons in the teleost Sternarchus albifrons was performed to test the hypothesis that, in regenerated spinal cord, incorrectly located electromotor neurons are eliminated because their axons do not reach the correct target area (electric organ). In each cross section examined, all of the regenerated electromotor neurons ipsilateral to the implantation site were labeled with horseradish peroxidase, including those ectopic cells located at the edge of the cord, which are later eliminated by selective cell death. Retrograde labeling of these ectopic neurons demonstrates that their axons do extend into the correct target area (the regenerated electric organ). Thus total misdirection of the axons cannot be the cause of their subsequent cell death. We conclude that selective neuronal death in this system does not reflect the absence of axonal projection to the correct target area.