"Cholinergic" postsynaptic membranes of bullfrog sympathetic ganglia: electron microscopy of thin sections and freeze-fracture replicas

Synaptic organization of amphibian sympathetic ganglia

The synaptic organization of the amphibian sympathetic ganglia was studied, especially in the last two abdominal paravertebral ganglia of the frog. These ganglia appear to form a monosynaptic relay, not containing interneurons. They consist of two systems working in parallel: the principal neurons, by far the most numerous, and a small number of chromaffin (i.e., SIF) cells, usually arranged in clusters. Each principal neuron is innervated by a preganglionic branch forming a set of cholinergic synapses which exhibit classical ultrastructure. The only peculiarity is the presence of a subsynaptic apparatus in a variable percentage of synaptic complexes. Electrophysiological studies have demonstrated that synaptic transmission is due to ACh release and involves several postsynaptic potentials. Moreover, the principal neurons are of two types, B and C, whose preganglionic axons and their own axons have different conduction velocities. C neurons tend to be small in diameter, and B neurons are larger, but the size distribution of the two populations overlaps. More recently, it was demonstrated that these two neuronal systems have different immunocytochemical features. The C preganglionic fibers contain an LHRH-like peptide, which is responsible for late synaptic events. The B preganglionic fibers contain CGRP, whose role has not yet been established. The principal neurons all contain adrenaline, but neuropeptide Y is also present in C neurons and could be a second transmitter at peripheral junctions. SP-containing fibers also pass through the ganglia, but give rise to intraganglionic synapses only rarely, except in the celiac plexus. Galanin can coexist with neuropeptide Y in certain C neurons. Numerous principal neurons are immunoreactive for VIP. Chromaffin cells contain noradrenaline and metenkephalin, and some contain SP or LHRH; they are endocrine cells controlled by preganglionic fibers and can have a modulatory effect on principal neurons endowed with appropriate receptors. The accessibility of frog abdominal ganglia and the anatomical separation of B and C preganglionic fiber pathways provide interesting systems in which to carry out experimentation on the stability and specificity of synaptic contacts. After postganglionic axotomy, the majority of synapses disappear by disruption of synaptic contacts. There is a certain discrepancy between the recovery of synaptic transmission and the reappearance of morphologically identifiable synapses, suggesting that a certain amount of transmission is possible at contacts devoid of synaptic complexes. The selective deafferentation of B or C neurons showed that the subsynaptic apparati are mainly found at B neuron synapses. The course of reinnervation following selective deafferentation reveals the existence of different specificities at B and C synapses: C neurons are easily reinnervated by B preganglionic fibers, whereas C fibers appear fairly ineffective at reinnervating B neurons, even after a long interval. Attempts were made to reinnervate ganglionic neurons with somatic motor nerve fibers. Reinnervation was achieved only rarely, and it is concluded that the ganglionic synapses in the frog have a higher specificity and lower plasticity than in mammals.

The response of rubrospinal neurons to axotomy at different stages of development in the North American opossum

Rubral axons can grow around a lesion of their pathway in the thoracic spinal cord of developing opossums and a critical period exists for that plasticity. The critical period probably begins when rubral axons first grow into the thoracic cord, and it extends until approximately postnatal day 30. We previously noted that most rubrospinal neurons die after transection of their axon during the critical period, suggesting that plasticity results primarily from growth of axons not damaged by the lesion (Xu and Martin, J. Comp. Neurol. 279, 368-381, 1989). That observation led us to study the response of rubrospinal neurons to axotomy in more detail and at additional stages of development, using a prelabeling paradigm. We first injected fast blue (FB) into the caudal thoracic or rostral lumbar spinal cord in animals ranging from estimated postnatal day 9 to 50 and, about 4 days later, lesioned the rubrospinal tract several segments rostral to the injection. Approximately 30 days later, the animals were killed so that the red nucleus could be searched for labeled neurons. During the critical period for plasticity, rubrospinal neurons showed signs of degeneration 1 week after their axon was cut. When animals were killed 2-3 weeks after lesioning, there was an obvious decrease in axotomized neurons within the red nucleus, and by 4 weeks, more than 75% of them had degenerated. The marked susceptibility of rubrospinal neurons to axotomy during the critical period for plasticity is consistent with the hypothesis that developmental plasticity of the rubrospinal tract results primarily from growth of axons that were not damaged by the lesion. Our results also suggest that survival of axotomized rubrospinal neurons increases with age.

Evidence for new growth and regeneration of cut axons in developmental plasticity of the rubrospinal tract in the North American opossum

We have shown previously that rubral axons can grow around a lesion of their spinal pathway in the developing opossum and that a critical period exists for that plasticity (Martin and Xu, Dev Brain Res 39:303, 1988). Since most rubrospinal neurons degenerate after axotomy during the critical period, we have proposed that plasticity results primarily from growth of late arriving axons around the lesion rather than regeneration of cut axons (Xu and Martin, J Comp Neurol 279:368, 1989). In the present study, we used a double-labeling paradigm to test that hypothesis. Four groups of pouch young opossums received bilateral or unilateral injections of Fast Blue (FB) into the caudal thoracic or rostral lumbar cord (T12-L2) at different ages in order to label rubrospinal neurons. Three or 4 days later, the rubrospinal tract was transected unilaterally, four to five segments rostral to the injection(s). If the injection was unilateral, the lesion was made ipsilateral to it. The animals were maintained for about 1 month before a second marker, Diamidino Yellow (DY), was injected, usually bilaterally, between the FB injection(s) and the lesion. The animals were maintained for about 5 days before sacrifice and sections through the red nucleus and spinal cord were examined with a fluorescence microscope. During the critical period for plasticity, only a few rubral neurons contralateral to the lesion were labeled by FB alone, supporting our previous contention that most axotomized neurons degenerate. In contrast, many neurons were labeled by DY alone, indicating that their axons were not present in the caudal cord at the time of the FB injection and that they grew around the lesion during the 1 month survival to incorporate DY. A few double-labeled neurons were also found. One interpretation of such neurons is that they survived axotomy, as evidenced by the presence of FB, and supported axons which grew around the lesion to take up DY. Another interpretation is that they supported late growing axons which incorporated residual FB as well as DY. In order to choose between these alternatives, a similar double-labeling paradigm was carried out, but with removal of FB at the time of the lesion. Since a few neurons were still double labeled, we conclude that regeneration of cut axons also contributed to rubrospinal plasticity. Our results support our previous suggestion that developmental plasticity of the rubrospinal tract results primarily from growth of late arriving axons around the lesion, but they also suggest that regeneration of cut axons occurs.

Deep-etch structure of astrocytes at the superficial glia limitans, with special emphasis on the internal and external organization of their plasma membranes

The cytoskeletal system in rat subpial astrocytes and the relationship between astrocytic plasma membrane and basal lamina or cytoplasmic components were examined with a quick-freeze deep-etch technique, mainly using chemically fixed tissues. Attention was focused on the way intramembrane particles (IMPs), particularly orthogonal arrays, are organized in the membranes and related to extramembrane components. The basal lamina was composed of a sheet-like network of strands (4-9 nm thick), some, which we have called 'trabecular' strands, extending through the lamina lucida to touch the astrocytic membrane at irregular intervals. The trabecular strands usually formed a bulbous structure where they touched the membrane, but sometimes appeared to intrude directly into the external lipid layer. The orthogonal arrays did not extend to the outer true surface, and no special structure was detectable in association with them. Small spherical protrusions (7-9 nm in diameter), related to neither the trabecular strands nor the arrays, were observed in the outer surface. Judging from their size and distribution, these are probably tops of tall globular IMPs. In the inner or cytoplasmic true surface, protrusions were relatively numerous; some were large, 15-20 nm in diameter, while others were small (8-10 nm). Some of the small protrusions were identified as transmembrane components. Although protrusions were more conspicuous in the inner than in the outer surface, none of them provided images related or similar to the orthogonal arrays. Some protrusions in the inner surface were connected with thin (4-5 nm) or thick (approximately 10 nm) filaments constituting the underlying network. The thin filaments were also anchored to the intermediate filaments which lay parallel with the astrocytic membranes. In the cytoplasm, the intermediate filaments were firmly packed to form bundles. Because the orthogonal arrays are probably embedded within the astrocytic membrane, they may not serve as a transmembrane channel but rather contribute to some stabilizing function for the membrane.

Specializations of plasma membranes in Pacinian corpuscles: implications for mechano-electric transduction

Pacinian corpuscles of cat mesentery were studied with freeze-fracture and thin sectioning methods after chemical fixation. Intramembranous particles (IMPs) exhibit differences in both density and pattern of distribution between the axolemma of the smooth short axis (x-axis) region and that of the axonal spine region of the long axis (y-axis) of the axon terminal. The axolemma of the x-axis has IMPs at a density of 2687 +/- 581 per micron2 (mean +/- S.E.M.), and these particles are 9.0 +/- 1.7 nm (mean +/- S.D.) in diameter. In contrast, the axolemma of the y-axis has a higher density of IMPs (3607 +/- 612 per micron2) which are larger (diameter, 10.0 +/- 1.7 nm). The particle distribution is not homogeneous in x-axis membranes as there are small patchy areas devoid of particles scattered throughout the entire surface. The E-face of the axolemma has a low density of IMPs (approximately 200 per micron2 in both x- and y-axes). However, IMPs in the E-face are smaller (approximately 9 nm) in the x-axis than in the y-axis (approximately 10 nm). The inner core lamellar cells have IMPs at a density of 3276 +/- 739 per micron2 and 553 +/- 169 per micron2 in the P- and E-faces, respectively. The particles are about 10 nm in diameter in both faces. Many gap junctions occur between lamellar cells especially near the clefts, suggesting that hemilamellae of each inner core half are kept at the same electrotonic potential. The outer core lamellar cells have IMPs at a density of 2239 +/- 403 per micron2 and 536 +/- 123 per micron2 in their P- and E-faces, respectively. The particles are approximately 10 nm in diameter in both faces. A noteworthy finding is that tight junctions are prominent at cell-to-cell appositions within individual lamellae, especially in the first and second (or sometimes third) innermost lamellae of the outer core. These tight junctions are considered to be a barrier to the leakage of fluid and/or ions between interlamellar spaces as well as between inner and outer cores. An intermediate cell layer is identified between the inner and outer cores. The connective tissue space of this cell layer corresponds to the endoneurium, indicating that intermediate layer cells are comparable to endoneurial fibroblasts.(ABSTRACT TRUNCATED AT 400 WORDS)