Sanford L. Palay

Ultrastructure of the small neurons in the superior cervical ganglion.

Abstract The granule-containing cells of the rat superior cervical sympathetic ganglion were examined by light and electron microscopy. They are sparsely distributed in small cluster of about 2–5 cells. Each group has an incomplete sheath of satellite cells and lies in apposition to a blood capillary. The granule-containing cells contain numerous microvilli that project into the intercellular clefts, and prominent in their cytoplasm are dark-cored vesicles with 1400Aaverage overall diameter. Available evidence suggests that these contain a catecholamine, probably noradrenalin or its close analogue dopamine. On their cell bodies and minor processes they receive synaptic boutons which, judged by the types of vesicles found in them, are pre-ganglionic. In addition, they posses longer processes: these form synaptic contacts with nerve elements that are likely to be dendrites of ganglionic neurons. The dark-cored vesicles that approach the pre-synaptic membrane differ from those described in other mammalian synapses. On the basis of this evidence, the granule-containing cells can be classified as interneurons which receive impulses from pre-ganglionic sympathetic fibers and transmit to small nerve processes within the ganglion. Although they are commonly referred to as chromaffin cells they should not be allowed membership of this class, since they reportedly do not give the chromaffin reaction. The function of the interneuron may be to provide negative feedback to ganglionic neurons, thus limiting their dependence on direct signals from the pre-ganglionic neuron. From the numbers of interneurons that are found in the ganglia, it may be inferred that their contacts are restricted to ganglion cells on a particular pathway. Information about the kind of sympathetic pathway that would require such an interneuron would be welcome.

The nucleus paragigantocellularis lateralis in the rat

SummaryThe nucleus paragigantocellularis lateralis (PGCL) is located in the ventral portion of the rostral medulla. Serial sections of the rat brainstem were examined in the three cardinal planes and the boundaries of the PGCL were determined. In order to visualize the shape and extent of the nucleus, a three-dimensional reconstruction of the PGCL was made from a series of coronal sections. Measurements of neuronal areas, lengths, and widths indicate that a number of neuronal types are present. Small neurons measure less than 150 μm2 and large neurons greater than 250 μm2.Some neuronal types are distributed preferentially throught the PGCL, and on this basis the nucleus may be divided into caudal and rostral subgroups. Most large neurons (>250 μm2) are found in the caudal portion. Certain neurons contain intranuclear rods, and these neurons are often disposed in small groups, especially common in the caudal PGCL. Two morphologically distinct neuronal types incorporate 3H-serotonin when this marker is infused into the ventricular system; the other neurons not marked by this method probably contain other, different transmitters. On the basis of neuronal measurements and staining qualities, it is ascertained that the PGCL is a parvocellular reticular nucleus characterized by many neuronal types.

The fine structure of the lateral vestibular nucleus in the rat. II. synaptic organization

Three types of axon terminal occur in the lateral vestibular nucleus of the rat: large, slender and small. Large terminals are often seated in cup-like depressions in giant cells and encircled by thorns. The junctional interface exhibits 3 kinds of specialization: (1) puncta adhaerentia, interpreted as simply adhesive; (2) ‘active’ zones, interpreted as sites of chemical transmission, and (3) ‘gap’ junctions, interpreted as sites of low electrical resistance between the apposed cells. The physiological implications of this complicated synaptic interface are considered and several possible functions are suggested. Primarily, it is suggested that the operation of such a synaptic interface would be more subtle and delicate than what has been found hitherto. Large terminals and these mixed synapses have not been seen in the cat. Slender terminals are long, relatively straight axons with several varicosities (boutons en passant) along their course, each with ‘active’ zones. These endings are the least common in the nucleus. The most common are the small endings, which occur on all cell types and are the only ones to appear on the small neurons. No correlation was found between sizes of synaptic vesicles and sizes of the endings as all kinds of terminals contained similar vesicles. Axo-axonal synapses and complex glomeruli are also encountered in this nucleus.

The form of velate astrocytes in the cerebellar cortex of monkey and rat: high voltage electron microscopy of rapid Golgi preparations.

SummaryHigh voltage electron microscopy of Golgi preparations vividly displays the veil-like appendages on certain protoplasmic astrocytes. These appendages are extremely thin sheets of cytoplasm or plasmalemmal films expanding from the larger processes of the cells. Because of the prominence of this structural feature, reminiscent of the appearance of astrocytes in tissue culture, we designate these cells as velate astrocytes, in order to distinguish them from those protoplasmic astrocytes that lack such appendages. In the cerebellar cortex, velate astrocytes are represented by two types of neuroglial cell: (1) the Golgi epithelial cell and (2) the common astrocyte of the granular layer. The first type not only gives rise to the Bergmann fibers, but also envelops the Purkinje cell and all of its processes. The second type divides up the granular layer into gross compartments containing individual glomeruli and single or clustered granule cells. The probable significance of this compartmentation is discussed.

Meynert cells in the primate visual cortex

SummaryThe solitary cells of Meynert are distinguished by their specific location in layer V of the striate cortex, very large size, argyrophilia, and the profusion of neurofilaments in their dendrites and perikarya. They occur with greater frequency in the macular region of the cortex, spaced a minimal distance of 110 μm apart, at a maximum density of about 8000/cm2. In the perifoveal cortex, Meynert cells are spaced about 400 μm apart and packed at a density of approximately 625/cm2. Each Meynert cell has an apical dendrite and many large basal dendrites. The perikaryon and primary segments of all dendrites are spine-free; however, more distally a total of 36000 spines are present, differentially disposed upon the dendritic surfaces. The basal dendrites bear over 77% of the spines on the Meynert cell, although they account for only 66% of the total length of the dendritic arborization. The first part of the apical dendrite is the most densely decorated with appendages, accounting for almost 10% of the spines on the whole dendritic tree. The apical dendrite becomes progressively less spiny as it passes through the superficial part of layers IV and III; less than 2.5% of the total number of spines of the Meynert cell project from this part of the apical dendrite. When the dendrite reaches layer II it bursts into an umbel of rapidly tapering branches. These are highly spinose, accounting for 8–13% of the cells total, dispersed over only 23% of the linear dendritic length. It is suggested that this differential distribution of thorns can be correlated with the axonal inputs in the various cortical layers, and that the Meynert cell is designed to receive maximal information from layers I and II, and from layers V and VI, which are sources mainly of intracortical inputs. Thus the Meynert cell may be principally concerned with integrative information. In the perifoveal cortex, the basal dendrites of adjacent Meynert cells overlap considerably, and the apical terminal bouquet dendrites do not. In the macular cortex, because of the increased frequency of these neurons, both basal and apical terminal dendritic fields overlap. A model is developed to illustrate these hypotheses.

Gamma-aminobutyric acid pathways in the cerebellum studied by retrograde and anterograde transport of glutamic acid decarboxylase antibody after in vivo injections

SummaryInjections of characterized antibody against glutamic acid decarboxylase (GAD), the enzyme responsible for the synthesis of γ-aminobutyric acid (GABA), were made into the cerebellum. Small cortical injections of anti-GAD antibody produced labeled stellate, basket, Purkinje, and Golgi cells and their processes at the injection site. Anterograde transport of GAD antigen-antibody complexes in Purkinje cell axons caused intense labeling of terminals in deep cerebellar and several vestibular nuclei. Small groups of mossy fiber rosettes labeled and produced retrograde labeling and GAD immunoreactivity in a small number of pleomorphic neurons in the deep cerebellar nuclei. Injections into the dentate nucleus produced retrograde labeling in Purkinje cell bodies and anterograde label in a small number of mossy fiber rosettes. All projections conformed to previously reported topographic distributions of corticonuclear and nucleocortical cerebellar pathways. These findings confirm the GABA content of most Purkinje cell-deep nuclei connections and provide new evidence for a GABA component in part of the nucleocortical pathway in the cerebellum. Immunocytochemical controls for specificity were conducted by injections of preimmune rabbit serum as a substitute for GAD antibody. Only nonspecific labeling was obtained in these cases. Colchicine caused a cumulative enhancement of GAD immunoreactivity in all cases. The present studies indicate that the method of in vivo antibody injections can be utilized to study chemically specific connections in nervous tissue.

Tendril and glomerular collaterals of climbing fibers in the granular layer of the rat's cerebellar cortex

SummaryRapid Golgi preparations show that two kinds of collaterals issue from the climbing fiber in its course through the granular layer. The first resembles the tendrils found in the molecular layer and consists of globose varcosities connected by a very fine thread. In electron micrographs these varicosities in the granular layer contain dense aggregates of round synaptic vesicles at least 520 Å in diameter and the connecting threads contain numerous microtubules. The varicosities synapse on the somata of Golgi II cells and on the shafts of dendrites belonging to both Golgi II and granule cells. The second type of collateral emerges from the main stem of the climbing fiber as a stout branch that sprays out abruptly into a large efflorescence. In electron micrographs this terminal appears as the central stellate structure in a glomerulus and is packed with round synaptic vesicles like those in the tendril varicosities. Granule cell dendrites encircle the terminal and occasionally synapse with it. Often the terminal in the glomerulus also forms an extensive junction—a synapse en marron—with the some of Golgi II cell. In this region the surface of the cell is wrinkled like a Spanish chestnut and the glomerular terminal is reciprocally ridged and furrowed to match. Synaptic complexes occur only in the furrows of the cell surface. A broad subsynaptic zone is filled with a fine filamentous matrix.This study provides the first morphological identification and description of climbing fiber terminations in the granular layer, the existence of which has been suggested by earlier Golgi studies and postulated by neurophysiologists.The fact that climbing fibers synapse on both granule cells and Golgi II cells complicates the analysis of the way in which the cerebellar cortex operates, because these two cells have postsynaptic effects of opposite sign. The climbing fiber is known to evoke a complex discharge from the Purkinje cell, consisting of a large primary spike and smaller secondary potentials. It is suggested that when a climbing fiber volley traverses the granular layer, the granule cells on which it synapses are induced to excite stellate and basket cells in the molecular layer which in their turn inhibit the secondary spikes of the climbing fiber response in the Purkinje cell. Meanwhile the Golgi II cells, stimulated by the same climbing fiber volley, suppress the granule cells and thus transsynaptically limit the duration of the inhibitory effects exerted by the interneurons in the molecular layer.

A Guide to the Synaptic Analysis of the Neuropil

A morphological analysis of the organization of the gray matter in the central nervous system depends on the discovery of consistent repetitive patterns. Without these, the gray matter remains a chaotic jungle. An hypothesis derived from the study of a few simple regions has been developed to serve as a guide in finding these patterns. It states that all nerve fibers and terminals arising from a particular group of nerve cells, or, more precisely, a particular nerve cell type, display similar axoplasmic configurations despite variations in size and shape of the terminations. This hypothesis is reminiscent of the so-called Dales principle that a nerve cell makes use of the same transmitter at all of its branches or terminations. These apparent rules of uniformity or congruity merely reflect the functional integrity of the nerve cell and the role of its parts in the nervous system. But as an hypothesis, it needs to be tested, and it needs to be tested anew in each region, since exceptions to the assumed rule can be expected. It is therefore proposed as the first working hypothesis in each new region. If it should prove to be true in general, it will facilitate and rationalize the analysis of the gray matter, as it has already done in the cerebellar cortex and the deep cerebellar nuclei. If it should prove to be false in a few regions, the analysis will become more difficult, and additional modes of marking nerve endings will have to be used. Experimental methods for identifying nerve terminals can be translated from the light microscopic to the electron microscopic level, but there are significant drawbacks at both levels: lack of precision, destruction of fibers of passage, and rapid evolution of the degenerative process may greatly restrict their usefulness. Labeling with tritiated amino acids or transmitters, or with horseradish peroxidase, provide new methods for tracing interneuronal connections at the electron microscopic level. These have the advantages of high specificity, nondestructiveness and a physiological mode of selective marking. However, they do not offer a solution to the problem of short-range connections. For these, careful reconstructions of serial sections may prove necessary, as Sjöstrand (1974) has demonstrated in a remarkable paper on the retina. The aim of all these methods is to discover patterns of synaptic connectivity in order to map the cellular organization of the nervous system. In the foregoing, nothing was said about synapses other than those articulating axons with somata or dendrites and their appendages. Clearly the same principles of recognition apply to axo-axonal and dendro-dendritic synapses. Although the synapses that have been considered here are chemical synapses, the same questions regarding the identity of the partners in electrotonic junctions must be asked as well.

Interrelations of basket cell axons and climbing fibers in the cerebellar cortex of the rat.

SummaryAn analytical study was undertaken with both electron microscopy and the rapid Golgi method in order to clarify the interrelations of climbing fibers, basket cell axons, and Purkinje cell dendrites. The two fibers are readily distinguished in electron micrographs by means of their differing content of microtubules and neurofilaments, the packing density of synaptic vesicles, and the disposition of their synaptic junctions on the Purkinje cell dendrite. Climbing fibers are generally thin and contain many microtubules. They give off attenuated collaterals, whose rounded varicosities are densely packed with vesicles and which form en passant synapses with clusters of thorns projecting from the major Purkinje dendrites. In contrast, basket axons are relatively thick and contain many neurofilaments. By means of slight dilatations containing loosely aggregated vesicles, the axon and its collaterals form numerous synapses en passant with the smooth dendritic shafts and the perikaryon of the Purkinje cell. Climbing fibers and basket cell axons run along parallel with each other but without forming axo-axonic synapses as they ascend over the surface of the Purkinje dendrites. Both fibers form especially elaborate intertwined festoons at the branching points of the major dendrites. The kinds of synapses found are described in detail, and the functional implications are discussed.The hypothesis is developed that the dendritic thorn is a device for isolating the subsynaptic membrane from electrical events in the rest of the dendrite at the cost of reducing the effectiveness of the synapse. This principle is incorporated in the Purkinje dendrite—parallel fiber synapses, in which an individual fiber can be expected to have little importance. The disadvantage of using thorns as postsynaptic surfaces can be mitigated by clustering them and increasing the number of thorns contacted by each presynaptic terminal. This method is utilized at the junctions between the climbing fiber and the Purkinje dendrite to produce one of the most powerful excitatory synapses known. It is furthermore suggested that the elaborate plexus of climbing fibers and basket cell axons synapsing in the crotches of branching dendrites is strategically located to control the flow of information in the Purkinje cell dendritic tree.

High voltage electron microscopy of rapid Golgi preparations. Neurons and their processes in the cerebellar cortex of monkey and rat

SummaryThe rapid Golgi reaction produces initially a red and secondarily a black precipitate in certain neurons, neuroglia, and endothelial cells of nerve tissue. High voltage electron microscopy shows that the red impregnation is a dense, intracellular, fibrillar meshwork of rodlets with a filamentous substructure. The black impregnation is due to the superimposition of very dense globular and polyhedral crystals that stud the surface of the underlying red fibrillar network. When present in quantity, these globules conglomerate to form a crust over the impregnated structure. This can cause distortion in the shape of the structure and considerable overestimation of its size in light microscope studies of black Golgi material.As seen in the high voltage electron microscope, the varicosities in the course of red-impregnated axons (for example, parallel fibers in the cerebellar cortex) show light circular patches that correspond to synaptic sites. Some varicosities have more than one site, an observation that is corroborated by conventional electron microscopy. The study of Golgi preparations in the high voltage electron microscope would enable one to make an accurate estimate of the number of synapses effected by an axon along its course. This information cannot otherwise be obtained, either by light microscopy or by electron microscopy of thin sections. Stereoscopic pairs show the three-dimensional interrelationships between neurons and neuroglia in the neuropil.Selected area electron diffraction studies on sections of Golgi material can provide in situ chemical information on the impregnated fibers. Thus, high voltage electron microscopy may provide an important auxiliary technique to complement light microscopy and standard electron microscopy in the study of the structure and organization of nervous tissue.