Principal neurons as local circuit neurons in the rat superior cervical ganglion: the synaptology of the neuronal processes revealed by intracellular injection of biocytin

Differential Inhibition of Ca2+ channels by alpha2-adrenoceptors in three functional subclasses of rat sympathetic neurons

A comparison of identified sympathetic neurons in the isolated intact superior cervical ganglion revealed that secretomotor, pilomotor, and vasoconstrictor cells differ in their action potential mechanisms and in their postsynaptic alpha(2)-adrenergic responses to 10 microM norepinephrine (NE). In normal saline, the half-width of the spike afterhyperpolarization (AHP) in secretomotor neurons (103.5 +/- 6.2 ms) was twofold that recorded in vasoconstrictor neurons (47.7 +/- 2.9 ms) and 1.5-fold that in pilomotor neurons (71.4 +/- 10.3 ms). Bath-applied NE reversibly inhibited the action potential repolarization shoulder, AHP amplitude, and AHP duration in secretomotor and pilomotor neurons to a similar extent, but had no effect on vasoconstrictor neurons. The insensitivity of vasomotor neurons to NE was not an artifact produced by microelectrode recording because all three cell groups were similar in terms of resting potential and input resistance. Moreover, NE insensitivity was not a natural consequence of briefer AHP duration in vasoconstrictor cells. Adding 10 mM TEA(+) caused marked accentuation of the shoulder and AHP duration in vasoconstrictor neurons and comparable changes in the other two cell types, but did not unmask any sign of NE sensitivity in the vasoconstrictors. However, the spike shoulder and AHP in vasoconstrictors were Cd(2+) sensitive, blocked by omega-conotoxin, an N-type calcium channel antagonist, and inhibited by oxotremorine-M, a muscarinic receptor agonist. These data show that NE can differentially modulate functional subsets of mammalian sympathetic neurons and that NE insensitivity can serve as a practical experimental criterion for identification of vasomotor neurons in the isolated ganglion.

Structure of peripheral synapses: autonomic ganglia

Final motor neurons in sympathetic and parasympathetic ganglia receive synaptic inputs from preganglionic neurons. Quantitative ultrastructural analyses have shown that the spatial distribution of these synapses is mostly sparse and random. Typically, only about 1%-2% of the neuronal surface is covered with synapses, with the rest of the neuronal surface being closely enclosed by Schwann cell processes. The number of synaptic inputs is correlated with the dendritic complexity of the target neuron, and the total number of synaptic contacts is related to the surface area of the post-synaptic neuron. Overall, most neurons receive fewer than 150 synaptic contacts, with individual preganglionic inputs providing between 10 and 50 synaptic contacts. This variation is probably one determinant of synaptic strength in autonomic ganglia. Many neurons in prevertebral sympathetic ganglia receive additional convergent synaptic inputs from intestinofugal neurons located in the enteric plexuses. The neurons support these additional inputs via larger dendritic arborisations together with a higher overall synaptic density. There is considerable neurochemical heterogeneity in presynaptic boutons. Some synapses apparently lack most of the proteins normally required for fast transmitter release and probably do not take part in conventional ganglionic transmission. Furthermore, most preganglionic boutons in the ganglionic neuropil do not form direct synaptic contacts with any neurons. Nevertheless, these boutons may well contribute to slow transmission processes that need not require conventional synaptic structures.

Simulations to derive membrane resistivity in three phenotypes of guinea pig sympathetic postganglionic neuron

The electrotonic behavior of three phenotypes of sympathetic postganglionic neuron has been analyzed to assess whether their distinct cell input capacitances simply reflect differences in morphology. Because the distribution of membrane properties over the soma and dendrites is unknown, compartmental models incorporating cell morphology were used to simulate hyperpolarizing responses to small current steps. Neurons were classified as phasic (Ph), tonic (T), or long-afterhyperpolarizing (LAH) by their discharge pattern to threshold depolarizing current steps and filled with biocytin to determine their morphology. Responses were simulated in models with the average morphology of each cell class using the program NEURON. Specific membrane resistivity, R(m), was derived in each model. Fits were acceptable when specific membrane capacitance, C(m), and specific resistivity of the axoplasm, R(i,) were varied within realistic limits and when underestimation of membrane area due to surface irregularities was accounted for. In all models with uniform R(m), solutions for R(m) that were the same for all classes could not be found unless C(m) or R(i) were different for each class, which seems unrealistic. Incorporation of a small somatic shunt conductance yielded values for R(m) for each class close to those derived assuming isopotentiality (R(m) approximately 40, 27, and 15 k omega cm(2) for T, Ph, and LAH neurons, respectively). It is concluded that R(m) is distinct between neuron classes. Because Ph and LAH neurons relay selected preganglionic inputs directly, R(m) generally affects function only in T neurons that integrate multiple subthreshold inputs and are modulated by peptidergic transmitters.

Spontaneous synaptogenesis in ex vivo sympathetic ganglion and the blockade by serum treatment

Central denervation for more than 1 month has been shown to cause an increase in the number of adrenergic synapses in sympathetic ganglia in vivo. Here, we report several lines of evidence that adrenergic synapses may be generated de novo in ex vivo superior cervical ganglion (SCG) of adult rats only several hours after the isolation. Structures immunoreactive for synaptophysin, a marker of presynaptic elements, were drastically decreased 6 days after the preganglionic denervation. A significant increase in number of synaptophysin positive boutons was observed over 3-8 hours in the denervated SCGs maintained ex vivo at 36 degrees C in oxygenated physiologic saline, and this increase was blocked by adding normal serum in the saline. Electron microscopic analysis confirmed that the number of adrenergic synapses specifically labeled with 5-hydroxydopamine was increased by several-fold under the same condition. Intracellular labeling of SCG neurons revealed an increase in the incidence (from 8 to 50%) of neurons having dendritic plexus after the in vitro incubation. No evidence of axonal sprouting within the ganglion was observed. Intracellular recordings from single neurons of denervated SCGs revealed that maximum amplitudes of inhibitory postsynaptic potentials, which were completely blocked by yohimbine, an alpha2-adrenoceptor antagonist, in response to focal stimulation were increased over the several hours. These results suggest that dendrites of SCG neurons rapidly develop and exhibit local efferent characteristics that underlie the inhibitory synaptic transmission once they are subjected to serum deprivation.

Optical imaging of the spontaneous neuronal activities in the male rat major pelvic ganglion following denervation of the pelvic nerve

Measurement of groups of neuronal activities following pelvic nerve transection to the major pelvic ganglion (MPG) of the rat was performed using voltage-sensitive dye (RH795) and an optical recording system. In control MPG, averaged neuron diameters were 32.0 +/- 0.6 x 22.6 +/- 0.4 microm. Application of KCl (10-50 mM) to the ganglia exhibited excitation which increased in a dose-dependent fashion. Fluctuating membrane potentials were not observed in control ganglion neurons. After the denervation of pelvic nerve chronically (2-6 weeks), the spontaneous neuronal activities were recorded in 91% of the experiments (n = 32). The activity was occurring somewhat periodically (2-15 s). Averaged neuron diameters were 41.3 +/- 1.3 x 24.7 +/- 0.9 microm in denervated MPG which is significantly larger than control. Since average neuron size increased 4 weeks after the denervation, the new excitatory activities could have influenced the change of the neuron size. The new activities might produce contraction of target organs in the pelvic viscera.

Noradrenaline-mediated lateral inhibition in the sympathetic ganglion

Intracellular recordings were made from neurons of denervated superior cervical ganglion (SCG) of the rat. Focal electrical stimulation near the recording site evoked a monosynaptic inhibitory postsynaptic potential (IPSP) in about 15% of SCG neurons. The IPSP was reversibly abolished by alpha2-adrenoceptor antagonist, yohimbine. The ionic mechanism of the IPSP appeared to be due to an increase of potassium conductance following the activation of alpha2-adrenoceptor. Electron microscopic observations confirmed that these cells received synapses of intrinsic type. It was concluded that some SCG neurons were under an inhibitory control from neighboring neurons via noradrenergic synapses.

Exocytotic release from neuronal cell bodies, dendrites and nerve terminals in sympathetic ganglia of the rat, and its differential regulation

Stimulant-induced exocytosis has been demonstrated in sympathetic ganglia of the rat by in vitro incubation of excised ganglia in the presence of tannic acid, which stabilizes vesicle cores after their exocytotic release. Sites of exocytosis were observed along non-synaptic regions of the surfaces of neuron somata and dendrites, including regions of dendrosomatic and dendrodendritic apposition, as well as along the surfaces of nerve terminals About half the exocytoses associated with nerve terminals were parasynaptic or synaptic, and these appeared mostly to arise from the presynaptic terminal, but occasionally from the postsynaptic element. The results demonstrated that the neurons of sympathetic ganglia release materials intraganglionically in response to stimulation, that release from different parts of the neuron is subject to independent regulation, at least via cholinergic receptors, and that release is partly diffuse, potentially mediating autocrine or paracrine effects, and partly targeted toward other neurons, but that the latter mode is not necessarily, and not evidently, synaptic. Specifically, exocytosis from all locations increased significantly during incubation in modified Krebs' solution containing 56 nm potassium. Observation of the effects of cholinergic agonists (nicotine, carbachol, oxotremorine) and antagonists (atropine, AF-DX 116) showed that nicotinic and muscarinic excitation each, independently, increased the incidence of exocytosis from somata and dendrites. Exocytosis from nerve endings was not altered by nicotine, but was enhanced or, at high initial rates of exocytosis, decreased, by muscarinic stimulation. Evidence was obtained for muscarinic auto-inhibition of exocytosis from nerve terminals, occurring under basal incubation conditions, and for a muscarinic excitatory component of somatic exocytosis, elicitable by endogenous acetylcholine. The M2-selective muscarinic antagonist AF-DX 116 was found to modify the exocytotic response of the dendrites to oxotremorine, widening the range of its variation; this effect is consistent with recent evidence for the presence of M2-like muscarinic binding sites, in addition to M1-like binding, upon these dendrites [Ramcharan E. J. and Matthews M. R. (1996) Neuroscience 71, 797-832]. Over all conditions, disproportionately more sites of somatic and dendritic exocytosis were found to be located in regions of dendrosomatic and dendrodendritic apposition than would be expected from the relative extent of the neuronal surface occupied by these relationships. Such mechanisms of intraganglionic release may be expected to contribute to the regulation and integration of the behaviour of the various functionally distinctive populations of neurons in these ganglia, by autocrine, paracrine, and focal, neuroneuronal, routes of action. Similar phenomena of exocytotic soma-dendritic release might prove to subserve integrative neuroneuronal interactions more widely throughout the nervous system.

Arborization pattern of sympathetic preganglionic axons in the rat superior cervical and stellate ganglia

Anterograde labeling technique with Phaseolus Vulgaris leucoagglutinin (PHA-L) was employed to observe how a single preganglionic axon arborizes in the superior cervical ganglion (SCG) and stellate ganglion (STG) of rats. PHA-L was injected into the intermediolateral nucleus of the spinal cord at the middle point between segments T1 and T2, and labeled axons were detected immunohistochemically in serial sections. We traced and drew three preganglionic axons over their full length in the SCG and STG. In SCG, the labeled axons bifurcated repeatedly and extended to a length of 600-700 microns in the rostrocaudal direction, and about 200 microns in the transverse direction. These three preganglionic axons made 11, 14 and 11 dense terminal plexus regions along their trajectory. The pattern of the most dense terminal plexus corresponded to the pericellular type dendritic plexus, one of the plexus patterns of dendritic collaterals of SCG neurons. In the STG, the extent of axonal arborization was more variable than that in the SCG, ranging from 400 to 800 microns in the rostrocaudal direction and about 400 microns in the transverse direction. The three analyzed axons made 21, 19 and 20 dense terminal plexus regions along their trajectory, with a similar pattern to those in SCG. These results indicated that there might be a columnar or ellipsoidal organization of postganglionic neurons which are innervated by single preganglionic axons.

Distribution of mRNAs for pituitary adenylate cyclase-activating polypeptide (PACAP), PACAP receptor, vasoactive intestinal polypeptide (VIP), and VIP receptors in the rat superior cervical ganglion

The distribution of mRNAs for pituitary adenylate cyclase-activating polypeptide (PACAP), PACAP receptor (PACAP-R), vasoactive intestinal polypeptide (VIP) and two subtypes of VIP receptors (VIP1-R and VIP2-R) was examined by in situ hybridization in the superior cervical ganglion (SCG) of the adult rat. PACAP-R mRNA was expressed intensely in virtually all principal neurons. PACAP mRNA was expressed in approximately half of the principal neurons, where the levels of expression vary extensively. Intense expression of VIP mRNA was observed only in a few principal neurons. Neither VIP1-R mRNA nor VIP2-R mRNA was detected in SCG cells. These findings suggest that PACAP, but not VIP, may function as a paracrine or autocrine regulatory factor through PACAP-R in the principal neurons of the SCG.

Plasticity of neuropeptide Y in the rat superior cervical ganglion in response to nerve lesion

Axotomy of the rat superior cervical ganglion results in a two-fold increase of neuropeptide tyrosine as determined by radioimmunoassay. On the other hand, treatment of sympathetic neuron cultures with leukemia inhibitory factor, a cytokine that is known to be involved in the up-regulation of galanin after axotomy in vivo, decreases neuropeptide tyrosine messenger RNA. These, apparently contradictory findings, prompted us to investigate the regulation of neuropeptide tyrosine in the axotomized superior cervical ganglion in vivo. For comparison, the regulation of galanin was examined under the same conditions. Compared to control ganglia, the number of neuropeptide tyrosine-positive cell bodies decreased while the density of immunoreactive neuronal processes increased one week after transection of the major postganglionic nerves. The nerve fibres were identified as axons by the absence of MAP2, a somatodendritic marker protein. They extended into both carotid nerves and ramified at the lesion site. In situ hybridization revealed that, although the number of neuropeptide tyrosine messenger RNA-positive neurons was not different from controls, the average grain density/neuron decreased by 40%. When axotomized ganglia were decentralized simultaneously, a three-fold elevation of neuropeptide tyrosine immunoreactivity was detectable by radioimmunoassay and an additional increase in numerical density of neuropeptide tyrosine-immunoreactive nerve fibres was observed. Levels of neuropeptide tyrosine messenger RNA were significantly reduced within postganglionic neurons. This synergistic effect of combined axotomy and decentralization on peptide content was also detected for the neuropeptide galanin that, in contrast to neuropeptide tyrosine, is induced by axotomy or decentralization on protein and messenger RNA level. Therefore, while neuropeptide tyrosine messenger RNA is reduced in axotomized ganglia (most likely in response to leukemia inhibitory factor), the peptide accumulates in axonal processes resulting in increased peptide levels as determined by radioimmunoassay.

Ultrastructure of neuronal circuitry in sympathetic ganglia

To elucidate the intraganglionic circuitry in sympathetic ganglia, attempts have been made to define the nature and source of those neuronal elements that establish synaptic connections there. Intracellular labeling of sympathetic cells is of particular value for this purpose. Dendrites of principal neurons in the rat superior cervical ganglion exhibit a varying complexity in their morphology and arborization. Some dendrites show specializations such as a glomerular plexus, where extensively-branched dendritic collaterals form synaptic connections comprising not only axodendritic synapses between preganglionic axons and principal cell dendrites, but also dendrodendritic synapses between principal cell dendrites. A few of these may represent reciprocal synapses. Most presynaptic elements of adrenergic synapses observed by conventional methods appear to represent these specialized dendritic collaterals of principal neurons. These presynaptic dendrites may be an important addition to the conventional scheme of intraganglionic synaptic organization. However, there seem to be extreme species and even strain differences in the number of these adrenergic synapses, and in the sophistication of the specialized local circuits within sympathetic ganglia. Sympathetic ganglia may thus function as more than a simple relay station, with specialized neuronal circuitry that may be involved in the modulation of cholinergic transmission.

Localization, regulation and functions of neurotransmitters and neuromodulators in cervical sympathetic ganglia

Cervical sympathetic ganglia represent a suitable model for studying the establishment and plasticity of neurochemical organization in the nervous system since sympathetic postganglionic neurons: (1) express several neuromediators, i.e., short acting transmitters, neuropeptide modulators and radicals, in different combinations; (2) receive synaptic input from a limited number of morphologically and neurochemically well-defined neuron populations in the central and peripheral nervous systems (anterograde influence on phenotype); (3) can be classified morphologically and neurochemically by the target they innervate (retrograde influence on phenotype); (4) regenerate readily, making it possible to study changes in neuromediator content after axonal lesion and their possible influence on peripheral nerve regeneration; (5) can be maintained in vitro in order to investigate effects of soluble factors as well as of membrane bound molecules on neuromediator expression; and (6) are easily accessible. Acetylcholine and noradrenaline, as well as neuropeptides and the recently discovered radical, nitric oxide, are discussed with respect to their localization and possible functions in the mammalian superior cervical and cervicothoracic (stellate) paravertebral ganglia. Furthermore, mechanisms regulating transmitter synthesis in sympathetic neurons in vivo and in vitro, such as soluble factors, cell contact or electrical activity, are summarized, since modulation of transmitter synthesis, release and metabolism plays a key role in the neuronal response to environmental influences.

Three electrophysiological classes of guinea pig sympathetic postganglionic neurone have distinct morphologies

Sympathetic postganglionic neurones can be differentiated electrophysiologically into three classes (phasic, Ph; tonic, T; and long-afterhyperpolarising, LAH) based on their potassium channel expression and consequent differences in excitability. We tested whether neuronal morphology differs between these classes. Neurones in coeliac, inferior mesenteric, and lower lumbar paravertebral ganglia of guinea pigs were filled with biocytin during in vitro experiments in which electrical properties were recorded. The dimensions of somata and dendrites were measured in approximately equal numbers of stained neurones of each class. The three electrophysiological classes were distinct in terms of soma shape, soma size (Ph < T = LAH), total dendritic length (LAH < Ph < T) and average length of dendrites (LAH < Ph < T) (P < 0.0001, multivariate analysis of variance). The mean number of primary dendrites also differed (LAH 13, Ph 16, T 20). The majority of dendrites did not branch, the ratios of terminations to primary dendrites being 1.36 (LAH), 1.63 (Ph) and 1.81 (T). Overall, LAH neurones, with medium-sized somata but the smallest dendritic trees, were more distinct morphologically than Ph and T neurones. The morphological differences between classes were not dependent on differences in location. Further, there was no apparent relation between morphology and the pattern of synaptic input each class receives. The results indicate that three distinct groups of sympathetic postganglionic neurone exist in adult guinea pigs, although more than three functions are subserved by these neurones.

Lack of evidence for P2X-purinoceptor involvement in fast synaptic responses in intact sympathetic ganglia isolated from guinea-pigs

Recordings were made from neurons in intact pre- and paravertebral guinea-pig sympathetic ganglia using intracellular microelectrodes. Fast excitatory synaptic responses were evoked by stimulation of preganglionic and peripheral nerve trunks. Suramin (0.1-1 mM) did not affect passive or active membrane properties, nor the amplitude or decay time-course of either synaptic potentials or synaptic currents. Synaptic responses were reversibly reduced in amplitude by hexamethonium (98.7 +/- 0.8%, 50-1000 microM) and d-tubocurarine (95.3 +/- 2.6%, 10-280 microM). ATP (0.5-1 mM) and alpha,beta-methylene ATP (1-40 microM) applied in the bathing solution produced no significant changes in resting membrane potential or input resistance. Prolonged application (up to 25 min) of either compound was also without effect on synaptic responses. These substances also did not affect ganglion cells axotomized one to five days in vivo. These data suggest that activation of P2X-purinoceptors is not involved in the generation of fast excitatory synaptic responses in intact guinea-pig sympathetic ganglia. It appears that dissociation of these neurons must markedly increase their sensitivity to purine nucleotides.

Correlation between dendrodendritic synapses of adrenergic type and synaptically evoked hyperpolarization in the sympathetic ganglion of adult rats

Intracellular recording and labeling with biocytin followed by electron microscopic observation were used to examine the nature and the morphological basis of a synaptically evoked hyperpolarization following spikes in the rat superior cervical ganglion neurons. A large hyperpolarization (the amplitude > 8 mV; the duration > 1 s following spikes) was elicited by repetitive stimulation of the preganglionic nerves in 8% of cells examined (n = 50). The alpha 2-adrenoceptor antagonist, yohimbine, reversibly attenuated the hyperpolarization, without affecting spikes. A nicotinic antagonist, hexamethonium, blocked both the hyperpolarization and spikes. Atropine had no effect of these responses. Electron microscopic observation of dendrites of these cells revealed that they received synaptic inputs of adrenergic type besides a cholinergic one from the preganglionic axons. Some dendrites served as presynaptic elements. These results strongly suggest that the hyperpolarization is an inhibitory postsynaptic potential and that this disynaptic response to the preganglionic stimulation is mediated mainly by two transmitters, acetylcholine and noradrenaline that are released from axodendritic and dendrodendritic synapses, respectively. We conclude that there appears to be an adrenergic inhibitory local circuit that modulates cholinergic transmission in the sympathetic ganglia.

Neuron labeling by extracellular delivery of horseradish peroxidase in vivo: a method for studying the local circuitry of projection and interneurons at physiologically characterized sites

An anatomical method is described that yields individual neurons with continuously labeled dendrites and axons following the extracellular deposition of horseradish peroxidase (HRP) at neurophysiological recording sites in vivo. The method is a logical evolution of previous methods for iontophoretic delivery of HRP: Parameters critical to the ultimate concentration of HRP at the labeling site are reduced by an order of magnitude relative to standard practice. In successful cases one neuron or two in the immediate vicinity (50 microns) of recording sites is/are labeled. Labeling of other processes traversing the injection site, if any, is subliminal at highest light microscopic magnification. Due to the labeling of so few cells and the absence of other labeled processes, dendritic trees and local axonal arbors can be reconstructed without ambiguity. In addition to recovering neurons at sites characterized with physiological (e.g., sensory) stimuli, the method offers the further advantage of being fully compatible with subsequent electron microscopy. Both large (> 20 microns) and small (approximately 8 microns) neuron types and glia have been labeled.

Neuronal and nonneuronal expression of neurotrophins and their receptors in sensory and sympathetic ganglia suggest new intercellular trophic interactions

Nerve growth factor promotes the survival of populations of sensory and sympathetic neurons. Although ganglia have been used for classical assays of neurotrophin action, knowledge is incomplete regarding the spatial arrangements through which neurotrophins are delivered to responsive cells within the ganglia and their attached nerve trunks. Whereas populations of ganglionic neurons may be capable of responding to a particular neurotrophin in vitro, the spectrum of receptor components and neurotrophins expressed by the various neuronal and nonneuronal cells comprising the ganglia in adult rats remains to be elucidated in vivo. Brain-derived neurotrophic factor (BDNF) mRNA was expressed by a population of small to medium sized neurons in all sensory ganglia except in the mesencephalic nucleus of the trigeminal nerve. Interestingly, BDNF immunoreactivity was detected in a more widespread population of neurons of these ganglia, as well as in scattered satellite cells of both sensory and sympathetic ganglia. These nonneuronal cells also expressed mRNA encoding a truncated form of the BDNF receptor, trkBtrunc, and full-length transcripts of trkB appeared to be confined to neuronal populations. Several other components of neurotrophin receptors (low-affinity neurotrophin receptor, trk, and trkC) were prominently expressed by different populations of neuronal cells in sympathetic and sensory ganglia, but they were not detected in nonneuronal cells. Neither nerve growth factor nor neurotrophin-3 mRNAs were detected in these ganglia. Unexpectedly, BDNF and trkBtrunc expression was detected in oligodendrocytes myelinating the central processes of sensory neurons. Schwann cells did not express detectable quantities of either entity, thereby establishing a dramatic boundary delineated by neurotrophin/neurotrophin receptor expression that coincided with the interface between the oligodendroglia of the central nervous system (CNS) and Schwann cells of the peripheral nervous system (PNS). Localization of BDNF expression to an additional population of nonneuronal cells--satellite cells within sensory and sympathetic ganglia--suggest a more extensive role for neurotrophic factors than originally encompassed by the target-derived neurotrophic-factor-concept paradigm. These data support the hypothesis of a possible autocrine or paracrine trophic interaction between populations of neuronal and nonneuronal cells in the peripheral nervous system. BDNF expression in oligodendrocytes but not in Schwann cells at the CNS/PNS junction may provide an additional means of maintaining cell-appropriate connections in the nervous system.

Origins of nerve terminals containing nitric oxide synthase in the guinea-pig coeliac ganglion

Nitric oxide synthase was localised immunohistochemically and by NADPH diaphorase activity in two groups of nerve terminals and in rare cell bodies in the guinea-pig coeliac ganglion. Strongly reactive varicose terminals surrounded a subgroup of principal ganglion cells, most of which were in the medial lobes of the ganglion and most of which were somatostatin immunoreactive. A second set of varicose terminals, which were less intensely reactive, were found throughout the ganglia. Nitric oxide synthase containing nerve cell bodies in the intermediolateral cell columns of the spinal cord were labelled by dye retrogradely transported from the coeliac ganglion. Lesion of nerve connections between abdominal viscera and the coeliac ganglion caused a loss of the strongly reactive fibres, while the widely distributed, less intensely reactive fibres persisted. It is concluded that nitric oxide synthase terminals in the coeliac ganglion come from two sources, sympathetic preganglionic neurons and intestinofugal neurons.

Ultrastructure of the macaque ciliary ganglion

The primate ciliary ganglion is an obligatory relay in the pathways that control the lens and pupil for the near response and the light reflex, two functions which have been the target of increasing inquiry in behavioural physiology paradigms. This investigation provides a comprehensive description of the ultrastructure of the ciliary ganglion in the rhesus monkey (Macaca mulatta). The results indicate that the ciliary ganglion contains a heterogeneous population of neurons in terms of somatic size, cytoplasmic contents and somatodendritic distribution of terminals. Variations in the clear and dense-cored vesicle content of the synaptic profiles present in the ganglion suggest that the synaptic inputs are also heterogeneous and may mediate separate functions. Several characteristic ultrastructural features of the macaque ciliary ganglion are noteworthy. Despite the large size of the neuronal somata, most cells do not exhibit contacts directly onto the somatic membrane. However, the few somata that do receive direct input often display several axosomatic contacts. The vast majority of synaptic interactions occur in the perisomatic neuropil, where the postsynaptic elements consist of simple and complex somatic appendages, as well as dendrites with their appendages. There is little neuropil independent of these immediately perisomatic regions. In some cases, axonal terminals form the central element of complex glomeruli, in which they are presynaptic to numerous spine-like profiles. In other cases, axon terminals and their postsynaptic targets are found within shallow depressions in the somatic membrane or, occasionally, deeply embedded within the borders of the postganglionic neuron. The somata and all the non-myelinated neuronal elements are surrounded by interdigitating, electron-dense processes of satellite cells. These glial cells are sometimes found in shallow recesses, or deeply embedded within the borders of the neuronal somata. The complexity of the ultrastructure of the ciliary ganglion in the macaque suggests that this ganglion may not be a simple relay in the parasympathetic outflow to the eye, but may instead be the site of neuronal processing of the preganglionic input.