The complexity and diversity of synaptic transmission in the prevertebral sympathetic ganglia

Progress in Neurobiology Vol. 24. pp. 43 to 93. 1985

0301-0082/85 $0.00 + .50 Copyright(~) 1985 Pergamon Press Ltd

Printed in Great Britain. All rights reserved

THE COMPLEXITY A N D DIVERSITY OF SYNAPTIC TRANSMISSION IN THE P R E V E R T E B R A L SYMPATHETIC GANGLIA MARK A. SIMMONS Department of Anatomy, Emory University School of Medicine, Atlanta, GA 30322, U.S.A. (Received 15 November 1984)

Contents Abbreviations 1. Introduction 2. Structure of the prevertebral ganglia 2.1. General organization of the autonomic ganglia 2.2. Anatomy of the prevertebral ganglia 2.2.1. Celiac ganglia 2.2.2. Superior mesenteric ganglion 2.2.3. Inferior mesenteric ganglion 2.2.4. Pelvic plexus 2.3. Morphology of the prevertebral ganglia 2.3.1. Histology 2.3.2. Ultrastructure 2.3.3. Small intensely fluorescent cells 2.3.4. Number of fibers 2.3.4.1. Visceral afferents 2.3.4.2. Preganglionic/postganglionicsympathetics 3. Function of the prevertebral ganglia 3.1. Visceral innervation 3.1.1. Celiac-superior mesenteric ganglia 3.1.2. Inferior mesenteric ganglion 3.1.3. Pelvic ganglia 3.2. Modulation of intestinal motility 3.2.1. In situ studies 3.2.2. In vitro studies 3.3. The prevertebral ganglia as reflex centers 4. Electrophysiology of the prevertebral ganglia 4.1. Extracellular recordings 4.1.1. Conduction velocity and synaptic delay 4.1.2. Pathways through the ganglion 4.1.3. Peripheral reflex pathways 4.2. Intracellular recordings 4.2.1. Basic properties of ganglion cells 4.2.2. Synaptic inputs to ganglion cells 4.2.2.1. Celiac-superior mesenteric ganglia 4.2.2.2. Inferior mesenteric ganglion 4.2.2.3. Pelvic ganglia 4.2.3. Prevertebral ganglia-colon preparations 5. Noncholinergic transmission in sympathetic ganglia 5.1. Bullfrog sympathetic ganglia 5.1.1. Description of slow potentials 5.1.2. The transmitter of the LS-EPSP in the bullfrog 5.1.3. Voltage clamp studies of the LS-EPSP 5.2. Guinea pig celiac ganglion 5.2.1. Characteristics of the LS-EPSP 5.2.2. Similarity of the LS-EPSP and 5-HT depolarization 5.2.3. Pharmacological manipulation of serotonergic system 5.3. Guinea pig inferior mesenteric ganglion 5.3.1. Characteristics of the LS-EPSP 5.3.2. The transmitter of the LS-EPSP in the guinea pig 5.3.2.1. Actions of substance P 5.3.2.2. Substance P mimics the LS-EPSP 5.3.2.3. Release of substance P 5.3.2.4. Release and depletion of SP by capsaicin 5.3.3. Effects of enkephalins in the guinea pig IMG 5.4. Rabbit inferior mesenteric ganglion 5.5. Noncholinergic inhibitory potentials 43

44 45 45 45 46 46 46 46 49 51 51 51 52 53 53 53 54 54 54 54 54 54 55 55 56 57 57 57 58 59 59 60 61 61 61 63 63 65 65 65 66 66 67 67 69 69 69 69 71 71 72 73 73 75 77 78

44

M. :\. SIMMONS

6. The localization of peptides in the prevertebrat ganglia 6.1. Bombesin 6.2. Cholecystokinin 6.3. Enkepbalins (~.4. Neuropeptide Y 6.5. Somatoslatin ~'~.6. Substance P 6.7. Vasoactive intestinal polypeptide 6.8. Coexistence of peptides and norepinephrine 7. Tracing pathways of the prevertebral ganglia 7.1. Celiac ganglion 7.2. Inferior mesenteric ganglion 7.3. Application of horseradish peroxidase to the inferior mesenteric ganglion 7.4. Application of HRP to the nerve trunks 7.5. Preganglionic fibers to the IMG 7.6. Enkephalin-containing preganglionic neurons 7,7, Substance P-containing sensory neurons 7.8. Peptidergic neurons from the colon 7.8,1, Bombesin 7.8.2. Cholecystokinin 7.8.3. Vasoactive intestinal polypeptide 8. Conclusions Acknowledgements References

Abbreviations ACh--acetylcholine APP--avian pancreatic peptide BOM--bombesin CCK--cholecystokinin CG---celiac ganglion DA~opamine DflH---dopamine-/3-hydroxylase DH/3E---dihydro-/3-erythroidine DRG---dorsal root ganglion d-TC---d-tubocurarine EAD---early afterdischarge E•--potassium equilibrium potential E,,,--membrane potential E~--resting membrane potential ENK--enkephalin EP1---epinephrine EPSC--excitatory postsynaptic current F-EPSP--fast excitatory postsynaptic potential GK--potassium conductance G,,,--membrane conductance GN,,--sodium conductance HRP--horseradish peroxidase 5-HT--5-hydroxytryptamine, serotonin IM--M current IMG--inferior mesenteric ganglion LAD--late afterdischarge LHRH--luteinizing hormone releasing hormone LLN--late late negative LN--late negative LS-EPSP--late slow excitatory postsynaptic potential LS-IPSP--late slow inhibitory postsynaptic potential NE--norepinephrine NPY--neuropeptide Y PG--pelvic ganglion PNMT--phenylethanolamine-N-methyltransferase R~n--input resistance R,,,--membrane resistance SCG--superior cervical ganglion S-EPSP--slow excitatory postsynaptic potential SIF--small intensely fluorescent S-IPSP--slow inhibitory postsynaptic potential SMG---superior mesenteric ganglion SOM--somatostatin SP--substance P TTX--tetrodotoxin VIP--vasoactive intestinal polypeptide

~2 83 83 83 84 84 84 85 85 85 86 86 86 88 88

SYNAPT1CTRANSMISSIONIN THE PREVERTEBRALSYMPATHETICGANGLIA

45

1. Introduction

The sympathetic ganglia have traditionally been considered to serve as simple relay stations from the central nervous system to the peripheral visceral organs. In this role the adrenergic ganglionic neurons received impulses exclusively from cholinergic preganglionic neurons and relayed these impulses in a diffuse fashion to the viscera; no integrative functions were thought to occur in the ganglia. More recently, the sympathetic ganglia, in particular the prevertebral ganglia, have been found to have a much more complex organization and, accordingly, a more integrative function has been attributed to these structures. Several recent findings illustrate the complexity of synaptic transmission in the prevertebral sympathetic ganglia. Firstly, the prevertebral ganglia receive a number of inputs from multiple nerve trunks. Not all of these inputs originate in the preganglionic nuclei of the spinal cord; fibers also arise from cell bodies in the viscera. Secondly, afferent fibers with cell bodies in the dorsal root ganglia traverse and apparently effect synapses in the prevertebral ganglia. Thirdly, immunoreactivity to a number of neuropeptides has been observed in neuronal structures in these ganglia. Some of these peptides exert specific neurotransmitter-like effects at selected sites in the central and peripheral nervous systems and also in the prevertebral ganglia. Results such as these demonstrate that the prevertebral ganglia do not function solely to relay cholinergic preganglionic impulses from the spinal cord to the peripheral visceral organs, but that these ganglia are also involved in the integration of both centrally and peripherally originating synaptic information transmitted by both acetylcholine and noncholinergic transmitters. These findings, which have led to a re-evaluation of the traditional concepts of sympathetic neurotransmission, are discussed in detail here. The abbreviations used are listed at the beginning of the article.

2. Structure of the Prevertebral Ganglia 2.1.

GENERAL ORGANIZATION OF THE AUTONOMIC GANGLIA

The autonomic ganglia are aggregates of nerve cells surrounded by glial cells and connective tissue: The ganglionic neurons receive preganglionic fibers from neurons in the central nervous system and send postganglionic fibers to the effector organs--smooth muscles, glands and the heart. The parasympathetic ganglia receive preganglionic fibers from neurons situated in the brain stem and sacral spinal cord. The sympathetic ganglia, of interest here, receive preganglionic fibers from neurons in the thoracic to lumbar spinal cord levels. The cell bodies of the sympathetic preganglionic neurons are mainly located in the nucleus intermediolateralis of the spinal cord. Sympathetic ganglia can be classified into two types according to anatomical position-paravertebral and prevertebral. The paravertebral ganglia are the sympathetic chain ganglia. These ganglia and the commissures connecting them form the sympathetic trunks situated bilaterally along the ventro-lateral aspect of the spinal column from the first thoracic segment caudally to upper lumbar segments. Each ganglion is connected with its corresponding spinal nerve by communicating rami. Preganglionic fibers enter the ganglia via the white rami; postganglionic axons exit through the gray rami and join the spinal nerves to travel to the effector organs. The prevertebral ganglia are distinct from the sympathetic trunk in the abdominal cavity (additionally the cervical ganglia may be considered prevertebral, but they will not be considered here). The abdominal prevertebral ganglia lie close to the median sagittal plane of the body, ventral to the abdominal aorta. They include the left and right celiac ganglia and the superior mesenteric ganglion which comprise the solar plexus; the inferior mesenteric ganglia or plexus; and the pelvic plexus. Spinal preganglionic fibers to the prevertebral ganglia traverse the paravertebral ganglia in long nerves (Skok, 1973; Gabella, 1976) but, while the paravertebral ganglia receive most of their preganglionic fibers from the interomediolaterai nuclei of the spinal cord through the white rami, the

46

M . m . SIMMONS

situation in the prevertebral ganglia is more complicated. The spinal preganglionic~ take various routes to the prevertebral ganglia, frequently entering the ganglia thl-ough a number of nerve trunks. Additionally, as will be shown below, the prcvertebral gangli~l also receive synaptic inputs of peripheral origin. Consequently, the use of the icr'm "preganglionic" with respect to the prevertebral ganglia is not limited to neurons with soma in the spinal cord, but includes any neuron which synapses on a prevertebral ganglionic neuron. 2.2 ANATOMY OF THE PREVERTEBRAL GANGLIA

Detailed studies of the gross anatomy of the autonomic nervous system were performed by Langley and Anderson (1896b) at the end of the last century and subsequent studies have confirmed the accuracy of their descriptions (Trumble, 1933). Drawings of the prevertebral ganglia of the rabbit and human are presented in Figs 1 and 2, respectively. A schematic diagram showing the relative locations of the prevertebral ganglia is shown in Fig. 3. There is considerable variability in the anatomy of the prevertebral ganglia both within and between species. The relative positions of the ganglia are similar in different species. The major difference is an increase in the size and number of ganglia as the size of the animal increases. There is also variability in the anatomy within a given species. In addition to the major ganglia, numerous smaller ganglia are found along the vasculature throughout the abdominal region (Kuntz and Jacobs, 1955). Sources for illustrations of the prevertebral ganglia are given in Table 1. 2.2.1. Celiac ganglia The celiac ganglia are found near the juncture of the celiac artery and the aorta, between the adrenal glands and behind the stomach (Mitchell, 1953; Syromyatnikov and Skok, 1968; Skok, 1973; Kreulen and Szurszewski, 1979b). These have also been termed the semilunar ganglia, though they are rarely so shaped. The left and right ganglia are of similar size and are usually joined medially by a commissure of ganglion cells or nerve fibers. The nerves from the spinal cord associated with each celiac ganglion include the left and right greater and lesser splanchnic nerves. The splanchnic nerves carry preganglionic fibers from the lower thoracic spinal cord through the paravertebral sympathetic trunk to enter the lateral aspect of the ganglia. The nerves leaving the CG have been variously termed. For example, the nerves labelled CN in Fig. 3 have been termed, respectively, the superior and inferior celiac nerves (Kreulen and Szurszewski, 1979b); the medial and lateral nerves of the celiac plexus (Syromyatnikov and Skok, 1968) or the superior and inferior hepatogastric bundles (Kuo and Krauthamer, 1981). These fibers run along the celiac artery and its branches to the viscera. 2.2.2. Superior mesenteric ganglion Caudally along the aorta near the superior mesenteric artery is the superior mesenteric ganglion (Mitchell, 1953; Kreulen and Szurszewski, 1979b). Numerous nerve fibers connect this ganglion with the left and right celiac ganglia. The intermesenteric nerve, which has also been termed the ascending mesenteric nerve and the celiac root of the IMG, connects the SMG and the IMG. Fibers also run along the superior mesenteric artery and its branches. 2.2.3. Inferior mesenteric ganglion The inferior mesenteric ganglion is found near the exit of the inferior mesenteric artery from the aorta. The inferior, or lumbar, splanchnic nerves exit from the lumbar spinal cord, traverse the sympathetic trunks and enter the IMG. Frequently, one or more of these fibers joins the intermesenteric nerve to enter the ganglion (Langley and Anderson, 1896b). The intermesenteric nerve emanates from the rostral pole of the IMG, running towards the SMG. Leaving the IMG laterally and continuing along the inferior mesenteric artery are the lumbar colonic nerves. Two nerve strands are usually seen emanating from

47

SYNAPTIC TRANSMISSION IN THE PREVERTEBRAL SYMPATHETICGANGLIA

/

.

,

--

Z~

/,

+,ZC'V

<

> 'FIG. 1. Drawing of the anatomy of the abdominal nerves of a rabbit. From Langley and Anderson (1896b). With permission of the Physiological Society.

SYNAPTIC TRANSMISSION IN THE PREVERTEBRALSYMPATHETICGANGLIA

Splanchnic

49

Ns

~ami

ref ve£

.anchnic Ns

NI

Cc Inf~

Infer (

Trunk N.[

t.hrc

; Ns

FIG. 2. Diagram of the abdominal autonomic nerves in man. From H. C. T r u m b l e (1933) Brit. J. Surg. 20, 67(I. With permission of the publishers, Butterworth & Co. (Publishers) Ltd. Copyright 1933.

the posterior end of the IMG. These are the hypogastric nerves. In the rabbit IMG, a number of filaments leave the upper part of the ganglion laterally and pass anteriorly along the inferior mesenteric vein; these have been termed the ascending mesenteric nerve (Langley and Anderson, 1896b). These branches join strands from the superior mesenteric ganglion and give off filaments along the vasculature to the colon. 2.2.4.

Pelvicplexus

The pelvic plexus is formed by the branches of the hypogastric nerves, branches of the pelvic nerves, fibers from the sacral spinal cord and numerous ganglia with fibers arising from them (Kuntz and Moseley, 1936; Owman et al., 1983; Wo2niak and Skowroriska, 1967). The number and position of ganglia in this region are extremely variable. This plexus has also been termed the hypogastric plexus or the hypogastric-pelvic plexus. Some of the ganglia in this region receive spinal preganglionics from lumbar levels, some receive them from sacral spinal levels and others receive them from both spinal regions (Kuntz and Moseley, 1936), hence, this plexus contains sympathetic, parasympathetic and mixed ganglia. In general, the cells in the ganglia along the hypogastric nerve to the hypogastric

50

M . A . SIMMONS

i

~

--~LSN

i i\,MN

Fi6.3. Schematic diagram of the general organization of the abdominal prevertebral ganglia. CA, celiac artery; CC, connecting commissures; CG, celiac ganglia; CN, celiac nerves; GSN, greater splanchnic nerves; H G N , hypogastric nerves; IMA, inferior mesenteric artery; IMG, inferior mesenteric ganglion; IMN, intermesenteric nerve; ISN, inferior splanchnic nerves; LCN, lumbar colonic nerves; LSN, lesser splanchnic nerves; PG, pelvic ganglia; SMA, superior mesenteric artery; SMG, superior mesenteric ganglion.

nerve-pelvic nerve junction receive inputs from the hypogastric nerve and would be considered sympathetic, while those along the pelvic nerve are innervated via the pelvic nerve and would be classified as parasympathetic (Crowcroft and Szurszewski, 1971). These ganglia are composed of a relatively few cells compared to the more rostral prevertebral ganglia (Blackman et al., 1969).

TABLE1. 1LLUSTRATIONSOFTHEANATOMYOFTHEPREVERTEBRALGANGLIA Ganglion All

CG-SMG

IMG PG

Species Cat Rabbit Man Monkey Man Cat

Guinea pig Cat Guineapig Rabbit Cat Dog Rabbit Man Monkey Rat

Referencc Langley and Anderson, 1896b Trumble. 1933 Mitchell, 1953 Christensen e t a l . , 1951 Kuo and Krauthamer, 1981 Skok, 1973 Syromyamikov and Skok. 1968 Kruelen and Szurszewski, 1979b Skok, 1973 Crowcroft e t a l . , 1971 Simmons and Dun, 1985 Wozniak and Skowronska, 1967 SjOstrand, 1965

Langworthy, 1965 Purington e t a l , , 1973 Owman e t a l . , 1983

SYNAPTIC TRANSMISSIONIN THE PREVERTEBRAL SYMPATHETICGANGLIA

2.3.

51

MORPHOLOGY OF THE PREVERTEBRAL GANGLIA

Studies on the detailed structure of the prevertebral ganglia have been rather limited. Histological examinations have been performed with respect to the CG and IMG of the human and of the cat. The ultrastructure of the cat IMG and guinea pig PG has been examined. More recent studies on the neural connections of the prevertebral ganglia have used the guinea pig. 2.3.1. Histology Histological examination of the human CG and IMG reveals ganglion cells of medium size which are stellate in form (Kuntz, 1938, 1940). The cells have long dendrites which branch relatively infrequently and radiate from the cell bodies in all directions. Some cells also have short, or accessory, dendrites with numerous short branches. Adjacent cells are intimately related through their dendrites. A single ganglion cell may become related with a relatively large number of neighboring cells. In some instances, terminal arborizations of dendrites form dendritic nests around the cell bodies of adjacent cells. Preganglionic axons run through these dendritic nests. Axons enter the IMG in bundles containing preganglionic fibers as well as visceral afferents traversing the ganglion. Axons leave these bundles and ramify among the dendrites. The preganglionic axons spiral around the dendrites and cell body. The postganglionic axons do not ramify among the ganglion cells but aggregate into bundles which emerge from the ganglion. The caliber of these axons is similar to that of the preganglionic fibers. The fiber tracts in the CG also carry vagal fibers which pass through the ganglion, apparently without synapsing (Kuntz, 1938). The morphology is essentially the same in the cat but cells with short dendrites are less common (Kuntz, 1940) and pericellular and peridendritic nests are observed less frequently (M'Fadden et al., 1935; Kuntz, 1940). Cat ganglion cells are about 20 to 40/xm in size (M'Fadden etal., 1935). Cells in the pelvic ganglia of guinea pigs are of similar size and occur in small groups separated by connective tissue septa (Blackman et al., 1969; Watanabe, 1971). Using the histochemical method of Falck and Hillarp for the fluorescent localization of monoamines (Falck, 1962), Hamberger et al. (1964) found noradrenergic terminals surrounding cat sympathetic ganglion cells in basket-like structures. These terminals apparently correspond to the peridendritic nests of neighboring cells observed by Kuntz (1940). Of all the sympathetic ganglia examined, in both rabbit and cat, the prevertebral ganglia had the most well-developed systems of such terminals (Hamberger et al., 1965). Removal of the prevertebral ganglia from L3 to L7, removal of the celiac ganglia, section of the colonic nerves or section of the hypogastric nerves did not alter the peridendritic nests in the IMG (Hamberger and Norberg, 1965), thus, the nests apparently originate within the ganglion. 2 . 3 . 2 . Ultrastructure The perikarya of IMG neurons have very few synapses (Elfvin, 1971a). Axodendritic synapses are most frequent (Archakova et al., 1982) while axosomatic synapses are seen only occasionally. Short and long processes project from the perikarya. The short processes are connected with the rest of the cytoplasm by a stalk about 100 nm in diameter. These short processes extend 1 to 3/zm from the surface of the cell body. The long processes are connected by a stalk which is 1 to 3 ttm wide. Often several synaptic sites are seen on both types of processes. Sometimes, a single preganglionic terminal synapses with both a short process and the surface of the perikaryon. The presynaptic regions contain clear vesicles and are presumably ACh-containing neurons. In the short cell processes small dense core vesicles are found. These processes form dendrodendritic and dendrosomatic contacts with adjacent ganglion cells. The number of dendrodendritic contacts is about the same as the number of axodendritic contacts made by the preganglionic fibers (Eifvin, 1971c). These short processes containing closely packed dense

52

M . A . SIMMONS

core vesicles correspond to the fluorescent basket-like networks seen with the FalkHillarp technique. In addition to the classical axodendritic synapses made by the preganglionic fibers with the IMG cells, specialized axoaxonic synaptic contacts are seen between separate proganglionic cholinergic fibers (Elfvin, 1971b). In the pelvic ganglia of the guinea pig, Blackman et al. (1969) reported that synapses were seen most often between varicose preganglionic axons and fine cell processes. On the other hand, Watanabe (1971) reported most contacts were axosomatic, but did observe axodendritic and axoaxonic contacts. Numerous contacts were made on a single cell by unmyelinated axons of 0.5 to 2.0/xm diameter. Some of the preganglionic fibers contained small agranutar vesicles, suggesting they were cholinergic, and others had small granular vesicles and were classified as catecholaminergic. The former type predominated. 2.3.3. Small intensely fluorescent cells SIF cells are practically absent in the cat IMG; they have, however, been found in the rabbit IMG (Elfvin, 1968). These cells are about 10 to 25/zm in diameter and occur in groups of three to five cells. The ratio of ganglion cells to SIF cells in the guinea pig IMG has been calculated at 4 : 1 (Crowcroft et al., 1971b). In this ganglion the SIF cells are located in small groups of cells separated from the principal ganglion cells (Elfvin et al., 1975). These groups of SIF cells have been termed paraganglia. The processes of these cells do not approach or contact the principal ganglionic neurons, but do come in close proximity to the blood vessels of the ganglion. At the points of contact between the SIF cells and vessels, dense core vesicles are present in the SIF cell processes. These authors used antibodies to DflH, the enzyme which converts D A to NE, to reveal NE-containing cells, and antibodies to PNMT, which converts NE to EPI, to reveal EPI-containing cells. All of the SIF cells stained for D/3H and a portion also for PNMT. Fluorescent fibers could be observed traveling from the SIF cells to the blood vessels. No fluorescent fibers were seen traveling from the paraganglia to surround the principal ganglionic neurons. The presence of both extremely dense and moderately dense granules in the SIF cells was taken as further support that the SIF cells contain both EPI and NE. (The dense granules are thought to indicate the storage of NE while the less dense indicate EPI.) No evidence was found for D A in the SIF cells of the guinea pig IMG. Consistent with this, no D A could be detected biochemically in the guinea pig IMG (Crowcroft et al., 1971b). The SIF cells in the guinea pig IMG were found to be innervated by two types of axons. The first type contained small agranular vesicles, was not affected by 5-hydroxydopamine and degenerated following section of the lumbar splanchnic and ascending mesenteric nerves (Furness and Sobels, 1976). These are characteristics of cholinergic preganglionic sympathetic fibers. The other type of axon was not affected by decentralization of the ganglion and contained small granular vesicles which increased in electron density following 5-hydroxydopamine, indicating that they were noradrenergic axons originating within the IMG, from the principal ganglionic neurons. These authors also studied the processes of the SIF cells and concluded that the only structures towards which the SIF cells could release transmitter were the capillaries. SIF cells have also been found in the guinea pig pelvic ganglia and are 10 to 20 ~m in diameter (Watanabe, 1971). Axons 1 to 3 tzm thick extend from these cells to terminate near capillaries. These cells receive unmyelinated preganglionic axons to their soma and dendrite and are found most often near the entering/exiting axon bundles (Blackman et al., 1969). In contrast, in the rat PG SIF cell processes are rarely seen near capillaries (Dail et al., 1975). Rather, the processes frequently run between the ganglionic neurons and contact ganglion cell processes. The question remains as to the role of the SIF cells in the prevertebral ganglia. These cells are apparently innervated by both preganglionic cholinergic fibers and noradrenergic ganglion cells, suggesting they would be apprised of ganglionic activity. It is not known whether this innervation is functional and what the resultant response is to ganglionic

SYNAPTIC TRANSMISSION IN THE PREVERTEBRAL SYMPATHETIC GANGLIA

53

activity. The main apparent release site for the SIF cell processes is the capillary. In this regard, the SIF cells have been categorized according to release site: interneurons which are thought to release onto other neurons and paraneurons which supposedly release into ganglion vessels (Taxi et al., 1983; Williams and Jew, 1983). In this context, most of the cells in the prevertebral ganglia would serve as paraneurons, like small endocrine glands releasing their catecholamines into the ganglion vasculature. 2.3.4. Number of fibers 2.3.4.1. Visceral afferents Harris (1943) counted the fibers in the nerve trunks of the cat IMG. In all of these experiments the fibers descending from the CG and SMG were cut. He counted 2,500 visceral afferent fibers in the splanchnic nerves and 5,000 in the peripheral nerves: 3,000 in the colonic nerve and 2,000 in the hypogastric nerve. The splanchnic vs peripheral nerve difference is explained by assuming that upon reaching the IMG each fiber from the spinal ganglion gives off one collateral. An alternative explanation is that some of these visceral afferent fibers do not have their cell bodies in the DRG, but are fibers of enteric origin which terminate in the IMG. This group contained both myelinated and unmyelinated fibers. Most of the myelinated fibers were of small (1-3/zm), some of medium (4-6/zm) and some of large (7-10/~m) diameter. The visceral afferent fibers reaching the colon were exclusively unmyelinated while some of those in the hypogastric nerve were myelinated.

2.3.4.2. Preganglionic/postganglionic sympathetics Approximately 3,000 postganglionic sympathetic fibers enter the cat IMG from the lumbar region of the spinal cord. The intermesenteric nerves had been severed to allow examination exclusively of the lumbar inputs. A total of 9,500 fibers exit the IMG in the colonic and hypogastric nerves. Thus, two out of three postganglionic sympathetic fibers in the colonic and hypogastric nerves have their origin in the IMG, if one assumes that none of the postganglionic fibers branch. Of the 4,000 preganglionic fibers entering the IMG from the lumbar nerves, 1,000 pass through and descend in the hypogastric nerve. If those fibers passing through the ganglion synapse, the ratio of preganglionics to postganglionics is 1 : 1.5, if not, the ratio is 1 : 3 (Harris, 1943). Widely varying values have been reported for the preganglionic: postganglionic ratio in other ganglia. Several factors contribute to this variation. First, it appears that the value obtained depends on the animal examined. For example, in the SCG of the cat reported values range from 1 : 11 to 1:32 (Billingsley and Ranson, 1918; Wolf, 1941), in the dog a ratio of 1:8 was calculated (Sanbe, 1981) and in the human values as low as 1 : 196 have been reported (Ebbesson, 1968). There is even considerable variation among primates as in the SCG of the squirrel monkey a value of 1 : 28 was found (Ebbesson, 1968). A second factor is the ganglion examined. In the human lumbar sympathetic chain ganglia ratios around 1 : 70 have been reported (Webber, 1955). These differences partly reflect variation in the number of neurons in the ganglion rather than in the number of preganglionic fibers. The number of neurons in the human SCG is about ten times greater than the number in the same ganglion in the cat and 20 times as much as the number in the rabbit, while the number of preganglionics in these species is of the same magnitude (Ebbesson, 1963). Brooks-Fournier and Coggeshall (1981) have reported a much higher ratio in the rat SCG (1 : 4). These authors point out that the preganglionic nerve trunks also contain sensory and postganglionic fibers. In previous work, the presence of these non-preganglionic axons in the preganglionic nerve trunks was not considered. Furthermore, accurate counts of the number of the unmyelinated postganglionic axons is only possible under the electron microscope. Low preganglionic:postganglionic ratios have been taken as support for a diffuse control of the viscera by the sympathetic system, as compared to a more precise control of the parasympathetic system. In view of the fact that many of the reported values of preganglionic:postganglionic ratio may overestimate the number of preganglionic fibers and underestimate the number of postganglionic fibers, this interpretation may not

54

M . A . SIMMONS

be justified. The preganglionic:postganglionic ratio in the IMG and the numbc~ oI prevertebral ganglion cells innervated by a single preganglionic fiber are discussed further in Section 4.2.2.

3. Function of the Prevertebral Ganglia

3.1. VISCERALINNERVATION

3.1.1. Celiac-superior mesenteric ganglia Fibers from the CG continue along the celiac artery and its branches to innervate the lower portion of the esophagus, stomach, duodenum, small intestine, liver, spleen, pancreas, kidneys and adrenals (Christensen et al., 1951; Jansson, 1969; Johansson and Langston, 1964; Kuntz and van Buskirk, 1941; Mitchell, 1953; Skok, 1973; Tiscornia, 1977). The SMG sends fibers along the superior mesenteric artery mainly to the intestine (Mitchell, 1953).

3.1.2. Inferior mesenteric ganglion Classical studies on the effector organ responses following electrical stimulation of the IMG and associated nerve trunks were conducted by Langley and Anderson in the 1890s. Visual observation was made of the various organs following electrical stimulation at various sites. The IMG was shown to be involved in the innervation of the base of the bladder. These fibers travel via the hypogastric nerve (Langley and Anderson, 1895b). Fibers in the hypogastric nerve also go to the internal genital organs--the vasa deferentia and seminal vesicles in the male and the uterus and vagina in the female (Langley and Anderson, 1895c). The majority of fibers from this ganglion innervate the gastrointestinal tract, particularly the large intestine (Furness and Costa, 1974; Johansson and Langston, 1964). The postganglionic fibers innervate the vasculature and intrinsic ganglia of the intestine. Few fibers act directly on the muscle coats, except in the sphincter regions where there is a dense innervation. Sympathetic activation results in a decrease in motility due to a decrease of ACh release from the intramural neurons and due to sphincter contraction. Vasoconstriction also occurs. As will be seen, the most well-studied interaction between prevertebral ganglia and viscera involves the IMG and the colon. 3.1.3. Pelvic ganglia The cells in the PG send fibers to the urinary bladder, urethra, uterus, vagina, prostate, vas deferens and seminal vesicles (Wo~niak and Skowrofiska, 1967). The hypogastric nerve contains both adrenergic postganglionic fibers which travel mainly to the bladder and cholinergic preganglionic fibers which innervate the short adrenergic neurons of the PG which themselves innervate the internal genitalia (Owman et al., 1983; Sj6strand, 1965). 3.2. MODULATION OF INTESTINAL MOTILITY Just prior to the beginning of the twentieth century, two laboratories had investigated the innervation of the large intestine and had come to the same conclusion--that stimulation of the lumbar splanchnics or of the colonic nerves results in an inhibition of the large intestine, while stimulation of the sacral nerves results in contraction (Bayliss and Starling, 1900; Langley and Anderson, 1895a, 1896a). There was apparently very little interest in this part of the autonomic nervous system until the late 1920s, when interest in the role of the IMG again increased as investigations on intestinal motility advanced.fin general, the visual observations of these early investigators have been subsequently confirmed with more quantitative methods.

SYNAPT1C TRANSMISSION IN THE PREVERTEBRAL SYMPATHETIC GANGLIA

55

3.2.1. In situ studies Following stimulation of the lumbar sympathetics, the colonic nerves or the hypogastric nerves a marked contraction of the internal anal sphincter occurs in anesthetized dogs (Learmonth and Markowitz, 1929). The effect is greatest following hypogastric nerve stimulation. The colonic nerves also exert inhibitory effects on the colon (Learmonth and Markowitz, 1930). Section of the colonic nerves is followed immediately by an increase in intracolonic pressure and, in some cases, by an increased amplitude of the contractions of the colon. This reveals a tonic inhibitory influence on the musculature of the distal colon. On this basis, section of the inferior mesenteric plexus was suggested as a treatment of Hirschsprung's disease (Rankin and Learmonth, 1930, 1932) and was used successfully in a few cases (Ross, 1935). It is now known that Hirschsprung's disease is due to a congenital absence of the enteric plexuses. Current therapy involves resection of the affected region of the colon (Kleinhaus et al., 1979). Others found that section of the preganglionics or postganglionics of the IMG or removal of the IMG had little effect on the movements of the colon in cats (M'Fadden et al., 1935). This study differed from the above in that examinations were not performed acutely, but observations were made at least one week following surgery. Colonic motility was measured by X-ray observation of the movements of radio-opaque meals or enemas. Although no hypermotility was observed, there was a decrease in transit time through the colon, apparently due to diminished resistance at the internal anal sphincter, a structure shown above to be contracted by stimulation of the fibers of the IMG. Garry (1933b) examined the responses in the large bowel of the cat to stimulation of the lumbar sympathetic outflow. Movements of the large bowel and the anal canal were recorded by two balloons inserted through the anus. Intestinal motility was stimulated by movement, but not distention, of either of the balloons. Stimulation of the lumbar sympathetics exerted an inhibitory influence on the response of both the large bowel and the internal anal sphincter to movement of either balloon. In another study, the behaviour of the large bowel of the cat was also measured by a balloon inserted through the anus (Garry, 1933a) when the lumbar sympathetic outflow was intact the bowel was inactive. Section of the entire lumbar outflow resulted in increased tone and rhythmical activity in the colon. Section of only the spinal rami of the IMG caused a slight increase in tone and rhythmic contractions appeared. Subsequent section of the colonic and hypogastric nerves led to a marked increase in both tone and rhythmicity. Division of the hypogastrics alone had little effect, but section of the colonics led to a marked increase in gut activity, even when the hypogastrics were intact. These studies indicated that, in the absence of lumber spinal inputs, inhibition of the colon was maintained from the IMG via the lumbar colonic nerves. The role of the IMG in the control of intestinal motility has also been examined in dogs (Lawson, 1934). The colonic or hypogastric nerves were sectioned for stimulation. After stimulation of the colonic nerves an inhibition of the colon was observed accompanied by a contraction of the internal anal sphincter. Stimulation of the hypogastric nerves produced consistent effects only in the distal segment of the colon and at the internal anal sphincter. When the spinal rami to the IMG were cut, the inhibitory phase was prolonged. Subsequent section of the colonic nerves when stimulating the hypogastric nerves or section of the hypogastric nerves while stimulating the colonic nerves further prolonged the inhibition. Lawson (1934) concluded that the cells of the IMG act independently of preganglionic influences and that the relationships of the response of one intestinal segment to other segments depends on either automatic or reflex activity in the cells of the IMG. 3.2.2. In vitro studies The rabbit isolated colon has been used to examine the effects of the colonic and pelvic nerves in vitro (Garry and Gillespie, 1955). Pelvic nerve stimulation resulted in contraction of the colon with the maximal response occurring at a 10 Hz stimulation frequency. A

56

M. A, SIMMONS

contraction could be obtained with frequencies as low as 1 Hz. As reported m s i t , . stimulation of the colonic nerves caused an inhibition of the colon. These fibers required a higher frequency of activation as the maximal response was seen at 100 Hz and no response could be elicited below 5 Hz. The inhibition showed no fatigue and was followed by ,1 sudden sharp contraction following cessation of stimulation. The effects of pelvic stimulation, but not the response to colonic nerve stimulation, were antagonized by hexamethonium and atropine. The inhibitory response involved a catecholamine transmitter since it disappeared in colons taken from reserpinized animals (Gillespie and Mackenna, 1961). 3.3. THE PREVERTEBRAI. GANGLIA AS REFLEX CENTERS Prior to 1937, the evidence for independent activity in mammalian sympathetic ganglia was unconvincing, particularly in view of the prevalent concept that the ganglia served merely as relay stations from the central nervous system to the peripheral visceral organs (Gaskell, 1916; Langley, 1921). Garry (1933a) and Lawson (1934), however; had offered evidence for continued inhibition of the intestine by the IMG following acute decentralization, but it was possible that this result was an artifact of trauma and exposure. To clarify these findings, Lawson and Holt (1937) recorded colonic motility as an index of the activity of the IMG in unanesthetized dogs. Two to four weeks following ramisection of the IMG, basal tone levels were decreased in the anal canal and proximal colon and raised in the middle colon. Contraction height was not uniformly affected except in the distal colon, where there was an increase. After removal of the IMG following ramisection basal tone levels were not uniformly affected and in some cases tone was lower; however, contractility was increased throughout the colon, most in the distal colon and anal canal, least in the proximal segment. The increase in contractility following ganglionectomy was established within twelve hours and no progressive changes were noted. These results were interpreted as further evidence that control of the colon by the decentralized IMG is a result of nervous activity originating within the ganglion itself, independent of inputs from the central nervous system. Degeneration experiments revealing the persistence of axons in the distal ends of the nerve trunks of the IMG following nerve section more centrally, further suggested that the IMG was a reflex center. Following extirpation of the lumbar segments of the sympathetic trunk and ascending mesenteric nerves a large percentage, but not all, of the axons in the ganglia degenerated (Harris, 1943; Kuntz, 1940). Following removal of the IMG (M'Fadden et al., 1935) or section of the colonic nerves (Kuntz, 1940), intact fibers were still seen in the distal part of the colonic nerves. Kuntz (1940) suggested that these were fibers of enteric origin which travel centripetally in the colonic nerves; thus, the IMG may be a reflex center. To test this hypothesis, Kuntz (1940) extirpated the lumbar segments of the sympathetic trunks and severed the ascending mesenteric nerves connecting the C G - S M G with the IMG of anesthetized cats. The colon was transected in order to interrupt the enteric plexuses and balloons were placed in the proximal and distal segments of the colon. When the proximal segment was quiescent, inflation of the distal colon had no effect, but when the proximal segment was undergoing rhythmic contractions, inflation of the balloon in the distal segment commonly resulted in inhibition of the proximal segment. In these experiments, all neural pathways from the distal segment of the large intestine to the proximal segment were interrupted, except pathways through the IMG. Thus, the response of the proximal segment following distention of the distal colon can be explained as a reflex response through the IMG. Others, however, were unable to show the persistence of intestino-intestinal reflexes through decentralized prevertebral ganglia (Youmans et al., 1942). In this study jejunal activity in dogs was measured. Inhibitory reflexes were only observed when the connections of the celiac ganglion with the spinal cord were intact. In view of these discrepancies, Kuntz and Saccomanno (1944) combined anatomical and physiological techniques to demonstrate that the IMG could modulate reflexes in the

SYNAPT1C TRANSMISSION 1N THE PREVERTEBRAL SYMPATHETIC GANGLIA

57

absence of central connections. Axon terminations persisted in the IMG following decentralization and intact fibers could be found in the distal segments of nerves from the IMG following section of these nerves. On the assumption that some of these distal fibers made synapses in the IMG, they were regarded as the afferent limb of a reflex arc. The physiological experiments showed that, following removal of the spinal cord from the first thoracic segment caudally, an inhibition of the proximal colon was still observed following distention of the distal segment. This response was not mediated by the enteric plexuses as the colon had been transected between the distending and recording balloons. In the acute experiments it was quite possible that the reponse was a pseudo-reflex, mediated through branches of the visceral afferent fibers; however, a similar response was also seen seven days after decentralization of the IMG by bilateral extirpation of the lumbar sympathetic trunk and section of the ascending and hypogastric nerves. By this time the dorsal root afferents had undergone degeneration. It is concluded that impulses arising in the colon are conducted to the IMG through axons of enteric ganglion cells. These axons synapse onto IMG neurons which transmit impulses back to the colon forming a true reflex arc. A similar series of experiments was conducted on the CG. Kuntz (1938) had shown intact nerve fibers in the CG following bilateral section of the splanchnic nerves and in the distal parts of the mesenteric nerves following section of these nerves near the CG, indicating the presence of fibers of enteric origin. In a further study (Kuntz and van Buskirk, 1941) it was shown that, in cats, distention of the ileum and colon or stimulation of the mesenteric nerves resulted in an inhibition of bile flow in control animals and also in animals whose splanchnic and vagus nerves had been sectioned and jejunum sectioned. Since the complete division of the splanchnic and vagus nerves was verified and since the intestine was transected, this reflex could not have been mediated by central reflex pathways nor through the enteric plexuses. This provided physiological evidence of peripheral reflex connections in the celiac ganglia. Others have confirmed the existence of intestino-intestinal inhibitory reflexes following transection of the intestine and decentralization of the CG (Semba, 1954). More direct evidence for the involvement of the IMG in peripheral reflex activity has been provided by the electrophysiological studies described below.

4. Electrophysiology of the Prevertebral Ganglia The first investigations of the electrophysiology of synaptic transmission through the IMG appeared shortly after studies which recorded the electrical activity of crayfish ganglia (Adrian, 1931), sympathetic nerves (Adrian et al., 1932) and the superior cervical ganglion (Eccles, 1935). 4.1. EXTRACELLULARRECORDINGS Through the use of extracellular recordings from the nerves associated with the prevertebral ganglia, information has been obtained on the conduction velocity in the associated nerve fibers, on the synaptic delay occurring in the ganglia and on the conduction/reflex pathways through the ganglia.

4.1.1. Conduction velocity and synaptic delay The conduction velocities of the various nerves associated with the prevertebral ganglia are listed in Table 2. In general, the preganglionic fibers to the abdominal prevertebral ganglia are slow conducting fibers. These fibers would therefore be classified as C fibers. The actual synaptic delay, from preterminal action potential to beginning of F-EPSP, has not been directly measured in the prevertebral ganglia. The method used has been to plot the distance from stimulating to recording electrode (x-axis) vs latency of response (y-axis) while placing the stimulating electrode at various points along the preganglionic nerve trunk. This gives a line with slope equal to the reciprocal of conduction velocity and JPN 24:1-D

58

~I

,"k. SIMMONS

TABI.~ 2. CC,NTnlCI ION VEI.OCIFIES OF IHF NI RVFS OF Fibers Splanchnic

Hypogastric

E x a m i n e d Species ( at

Dog Rabbit ('at

G u i n e a pig

Intermesenteric

Dog Rabbit

Preganglionic

Rabbit

Postganglionic

Rabbit

Value (m/sec) < I(I <:i),9 2-7 0.5-2.0 II. 1-4.1~ >5 9,7 12.8 {1,9 1.6 0.8 1.5- 10 11.5 11,1 1 . 1 1 2 (1.45 o.25 I). 15-1).6 I. 5 , 6 . 0 (i.25-1).5

I'H!

PREVERIEBRAI (iANGI I:~. Refercncc 1 lord, 1937 Kriereta[.. 1982 King and Szurszewski, 1~84a Brown and Pascoe, 1952 Lloyd, 1937 Adrianet al., 1932 Ferry, 1967 K i n g a n d S z u r s z e w s k i , 1984a Brown and Pascoc, 1952

S i m m o n s and D u n , 1985 S i m m o n s and Dun. 1985

y-intercept equal to the delay. Thus, synaptic delay measured by this method includes not only the delay at the synapse, but also any decrement of conduction velocity as the axon branches or ramifies through the ganglion. Consequently, it is not simply the synaptic delay and is more accurately termed ganglionic delay. Estimates of the ganglionic delay in the prevertebral ganglia have varied from around 5 to 33 msec. Ganglionic delay in the IMG of the cat in situ was determined to be between 4.5 and 6 msec (Lloyd, 1937). A similar value was reported in a later study where the minimum ganglionic delay was calculated to be 5 msec (Job and Lundberg, 1952). Using extracellular recordings, a delay of 5,2 msec was calculated in the cat C G (Syromyatnikov and Skok, 1968). Using intracellular recordings from cells of the celiac-superior mesenteric ganglion of the guinea pig in vitro at 37°C a ganglionic delay of 9.3 msec has been reported (Kreulen and Szurszewski, 1979a). In the rabbit IMG in vitro a much longer delay of 33 msec was found at 20 to 22°C (Brown and Pascoe, 1952). These values are much longer than the more direct measurement of synaptic delay. In the rabbit SCG the delay from stimulation of the preganglionic terminal to the F - E P S P was only 1.5 to 2.0 msec (Christ and Nishi, 1971 ).

4.1.2. Pathways through the ganglion There is a complex network of conduction pathways through the prevertebral ganglia. Stimulation of the splanchnic nerves to the cat CG resulted in the appearance of a synaptic response in all of the other attached nerves except for the other splanchnic nerves. Similarly, stimulation of the nerves other than the splanchnics resulted in synaptic potentials in all the other nerves, including the splanchnics (Syromyatnikov and Skok, 1968). All of these responses, except for the response in the splanchnic nerve, were blocked by a cholinergic antagonist, indicating their synaptic nature. The response in the splanchnic nerve following stimulation of another nerve was not cholinergic; however, the delay suggested that the response was synaptic. These results show that the nerves of the C G contain both pre- and postganglionic fibers. In the cat IMG in situ, stimulation of the inferior splanchnic nerves resulted in synaptic responses in the colonic and hypogastric nerves (Lloyd, 1937). Following stimulation of the hypogastric nerve synaptic responses were recorded in both the same nerve and in the colonic nerve while an antidromic response was seen in the inferior splanchnic nerves (Job and Lundberg, 1952; Lloyd, 1937). In the rabbit IMG in vitro synaptic responses were recorded in the ascending mesenteric nerve following stimulation of the same nerve or following stimulation of the inferior splanchnic nerves (Brown and Pascoe, 1952). Extensive convergence of fibers from the lumbar splanchnic nerves and from the hypogastric nerve was demonstrated by Oscarsson (1955).

SYNAPTIC TRANSMISSION IN THE PREVERTEBRAL SYMPATHETIC GANGLIA

59

4.1.3. Peripheral reflex pathways Extracellular recordings have confirmed the existence of reflex pathways postulated as a result of the earlier physiological studies. Job and Lundberg (1952) examined the responses in the colonic and hypogastric nerves of the cat following hypogastric nerve stimulation. Degeneration after excision of the sympathetic ganglia from the ninth thoracic to the fifth lumbar level, section of the dorsal and ventral roots or section of the preganglionic nerves did not alter the responses; thus, the presynaptic fibers which carry the reflex do not originate central to the IMG. If the hypogastric nerve is sectioned and allowed to degenerate, stimulation and recording on the peripheral part of this nerve reveals an action potential. Consequently, the origin of these fibers must be located peripherally to the IMG. Similar findings have been obtained in other studies. McLennan and Pascoe (1954) examined the origin of the high-threshold, slow-conducting fibers observed in the ascending mesenteric nerve by Brown and Pascoe (1952). Degeneration following section of the inferior splanchnic fibers, vagal fibers or the sympathetic trunks did not affect these fibers. Variable effects were seen following removal of the C G - S M G . As noted above, some of the fibers in the ascending branches do intermingle with the C G - S M G while others run towards the colon; hence, removal of the C G - S M G would remove some of these fibers. Section of the ascending branches themselves was the only manipulation which resulted in a disappearance of the response; however, section of the ascending mesenteric nerve did not exclusively involve the slow preganglionic fibers but also involved postganglionic fibers. To solve this problem, the ascending branches were severed close to the ganglion and, after degeneration of the postganglionic fibers, the segment of the nerve rostral to the section was examined. The slow fibers remained; therefore, the cells of origin must lie in the rostral distribution of the ascending mesenteric nerves. These fibers definitely do not enter the abdomen in the vagi or thoracic splanchnics and are probably not collaterals of dorsal root afferents. As has been suggested for the cat IMG, the most likely explanation in the rabbit is that these are fibers of enteric origin. Recording the activity in the colonic nerves in vivo in anesthetized cats most commonly revealed irregularly grouped discharges (De Groat and Krier, 1979). Injection of ganglionic blocking agents blocked the activity while section of the inferior splanchnic nerves reduced the activity in the colonic nerves. The activity in the colonic nerves was found to depend on activity from the upper lumbar levels of the cord. This was illustrated by the finding that transection of the cervical (C2-C3) or thoracic (T10-T13) spinal cord did not alter colonic nerve activity, but procaine injected at L3 to L4 or bilateral section of the lumbar ventral roots markedly depressed colonic nerve activity. Furthermore, colonic nerve activity was positively correlated with intestinal motility. Procaine injection or ablation of the spinal cord at L1 to L5 decreased colonic nerve activity and thereby increased intraluminal pressure and unmasked or enhanced slow rhythmic waves in the colon. Subsequent section of the colonic nerves resulted in a further increase in basal pressure and of the intestinal contractions. This indicates that the isolated lumbar spinal cord as well as the decentralized I M G can generate tonic inhibitory input to the large intestine (De Groat and Krier, 1979). Following stimulation of the colonic nerves an outcoming asynchronous reflex activity could be recorded in the same nerves. The reflexes were depressed by hexamethonium. Section of the lumbar spinal dorsal roots or ablation of the cord from L1 to L5 markedly reduced, but did not abolish, the response to colonic nerve stimulation (De Groat and Krier, 1979). This indicates at least two possible reflexes involved in colonic nerve activity. First, a spinal reflex which depends on the integrity of the dorsal root afferents and second, a peripheral reflex through the IMG. 4.2. INTRACELLULARRECORDINGS Several studies have examined the electrical properties of single cells in the prevertebral ganglia. These studies also serve to illustrate the complexity of synaptic transmission

M. A, SIMMONS

60

TABLE 3. ELECTRICALPROPERTIESOF PREVERTEBRAI.GANGLION CELLS Guineapig Guineapig Guineapig CG r I M G 2-~ p G 4.s Resting m e m b r a n e potential

-49

Cat C G ~,

Cat 1MG 7.

Dog I M G ~'

Rabbit I M G ~"

-52

-52

--.49

-40

-52

--411Io-711

4(1 to 150

40 to 124 10

22 13

46 to 78 4 to 7

10to20

16

10

5to18 t).5

Gumeapig S('G I 45to

q(5

(mV) Input resistance (MI'I) M e m b r a n e time constant (msec) Threshold depolarization ( m V ) Rheobase (nA) C h r o n a x i e (msec) Action potential amplitude (mV) Action potential afterhype rpolarization amplitude (mY)

26 7

0.3 10{I

76

15

22

7

37 8

~i .L 4

12 0.2 13 65

17 ~. 11 ~4

II

' Kreulen and SzurszewskL 1979a; -"Crowcroft and Szurszewski, 1971 ; 3 Szurszewski and W e e m s . 1976; a B l a c k m a n et al., 1969: s H o l m a n et al., 1971; " D e e k t o r and W e e m s , 1983; 7 K r i e r et al., 1982: x Jul~ and Szurszewski, 1983; 9 King and Szurszewski, 1984a; u, S i m m o n s and D u n . 1985: ' i Perri et al., 1971).

through these ganglia with respect to multiple input pathways. Additionally, the use of an isolated prevertebral ganglia-colon preparation has allowed direct comparison of intracellular responses to the activity of the colon. Most recently, intracellular recordings have revealed that peptides exert specific effects on these cells and may play a role in synaptic transmission in this ganglion. 4.2.1. Basic properties of ganglion cells Reported values for some basic membrane electrical properties of the prevertebral ganglion neurons are listed in Table 3. These studies have been conducted in vitro. The first intracellular studies on this ganglion used the guinea pig IMG (Crowcroft and Szurszewski. 1971). Three types of cells were found. Most were of the first type, which responded to direct intracellular or presynaptic stimulation with an action potential. The maximum E, in these cells was -65 mV. The second type of cell had a more negative E,, ranging from -75 to -85 mV and required larger depolarizations to initiate action potentials. The third type of cell also had a high Er, but was inexcitable; these were presumably glial cells. There is evidence that the Type II cells may have actually been damaged Type I cells. Neild (1978) observed that during the first minute of penetration, some Type I cells spontaneously hyperpolarized to become Type II cells. These cells would then gradually depolarize over a period of 5 to 10 min and again become Type I, although the impalement was usually lost before complete recovery. A similar phenomenon was seen in the rabbit IMG (personal observation). Most subsequent studies have described only one type of excitable cell with Er around - 5 0 mV (Szurszewski and Weems, 1976; Krier et al.. 1982). With hyperpolarizing currents, some cells of the guinea pig and rabbit IMG exhibited linear current-voltage relationships, while other cells rectified beyond 15 to 20 mV more negative than Er (Crowcroft and Szurszewski, 1971; Simmons and Dun. 1985). With depolarizing pulses, delayed rectification was evident. All pelvic ganglion cells innervated by the hypogastric nerve showed anomalous rectification for hyperpolarizations > 10 mV negative to E, (Blackman et al., 1969). Those innervated by the pelvic nerve, or the pelvic nerve and the hypogastric nerve, did not rectify (Crowcroft and Szurszewski. 1971). The reason for this difference remains unexplained. In the rabbit IMG the orthodromic spike and antidromic spike did not differ in amplitude (Simmons and Dun, 1985). In the pelvic ganglia the orthodromic spike was found to be smaller than the antidromic response (Blackman et al., 1969). An explanation for the orthodromic spike being smaller than the direct spike is that during an orthodromic spike the synaptic membrane is short-circuited by ACh (Fatt and Katz. 1951). In the IMG there are few axosomatic synapses (Elfvin, 1971a) and a decrease in resistance out on the dendritic membrane would provide little shunting effect on the soma spike. In the PG.

SYNAFHC TRANSMISSION IN THE PREVERTEBRAL SYMPATHETIC GANGLIA

61

however, there are mainly axosomatic synapses so that the soma spike is shunted through the low resistance of these synaptic areas of membrane. Antidromic responses are rare in the IMG. Approximately 20% of the cells in the guinea pig IMG showed an antidromic response (Crowcroft and Szurszewski, 1971) and only 8% of the cells in the rabbit IMG exhibited an antidromic response (Simmons and Dun, 1985). Antidromic spikes were not observed in the dog IMG (King and Szurszewski, 1984a). 30 to 34% of the neurons in the CG could be antidromicaUy activated (Kreulen and Szurszewski, 1979b; Decktor and Weems, 1983). The reasons for the infrequent occurrence of antidromic spikes is not clear. It is possible that the axons of these cells have electrical properties which make antidromic propagation difficult. Branch points in axons would be one possibility; however, it has not been established whether the postganglionic fibers give off collaterals prior to the terminal arborizations at the effector organ. Another point would be the sudden enlargement at the axon-soma junction. It is also likely that the nerve containing the axon of the cell being recorded from would not be stimulated in a given experiment. The IMG has several nerve trunks associated with it and stimulation of every fiber in a single experiment would be unusual and practically impossible. 4.2.2. Synaptic inputs to ganglion cells 4.2.2.1. Celiac-superior mesenteric ganglia Maximal electrical stimulation of any of the nerves of the CG resulted in multiple F-EPSPs and action potentials (Kreulen and Szurszewski, 1979b). The F-EPSP peaked in 3 to 6 msec, decayed exponentially with a time constant of 18 msec and was blocked by d-TC or DH/3E. A topographical organization of inputs was apparent here. The number and intensity of inputs was related to the cell's position in the ganglion. Cells near the entrance of a nerve to the ganglion responded maximally to stimulation of that nerve while only showing sub-threshold F-EPSPs to stimulation of the other nerves. A similar organization was apparent in the SMG. 4.2.2.2. Inferior mesenteric ganglion In contrast to the CG-SMG, most cells in the cat, dog, guinea pig and rabbit IMG receive synaptic input from all of the nerve trunks entering the ganglion (Crowcroft and Szurszewski, 1971; Simmons and Dun, 1985; Jul6 et al., 1983; King and Szurszewski, 1984a). The F-EPSP in the rabbit IMG peaked, on average, in 7 msec and lasted 42 msec, the time constant of decay was 18 msec, identical to that in the guinea pig CG (Simmons and Dun, 1985). This is twice as long as the membrane time constant in the rabbit IMG and has been taken as evidence that the transmitter action outlasts the peak of the F-EPSP (Coombs et al., 1956). The F-EPSP in these cells was abolished in low Ca2+/high Mg 2÷ solution and by d-TC (50/zM). Examination of the increase in the size and number of F-EPSPs with increasing stimulus intensity revealed that each neuron in the guinea pig IMG receives at least ten presynaptic inputs from each of the colonic and ascending mesenteric nerves, three to five inputs via each hypogastric nerve and one to three fibers from each splanchnic nerve; therefore, the total number of inputs to a single cell in this ganglion would be approximately 40 (Crowcroft and Szurszewski, 1971). In the rabbit IMG the cells received, on average, 12 inputs from the aortic branch of the ascending mesenteric nerve, 19 from the ascending mesenteric nerve and 5 from the hypogastric nerve. This gives an average of 46 inputs (Simmons and Dun, 1985). Cells of the cat IMG receive synaptic input from two or three different lumbar sympathetic rami, most often from adjacent segments (Jul6 et al., 1983; Krier et al., 1982). The number of inputs from a single ramus to one cell ranged from 1 to 13 with a mean of 5 per cell (Krier et al., 1982). Some fibers from the lumbar spinal cord synapse in the paravertebral sympathetic ganglia and then continue to the prevertebral ganglia (Hartman and Krier, 1984). These results confirmed the findings of extracellular recordings that all of the nerves associated with the IMG carry preganglionic fibers.

M. ; \ . SIMMO~ns

f~2

B

A

r'rrrr q"-r!:li i

illlilillilllll!l]lllil]ll~"

UTT~TlitIIIIiI-'~TFIJjiiiiilEIiI~

60 sec

Jl0 mV

0.SnA

FIG. 4. S-EPSP in the rabbit IMG. (A) Following stimulation of the intermesenteric nerve ( 16 Hz for 3 sec) a slow depolarization ensues following the F-EPSPs (initial vertical tracings). (B) Thc depolarization is abolished 4 rain after addition of atropine ( 1 /xM) to the perfusatc. From Simmons and Dun (1985).

The number of inputs to ganglion cells may be correlated with the degree of dendritic branching of the cells. In the rabbit ciliary ganglion it has been found that cells which lack dendrites are generally innervated by a single axon, whereas neurons with increasing numbers of dendrites are innervated by a proportionally greater number of axons (Purves and H u m e , 1981). This is consistent with the morphological studies showing dendritic branching of the IMG neurons. The marked convergence of F - E P S P inputs distinguishes the IMG from other sympathetic ganglia. For example, in the SCG an average of ten inputs per cell was calculated (Nja and Purves, 1977). The calculation of inputs to the IMG cells may be subject to several errors, however. First, some preganglionic fibers traverse the ganglion en route to cells in the PG. These fibers would be present in two of the nerve trunks of the IMG and, during a study of the inputs, would be stimulated both orthodromically and antidromicalty, resulting in that particular input being counted twice. Other factors may result in the count being underestimated. It is possible that in a given experiment, not all of the nerve fibers in that particular nerve would be responsive to electrical stimulation. Also, some fibers enter the IMG separately from the major nerve trunks and would not have been studied. In any case, the degree of convergence of presynaptic inputs to cells in the prevertebral ganglia is clearly significant. As noted above, most of the inputs to C G and IMG cells are sub-threshold. In order for an impulse to be relayed through these ganglia, several inputs must sum to generate an action potential in the ganglion cell. This indicates that these ganglia do not serve as simple relay stations from central nervous system to peripheral viscera, but are involved in the integration of nervous activity. On the basis of results obtained in the rabbit SCG, it was calculated that each preganglionic fiber must innervate 240 postganglionic cells (Wallis and North, 1978). This has been taken as a further illustration of the diffuse control exerted by the sympathetic nervous system on the viscera. In order to determine the number of cells innervated by each preganglionic axon the ratio of preganglionic to postganglionic fibers must be known. As discussed in Section 2.3.4, a value as low as 1 : 1.5 has been suggested in the cat IMG (Harris, 1943). This study only counted fibers of lumbar spinal origin. We now know that the IMG also receives preganglionic cholinergic inputs of peripheral origin and preganglionics probably also descend to the IMG via the intermesenteric nerve. Interspecies differences in the preganglionic: postganglionic ratio have been shown to be related more to differences in the number of postganglionic cells than to differences in the number of preganglionic fibers (Ebbeson, 1963). Since the rabbit IMG is smaller than the cat IMG. and assuming size is related to cell number, one would expect a higher ratio in the rabbit

63

SYNAPT1C TRANSMISSION IN THE PREVERTEBRAL SYMPATHETIC GANGLIA

o

I

I 2 sec

FIG. 5. Electrical activity in a guinea pig IMG cell. Colon and IMG were connected by the lumbar colonic nerves. The dot designates the cutting of the colonicnerves. From Crowcroftetal. (1971a). With permission of the authors and the PhysiologicalSociety. IMG and an even higher ratio in the guinea pig IMG. Considering these factors a ratio near 1 : 1 does not seem unreasonable in the IMG. If we accept this ratio then each preganglionic fiber would innervate approximately 40 postganglionic cells, a number considerably smaller than the 240 obtained in the SCG. In addition to the F - E P S P , a S - E P S P was also observed in some cells of the rabbit IMG (Fig. 4) (Simmons and Dun, 1985). This response was evident following repetitive presynaptic stimulation and disappeared following perfusion with atropine (1/xu). Only 7% of the cells exhibited an atropine-sensitive S - E P S P in the absence of a noncholinergic depolarization. The atropine-sensitive depolarization is apparently comparable to the muscarinic cholinergic S - E P S P previously recorded from other sympathetic ganglion cells (Libet, 1970). In many cells, both cholinergic and noncholinergic depolarizations were observed (see below). 4.2.2.3. Pelvic ganglia 60% of the cells in the pelvic plexus exhibited an orthodromic response following stimulation of the hypogastric nerve (Blackman etal., 1969). This input was usually a single "all-or-nothing" synaptic input, but sometimes up to five inputs were revealed (Crowcroft and Szurszewski, 1971). This is consistent with the view that the number of presynaptic inputs correlates with cell processes, as these cells have few processes. No S - E P S P was observed in the pelvic ganglia of the guinea pig (Holman et al., 1971). 4.2.3. Prevertebral ganglia-colon preparations The hypothesis that the IMG receives input from neurons in the colon has been confirmed using an in vitro IMG-colon preparation (Crowcroft et al., 1971a). This preparation consists of the IMG attached to the colon by the colonic nerves. The preparation is placed in a two-compartment organ bath, the IMG in one compartment and the colon in the other. In this manner the input to I M G neurons can be examined by intracellular recording and drugs can be applied selectively to the IMG or to the colon. With this preparation, continuous electrical activity consisting of action potentials or F - E P S P s was recorded from neurons in all regions of the IMG (Fig. 5). This spontaneous activity was indistinguishable from that elicited by sub-maximal stimulation of the nerve trunks connected to the IMG. When the colonic nerves were cut, an immediate and irreversible abolition of the spontaneous activity occurred (Fig. 5). When the colonic nerves were left intact, the spontaneous activity was reversibly abolished by application of T F X to the colon. At this time, stimulation of the other nerve trunks to the IMG was still effective in eliciting a synaptic response; thus, the spontaneous activity must have been initiated in the colon. Application of the cholinergic nicotinic antagonist DH/3E to the IMG blocked both spontaneous and evoked activity indicating that spontaneous activity in

64

M . A . SlmMONS

the IMG was mediated by the activation of nicotinic receptors. DH/3E applied to the colon depressed, but did not completely abolish, the spontaneous activity recorded from the IMG cells while not affecting nerve-evoked responses. Part of the spontaneous activity from the colon, therefore, must result from cholinergic activation in the colon. The activity in the IMG cells is related to intestinal intraluminal pressure. The input to the IMG from the colon in vitro can be activated by raising the intraluminal pressure of the colon above 5 cm H20; the input to the IMG increased as pressure increased above this level (Weems and Szurszewski, 1977). At least part of this input, therefore, seems to be dependent on mechanoreceptors in the gut. Mechanoreceptors have been classified as slowly adapting or rapidly adapting and both types are found in the gastrointestinal tract. Since a continuous increase in intraluminal pressure did not result in an adaptation of the afferent input it was presumed to be from the slowly adapting type of mechanoreceptor. Agents which have a relaxant effect on the intestine also resulted in a decrease in the spontaneous activity recorded from IMG cells. Papaverine, isoproterenol, adenosine triphosphate and atropine all exhibited this effect. Carbachol, ACh and 5-HT. which excite the colon, increased the synaptic activity. The effects of carbachol and ACh were blocked by atropine applied to the colon (Crowcroft etal., 1971a; Szurszewski and Weems, 1976). A negative feedback circuit was found between the IMG and the colon. Stimulation of the nerve trunks of the IMG not only resulted in an inhibition of colonic motility, but also an inhibition of the spontaneous activity from the colon. Following stimulation of the colonic nerves at 20 Hz for 1 sec the inhibition lasted 1.5 sec. Increasing the frequency or duration of the nerve stimulation resulted in a longer period of inhibition of the spontaneous activity. This inhibition was mimicked by application of N E to the colon but only a slight inhibitory effect was seen when N E was added to the IMG. Both the synaptic and the NE inhibition of the spontaneous activity were antagonized by addition of the adrenergic antagonists phentolamine and phenoxybenzamine to the colon (Crowcroft et al., 1971a). In reserpinized animals repetitive stimulation of the nerve trunks failed to inhibit the spontaneous synaptic activity originating in the colon, further indicating its adrenergic nature (Szurszewski and Weems, 1976). King and Szurszewski (1984b) have obtained evidence that the cholinergic mechanoreceptor neurons from the gut project to the IMG, and no further centripetally. They dissected the spinal cord. D R G , paravertebral ganglia, prevertebral ganglia and colon, with all associated nerve trunks, for in vitro electrophysiology. Their evidence is as follows: (1) While stimulation of the nerve trunks of the IMG or distention of the colon did result in cholinergic excitation in the IMG, stimulation of the dorsal roots from T13-L4 did not, indicating that the cholinergic excitatory fibers from the colon do not continue more centrally. Tsunoo et al. (1982) had also found that dorsal root stimulation did not elicit F-EPSPs in the IMG cells. (2) When recordings were made from D R G cells, stimulation of the nerve trunks of the IMG or distention of the colon failed to produce action potentials in the D R G cells. (3) Extracellular recordings from the dorsal or ventral roots detected no discharges following stimulation of the nerve trunks to the IMG or following colonic distention. Discordantly, in the guinea pig IMG it has been suggested that SP-containing sensory neurons with cell bodies in the D R G project through the IMG to the colon (Section 5.3). King and Szurszewski (1984b) suggest that perhaps the number of such SP-containing cells was too small to be detected by random sampling of D R G cells or by extracellular recordings from the nerve roots. In this regard, the tracers H R P or True Blue labeled only a few % of the D R G cells when applied to the IMG or the colonic nerves (Dalsgaard et al., 1982a; Dalsgaard and Elfvin, 1979; King and Szurszewski, 1984b). A preparation consisting of the CG and SMG attached to the colon has also been used (Kreulen and Szurszewski, 1979b). The colon was cut so that oral and aboral segments could be distended separately. 68% of the neurons in the C G exhibited F-EPSPs when the oral section was distended and 37% showed F-EPSPs when the aboral section was distended. 13% responded to distention of both segments (Kreulen and Szurszewski, 1979c). In the SMG, 57% of the cells responded to oral distention while 43% responded to

SYNAPTICTRANSMISSIONIN THE PREVERTEBRALSYMPATHETICGANGLIA

65

aboral distention. This shows a viserotopic organization of inputs from colon to CG-SMG. This correlates with the findings that the inputs to the cells in the CG-SMG are topographically organized (Section 4.2.2.1). In a similar vein, nerve trunks from the different prevertebral ganglia affect different portions of the intestine (Kreulen and Szurszewski, 1979c). Stimulation of the nerves from the CG caused a 38% reduction in pressure in the oral segment but no measurable decrease in the caudal segment. Such stimulation also eliminated propulsive contractions. Stimulation of the intermesenteric nerve inhibited oral pressure 30% and caudal pressure 19% with a greater effect on contractions in the caudal segment. Stimulation of the colonic nerves, from the IMG to the colon, inhibited oral pressure 15 %, caudal pressure 22% and all contractions in the caudal segment. In this manner, neurons in certain prevertebral ganglia receive mechanoreceptor inputs from specific areas of the colon and in return send axons to act on that section of the colon. To test for inhibitory reflexes through the prevertebral ganglia, one segment was distended and the pressure in the other segment monitored. Distention of one segment resulted in inhibition of the other segment. This reflex worked in both directions and ceased when the nerves between one of the segments and the prevertebral ganglia were disconnected. With the colon attached, continuous spontaneous F-EPSPs were observed in 32% of the cells of the right celiac ganglion, in 34% of the left CG and in 54% of the SMG. These EPSPs were blocked by addition of d-TC to the ganglion. This activity was increased by colonic distention. Section of the nerves attaching the ganglia to the colon was used to determine the area of the colon from which the ganglion cell received mechanosensory information. The input to the cells in the oral regions generally was abolished following section of the inferior celiac nerve while the responses in the cells in the SMG ceased following lumbar colonic or ascending mesenteric nerve section. Taken with the above findings that the input to cells in the CG and SMG depended on the position of the cell in the ganglion, these results reveal a topographic organization with the oral colon sending information to the CG and the aboral colon sending information to the SMG and IMG. These results provide direct evidence for a peripheral reflex arc involving the IMG and the distal colon. The noradrenergic cells of the IMG act to depress the activity of excitatory neurons in the colon. In turn, these neurons send fibers t o t h e IMG in a feedback loop similar to recurrent inhibition. Since blockade of the nicotinic receptors in the IMG suppresses the spontaneous activity, the fibers from the colon must release ACh. Similarly, DHflE was able to suppress, but not completely, the spontaneous activity when applied to the colon; thus, some, but not all, of the enteric cholinergic neurons must themselves be driven by cholinergic inputs.

5. Noncholinergic Transmission in Sympathetic Ganglia Excitatory transmission in sympathetic ganglia not blocked by cholinergic antagonists was first identified in bullfrog paravertebral ganglia. Henceforth, slow noncholinergic depolarizations have been identified in mammalian prevertebral ganglia. 5.1. BULLFROG SYMPATHETIC GANGLIA

5.1.1. Description of slow potentials With extracellular electrodes on the postganglionic fibers of the ninth or tenth ganglion of the bullfrog sympathetic chain, Nishi and Koketsu (1968) observed EAD and LAD following repetitive stimulation of the preganglionic fibers. The EAD appeared during stimulation, reached a maximum immediately following cessation of stimulation and diminished in about 30 sec (at 25°C). This response was blocked by atropine. In contrast, the LAD occurred immediately or shortly after the cessation of stimulation, reached a maximum in about 30 sec and lasted for more than 2 min; furthermore, this response was insensitive to cholinergic antagonists. Using sucrose gap recordings, potential changes of

6¢~

X! \

~lm~qons

the whole ganglion were measured which correspond in time to the E A D and LAD, these are the LN and LLN potentials, respectively. Direct evidence that the LN and LLN waves are postsynaptic potentials was obtained by intracellular recordings which revealed a long-lasting depolarization of the ganglion cells, the amplitude of which was dependent on the frequency and duration of preganglionic stimulation. Atropine depressed the initial phase of this slow depolarization while enhancing the late phase. Hence, the early component of the low depolarization is the S - E P S P (Libet and Tosaka, 1966); it relates to the E A D and LN wave. The latter component of the depolarization, which corresponds to the L A D and LLN wave, is not blocked by cholinergic antagonists and is referred to as the LS-EPSP. A L S - E P S P is observed in 90% of the B cells and 75% of the C cells in the bullfrog ninth and tenth lumbar ganglia (Jan and Jan, 1982). The optimal frequency for the L S - E P S P is 5 to 10 Hz; with 20 to 5(1 stimuli, the LS-EPSP amplitude is 5 to 10 mV. In 80% of the cells, the L S - E P S P is accompanied by an increase in R m.

5.1.2. The transmitter of the LS-EPSP in the bullfrog The transmitter mediating the L S - E P S P in the bullfrog has been recently identified to be a LHRH-like peptide (Jan et al., 1979. 1980; Jan and Jan, 1982). Several findings have led to this conclusion. (1) An LHRH-Iike substance is found in the bullfrog sympathetic chain by radioimmunoassay. This substance is not identical to mammalian L H R H . (2) After section of the preganglionic nerves, 95% of this LHRH-Iike peptide disappears from the ganglion while the content triples in the spinal nerves proximal to the cut; thus, the LHRH-Iike peptide is contained in the preganglionic fibers. (3) L H R H - i m m u n o r e a c t i v e fibers have been identified surrounding the cell bodies of the ganglionic neurons. (4) An L H R H - l i k e peptide can be released in a Ca2+-dependent manner by raising K + concentration or by stimulation of the preganglionic nerves. (5) When applied to the ganglion cells by pressure ejection, L H R H causes a slow depolarization. (a) This depolarization is accompanied by an increased Rin, as is the LS-EPSP. (b) When Ca x* in the Ringer solution is replaced by M f +, in order to prevent transmitter release, the response following presynaptic stimulation is blocked. The response to applied L H R H is unchanged under these conditions, indicating a direct action of L H R H on the postsynaptic membrane. (6) The effects of various analogs of L H R H in the bullfrog ganglia are the same as the effects of these analogs on the release of luteinizing hormone in rats. Analog antagonists block the L S - E P S P and the response to applied L H R H while analog agonists mimic the LS-EPSP. None of these compounds affected cholinergic transmission. Another peptide, substance SP, also depolarizes and increases R~,, of neurons in the bullfrog ganglia by a direct postsynaptic action (Jan and Jan, 1982). SP does not appear to be involved in the generation of the L S - E P S P since no cross-desensitization is observed between SP and the L S - E P S P in the bullfrog and L H R H analogs have no effect on the SP responses. Also, L H R H is still effective after desensitization of the cells to SP; therefore, L H R H and SP must act at different sites, Although SP immunoreactivity is contained in bundles of axons passing through the ganglion, no SP-positive synaptic boutons are seen and the distribution of SP fibers is distinct from that of the L H R H fibers which surround C cells. 5.1.3. Voltage clamp studies of the LS-EPSP Voltage clamp experiments have been conducted on the slow potentials in bullfrog sympathetic neurons to examine the ionic currents affected by L H R H and nerve stimulation. Bullfrog sympathetic neurons are relatively large and have few processes, making

SYNAPTIC TRANSMISSION IN THE PREVERTEBRAL SYMPATHETIC GANGLIA

67

them suitable for two-electrode voltage clamp studies. Adams and Brown (1980) reported that at Em = - 3 0 mV, L H R H induced a steady inward current and decreased G,,. At Em = - 6 0 mV, there was no change in current but there was a decrease in G m . This current has also been shown to be reduced by muscarinic agonists, hence it has been termed the M current, IM (Brown and Adams, 1980). The M current consists of a time- and voltagedependent outward K + current which is activated in the potential range of - 6 0 to - 10 mV and is inactivated at more negative potentials and in the presence of muscarinic agonists or L H R H . Consequently, the reduction of this current results in a net inward current and a decreased Gm resulting in a depolarization of the cell. An increase in R m is not observed in all cells, however, indicating that suppression of the M current alone may not be sufficient to account for the mechanism of the LS-EPSP. Following presynaptic stimulation or L H R H application only 80% of the cells exhibit an increased R m (Jan and Jan, 1982; Katayama and Nishi, 1982). Also, under voltage clamp, one would expect the response to decrease at holding potentials more negative than - 6 0 mV since beyond this potential the M current should be inactivated and any potential due strictly to a change in GK would disappear at EK (approximately --80 mV). However, this does not occur in all cells. Furthermore, in some cells the early portion, but not the late portion, of the response appeared to reverse at EK (Jan and Jan, 1982; Kuffler and Sejnowski, 1983). The explanation offered is that changes occur in more than one conductance, each with a characteristic time course and voltage sensitivity, and that at least one of these conductances is not selective only for potassium. Katayama and Nishi (1982) have divided the membrane currents during the L S - E P S P (excitatory postsynaptic currents--EPSCs) into two types. The Type I EPSC is associated with a reduction of membrane conductance. With hyperpolarization this current is reduced. The Type II EPSC involves an increase in membrane conductance at Er with a marked enhancement of both the EPSC and the conductance change at hyperpolarized levels. Many cells exhibited both Type I and Type II responses. At potentials more positive than - 6 0 mV a typical Type I response was observed while hyperpolarization to levels more negative than - 6 0 mV resulted in a change of the properties of the reponse to Type II. The Type I response appears similar to that of the muscarinic depolarization generated by suppression of IM. However, even in these cells a small L H R H - i n d u c e d response was evident at levels of hyperpolarization when the M channel should be closed. It is concluded that the Type I response also involved a decrease in the resting GK. The Type II response was abolished in Na+-free solution indicating the involvement of a GNa in this response. An interesting feature of this response was that steady hyperpolarization enhanced the response while 1 sec hyperpolarizing pulses did not, again indicating the involvement of unique time- and voltage-dependent channels. While all of the above studies have concluded that suppression of IM alone cannot account for the diversity of responses involved, Adams et al. (1982) have concluded that selective inhibition of IM is sufficient to explain the actions of muscarine and L H R H . This is based on the premise that the results of other workers are inconsistent with inactivation of IM only at hyperpolarized levels and even then the observed responses are small. Hence, at the normal E, these other currents would make little contribution to the depolarization. As to its role in the modulation of synaptic transmission, IM is regarded as a "braking" current which, when activated with depolarization, controls excitability and limits repetitive spiking. The removal of this brake by a released transmitter substance therefore results in increased excitability and represents the physiologically significant aspect of IM suppression. 5.2. GUINEA PIG CELIAC GANGLION

5.2.1. Characteristics of the LS-EPSP A noncholinergic L S - E P S P has recently been recorded from cells in the guinea pig celiac ganglion (Dun and Ma, 1984). This depolarization was unaffected by d-TC or atropine, but was blocked by a low Cae+/high Mg 2+ solution. Following repetitive nerve stimulation

68

M.A. SIMMONS

A CON

i

Hill[

d-Tc .t-

ATRO

bl l J ]l itiJ [ II1111111l',lililtiillfllllllllllllllllllllll Illll:llelllillillllllllltlti'i .....................

1111tlt ;t111111

LOW Co

WASH

IIMiIITHIIIIIIIIh,IlltlI;II,Illltllllltl'lllllrllliilllllll mt ll+liiitii ~L

"

b

iiii+ '

I

t

~

J

ilillilllllititiilliillillllliiiiiMllllltltll

B

FIG.6. LS-EPSP in the guinea pig CG. (A) Con, LS--EPSPfollowingstimulationof the left greater splanchnic nerve (20 Hz for 2 sec); d-TC and atro,qn the presence of d-TC (50/,L~) and atropine (1 tiM) the F-EPSPs were suppressed but the LS-EPSP remained; Low Ca, low Ca2+(0.25 mM)/ high Mg2+(12 mM)attenuated the LS-EPSP; Wash, responses returned after perfusion with normal Krebs solution. (B) Biphasie LS--EPSPin another CG cell. Horizontal calibration, 60 sec. Vertical calibration, l0 inV. From Dun and Ma (1984). With permissionof the authors and the Physiological Society. (10-20 Hz, 1-2 sec), 70% of the cells in this ganglion exhibited a L S - E P S P . Two types of response were observed in the C G neurons. In most cells, the depolarization was monophasic, but in 10% the response consisted of two distinct peaks. Typical responses are shown in Fig. 6. T h e r e was considerable variability in the responses with durations ranging from 40 sec to 15 rain (mean duration = 160 sec) and amplitudes from 2 to 24 m V (mean about 5 mV). As in the rabbit I M G (see Fig. 16), the amplitude of the response increased with increasing numbers of preganglionic stimuli. An increase (mean = 2 9 % ) in Ri, was observed during the L S - E P S P in 83% of the C G neurons. When the m e m b r a n e was clamped at Er an increase of similar magnitude was observed in 96% of the cells. M e m b r a n e depolarization and m e m b r a n e hyperpolarization decreased or increased, respectively, the amplitude of the L S - E P S P , giving an extrapolated reversal potential of - 3 7 mV. A L S - E P S P could be elicited in some cells following stimulation of any of the four nerve trunks tested. O t h e r cells responded after stimulation of two or three of the

SYNAPTICTRANSMISSIONIN THE PREVERTEBRALSYMPATHETICGANGLIA

69

nerves. As observed during the LS-EPSP in bullfrog ganglia, membrane excitability was increased during the LS-EPSP, even in the absence of membrane potential changes.

5.2.2. Similarity of LS-EPSP and 5-HT depolarization Pharmacological analysis of this response has led to the suggestion that, in some of the celiac ganglion cells, the slow depolarization may be mediated by 5-HT (Dun et al., 1984). Of the cells which exhibited a LS--EPSP, 63% were also depolarized by 5-HT. The 5-HT depolarization was not affected by d-TC, atropine, or low Ca2+/high Mg 2÷, indicating that 5-HT was not acting presynaptically to cause the release of ACh or some other substance. Further study of these cells revealed that the membrane resistance change during the 5-HT-induced depolarization and during the nerve evoked response was always the same in a given cell. Similarly, when Em was altered, the change in amplitude of the LS-EPSP and 5-HT depolarization were in the same direction. Perfusion of the ganglion with 5-HT resulted in an apparent desensitization of the LS-EPSP. 5.2.3. Pharmacological manipulation of serotonergic system Two 5-HT antagonists were tested on these cells. While methysergide depressed the F-EPSP more effectively than the LS-EPSP, cyproheptadine suppressed both the LSEPSP and the 5-HT-induced response with less effect on the F-EPSP. In five cells which were not depolarized by 5-HT, cyproheptadine did not affect the LS--EPSP but still depressed the F-EPSP. Fluoxetine, a 5-HT uptake inhibitor, increased the amplitude of the LS-EPSP by 90% and the 5-HT response by 116% in the 5-HT-responsive cells. The 5-HT precursor, L-tryptophan, exerted a long lasting facilitatory effect on the LS-EPSP, but did not enhance the action of 5-HT. Further support for an involvement of 5-HT in this response was obtained from immunohistochemical studies which showed the presence of 5-HT immunoreactive fibers in this ganglion. These results led these authors to suggest that 5-HT may be involved in the LS-EPSP in a proportion of celiac ganglion neurons (Dun et al., 1984). However, 40% of these cells were unresponsive to 5-HT and unaffected by cyproheptadine, fluoxetine or tryptophan. This may indicate that the LS--EPSP in these cells involves a transmitter other than 5-HT. 5.3. GUINEA PIG INFERIOR MESENTERIC GANGLION

5.3.1. Characteristics of the LS-EPSP A mammalian counterpart of the LS-EPSP observed in amphibian ganglia was first reported in the guinea pig IMG (Neild, 1978). Stimulation of the hypogastric nerves at 20 to 30 Hz for 1 to 5 sec resulted in a depolarization of 2 to 20 mV (Neild, 1978; Dun and Jiang, 1982), the average value being about 4 mV (Dun and Jiang, 1982). The depolarization usually appeared 0.5 to 5 sec after the stimulus burst, reaching a peak in 10 to 25 sec (Neild, 1978; Tsunoo et al., 1982). The half decay time of the response has been reported to be 51 sec (Tsunoo et al., 1982), although the mean total duration has been found to be only 54 sec (Dun and Jiang, 1982) or 90 sec (Neild, 1978). The time course ranges from 20 sec to 4 min following hypogastric nerve stimulation at 10 to 30 Hz for 1 to 5 sec. A slow depolarization was observed in 70 to 80% of the cells in the guinea pig IMG (Neild, 1978; Tsunoo et al., 1982). This response is very similar to the LS-EPSP observed in bullfrog sympathetic ganglia and will be referred to also as the LS--EPSP. There is a great diversity in the types of membrane resistance changes which have been reported to accompany the LS-EPSP in the guinea pig IMG. In the first study on this slow depolarization (Neild, 1978), a decrease of R~, was observed in 96% of the cells. A decrease of 15% was observed for responses <7 mV in amplitude and of 26% when the response was >7 mV. This seems to be consistent with a rectification of the membrane with depolarization. The author claims, however, that while a depolarization of the cell by passing current through the electrode did result in a decreased Rin, i.e. rectification, the decrease was never as great as that seen during the synaptically evoked depolarization. In the remaining cells

~,I A , Slmmo~,s

?{i

I, I

11 II I

l

11 ,

. . . . .

~

,

,,T

I1,'

T ....

A

........ "-~-'~-'-

...... F" :

' " t " : . . . . . ~*' li!!l '311;": T : ' ~

~

~

]

w

]

~

B

-

i

FIG. 7. L S - E P S P s and accompanying Rin changes in two different guinea pig IMG cells, (A) and (B). T o p tracing, current applied. Bottom tracing, m e m b r a n e potential. First. a L S - E P S P was elicited in each cell. T h e n the m e m b r a n e depolarization was nullified by the application of hyperpolarizing current. In (A) an initial decrease in Ri, was followed by an increase. In (B) only an increase was observed. Horizontal calibration, 40 sec. Vertical calibration. 10 mV/5 n A . From D u n and Jiang (1982). With permission of the authors and the Physiological Society.

(4%) an increase in R~, was observed. The membrane potential was not clamped in a study which found that, for a small depolarization, an increase in Rm was usually observed (Konishi et al., 1979b), but if the depolarization exceeded 15 mV Rm usually decreased. This study concluded that the observed decrease was probably due to delayed rectification of the membrane. Unfortunately, when the cell was held at Er during the response, so that resistance changes could be examined in the absence of E m changes, a variety of results were obtained and the nature of the resistance change differs according to different investigators. Dun and Jiang (1982) found three types of resistance change: (1) an increase of 21% beginning 20 to 30 sec following stimulation was seen in 45% of the cells; (2) a small (16%) decrease followed by a sustained increase of about 25%, was seen in 32% of the cells and (3) a monophasic increase averaging 22% occurred in the remainder (Fig. 7). Tsunoo et al. (1982) reported only two types of changes when the cells were held at Er: an increase of 30% in two-thirds of the cells and no change in the remainder. The relationship between Em and the L S - E P S P also varies. In some cells, membrane hyperpolarization resulted in an increased L S - E P S P amplitude (Dun and Jiang, 1982). In other cells the amplitude decreased at moderately hyperpolarized levels while further hyperpolarization resulted in a return of the LS-EPSP. In conclusion, it appears that the ionic mechanism of the L S - E P S P is complex, involving changes in membrane conductance to more than one ion species. The amplitude of the slow depolarization is positively correlated with the frequency and duration of presynaptic stimulation, Neild (1978) found that with a 2 sec duration the lowest effective frequency was 10 Hz, but if a 1 Hz stimulus is continued for several sec a small, but detectable, depolarization can be evoked (Dun and Jiang, 1982). In most cells,

SYNAPTIC TRANSMISSION IN THE PREVERTEBRAL SYMPATHETIC GANGLIA

71

the maximum depolarization is elicited at 20 to 30 Hz for 2 to 5 sec while at higher frequencies the depolarization becomes progressively smaller (Dun and Jiang, 1982). The slow depolarization was not affected by d-TC, atropine or guanethidine, indicating that it was not mediated by cholinergic or adrenergic receptors, but was reversibly abolished by the replacement of external Ca 2+ with Mg 2+ (Neild, 1978), indicating that it was probably a result of the action of a synaptically released transmitter. Also, the occurrence of a LS-EPSP in the guinea pig IMG did not follow stimulation of specific nerve trunks associated with the ganglion, but a LS-EPSP could be seen in some cells following stimulation of any of the nerve trunks entering the ganglion (Dun and Jiang, 1982; Tsunoo et al., 1982). Different outcomes did, however, result from dorsal and ventral root stimulation: ventral root stimulation (L2, L3) produced only F-EPSPs, while dorsal root stimulation elicited only a LS-EPSP (Tsunoo et al., 1982). 5.3.2. The transmitter of the LS-EPSP in the guinea pig When SP was applied to the guinea pig IMG a depolarization ensued which was electrophysiologically and pharmacologically similar to the LS-EPSP elicited by repetitive stimulation of the hypogastric nerves (Dun and Karczmar, 1979). Following densensitization by continuous application of SP, presynaptic stimulation failed to elicit the LS-EPSP. Thus, it was suggested that SP may be involved in LS-EPSP in the guinea pig IMG. These findings have been confirmed and extended. 5.3.2.1. Actions of substance P When applied to the guinea pig IMG by superfusion, SP (0.5/.tu) caused a depolarization in about 85% of the ganglionic neurons, a biphasic response in 5% and no effect in the remaining 10% (Dun and Minota, 1981). The biphasic response consisted of an initial small hyperpolarization followed by a depolarization. The average amplitude of the response in those cells exhibiting a monophasic depolarization was 9 mV. With a 10 to 20 sec application of SP the response lasted a little over 3 min. SP was acting directly on the postsynaptic membrane since the SP response was not altered in low Ca2+/high Mg 2+ or by TTX (Dun and Minota, 1981; Tsunoo et al., 1982). D-TC (50/~i), atropine (1/zM), DH/3E and hexamethonium had no effect, ruling out an action of ACh (Dun and Minota, 1981; Tsunoo et al., 1982). In neurons clamped at E, of around - 5 0 to - 6 0 mV, SP elicited two types of changes in Ri°. In half of the cells Ri, increased briefly by 20%, followed by a more prolonged decrease of 23%. In the other half of the cells, an increase of about 25% was observed. In cells with Er higher than - 7 0 mV, the depolarization caused by SP was much larger and accompanied by a large increase in R~,. Much of this increase was probably due to rectification of the membrane since the current-voltage relations of these cells exhibited rectification and a much smaller increase in R~, occurred when the membrane potential was clamped at E,. The effects of Em on the SP response were of two types. In some cells, conditioning hyperpolarization of the membrane enhanced the SP depolarization. The extrapolated equilibrium potential for these cells averaged -32 mV. In the second type of cell, the SP depolarization was attenuated as Em approached the potassium equilibrium potential. When external Na + was replaced by sucrose or Tris, the SP depolarization was reduced to about 20% of the control response. In addition, high K + reduced the response by about one-half (Dun and Minota, 1981). The action of SP on the neurons of the guinea pig IMG thus seems to involve changes in conductance to more than one ion including a combination of increased GNa and decreased GK. Substance P has been shown to depolarize neurons at various sites in the central and peripheral nervous systems (Katayama and North, 1978; Krnjevic, 1977; Nicoll, 1978; Otsuka and Konishi, 1977). However, the mechanism of action depends on the neurons being examined. An increased R~n was observed in the myenteric plexus of the guinea pig (Katayama and North, 1978) and in cat motoneurons (Krnjevic, 1977), while both a decrease and increase was seen in the guinea pig IMG (Dun and Minota, 1981). A decreased Rin was observed in frog spinal motoneurons (Nicoll, 1978).

M. A. SIMMONS

72

A

13 lOmV I rain FiG. 8. Rin changes of a guinea pig IMG cell during a LS-EPSP, (A), and during SP application, (B). Electrotonic potentials elicited by 0.3 hA, 200 msec hyperpolarizing pulses every 4 sec. LS-EPSP evoked by splanchnic nerve stimulation at 20 Hz during period indicated by bar in (A), SP (0.07 /~M) applied during period indicated by bar in (B). Note similarity of R~, changes as indicated by increased amplitude of electronic potentials. From Tsunoo et al. (1982). Copyright 1982, Pergamon Press, Ltd. With permission of the authors and Pergamon Press, Inc.

One suggestion for the different conductance changes observed in the IMG in response to SP is that GNa activation and GK inactivation occur at different locations on the cell (Dun and Minota, 1981). Whenever GK was found to constitute the primary effect of SP, the peptide was applied iontophoretically to the soma of the neuron, while in those studies where the entire neuronal surface was exposed to SP, various resistance changes have been observed. It is possible then that GK inactivation occurs primarily at the soma membrane while GNa activation is the predominant effect away from the soma membrane, for instance on the dendritic membrane. 5.3.2.2. Substance P mimics the L S - E P S P Several lines of evidence support the hypothesis that SP is the transmitter of the LS-EPSP in the guinea pig IMG. First, changes in Rin during an SP depolarization are similar to those which occur during a synaptically-induced depolarization. For a particular cell both the LS-EPSP and SP depolarization are associated with a membrane resistance change of the same type (Fig. 8) (Dun and Jiang, 1982; Tsunoo et al., 1982). Second, the LS--EPSP has been shown to disappear following desensitization with SP (Fig. 9), With prolonged application of SPa LS-EPSP could not be elicited. Following washout of SP the LS-EPSP returned (Dun and Jiang, 1982; Tsunoo et al., 1982). Also, an SP analog which differentially antagonizes the effects of SP on smooth muscle has a depressant effect on the

t,

Substance P (1/,tM)

|

, ,, 110 mV

40mc

Fio. 9, Effects of SP on the LS--EPSP in a guinea pig IMG cell. Intervals between recordings marked in minutes. First, a LS-EPSP was induced by repetitive stimulation of the hypogastric nerves. Application of SP (lp, M, between arrows) caused a depolarization. With the continued presence of SP the depolarization gradually subsided. At this point nerve stimulation failed to elicit a depolarization. In the final record, 5 rain following washout of SP, a LS-EPSP could again be elicited. From Dun and Jiang (1982). With permission of the authors and the Physiological Society.

SYNAPTIC TRANSMISSION IN THE PREVERTEBRAL SYMPATHETIC GANGLIA

A

E

13

13

D

H

73

.

FIG. 10. Effects of (D-Pro2,D-PheT,D-Trp~)-SPon the LS-EPSP in a guinea pig IMG neuron. LS-EPSPs elicited by stimulation of the hypogastricnerves (30 Hz for 4 sec, F-EPSPs appear as initial vertical bar). (A) Control. (B) 5 min after perfusion with the SP analog (1 ~M). The amplitude of the LS-EPSP is depressed by about 30%. (C), (D) and (E) Responses 10, 25 and 60 min following washing with normal Krebs solution. (F) 5 min after perfusion with SP analog (10/XM). The amplitude of the LS-EPSP is depressed by about half. (G) and (H) Responses 5 and 30 min after return to normal Krebs solution. From Jiang et al. (1982). Copyright 1982 by the AAAS. With permission of the authors and the AAAS. L S - E P S P in the guinea pig IMG in a concentration-dependent manner (Fig. 10) (Jiang et al., 1982a). This effect seems specific for the L S - E P S P since the nicotinic F - E P S P is unaffected as is the muscarinic slow excitatory postsynaptic potential in the rabbit SCG. The depression of the L S - E P S P by SP antagonists has been confirmed in another laboratory (Konishi et al., 1983). Other possible transmitters have been tested on the guinea pig IMG, but none affect the L S - E P S P (Tsunoo et al., 1982). Angiotensin II depolarized cells in the IMG; an angiotensin II antagonist blocked the depolarizing action of angiotensin II but did not affect the L S - E P S P or SP depolarization. L H R H failed to depolarize IMG cells. 5-HT, however, did depolarize the cells; this action was antagonized by methysergide, but methysergide did not affect L S - E P S P or SP depolarization. Norepinephrine and gamma-amino butyric acid also depolarized I M G cells and were antagonized by phentolamine and picrotoxin, respectively. Neither of these antagonists affected the L S - E P S P or SP depolarization. The following substances were without effect on IMG cells: L-glutamate, glycine, thyrotropinreleasing hormone and somatostatin. 5.3.2.3. Release o f substance P Using radioimmunoassay techniques it has been shown that an SP immunoreactive material can be released from the IMG following incubation of the ganglion in a high K ÷ solution (Konishi et al., 1979b) and that this release is Ca2+-dependent. Following section of the inferior splanchnic nerves to the IMG, an accumulation of SP was found in the central stump of these nerves. Proximal to the section the SP content was four times greater than that in the ganglionic segments indicating that the cell bodies of origin of the SP fibers are not located within the IMG. Similar effects were observed following hypogastric nerve section (Tsunoo et al., 1982). 5.3.2.4. Release and depletion o f S P by capsaicin Further evidence for the involvement of SP in t h e L S - E P S P is provided by studies on JPN 24:1-E

74

M . A . SIMMONS

11

Capsaicin (100 #M)

Post-wash 3 hr

'" ~10mV 80 sec 7-#

FIG. 11. Effects of capsaicin on the LS--EPSP in a guinea pig IMG neuron. Nerve stimulation (curved arrows) elicited a burst of action potentials followed by a LS-EPSP. Capsaicin (100 ~M) applied between the two straight arrows, resulted in a large depolarization followed by a disappearance of the LS-EPSP. The LS-EPSP has not returned after 3 hr of washing with normal Krebs solution (bottom tracing). The F-EPSPs were attenuated during the capsaicin depolarization but recovered afterwards. From Dun and Kiraly (1983). With permission of the authors and the Physiological Society.

capsaicin. This compound has been shown to cause the release and subsequent depletion of SP (Gamse et al., 1981a,b; Jessell et al., 1978). This substance also results in a marked reduction in SP immunoreactive nerve fibers in the guinea pig IMG (Dalsgaard et al.. 1983b). Application of capsaicin (0.5-100/z~) to the guinea pig IMG in vitro resulted in a depolarization which was similar to the LS-EPSP (Fig. 11) (Dun and Kiraly, 1983" Tsunoo et al., 1982). The peak amplitude of the LS-EPSP which could be elicited by preganglionic stimulation was positively correlated with the amplitude of the depolarization following capsaicin application in a given cell, while the duration of the capsaicin depolarization was significantly longer than the LS-EPSP. In any given cell, the three depolarizations. synaptic, SP or capsaicin, always caused similar resistance changes. Capsaicin only caused a depolarization in those cells exhibiting a LS-EPSP and did not affect Era, Rin or the F-EPSP in those cells that did not show a LS-EPSP. Capsaicin apparently resulted in a depletion of the transmitter of the LS-EPSP since, following the application of capsaicin. nerve stimulation failed to elicit a LS--EPSP. In a few cells, however, a partial restoration of the LS-EPSP was seen about an hour after washout of capsaicin. A second application of capsaicin was usually ineffective in eliciting a depolarization. After capsaicin. SP was still effective in eliciting a slow depolarization, indicating no change in postsynaptic sensitivity. Further evidence for a presynaptic site of action of capsaicin was obtained by experiments in Ca2+-free perfusing solution. In this case, capsaicin was ineffective. In ganglia pretreated with TTX, however, a capsaicin depolarization could still be induced. This supports the hypothesis that capsaicin causes a Ca2+-dependent release of transmitter that does not depend on the generation of action potentials. Capsaicin was ineffective in a few cells (8%). In these cells, the LS-EPSP was not affected by prolonged superfusion with SP. In bullfrog ganglia, where the mediator of the LS-EPSP is presumed to be an LHRH-like peptide, capsaicin had no effect (Dun and Kiraly, 1983). Capsaicin (0.9-1.5/x~) also stimulated the release of SP into a perfusate of the IMG in vitro (Tsunoo et al., 1982). The amount of SP released following incubation in capsaicincontaining solution was about five times greater than spontaneous release. The release of another neuropeptide, VIP, which has also been localized to the IMG by immunohistochemistry, was not affected by capsaicin. As with the intracellular recordings, the effects of capsaicin on the release of SP were not observed in a low Ca2+/high Mg2+ medium. The effects of in vivo capsaicin on the LS-EPSP in the guinea pig IMG have been less conclusive (Peters and Kreulen, 1984). In contrast to the above studies which applied capsaicin in a perfusate of the ganglion in vitro, these authors administered subcutaneous

SYNAPT1C TRANSMISSION

IN T H E P R E V E R T E B R A L

SYMPATHETIC

A

GANGLIA

75

B Met-enke Ynolin

mV

DAEA O.UJM -~

eQ

60e

a

......

°

_

• o

_

b

,ooooooooo-,

'OOO

O00l

d. 0

J

5

d, d 30 +.

I ,/

I

15" 25

10

20-t

I

Nolox0ne 1.5pM

OAEAO.3pu O0

" ° " ooP ° ° ° ° ° o o

oo,

°°ooolooolooo

oo°~ bl

•

C

'~1

o -50 I°°°°°°°

I

5o

mV

t

o •

'

a

~5

II

DAEA O.3pM

ooo

~o-

I

I

55

I //

60

I

85

I

I

90

-~

30 mY

30 ms

t

I

95 min

FIG. 12. Effects of enkephalins on F-EPSPs, Er and Rin of a guinea pig IMG cell and antagonism by naloxone. In 0.7 mM Ca 2+ and 2 mM Mg 2+. DAEA-(D-Ala)-methionine enkephalinamide. F-EPSPs elicited by stimulation of the splanchnic nerve at 1 rain intervals for 16 sec at 2.5 Hz were superimposed on electronic potentials elicited by hyperpolarizing pulses of 70 msec and 0.5 hA. (A) Filled circles indicate the amplitude of 32 average F-EPSPs. Open circles indicate E,. Ordinate-left, F-EPSP amplitude; right, E,. Drugs were applied for the periods indicated by the bars at the concentrations indicated. (B) Averaged records obtained at arrows a-c of (A). The enkephatins depressed the amplitude of the F-EPSP without significant effect on E, or Rin. Reprinted with permission from Nature, 282, 515. Copyright 1979, Macmillan Journals Limited.

doses of 50 to 350 mg/kg capsaicin. Four to twelve days later, the animals were killed and the IMG removed for intracellular recording in vitro. The neurons in the ganglia of these animals were similar to control ganglia with respect to E, and F-EPSP amplitude and frequency. LS-EPSP amplitude in the capsaicin-treated animals was decreased to 50% of control, but was not completely abolished. SP was then constantly perfused in an attempt to desensitize the remaining depolarization. This treatment decreased the LS-EPSP by 50% in both control and capsaicin-treated animals, but again the LS-EPSP was not completely abolished, in contrast to the previous results. These authors conclude that the remaining depolarization is evidence for the involvement of some transmitter other than SP. An alternative explanation is that, as has been shown immunohistochemically (Dalsgaard et al., 1983b), not all the SP fibers were destroyed by in vivo capsaicin, thereby resulting in a reduction, but not a complete disappearance of the response. In conclusion, the electrophysiological evidence above supports the hypothesis that peptides mediate the L S - E P S P - - L H R H in the bullfrog paravertebral ganglia and SP in the guinea pig IMG. These data are further supported by the immunohistochemical tracing experiments discussed in following sections.

5.3.3. Effects of enkephalins in the guinea pig IMG The opiate peptides have also been shown to exert specific effects on neural transmission in the guinea pig IMG. Met-enkephalin, Leu-enkephalin and (o-Ala2)-Met-enkephalina mide, a metabolically stable Met-enkephalin analog, all exert similar effects, the only quantitative difference being that (o-AlaZ)-Met-enkephalinamide is more potent (Konishi et al., 1979a). Thus, these substances will be referred to collectively as ENK. When the IMG is pefused with a solution containing ENK, the amplitude of the F-EPSP evoked by presynaptic stimulation was depressed by about 50% as shown in Fig. 12

M. A, SIMMONS

76

Jl0mv 4 0 $eC

a

4

.... k

I

0

©

g

i

5

I0 [~

2 ,a~

I

15

I

20

I

25

I

30

l

35

I

40

i

45

~

50

55 ....

I~1 2 u M ~nl 2 uM

F1G. 13. Effects of methionine-enkephalin on the LS-EPSP in a guinea pig IMG cell. In the presence of d-TC (5(I/ZM). LS-EPSPs elicited by hypogastric nerve stimulation (3(1 Hz for 5 sec). Records taken every 5 min correspond to the labeled points on the graph below. (a) Control. (b) 5 min after Met-ENK (2/xu). (c), (d), (e) and (f) 5, 10, 15 and 20 rain after wash. (g) 5 min after naloxone (2/.tM). (h) 5 min after Met-ENK. (i), (j) and (k) responses after washing. From Jiang et al. (1982). With permission of the authors and Elsevier Biomedical Press.

(Konishi et al., 1979a). ENK had little or no effect on either Em or R~,. The ENK effect was mediated by opiate receptors since it was blocked by the specific opiate antagonist naloxone. Naloxone itself did not produce a significant change in the amplitude of the F-EPSP. ENK did not affect the response to iontophoretically applied ACh or SP, indicating no change in the sensitivity of the postsynaptic membrane. Quanta! analysis revealed that the effect of ENK was to increase the number of failures of F-EPSPs and to reduce quantal content. These results suggested a presynaptic site of action of ENK in this preparation. A presynaptic site of action has also been suggested for ENK effects on the LS-EPSP in this ganglion (Jiang et al., 1982b). ENK depressed the amplitude of the LS-EPSP in about 70% of the cells tested by an average of 30% (Fig. 13). In 90% of these cells ENK was without effect on either Em o r Rin. Pretreatment of the ganglion with naloxone or naltrexone effectively prevented the effects of ENK in all cells tested, The direct effects of SP on the postsynaptic membrane were not affected by ENK or the antagonists. Additionally, it was found that the administration of naloxone or naltrexone alone to the IMG resulted in an increased amplitude of the LS-EPSP in about one-half of the cells (Fig, 14), indicating that endogenous ENK may be acting tonically to inhibit the LS-EPSP. These findings suggest that ENK may play a role in the modulation of the LS-EPSP in this ganglion and that the site of action of ENK would be presynaptic.

77

SYNAPTICTRANSMISSION1N THE PREVERTEBRALSYMPATHETICGANGLIA

CON CON

_

SP

Nalomone

I uM

~k~lolo~ 1~M

5'

5'

WASH 5'

¢-.___

........... " SP

J ~

WASH 10'

10° ~ _ :

SP

Jlo.v 20

Se(

I I0 mV

(~

40

sec

Fro. 14. Effects of naloxone on the LS-EPSP (left) and SP-induced depolarization (right) in a guinea pig IMG cell. In the presence of d-TC (50 p.m). LS-EPSPs elicited by stimulation of the hypogastric nerve (30 Hz for 5 sec). The presence of naloxone (1/zM) resulted in an enhancement of the LS-EPSP but did not affect the SP depolarization. From Jiang et al. (1982). With permission of the authors and Elsevier Biomedical Press.

5.4. RABBIT INFERIOR MESENTERIC GANGLION A typical LS-EPSP recorded from the rabbit IMG is shown in Fig. 15. As mentioned previously (Section 4.2.2.2), a muscarinic S-EPSP was observed in some of the cells in this ganglion. Most frequently, this response was contiguous with a noncholinergic response. As illustrated in Fig. 15, atropine increased the time to peak of the slow depolarization, on average, from 3 to 9 sec, but it did not affect the amplitude, Rin change or duration of the slow response. The LS-EPSP in the rabbit IMG averaged 4.3 mV, lasted 159 sec and was accompanied by a 27% increase in Rin (Simmons and Dun, 1985). A LS-EPSP was observed in 63% of the cells in this ganglion. Following separate stimulation of two different nerve trunks a LS-EPSP could be elicited in 36% of the cells following stimulation of each of the nerves. Several findings indicated that this was a synaptic response. First, the response was not mimicked by direct intracellular stimulation that elicited repetitive firing of the cell comparable to that seen during the preganglionic stimulus train. Second, the LS-EPSP was not affected following blockade of the spikes by d-TC (50 p.i), but was blocked by TTX (1 /xi). Finally, the response was effectively abolished in a low Ca2+/high Mg2÷ solution. Like the F-EPSP inputs to the IMG, the LS-EPSP inputs were found not to be topographically organized (Simmons and Dun, 1985). LS-EPSPs could be elicited in cells located throughout the IMG irrespective of the nerve trunk being stimulated. This study also examined the relationship between the amplitude of the LS-EPSP and the parameters of presynaptic stimulation. The amplitude was found to increase with

78

M . A . SIMMONS

A

i ii r r i-fh-iY

iilillltl~lllllllllllll

'17

IIIIIIIIIII1111

.

;I,IIIT~ilIIIIIIITII!!ilIEi

,

,']--.~'~ll

IIlllllllJllltll

i l l l l l t l l l l l

] l l l l l l l l l l l l ~ l l l l l [ l l i l i , ~

III

I l I l l l I l f l i l l J i t l l l i l l l l l l l i i [ ' - ~ -

60sec

J 5mV/0.SnA

FIG. 15. Noncholinergic slow depolarization m a rabbit IMG neuron. Top tracings, membrane potential. Bottom tracings, current pulses. (A) Following stimulation of the ascending mesenteric nerve at 16 Hz for 3 sec a long lasting depolarization was observed following the F-EPSP (initial vertical tracings). (B) After 5 min perfusion with d-tubocurarine (50/xM) the initial F-EPSPs were completely blocked, but the slow depolarization remained. (C) Addition of atropine (1/XM) to the perfusate resulted in an increase in the rise time of the response, but did not antagonize the long lasting depolarization. From Simmons and Dun (1985).

increasing numbers and duration of presynaptic stimuli (Fig. 16). This relationship for many cells is plotted in Fig. 17. The LS-EPSP in the rabbit IMG differs from the LS-EPSP in the guinea pig IMG with regard to its pharmacology. Tests to determine an involvement of SP in the LS-EPSP in the rabbit IMG gave negative results (Simmons and Dun, 1985). (1) SP depolarized only 33% of the cells in the rabbit IMG compared to 90% in the guinea pig (Dun and Minota, 1981); (2) the SP depolarization in 60% of those cells affected was accompanied by a decrease in Ri, unlike the LS-EPSP which was never accompanied by a Rio decrease; (3) SP application did not result in a densensitization of the LS-EPSP; (4) SP antagonists did not affect the LS-EPSP and (5) capsaicin had no effect on the rabbit IMG cells or on the LS-EPSP. These data suggest that SP is not the chemical mediator of the LS--EPSP in this ganglion. The compound involved remains to be determined. 5.5. NONCHOLINERGIC INHIBITORY POTENTIALS

A hyperpolarizing response similar in time course to the S-IPSP seen in other sympath-

SYNAPTIC

TRANSMISSION

IN THE

3 sec train

Hz

I

•

---11,

I

--

| --

1

I

_

SYMPATHETIC

~|

!

. .

GANGLIA

79

6 sec train

,_1 4

PREVERTEBRAL

,

,

....

--

-r

-

--ql'

75 - _

I

100 - - - I_----. ____ :_ |

.

FIG. 16. Response of a cell in the rabbit IMG to presynaptic stimulation of increasing frequency and duration. In the presence of atropine (1 ttr,l). Left, responses following a 3 sec train of presynaptic

pulses. Right, responses following a 6 see train of presynaptic pulses. The frequency of the presynaptic pulses for these train durations is increased from 2-100 Hz as labeled on the left. Note the dependency of the amplitude of the LS--EPSPon the frequency and duration of nerve stimuli. Horizontal calibration, 30 sec. Vertical calibration, 10 mV. From Simmons and Dun (1985).

etic neurons (see K u b a and Koketsu, 1978) has been observed following repetitive stimulation of the intermesenteric or hypogastric nerves in the guinea pig I M G (Ma et al., 1983). A S--IPSP was seen in 16% of these cells and averaged 4.7 m V in amplitude and 4.3 sec in duration. This response differed, however, from that seen in other ganglia in that it was unaffected by atropine nor was it affected by a-adrenergic antagonists. A noncholinergic L S - I P S P , having a much longer time course than the S-IPSP, was observed in the rabbit I M G (Simmons and Dun, 1985). This response was apparent in 13% of the cells I M G and in most of these cells was followed by a L S - E P S P .

6. The Localization of Peptides in the Prevertebral Ganglia Immunoreactivity to a n u m b e r of peptides has been examined in the sympathetic ganglia of the guinea pig, cat and rat. Bombesin, cholecystokinin, enkephalin, neuropeptide Y, somatostatin and vasoactive intestinal polypeptide immunoreactivity have been demonstrated in the prevertebral ganglia of several species. Immunohistochemical results must be interpreted with caution since antisera m a y react not only with the antigen against which they were raised but also with structurally related compounds, e.g. precursors of the antigen or peptides of similar structure. For this reason, immunohistochemical results are

~(1

M,

A.

SIMMONS

% ~50A

M a25P L

I a00T O

O

75-

E 50-

25-

Oi

20

40

60

80

100

i

i

i

I

I

I

t

[

t

1

120

140

160

180

200

220

240

260

280

300

PULSES

FIG. 17. Relationship between the number of presynaptic pulses and amplitude of the LS-EPSP. Ordinate, % amplitude of the response obtained with 90 pulses. Abscissa, number of pulses in the stimulus train. The graph shows the mean + 1 standard deviation of the responses obtained at each value of pulses.

discussed in terms of "immunoreactivity" or "iike-immunoreactivity" (H6kfelt et al., t 980) and must be confirmed by characterization of both the antigen and antibody. 6.1. BOMBESIN Dense networks of bombesin immunoreactive nerve fibers have been observed in the CG-SMG of the rat and guinea pig and in the guinea pig IMG (Schultzberg, t983). No mention is made of BOM immunoreactivity in the rat IMG. In the rat CG-SMG and in the guinea pig IMG the bombesin fibers were evenly distributed. In the guinea pig CG-SMG fibers were not seen in all regions of the ganglion. Radioimmunoassay also revealed differences between the rat and guinea pig CG with regard to BOM content--the rat ganglion contained four times more BOM per gram of tissue than the guinea pig ganglion. 6.2. CHOLECYSTOKININ CCK has also been observed in the guinea pig IMG and CG (Dalsgaard et al., 1983a). CCK fibers form a dense network surrounding the ganglion cells in the guinea pig IMG. 6.3. ENKEPHALINS In the IMG of the guinea pig dense networks of ENK immunoreactive nerve terminals exhibiting strongly fluorescent varicosities and thin intervaricose fibers were observed following incubation with either met- or Ieu-ENK antisera (Schultzberg et al., 1978: Schultzberg et al., 1979). The majority of principal ganglionic neurons were surrounded by strands of fluorescent fibers but sometimes small groups of cells were seen which lacked fluorescence. No ENK immunoreactivity was present in the principal ganglionic neurons but a few SIF cells apparently contained ENK. A smaller number of strongly fluorescent fibers were seen in the CG-SMG surrounding a smaller percentage of the ganglion cells. A few ENK immunoreactive fibers were also present in the SCG.

SYNAPTIC TRANSMISSION IN THE PREVERTEBRAL SYMPATHETIC GANGLIA

81

The rat CG-SMG and IMG contain medium to strongly fluorescent varicose nerve terminals using met-ENK antisera while a weaker fluorescence was seen after incubation with Ieu-ENK antisera (Schultzberg et al., 1979). The density of the ENK fluorescence varied throughout the ganglion with many fluorescent strands in some areas and few in other parts. The rat SCG, in addition to a patchy network of ENK fibers similar to the prevertebral ganglia, also exhibited a medium intensity fluorescence in a minority of the ganglionic neurons. 6.4. NEUROPE~nDEY A population of prevertebral ganglion cells were originally found to be immunoreactive to antisera to APP (Lundberg et al., 1980, 1982a). Subsequent attempts to measure the presence of APP by radioimmunoassay in these ganglia were unsuccessful, however (Lundberg et al., 1982b, 1984). Another peptide in this family, neuropeptide Y, was characterized (Tatemoto, 1982) and antisera to NPY were found to react with the previously described APP-immunoreactive cells (Lundberg et al., 1982b, 1984). Also, radioimmunoassay for NPY revealed the presence of NPY in areas which had stained positively for APP but in which no APP could be detected. It is now thought that the earlier descriptions of APP-containing cells were actually NPY-containing cells. This has been taken into account in the following discussion. In the rat, NPY-like immunoreactivity was observed in the cell bodies of the SCG, stellate and CG. NPY immunoreactivity was also observed in the sympathetic ganglia of the cat. The highest proportion of labeled cells was found in the CG. Many cells were seen in the SCG and seventh lumbar ganglia. Somewhat lower numbers were found in the SMG and IMG. No NPY immunoreactivity was seen to occur in varicose nerve terminals in these ganglia (Lundberg et al., 1980; Lundberg et al., 1982a). A large number of the cells in the SCG of the guinea pig were also found to be NPY-positive, NPY-like immunoreactive cell bodies were also found in the prevertebral ganglia of the guinea pig. In the CG-SMG complex the number of NPY-labeled cells varied in different regions of the ganglion, while in the IMG cells were distributed more evenly (Lundberg et al., 1982). NPY immunoreactive cells were also seen in cat sympathetic cells. 6.5. SOMATOSTATIN

Unlike the above peptides which were found in nerve fibers in the IMG, SOM has been found to be present in cell bodies of most sympathetic ganglia studied, including the SCG, CG-SMG and IMG of the guinea pig and the middle cervical, CG-SMG and IMG of the rat (H6kfelt et al., 1977a) The immunofluorescence was localized in the cytoplasm, often forming a continuous ring around the nucleus and extending short distances into the cell processes. As with the other neuropeptides, the most intense labeling was seen in the abdominal prevertebral ganglia, particularly in the guinea pig (H6kfelt etal., 1977a) About two-thirds of the principal ganglion cells in the IMG and SMG were labeled. In the anterior superior part of the CG only 25% of the cells were labeled. Very few cells were labeled in the SCG. Incubation of adjacent sections with antibodies to DflH revealed that the cells containing SOM were noradrenergic ganglion cells. The presence of SOM immunoreactivity in the guinea pig CG has been confirmed with electron microscopy (L6rfinth et al., 1980). Again, a greater portion of labeling was seen in the CG than in the SCG. Within the CG, unlabeled axons formed presynaptic contacts with labeled dendrites. In the nerves of the CG, no labeled axons could be found in the peripheral nerves, indicating that the SOM fibers do not originate nor project peripherally, while only a very few labeled axons could be found in the splanchnic nerves. It is concluded that most of the SOM immunoreactive axons in the CG come from the principal ganglionic neurons, although a few may originate centrally. In the rat, the appearance and distribution of SOM immunoreactivity was similar to that seen in the guinea pig. Although not quantified, the number of cells labeled was less (H6kfelt et al., 1977a). JPN 24:1-F

,~2

M.A.

SIMMONS

6.6. SUBSTANCEP All sympathetic ganglia studied to date, including the SCG, stellate, C G - S M G and IMG of the guinea pig; the SCG, C G - S M G and IMG of the cat and the SCG, middle cervical, stellate, C G - S M G and IMG of the rat, have been found to contain SP immunoreactive fibers in varying amounts (H6kfelt et al., 1977c). No SP immunoreactivity was observed in the cell bodies of the sympathetic ganglionic neurons in this study (H6kfelt et al., 1977c). However, SP irnmunofluorescence has been observed in most of the cell bodies in cultures of the SCG of neonatal rats (Kessler et al., 1981). In the guinea pig, the highest density of SP immunoreactive fibers was found in the CG, SMG and IMG (H6kfelt et al., 1977c). In the IMG strongly fluorescent beaded fibers were seen with the number of fibers varying in different parts of the ganglion. Fluorescent nerve bundles frequently apposed and surrounded ganglion cell somata. In the C G - S M G , the appearance was somewhat different with the axons running singly, rather than in bundles, and forming a more delicate network. Comparatively few SP immunoreactive fibers were seen in the stellate and SCG. In the cat CG, SMG and IMG, SP immunoreactivity was found to be of a more moderate intensity than in the same ganglia in the guinea pig (H6kfelt et al., 1977c). Thin, non-varicose axons of low fluorescence intensity and nerve terminals with strongly fluorescent varicosites were frequently seen surrounding the ganglion cells. In the SCG only a few fluorescent fibers were present. In the rat CG, SMG and IMG, a moderate SP fluorescence formed a plexus surrounding almost all of the ganglion cells. In the stellate, SCG and middle cervical ganglia only singly occurring fibers were present (H6kfelt et al., 1977c). After treatment of guinea pigs with capsaicin (4 × 50 mg/kg, s.c.), SP immunoreactive fibers were greatly depleted in the IMG (Dalsgaard et al., 1983b), although a few immunofiuorescent fibers were still observed. BOM, CCK, E N K and VIP fibers were unaffected. This result agrees with the electrophysiological studies showing attenuation of the L S - E P S P following capsaicin application (see Section 5.3.2.4). 6.7. VASOACTIVEINTESTINALPOLYPEPTIDE The vast majority of cells in the guinea pig IMG were surrounded by a very dense network of strongly VIP immunoreactive varicose nerve fibers (HOkfelt et al., 1977b), Occasionally, small groups of principal ganglionic neurons were seen devoid of fluorescent fibers. The inferior part of the C G - S M G had a similar distribution of VIP immunoreactive fibers. Additionally, single cell bodies were labeled, usually occurring in areas where relatively few fluorescent fibers were seen. The superior part of the C G contained similar dense networks interspersed with more frequently occurring areas containing a few or no fibers. In the SCG a sparse network of thin fibers was present, but no cell bodies were immunoreactive. The VIP immunoreactive fibers have been shown to form synaptic contacts in the guinea pig CG (Kondo and Yui, 1982). In most cases, 72%, the VIP fibers were presynaptic to dendrites and in a few cases (9%) made contacts with ganglion cell soma. The remaining contacts were with vesicle-containing profiles. These were most likely axo-axonic contacts with preganglionic fibers. VIP immunofluorescence was generally much less intense in the rat ganglia and cell bodies were never labeled (H6kfelt et al., 1977b). In the CG, SMG and IMG, a regular network of weakly fluorescent fibers was seen surrounding most of the ganglionic neurons. In the SCG, single varicose fibers were observed while the middle cervical ganglion showed no VIP immunoreactivity. In the cat, singly-occurring VIP immunoreactive cell bodies were present in the prevertebral ganglia (Lundberg et al., 1979, 1982a). The VIP immunoreactive fibers were less than in the guinea pig, but were still dense and decreased in fluorescence intensity following vinblastine treatment of the ganglion. The VIP cell bodies did not fluoresce following incubation with tyrosine hydroxylase or D/3H antibodies, indicating that they do not contain D A or NE, but most did stain for acetylcholinesterase, indicating they also contained ACh.

SYNAPTICTRANSMISSIONIN THE PREVERTEBRALSYMPATHETICGANGLIA

83

6.8. COEXISTENCE OF PEPTIDES AND NOREPINEPHRINE As seen above, NPY, SOM and VIP immunoreactive principal ganglion cell somata have been found in the prevertebral ganglia. Recently, the pattern of coexistence of the various peptides and NE has been studied and the cells have been classified according to the presence or absence of these compounds (Lundberg et al., 1982a). In the guinea pig prevertebral ganglia, the principal ganglion cells have been categorized into four types according to peptide immunoreactivity (Lundberg et al., 1982a). The first type of cell contains both SOM and NE. The second type of cell contains both NPY and NE. These two types were found to be mutually exclusive in that NPY and SOM immunoreactivity were never observed in the same neurons. The third type of cell contains only NE, showing no immunoreactivity to NPY, SOM or VIP. The fourth type of cell is the singly-occurring VIP immunoreactive cell. This type was not labeled by antibodies to tyrosine hydroxylase or DflH and is, therefore, thought to contain VIP but not NE.

7. Tracing Pathways of the Prevertebral Ganglia 7.1. CELIACGANGLION When HRP was applied to the greater splanchnic nerve, labeled cell bodies were found in three areas, spinal cord, DRG and paravertebral ganglia, labeling the cell bodies of preganglionic sympathetic neurons, sensory neurons and postganglionic sympathetic neurons, respectively (Kuo and Krauthamer, 1981). The largest number of labeled paravertebral cells was at T13, the level at which the greater splanchnic nerve branches from the sympathetic trunk. When the peripherally-running nerves of the CG were exposed to HRP, labeled cells were observed in the CG, paravertebral ganglia and DRG. No sympathetic preganglionic fibers were labeled, thus, the preganglionic fibers do not run further peripherally than the CG. In agreement with the physiological studies, the labeled CG cells were clustered near the exit of the treated nerve from the ganglion (Kuo and Krauthamer, 1981). Autoradiographic tracing has also revealed connections between left and right celiac ganglia of the cat (Kelts et al., 1979) with processes from one CG crossing to surround cells in the opposite CG. 7.2. INFERIOR MESENTERIC GANGLION The physiological studies mentioned above provide evidence for a complicated system of pathways to and from the IMG. Additionally, the immunohistochemical findings have revealed neuropeptide pathways in this ganglion. Anatomical studies on the connectivity of the IMG have recently been combined with peptide immunohistochemistry to further examine these pathways. These studies provide direct evidence for inputs to the IMG from a variety of sites and for the presence of peptides in specific pathways of the ganglion. These findings also support the electrophysiological findings which suggest a neurotransmitter function of peptides in this ganglion. All of the following studies have used the guinea pig. 7.3. APPLICATION OF HORSERADISH PEROXIDASE TO THE INFERIOR MESENTERIC GANGLION Following injection of HRP into the IMG, widely-distributed labeled cell bodies were seen (Dalsgaard and Elfvin, 1982). In the CG-SMG labeled cells appeared singly or in small groups and were apparently distributed at random. HRP-labeled cell bodies were also observed in the PG and usually occurred singly. A small number of HRP-positive cell bodies were also observed in the nodose ganglia. After injection of the fluorescent tracer True blue into the IMG, labeled cell bodies were also seen in the enteric nervous system and in the DRG. Some of the cell bodies of the ganglia in the myenteric plexus of the distal colon were clearly labeled. Also, in the submucus plexus an occasional fluorescent cell was seen.

;';4

M . A . SIMM()NS 7.4. APPLICATION OF H R P TO THE NERVE TRUNKS

Following application of H R P to the severed ends of the hypogastric nerve, labeled cell bodies could be seen in the IMG, in the C G - S M G , in the spinal cord and also in the D R G (Dalsgaard and Elfvin, 1982). A number of randomly-distributed postganglionic neurons were labeled in the IMG, while only a few labeled cells appeared in the C G - S M G . Confirming the physiological findings that a significant number of preganglionic fibers descend in the hypogastric nerve, sympathetic preganglionic cell bodies labeled with HRP were found in the nucleus intermediolateralis and nucleus intercalatus of the spinal cord. When H R P was applied to the cut colonic nerves, heavily labeled cells were seen in the IMG (Dalsgaard and Elfvin, 1982). Cells were also labeled in the caudal part of the SMG. Labeled axons could be detected in the intermesenteric nerve connecting the IMG with the SMG. Labeled cells could also be found in the D R G , but no cells filled with H R P were seen in the spinal cord, indicating that visceral afferents, but not postganglionic sympathetics, run in this nerve. H R P tracing studies have also shown that the connections of the IMG, the CG and the SMG with the colon are different (Kreulen and Szurszewski, 1979a). When H R P was applied to the cut ends of the celiac nerves, which travel to the colon from the CG, labeled cells were observed in the CG, a few cells in the S M G , but no cells in the IMG were labeled. When applied to the colonic nerves, dense filling was observed throughout the IMG, in some cells in the SMG and very few in the CG. This anatomical finding correlates with electrophysiological studies. A higher percentage of neurons in the rostral part of the CG receive input from celiac nerves than from the intermesenteric nerves (Kreulen and Szurszewski, 1979c). In the IMG, however, the inputs are not patterned, cells in all parts of the IMG receive synaptic inputs from all of the associated nerve trunks (Crowcroft and Szurszewski, t971).

7.5. PREGANGL1ONICFIBERS TO THE IMG The H R P technique has been used to study, in detail, the origin of the preganglionic fibers to the IMG (Dalsgaard and Elfvin, 1979). The sympathetic nuclei in the spinal cord include the nucleus intermediolateralis, the nucleus intercalatus and the nucleus intermediomedialis. After application of H R P to the guinea pig IMG, labeled preganglionic neurons were found in the intermediolateralis and intercalatus nuclei. The cells were located bilaterally in the spinal cord, in contrast to the SCG which receive fibers only from the ipsilateral cord. Labeled cells could be found from the level of T13 to L4. Only a few cells were seen at these limiting levels. In agreement with the physiological studies, the highest proportion of labeled cells were seen at L2 and L3. 7.6. ENKEPHALIN-CONTAINING PREGANGLIONIC NEURONS

As mentioned previously, a dense network of ENK-containing nerve terminals was found to surround the ganglion cells in the guinea pig IMG (Schultzberg etal., 1978, 1979). The origin of these fibers has been studied using a combination of retrograde tracing and immunohistochemistry. Following incubation of slices from the L2 and L3 levels of the spinal cord with antibodies to E N K many fluorescent cell bodies were seen in the nuclei which contain the sympathetic preganglionic neurons (Dalsgaard et al., 1982b). Most of the E N K immunoreactive cells were in the nucleus intermediolateralis with a few in the nucleus intercalatus. These cells have been compared with the labeling obtained following application of a retrogradely transported dye to the IMG. About 40% of the cells which contained enkephalin were also found to be preganglionic to the IMG. Of the remaining cells some were labeled with E N K only and some with the tracer only. At the electron microscopic level numerous axo-axonic contacts were observed in the guinea pig IMG, as reported previously for the cat IMG (Elfvin, 1971b).

SYNAPT1CTRANSMISSIONIN THE PREVERTEBRALSYMPATHETICGANGLIA

85

These results give strong evidence for the existence of ENK-containing preganglionic neurons originating in the L2 and L3 levels of the spinal sympathetic nuclei projecting to the IMG. Again, the anatomical findings correspond well to the physiological findings as ENKs have been shown to exert a presynaptic inhibitory effect on synaptic transmission in the IMG (Konishi et al., 1979a; Jiang et al., 1982b). 7.7. SUBSTANCEP-CONTAINING SENSORYNEURONS

When HR P was applied to the IMG, labeled cell bodies were also found in the D R G at the T13 to L5 levels (Elfvin and Dalsgaard, 1977). Labeled cells occurred in the L2 and L3 D R G in rather large numbers. A similar distribution of labeled cells in the D R G was seen following application of HRP to the cut ends of the hypogastric or colonic nerves. This indicates that the fibers of the D R G traverse the IMG, probably on a course to the viscera. When the inferior splanchnic nerves were severed, there was a marked accumulation of SP immunoreactive material in the central stumps of these nerves indicating that SP was being transported from cell bodies central to the IMG (Matthews and Cuello, 1982). When combined with section of the intermesenteric nerve there was a complete loss of SP immunoreactive fibers from the IMG. When the hypogastric or colonic nerves were sectioned the appearance of the SP network in the IMG did not change, but there was an accumulation of immunoreactivity in the ganglionic stumps of these nerves. Additionally, capsaicin treatment almost completely abolished the SP immunoreactivity from the IMG. At the ultrastructural level using peroxidase-antiperoxidase immunohistochemistry SP immunoreactivity was seen in vesicle-containing nerve varicosities in the IMG. Some of these fibers made synapses with dendrites in the ganglion and showed features typical of axo-dendritic synapses. In a further study, retrograde tracing was again combined with immunohistochemistry for the localization of substance P in the D R G (Dalsgaard etal., 1982a). As seen following HRP transport, the fluorescent dye retrogradely transported from the IMG was seen in the cell bodies of the D R G , usually in singly-occurring small cells, but occasionally in small groups. Incubation of the sections of the D R G with antisera to SP revealed the occurrence of small evenly-distributed SP-labeled cells. Comparison of the retrogradely-labeled cells with the SP immunoreactive cells revealed that some of the cells were labeled with both techniques. There were also D R G cells labeled by only one of the techniques. Electrophysiologicai studies have demonstrated that SP is the probable mediator of the LS-EPSP in this ganglion and have suggested that the SP fibers are collaterals of primary sensory neurons (Dun and Jiang, 1982; Tsunoo et al., 1982). The anatomical studies combining SP immunohistochemistry and retrograde tracing strongly support this hypothesis. 7.8.

PEPTIDERGIC NEURONS FROM THE COLON

In contrast to ENK and SP fibers which originate centrally to the IMG in the spinal cord and DR G, respectively, the other peptides which have been found in the IMG appear to have a peripheral origin, at least in the guinea pig IMG. These include BOM, CCK and VIP. 7.8.1. B o m b e s i n In the normal guinea pig, a dense network o f B O M immunoreactive fibers was observed in the guinea pig IMG (Dalsgaard et al., 1983a; Schultzberg, 1983). These fibers remained following section of the splanchnic, intermesenteric and hypogastric nerves, leaving the colonic nerves. When the colonic nerves were ligated BOM immunoreactivity accumulated distal to the ligation. BOM fibers were not observed in the other nerves. These results indicated that the BOM fibers reach the IMG exclusively via the colonic nerves (Dalsgaard et al., 1983a). These fibers may be involved in an afferent projection from the colon to the IMG since BOM immunoreactivity has been observed in the gastrointestinal tract (Dockray et al., 1979).

~

M.A.

SIMMONS

7.8.2. Cholecystokinin The normally dense network of CCK immunoreactive fibers in the guinea pig IM(; disappears following complete denervation of this ganglion (Dalsgaard et al.. 1983a). Similar to the case with B O M , cutting all the nerves except the colonic nerves revealed a pattern of CCK fibers similar to that seen in the normal animals. Ligation of the colonic nerves resulted in an accumulation of CCK immunoreactivity distally. Although a few CCK fibers were found in the other nerves, CCK fibers to the IMG travel mainly in the colonic nerves. CCK immunoreactive cell bodies have been found in the gastrointestinal tract (Schultzberg et al., 1980) and may project from the colon to the IMG. 7.8.3. Vasoactive intestinal polypeptide

The VIP fibers appear to take several routes to the IMG. The major supply of VIP fibers was found to travel in the hypogastric and colonic nerves. Some VIP fibers also enter via the intermesenteric nerve and a few via the inferior splanchnic nerves. When the colonic or hypogastric nerves were ligated, a heavy accumulation of immunoreactivity was seen distal to the ligature, with a small amount seen proximally. In the intermesenteric nerve there was accumulation on both sides of the ligature and in the inferior splanchnic nerves the accumulation was proximal to the IMG (Dalsgaard et al., 1983a). Thus, the VIP fibers reach the ganglion through several different pathways. Most fibers originate in the distal colon, suggesting yet another peptidegic afferent pathway from the colon to the IMG. Some neurons must also enter through the hypogastric nerve; these probably originate from VIP cells in the pelvic plexus. The fibers entering in the intermesenteric nerve may originate in the C G - S M G since VIP-containing cells have been described there (H6kfelt et al. , 1977b).

8. Conclusions

The above results show that the abdominal prevertebral ganglia cannot be regarded as simple relay stations from the central nervous system to the peripheral visceral organs. The cholinergic inputs to the cells in the CG, SMG and IMG generally require summation to generate an action potential in the postganglionic neuron. Figure 18 is a schematic diagram of the cholinergic and adrenergic pathways of the IMG surmised from the studies cited above. These pathways illustrate the various origins of inputs to these cells and the involvement of peripherally originating information in the input-output relationships of this ganglion. In addition to these pathways, a complement of peptide-containing fibers exists in the prevertebral ganglia. Again using the IMG as an example, Fig. 19 is a schematic of the peptide pathways revealed through the above-mentioned studies. This diagram shows that the presence of peptides, some of which have been shown to have specific actions in the prevertebral ganglia, adds another level of complexity to ganglionic neurotransmission. The exact significance of the presence of most of these peptides and the specific actions of these peptides on ganglionic transmission has not been established, however, some conjecture has been made with regard to the L S - E P S P in the prevertebral ganglia and its possible functional significance. The most obvious implication of a L S - E P S P would be to increase the probability of an action potential being generated following impingement of a F - E P S P onto the postganglionic cell. The depolarization of the L S - E P S P would bring the cell closer to threshold. The increase in Rin would make the current flow generated during a F - E P S P result in a greater voltage change. That this does occur has been demonstrated in the bullfrog preparation. Injection of current pulses which did not cause spikes in the resting state frequently did so during the L S - E P S P (Jan et al., 1979). Further analysis revealed, in addition to the two factors mentioned above, a lowering of the action potential threshold by 1 to 3 mV during a LS-EPSP. Increased excitability also occurs during the S - E P S P (Schulman and Weight, 1976). Since most of the nicotinic cholinergic inputs to these cells

SYNAPTIC TRANSMISSION IN THE PREVERTEBRAL SYMPATHETIC GANGLIA

cord

DRG

we

~re

viscera /

\

I

I

~k-'O

FIG. 18. Schematic diagram of the cholinergic and adrenergic pathways of the IMG. Solid circles indicate populations of ACh-containing cells. Open circles indicate populations of NE-containing cells. Crossed cell is a SIF cell population, innervating blood vessels. Para, paravertebral sympathetic ganglia; pre, prevertebral sympathetic ganglia. Cholinergic neurons in the spinal cord and the viscera project to the IMG cells either directly or after passing through the paravertebral ganglia or other prevertebral ganglia. The adrenergic |MG cells then project to the viscera. Anatomical evidence indicates that the IMG cells also contact each other and that the SIF cells are innervated by both ACh- and NE-containing cells, spinel cord

DRG

pare

pre

viscera

/

FIG. 19. Schematic diagram of the peptidergic pathways of the IMG. Solid circle indicates population of ACh-containing cells. Empty circle indicates population of NE-containing ganglion cells (projections not shown). Para, paravertebral sympathetic ganglia; pre, prevertebral sympathetic ganglia; B, BOM-containing cell population; C, CCK-containing cell population; E, ENK-containing cell population; S, SOM-containing cell population; SP, Spocontaining cell population; V, VIP-containing cell population. ENK-containing preganglionic neurons act presynaptically on ACh-containing cells to inhibit the F-EPSP or on SP-containing afferents to inhibit the LS-EPSP, SP-containing cell bodies in the DRG project through the IMG to the viscera. BOM-, CCK- and VIP-containing cells project from the viscera to the IMG. VIP fibers also originate in the other prevertebral ganglia. SOM-containing cell bodies are present in the prevertebrai ganglia and form synapses within a ganglion.

87

S8

M. A, SIMMONS

are sub-threshold and a summation of these inputs is required to elicit an action polcntial. even a slight increase in excitability would be important to the prevertebral ganglia cells. The summation of sub-threshold F-EPSPs would necessarily involve synchronism within a few msec of converging inputs. During a LS-EPSP the time span of the activity or' the synapse would be increased, thereby increasing the likelihood of temporal summation (Hartzell, 1981 ). The likelihood of spatial summation of inputs would also be increased. A decrease in membrane conductance would increase the space constant of the dendrite: therefore, a given potential would be conducted along the process with less decrement per unit length, increasing spatial summation. In these ways the LS-EPSP could act to modulate the cholinergic potentials and play an important role in the informationprocessing of the prevertebral ganglia. In the bullfrog paravertebral ganglia the LS-EPSP is thought to be mediated by LHRH, in the guinea pig IMG by SP and in the guinea pig CG possibly by 5-HT. This emphasizes the interspecies diversity of synaptic transmission in the sympathetic ganglia. In each of these ganglia the electrical properties of the noncholinergic response was similar; however, the pharmacology of the response is diverse. This may be indicative of a specialization of function with regard to noncholinergic transmission in the different ganglia. Alternatively, it may indicate an evolutionary divergence. In any case, the F-EPSPs in all these ganglia are mediated by ACh and the variable nature of the noncholinergic response remains unexplained. The above data have led to a re-evaluation of our concepts of synaptic transmission in the sympathetic ganglia. These studies have shown that the prevertebral ganglia comprise sophisticated neural networks involved in the integration of neural inputs. Some postsynaptic potentials in these ganglia have been shown to be mediated (in the case of SP) or modulated (in the case of ENK) by peptides in addition to the "classical'" transmitter. ACh. Prevertebral ganglion preparations may serve as suitable models for studying similar phenomena believed to occur in the central nervous system.

Acknowledgements Thanks are extended to N. J. Dun for stimulating this work and to H. C. Hartzell and S. A. Shefner for helpful criticisms. Supported by the Department of Pharmacology, Loyola University Medical Center, Maywood, Illinois; the Arthur J. Schmitt Foundation and National Research Service Award 5F32 NS07315.

References ADAMS,P. R. and BROWN, D. A. (1980) Luteinizing hormone-releasing factor and muscarinic agonists act on the same voltage-sensitive K+-current in bullfrog sympathetic neurons. Br. J. Pharmac. 68, 353-355. ADAMS, P. R., BROWN, D. A. and CONSTANTI, A. (1982) M-currents and other potassium currents in bullfrog sympathetic neurons. J. Physiol., Lond. 330, 537-572. ADRIAN,E. D. (1931) Potential changes in the isolated nervous system of Dytiscus marginalis. J. Physiol.. Lond. 72, 132-151. ADRIAN, E. U., BRONK,D. W. and PHILLIPS,G. (1932) Discharges in mammalian sympathetic nerves. J. Physiol., Lond. 74, 115-133. ARCHAKOVA,L. l., BULYGIN, I. A. and NETUKOVA,N. ]. (t982) The ultrastructural organization of sympathetic ganglia of the cat. J. Autonom. Nerv. Sys. 6, 83-93. BAYL1SS, W. M. and STARLING, E. A. (1900) The movements and the innervation of the large intestine. J. Physiol., Lond. 26, 107-118. B1LLINGSLEY,P. R. and RANSON,S. W. (1918) On the number of nerve cells in the ganglion cervicale superius and of nerve fibers in the cephalic end of the truncus sympatheticus in the cat and on the numerical relations of preganglionic and postganglionic neurons. J. comp. Neurol. 29, 359-366. BLACKMAN,J. G., CROWCROFT, P. J., DEVINE, C. E., HOLMAN,M. E. and YONEMURA,K. (1969) Transmission from preganglionic fibers in the hypogastric nerve to peripheral ganglia in male guinea-pigs. J. Physiol.. Lond. 201,723-743. BROOKS-FOURNIER,R. and COGGESnALL,R. E. (1981) The ratio of preganglionic axons to postganglionic cells in the sympathetic nervous system of the rat. J. eomp. Neurol. 197, 217-236. BROWN, D. A. and ADAMS, P. R. (1980) Muscarinic suppression of a novel voltage-sensitive K + current in a vertebrate neurone. Nature 283, 673-676. BROWN, G. L. and PASCOE,J. E. (1952) Conduction through the inferior mesenteric ganglion of the rabbit. J. Physiol., Lond. 118, 113-123.

SYNAPTIC TRANSMISSION IN THE PREVERTEBRALSYMPATHETICGANGLIA

89

CHRIST, D. D. and NIsnl, S. (1971) Effects of adrenaline on nerve terminals in the superior cervical ganglion of the rabbit. Br. J. Pharmac. 41,331-338. CHRISTENSEN, K., LEWIS, E. and KUNTZ, A. (1951) Innervation of the renal blood vessels in the cat. J. comp. Neurol. 95, 373-385. COOMBS, J. S., CURTIS, D. R. and ECCLES, J. C. (1956) Time course of motoneuronal responses. Nature 178, 1049-1050. CROWCROVr, P. J. and SZURSZEWSrl, J. H. (1971) A study of the inferior mesenteric and pelvic ganglia of guinea-pigs with intracellular electrodes. J. Physiol., Lond. 219, 421-441. CROWCROFF, P. J., HOLMAN, M. E. and SZURSZEWSKi,J. H. (1971a) Excitatory input from the distal colon to the inferior mesenteric ganglion in the guinea-pig. J. Physiol., Lond. 219, 443-461. CROWCROFr, P. J., JARROT, B. and OSTBERG, A. (1971b) Are there inhibitory mechanisms in mammalian autonomic ganglia? Proc. Aust. Physiol. Pharmac. Soc. 2, 31-32. DALE, W. G. JR., EVAN, A. P. JR. and EASON, H. R. (1975) The major ganglion in the pelvic plexus of the male rat. A histochemical and ultrastructurai study. Cell Tiss. Res. 159, 49--62. DALSGAARD, C.-J. and ELFVIN, L.-G. (1979) Spinal origin of preganglionic fibers projecting onto the superior cervical ganglion and inferior mesenteric ganglion of the guinea pig, as demonstrated by the horseradish peroxidase technique. Brain Res. 172, 139-143. DALSGAARD, C.-J. and ELFVlN, L.-G. (1982) Structural studies on the connectivity of the inferior mesenteric ganglion of the guinea pig. J. Autonom. Nerv. Sys. 5, 265-278. DALSGAARD, C.-J., H6KFELT, T., ELFVlN, L.-G., SKIRBOLL, L. and EMSON, P. (1982a) Substance P-containing primary sensory neurons projecting to the inferior mesenteric ganglion: evidence from combined retrograde tracing and immunohistochemistry. Neuroscience 7, 647-654. DALSGAARD, C.-J., H(JKFELT, T., ELFVIN, L.-G. and TERENIUS, L. (1982b) Enkephalin-containing sympathetic preganglionic neurons projecting to the inferior mesenteric ganglion: evidence from combined retrograde tracing and immunohistochemistry. Neuroscience 7, 2039-2050. DALSGAARD, C.-J., H(}KFELT, T., SCHULTZBERG, M., LUNDBERG, J. M., TERENIUS, L., DOCKRAY, G. J. and GOLDSTEIN, M. (1983a) Origin of peptide-containing fibers in the inferior mesenteric ganglion of the guinea pig: immunohistochemical studies with antisera to substance P, enkephalin, vasoactive intestinal polypeptide, cholecystokinin and bombesin. Neuroscience 9, 191-211. DALSGAARD, C.-J., VINCENT, S. R., SCHULTZBERG,M., HOKFELT, T., ELFVIN, L.-G., TERENIUS, L. and DOCKRAY, G. J. (1983b) Capsaicin-induced depletion of substance P-like immunoreactivity in guinea pig sympathetic ganglia. J. Autonom. Nerv. Sys. 9, 595-606. DECrTOR, D. L. and WEEMS, W. A. (1983) An intracellular characterization of neurones and neural connexions within the left coeliac ganglion of cats. J. Physiol., Lond. 341,197-211. DE GROAT, W. C. and KRIER, J. (1979) The central control of the lumbar sympathetic pathway to the large intestine of the cat. J. Physiol., Lond. 289, 449-468. DOCKRAY, G. J., VAILLANT, C. and WALSH, [. H. (1979) Bombesin-like immunoreactivity in the rat gastrointestinal tract. Neuroscience 4, 1561-1568. DUN, N. J. and KARCZMAR,A. G. (1979) Actions of substance P on sympathetic neurons. Neuropharmacology 13,219-223. DUN, N. J. and JIANG, Z. G. (1982) Non-cholinergic excitatory transmission in inferior mesenteric ganglia of the guinea pig: possible mediation by substance P. J. Physiol., Lond. 325, 145-159. DUN, N. J. and KIRALY, M. (1983) Capsaicin causes release of a substance P-like peptide in guinea-pig inferior mesenteric ganglia. J. Physiol.. Lond. 340, 107-120. DUN, N. J. and MA, R. C. (1984) Slow non-cholinergic excitatory potentials in neurones of the guinea-pig coeliac ganglia. J. Physiol., Lond. 351, 47-60. DUN, N. J. and MINOTA, S. (1981) Effects of substance P on neurones of the inferior mesenteric ganglia of the guinea pig. J. Physiol., Lond. 321,259-271. DUN, N. J., KIRALY, M. and MA, R. C, (1984) Evidence for a serotonin mediated slow excitatory potential in the guinea-pig coeliac ganglia. J. Physiol., Lond. 351, 61-76. EBBESSON, S. O. E. (1963) A quantitative study of human superior cervical sympathetic ganglia. Anat. Rec. 146, 353-356. EBBESSON, S. O. E. (1968) Quantitative studies of superior cervical sympathetic ganglia in a variety of primates including man. I. The ratio of preganglionic fibers to ganglionic neurons. J. Morphol. 124, 117-132. ECCLES, J. C. (1935) The ,~actionpotential of the superior cervical ganglion. J. Physiol., Lond. 85, 179-206. ELFVIN, L.-G. (1968) A new granule-containing nerve cell in the inferior mesenteric ganglion of the rabbit. J. Ultrastruct. Res. 22, 37-44. ELEVIN, L.-G. (1971a) UItrastructural studies on the synaptology of the inferior mesenteric ganglion of the cat. I. Observations on the cell surface of the postganglionic perikarya. J. Ultrastruct. Res. 37, 411-425. E LFVIN, L.-G. ( 1971 b) Ultrastructural studies on the synaptology of the inferior mesenteric ganglion of the cat. I I Specialized serial neuronal contacts between preganglionic end fibers. J. Ultrastruct. Res. 37,426-431. EI.FVIN, L.-G. (1971c) Ultrastructural studies on the synaptology of the inferior mesenteric ganglion of the cat. IlI. The structure and distribution of the axodendritic and dendrodendritic contacts. J. Ultrastruct. Res. 37, 432-448. ELFWN, L.-G. and DALSGAARD,C.-J. (1977) Retrograde axonal transport of horseradish peroxidase in afferent fibers of the inferior mesenteric ganglion of the guinea pig. Identification of the cells of origin in dorsal root ganglia. Brain Res. 126, 149-153. E LFVIN, L.-G., HOKFELT, T. and GOLDSTEIN, M. (1975) Fluorescence microscopical, immunohistochemical and ultrastructural studies on sympathetic ganglia of the guinea pig, with special reference to the SIF cells and their catecholamine content. J. Ultrastruct. Res. 51,377-396. FALCK, B. (1962) Observations on the possibilities of the cellular localization of monoamines by a fluorescence method. Acta physiol, scand. 56, Suppl. 197.

90

M . A . SIMMONS

FAIT, P. and KArZ, B. (1951) An analysis of the end-plate potential recorded with an intracellular electrode. J. Physiol., Lond. 115,320-37[L FERRY, C. B. (1967) The innervation of the vas deferens of the guinea-pig. J. Physiol., Lond. 192,463--478. FURNESS, J. B. and Cos~'A, M. (1974) The adrenergic innervation of the gastrointestinal tract. Ergebn. Phystol. 69, 1-51. FURNESS, J. B. and SOBELS, G. (1976) The ultrastructure of paraganglia associated with the inferior mesenteric ganglia in the guinea-pig. Cell Tiss. Res. 171,123-129. GABEEt.A, G. (1976) Structure o f the Autonomic Nervous System. Wiley, New York. GAMSE, R., LEEMAN, S. E., HOLZER,P. and LEMBECK,F. (1981a) Differential effects of capsaicin on the content of somatostatin, substance P, and neurotensin in the nervous system of the rat. Naunyn Schmied. Arch. Pharmac. 317, 140-148. GAMSE, R., WAX, A., ZIGMOND, R. E. and LEEMAN, S. E. (1981b) Immunoreactive substance P in sympathetic ganglia: distribution and sensitivity towards capsaicin. Neuroscience 6, 437-441. GARRY, R. C. (1933a) The nervous control of the caudal region of the large bowel in the cat. J. Physiol., Lond. 77,422-431. GARRY, R. C. (1933b) The response to stimulation of the caudal end of the large bowel in the cat. J. Physiol., Lond. 78,208-224. GARRY, R. C. and GILLESPIE, J. S. (1955) The responses of the musculature of the colon of the rabbit to stimulation, in vitro, of the parasympathetic and of the sympathetic outflows. J. Physiol., Lond. 128, 557-576. GASKELL,W. H. (1916) The Involuntary Nervous System. Longmans Green, London. GILLESPIE,J. S. and MACKENNA,B. R. (1961) The inhibitory action of the sympathetic nerves on the smooth muscle of the rabbit gut, its reversal by reserpine and restoration by catecholamines and by dopa. J. Physiol., Lond. 156, 17-34. HAMBURGER,B. and NORBERG, K.-A. (1965) Studies on some systems of adrenergic synaptic terminals in the abdominal ganglia of the cat. Acta physiol, scand. 65, 235-242. HAMBURGER,B., NORBERG, K.-A. and SJOQVlST,F. (1964) Evidence for adrenergic nerve terminals and synapses in sympathetic ganglia. Int. J. Neuropharmac. 2, 279-282. HAMBURGER, B., NORBERG, K.-A. and UNGERSTEDT, U. (1965) Adrenergic synaptic terminals in autonomic ganglia. Acta physiol, scand. 64, 285-286. HARRIS, A. J. (1943) An experimental analysis of the inferior mesenteric plexus. J. comp. Neurol. 79, 1-17. HARTMAN,D. A. and KRIER, J. (1984) Synaptic and antidromic potentials of visceral neurons in ganglia of the lumbar sympathetic chain of the cat. J. Physiol., Lond. 350, 413-428. HARTZELL, H. C. (1981) Mechanisms of slow postsynaptic potentials. Nature 291,539-544. HOKFELT, T., ELFV1N, L.-G., ELDE, R., SCHULTZBERG,M., GOLDSTEtN, M. and LuFr, R. (1977a) Occurrence of somatostatin-like immunoreactivity in some peripheral sympathetic noradrenergic neurons. Proc. natn. Acad. Sci. U.S.A. 74, 3587-3591. HOKFELT, T., ELFVIN, L.-G., SCHULTZBERG,M., FUXE, K., SAID,S. I., MUTT, V. and GOLDSTEIN,M. (1977b) lmmunohistochemical evidence of vasoactive intestinal polypeptide-containing neurons and nerve fibers in sympathetic ganglia. Neuroscience 2,885-896. H6KFELT, T., ELFVIN, L.-G., SCHULTZBERG,M., GOLDSTEIN, M. and N1LSSON, G. (1977c) On the occurrence of substance P-containing fibers in sympathetic ganglia: immunohistochemical evidence. Brain Res. 132, 29-41. H(JKFELT, T., LUNDBERG,J. M., SCHULTZBERG,M., JOHANSSON, O., SKIRBOLL,L., ANGGARD,g . , FREDHOLM, B., HAMBURGER, B., PERNOW, B., REHFELD, J. and GOLDSTEIN, M. (1980) Cellular localization of peptides in neural structures. Proc. R. Soc. Lond. B 210, 63-77. HOLMAN, M. E., MUIR, T. C., SZURSZEWSKI,J. H. and YONEMURA, K. (1971) Effect of iontophoretie application of cholinergic agonists and their antagonists to guinea-pig pelvic ganglia. Br. J. Pharmae. 41, 26-40. JAN, L. Y. and JAN, Y. N. (1982) Peptidergic transmission in sympathetic ganglia of the frog. J. Physiol., Lond. 327, 219-246. JAN, Y. N., JAN, L. Y. and KUFFLER(1979) A peptide as a possible transmitter in sympathetic ganglia of the frog. Proc. natn. Acad. Sci. U.S.A. 76, 1501-1505. JAN, xf-.N., JAN, L. Y. and KUFFLER(1980) Further evidence for peptidergic transmission in sympathetic ganglia. Proc. natn. Acad. Sci. U.S.A. 77, 5008-5012. JANSSON, G. (1969)Extrinsic nervous control of gastric motility. Acta physiol, scand. Suppl. 326, 1-42. JESSELL, T. M., IVERSEN, L. L. and CUELLO, A. C. (1978) Capsaicin-induced depletion of substance P from primary sensory neurons. Brain Res. 152, 183-188. J1ANG, Z. G., DUN, N. J. and KARCZMAR, A. G. (1982a) Substance P: a putative sensory transmitter in mammalian autonomic ganglia. Science 217,739-741. J1ANG, Z. G., SIMMONS, M. A. and DUN, N. J. (1982b) Enkephalinergic modulation of non-cholinergic transmission in mammalian prevertebral ganglia. Brain Res. 235, 185-191. JoB, C. and LUNDBERG, A. (1952) Reflex excitation of cells in the inferior mesenteric ganglion on stimulation of the hypogastric nerve. Acta physiol, scand. 26, 366-382. JOHANSSON, B. and LANGSTON, J. B. (1964) Reflex influence of mesenteric afferents on renal, intestinal and muscle blood flow and on intestinal motility. Acta physiol, scand. 61,400-412. JULL Y. and SZURSZEWSKI,J. H. (1983) Electrophysiology of neurones of the inferior mesenteric ganglion of the cat. J. Physiol., Lond. 344, 277-292. JULY, Y., KR1ER, J. and SZURSZEWSKI,J. H. (1983) Patterns of innervation of neurones in the inferior mesenteric ganglion of the cat. J. Physiol., Lond. 344, 293-304. KATAYAMA,Y. and NORTH, R. A. (1978) Does substance P mediate slow synaptic excitation within the myenteric plexus? Nature 274, 387-388. KATAYAMA, Y. and NISHI, S. (1982) Voltage-clamp analysis of peptidergic slow depolarizations in bullfrog sympathetic ganglion cells. J. Physiol., Lond. 333, 305-313.

SYNAPTICTRANSMISSION1N THE PREVERTEBRALSYMPATHETICGANGLIA

91

KELTS, K. A., WHrrLOCK, D. G., LEDBURY, P. A. and REESE, B. A. (1979) Postganglionic connections between sympathetic ganglia in the solar plexus of the cat demonstrated autoradiographicaily. Expl Neurol. 63, 120-134. KESSLER, J. A., ADLER, J. E., BOHN, M. C. and BLACK, I. B. (1981) Substance Pin principal sympathetic neurons: regulation by impulse activity. Science 214, 335-336. KING, B. F. and SZURSZEWSKI,J. H. (1984a) An electrophysiological study of inferior mesenteric ganglion of the dog. J. Neurophysiol. 51,607-615. KING, B. F. and SZURSZEWSKI,J. H. (1984b) Mechanoreceptor pathways from the distal colon to the autonomic nervous system in the guinea-pig. J. Physiol., Lond. 350, 93-107. KLEINHAUS, S., BOLEY, S. J., SHERAN, M. and SmBER, W. K. (1979) Hirschprung's disease: a survey of the members of the surgical section of the American Academy of Pediatrics. J. Ped. Surg. 14, 588-597. KONDO, H. and YuI, R. (1982) An electron microscopic study on VIP-like immunoreactive nerve fibers in the celiac ganglion of guinea pigs. Brain Res. 237, 227-231. KoNIsm, S., TSUNOO, A. and OTsur~, M. (1979a) Enkephalins presynaptically inhibit cholinergic transmission in sympathetic ganglia. Nature 282, 515-516. KoNism, S,, TSUNOO, A. and Orsor~, M. (1979b) Substance P and noncholinergic excitatory synaptic transmission in guinea pig sympathetic ganglia. Proc. Japan Acad. 55, 525-530. KONISHI, S., OrSUKA, M., FOLKERS, K. and ROSELL, S. (1983) A substance P antagonist blocks non-cholinergic slow excitatory potential in guinea-pig sympathetic ganglia. Actaphysiol. scand. 117, 157-160. KREULEN, D. L. and SZURSZEWSKI,J. H. (1979a) Electrophysiologic and morphologic basis for organization of neurons in prevertebral ganglia. In: Frontiers o f Knowledge in the Diarrheal Diseases, pp. 211-226. Eds. H. D. JANOWrrZ and D. B. SACHAR.Projects in Health, Upper Montclair, New York. KREULEN, D. L. and SZURSZEWSKI, J. H. (1979b) Nerve pathways in celiac plexus of the guinea pig. Am. J. Physiol. 237, E90-E97. KREULEN, D. L. and SZUESZEWSK1, J. H. (1979c) Reflex pathways in the abdominal prevertebral ganglia: evidence for a colD-colonic inhibitory reflex. J. Physiol., Lond. 295, 21-32. KRIER, J., SCHMALZ, P. F. and SZURSZEWSKI, J. H. (1982) Central innervation of neurones in the inferior mesenteric ganglion and of the large intestine of the cat. J. Physiol., Lond. 332, 125-138. KRNJEVfC, K. (1977) Effects of substance P on central neurons in cats. In: Substance P, pp. 217-230. Eds. U. S. YON EULER and B. PERNOW. Raven Press, New York. KuRA, K. and KOKETSU, K. (1978) Synaptic events in sympathetic ganglia. Prog. Neurobiol. 11, 77-169. KUFFLER, S. W. and SEJNOWSm, T. J. (1983) Peptidergic and muscarinic excitation at amphibian sympathetic synapses. J. Physiol., Lond. 341,257-278. KUNTZ, A. J. (1938) The structural organization of the celiac ganglia. J. comp. Neurol. 69, 1-12. KUNTZ, A. J. (1940) The structural organization of the inferior mesenteric ganglia. J. comp. Neurol. 72,371-382. KUNTZ, A. and MOSELEY, R. L. (1936) An experimental analysis of the pelvic autonomic ganglia in the cat. J. comp. Neurol. 64, 63-75. KUNTZ, A. and VAN BtJSKmK, C. (1941) Reflex inhibition of bile flow and intestinal motility mediated through decentralized celiac plexus. Proc. Soc. expl. Biol. 46, 519-523. KUNTZ, A. J. and SACCOMANNO,G. (1944) Reflex inhibition of intestinal motility mediated through decentralized prevertebral ganglia. J. Neurophysiol. 7, 163-170. KUNTZ, A. and JACOBS, M. W. (1955) Components of periarterial extensions of celiac and mesenteric plexuses. Anat. Rec. 123, 509-520. Kuo, D. C. and KRAUTHAMER,G. M. (1981) Paravertebral origin of postganglionic sympathetic fibers in the major splanchnic and distal coeliac nerves as demonstrated by horseradish peroxidase (hrp) retrograde transport method. J. Autonom. Nerv. Sys. 4, 25-32. LANGLEY, J. N. (1921) The Autonomic Nervous System. Heffer, Cambridge, Mass. LANGLEY, J. N. and ANDERSON, H. K. (1895a) On the innervation of the pelvic and adjoining viscera. Part I. The lower portion of the intestine. J. Physiol., Lond. 18, 67-105. LANGLEY,J. N. and ANDERSON, H. K. (1895b) On the innervation of the pelvic and adjoining viscera. Part II. The bladder. J. Physiol., Lond. 19, 71-84. LANGLEY, J. N. and ANDERSON, H. K. (1895c) On the innervation of the pelvic and adjoining viscera. Part IV. The internal generative organs. J. Physiol., Lond. 19, 122-130. LANGLEY, J. N. and ANDERSON, H. K. (1896a) On the innervation of the pelvic and adjoining viscera. Part VI. Histological and physiological observations upon the effects of section of the sacral nerves. J. Physiol., Lond. 19,372-384. LANGLEY, J. N. and ANDERSON, H. K. (1896b) On the innervation of the pelvic and adjoining viscera. Part VII. Anatomical observations. J. Physiol., Lond. 20, 372-406. LANGWORTHY, O. R. (1965) Innervation of the pelvic organs of the rat. Invest. Urol. 2, 491-511. LAWSON, H. (1934) The role of the inferior mesenteric ganglia in the diphasic response of the colon to sympathetic stimuli. Am. J. Physiol. 109, 257-273. LAWSON, H. and HOLT, J. P. (1937) The control of the large intestine by the decentralized inferior mesenteric ganglion. Am. J. Physiol. 118, 780-785. LEARMONTH, J. R. and MARKOWITZ,J. (1929) Studies on the function of the lumbar sympathetic outflow. I. The relation of the lumbar sympathetic outflow to the sphincter ani internus. Am. J. Physiol. 89, 686--691. LEARMONTH, J. R. and MARKOWITZ,J. (1930) Studies on the innervation of the large bowel. II. The influence of the lumbar colonic nerves on the distal portion of the colon. Am. J. Physiol. 94, 501-504. LI~I~NTH, C., WILLIAMS,T. H., JEW, J. Y. and ARIMURA, A. (1980) Immuno-electron microscopic identification of somatostatin in cells and axons of sympathetic ganglia in the guinea pig. Cell Tiss. Res. 212, 83-89. LIBET, B. (1970) Generation of slow inhibitory and excitatory postsynaptic potentials. Fed. Proc. 29, 1945-1956. LIBET, B. and TOSAKA, T. (1966) Slow postsynaptic potentials recorded intracellularly in sympathetic ganglia. Fed. Proc. 25, 270.

92

M . A . SIMMONS

LLOYD, D. P. C. (1937) The transmission of impulses through the inferior mesenteric ganglia. J. Physiol.. Lond. 91,296-313. LUNDBERG, J. M., HOKFELT, T., SCHULTZBERG,M., UVNAS-WALLENSTEN,K., KOHLER. C. and SAID, S. t. B. (1979) Occurrence of vasoactive intestinal polypeptide (VIP)-like immunoreactivity in certain cholinergic neurons of the cat. Evidence from combined immunohistochemistry and acetylcholinesterase staining. Neuroscience 4, 1539-1559. LUNDBERG, J. M., HOKFELE,T., JStNGGARD,A., KIMMEL,J., GOLDSTEIN,M. and MARKEY,K. (1980) Coexistence of an avian pancreatic polypeptide (APP) immunoreactive substance and catecholamines in some peripheral and central neurons. Acta physiol, scand. 110, 107-109. LUNDBERG, J, M.. HOKFELT, T., fitNGGARD, A., TERENIUS, L., ELDE, R., MARKEY, K., GOLDSTE1N, M. and KIMMEL,J. (1982a) Organizational principles in the peripheral sympathetic nervous system: subdivision by coexisting peptides (somatostatin-, avian pancreatic polypeptide- and vasoactive intestinal polypeptide-like immunoreactive materials). Proc. hath. Acad. Sci. U.S.A. 79, 1303-1307. LUNDBERG,J. M., TERENIUS, L., HOKFELT,T., MARTLING,C. R., TATEMOTO,K., Mutt, V., POLAK,J., BLOOM,S. and GOLDSrEIN, M. (1982b) Neuropeptide Y (NPY)-like immunoreactivity in peripheral nonadrenergic neurons and effects of NPY on sympathetic function. Acta physiol, scand. 116, 477-480. LUNDBERG,J. M., TEREN1US,L., HOKFELT,T. and TATEMOTO,K. (1984) Comparative immunohistochemical and biochemical analysis of pancreatic polypeptide-like peptides with special reference to presence of neuropeptidc Y in central and peripheral neurons. J. Neuroscience 4, 2376-2386. MA, R. C., DUN, N. J. and JIANG, Z. G. (1983)Evidence of slow IPSP in mammalian prevertebral ganglia. Brain Res. 270, 350-354. MA~HEWS, M. R. and CUELLO, A. C. (1982) Substance P-immunoreactive peripheral branches of sensory neurons innervate guinea pig sympathetic neurons. Proc. natn. Acad. Sci, U.S.A. 79, 1668-1672. MCLENNAN, H. and PASCOE,J. E. (1954) The origin of certain non-medullated nerve fibers which form synapses in the inferior mesenteric ganglion of the rabbit. J. Physiol., Lond. 124, 145-156. M'FADDEN. G. D. F., LOUGHRIDGE,J. S. and MILROY,T. H. (1935) The nerve control of the distal colon. Q. J. exp. Physiol. 25,315-327. MITCHE1.E, G. A. G. (1953) Anatomy o f the Autonomic Nervous System. Livingstone Ltd., London. NEILD, T. O. (1978) Slowly-developing depolarization of neurons in the guinea-pig inferior mesenteric ganglion following repetitive stimulation of the preganglionic nerves. Brain Res. 1411,231-239. N~COEL, R. A. (1978) The action of thyrotropin-releasing hormone, substance P and related peptides on frog spinal motoneurons. J. Pharmac. exp. Ther. 207,817--824. NISHI, S. and KOKETSU, K. (1968) Early and late afterdischarges of amphibian sympathetic ganglion cells. J. Neurophysiol. 31,109-121. NJA, A. and PURVES,D. (1977) Specific innervation of guinea-pig superior cervical ganglion cells by preganglionic fibres arising from different levels of the spinal cord. J. Physiol., Lond. 264, 565-583. OSCARSSON,O. (1955) On the functional organization of the two presynaptic systems to the colonic nerve neurons of the inferior mesenteric ganglion of the cat. Acta physiol, stand. 35, 153-166. OTSUKA. M. and KoNmnh K. (1977) Electrophysiological and neurochemical evidence for substance P as a transmitter of primary sensory neurons. In: Substance P, pp. 207-214. Eds. 1.1. S. voN EULERand B. PERNOW. Raven Press. New York. OWMAN. C., ALM, P. and SJOBERG,N.-O. (1983) Pelvic autonomic ganglia: structure, transmitters, function and steroid influence. In: Autonomic Ganglia, pp. 125-143. Ed. L.-G. EEFVIN. Wiley, New York. PERRI, V., SACCHI,O. and CASEEt,A,C. (1970) Electrical properties and synaptic connections of the sympathetic neurons in the rat and guinea-pig superior cervical ganglion. Pfliigers Arch. 314, 40-54. PETERS, S. and KREULEN,D. L. (1984) A slow EPSP in mammalian inferior mesenteric ganglion persists after in vivo capsaicin. Brain Res. 303, 186-189. PUR1NGTON, T., FI,ErCHER, T. and BBADLEV,W. (1973) Gross and light microscopic features of the pelvic plexus in the rat. Acta Anat. 175,697-706. PURVES, D. and HUME, R. I. (1981) The relation of postsynaptic geometry to the number of presynaptic axons that innervate autonomic ganglion cells. J. Neurosci. 1,441-452. RANKtN, F. W. and LEARMONTH,J. R. (1930) Section of the sympathetic nerves of the distal part of the colon and the rectum in the treatment of Hirschprung's disease and certain types of constipation. Ann. Surg. 92, 710-720. RANKIN, F. W. and LEARMONTH,J. R. (1932) The present status of the treatment of Hirschprung's disease. Am. J. Surg. 15,219-227. Ross, J. P. (1935) The results of sympathectomy: an analysis of the cases reported by fellows of the association of surgeons. Br. J. Surg. 23,433-443. SANB|:~, S. (1961) Studies on the incoming and outgoing myelinated fibers of the sympathetic superior cervical ganglion. Fukushima J. med. Sci. 8, 10%129. SCHULMAN, J. A. and WEIGHT, F. F. (1976) Synaptic transmission: long lasting potentiation by a postsynaptic mechanism. Science 194, 1437-1439. SCHULTZBERC,,M. (1983) Bombesin-like immunoreactivity in sympathetic ganglia. Neuroscience 8, 363-374. SCHULTZBERG, M., HOKFELT, T., LUNDBERG, J. M., TERENIUS, L., ELIWIN, L.-G. and ELDE, R. (1978) Enkephalin-like immunoreactivity in nerve terminals in sympathetic ganglia and adrenal medulla and in adrenal medullary gland cells. Acta physiol, scand. 103, 475-477. SCHULTZBERG,M., HOKFELr,T., TERENIUS,L., ELFVIN,L.-G., LUNDBERG,J. M., BRANDT,J., ELDE, R: P. and GOLDSTEIN,M. (1979) Enkephalin immunoreactive nerve fibres and cell bodies in sympathetic ganglia of the guinea pig and rat. Neuroscience 4, 249-270. SCHULTZBERG, M., HOKFELT,T., NILSSON, G., TEREN1US, L., REHFELD, J. F., BROWN, M., ELDE, R. P., GOLDSTEIN, M. and SAID, S. (1980) Distribution of peptide- and catecholamine-containing neurons in the gastrointestinal tract of rat and guinea-pig: immunohistochemical studies with antisera to substance P,

SYNAPTIC TRANSMISSION IN THE PREVERTEBRALSYMPATHETICGANGLIA

93

vasoactive intestinal polypeptide, enkephalin, somatostatin, gastrin/cholecystokinin, neurotensin and dopamine-/3-hydroxylase. Neuroscience 5,689-744. SEMBA, T. (1954) lntestino-intestinal inhibitory reflexes. Japan. J. Physiol. 4, 241-245. SIMMONS, M. A. and DUN, N. J. (1985) Synaptic transmission in the rabbit inferior mesenteric ganglion. Submitted for publication. SJ0STRAND, N. O. (1965) The adrenergic innervation of the vas deferens and the accessory male genital gland. Acta physiol, scand. 65, Suppl. 257. SKOK, V. I. (1973) Physiology o f Autonomic Ganglia. Igaku Shoin, Tokyo. SYROMYATNIKOV,A. V. and SKOK, V. 1. (1968) Conducting pathways of sympathetic ganglia of the cat solar plexus. Neurosci. Transl. 8, 918--929. SZURSZEWSKI, J. H. and WEEMS, W. A. (1976) A study of peripheral input to and its control by post-ganglionic neurones of the inferior mesenteric ganglion. J. Physiol., Lond. 256, 541-556. TATEMOTO, K. (1982) Neuropeptide Y: Complete amino acid sequence of the brain peptide. Proc. hath. Acad. Sci. U.S.A. 79, 5485-5489. TAXI, J., DERER, M. and DOMICH, A. (1983) Morphology and histophysiolbgy of SIF cells in the autonomic ganglia. In: Autonomic Ganglia, pp. 67-95. Ed. L.-G. ELFVIN. Wiley, New York. TlSCORNIA, O. M. (1977) The neural control of exocrine and endocrine pancreas. Am. J. Gastroent. 67, 541-560. TRUMBLE, H. C. (1933) The plan of the visceral nerves in the lumbar and sacral outflows of the autonomic nervous system. Br. J. Surg. 21,664-676. TSUNOO, A., KONISHI, S. and OTSUKA, M. (1982) Substance P as an excitatory transmitter of primary afferent neurons in guinea-pig sympathetic ganglia. Neuroscience 7, 2025-2037. WALLIS, D. I. and NORTH, R. A. (1978) Synaptic input to cells of the rabbit superior cervical ganglion. Pfl~gers Arch. 374, 145-152. WATANABE,H. (1971) Adrenergic nerve elements in the hypogastric ganglion of the guinea pig. Am. J. Anat. 130, 305-330. WEBBER, R. H. (1955) Analysis of the sympathetic trunks, communicating rami, sympathetic roots and visceral rami in the lumbar region in man. Ann. Surg. 141,398-413. WEEMS. W. A. and SZURSZEWSK1,J. H. (1977) Modulation of colonic motility by peripheral neural inputs to neurons of the inferior mesenteric ganglion. Gastroenterology 73, 273-278. WILLIAMS,T. and JEW, J. (1983) Monoamine connections in sympathetic ganglia. In: Autonomic Ganglia, pp. 235-264. Ed. L.-G. ELFVIN. Wiley, New York. WOLF, G. A. JR. (1941) The ratio of preganglionic neurons to postganglionic neurons in the visceral nervous system. J. comp. Neurol. 75, 235-243. WO~NIAK,W. and SKOWROI(ISKA,U, (1967) Comparative anatomy of pelvic plexus in cat, dog, rabbit, macaque and man. Anat. Anz. 120, 457-473. YOUMANS, W. B., KARSTENS, A. I. and AUMANN, K. W. (1942) Nervous pathways for the reflex regulation of intestinal pressure. Am. J. Physiol. 135,619-627.