- Email: [email protected]
Morphology and ultrastructure of the sympathetic celiac ganglion neurons projecting to the cardia and pylorus of the rat stomach
Autonomic Neuroscience: Basic and Clinical 134 (2007) 8 – 17 www.elsevier.com/locate/autneu
Morphology and ultrastructure of the sympathetic celiac ganglion neurons projecting to the cardia and pylorus of the rat stomach Tetsu Hayakawa ⁎, Sachi Kuwahara, Seishi Maeda, Koichi Tanaka, Makoto Seki Department of Anatomy, Hyogo College of Medicine, Mukogawa, Nishinomiya, Hyogo 663-8501, Japan Received 27 October 2006; received in revised form 19 January 2007; accepted 27 January 2007
Abstract The stomach receives sympathetic projections from the celiac ganglion. To determine what kinds of neurons in the celiac ganglion project to the cardia or the pylorus of the stomach, we injected the retrograde tracer Fluoro-Gold into the cardia and the retrograde tracer cholera toxin subunit b into the pylorus of the same animal. A few neurons (about 10%) innervating the cardia sent collateral projections to the pylorus. Ultrastructural observations revealed that the celiac ganglion contained oval, medium-sized to large neurons. They had a dark cytoplasm containing numerous free ribosomes, rough endoplasmic reticulum, mitochondria, lysosomes, several Golgi apparatuses, and an oval nucleus. The axon terminals were small and usually contacted thin processes extending from the dendrites or the soma. About half of the terminals contained round vesicles, while the rest contained pleomorphic vesicles. Both types of terminals made asymmetric synaptic contacts. We then retrogradely labeled the neurons projecting to the cardia and the pylorus with wheat germ agglutinin conjugated horseradish peroxidase to examine their ultrastructural characteristics. The neurons projecting to the cardia (33.3 × 22.4 μm) were similar to the neurons projecting to the pylorus (33.4 × 24.7 μm) in their size and ultrastructural appearance. The neurons not projecting to the stomach (40.4 × 28.0 μm) were significantly larger than the neurons projecting to the cardia or the pylorus. Only a few axosomatic terminals were found on the neurons projecting to the cardia (1.6 per somatic profile), the pylorus (1.3) or the neurons not projecting to the stomach (0.9). These results provide morphological bases for the sympathetic motor neurons innervating the stomach. © 2007 Elsevier B.V. All rights reserved. Keywords: Sympathetic neurons; Gastric motility; Retrograde labeling; Synaptic organization; Electron microscopy
1. Introduction The neurons of the sympathetic celiac ganglion innervate the stomach and regulate the motility of the stomach wall. The celiac ganglion of the rat lies on the ventral wall of the abdominal aorta just rostral to the root of the superior mesenteric artery, and joins the superior mesenteric ganglia to form the celiac superior mesenteric ganglion complex. Light microscopic observation revealed that the neurons of the celiac ganglion are large, round in shape with several dendrites, and surrounded by satellite cells (Szurszewski and Miller, 1994). Ultrastructural studies of the prevertebral ⁎ Corresponding author. Tel.: +81 798 45 6484; fax: +81 798 45 6485. E-mail address: [email protected] (T. Hayakawa). 1566-0702/$ - see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.autneu.2007.01.011
sympathetic neurons including the celiac, superior mesenteric and inferior mesenteric ganglia of the mouse (Miller et al., 1996), the cat (Elfvin, 1971a), and the guinea pig (Gibbins et al., 2003) showed that the sympathetic neurons are large, and round in shape with four to five primary dendrites. The soma and primary dendrites are surrounded by satellite cells and have only a few axosomatic terminals. Sometimes short fine processes (accessory dendrites) extend from the soma, and make synaptic contacts with axon terminals. In the neuropil, there are many axodendritic terminals containing round or pleomorphic synaptic vesicles. Axoaxonic contacts and dendrodendritic contacts are present in the cat (Elfvin, 1971c) and the mouse (Miller et al., 1996). The ultrastructural features and synaptic organization of the celiac ganglia of the mouse are somewhat different from the
T. Hayakawa et al. / Autonomic Neuroscience: Basic and Clinical 134 (2007) 8–17
other species that have been studied. However, there have been few studies about the ultrastructure and synaptic organization of the rat celiac ganglion neurons. Immunohistochemical studies of the celiac ganglion (Macrae et al., 1986; Jobling and Gibbins, 1999) revealed that almost all neurons contain noradrenaline (NA). Three major subpopulations of noradrenergic neurons have been described in the guinea pig celiac ganglion, that is, NA neurons that also contain somatostatin (SOM), NA neurons that also contain neuropeptide Y (NPY), and NA neurons that do not contain either SOM or NPY (Lundberg et al., 1982; Macrae et al., 1986; Szurszewski and Miller, 1994). The celiac ganglion neurons receive many fibers from the intermediolateral nucleus of the thoracic spinal cord, the dorsal root ganglion, and the intestinofugal neurons of the myenteric ganglia in the gastrointestinal tract (Kuramoto and Furness, 1989; Messenger and Furness, 1992; Domoto et al., 1995). These fibers and terminals contain several kinds of peptides such as vasoactive intestinal peptide (VIP), cholecystokinin (CCK), enkephalins (ENK), dynorphin (DYN), neurotensin (NT), calcitonin gene-related peptide (CGRP), and substance P (SP) in addition to nitric oxide (NO) and acetylcholine (Kondo and Yui, 1981, 1982a,b; Kondo and Yamamoto, 1988; Elfvin et al., 1993; Szurszewski and Miller, 1994). The axon terminals contain small clear vesicles, small pleomorphic vesicles, and small or large granular vesicles (Smolen, 1988). However, it is not clear what kind of axon terminals contact the celiac ganglion neurons that innervate the stomach of the rat. Physiological studies have revealed that stimulation of sympathetic innervation of the pylorus elicits a contraction of sphincter muscles, while stimulation of the splanchnic nerve inhibits the gastric contraction of the fundus and the cardia of the stomach (Roman and Gonella, 1981). A retrograde tracing study using two kinds of tracers showed that the neurons of the dorsal motor nucleus of the vagus (DMV) send collateral projections to the cardia and the corpus, whereas the cardia and the pylorus received projections mostly from different neurons in the DMV (Hayakawa et al., 2003). Thus, it is likely that different groups of neurons in the celiac ganglion project to and have different effects on the cardia and the pylorus. In the present study, we attempted to clarify whether the neurons projecting to the cardia of the stomach in the celiac ganglia have collateral projections to the pylorus, using two kinds of retrograde tracers, Fluoro-Gold (FG) and cholera toxin subunit b (CTb). We also identified the neurons projecting to the cardia and the pylorus by retrograde labeling with wheat germ agglutinin conjugated horseradish peroxidase (WGA-HRP), and determined their ultrastructural features and synaptic organization with the electron microscope. 2. Materials and methods We used sixteen male Sprague–Dawley rats weighing 250– 300 g. All surgical procedures were carried out with the animals
9
under sodium pentobarbital anesthesia (50 mg/kg, i.p.). The Animal Care and Use Committee of Hyogo College of Medicine approved the procedures. The abdominal wall was opened and the liver was pushed aside to expose the stomach. Using a glass micropipette (tip diameter 80 μm) affixed to a 10-μl Hamilton syringe, four aliquots of 2 μl each of 1% FG (Fluorochrome LLC, Denver CO) in distilled water were injected into the wall of the ventral surface of cardia. Then four aliquots of 2 μl each of 2% CTb (List Biological Laboratories, Campbell CA) in 0.1 M phosphate buffer (PB) at pH 7.4 were injected by pressure into the wall of the ventral surface of the pylorus (5 cases) in the same animal. Cotton swabs were used during the injections to prevent the spread of the tracer to adjacent structures. Three days after the injections, the animals were anesthetized again and perfused first with 100 ml of saline and then with 500 ml of 4% paraformaldehyde-15% picric acid in PB (pH 7.4). The celiac ganglion together with the abdominal aorta was immediately removed and placed in the same fixative for 1 h. Serial frozen sections were cut at 40 μm, and processed for immunohistochemistry. After a rinse with PB (pH 7.4), the sections were incubated with 1% bovine serum albumin in PB at pH 7.4 containing 0.9% NaCl and 0.3% Triton X-100 (PBST) for 1 h at room temperature. The sections were then incubated with a goat anti-CTb serum (List Biological Laboratories; 1:20,000 dilution) for 1 day at 4 °C. The primary antibodies were localized by incubation with Cy3-conjugated donkey anti-goat IgG (Jackson, West Grove PA; 1:400 dilution) for 5 h at room temperature. After rinsing with PBST, the sections were mounted onto gelatin-coated slides and dried. FG was viewed with a U excitation filter (blue), and Cy3 with a G excitation filter (red). Fluorescence photomicrographs were taken using an Olympus AX80. We made merged photographs using Photoshop® CS, and observed blue-green FG-, red CTb- and double-labeled (yellow) neurons. Then, we counted the labeled neurons in alternate serial sections of the ganglion to avoid double counting, and calculated the percentage of the double-labeled neurons for the FG- or the CTb-labeled neurons. To investigate the ultrastructure of the celiac ganglion neurons, three rats were anesthetized and perfused first with 100 ml of saline and then with 500 ml of 1% glutaraldehyde1% paraformaldehyde in PB (pH7.4). The celiac ganglion together with the abdominal aorta was removed and immersed in the same fixative. Serial frontal sections were cut at 100 μm with a Vibratome. The sections were rinsed with PB, and postfixed with 2% OsO4 in PB (pH 7.4) for 2 h at 45 °C. Then they were dehydrated with methanol and embedded between Aclar films (Nisshin EM, Tokyo) with Epon 812. The ultrathin sections were then made and collected on Formvar-coated single-slot grids. They were stained with uranyl acetate and Reynold's solution, and examined with a JEOL 1220EX transmission electron microscope. To investigate the ultrastructure of the celiac ganglion neurons innervating the stomach, 5 μl of 4% WGA-HRP
10
T. Hayakawa et al. / Autonomic Neuroscience: Basic and Clinical 134 (2007) 8–17
Fig. 1. Schematic diagram showing the injection sites of FG (cross stripe) into the cardia and CTb (shaded area) into the pylorus of the stomach. A—antrum; Ca—cardia; Cp—corpus; F—fundus; P—pylorus.
(Vector Laboratories, Burlingame, CA) was injected into the pylorus (3 cases), or the cardia (3 cases) as described above. We also injected 10 μl of 4% WGA-HRP into the whole area of the ventral wall of the stomach including the cardia and the pylorus (2 cases). Three days after the injection, the animals were anesthetized again and perfused first with 100 ml of saline and then with 500 ml of 1% paraformaldehyde–1% glutaraldehyde in PB ( pH 7.4). The celiac ganglion was immediately removed and placed in the same fixative for 1 h. Serial frontal sections were cut at 100 μm with a Vibratome. The sections were presoaked for 5 min with a solution containing 195 mg of ammonium heptamolybdate, 4 mg of 3,3′,5,5′-tetramethylbenzidine in 2 ml of ethanol, and 78 ml of PB (pH 5.0). The sections were processed for 30 min by adding 0.8 ml of 0.3% H2O2 per 80 ml of solution every 5 min. Some sections were rinsed with PB and mounted on gelatin-coated glass slides for light microscopic observation. The rest of the sections were rinsed briefly with PB at pH 5.0, and immediately postfixed with 2% OsO4 in PB (pH 6.0) for 2 h at 45 °C. The sections were then dehydrated in methanol and embedded between Aclar films with Epon 812 (Hayakawa and Zyo, 1992). After confirming that the neurons projecting to the pylorus and the
cardia were labeled retrogradely, the celiac ganglion including the labeled neurons was trimmed under microscopic observation. The ultrathin sections were made, and collected on Formvar-coated single-slot grids. Then they were stained with uranyl acetate and Reynold's solution, and examined with a JEOL 1220EX transmission electron microscope. The retrogradely labeled or unlabeled neurons cut through the plane of the nucleolus were sampled randomly, and photographed at a final magnification of ×10,000 for the quantitative electron microscopic analyses. Prints were used to make montages of somatic profiles. A planimeter and a curvimeter were used to calculate the width, length and area of the neuronal somata, the number of axosomatic terminals, and the percentages of the two different morphological classes of terminals, i.e., the terminals containing round vesicles and the terminals containing pleomorphic vesicles. We then measured the cross-sectional diameter of the dendrites contacting the axon terminals and classified the diameters into three groups: less than 0.5 μm, 0.5–1 μm and more than 1 μm. The distribution of axodendritic terminals contacting these three sizes of labeled dendrites was calculated. The Student's t-test ( P b 0.05) was used to compare the mean values for the neurons projecting to the different regions of the stomach. 3. Results When FG was injected into the cardia and CTb into the pylorus of the same animal, the injection site of FG extended over a region of approximately 1 cm in diameter including the ventral wall of the cardia, the corpus and the fundus. The injection site of CTb extended over a region of approximately 1 cm in diameter including the ventral wall of the pylorus and the antrum (Fig. 1). Fluorescent microscopic observations revealed that retrogradely Cy3-conjugated CTb-labeled neurons were seen as golden-yellow (Fig. 2A), while retrogradely FG-labeled neurons were seen as white-blue. The merged
Fig. 2. Retrogradely labeled neurons in the celiac ganglion projecting to the pylorus or the cardia. (A) Fluorescent photomicrograph of neurons retrogradely labeled with Cy3-conjugated CTb in the celiac ganglion after injection of CTb into the pylorus of the stomach in Case S304. Arrows indicate the splanchnic nerve. (B) Merged photomicrograph of retrogradely CTb-labeled neurons (red), FG-labeled neurons (green), and neurons double-labeled (yellow) with Cy3-CTb and FG (arrow) in the celiac ganglion after injections of CTb into the pylorus and FG into the cardia of the stomach in the same animal in Case S302. Scale bars = 200 μm in A and 50 μm in B.
T. Hayakawa et al. / Autonomic Neuroscience: Basic and Clinical 134 (2007) 8–17
11
Fig. 3. Electron micrograph of a large unlabeled neuron (L) and a medium-sized neuron retrogradely labeled with WGA-HRP (M) in the celiac ganglion after injection of WGA-HRP into the cardia of the stomach in Case S300 (ER—rough endoplasmic reticulum; BV—blood vessel; G—glial cell; p—axon terminals and dendrites surrounded by glial process; S—satellite cell). Scale bar = 10 μm.
photographs using Photoshop® CS showed the CTb-labeled neurons as red, FG-labeled neurons as green, and doublelabeled neurons as yellow (Fig. 2B). Many CT-b labeled neurons and many FG-labeled neurons were found throughout the celiac ganglion, and intermingled each other (Fig. 2B). A
few double-labeled neurons were found in the celiac ganglion (Fig. 2B). We counted FG-labeled, CTb-labeled and doublelabeled neurons in alternate serial sections through the ganglion. The total number of FG-labeled neurons was 449.2 ± 115.4 (mean ± SE, n = 5), the total number of CTb-
Fig. 4. A finger-like somatic process (F) extending from a medium-sized neuron receives an axon terminal. Note that the cytoplasm contains numerous free ribosomes and short strands of rough endoplasmic reticulum (ER), and the somatic membrane and the process are surrounded by glial cytoplasm (arrowheads). Arrows indicate axon terminals and fine dendrites surrounded by glial cytoplasm (D—dendrite; T—axon terminal). Scale bar = 1 μm.
12
T. Hayakawa et al. / Autonomic Neuroscience: Basic and Clinical 134 (2007) 8–17
Fig. 5. Two kinds of axon terminals containing round vesicles (A, B, D) and pleomorphic vesicles (C) in the celiac ganglion. (A) An axosomatic terminal containing round synaptic vesicles and forming asymmetric synaptic contacts with a somatic process. (B) An axodendritic terminal containing round synaptic vesicles and forming asymmetric synaptic contacts with a process extending from a primary dendrite (PD). (C) An axodendritic terminal containing pleomorphic synaptic vesicles and forming asymmetric synaptic contacts with a process extending from a small dendrite (D). (D) Two axodendritic terminals containing round synaptic vesicles and forming asymmetric synaptic contacts with small dendrites (D). In all panels, terminals are surrounded by glial process (small arrows). Large arrows indicate asymmetric synaptic contacts, arrowheads indicate dense-cored vesicles, and double-arrows indicate small glycogen-like granules. Scale bars = 0.5 μm in A, C and D, and 1 μm in B.
labeled neurons was 843.6 ± 146.7, while the total number of double-labeled neurons was 67.0 ± 20.7. The average percentages of double-labeled neurons were 14.9 ± 2.2% for the FGlabeled neurons, and 7.4 ± 1.3% for the CT-b labeled neurons. Thus, the cardia and the pylorus received projections mostly from different neurons in the celiac ganglion. Electron microscopic observations revealed that the celiac ganglion contained medium-sized to large neurons, satellite cells, glial cells, and neuropil including many fine dendrites and axon terminals surrounded by glial cytoplasm, and collagen fibers (Figs. 3 and 4). Only a few small neurons less than 10 μm in diameter, the so-called small intensely fluorescent (SIF) cells, were found. Most of the somatic surface of the neurons was covered by the thin cytoplasm of the satellite cells. Thus, the neurons were separated from each other. A few axosomatic terminals were present. Sometimes a finger-like process (about 0.3–0.5 μm in width) extended from the soma, and made asymmetric synaptic contacts with axon terminals surrounded by glial
Fig. 6. A neuron with four dendrites (arrows) in the celiac ganglion that was retrogradely labeled with WGA-HRP after injection of WGA-HRP into the cardia of the stomach in Case S300. Scale bar = 30 μm.
T. Hayakawa et al. / Autonomic Neuroscience: Basic and Clinical 134 (2007) 8–17
13
Fig. 7. A retrogradely WGA-HRP-labeled neuron in the celiac ganglion after injection of WGA-HRP into the pylorus of the stomach in Case S305 (ER—rough endoplasmic reticulum; g—Golgi apparatus; G—glial cell; N—nucleus; S—satellite cell). Arrows indicate retrogradely transported WGA-HRP reaction products. Scale bar = 5 μm.
cytoplasm (Fig. 4). The axon terminals contained several kinds of vesicles, such as round clear vesicles (40 to 50 nm in diameter, Fig. 5A, B, D), pleomorphic clear vesicles (20 to 60 nm long, Fig. 5C), large dense-cored vesicles (100 to 120 nm in diameter, Figs. 5 and 8), and glycogen-like small dark vesicles (less than 20 nm in diameter, Fig. 5A and B). Dense-cored vesicles were often found in both the terminals containing round vesicles and the terminals containing pleomorphic vesicles (Figs. 5 and 8). About half of the axodendritic terminals (50.2%) contained round synaptic vesicles. Both the terminals containing round vesicles and the terminals containing pleomorphic vesicles formed asymmetric synaptic contacts with the dendrites (Fig. 5). In the neuropil, several small dendrites and axon terminals merged and formed a bundle. The bundles were wrapped by the cytoplasm of glial cells, and were separated from each other (Fig. 4). There were many glial cells among the bundles. The primary dendrites were also covered with glial cytoplasm and received a few axon terminals (Fig. 5B). Unlike axon terminals in the brain, no terminals were found to make symmetric synaptic contacts. Sometimes the axon terminals contacted the smooth somatic membrane (Fig. 8B). The length of axodendritic terminals as a whole was 1.45 ± 0.04 μm (mean ± SE, n = 231). Most of them (67.6%) contacted small dendrites (less than 1 μm in diameter). A
few myelinated fibers were present. Only a few terminals (4.8%) contacted large or primary dendrites (more than 1.5 μm in diameter). There were no axoaxonic terminals, synapses en passant, or dendrodendritic appositions. When WGA-HRP was injected into the cardia, many retrogradely labeled neurons were found throughout the celiac ganglion. The labeled neurons were round or oval, and had four or five primary dendrites (Figs. 3 and 6). The neurons projecting to the cardia were medium-sized (33.3 ± 1.0 × 22.4 ± 0.7 μm, n = 24), round or oval in shape, with a mean area of 593.1 ± 27.0 μm2 per section. The labeled neurons had a dark cytoplasm containing an oval nucleus (13.5 ± 0.4 × 10.0 ± 0.3 μm) having one or two nucleoli, numerous free ribosomes, many mitochondria, many short strands of rough endoplasmic reticulum (rER), smooth endoplasmic reticulum (sER), lysosomes, and several Golgi apparatuses (Fig. 3). There were a few neurons containing aggregation of various sizes of granular vesicles. The cell surface was mostly covered with the thin cytoplasm of satellite cells. Thus, the average number of axosomatic terminals was small and 1.6 ± 0.3 per section through the nucleolus. The axosomatic terminals often contacted small processes extending from the soma (Fig. 8A). About 60% of the axosomatic terminals contained round synaptic vesicles, while the rest of them contained pleomorphic vesicles.
14
T. Hayakawa et al. / Autonomic Neuroscience: Basic and Clinical 134 (2007) 8–17
Fig. 8. Axon terminals contacting retrogradely labeled neurons that project to the pylorus or the cardia. (A) A somatic process extending from a neuron retrogradely labeled with WGA-HRP receiving two axon terminals (T) that contain round synaptic vesicles and that form asymmetric synaptic contacts (arrow) in the case of injection of WGA-HRP into the cardia of the stomach. (B) A terminal (T) containing round synaptic vesicles and forming asymmetric contacts (arrow) with the somatic membrane of a neuron retrogradely labeled with WGA-HRP in the case of injection of WGA-HRP into the pylorus of the stomach. (C) A terminal with anterogradely transported WGA-HRP reaction products (small arrow) forming asymmetric contacts (arrow) with a somatic process of a neuron retrogradely labeled with WGA-HRP in the case of injection of WGA-HRP into the pylorus of the stomach. Arrowheads indicate dense-cored vesicles, doublearrows indicate small glycogen-like granules, and open arrows indicate retrogradely transported WGA-HRP reaction products. Scale bars = 1 μm in A, B and C.
When WGA-HRP was injected into the pylorus, many retrogradely labeled neurons were found in the celiac ganglion. The neurons projecting to the pylorus were medium-sized (33.4 ± 0.9 × 24.7 ± 0.8 μm, n = 24), round or oval in shape, with a mean area of 650.9 ± 34.5 μm2 per section (Fig. 7). A lightly stained, prominent round or oval nucleus was present at the center of the soma, and the cells had one or two nucleoli. The neurons had a dark cytoplasm containing numerous free ribosomes, many short strands of rER, many sER, many mitochondria, lysosomes, several Golgi apparatuses and lysosomes, but only a few granular vesicles of various sizes. The somatic membrane was generally smooth and covered with the thin cytoplasm of the satellite cells. The neurons projecting to the pylorus received only a few axosomatic terminals (1.3 ± 0.3 per section through the nucleolus, n = 24). The axosomatic terminals contacted the smooth somatic membrane (Fig. 8B) or small somatic processes (Fig. 8C). About 60% of them contained round synaptic vesicles, and the rest of them contained pleomorphic vesicles (Table 1). Only two terminals were found to contain anterogradely WGA-HRP reaction products, which contact the retrogradely labeled neurons (Fig. 8C).
When WGA-HRP was injected into the whole ventral wall of the stomach including the cardia, the corpus, the antrum, and the pylorus, many retrogradely labeled neurons Table 1 Ultrastructural characteristics and the number of axosomatic terminals on the neurons in the celiac ganglia that innervate the pylorus, the cardia, or do not innervate the stomach (unlabeled neurons) Cardiac neurons (n = 24) 593.1 ± 27.0 Area (μm2) Cell length (μm) 33.3 ± 1.0 Cell width (μm) 22.4 ± 0.7 Nuclear length (μm) 13.5 ± 0.4 Nuclear width (μm) 10.0 ± 0.3 Axosomatic terminals 1.6 ± 0.3 Terminals containing 1.0 ± 0.2 round synaptic vesicles Terminals containing 0.6 ± 0.1 pleomorphic synaptic vesicles
Pyloric neurons (n = 24)
Unlabeled neurons (n = 22)
650.9 ± 34.5 33.4 ± 0.9 24.7 ± 0.8 13.5 ± 0.5 10.3 ± 0.4 1.3 ± 0.3 0.8 ± 0.2
872.8 ± 41.9⁎ 40.4 ± 1.0⁎ 28.0 ± 1.0⁎ 14.5 ± 0.4 11.7 ± 0.3 0.9 ± 0.2 0.4 ± 0.1
0.6 ± 0.1
0.5 ± 0.2
The values are the mean ± S.E. ⁎ Indicates statistically significant differences between the unlabeled neurons and the neurons innervating the pylorus or the cardia (P b 0.05).
T. Hayakawa et al. / Autonomic Neuroscience: Basic and Clinical 134 (2007) 8–17
were found throughout the celiac ganglion. Nevertheless, there were still many neurons in the celiac ganglion that were not labeled retrogradely. These unlabeled neurons were large (40.4 ± 1.0 × 28.0 ± 1.0 μm), and had a dark cytoplasm containing many mitochondria, numerous free ribosomes, many rough and smooth ER, several Golgi apparatuses, many lysosomes, and a light nucleus (14.5 ± 0.4 × 11.7 ± 0.3 μm) with one or two prominent nucleoli (Fig. 3). Sometimes they contained many lysosomes, lipofuscin granules, and granular vesicles of various sizes in the cytoplasm. Since the somatic membrane was mostly covered with the cytoplasm of the satellite cells, they received only a few axosomatic terminals (0.9 ± 0.2 per section through the nucleolus, n = 22). About 44% of the axosomatic terminals contained round synaptic vesicles, and the rest of them contained pleomorphic synaptic vesicles (Table 1). There were no significant differences in the area, the size, and the number of axosomatic terminals between the neurons projecting to the cardia and to the pylorus. In contrast, there were significant differences in the area (t = 5.61, P b 0.01), in the length (t = 4.93, P b 0.01), and in the width (t = 4.46, P b 0.01) between the neurons not projecting to the stomach and the neurons projecting to the cardia. Significant differences were also present in the area (t = 4.09, P b 0.01), in the length (t = 5.03, P b 0.01), and in the width (t = 2.49, P b 0.05) between the neurons not projecting to the stomach and the neurons projecting to the pylorus. However, there were no significant differences in the number of axosomatic terminals, the number of axosomatic terminals containing round vesicles or pleomorphic vesicles among these three kinds of the neurons. 4. Discussion The present study demonstrated that the cardia and the pylorus are innervated mostly by different neurons in the celiac ganglion. The numbers of retrogradely labeled neurons were variable between animals injected with FG and CTb into the cardia and the pylorus of the stomach, respectively. Nevertheless, the number of double-labeled neurons was small, and the percentages of double-labeled neurons for the FG-labeled (14.9 ± 2.2%) or the CTb-labeled neurons (7.4 ± 1.3%) were small and constant. Thus, there were only a few neurons in the celiac ganglion that send collateral fibers to both the cardia and the pylorus. The vagus nerve sends parasympathetic fibers to the stomach (Rogers and Hermann, 1983; Shapiro and Miselis, 1985; Norgren and Smith, 1988; Holst et al., 1997; Hayakawa et al., 2004, 2006). Retrograde tracing studies using two kinds of tracers showed that the cardia and the antrum of the stomach received collateral projections from the neurons in the DMV, while the pylorus was innervated by neurons that were different from the neurons projecting to the cardia in the DMV (Hayakawa et al., 2003). Based on these results and our observations, the cardia appears to be innervated and regulated by parasympathetic neurons and sympathetic
15
neurons that are different from the neurons projecting to the pylorus. Berthoud et al. (1991) reported that the cardia receives parasympathetic vagal projections by way of the gastric branch, and the pylorus receives projections by way of the hepatic branch of the vagus nerve. However, it is not clear if the neurons in the celiac ganglion run through the celiac branch and project to the cardia, and the neurons projecting to the pylorus traverse the hepatic branch. We also demonstrated that the celiac ganglion is composed of medium-sized to large neurons having a dark cytoplasm with many cell organelles, and a few axosomatic terminals. Sometimes small processes protruded from the soma and received axon terminals forming asymmetric synaptic contacts. The small processes corresponded to the accessory dendrites in the cat (Elfvin, 1971b) or the short spine-like processes in the mouse prevertebral ganglia (Miller et al., 1996). The present ultrastructural features of the celiac ganglion neurons of the rat are similar to those of the neurons of the superior mesenteric ganglion of the mouse (Miller et al., 1996), and the neurons of the inferior mesenteric ganglion of the guinea pig (Gibbins et al., 2000, 2003) and the cat (Elfvin, 1971a). Serial ultrathin sections of the neuron of the inferior mesenteric ganglion of the cat revealed that the neurons have a few axosomatic terminals and short processes (Elfvin, 1971a). In the prevertebral ganglia, there are small intensely fluorescent cells with many large granules in there sparse cytoplasm, the so-called SIF cells (Yokota and Burnstock, 1983; Elfvin et al., 1997). Only a few SIF cells were found in the present study. Szurszewski and Miller (1994) reported that the number and appearance of SIF cells are quite variable among species. In the rat neuropil, many axodendritic terminals were located around the neuronal soma. Unlike the guinea pig and the mouse prevertebral ganglia (Miller et al., 1996; Gibbins et al., 2003), several terminals and small dendrites coalesced to form a bundle surrounded by the glial cytoplasm. Light microscopic immunohistochemical studies showed that many terminal-like boutons are distributed around the soma, and contain nitric oxide synthase (NOS) or acetylcholine (Hamberger and Norberg, 1965; Alm et al., 1995; Elfvin et al., 1997; Gibbins et al., 2003). These terminals were considered to contact small dendrites near the soma, but not the somatic membrane, judging from the ultrastructural observations. There were many axoaxonic contacts and dendrodendritic appositions in the cat inferior mesenteric ganglion (Elfvin, 1971c; Miller et al., 1996), while in the rat celiac ganglion there were few axoaxonic terminals and dendrodendritic appositions. Immunohistochemical studies of the rat, the cat and the guinea pig have demonstrated that the terminals containing NOS, ENK, NT, and choline acetyltransferase (ChAT) come from the thoracic spinal cord, and the terminals containing VIP, CCK, DYN, bombesin, and CGRP come from the myenteric ganglion of the intestine (Hamaji et al., 1987; Messenger and Furness, 1992; Elfvin et al., 1993; Domoto et al., 1995). Electron microscopic studies combined with
16
T. Hayakawa et al. / Autonomic Neuroscience: Basic and Clinical 134 (2007) 8–17
immunohistochemistry of the prevertebral ganglia showed that the terminals containing pleomorphic vesicles also contain VIP in the guinea pig (Kondo and Yui, 1982b; Hamaji et al., 1989), while the terminals containing small round vesicles also contain NOS in the rat (Domoto et al., 1994) or ChAT, and the terminals containing large granular vesicles or dense-cored vesicles also contain tyrosine hydroxylase, the rate-limiting enzyme for NA synthesis in the rat, the cat, and the guinea pig (Leranth and Ungvary, 1980; Yokota and Burnstock, 1983; Smolen, 1988; Alm et al., 1995; Elfvin et al., 1997). Both the neurons projecting to the cardia and the neurons projecting to the pylorus had terminals containing round vesicles or pleomorphic vesicles, and these terminals often contained dense-cored vesicles in the present study. Thus, they may receive terminals using acetylcholine, VIP or NO as neurotransmitters or neuromodulators. The celiac ganglia neurons projecting to the stomach received projections from the thoracic spinal cord, but few from the neurons of the myenteric ganglion of the stomach (Messenger and Furness, 1992). Thus, the axosomatic terminals containing pleomorphic vesicles on the neurons projecting to the stomach may come from the intrinsic neurons in the prevertebral ganglia, but not from the myenteric ganglia of the alimentary canal. The retrogradely labeled neurons projecting to the cardia were similar to the labeled neurons projecting to the pylorus in their ultrastructural appearances, their size, the number of axosomatic terminals, and the ratio of the two kinds of terminals (Table 1). Based on peptide content, there are three kinds of neurons in the celiac ganglion: NA neurons that also contain NPY; NA neurons that also contain SOM; and NA neurons that do not contain either NPY or SOM (Macrae et al., 1986). Lindh et al. (1986) reported that the retrogradely labeled neurons projecting to the pylorus contain NA and NPY, or NA and SOM, but only a few labeled neurons contain only NA. There have been, however, no immunohistochemical studies about the neurons projecting to the cardia. There were large neurons in the celiac ganglion that do not project to the stomach in the present study. Although we could not label all neurons projecting to the stomach, the neurons projecting to the pylorus or the cardia were significantly smaller than the unlabeled neurons. Since the celiac ganglion neurons project not only to the stomach but also to the ileum and colon in the rat, most of these unlabeled neurons may project to the small and large intestine (Luckensmeyer and Keast, 1994; Quinson et al., 2001). Kuramoto and Furness (1989) demonstrated that the celiac ganglion neurons receive projections from the neurons in the myenteric ganglia in the duodenum, and the small and large intestine. Then they formed reciprocal neuronal circuitry for the regulation of alimentary canal activity (Furness, 2003). Thus, the unlabeled large neurons may project to the duodenum and the intestine, and receive projections from the intestinofugal neurons in the myenteric ganglion of the small intestine. The ultrastructural features of the large neurons were also similar to those of the neurons projecting
to the stomach except for their cell sizes. Sometimes there were large neurons, which contain many lysosomes and large granules. These neurons were often present in the inferior mesenteric, and the superior mesenteric ganglia of the mouse (Miller et al., 1996) and the cat (Elfvin, 1971a). Physiological studies showed that neurons innervating the pylorus cause contractions of the sphincter muscle using NA (Edin et al., 1979), while stimulation of the splanchnic nerve inhibited the contraction of the fundus and the cardia of the stomach (Fandriks et al., 1987). Electrophysiological studies have reported that there are three classes of neurons based on their potassium channel expression in excitability in the guinea pig celiac ganglion (Boyd et al., 1996). These three classes of neurons had a different number of dendrites and sizes of their neuronal soma. The neurons projecting to the pylorus and the cardia may have different functions, and immunohistochemical and morphological properties, though we could not see obvious ultrastructural differences between these neurons. Further immunohistochemical and electrophysiological studies will be needed on the neurons projecting to the cardia and the pylorus in the celiac ganglion. Acknowledgements The authors thank Ms. M. Hatta and Mr. K. Gion for their technical assistance. References Alm, P., Uvelius, B., Ekstrom, J., Holmqvist, B., Larsson, B., Andersson, K.E., 1995. Nitric oxide synthase-containing neurons in rat parasympathetic, sympathetic and sensory ganglia: a comparative study. Histochem. J. 27, 819–831. Berthoud, H.R., Carlson, N.R., Powley, T.L., 1991. Topography of efferent vagal innervation of the rat gastrointestinal tract. Am. J. Physiol. 260, R200–R207. Boyd, H.D., McLachlan, E.M., Keast, J.R., Inokuchi, H., 1996. Three electrophysiological classes of guinea pig sympathetic postganglionic neurone have distinct morphologies. J. Comp. Neurol. 369, 372–387. Domoto, T., Teramoto, M., Tamura, K., Yasui, Y., 1994. Ultrastructural study on NOS-immunoreactive nerve terminals in the rat coeliac ganglion. NeuroReport 169–172. Domoto, T., Teramoto, M., Tanigawa, K., Tamura, K., Yasui, Y., 1995. Origins of nerve fibers containing nitric oxide synthase in the rat celiacsuperior mesenteric ganglion. Cell Tissue Res. 281, 215–221. Edin, R., Ahlman, H., Kewenter, J., 1979. The vagal control of the feline pyloric sphincter. Acta Physiol. Scand. 107, 169–174. Elfvin, L.G., 1971a. Ultrastructural 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. Elfvin, L.G., 1971b. Ultrastructural studies on the synaptology of the inferior mesenteric ganglion of the cat: II. Specialized serial neuronal contacts between preganglionic end fibers. J. Ultrastruct. Res. 37, 426–431. Elfvin, L.G., 1971c. Ultrastructural studies on the synaptology of the inferior mesenteric ganglion of the cat: 3. The structure and distribution of the axodendritic and dendrodendritic contacts. J. Ultrastruct. Res. 37, 432–448. Elfvin, L.G., Lindh, B., Hokfelt, T., 1993. The chemical neuroanatomy of sympathetic ganglia. Annu. Rev. Neurosci. 16, 471–507. Elfvin, L.G., Holmberg, K., Emson, P., Schemann, M., Hokfelt, T., 1997. Nitric oxide synthase, choline acetyltransferase, catecholamine enzymes
T. Hayakawa et al. / Autonomic Neuroscience: Basic and Clinical 134 (2007) 8–17 and neuropeptides and their colocalization in the anterior pelvic ganglion, the inferior mesenteric ganglion and the hypogastric nerve of the male guinea pig. J. Chem. Neuroanat. 14, 33–49. Fandriks, L., Jonson, C., Nylander, O., 1987. Effects of splanchnic nerve stimulation and of clonidine on gastric and duodenal HCO3-secretion in the anaesthetized cat. Acta Physiol. Scand. 130, 251–258. Furness, J.B., 2003. Intestinofugal neurons and sympathetic reflexes that bypass the central nervous system. J. Comp. Neurol. 455, 281–284. Gibbins, I.L., Jobling, P., Messenger, J.P., Teo, E.H., Morris, J.L., 2000. Neuronal morphology and the synaptic organisation of sympathetic ganglia. J. Auton. Nerv. Syst. 81, 104–109. Gibbins, I.L., Teo, E.H., Jobling, P., Morris, J.L., 2003. Synaptic density, convergence, and dendritic complexity of prevertebral sympathetic neurons. J. Comp. Neurol. 455, 285–298. Hamaji, M., Kawai, Y., Kawashima, Y., Tohyama, M., 1987. Projections of bombesin-like immunoreactive fibers from the rat stomach to the celiac ganglion revealed by a double-labeling technique. Brain Res. 416, 192–194. Hamaji, M., Kawai, Y., Kawashima, Y., Tohyama, M., 1989. An electron microscopic study on VIP-, BOM- and CCK-like immunoreactive terminals in the celiac-superior mesenteric ganglion complex of the guinea pig. Brain Res. 488, 283–287. Hamberger, B., 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. Hayakawa, T., Zyo, K., 1992. Ultrastructural study of ascending projections to the lateral mammillary nucleus of the rat. Anat. Embryol. (Berl.) 185, 547–557. Hayakawa, T., Takanaga, A., Tanaka, K., Maeda, S., Seki, M., 2003. Cells of origin of vagal motor neurons projecting to different parts of the stomach in the rat: confocal laser scanning and electron microscopic study. Anat. Embryol. (Berl.) 207, 289–297. Hayakawa, T., Takanaga, A., Tanaka, K., Maeda, S., Seki, M., 2004. Distribution and ultrastructure of dopaminergic neurons in the dorsal motor nucleus of the vagus projecting to the stomach of the rat. Brain Res. 1006, 66–73. Hayakawa, T., Kuwahara, S., Maeda, S., Tanaka, K., Seki, M., 2006. Direct synaptic contacts on the myenteric ganglia of the rat stomach from the dorsal motor nucleus of the vagus. J. Comp. Neurol. 498, 352–362. Holst, M.C., Kelly, J.B., Powley, T.L., 1997. Vagal preganglionic projections to the enteric nervous system characterized with Phaseolus vulgaris-leucoagglutinin. J. Comp. Neurol. 381, 81–100. Jobling, P., Gibbins, I.L., 1999. Electrophysiological and morphological diversity of mouse sympathetic neurons. J. Neurophysiol. 82, 2747–2764. Kondo, H., Yamamoto, M., 1988. The ontogeny and fine structure of calcitonin gene-related peptide (CGRP)-immunoreactive nerve fibers in the celiac ganglion of rats. Arch. Histol. Cytol. 51, 91–98. Kondo, H., Yui, R., 1981. An electron microscopic study on substance P-like immunoreactive nerve fibers in the celiac ganglion of guinea pigs. Brain Res. 222, 134–137. Kondo, H., Yui, R., 1982a. An electron microscopic study on enkephalinlike immunoreactive nerve fibers in the celiac ganglion of guinea pigs. Brain Res. 252, 142–145.
17
Kondo, H., Yui, R., 1982b. An electron microscopic study on VIP-like immunoreactive nerve fibers in the celiac ganglion of guinea pigs. Brain Res. 237, 227–231. Kuramoto, H., Furness, J.B., 1989. Distribution of enteric nerve cells that project from the small intestine to the coeliac ganglion in the guinea-pig. J. Auton. Nerv. Syst. 27, 241–248. Leranth, C., Ungvary, G., 1980. Axon types of prevertebral ganglia and the peripheral autonomic reflex arc. J. Auton. Nerv. Syst. 1, 265–281. Lindh, B., Hokfelt, T., Elfvin, L.G., Terenius, L., Fahrenkrug, J., Elde, R., Goldstein, M., 1986. Topography of NPY-, somatostatin-, and VIPimmunoreactive, neuronal subpopulations in the guinea pig celiacsuperior mesenteric ganglion and their projection to the pylorus. J. Neurosci. 6, 2371–2383. Luckensmeyer, G.B., Keast, J.R., 1994. Projections from the prevertebral and major pelvic ganglia to the ileum and large intestine of the male rat. J. Auton. Nerv. Syst. 49, 247–259. Lundberg, J.M., Hokfelt, T., Anggard, A., Terenius, L., Elde, R., Markey, K., Goldstein, M., Kimmel, J., 1982. Organizational principles in the peripheral sympathetic nervous system: subdivision by coexisting peptides (somatostatin-, avian pancreatic polypeptide-, and vasoactive intestinal polypeptide-like immunoreactive materials). Proc. Natl. Acad. Sci. U. S. A. 79, 1303–1307. Macrae, I.M., Furness, J.B., Costa, M., 1986. Distribution of subgroups of noradrenaline neurons in the coeliac ganglion of the guinea-pig. Cell Tissue Res. 244, 173–180. Messenger, J.P., Furness, J.B., 1992. Distribution of enteric nerve cells that project to the coeliac ganglion of the guinea-pig. Cell Tissue Res. 269, 119–132. Miller, S.M., Hanani, M., Kuntz, S.M., Schmalz, P.F., Szurszewski, J.H., 1996. Light, electron, and confocal microscopic study of the mouse superior mesenteric ganglion. J. Comp. Neurol. 365, 427–444. Norgren, R., Smith, G.P., 1988. Central distribution of subdiaphragmatic vagal branches in the rat. J. Comp. Neurol. 273, 207–223. Quinson, N., Robbins, H.L., Clark, M.J., Furness, J.B., 2001. Locations and innervation of cell bodies of sympathetic neurons projecting to the gastrointestinal tract in the rat. Arch. Histol. Cytol. 64, 281–294. Rogers, R.C., Hermann, G.E., 1983. Central connections of the hepatic branch of the vagus nerve: a horseradish peroxidase histochemical study. J. Auton. Nerv. Syst. 7, 165–174. Roman, C., Gonella, J., 1981. In: Johnson, L.R. (Ed.), Extrinisic Control of Digestive Tract Motility. Rave Press, New York. 289–333 pp. Shapiro, R.E., Miselis, R.R., 1985. The central organization of the vagus nerve innervating the stomach of the rat. J. Comp. Neurol. 238, 473–488. Smolen, A.J., 1988. Morphology of synapses in the autonomic nervous system. J. Electron. Microsc. Tech. 10, 187–204. Szurszewski, J.H., Miller, S.M., 1994. In: Johnson, L.R. (Ed.), Physiology of Prevertebral Ganglion. Raven Press, New York. 795-877 pp. Yokota, R., Burnstock, G., 1983. Synaptic organisation of the pelvic ganglion in the guinea-pig. Cell Tissue Res. 232, 379–397.
Morphology and ultrastructure of the sympathetic celiac ganglion neurons projecting to the cardia and pylorus of the rat stomach Tetsu Hayakawa ⁎, Sachi Kuwahara, Seishi Maeda, Koichi Tanaka, Makoto Seki Department of Anatomy, Hyogo College of Medicine, Mukogawa, Nishinomiya, Hyogo 663-8501, Japan Received 27 October 2006; received in revised form 19 January 2007; accepted 27 January 2007
Abstract The stomach receives sympathetic projections from the celiac ganglion. To determine what kinds of neurons in the celiac ganglion project to the cardia or the pylorus of the stomach, we injected the retrograde tracer Fluoro-Gold into the cardia and the retrograde tracer cholera toxin subunit b into the pylorus of the same animal. A few neurons (about 10%) innervating the cardia sent collateral projections to the pylorus. Ultrastructural observations revealed that the celiac ganglion contained oval, medium-sized to large neurons. They had a dark cytoplasm containing numerous free ribosomes, rough endoplasmic reticulum, mitochondria, lysosomes, several Golgi apparatuses, and an oval nucleus. The axon terminals were small and usually contacted thin processes extending from the dendrites or the soma. About half of the terminals contained round vesicles, while the rest contained pleomorphic vesicles. Both types of terminals made asymmetric synaptic contacts. We then retrogradely labeled the neurons projecting to the cardia and the pylorus with wheat germ agglutinin conjugated horseradish peroxidase to examine their ultrastructural characteristics. The neurons projecting to the cardia (33.3 × 22.4 μm) were similar to the neurons projecting to the pylorus (33.4 × 24.7 μm) in their size and ultrastructural appearance. The neurons not projecting to the stomach (40.4 × 28.0 μm) were significantly larger than the neurons projecting to the cardia or the pylorus. Only a few axosomatic terminals were found on the neurons projecting to the cardia (1.6 per somatic profile), the pylorus (1.3) or the neurons not projecting to the stomach (0.9). These results provide morphological bases for the sympathetic motor neurons innervating the stomach. © 2007 Elsevier B.V. All rights reserved. Keywords: Sympathetic neurons; Gastric motility; Retrograde labeling; Synaptic organization; Electron microscopy
1. Introduction The neurons of the sympathetic celiac ganglion innervate the stomach and regulate the motility of the stomach wall. The celiac ganglion of the rat lies on the ventral wall of the abdominal aorta just rostral to the root of the superior mesenteric artery, and joins the superior mesenteric ganglia to form the celiac superior mesenteric ganglion complex. Light microscopic observation revealed that the neurons of the celiac ganglion are large, round in shape with several dendrites, and surrounded by satellite cells (Szurszewski and Miller, 1994). Ultrastructural studies of the prevertebral ⁎ Corresponding author. Tel.: +81 798 45 6484; fax: +81 798 45 6485. E-mail address: [email protected] (T. Hayakawa). 1566-0702/$ - see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.autneu.2007.01.011
sympathetic neurons including the celiac, superior mesenteric and inferior mesenteric ganglia of the mouse (Miller et al., 1996), the cat (Elfvin, 1971a), and the guinea pig (Gibbins et al., 2003) showed that the sympathetic neurons are large, and round in shape with four to five primary dendrites. The soma and primary dendrites are surrounded by satellite cells and have only a few axosomatic terminals. Sometimes short fine processes (accessory dendrites) extend from the soma, and make synaptic contacts with axon terminals. In the neuropil, there are many axodendritic terminals containing round or pleomorphic synaptic vesicles. Axoaxonic contacts and dendrodendritic contacts are present in the cat (Elfvin, 1971c) and the mouse (Miller et al., 1996). The ultrastructural features and synaptic organization of the celiac ganglia of the mouse are somewhat different from the
T. Hayakawa et al. / Autonomic Neuroscience: Basic and Clinical 134 (2007) 8–17
other species that have been studied. However, there have been few studies about the ultrastructure and synaptic organization of the rat celiac ganglion neurons. Immunohistochemical studies of the celiac ganglion (Macrae et al., 1986; Jobling and Gibbins, 1999) revealed that almost all neurons contain noradrenaline (NA). Three major subpopulations of noradrenergic neurons have been described in the guinea pig celiac ganglion, that is, NA neurons that also contain somatostatin (SOM), NA neurons that also contain neuropeptide Y (NPY), and NA neurons that do not contain either SOM or NPY (Lundberg et al., 1982; Macrae et al., 1986; Szurszewski and Miller, 1994). The celiac ganglion neurons receive many fibers from the intermediolateral nucleus of the thoracic spinal cord, the dorsal root ganglion, and the intestinofugal neurons of the myenteric ganglia in the gastrointestinal tract (Kuramoto and Furness, 1989; Messenger and Furness, 1992; Domoto et al., 1995). These fibers and terminals contain several kinds of peptides such as vasoactive intestinal peptide (VIP), cholecystokinin (CCK), enkephalins (ENK), dynorphin (DYN), neurotensin (NT), calcitonin gene-related peptide (CGRP), and substance P (SP) in addition to nitric oxide (NO) and acetylcholine (Kondo and Yui, 1981, 1982a,b; Kondo and Yamamoto, 1988; Elfvin et al., 1993; Szurszewski and Miller, 1994). The axon terminals contain small clear vesicles, small pleomorphic vesicles, and small or large granular vesicles (Smolen, 1988). However, it is not clear what kind of axon terminals contact the celiac ganglion neurons that innervate the stomach of the rat. Physiological studies have revealed that stimulation of sympathetic innervation of the pylorus elicits a contraction of sphincter muscles, while stimulation of the splanchnic nerve inhibits the gastric contraction of the fundus and the cardia of the stomach (Roman and Gonella, 1981). A retrograde tracing study using two kinds of tracers showed that the neurons of the dorsal motor nucleus of the vagus (DMV) send collateral projections to the cardia and the corpus, whereas the cardia and the pylorus received projections mostly from different neurons in the DMV (Hayakawa et al., 2003). Thus, it is likely that different groups of neurons in the celiac ganglion project to and have different effects on the cardia and the pylorus. In the present study, we attempted to clarify whether the neurons projecting to the cardia of the stomach in the celiac ganglia have collateral projections to the pylorus, using two kinds of retrograde tracers, Fluoro-Gold (FG) and cholera toxin subunit b (CTb). We also identified the neurons projecting to the cardia and the pylorus by retrograde labeling with wheat germ agglutinin conjugated horseradish peroxidase (WGA-HRP), and determined their ultrastructural features and synaptic organization with the electron microscope. 2. Materials and methods We used sixteen male Sprague–Dawley rats weighing 250– 300 g. All surgical procedures were carried out with the animals
9
under sodium pentobarbital anesthesia (50 mg/kg, i.p.). The Animal Care and Use Committee of Hyogo College of Medicine approved the procedures. The abdominal wall was opened and the liver was pushed aside to expose the stomach. Using a glass micropipette (tip diameter 80 μm) affixed to a 10-μl Hamilton syringe, four aliquots of 2 μl each of 1% FG (Fluorochrome LLC, Denver CO) in distilled water were injected into the wall of the ventral surface of cardia. Then four aliquots of 2 μl each of 2% CTb (List Biological Laboratories, Campbell CA) in 0.1 M phosphate buffer (PB) at pH 7.4 were injected by pressure into the wall of the ventral surface of the pylorus (5 cases) in the same animal. Cotton swabs were used during the injections to prevent the spread of the tracer to adjacent structures. Three days after the injections, the animals were anesthetized again and perfused first with 100 ml of saline and then with 500 ml of 4% paraformaldehyde-15% picric acid in PB (pH 7.4). The celiac ganglion together with the abdominal aorta was immediately removed and placed in the same fixative for 1 h. Serial frozen sections were cut at 40 μm, and processed for immunohistochemistry. After a rinse with PB (pH 7.4), the sections were incubated with 1% bovine serum albumin in PB at pH 7.4 containing 0.9% NaCl and 0.3% Triton X-100 (PBST) for 1 h at room temperature. The sections were then incubated with a goat anti-CTb serum (List Biological Laboratories; 1:20,000 dilution) for 1 day at 4 °C. The primary antibodies were localized by incubation with Cy3-conjugated donkey anti-goat IgG (Jackson, West Grove PA; 1:400 dilution) for 5 h at room temperature. After rinsing with PBST, the sections were mounted onto gelatin-coated slides and dried. FG was viewed with a U excitation filter (blue), and Cy3 with a G excitation filter (red). Fluorescence photomicrographs were taken using an Olympus AX80. We made merged photographs using Photoshop® CS, and observed blue-green FG-, red CTb- and double-labeled (yellow) neurons. Then, we counted the labeled neurons in alternate serial sections of the ganglion to avoid double counting, and calculated the percentage of the double-labeled neurons for the FG- or the CTb-labeled neurons. To investigate the ultrastructure of the celiac ganglion neurons, three rats were anesthetized and perfused first with 100 ml of saline and then with 500 ml of 1% glutaraldehyde1% paraformaldehyde in PB (pH7.4). The celiac ganglion together with the abdominal aorta was removed and immersed in the same fixative. Serial frontal sections were cut at 100 μm with a Vibratome. The sections were rinsed with PB, and postfixed with 2% OsO4 in PB (pH 7.4) for 2 h at 45 °C. Then they were dehydrated with methanol and embedded between Aclar films (Nisshin EM, Tokyo) with Epon 812. The ultrathin sections were then made and collected on Formvar-coated single-slot grids. They were stained with uranyl acetate and Reynold's solution, and examined with a JEOL 1220EX transmission electron microscope. To investigate the ultrastructure of the celiac ganglion neurons innervating the stomach, 5 μl of 4% WGA-HRP
10
T. Hayakawa et al. / Autonomic Neuroscience: Basic and Clinical 134 (2007) 8–17
Fig. 1. Schematic diagram showing the injection sites of FG (cross stripe) into the cardia and CTb (shaded area) into the pylorus of the stomach. A—antrum; Ca—cardia; Cp—corpus; F—fundus; P—pylorus.
(Vector Laboratories, Burlingame, CA) was injected into the pylorus (3 cases), or the cardia (3 cases) as described above. We also injected 10 μl of 4% WGA-HRP into the whole area of the ventral wall of the stomach including the cardia and the pylorus (2 cases). Three days after the injection, the animals were anesthetized again and perfused first with 100 ml of saline and then with 500 ml of 1% paraformaldehyde–1% glutaraldehyde in PB ( pH 7.4). The celiac ganglion was immediately removed and placed in the same fixative for 1 h. Serial frontal sections were cut at 100 μm with a Vibratome. The sections were presoaked for 5 min with a solution containing 195 mg of ammonium heptamolybdate, 4 mg of 3,3′,5,5′-tetramethylbenzidine in 2 ml of ethanol, and 78 ml of PB (pH 5.0). The sections were processed for 30 min by adding 0.8 ml of 0.3% H2O2 per 80 ml of solution every 5 min. Some sections were rinsed with PB and mounted on gelatin-coated glass slides for light microscopic observation. The rest of the sections were rinsed briefly with PB at pH 5.0, and immediately postfixed with 2% OsO4 in PB (pH 6.0) for 2 h at 45 °C. The sections were then dehydrated in methanol and embedded between Aclar films with Epon 812 (Hayakawa and Zyo, 1992). After confirming that the neurons projecting to the pylorus and the
cardia were labeled retrogradely, the celiac ganglion including the labeled neurons was trimmed under microscopic observation. The ultrathin sections were made, and collected on Formvar-coated single-slot grids. Then they were stained with uranyl acetate and Reynold's solution, and examined with a JEOL 1220EX transmission electron microscope. The retrogradely labeled or unlabeled neurons cut through the plane of the nucleolus were sampled randomly, and photographed at a final magnification of ×10,000 for the quantitative electron microscopic analyses. Prints were used to make montages of somatic profiles. A planimeter and a curvimeter were used to calculate the width, length and area of the neuronal somata, the number of axosomatic terminals, and the percentages of the two different morphological classes of terminals, i.e., the terminals containing round vesicles and the terminals containing pleomorphic vesicles. We then measured the cross-sectional diameter of the dendrites contacting the axon terminals and classified the diameters into three groups: less than 0.5 μm, 0.5–1 μm and more than 1 μm. The distribution of axodendritic terminals contacting these three sizes of labeled dendrites was calculated. The Student's t-test ( P b 0.05) was used to compare the mean values for the neurons projecting to the different regions of the stomach. 3. Results When FG was injected into the cardia and CTb into the pylorus of the same animal, the injection site of FG extended over a region of approximately 1 cm in diameter including the ventral wall of the cardia, the corpus and the fundus. The injection site of CTb extended over a region of approximately 1 cm in diameter including the ventral wall of the pylorus and the antrum (Fig. 1). Fluorescent microscopic observations revealed that retrogradely Cy3-conjugated CTb-labeled neurons were seen as golden-yellow (Fig. 2A), while retrogradely FG-labeled neurons were seen as white-blue. The merged
Fig. 2. Retrogradely labeled neurons in the celiac ganglion projecting to the pylorus or the cardia. (A) Fluorescent photomicrograph of neurons retrogradely labeled with Cy3-conjugated CTb in the celiac ganglion after injection of CTb into the pylorus of the stomach in Case S304. Arrows indicate the splanchnic nerve. (B) Merged photomicrograph of retrogradely CTb-labeled neurons (red), FG-labeled neurons (green), and neurons double-labeled (yellow) with Cy3-CTb and FG (arrow) in the celiac ganglion after injections of CTb into the pylorus and FG into the cardia of the stomach in the same animal in Case S302. Scale bars = 200 μm in A and 50 μm in B.
T. Hayakawa et al. / Autonomic Neuroscience: Basic and Clinical 134 (2007) 8–17
11
Fig. 3. Electron micrograph of a large unlabeled neuron (L) and a medium-sized neuron retrogradely labeled with WGA-HRP (M) in the celiac ganglion after injection of WGA-HRP into the cardia of the stomach in Case S300 (ER—rough endoplasmic reticulum; BV—blood vessel; G—glial cell; p—axon terminals and dendrites surrounded by glial process; S—satellite cell). Scale bar = 10 μm.
photographs using Photoshop® CS showed the CTb-labeled neurons as red, FG-labeled neurons as green, and doublelabeled neurons as yellow (Fig. 2B). Many CT-b labeled neurons and many FG-labeled neurons were found throughout the celiac ganglion, and intermingled each other (Fig. 2B). A
few double-labeled neurons were found in the celiac ganglion (Fig. 2B). We counted FG-labeled, CTb-labeled and doublelabeled neurons in alternate serial sections through the ganglion. The total number of FG-labeled neurons was 449.2 ± 115.4 (mean ± SE, n = 5), the total number of CTb-
Fig. 4. A finger-like somatic process (F) extending from a medium-sized neuron receives an axon terminal. Note that the cytoplasm contains numerous free ribosomes and short strands of rough endoplasmic reticulum (ER), and the somatic membrane and the process are surrounded by glial cytoplasm (arrowheads). Arrows indicate axon terminals and fine dendrites surrounded by glial cytoplasm (D—dendrite; T—axon terminal). Scale bar = 1 μm.
12
T. Hayakawa et al. / Autonomic Neuroscience: Basic and Clinical 134 (2007) 8–17
Fig. 5. Two kinds of axon terminals containing round vesicles (A, B, D) and pleomorphic vesicles (C) in the celiac ganglion. (A) An axosomatic terminal containing round synaptic vesicles and forming asymmetric synaptic contacts with a somatic process. (B) An axodendritic terminal containing round synaptic vesicles and forming asymmetric synaptic contacts with a process extending from a primary dendrite (PD). (C) An axodendritic terminal containing pleomorphic synaptic vesicles and forming asymmetric synaptic contacts with a process extending from a small dendrite (D). (D) Two axodendritic terminals containing round synaptic vesicles and forming asymmetric synaptic contacts with small dendrites (D). In all panels, terminals are surrounded by glial process (small arrows). Large arrows indicate asymmetric synaptic contacts, arrowheads indicate dense-cored vesicles, and double-arrows indicate small glycogen-like granules. Scale bars = 0.5 μm in A, C and D, and 1 μm in B.
labeled neurons was 843.6 ± 146.7, while the total number of double-labeled neurons was 67.0 ± 20.7. The average percentages of double-labeled neurons were 14.9 ± 2.2% for the FGlabeled neurons, and 7.4 ± 1.3% for the CT-b labeled neurons. Thus, the cardia and the pylorus received projections mostly from different neurons in the celiac ganglion. Electron microscopic observations revealed that the celiac ganglion contained medium-sized to large neurons, satellite cells, glial cells, and neuropil including many fine dendrites and axon terminals surrounded by glial cytoplasm, and collagen fibers (Figs. 3 and 4). Only a few small neurons less than 10 μm in diameter, the so-called small intensely fluorescent (SIF) cells, were found. Most of the somatic surface of the neurons was covered by the thin cytoplasm of the satellite cells. Thus, the neurons were separated from each other. A few axosomatic terminals were present. Sometimes a finger-like process (about 0.3–0.5 μm in width) extended from the soma, and made asymmetric synaptic contacts with axon terminals surrounded by glial
Fig. 6. A neuron with four dendrites (arrows) in the celiac ganglion that was retrogradely labeled with WGA-HRP after injection of WGA-HRP into the cardia of the stomach in Case S300. Scale bar = 30 μm.
T. Hayakawa et al. / Autonomic Neuroscience: Basic and Clinical 134 (2007) 8–17
13
Fig. 7. A retrogradely WGA-HRP-labeled neuron in the celiac ganglion after injection of WGA-HRP into the pylorus of the stomach in Case S305 (ER—rough endoplasmic reticulum; g—Golgi apparatus; G—glial cell; N—nucleus; S—satellite cell). Arrows indicate retrogradely transported WGA-HRP reaction products. Scale bar = 5 μm.
cytoplasm (Fig. 4). The axon terminals contained several kinds of vesicles, such as round clear vesicles (40 to 50 nm in diameter, Fig. 5A, B, D), pleomorphic clear vesicles (20 to 60 nm long, Fig. 5C), large dense-cored vesicles (100 to 120 nm in diameter, Figs. 5 and 8), and glycogen-like small dark vesicles (less than 20 nm in diameter, Fig. 5A and B). Dense-cored vesicles were often found in both the terminals containing round vesicles and the terminals containing pleomorphic vesicles (Figs. 5 and 8). About half of the axodendritic terminals (50.2%) contained round synaptic vesicles. Both the terminals containing round vesicles and the terminals containing pleomorphic vesicles formed asymmetric synaptic contacts with the dendrites (Fig. 5). In the neuropil, several small dendrites and axon terminals merged and formed a bundle. The bundles were wrapped by the cytoplasm of glial cells, and were separated from each other (Fig. 4). There were many glial cells among the bundles. The primary dendrites were also covered with glial cytoplasm and received a few axon terminals (Fig. 5B). Unlike axon terminals in the brain, no terminals were found to make symmetric synaptic contacts. Sometimes the axon terminals contacted the smooth somatic membrane (Fig. 8B). The length of axodendritic terminals as a whole was 1.45 ± 0.04 μm (mean ± SE, n = 231). Most of them (67.6%) contacted small dendrites (less than 1 μm in diameter). A
few myelinated fibers were present. Only a few terminals (4.8%) contacted large or primary dendrites (more than 1.5 μm in diameter). There were no axoaxonic terminals, synapses en passant, or dendrodendritic appositions. When WGA-HRP was injected into the cardia, many retrogradely labeled neurons were found throughout the celiac ganglion. The labeled neurons were round or oval, and had four or five primary dendrites (Figs. 3 and 6). The neurons projecting to the cardia were medium-sized (33.3 ± 1.0 × 22.4 ± 0.7 μm, n = 24), round or oval in shape, with a mean area of 593.1 ± 27.0 μm2 per section. The labeled neurons had a dark cytoplasm containing an oval nucleus (13.5 ± 0.4 × 10.0 ± 0.3 μm) having one or two nucleoli, numerous free ribosomes, many mitochondria, many short strands of rough endoplasmic reticulum (rER), smooth endoplasmic reticulum (sER), lysosomes, and several Golgi apparatuses (Fig. 3). There were a few neurons containing aggregation of various sizes of granular vesicles. The cell surface was mostly covered with the thin cytoplasm of satellite cells. Thus, the average number of axosomatic terminals was small and 1.6 ± 0.3 per section through the nucleolus. The axosomatic terminals often contacted small processes extending from the soma (Fig. 8A). About 60% of the axosomatic terminals contained round synaptic vesicles, while the rest of them contained pleomorphic vesicles.
14
T. Hayakawa et al. / Autonomic Neuroscience: Basic and Clinical 134 (2007) 8–17
Fig. 8. Axon terminals contacting retrogradely labeled neurons that project to the pylorus or the cardia. (A) A somatic process extending from a neuron retrogradely labeled with WGA-HRP receiving two axon terminals (T) that contain round synaptic vesicles and that form asymmetric synaptic contacts (arrow) in the case of injection of WGA-HRP into the cardia of the stomach. (B) A terminal (T) containing round synaptic vesicles and forming asymmetric contacts (arrow) with the somatic membrane of a neuron retrogradely labeled with WGA-HRP in the case of injection of WGA-HRP into the pylorus of the stomach. (C) A terminal with anterogradely transported WGA-HRP reaction products (small arrow) forming asymmetric contacts (arrow) with a somatic process of a neuron retrogradely labeled with WGA-HRP in the case of injection of WGA-HRP into the pylorus of the stomach. Arrowheads indicate dense-cored vesicles, doublearrows indicate small glycogen-like granules, and open arrows indicate retrogradely transported WGA-HRP reaction products. Scale bars = 1 μm in A, B and C.
When WGA-HRP was injected into the pylorus, many retrogradely labeled neurons were found in the celiac ganglion. The neurons projecting to the pylorus were medium-sized (33.4 ± 0.9 × 24.7 ± 0.8 μm, n = 24), round or oval in shape, with a mean area of 650.9 ± 34.5 μm2 per section (Fig. 7). A lightly stained, prominent round or oval nucleus was present at the center of the soma, and the cells had one or two nucleoli. The neurons had a dark cytoplasm containing numerous free ribosomes, many short strands of rER, many sER, many mitochondria, lysosomes, several Golgi apparatuses and lysosomes, but only a few granular vesicles of various sizes. The somatic membrane was generally smooth and covered with the thin cytoplasm of the satellite cells. The neurons projecting to the pylorus received only a few axosomatic terminals (1.3 ± 0.3 per section through the nucleolus, n = 24). The axosomatic terminals contacted the smooth somatic membrane (Fig. 8B) or small somatic processes (Fig. 8C). About 60% of them contained round synaptic vesicles, and the rest of them contained pleomorphic vesicles (Table 1). Only two terminals were found to contain anterogradely WGA-HRP reaction products, which contact the retrogradely labeled neurons (Fig. 8C).
When WGA-HRP was injected into the whole ventral wall of the stomach including the cardia, the corpus, the antrum, and the pylorus, many retrogradely labeled neurons Table 1 Ultrastructural characteristics and the number of axosomatic terminals on the neurons in the celiac ganglia that innervate the pylorus, the cardia, or do not innervate the stomach (unlabeled neurons) Cardiac neurons (n = 24) 593.1 ± 27.0 Area (μm2) Cell length (μm) 33.3 ± 1.0 Cell width (μm) 22.4 ± 0.7 Nuclear length (μm) 13.5 ± 0.4 Nuclear width (μm) 10.0 ± 0.3 Axosomatic terminals 1.6 ± 0.3 Terminals containing 1.0 ± 0.2 round synaptic vesicles Terminals containing 0.6 ± 0.1 pleomorphic synaptic vesicles
Pyloric neurons (n = 24)
Unlabeled neurons (n = 22)
650.9 ± 34.5 33.4 ± 0.9 24.7 ± 0.8 13.5 ± 0.5 10.3 ± 0.4 1.3 ± 0.3 0.8 ± 0.2
872.8 ± 41.9⁎ 40.4 ± 1.0⁎ 28.0 ± 1.0⁎ 14.5 ± 0.4 11.7 ± 0.3 0.9 ± 0.2 0.4 ± 0.1
0.6 ± 0.1
0.5 ± 0.2
The values are the mean ± S.E. ⁎ Indicates statistically significant differences between the unlabeled neurons and the neurons innervating the pylorus or the cardia (P b 0.05).
T. Hayakawa et al. / Autonomic Neuroscience: Basic and Clinical 134 (2007) 8–17
were found throughout the celiac ganglion. Nevertheless, there were still many neurons in the celiac ganglion that were not labeled retrogradely. These unlabeled neurons were large (40.4 ± 1.0 × 28.0 ± 1.0 μm), and had a dark cytoplasm containing many mitochondria, numerous free ribosomes, many rough and smooth ER, several Golgi apparatuses, many lysosomes, and a light nucleus (14.5 ± 0.4 × 11.7 ± 0.3 μm) with one or two prominent nucleoli (Fig. 3). Sometimes they contained many lysosomes, lipofuscin granules, and granular vesicles of various sizes in the cytoplasm. Since the somatic membrane was mostly covered with the cytoplasm of the satellite cells, they received only a few axosomatic terminals (0.9 ± 0.2 per section through the nucleolus, n = 22). About 44% of the axosomatic terminals contained round synaptic vesicles, and the rest of them contained pleomorphic synaptic vesicles (Table 1). There were no significant differences in the area, the size, and the number of axosomatic terminals between the neurons projecting to the cardia and to the pylorus. In contrast, there were significant differences in the area (t = 5.61, P b 0.01), in the length (t = 4.93, P b 0.01), and in the width (t = 4.46, P b 0.01) between the neurons not projecting to the stomach and the neurons projecting to the cardia. Significant differences were also present in the area (t = 4.09, P b 0.01), in the length (t = 5.03, P b 0.01), and in the width (t = 2.49, P b 0.05) between the neurons not projecting to the stomach and the neurons projecting to the pylorus. However, there were no significant differences in the number of axosomatic terminals, the number of axosomatic terminals containing round vesicles or pleomorphic vesicles among these three kinds of the neurons. 4. Discussion The present study demonstrated that the cardia and the pylorus are innervated mostly by different neurons in the celiac ganglion. The numbers of retrogradely labeled neurons were variable between animals injected with FG and CTb into the cardia and the pylorus of the stomach, respectively. Nevertheless, the number of double-labeled neurons was small, and the percentages of double-labeled neurons for the FG-labeled (14.9 ± 2.2%) or the CTb-labeled neurons (7.4 ± 1.3%) were small and constant. Thus, there were only a few neurons in the celiac ganglion that send collateral fibers to both the cardia and the pylorus. The vagus nerve sends parasympathetic fibers to the stomach (Rogers and Hermann, 1983; Shapiro and Miselis, 1985; Norgren and Smith, 1988; Holst et al., 1997; Hayakawa et al., 2004, 2006). Retrograde tracing studies using two kinds of tracers showed that the cardia and the antrum of the stomach received collateral projections from the neurons in the DMV, while the pylorus was innervated by neurons that were different from the neurons projecting to the cardia in the DMV (Hayakawa et al., 2003). Based on these results and our observations, the cardia appears to be innervated and regulated by parasympathetic neurons and sympathetic
15
neurons that are different from the neurons projecting to the pylorus. Berthoud et al. (1991) reported that the cardia receives parasympathetic vagal projections by way of the gastric branch, and the pylorus receives projections by way of the hepatic branch of the vagus nerve. However, it is not clear if the neurons in the celiac ganglion run through the celiac branch and project to the cardia, and the neurons projecting to the pylorus traverse the hepatic branch. We also demonstrated that the celiac ganglion is composed of medium-sized to large neurons having a dark cytoplasm with many cell organelles, and a few axosomatic terminals. Sometimes small processes protruded from the soma and received axon terminals forming asymmetric synaptic contacts. The small processes corresponded to the accessory dendrites in the cat (Elfvin, 1971b) or the short spine-like processes in the mouse prevertebral ganglia (Miller et al., 1996). The present ultrastructural features of the celiac ganglion neurons of the rat are similar to those of the neurons of the superior mesenteric ganglion of the mouse (Miller et al., 1996), and the neurons of the inferior mesenteric ganglion of the guinea pig (Gibbins et al., 2000, 2003) and the cat (Elfvin, 1971a). Serial ultrathin sections of the neuron of the inferior mesenteric ganglion of the cat revealed that the neurons have a few axosomatic terminals and short processes (Elfvin, 1971a). In the prevertebral ganglia, there are small intensely fluorescent cells with many large granules in there sparse cytoplasm, the so-called SIF cells (Yokota and Burnstock, 1983; Elfvin et al., 1997). Only a few SIF cells were found in the present study. Szurszewski and Miller (1994) reported that the number and appearance of SIF cells are quite variable among species. In the rat neuropil, many axodendritic terminals were located around the neuronal soma. Unlike the guinea pig and the mouse prevertebral ganglia (Miller et al., 1996; Gibbins et al., 2003), several terminals and small dendrites coalesced to form a bundle surrounded by the glial cytoplasm. Light microscopic immunohistochemical studies showed that many terminal-like boutons are distributed around the soma, and contain nitric oxide synthase (NOS) or acetylcholine (Hamberger and Norberg, 1965; Alm et al., 1995; Elfvin et al., 1997; Gibbins et al., 2003). These terminals were considered to contact small dendrites near the soma, but not the somatic membrane, judging from the ultrastructural observations. There were many axoaxonic contacts and dendrodendritic appositions in the cat inferior mesenteric ganglion (Elfvin, 1971c; Miller et al., 1996), while in the rat celiac ganglion there were few axoaxonic terminals and dendrodendritic appositions. Immunohistochemical studies of the rat, the cat and the guinea pig have demonstrated that the terminals containing NOS, ENK, NT, and choline acetyltransferase (ChAT) come from the thoracic spinal cord, and the terminals containing VIP, CCK, DYN, bombesin, and CGRP come from the myenteric ganglion of the intestine (Hamaji et al., 1987; Messenger and Furness, 1992; Elfvin et al., 1993; Domoto et al., 1995). Electron microscopic studies combined with
16
T. Hayakawa et al. / Autonomic Neuroscience: Basic and Clinical 134 (2007) 8–17
immunohistochemistry of the prevertebral ganglia showed that the terminals containing pleomorphic vesicles also contain VIP in the guinea pig (Kondo and Yui, 1982b; Hamaji et al., 1989), while the terminals containing small round vesicles also contain NOS in the rat (Domoto et al., 1994) or ChAT, and the terminals containing large granular vesicles or dense-cored vesicles also contain tyrosine hydroxylase, the rate-limiting enzyme for NA synthesis in the rat, the cat, and the guinea pig (Leranth and Ungvary, 1980; Yokota and Burnstock, 1983; Smolen, 1988; Alm et al., 1995; Elfvin et al., 1997). Both the neurons projecting to the cardia and the neurons projecting to the pylorus had terminals containing round vesicles or pleomorphic vesicles, and these terminals often contained dense-cored vesicles in the present study. Thus, they may receive terminals using acetylcholine, VIP or NO as neurotransmitters or neuromodulators. The celiac ganglia neurons projecting to the stomach received projections from the thoracic spinal cord, but few from the neurons of the myenteric ganglion of the stomach (Messenger and Furness, 1992). Thus, the axosomatic terminals containing pleomorphic vesicles on the neurons projecting to the stomach may come from the intrinsic neurons in the prevertebral ganglia, but not from the myenteric ganglia of the alimentary canal. The retrogradely labeled neurons projecting to the cardia were similar to the labeled neurons projecting to the pylorus in their ultrastructural appearances, their size, the number of axosomatic terminals, and the ratio of the two kinds of terminals (Table 1). Based on peptide content, there are three kinds of neurons in the celiac ganglion: NA neurons that also contain NPY; NA neurons that also contain SOM; and NA neurons that do not contain either NPY or SOM (Macrae et al., 1986). Lindh et al. (1986) reported that the retrogradely labeled neurons projecting to the pylorus contain NA and NPY, or NA and SOM, but only a few labeled neurons contain only NA. There have been, however, no immunohistochemical studies about the neurons projecting to the cardia. There were large neurons in the celiac ganglion that do not project to the stomach in the present study. Although we could not label all neurons projecting to the stomach, the neurons projecting to the pylorus or the cardia were significantly smaller than the unlabeled neurons. Since the celiac ganglion neurons project not only to the stomach but also to the ileum and colon in the rat, most of these unlabeled neurons may project to the small and large intestine (Luckensmeyer and Keast, 1994; Quinson et al., 2001). Kuramoto and Furness (1989) demonstrated that the celiac ganglion neurons receive projections from the neurons in the myenteric ganglia in the duodenum, and the small and large intestine. Then they formed reciprocal neuronal circuitry for the regulation of alimentary canal activity (Furness, 2003). Thus, the unlabeled large neurons may project to the duodenum and the intestine, and receive projections from the intestinofugal neurons in the myenteric ganglion of the small intestine. The ultrastructural features of the large neurons were also similar to those of the neurons projecting
to the stomach except for their cell sizes. Sometimes there were large neurons, which contain many lysosomes and large granules. These neurons were often present in the inferior mesenteric, and the superior mesenteric ganglia of the mouse (Miller et al., 1996) and the cat (Elfvin, 1971a). Physiological studies showed that neurons innervating the pylorus cause contractions of the sphincter muscle using NA (Edin et al., 1979), while stimulation of the splanchnic nerve inhibited the contraction of the fundus and the cardia of the stomach (Fandriks et al., 1987). Electrophysiological studies have reported that there are three classes of neurons based on their potassium channel expression in excitability in the guinea pig celiac ganglion (Boyd et al., 1996). These three classes of neurons had a different number of dendrites and sizes of their neuronal soma. The neurons projecting to the pylorus and the cardia may have different functions, and immunohistochemical and morphological properties, though we could not see obvious ultrastructural differences between these neurons. Further immunohistochemical and electrophysiological studies will be needed on the neurons projecting to the cardia and the pylorus in the celiac ganglion. Acknowledgements The authors thank Ms. M. Hatta and Mr. K. Gion for their technical assistance. References Alm, P., Uvelius, B., Ekstrom, J., Holmqvist, B., Larsson, B., Andersson, K.E., 1995. Nitric oxide synthase-containing neurons in rat parasympathetic, sympathetic and sensory ganglia: a comparative study. Histochem. J. 27, 819–831. Berthoud, H.R., Carlson, N.R., Powley, T.L., 1991. Topography of efferent vagal innervation of the rat gastrointestinal tract. Am. J. Physiol. 260, R200–R207. Boyd, H.D., McLachlan, E.M., Keast, J.R., Inokuchi, H., 1996. Three electrophysiological classes of guinea pig sympathetic postganglionic neurone have distinct morphologies. J. Comp. Neurol. 369, 372–387. Domoto, T., Teramoto, M., Tamura, K., Yasui, Y., 1994. Ultrastructural study on NOS-immunoreactive nerve terminals in the rat coeliac ganglion. NeuroReport 169–172. Domoto, T., Teramoto, M., Tanigawa, K., Tamura, K., Yasui, Y., 1995. Origins of nerve fibers containing nitric oxide synthase in the rat celiacsuperior mesenteric ganglion. Cell Tissue Res. 281, 215–221. Edin, R., Ahlman, H., Kewenter, J., 1979. The vagal control of the feline pyloric sphincter. Acta Physiol. Scand. 107, 169–174. Elfvin, L.G., 1971a. Ultrastructural 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. Elfvin, L.G., 1971b. Ultrastructural studies on the synaptology of the inferior mesenteric ganglion of the cat: II. Specialized serial neuronal contacts between preganglionic end fibers. J. Ultrastruct. Res. 37, 426–431. Elfvin, L.G., 1971c. Ultrastructural studies on the synaptology of the inferior mesenteric ganglion of the cat: 3. The structure and distribution of the axodendritic and dendrodendritic contacts. J. Ultrastruct. Res. 37, 432–448. Elfvin, L.G., Lindh, B., Hokfelt, T., 1993. The chemical neuroanatomy of sympathetic ganglia. Annu. Rev. Neurosci. 16, 471–507. Elfvin, L.G., Holmberg, K., Emson, P., Schemann, M., Hokfelt, T., 1997. Nitric oxide synthase, choline acetyltransferase, catecholamine enzymes
T. Hayakawa et al. / Autonomic Neuroscience: Basic and Clinical 134 (2007) 8–17 and neuropeptides and their colocalization in the anterior pelvic ganglion, the inferior mesenteric ganglion and the hypogastric nerve of the male guinea pig. J. Chem. Neuroanat. 14, 33–49. Fandriks, L., Jonson, C., Nylander, O., 1987. Effects of splanchnic nerve stimulation and of clonidine on gastric and duodenal HCO3-secretion in the anaesthetized cat. Acta Physiol. Scand. 130, 251–258. Furness, J.B., 2003. Intestinofugal neurons and sympathetic reflexes that bypass the central nervous system. J. Comp. Neurol. 455, 281–284. Gibbins, I.L., Jobling, P., Messenger, J.P., Teo, E.H., Morris, J.L., 2000. Neuronal morphology and the synaptic organisation of sympathetic ganglia. J. Auton. Nerv. Syst. 81, 104–109. Gibbins, I.L., Teo, E.H., Jobling, P., Morris, J.L., 2003. Synaptic density, convergence, and dendritic complexity of prevertebral sympathetic neurons. J. Comp. Neurol. 455, 285–298. Hamaji, M., Kawai, Y., Kawashima, Y., Tohyama, M., 1987. Projections of bombesin-like immunoreactive fibers from the rat stomach to the celiac ganglion revealed by a double-labeling technique. Brain Res. 416, 192–194. Hamaji, M., Kawai, Y., Kawashima, Y., Tohyama, M., 1989. An electron microscopic study on VIP-, BOM- and CCK-like immunoreactive terminals in the celiac-superior mesenteric ganglion complex of the guinea pig. Brain Res. 488, 283–287. Hamberger, B., 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. Hayakawa, T., Zyo, K., 1992. Ultrastructural study of ascending projections to the lateral mammillary nucleus of the rat. Anat. Embryol. (Berl.) 185, 547–557. Hayakawa, T., Takanaga, A., Tanaka, K., Maeda, S., Seki, M., 2003. Cells of origin of vagal motor neurons projecting to different parts of the stomach in the rat: confocal laser scanning and electron microscopic study. Anat. Embryol. (Berl.) 207, 289–297. Hayakawa, T., Takanaga, A., Tanaka, K., Maeda, S., Seki, M., 2004. Distribution and ultrastructure of dopaminergic neurons in the dorsal motor nucleus of the vagus projecting to the stomach of the rat. Brain Res. 1006, 66–73. Hayakawa, T., Kuwahara, S., Maeda, S., Tanaka, K., Seki, M., 2006. Direct synaptic contacts on the myenteric ganglia of the rat stomach from the dorsal motor nucleus of the vagus. J. Comp. Neurol. 498, 352–362. Holst, M.C., Kelly, J.B., Powley, T.L., 1997. Vagal preganglionic projections to the enteric nervous system characterized with Phaseolus vulgaris-leucoagglutinin. J. Comp. Neurol. 381, 81–100. Jobling, P., Gibbins, I.L., 1999. Electrophysiological and morphological diversity of mouse sympathetic neurons. J. Neurophysiol. 82, 2747–2764. Kondo, H., Yamamoto, M., 1988. The ontogeny and fine structure of calcitonin gene-related peptide (CGRP)-immunoreactive nerve fibers in the celiac ganglion of rats. Arch. Histol. Cytol. 51, 91–98. Kondo, H., Yui, R., 1981. An electron microscopic study on substance P-like immunoreactive nerve fibers in the celiac ganglion of guinea pigs. Brain Res. 222, 134–137. Kondo, H., Yui, R., 1982a. An electron microscopic study on enkephalinlike immunoreactive nerve fibers in the celiac ganglion of guinea pigs. Brain Res. 252, 142–145.
17
Kondo, H., Yui, R., 1982b. An electron microscopic study on VIP-like immunoreactive nerve fibers in the celiac ganglion of guinea pigs. Brain Res. 237, 227–231. Kuramoto, H., Furness, J.B., 1989. Distribution of enteric nerve cells that project from the small intestine to the coeliac ganglion in the guinea-pig. J. Auton. Nerv. Syst. 27, 241–248. Leranth, C., Ungvary, G., 1980. Axon types of prevertebral ganglia and the peripheral autonomic reflex arc. J. Auton. Nerv. Syst. 1, 265–281. Lindh, B., Hokfelt, T., Elfvin, L.G., Terenius, L., Fahrenkrug, J., Elde, R., Goldstein, M., 1986. Topography of NPY-, somatostatin-, and VIPimmunoreactive, neuronal subpopulations in the guinea pig celiacsuperior mesenteric ganglion and their projection to the pylorus. J. Neurosci. 6, 2371–2383. Luckensmeyer, G.B., Keast, J.R., 1994. Projections from the prevertebral and major pelvic ganglia to the ileum and large intestine of the male rat. J. Auton. Nerv. Syst. 49, 247–259. Lundberg, J.M., Hokfelt, T., Anggard, A., Terenius, L., Elde, R., Markey, K., Goldstein, M., Kimmel, J., 1982. Organizational principles in the peripheral sympathetic nervous system: subdivision by coexisting peptides (somatostatin-, avian pancreatic polypeptide-, and vasoactive intestinal polypeptide-like immunoreactive materials). Proc. Natl. Acad. Sci. U. S. A. 79, 1303–1307. Macrae, I.M., Furness, J.B., Costa, M., 1986. Distribution of subgroups of noradrenaline neurons in the coeliac ganglion of the guinea-pig. Cell Tissue Res. 244, 173–180. Messenger, J.P., Furness, J.B., 1992. Distribution of enteric nerve cells that project to the coeliac ganglion of the guinea-pig. Cell Tissue Res. 269, 119–132. Miller, S.M., Hanani, M., Kuntz, S.M., Schmalz, P.F., Szurszewski, J.H., 1996. Light, electron, and confocal microscopic study of the mouse superior mesenteric ganglion. J. Comp. Neurol. 365, 427–444. Norgren, R., Smith, G.P., 1988. Central distribution of subdiaphragmatic vagal branches in the rat. J. Comp. Neurol. 273, 207–223. Quinson, N., Robbins, H.L., Clark, M.J., Furness, J.B., 2001. Locations and innervation of cell bodies of sympathetic neurons projecting to the gastrointestinal tract in the rat. Arch. Histol. Cytol. 64, 281–294. Rogers, R.C., Hermann, G.E., 1983. Central connections of the hepatic branch of the vagus nerve: a horseradish peroxidase histochemical study. J. Auton. Nerv. Syst. 7, 165–174. Roman, C., Gonella, J., 1981. In: Johnson, L.R. (Ed.), Extrinisic Control of Digestive Tract Motility. Rave Press, New York. 289–333 pp. Shapiro, R.E., Miselis, R.R., 1985. The central organization of the vagus nerve innervating the stomach of the rat. J. Comp. Neurol. 238, 473–488. Smolen, A.J., 1988. Morphology of synapses in the autonomic nervous system. J. Electron. Microsc. Tech. 10, 187–204. Szurszewski, J.H., Miller, S.M., 1994. In: Johnson, L.R. (Ed.), Physiology of Prevertebral Ganglion. Raven Press, New York. 795-877 pp. Yokota, R., Burnstock, G., 1983. Synaptic organisation of the pelvic ganglion in the guinea-pig. Cell Tissue Res. 232, 379–397.