Edward A. Fox

Longitudinal columnar organization within the dorsal motor nucleus represents separate branches of the abdominal vagus

To identify the distribution of central preganglionics associated with each branch of the subdiaphragmatic vagus, the fluorescent tracer True Blue (TB) was administered intraperitoneally to rats with 4 out of 5 branches cauterized, and then, after 72 h, the animals were sacrificed for histological analysis. Each vagal branch contained the axons of a topographically distinct column of cells within the dorsal motor nucleus of the vagus (DMN). The columns representing the 4 branches with the largest numbers of efferents are organized as paired, bilaterally symmetrical, longitudinal distributions on either side of the medulla. Each DMN side contains a column occupying the medial two-thirds or more of the nucleus and corresponding to one of the gastric branches (left DMN, anterior gastric; right DMN, posterior gastric). Also on each side, the lateral pole of the DMN consists of a coherent cell column corresponding to one of the celiac branches (left DMN, accessory celiac; right DMN, celiac). The fifth branch, the hepatic, is represented by a limited number of somata forming a diffuse column largely coextensive with that representing the anterior gastric branch. At some levels of the DMN, the columns overlap. Labeled cells observed in the reticular formation were correlated in number, left-right ratios and response to vagotomy with those in the DMN, which suggests that they are displaced cells of the nucleus. Distributions of labeled cells in the nucleus ambiguus and the retrofacial nucleus were not tightly correlated with those of the DMN. An analysis of cell counts obtained for each of the individual branches suggests that vagal axons do not generally send collaterals through more than one branch.

Vagal afferent innervation of smooth muscle in the stomach and duodenum of the mouse: morphology and topography.

Intraganglionic laminar endings (IGLEs) and intramuscular arrays (IMAs), the two putative mechanoreceptors that the vagus nerve supplies to the gastrointestinal smooth muscle, have been characterized almost exclusively in the rat. To provide normative inventories of these afferents for the mouse, the authors examined the endings in the stomach and small intestine of three strains used as backgrounds for gene manipulations (i.e., C57, 129/SvJ, and WBB6). Animals received nodose ganglion injections of wheat germ agglutinin‐horseradish peroxidase or dextran‐tetramethylrhodamine conjugated to biotin. The horseradish peroxidase tissue was processed with tetramethylbenzidine and was used to map the distributions and densities of the two endings; the dextran material was counterstained with c‐Kit immunohistochemistry to assess interactions between intramuscular arrays and interstitial cells of Cajal. IGLEs and IMAs constituted the vagal innervation of mouse gastric and duodenal smooth muscle. IGLE morphology and distributions, with peak densities in the corpus‐antrum, were similar in the three strains of mice and comparable to those observed in rats. IMAs varied in complexity from region to region but tended to be simpler (fewer telodendria) in mice than in rats. IMAs were most concentrated in the forestomach and sphincters in mice, as in rats, but the topographic distributions of the endings varied both between strains of mice (subtly) and between species (more dramatically). IMAs appeared to make appositions with both interstitial cells and smooth muscle fibers. This survey should make it practical to assay the effects of genetic (e.g., knockout) and experimental (e.g., regeneration) manipulations affecting visceral afferents and their target tissues. J. Comp. Neurol. 428:558–576, 2000.

Selective loss of vagal intramuscular mechanoreceptors in mice mutant for steel factor, the c-Kit receptor ligand

Vagal intramuscular arrays are mechanoreceptors that innervate smooth muscle fibers and intramuscular interstitial cells of Cajal of the proximal GI tract. C-Kit mutant mice that lack intramuscular interstitial cells of Cajal also lack intramuscular arrays. Mice mutant for steel factor, the ligand for the c-Kit receptor, were studied to extend and validate these previous findings and to characterize associated changes in food intake. Injections of wheat germ agglutinin-horseradish peroxidase and of dextran into the nodose ganglion were employed to label intramuscular arrays and intraganglionic laminar endings, the other vagal mechanoreceptors found in the gut wall. These two receptor types were inventoried in wholemounts of the stomach and duodenum using a standardized sampling and quantification regime. Steel mutants exhibited a paucity of normal intramuscular arrays and lacked intramuscular interstitial cells of Cajal in the forestomach, whereas their intraganglionic laminar endings appeared normal in number, distribution, and morphology. These observations suggest that intramuscular array losses in steel and c-Kit mutants are specific and result from the elimination of the intramuscular interstitial cells of Cajal, the effect common to both mutations, not from interactions peculiar to background strains or non-specific effects. Double-labeling analyses of intramuscular arrays and intramuscular interstitial cells of Cajal reinforced the hypothesis based on previous findings in the c-Kit mice that these interstitial cells have a trophic effect on intramuscular array development and/or maintenance. Finally, meal pattern analyses revealed decreased meal size and increased meal frequency in steel mutants, with normal daily intake. These alterations suggest short-term feeding controls are affected by the loss of intramuscular arrays and/or intramuscular interstitial cells of Cajal, though long-term controls are unimpaired.

C-Kit mutant mice have a selective loss of vagal intramuscular mechanoreceptors in the forestomach.

Intramuscular arrays are one of two major classes of vagal afferent mechanoreceptors that innervate the smooth muscle wall of the proximal gastrointestinal tract. They consist of rectilinear telodendria that distribute in the muscle sheets, parallel to the long axes of muscle fibers. Intramuscular arrays appear to make direct contact with the muscle fibers, but they also course on, and form appositions with, intramuscular interstitial cells of Cajal. These complexes formed by intramuscular arrays and intramuscular interstitial cells of Cajal suggest that intramuscular arrays might require either structural or trophic support of the interstitial cells of Cajal for normal differentiation and/or maintenance. To evaluate this hypothesis, we have examined the morphology and distribution of vagal afferent endings in the c-Kit mutant mouse that lacks intramuscular interstitial cells of Cajal. Vagal afferents were labeled by nodose ganglion injection of either wheat germ agglutinin-horseradish peroxidase conjugate or a tagged dextran, and the labeled afferent terminals in the stomach were mapped using a standardized quantitative sampling scheme. Intramuscular arrays were dramatically reduced (in circular muscle by 63%; in longitudinal muscle by 78%) in the c-Kit mutant mice relative to their wild-type littermates. Additionally, a substantial number of the surviving axons and terminals in the mutant stomachs were morphologically aberrant. Moreover, the loss of intramuscular arrays in mutants appeared to be selective: the structure, distribution and density of intraganglionic laminar endings, i.e., the other vagal mechanoreceptors in smooth muscle, were not significantly altered. Finally, the conspicuous decrease in intramuscular array density in mutants was associated with a non-significant trend toward loss of nodose ganglion neurons. Collectively these findings suggest that interstitial cells are required for the normal development or maintenance of vagal intramuscular arrays. Therefore, the c-Kit mutant mouse will be valuable for determining the role(s) of interstitial cells in intramuscular array development as well as for providing an animal model with the intramuscular array class of vagal afferents selectively ablated.

Ultrastructural evidence for communication between intramuscular vagal mechanoreceptors and interstitial cells of Cajal in the rat fundus

Abstract  To assess whether afferent vagal intramuscular arrays (IMAs), putative gastrointestinal mechanoreceptors, form contacts with interstitial cells of Cajal of the intramuscular type (ICC‐IM) and to describe any such contacts, electron microscopic analyses were performed on the external muscle layers of the fundus containing dextran‐labelled diaminobenzidin (DAB)‐stained IMAs. Special staining and embedding techniques were developed to preserve ultrastructural features. Within the muscle layers, IMA varicosities were observed in nerve bundles traversing major septa without contact with ICC‐IM, contacting unlabelled neurites and glial cells. IMA varicosities were encountered in minor septa in contact with ICC‐IM which were not necessarily in close contact with muscle cells. In addition, IMA varicosities were observed within muscle bundles in close contact with ICC‐IM which were in gap junction contact with muscle cells. IMAs formed varicosities containing predominantly small agranular vesicles, occasionally large granular vesicles and prejunctional thickenings in apposition to ICC‐IM processes, indicating communication between ICC and IMA via synapse‐like contacts. Taken together, these different morphological features are consistent with a hypothesized mechanoreceptor role for IMA‐ICC complexes. Intraganglionic laminar ending varicosities contacted neuronal somata and dendrites in the myenteric plexus of the fundus, but no contacts with ICC associated with Auerbach’s plexus were encountered.

Anatomical considerations for surgery of the rat abdominal vagus: distribution, paraganglia and regeneration.

In order to provide a detailed surgical anatomy of the rat abdominal vagus, we examined pyridine silver-stained tissue from one group of normal animals and a second group that survived 9 months after vagotomy. In the normal sample, as has been established for man, there was considerable variability in the levels at which each of the vagal branches separated from the main trunks. Contrary to reports from dissection studies, most of the branches were not single fiber bundles but rather consisted of two or more separate bundles. At the extreme, the posterior gastric and coeliac branches each consisted of as many as 15 individual bundles. Even the main trunks of the subdiaphragmatic vagus were occasionally observed to have multiple components (anterior trunk, 13% of the cases; posterior, 25%). In addition to the classically recognized hepatic, anterior gastric, coeliac, and posterior gastric branches, we also observed an accessory coeliac branch of the anterior trunk in all animals. This accessory coeliac division originated just caudal to the hepatic branching and extended first laterally and then dorsally while running caudally to exit from the esophagus just before the separation of the coeliac branch from the posterior trunk. The vagi were observed to contain paraganglia consisting of islands of glomus cells, neurons, and extensive capillary beds, all situated within the perineurium. The paraganglia occurred in greatest frequency at the sites where the hepatic and coeliac branches divide from their respective trunks. Paraganglia were also observed peripherally within vagal branches; there they were most numerous within the coeliac branch and least numerous in the accessory coeliac. Other studies yielded evidence that regeneration had occurred after complete vagotomy. First, stumps of the branches distal to the resection scar contained axons. Central to the scar, axons grew out in all directions from the neuroma; some of them appeared to cross the scar and to reinnervate the distal stumps. Secondly, 30% of the animals in which regeneration was thought to be possible increased their insulin secretion in response to electrical stimulation of the cervical vagus. The implications of the above findings for experiments that involve manipulation or recording of the vagus are discussed.

Tracer diffusion has exaggerated CNS maps of direct preganglionic innervation of pancreas.

Small injections of True Blue (TB) into 4 different segments of the pancreas of the rat resulted in characteristic and different numbers and distributions of labeled cells within the dorsal motor nucleus of the vagus (DMN). Composites of these patterns of labeled cells in the DMN closely matched the distributions previously reported for more extensive injections of retrograde tracers into the pancreas. However, the application of a diffusion barrier (formed with a plastic wound spray) on the outer surface of the stomach and intestines adjacent to the pancreas segment which contained the TB injection depot prevented virtually all of the labeling of DMN cells. Similarly, applying a diffusion barrier directly to the injected pancreas segment itself prevented all or most (mean greater than 99%) of the DMN labeling. In contrast to this effect on DMN label, the barrier reduced more modestly the labeling of celiac ganglion somata after pancreas injections. Several additional control experiments also suggest that the absence of DMN label after the barrier application resulted from interference with tracer diffusion from the injected organ and not from neurotoxic effects. These include the following demonstrations in the presence of the barrier: (1) observation of unimpaired vagally stimulated insulin secretion, (2) uncompromised cell labeling of DMN from other organs treated with TB plus the barrier, and (3) normal hematoxylin and eosin stained pancreas tissue which had received TB injections and barrier application. It was concluded that both the number of parasympathetic preganglionic neurons that project monosynaptically to the pancreas and their distribution in the medulla may have been very significantly overestimated in previous tracer studies.

A genetic approach for investigating vagal sensory roles in regulation of gastrointestinal function and food intake

Sensory innervation of the gastrointestinal (GI) tract by the vagus nerve plays important roles in regulation of GI function and feeding behavior. This innervation is composed of a large number of sensory pathways, each arising from a different population of sensory receptors. Progress in understanding the functions of these pathways has been impeded by their close association with vagal efferent, sympathetic, and enteric systems, which makes it difficult to selectively label or manipulate them. We suggest that a genetic approach may overcome these barriers. To illustrate the potential value of this strategy, as well as to gain insights into its application, investigations of CNS pathways and peripheral tissues involved in energy balance that benefited from the use of gene manipulations are reviewed. Next, our studies examining the feasibility of using mutations of developmental genes for manipulating individual vagal afferent pathways are reviewed. These experiments characterized mechanoreceptor morphology, density and distribution, and feeding patterns in four viable mutant mouse strains. In each strain a single population of vagal mechanoreceptors innervating the muscle wall of the GI tract was altered, and was associated with selective effects on feeding patterns, thus supporting the feasibility of this strategy. However, two limitations of this approach must be addressed for it to achieve its full potential. First, mutation effects in tissues outside the GI tract can contribute to changes in GI function or feeding. Additionally, knockouts of developmental genes are often lethal, preventing analysis of mature innervation and ingestive behavior. To address these issues, we propose to develop conditional gene knockouts restricted to specific GI tract tissues. Two genes of interest are brain-derived neurotrophic factor (BDNF) and neurotrophin-3 (NT-3), which are essential for vagal afferent development. Creating conditional knockouts of these genes requires knowledge of their GI tract expression during development, which little is known about. Preliminary investigation revealed that during development BDNF and NT-3 are each expressed in several GI tract regions, and that their expression patterns overlap in some tissues, but are distinct in others. Importantly, GI tissues that express BDNF or NT-3 are innervated by vagal afferents, and expression of these neurotrophins occurs during the periods of axon invasion and receptor formation, consistent with roles for BDNF or NT-3 in these processes and in receptor survival. These results provide a basis for targeting BDNF or NT-3 knockouts to specific GI tract tissues, and potentially altering vagal afferent innervation only in that tissue (e.g., smooth muscle vs. mucosa). Conditional BDNF or NT-3 knockouts that are successful in selectively altering a vagal GI afferent pathway will be valuable for developing an understanding of that pathways roles in GI function and food intake.

Development of the Vagal Innervation of the GUT: Steering the Wandering Nerve

Background  The vagus nerve is the major neural connection between the gastrointestinal tract and the central nervous system. During fetal development, axons from the cell bodies of the nodose ganglia and the dorsal motor nucleus grow into the gut to find their enteric targets, providing the vagal sensory and motor innervations respectively. Vagal sensory and motor axons innervate selective targets, suggesting a role for guidance cues in the establishment of the normal pattern of enteric vagal innervation.

Abdominal pathways and central origin of rat vagal fibers that stimulate gastric acid

The brainstem location and peripheral course of the vagal preganglionic fibers that stimulate gastric acid secretion were identified using electrical stimulation combined with retrograde (True Blue; Dr. K. G. Illing, Gross Umstadt, Germany) and anterograde (Dil; Molecular Probes) fluorescent neural tracers in rats with various selective vagotomies. Animals with only one or both gastric branch(es) spared had normal, large gastric acid responses to electrical stimulation of the ipsilateral cervical vagus and showed an abundance of Dil-labeled vagal fibers and terminals in the gastric myenteric plexus. Rats with only the unpaired hepatic branch spared had a much smaller but significant gastric acid response and a few labeled vagal profiles in the antral region of the stomach. In contrast, rats with only one or both celiac branch(es) intact had neither a gastric acid response, nor evidence for Dil transport to the stomach. Retrograde transport of True Blue through the spared vagal axons to the brainstem indicated that the cell bodies of the preganglionics that send their axons through the acid-positive gastric and hepatic branches occupy the medial longitudinal columnar subnuclei of the dorsal motor nucleus. It is concluded that besides the long-recognized gastric branches, which are the major access route to the parietal cells, the hepatic branch contains a small number of fibers that most likely reach the antrum through the right gastroepiploic artery along the greater curvature and/or the right gastric artery.