Distribution, morphology and projections of nitrergic and non-nitrergic submucosal neurons in the pig small intestine

 Sequential nitric oxide synthase immunohistochemistry and nicotinamide adenine dinucleotide phosphate diaphorase (NADPHd) histochemistry in pig small intestinal wholemounts revealed a complete colocalisation of the two nitrergic markers in submucous neurons. The external submucous plexus (ESP) contained nitrergic neurons throughout. In the internal submucous plexus (ISP) we found a moderate number of nitrergic neurons in the duodenum, while they were rare in the jejunum and nearly absent in the ileum. Combined NADPHd histochemistry and silver impregnation showed morphological ESP type III and VI neurons to be NADPHd positive whereas ESP type II, IV and V neurons were NADPHd negative. Axons of ESP type III, IV and VI neurons were often observed to enter interconnecting strands directed abluminally. ESP type II neurons projected mainly to the ISP. In special silver-impregnated wholemounts containing both external muscle layers and the abluminal part of the submucous layer, i.e. the myenteric plexus and the ESP, the great majority of impregnated axons within the interconnecting strands were observed to run between both plexuses and did not enter the circular muscle layer. We conclude that ESP type III and VI neurons are nitrergic while ESP type II, IV and V neurons are non-nitrergic. Furthermore, we assume that ESP type III, IV and VI neurons may represent a submucosal input to the myenteric plexus. Accepted: 26 August 1997     

Spiny versus stubby: 3D reconstruction of human myenteric (type I) neurons

We have compared the three-dimensional (3D) morphology of stubby and spiny neurons derived from the human small intestine. After immunohistochemical triple staining for leu-enkephalin (ENK), vasoactive intestinal peptide (VIP) and neurofilament (NF), neurons were selected and scanned based on their immunoreactivity, whether ENK (stubby) or VIP (spiny). For the 3D reconstruction, we focused on confocal data pre-processing with intensity drop correction, non-blind deconvolution, an additional compression procedure in z-direction, and optimizing segmentation reliability. 3D Slicer software enabled a semi-automated segmentation based on an objective threshold (interrater and intrarater reliability, both 0.99). We found that most dendrites of stubby neurons emerged only from the somal circumference, whereas in spiny neurons, they also emerged from the luminal somal surface. In most neurons, the nucleus was positioned abluminally in its soma. The volumes of spiny neurons were significantly larger than those of stubby neurons (total mean of stubbies 806 ± 128 μm³, of spinies 2,316 ± 545 μm³), and spiny neurons had more dendrites (26.3 vs. 11.3). The ratios of somal versus dendritic volumes were 1:1.2 in spiny and 1:0.3 in stubby neurons. In conclusion, 3D reconstruction revealed new differences between stubby and spiny neurons and allowed estimations of volumetric data of these neuron populations.     

Structure of the tertiary component of the myenteric plexus in the guinea-pig small intestine

The tertiary component of the myenteric plexus consists of interlacing fine nerve fibre bundles that run between its principal ganglia and connecting nerve strands. It was revealed by zinc iodide-osmium impregnation and substance P immunohistochemistry at the light-microscope level. The plexus was situated against the inner face of the longitudinal muscle and was present along the length of the small intestine at a density that did not vary markedly from proximal to distal. Nerve bundles did not appear to be present in the longitudinal muscle as judged by light microscopy, although numberous fibre bundles were encountered within the circular muscle layer. At the ultrastructural level, nerve fibre bundles of the tertiary plexus were found in grooves formed by the innermost layer of longitudinal smooth muscle cells. In the distal parts of the small intestine, some of these nerve fibre bundles occasionally penetrated the longitudinal muscle coat. Vesiculated profiles in nerve fibre bundles of the tertiary plexus contained variable proportions of small clear and large granular vesicles; they often approached to within 50–200 nm of the longitudinal smooth muscle cells. Fibroblast-like cells lay between strands of the tertiary plexus and the circular muscle but were never intercalated between nerve fibre varicosities and the longitudinal muscle. These anatomical relationships are consistent with the tertiary plexus being the major site of neurotransmission to the longitudinal muscle of the guinea-pig small intestine.     

Immunohistochemical analysis of neuron types in the mouse small intestine

The definition of the nerve cell types of the myenteric plexus of the mouse small intestine has become important, as more researchers turn to the use of mice with genetic mutations to analyze roles of specific genes and their products in enteric nervous system function and to investigate animal models of disease. We have used a suite of antibodies to define neurons by their shapes, sizes, and neurochemistry in the myenteric plexus. Anti-Hu antibodies were used to reveal all nerve cells, and the major subpopulations were defined in relation to the Hu-positive neurons. Morphological Type II neurons, revealed by anti-neurofilament and anti-calcitonin gene-related peptide antibodies, represented 26% of neurons. The axons of the Type II neurons projected through the circular muscle and submucosa to the mucosa. The cell bodies were immunoreactive for choline acetyltransferase (ChAT), and their terminals were immunoreactive for vesicular acetylcholine transporter (VAChT). Nitric oxide synthase (NOS) occurred in 29% of nerve cells. Most were also immunoreactive for vasoactive intestinal peptide, but they were not tachykinin (TK)-immunoreactive, and only 10% were ChAT-immunoreactive. Numerous NOS terminals occurred in the circular muscle. We deduced that 90% of NOS neurons were inhibitory motor neurons to the muscle (26% of all neurons) and 10% (3% of all neurons) were interneurons. Calretinin immunoreactivity was found in a high proportion of neurons (52%). Many of these had TK immunoreactivity. Small calretinin neurons were identified as excitatory neurons to the longitudinal muscle (about 20% of neurons, with ChAT/calretinin/± TK chemical coding). Excitatory neurons to the circular muscle (about 10% of neurons) had the same coding. Calretinin immunoreactivity also occurred in a proportion of Type II neurons. Thus, over 90% of neurons in the myenteric plexus of the mouse small intestine can be currently identified by their neurochemistry and shape.     

Morphology of enkephalin-immunoreactive myenteric neurons in the human gut

The aim of this study was the morphological and further chemical characterisation of neurons immunoreactive for leu-enkephalin (leuENK). Ten wholemounts of small and large intestinal segments from nine patients were immunohistochemically triple-stained for leuENK/neurofilament 200 (NF)/substance P (SP). Based on their simultaneous NF-reactivity and 3D reconstruction of single NF-reactive cells, 97.5% of leuENK-positive neurons displayed the appearance of stubby neurons: small somata; short, stubby dendrites and one axon. Of these leuENK-reactive stubby neurons, 91.3% did not display co-reactivity for SP whereas 8.7% were SP-co-reactive. As to their axonal projection pattern, 50.4% of the recorded leuENK stubby neurons had axons running orally whereas in 29.4% they ran anally; the directions of the remaining 20.2% could not be determined. No axons were seen to enter into secondary strands of the myenteric plexus. Somal area measurements revealed clearly smaller somata of leuENK-reactive stubby neurons (between 259±47 m² and 487±113 m²) than those of putative sensory type II neurons (between 700±217 m² and 1,164±396 m²). The ratio dendritic field area per somal area of leuENK-reactive stubby neurons was between 2.0 and 2.8 reflecting their short dendrites. Additionally, we estimated the proportion of leuENK-positive neurons in comparison to the putative whole myenteric neuron population in four leuENK/anti-Hu doublestained wholemounts. This proportion ranged between 5.9% and 8.3%. We suggest leuENK-reactive stubby neurons to be muscle motor neurons and/or ascending interneurons. Furthermore, we explain why we do not use the term Dogiel type I neurons for this population.     

c-kit immunoreactive interstitial cells of Cajal in the human small and large intestine

 c-kit immunohistochemistry was performed on unfixed frozen sections of human small (duodenum, jejunum, and ileum) and large intestine (ascending, transverse, descending, and sigmoid colon). The c-kit immunoreactive cells in the muscularis externa of the intestinal wall were identified as interstitial cells of Cajal (ICC) and mast cells. ICC were identified by their morphology, localization, and organization based on previous light and electron microscopic studies. In the small intestine, ICC were located primarily in relation to the myenteric plexus of Auerbach, but also in septa between circular muscle lamellae. In the large intestine, ICC were seen in relation to Auerbach’s plexus, but also and in great numbers in the circular muscle layer and in teniae of the longitudinal muscle layer. The morphology of the ICC was similar in the small and large intestine, but the pattern of distribution was obviously different. c-kit immunoreactive mast cells were found predominantly in the inner part of the circular muscle layer. The anti-c-kit method is found to be an easy and reliable method to study at least most of the interstitial cells of Cajal and thereby contribute to further normal and pathological studies. Accepted: 31 July 1997     

Morphology of VIP/nNOS-immunoreactive myenteric neurons in the human gut

In this study, we characterized human myenteric neurons co-immunoreactive for neuronal nitric oxide synthase (nNOS) and vasoactive intestinal peptide (VIP) by their morphology and their proportion as related to the putative entire myenteric neuronal population. Nine wholemounts (small and large intestinal samples) from nine patients were triple-stained for VIP, neurofilaments (NF) and nNOS. Most neurons immunoreactive for all three markers displayed radially emanating, partly branching dendrites with spiny endings. These neurons were called spiny neurons. The spiny character of their dendrites was more pronounced in the small intestinal specimens and differed markedly from enkephalinergic stubby neurons described earlier. Exclusively in the duodenum, some neurons displayed prominent main dendrites with spiny side branches. Of the axons which could be followed from the ganglion of origin within primary strands of the myenteric plexus beyond the next ganglion (70 out of 140 traced neurons), 94.3% run anally and 5.7% orally. Very few neurons reactive for both VIP and nNOS could not be morphologically classified due to weak or absent NF-immunoreactivity. Another six wholemounts were triple-stained for VIP, nNOS and Hu proteins (HU). The proportion of VIP/nNOS-coreactive neurons in relation to the number of HU-reactive neurons was between 5.8 and 11.5% in the small and between 10.6 and 17.5% in the large intestinal specimens. We conclude that human myenteric spiny neurons co-immunoreactive for VIP and nNOS represent either inhibitory motor or descending interneurons.     

Quantitative estimation of putative primary afferent neurons in the myenteric plexus of human small intestine

This study aimed at estimating the proportion of human myenteric Dogiel type II neurons, putative intrinsic primary afferent neurons (IPANs), in relation to the entire myenteric neuron population. Since, at present, there is no known single marker, which specifically labels these neurons, we tried to identify the most appropriate marker combination based on the results of an earlier study. For this purpose, 10 wholemounts derived from human small intestinal segments were immunohistochemically triple-stained for calretinin (CALR), somatostatin (SOM) and neurofilaments (NF) and 9 were stained for substance P (SP), SOM and NF. In each wholemount, 15 ganglia selected randomly were evaluated. On the basis of their NF-reactivity, neurons reactive for one or co-reative for both of the other two markers, respectively, were morphologically classified as type II or non-type II neurons. We found that the majorities of neurons co-reactive for CALR/SOM and SP/SOM, respectively, were type II neurons whereas this was not the case for neurons, which were reactive for only one of the two markers. One of the statistical parameters estimated was the positive predictive value, the probability that a neuron displaying CALR/SOM- or SP/SOM-co-reactivity, respectively, is a type II neuron. This value was 97% in case of CALR/SOM- and 95% in case of SP/SOM-co-staining. Although the difference of the statistical parameters between the two stainings was not significant, CALR and SOM were used to estimate indirectly the proportion of type II neurons, in wholemounts co-stained with the pan-neuronal marker neuronal protein HuC/HuD (HU). In these wholemounts, altogether 9.1% of neurons were coreactive for CALR/SOM. We suggest that the proportion of myenteric type II neurons in the human small intestine is related to the proportion of CALR/SOM-co-reactive neurons and may be near to one tenth of the total myenteric neuronal population.     

Intramural distribution of immunoreactive vasoactive intestinal polypeptide (VIP), substance P,somatostatin and mammalian bombesin in the oesophago-gastro-pyloric region of the human gut

Summary The intramural distribution of vasoactive intestinal polypeptide (VIP), substance P, somatostatin and mammalian bombesin was studied in the oesophago-gastro-pyloric region of the human gut. At each of 21 sampling sites encompassing this entire area, the gut wall was separated into mucosa, submucosa and muscularis externa, and extracted for radioimmunoassay. VIP levels in the mucosa were very high in the proximal oesophagus (1231±174 pmol/g, mean±SEM) and showed varied, but generally decreasing concentrations towards the stomach, followed by a clear-cut increase across the pyloric canal (distal antrum: 73±16 pmol/g, proximal duodenum: 366±62 pmol/ g); consistent levels were found in submucosa and muscle (200–400 pmol/g) at most sites, the stomach again showing lower concentrations. By contrast, substance P was present in small amounts as far as the proximal stomach, but sharply increased across the pyloric canal, especially in mucosa and submucosa (distal antrum: 20±6.5 and 5.5±1.3 pmol/g; proximal duodenum: 62±8.5 and 34±11 pmol/g, respectively). Somatostatin concentrations were very low in the mucosa of the oesophagus and stepwise increased in the cardiac, mid-gastric and pyloric mucosa (cardia: 224±72 pmol/g; distal antrum: 513±152 pmol/g; proximal duodenum: 1013±113 pmol/g); concentrations in the submucosa and muscularis were generally low, with the exception of antrum and duodenum. Mammalian bombesin was comparatively well represented throughout the oesophageal muscularis (5–8 pmol/g), but most abundant in the stomach in all layers (oxyntic mucosa: 24±2.7 pmol/g; submucosa: 20±5.7 pmol/g; muscle: 28±5.0 pmol/g). In conclusion, a distinct differential distribution of the four peptides studied was revealed, indicating a diffuse, but highly differentiated peptide-containing innervation of the proximal human gut.     

Different distribution of S-100 protein and glial fibrillary acidic protein (GFAP) immunoreactive cells and their relations with nitrergic neurons in the human fetal small intestine.

The appearance, distribution and some histochemical features of non-neuronal cells (NN cells) associated with the myenteric plexus of human fetal small intestine have been studied by means of S-100 protein and GFAP immunocytochemistry between the 10th and 17th week of gestation. In addition, double labelling immunocytochemistry using an antibody raised against a constitutive isoform of nitric oxide synthase (bNOS) in combination with an S-100 protein antibody was applied to investigate the morphological relations between NN cells and nitrergic neurons in the developing gut wall. Cells with immunoreactivity for both glial-specific proteins are already present in the 10th week of gestation. While cells with S-100 protein immunoreactivity are located within the circular muscle layer as well as in the myenteric, and submucous plexuses, cells with GFAP immunopositivity are mainly restricted to the side of the myenteric plexus adjacent to the longitudinal muscle layer. In contrast to the dense network formed by S-100 protein immunopositive structures the GFAP immunopositive cells appear singly and do not connect into a network. Double-labelling immunocytochemistry reveals nitrergic fibers (NOS-IR) in close relation to the S-100 protein immunoreactive glial network. NOS-IR varicosities are in close association with the surface of those cells both in the circular muscle layer (CM) and in the tertiary plexus. It is concluded that two populations of NN cells with different locations and different immunohistochemical characters appear and develop together with the enteric ganglia in the human fetal intestine. The close morphological relation of NOS-IR fibers with S-100 protein immunopositive cellular network indicate a functional relationship between S-100 protein immunopositive cells and the nitrergic nerves during the early development of human enteric nervous system (ENS).     

Developmental pattern of three vesicular glutamate transporters in the myenteric plexus of the human fetal small intestine

Three vesicular glutamate transporters (VGLUT1-3) have previously been identified in the central nervous system, where they define the glutamatergic phenotype, and their expression is tightly regulated during brain development. In the present study we applied immunocytochemistry to examine the distribution of the immunoreactivity of all three VGLUTs during prenatal development of the myenteric plexus in the human small intestine. We also investigated changes in their localization in the different segments of the small intestine and in the different compartments of the developing myenteric ganglia. Immunoreactivity against all three VGLUTs was found predominantly in the ganglionic neuropil, interganglionic varicose fibers and perisomatic puncta, but cytoplasmic labeling with different intensities also occurred. Each transporter displayed a characteristic spatiotemporal expression pattern, with the transient increase or decrease of immunoreactive cell bodies, varicosities or perisomatic puncta, depending on the fetal age, the gut segment or the ganglionic compartment. Throughout gestational weeks 14-23, VGLUT1 immunoreactivity always predominated over VGLUT2 immunoreactivity, though both peaked around week 20. VGLUT3 immunoreactivity was less abundant in the developing myenteric plexus than those of VGLUT1 and VGLUT2 immunoreactivity. It was mainly expressed in the ganglionic neuropil and in the perisomatic puncta throughout the examined gestational period. Neuronal perikarya immunoreactive for VGLUT3 were restricted to between weeks 18 and 20 of gestation and exclusively to the oral part of the small intestine.     

Total numbers of neurons in myenteric ganglia of the guinea-pig small intestine

Two techniques that are thought to stain all of the neurons in the myenteric ganglia of the intestine are NADH diaphorase histochemistry and immunhistochemistry using a nerve cell body antiserum. However, this assumption has never been directly verified. In the present study myenteric ganglia of the guinea-pig ileum were prepared as whole-mounts and stained with either of these techniques. All nerve cells that could be identified in the whole-mounts were counted. The whole-mounts were then embedded flat in resin and serially sectioned at 1 m. Nerve cells were identified and counted from the serial sections, and the data compared to those obtained from the whole-mounts. NADH diaphorase histochemistry did not reveal all the neurons at incubation times that gave selective staining. In contrast, nerve cell body antiserum stained the entire neuronal population. To determine the total number of nerve cell bodies/ganglion and the proportion of nerve cell bodies with calbindin immunoreactivity, whole-mounts that had been processed for calbindin immunohistochemistry were serially sectioned and reconstructed. The total number of neurons per myenteric ganglion was 105±10 (SE). Calbindin-immunoreactive neurons comprised about 20% of the myenteric neurons, which is considerably less than previous estimates, because previously the total population has been underestimated. The spatial density of myenteric neurons in the undistended ileum of the guinea-pig is 17300 nerve cells/cm².     

Nonsulfated cholecystokinins in the small intestine of pigs and rats

Cholecystokinin (CCK) is a gut hormone that acts via two receptors. The CCK_A-receptor requires the tyrosyl residue in the C-terminal bioactive site of CCK to be O-sulfated, whereas, the CCK_B-receptor binds irrespective of sulfation. Consequently, unsulfated CCK peptides – if present – may constitute a hormone system that acts only through the CCK_B-receptor. Therefore, we have now examined whether, CCK peptides occur in nonsulfated form in the small intestine of pigs and rats. The concentrations of sulfated and nonsulfated CCK were measured by RIAs, one specific for sulfated CCKs and a new two-step assay specific for nonsulfated CCK. For further characterization, the intestinal extracts were subjected to size- and ion exchange-chromatography.The intestinal concentrations of sulfated and nonsulfated CCK were highest in the duodenum and the proximal part of jejunum both in the pig and the rat. The porcine duodenal mucosa contained 193 ± 84 pmol/g sulfated CCK and 31 ± 10 pmol/g nonsulfated CCK, and the upper rat intestine 70 ± 19 pmol/g and 8 ± 2 pmol/g, respectively. The degree of sulfation correlated with the endoproteolytic proCCK processing. Thus, 38% of porcine CCK-58 was unsulfated, whereas, only 12% of CCK-8 was unsulfated.The results show that a substantial part of intestinal CCK peptides in rats and pigs are not sulfated, and that the longer peptides (CCK-58 and CCK-33) are less sulfated than the shorter (CCK-22 and CCK-8). Hence, the results demonstrate that proCCK in the gut is processed both to sulfated and unsulfated α-amidated peptides of which the latter are assumed to act via the CCK_B-receptor.     

High- and medium-molecular-weight neurofilament proteins define specific neuron types in the guinea-pig enteric nervous system

Previous studies have demonstrated that neurofilament proteins are expressed by type II neurons in the enteric plexuses of a range of species from mouse to human. However, two previous studies have failed to reveal this association in the guinea-pig. Furthermore, immunohistochemistry for neurofilaments has revealed neurons with a single axon and spiny dendrites in human and pig but this morphology has not been described in the guinea-pig or other species. We have used antibodies against high- and medium-weight neurofilament proteins (NF-H and NF-M) to re-examine enteric neurons in the guinea-pig. NF-H immunoreactivity occurred in all type II neurons (identified by their IB4 binding) but these neurons were never NF-M-immunoreactive. On the other hand, 17% of myenteric neurons expressed NF-M. Many of these were uni-axonal neurons with spiny dendrites and nitric oxide synthase (NOS) immunoreactivity. NOS immunoreactivity occurred in surface expansions of the cytoplasm that did not contain neurofilament immunoreactivity. Thus, because of their NOS immunoreactivity, spiny neurons had the appearance of type I neurons. This indicates that the apparent morphologies and the morphological classifications of these neurons are dependent on the methods used to reveal them. We conclude that spiny type I NOS-immunoreactive neurons have similar morphologies in human and guinea-pig and that many of these are inhibitory motor neurons. Both type II and neuropeptide-Y-immunoreactive neurons in the submucosal ganglia exhibit NF-H immunoreactivity. NF-M has been observed in nerve fibres, but not in nerve cell bodies, in the submucosa. This work was supported by a grant from the National Health and Medical Council of Australia (grant number 400020).     

Neuropeptides in the gut: quantification and characterization of cholecystokinin octapeptide-, bombesin- and vasoactive intestinal polypeptide-like immunoreactivities in the myenteric plexus of the guinea-pig small institute

The localization of CCK8-, bombesin- and VIP-like immunoreactivities in the myenteric plexus of the guinea pig small intestine has been studied by radioimmunoassay of extracts of longitudinal muscle strips obtained with and without adherent myenteric plexus; concentrations were compared with those in other regions of the gut. In innervated strips of longitudinal muscle of ileum there was approximately 14 pmol/g CCK8-, 32 pmol/g bombesin- and 135 pmol/g VIP-like immunoreactivity; concentrations were reduced by over 70% in denervated strips. Gel filtration and ion exchange chromatography indicated that over 80% of CCK immunoreactivity was due to CCK8; no evidence was found of significant amounts of smaller COOH-terminal fragments. Bombesin immunoreactivity occurred in two forms, the major one resembling the amphibian tetradecapeptide in its elution from gel filtration columns. Immunoreactive VIP differed markedly from porcine VIP in immunochemical and chromatographic properties; the data suggest that guinea pig VIP is less basic than porcine VIP and that the two peptides differ in structure in their NH2-terminal regions. Some functional implications of these findings are discussed.     

Salmonella typhimurium and Salmonella enteritidis induce gut growth and increase the polyamine content of the rat small intestine in vivo

Abstract The effects of infection by Salmonella enteritidis and S. typhimurium on the small and large intestines, liver, spleen and mesenteric nodules of rats were studied in vivo. Both Salmonella serotypes persisted and proliferated in the gastrointestinal tract and invaded sub-epithelial tissues, mainly the ileum, leading to the systemic distribution of these pathogens. Coincidental with the infection, the rate of crypt cell proliferation increased resulting in substantial growth of the small intestine. The extent of this and the accompanying accumulation of polyamines was particularly dramatic in the ileum where there was also some disruption of the villus epithelium. It is possible that these effects of the infection on the metabolism and morphology of the small bowel, which strongly resembled the changes induced by some plant lectins, may facilitate the colonisation and invasion of the gut by Salmonellae .     

微生态学与中医“心与小肠相表里”新论

本文阐述了中医“心与小肠相表里”的理论内涵。中医“心与小肠相表里”与现代医学提出的“脑—肠轴”理论是相通的,具体可理解为中枢神经系统与肠道菌群的密切关系。脑与肠道菌群是相关的,二者在生理上相互配合、在病理上互相影响。同时探讨了微生态学与中医“心与小肠相表里”的相关性,阐述了脑对肠道菌群的调控和肠道菌群对中枢神经系统疾病的发生、发展的影响。依据中医“心与小肠相表里”理论和微生态学的临床应用,提出了用改善肠道菌群的微生态方法治疗和预防一些神经系统疾病的新理念。