The morphology and projections of retrogradely labeled myenteric neurons in the human intestine

BACKGROUND & AIMS

Myenteric ganglia in the human gastrointestinal tract contain a mixture of many different types of nerve cells that cannot be distinguished by their location. The aim of this study was to characterize different functional types of cells by using retrograde labeling in vitro to identify neurons according to their targets.

METHODS

The retrograde label 1,1'-didodecyl 3,3,3',3'-indocarbocyanine perchlorate (Dil) was applied to different target layers of human small or large intestine. After 3-5 days in organotypic culture, myenteric neurons projecting to the Dil application site were visualized and mapped using fluorescence microscopy.

RESULTS

Myenteric motor neurons projecting to the external muscle layer were typically unipolar cells with lamellar dendrites (Dogiel type I) and had short projections up to 16 mm long. In contrast, presumed interneurons with Dogiel type I morphology were shown to project up to 68 mm aborally or up to 38 mm orally. Multipolar Dogiel type II neurons with smooth cell bodies were labeled most frequently from the submucous plexus. No myenteric neurons were labeled by Dil applied to the mucosa.

CONCLUSIONS

Myenteric neurons labeled from each target had characteristic size, morphology, polarity, and length of projections, indicating that there is a high degree of organization in the human enteric nervous system.

Ultrastructural studies of the myenteric plexus and smooth muscle in organotypic cultures of the guinea-pig small intestine

External muscle and myenteric plexus from the small intestine of adult guinea-pigs were maintained in vitro for 3 or 6 days. Myenteric neurons and smooth muscle cells from such organotypic cultures were examined at the electron-microscopic level. An intact basal lamina was found around the myenteric ganglia and internodal strands. Neuronal membranes, nuclei and subcellular organelles appeared to be well preserved in cultured tissues and ribosomes were abundant. Dogiel type-II neurons were distinguishable by their elongated electron-dense mitochondria, numerous lysosomes and high densities of ribosomes. Vesiculated nerve profiles contained combinations of differently shaped vesicles. Synaptic membrane specializations were found between vesiculated nerve profiles and nerve processes and cell bodies. The majority of nerve fibres were well preserved in the myenteric ganglia, in internodal strands and in bundles running between circular muscle cells. No detectable changes were found in the ultrastructure of the somata and processes of glial cells. Longitudinal and circular muscle cells from cultured tissue had clearly defined membranes with some close associations with neighbouring muscle cells. Caveolae occurred in rows that ran parallel to the long axis of the muscle cells. These results indicate that the ultrastructural features of enteric neurons and smooth muscle of the guinea-pig small intestine are well preserved in organotypic culture.

Long aboral projections of Dogiel type II, AH neurons within the myenteric plexus of the guinea pig small intestine

Enteric AH neurons, with multipolar Dogiel type II morphology, project around the circumference of the intestine to myenteric ganglia, the submucosa and mucosa. Using retrograde labeling in vitro, intracellular recording, dye filling and immunohistochemistry, the projections of these neurons along the intestine were studied. When the retrograde tracer, Dil, was applied to the myenteric plexus, labeled nerve cell bodies were located up to 111 mm orally but only 13 mm aborally, demonstrating a marked difference in the lengths of projections up and down the small intestine. Of labeled nerve cell bodies located 2-110 mm orally, 43% had Dogiel type II morphology and of these, 70% were immunoreactive for calbindin, a calcium binding protein exclusive to Dogiel type II neurons. Intracellular filling with neurobiotin revealed several long circumferentially directed nerve fibers and short, filamentous dendrites; thus these were "dendritic" Dogiel type II neurons. This class accounts for approximately 3-4% of all myenteric neurons, and about 10% of all Dogiel type II neurons. Intracellular recordings revealed AH cell characteristics, with long afterhyperpolarizations following their action potentials, pronounced slow excitatory synaptic inputs and a lack of fast excitatory synaptic inputs. Antidromic action potentials could be evoked from the Dil application site in some cells, confirming their aboral projection. This is the first account of a major aboral projection of AH/Dogiel type II neurons and suggests an important role in aborally directed reflexes in the intestine.

The chemistry of enamel development

The central problems of enamel biochemistry are the mechanisms concerned with initiation and development of the mineral crystals, together with their architectural arrangement within the tissue. These processes are mediated by the extracellular matrix as well as the composition of the mineral itself. Initial mineral deposition occurs at the dentine surface, nucleated either by dentinal components or early enamel matrix, possibly non-amelogenin molecules. The early crystals are small in size and rich in magnesium and carbonate resulting in relatively poor crystallinity. This is in spite of the fact that fluoride is high at this stage. Crystal development includes a reduction in magnesium, carbonate and fluoride as crystals increase in length following the retreating ameloblasts from the dentine. The matrix acquires increasing concentrations of amelogenin and albumin. Prismatic structure begins to develop together with some growth of crystals in width and thickness. Degradation of amelogenin and non-amelogenin molecules generates a series of specific molecular fragments possibly concerned with modulating crystal growth and morphology and the creation of prismatic and interprismatic structures. Towards the end of secretion, matrix, now almost completely degraded, is replaced by fluid followed by massive crystal growth during maturation. Degradation of albumin also occurs at this stage, probably as a result of comprehensive destruction of molecules which might impair crystal growth. Selective acquisition of magnesium and fluoride at this stage may reflect the hydrated state of the tissue as well as cell changes. Fluid is displaced as crystals grow and the enamel acquires concentrations of mineral characteristic of mature tissue.

The control of ingress of albumin into developing enamel from adjacent dentine of the rat incisor

The present study used an immunohistochemical approach to map the distribution of albumin within the dentine adjacent to the developing enamel of both impeded and unimpeded rat incisors to determine if the dentine could be a possible route of entry for this protein into the developing enamel matrix. In dentine adjacent to the secretory and transition stages of the developing enamel, the dentinal tubules were labelled only over approximately the pulpal quarter of their length. The bulk of the dentine showed no labelling. However, labelling within the dentine appeared at the ADJ at a position approximately 2.5 mm occlusal to the distal root of the first molar, adjacent to enamel with no visible residual matrix. The results of this study suggest that adventitious ingress of albumin into enamel from the dentine is restricted during enamel secretion and can only potentially occur once enamel maturation has been initiated.

The effect of glycosylaminoglycans on the mineralization of sheep periodontal ligament in vitro

The effect of removal of glycosylaminoglycans on the mineralization of sheep periodontal ligament was determined using enzyme digests followed by incubation in solutions supersaturated with respect to hydroxyapatite at pH 7.4. TEM revealed that control periodontal ligament remained unmineralized. However, tissue from which glycosylaminoglycans had been removed contained plate-like crystals arranged parallel to and within the collagen fibrils. Electron probe and electron diffraction studies suggested that the crystals were apatitic with a similar order of crystallinity to dentine, and a Ca:P ratio of 1.61. In addition, the glycosylaminoglycan content of periodontal ligament, cementum and alveolar bone was compared using cellulose acetate electrophoresis. Periodontal ligament contained predominantly dermatan sulfate while cementum and alveolar bone contained mostly chondroitin sulfate. A role for glycosylaminoglycans in maintaining the unmineralized state of the periodontal ligament is suggested. Control of expression of specific proteoglycan species on a spatially restricted basis is presumably central to this role.

Uptake and metabolism of albumin by rodent incisor enamel in vivo and postmortem: implications for control of mineralization by albumin

The distribution of albumin throughout enamel development in the rat mandibular incisor was investigated using sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS PAGE) and Western blotting employing an anti-rat albumin antibody. Intact albumin was detectable at all stages of enamel development but was most evident during late secretion/transition. Its concentration was subsequently reduced during the maturation stage. Albumin degradation products appeared during the transition/early maturation stage indicating that albumin breakdown preceded its removal. As albumin inhibits apatite crystal growth, its degradation and removal may be a necessary prerequisite for normal enamel crystal growth, perhaps reflecting a general mechanism for removal of residual endogenous matrix or adventitious crystal growth inhibitors. Additional studies revealed that the maturation stage was particularly susceptible to albumin influx postmortem. Albumin could therefore form part of the natural crystal growth control process, which, if not removed, could hamper maturation and lead to white spot hypoplasias.

Characterization of alkaline phosphatase-reactive neurons in the guinea-pig small intestine

Endogenous alkaline phosphatase activity has been localized histochemically on the surface of enteric neurons of the guinea-pig small intestine by both light and electron microscopy. The enzyme activity was associated with some myenteric neurons that had Dogiel type I morphology, and the histochemical reaction products typically formed a honeycomb-like structure on labelled cell bodies. No Dogiel type II neurons in the myenteric plexus or submucous neurons showed alkaline phosphatase reactivity. Nerve fibres reactive for alkaline phosphatase were present in the myenteric plexus and ran in bundles in the circular muscle and deep muscular plexus. In addition, reactive varicose axons supplied the submucous plexus and non-ganglionated plexus of the mucosa. The results of interruption of the enteric neuronal pathways demonstrated that alkaline phosphatase-reactive myenteric neurons project anally to other myenteric ganglia, to the circular muscle and to the submucous plexus. Sequential enzyme histochemistry showed that virtually all alkaline phosphatase-reactive neurons also contained nitric oxide synthase, revealed by NADPH-diaphorase reactivity. It was estimated that 14-18% of all myenteric neurons showed alkaline phosphatase reactivity. About one-third of nitric oxide synthase-containing myenteric neurons, however, did not contain alkaline phosphatase activity. At the ultrastructural level, alkaline phosphatase activity was associated specifically with the plasma membranes of nerve cell bodies, axons and dendrites of some myenteric neurons. Reactive nerve fibres made close appositions with non-reactive submucous neurons and, within myenteric ganglia, predominantly with other alkaline phosphatase-reactive neurons. In addition to its presence in neurons, alkaline phosphatase reactivity was also present in some endothelial cells in blood vessels in the submucosa and in capillary pericytes. It is concluded, on the basis of the projections and neurochemistry, that in the guinea-pig small intestine alkaline phosphatase activity is associated with nitric oxide synthase-containing neurons which include inhibitory motor neurons to the circular muscle, and anally-directed interneurons to other myenteric and submucous neurons.

Isolation and characterisation of an alternatively-spliced rat amelogenin cDNA: LRAP--a highly conserved, functional alternatively-spliced amelogenin?

A cDNA coding for a 59 amino acid polypeptide containing both the carboxy- and amino-termini, but lacking the central domain, of the rat tooth enamel matrix protein, amelogenin, was cloned and sequenced. The deduced polypeptide sequence indicates that this cDNA was derived from an amelogenin RNA molecule by using an alternative intra-exonic 3' splice acceptor site. This alternatively spliced product is almost identical to products previously identified in both cow and mouse enamel organs: the leucine-rich amelogenin peptide (LRAP). The conservation of this truncated polypeptide across the species suggests that it may have an important role in the formation of tooth enamel.

All calbindin-immunoreactive myenteric neurons project to the mucosa of the guinea-pig small intestine

The projections of Dogiel type II myenteric neurons to the mucosa of the guinea-pig ileum were quantified by combining retrograde transport of DiI, in vitro, with immunohistochemistry. After DiI application to the mucosa over an area of 1.5 x 10 mm2, virtually all (> 97%) calbindin-immunoreactive Dogiel type II neurons in the myenteric plexus underneath the mucosal DiI application site were labelled, indicating that essentially all of these neurons project to the mucosa. From cell counts, on average 5 calbindin-immunoreactive neurons project to each villus, and each calbindin-immunoreactive neuron supplies on average 10 villi. Since Dogiel type II neurons that were not immunoreactive for calbindin (19% of all labelled nerve cells) also projected to the mucosa, it is likely that all Dogiel type II neurons, which are putative sensory neurons of the gut, project to the mucosa.

NADPH-diaphorase and other neuronal markers in nerves and ganglia supplying the guinea-pig vas deferens

Enzyme histochemistry, in combination with immunohistochemistry was used to establish the neurochemistry of neurons in the vas deferens and pelvic ganglia of the guinea-pig. Nerve fibres characterised by reactivity for reduced nicotinamide adenine dinucleotide phosphate (NADPH)-diaphorase reactivity formed a dense network in the lamina propria and circular muscle layer of the vas deferens, but were very sparse in the longitudinal muscle layer of the vas deferens. NADPH-diaphorase reactivity was also present in nerve fibres forming a dense perivascular plexus in many of the arteries in the pelvic region and in some of the endothelial cells, especially near the origin of the capillaries. Nerves with vasoactive intestinal polypeptide (VIP)-immunoreactivity had a similar distribution to NADPH-diaphorase reactive nerves. Tyrosine hydroxylase (TH)-immunoreactive nerve fibres were found in both muscle layers of the vas deferens. There was no coexistence of VIP- and TH-immunoreactivities in nerve fibres in the vas deferens. In the anterior pelvic ganglia, the origin of the nerve fibres in the vas deferens, several classes of neurons could be identified by the presence or absence of the reactivity for NADPH-diaphorase and immunoreactivity for VIP and TH. Neurons containing both VIP and NADPH-diaphorase reactivity accounted for 40% of neurons in the ganglia. Neurons with VIP-immunoreactivity but not NADPH-diaphorase reactivity accounted for 6%. TH-immunoreactive neurons accounted for 22% of neurons in the anterior pelvic ganglia. Very rare cells (< 1%) contained both VIP- and TH-immunoreactivities. The remaining neurons, which were not labelled by any of these markers, comprised 31% of neurons in anterior pelvic ganglia. These results demonstrate the existence of NADPH-diaphorase reactivity in neurons containing VIP-immunoreactivity, thus suggest that nitric oxide may be a neurotransmitter in guinea-pig vas deferens, especially in the circular muscle layer, in the arteries, and in other pelvic organs innervated by pelvic ganglia.

Nitric oxide mediates nonadrenergic, noncholinergic neural relaxation in the Australian possum

BACKGROUND

Nitric oxide has been shown to play an important role in neurally mediated relaxations of gastrointestinal smooth muscle. The aim of this study was to determine whether NO may be the inhibitory transmitter to circular smooth muscle from the sphincter of Oddi of the Australian brush-tailed possum (Trichosurus vulpecula).

METHODS

The effects of drugs on relaxations evoked by electrical-field stimulation of circular muscle strips precontracted with either erythromycin or carbachol were studied. Preparations were also processed histochemically to determine the presence of nicotinamide adenine dinucleotide phosphate (NADPH)-diaphorase reactivity in the tissue.

RESULTS

NG-nitro-L-arginine methyl ester reduced the amplitude of relaxations; this effect could be partially reversed by millimolar concentrations of L-arginine but not by D-arginine. Oxyhemoglobin also reduced the amplitude of the relaxations, and sodium nitroprusside mimicked the relaxations in precontracted strips. Histochemical processing revealed the presence of nerve cell bodies and nerve fibers associated with the circular muscle layer, which are reactive for NADPH-diaphorase and are thus likely to contain NO synthase.

CONCLUSIONS

These results are all consistent with NO released from nerve cells mediating a significant part of the nonadrenergic, noncholinergic relaxation of the circular muscle layer of the sphincter of Oddi.

Neuronal nitric oxide in the gut

Motility of the gastrointestinal tract is directly controlled by enteric inhibitory and excitatory motor neurons that innervate the layers of smooth muscle. Inhibitory motor neurons mediate receptive and accommodative relaxations and control the opening of sphincters, thus playing an important role in normal gut motility. Recent studies have demonstrated that nitric oxide (NO) is an important neurotransmitter released by inhibitory motor neurons in animal and human gut. Antagonists of nitric oxide synthase (NOS), the synthetic enzyme for NO, reduce the effectiveness of transmission from inhibitory motor neurons. Exogenous NO mimics inhibitory nerve activation, and a variety of compounds that affect the availability of endogenously produced NO modulate relaxations of gastrointestinal smooth muscle. It is clear, however, that NO is unlikely to be the only transmitter released by enteric inhibitory motor neurons: several other substances such as vasoactive intestinal polypeptide (VIP), or related peptides, and adenosine triphosphate (ATP) are also likely to contribute to nerve-mediated inhibition. The identification of NO as a major inhibitory neurotransmitter to gastrointestinal smooth muscle fills an important gap in our understanding of the physiological control of motility and opens up a wide range of new experimental possibilities. It may eventually lead to the development of new drugs for motility disorders. It should be noted, however, that NO is important in the brain, in cardiovascular control, in blood cell function and in many other organ systems, suggesting that it may be difficult to achieve specific pharmacological intervention targeted on inhibitory neurotransmission in the gut, without undesirable side effects.

NADPH-diaphorase reactivity in nerves supplying the rat anococcygeus muscle

NADPH-diaphorase histochemistry was used to establish the nature of neurons and nerve fibers in the rat anococcygeus muscle. NADPH-diaphorase-reactive nerve cell bodies were found in ganglia in the caudal dorsal fascia of the muscle and within the muscle itself. The reactive nerve cells had a relatively smooth surface and a single axon. They formed bundles of positive nerve fibers which entered the muscle and branched further to form a network extending towards the cranial end. In addition, bundles of NADPH-diaphorase-reactive fibers of extrinsic origin joined the loose plexus of local bundles and microganglia. The NADPH-diaphorase-reactive neurons supplying the rat anococcygeus muscle provide a neuroanatomical basis for the non-adrenergic non-cholinergic relaxation of the muscle mediated by nitric oxide.

Projections and chemical coding of neurons with immunoreactivity for nitric oxide synthase in the guinea-pig small intestine

The distribution of nitric oxide synthase (NOS) immunoreactivity was investigated in the guinea-pig small intestine. There were many immunoreactive nerve cell bodies in the myenteric plexus but very few in submucous ganglia. NOS immunoreactivity was not found in non-neuronal cells except for rare mucosal endocrine cells. Abundant immunoreactive nerve fibres in both myenteric and submucous ganglia, and in the circular muscle, arose from myenteric nerve cells whose axons projected anally along the intestine. NOS immunoreactivity coexisted with VIP-immunoreactivity, but not with substance P immunoreactivity. We conclude that nitric oxide synthase is located in a sub-population of enteric neurons, amongst which are inhibitory motor neurons that supply the circular muscle layer.

Identification of motor neurons to the longitudinal muscle of the guinea pig ileum

Motor neurons that innervate the longitudinal muscle of the guinea pig ileum were identified by retrograde transport from the longitudinal muscle plexus in organotypic culture. Motor neurons had short projections, less than 3.5 mm long, and never had Dogiel type II morphology; most labeled neurons had morphological characteristics of Dogiel type I neurons. Immunoreactivity for choline acetyltransferase was present in 97% of retrogradely labeled nerve cell bodies, reflecting the dominant cholinergic input to the longitudinal muscle layer. Substance P immunoreactivity was present in 48% of motor neurons, indicating that it or a similar tachykinin that mediates noncholinergic excitatory transmission is likely to be released by a subset of cholinergic motor neurons. This strongly suggests that the difference in frequency dependence of substance P and acetylcholine release is attributable to different release mechanisms rather than to activation of separate populations of motor neurons. Immunoreactivity for the calcium-binding protein calretinin was present in 87% of longitudinal muscle motor neurons. The neurochemical coding of longitudinal muscle motor neurons indicated that they constitute about one quarter of all myenteric neurons and are distinct from circular muscle motor neurons.

Projections and pathways of submucous neurons to the mucosa of the guinea-pig small intestine

Double-labelling immunohistochemistry and retrograde transport of the carbocyanine dye, DiI, were used to establish the pathways of submucous neurons to the mucosa of the guinea-pig small intestine. Following the application of DiI to a villus, DiI-labelled nerve cell bodies were found in the submucous plexus up to 8.3 mm circumferentially and 3.8 mm longitudinally. The size of each of the four characterised classes of submucous neurons was determined and their distributions and projections mapped. Cells characterised by vasoactive intestinal polypeptide immunoreactivity accounted for 52% of DiI-labelled cells and had the longest projections. Cells characterised by neuropeptide Y (19%) or by calretinin immunoreactivity (13% of all DiI-labelled neurons) had relatively short projections and cells with substance P immunoreactivity (20%) had intermediate lengths of projection. When DiI was applied directly to the submucous plexus, filled neurons of all classes had significantly shorter projections, indicating that they must run for considerable distances in other pathways to the mucosa, probably via the non-ganglionated plexus. On average, each villus is innervated by at least 70 submucous neurons. From quantitative estimates there are 9 submucous neurons per villus. Thus, each submucous neuron is likely to supply about 8 villi. This demonstrates a high degree of convergence and divergence in the innervation of the mucosa.

The role of albumin in developing rodent dental enamel: a possible explanation for white spot hypoplasia

The uptake of serum albumin by maturation-stage rodent enamel and the resulting effects on the growth of enamel crystallites were investigated in vitro. Albumin uptake was demonstrated by means of gel electrophoresis and confirmed by Western blotting with use of monoclonal antibodies. Measurement of crystal size was carried out by direct TEM measurement of enamel crystallite outlines after incubations in metastable solutions of calcium phosphate. The ability of endogenous enamel enzymes to degrade albumin was investigated by substrate-specific zymography. The results showed that albumin could be taken up by maturation-stage enamel and produce inhibition of crystallite growth. There was no detectable proteolytic activity in the enamel against albumin substrate, which suggests that albumin entering enamel by extravasation in vivo may produce incomplete tissue maturation, resulting in a white, opaque appearance on eruption.

Identification of myenteric neurons which project to the mucosa of the guinea-pig small intestine

Myenteric neurons which innervate the mucosa of the guinea-pig ileum were characterized by combining retrograde transport of DiI in vitro with immunohistochemistry. Of DiI-labelled myenteric neurons, 43% were immunoreactive for calbindin and substance P, 25% were immunoreactive for calbindin alone, and 18% were immunoreactive for substance P alone. These 3 classes of neurons had Dogiel Type II morphology and are probably sensory neurons. Two classes of probable secretomotor neurons were characterized by immunoreactivity for neuropeptide Y (4%) and vasoactive intestinal peptide (2%). These 5 classes of myenteric neurons represent over 90% of the retrogradely labelled myenteric neurons that project to the mucosa.

Calretinin immunoreactivity in cholinergic motor neurones, interneurones and vasomotor neurones in the guinea-pig small intestine

Immunoreactivity for calretinin, a calcium-binding protein, was studied in neurones in the guinea-pig small intestine. 26 +/- 1% of myenteric neurones and 12 +/- 3% of submucous neurones were immunoreactive for calretinin. All calretinin-immunoreactive neurones were also immunoreactive for choline acetyltransferase and hence are likely to be cholinergic. In the myenteric plexus, two subtypes of Dogiel type-I calretinin-immunoreactive neurones could be distinguished from their projections and neurochemical coding. Some calretinin-immunoreactive myenteric neurones had short projections to the tertiary plexus, and hence are likely to be cholinergic motor neurones to the longitudinal muscle. Some of these cells were also immunoreactive for substance P. The remaining myenteric neurones, immunoreactive for calretinin, enkephalin, neurofilament protein triplet and substance P, are likely to be orad-projecting, cholinergic interneurones. Calretinin immunoreactivity was also found in cholinergic neurones in the submucosa, which project to the submucosal vasculature and mucosal glands, and which are likely to mediate vasodilation. Thus, calretinin immunoreactivity in the guinea-pig small intestine is confined to three functional classes of cholinergic neurones. It is possible, for the first time, to distinguish these classes of cells from other enteric neurones.

Immunohistochemical identification of cholinergic neurons in the myenteric plexus of guinea-pig small intestine

It is well established that acetylcholine is a neurotransmitter at several distinct sites in the mammalian enteric nervous system. However, identification of the cholinergic neurons has not been possible due to an inability to selectively label enteric cholinergic neurons. In the present study an immunohistochemical method has been developed to localize choline acetyltransferase, the synthetic enzyme for acetylcholine, in order that cholinergic neurons can be visualized. The morphology, neurochemical coding and projections of cholinergic neurons in the guinea-pig small intestine were determined using double-labelling immunohistochemistry. These experiments have revealed that many myenteric neurons are cholinergic and that they can be distinguished by their specific combinations of immunoreactivity for neurochemicals such as calretinin, neurofilament protein triplet, substance P, enkephalin, somatostatin, 5-hydroxytryptamine, vasoactive intestinal peptide and calbindin. On the basis of their previously described projections, functional roles could be attributed to each of these populations. The identified cholinergic neurons are: motorneurons to the longitudinal muscle (choline acetyltransferase/calretinin); motorneurons to the circular muscle (choline acetyltransferase/neurofilament triplet protein/substance P, choline acetyltransferase/substance P and choline acetyltransferase alone); orally directed interneurons in the myenteric plexus (choline acetyltransferase/calretinin/enkephalin); anally directed interneurons in the myenteric plexus (choline acetyltransferase/somatostatin, choline acetyltransferase/5-hydroxytryptamine, choline acetyltransferase/vasoactive intestinal peptide); secretomotor neurons to the mucosa (choline acetyltransferase/somatostatin); and sensory neurons mediating myenteric reflexes (choline acetyltransferase/calbindin). This information provides a unique opportunity to identify functionally distinct populations of cholinergic neurons and will be of value in the interpretation of physiological and pharmacological studies of enteric neuronal circuitry.

Identification and immunohistochemistry of cholinergic and non-cholinergic circular muscle motor neurons in the guinea-pig small intestine

Motor neurons which innervate the circular muscle layer of the guinea-pig small intestine were retrogradely labelled, in vitro, with the carbocyanine dye, DiI, applied to the deep muscular plexus. By combining retrograde tracing and immunohistochemistry, the chemical coding of motor neurons was investigated. Five classes of neuron could be distinguished on the basis of the co-localization of immunoreactivity for the different antigens; the five classes were also characterized by different lengths and polarities of their axonal projections and by their cell body shapes. Two classes with local or orally directed axons were immunoreactive for choline acetyltransferase and substance P and are likely to be cholinergic excitatory motor neurons. Two other classes had anally directed axons; they were immunoreactive for vasoactive intestinal polypeptide and are likely to be inhibitory motor neurons. A small proportion of neurons with short projections to the circular muscle were immunoreactive for neither substance P nor for vasoactive intestinal polypeptide, but are likely to be cholinergic. The morphological and histochemical identification of excitatory and inhibitory motor neurons provides a neuroanatomical basis for the final motor pathways involved in the polarized reflex motor activity of the gut.