THE ANATOMICAL RECORD PART A 272A:475– 483 (2003)
Morphologies and Projections of Defined Classes of Neurons in the Submucosa of the Guinea-Pig Small Intestine JOHN B. FURNESS,1,2* GEORGE ALEX,1 MELANIE J. CLARK,1 AND VARSHA V. LAL1 1 Department of Anatomy and Cell Biology, University of Melbourne, Parkville, Australia 2 Centre for Neuroscience, University of Melbourne, Parkville, Australia
ABSTRACT Four types of neurons have previously been identified by neurochemical markers in the submucosal ganglia of the guinea-pig small intestine, and functional roles have been ascribed to each type. However, morphological differences among the classes have not been determined, and there is only partial information about their projections within the submucosa. In the present work, we used intracellular microelectrodes to fill neurons of each type with biocytin, which was then converted to a permanent dye, so that the shapes of the neurons could be determined and their projections within the submucosa could be followed. Cell bodies of noncholinergic secretomotor/ vasodilator neurons had Dogiel type I morphology. These neurons, which are vasoactive intestinal peptide immunoreactive, had single axons that ran through many ganglia without providing terminals around other neurons. Cholinergic secretomotor neurons with neuropeptide Y immunoreactivity had Stach type IV morphology, and cholinergic secretomotor/vasodilator neurons had stellate cell bodies. The axons of these two types ran short distances in the plexus and did not innervate other submucosal neurons. Neurons of the fourth type, intrinsic primary afferent neurons, had cell bodies with Dogiel type II morphology and their processes supplied networks of varicose processes around other nerve cells. It is concluded that each functionally defined type of submucosal neuron has a characteristic morphology and that intrinsic primary afferent neurons synapse with secretomotor neurons to form monosynaptic secretomotor reflex circuits. Anat Rec Part A 272A: 475– 483, 2003. © 2003 Wiley-Liss, Inc.
Key words: enteric nervous system; neuron morphology; secretomotor neurons; sensory neurons
Analysis of the organisation of enteric nerve circuits has been greatly assisted by studies that have identified different classes of neurons by their morphologies, neurochemistries, and connections with other neurons. These studies used a range of methods, singly or in combination, including immunohistochemistry, electron microscopy, intracellular dye filling, and retrograde labelling from restricted targets (Costa et al., 1996; Furness, 2000; Furness et al., 2000; Brookes, 2001). In the myenteric plexus of the guinea-pig ileum, these investigations have led to a full accounting of all neurons by shape, projection, and chemistry, and each morphologically and chemically defined phenotype can be related to a functional class of neuron. Three immunohistochemically-defined types and one morphologically-defined type account for almost 100% of submucosal neurons (Furness et al., 1984; Song et al., ©
2003 WILEY-LISS, INC.
1992; Evans et al., 1994; Quinson et al., 2001). These four types include 1) neurons with immunoreactivity for vasoactive intestinal peptide (VIP) and several other peptides, including galanin, dynorphin, and neuromedin U; 2) neurons immunoreactive for neuropeptide Y (NPY), choline
*Correspondence to: Professor John B. Furness, Department of Anatomy and Cell Biology, University of Melbourne, Parkville, VIC 3010, Australia. Fax: ⫹61-3-93475219. E-mail:
[email protected] Received 25 October 2002; Accepted 16 December 2002 DOI 10.1002/ar.a.10064
476
FURNESS ET AL.
acetyltransferase (ChAT), somatostatin, and calcitonin gene-related peptide; 3) neurons immunoreactive for calretinin, ChAT, and dynorphin; and 4) neurons with a typical Dogiel type II morphology that are immunoreactive for ChAT and tachykinins, about half of which are also immunoreactive for calbindin. NPY immunoreactivity defines the shapes of one phenotype (Furness et al., 1985), but neither VIP nor calretinin, nor other markers of the VIP and calretinin phenotypes, reveal the dendritic morphologies of these types (Furness et al., 1984; Brookes et al., 1991). Thus, the shapes of two of the types (VIP and calretinin neurons) have not been properly determined, and they have not been distinguished from each other or from the NPY neurons. In one study, submucosal neuron types were examined by fluorescence microscopy after cell filling (Evans et al., 1994). In that study, both NPY and VIP neurons are described as having Dogiel type III morphology, although few examples are shown. Calretinin neurons were not identified after dye filling. In a separate investigation, neurons were filled with dye and the VIP and NPY neurons were identified in normal submucosa and in submucosa after 3–5 days in organ culture (Song et al., 1997), but neuron shapes were only briefly described. Projections of the neurons within the submucosal layer have been examined after dye filling in a restricted number of studies (Evans et al., 1994; Reed and Vanner, 2001). In one of these studies (Evans et al., 1994), collaterals of the axons of VIP neurons were reported in submucosal ganglia; in another (Reed and Vanner, 2001), the axons of Dogiel type II neurons were shown to provide networks of varicose fibres in the ganglia. In the present work, we used dye filling from intracellular microelectrodes and immunohistochemical methods to define the shapes and projections of each type of submucosal neuron. This allowed us to make further deductions about their connections and functions.
MATERIALS AND METHODS Guinea-pigs from the inbred Hartley strain colony of the Department of Anatomy and Cell Biology at the University of Melbourne were used in this study. All efforts were made to minimise animal suffering and the number of animals used. The experiments conformed to National Health and Medical Research Council of Australia guidelines and were approved by the University of Melbourne Animal Experimentation Ethics Committee. Guinea-pigs (180 –250 g) of either sex were stunned by a blow to the head and killed by severing the carotid arteries and spinal cord. Segments of ileum were immediately removed 5–10 cm oral to the ileocaecal junction. The segments were placed in a recording dish lined with silicone elastomer, opened along the line of the mesenteric attachment, and pinned flat under moderate tension with the mucosa uppermost. The mucosa was gently removed from the underlying submucosa with fine forceps, and the preparation was turned over and repinned. The serosa and external musculature were then removed to yield a preparation of submucosa. During dissection the tissue was immersed in physiological saline (composition in mM: NaCl 118, KCl 4.8, NaHCO3 25, NaH2PO4 1.0, MgSO4 1.2, glucose 11.1, CaCl2 2.5) and kept at room temperature. The recording dish (volume 4 mL) was placed on the stage of an inverted microscope and continuously superfused (4 mL/min) with physiological saline that had been preheated to yield a
bath temperature of 35°C. The tissue was equilibrated with perfusate for 1 hr before recording was commenced. Neurons were impaled with conventional borosilicate glass micropipettes filled with 1% biocytin (Sigma, Sydney, Australia) in 1 M KCl. Electrode resistances were 80 –200 M⍀. The signals were amplified using an AxoClamp 2B amplifier (Axon Instruments, Foster City, CA), digitised at 1–10 kHz, and stored using PC-based data acquisition software (Axoscope 8.0). The positions of cells in the ganglia were mapped, and the ganglia were drawn for later cell localisation and identification. The tissue was then fixed overnight in 2% formaldehyde plus 0.2% picric acid in 0.1 M sodium phosphate buffer, pH 7.0, cleared in three changes of dimethylsulfoxide, and washed three times in phosphate-buffered saline (PBS; 0.9% NaCl in 0.01 M sodium phosphate buffer, pH 7.0). The tissue was then reacted with streptavidin coupled to Texas Red to reveal biocytin fluorescence, and processed for the immunohistochemical demonstration of calretinin or VIP immunoreactivity. For calretinin localisation, we used rabbit anti-calretinin (Swant Antibodies, Bellizona, Switzerland) at 1:1,000; for VIP, we used rabbit anti-VIP, code 7913 (Furness et al., 1981), at 1:200. After the tissue was viewed to determine whether intracellularly labelled neurons were immunoreactive, the streptavidin was converted to a permanent deposit using goat anti-streptavidin antiserum coupled to biotin (Vector Laboratories, Burlingame, CA), diluted 1:50, incubated with the preparations for 16 –24 hr at room temperature. The biotin was in turn localised using an avidin-biotinHRP kit (Vectastain; Vector Laboratories). The HRP was reacted with diaminobenzidine (DAB) and H2O2 to yield the permanent deposit. In order to determine the shapes of NPY immunoreactive neurons in the submucosal ganglia, tissue samples were prepared in which immunoreactivity was revealed with DAB. Wholemounts, fixed and prepared as above, were incubated in primary antisera against NPY (sheep anti-NPY, E2210, 1:400 (Furness et al., 1985)). After the preparations were washed they were incubated with biotinylated donkey anti-sheep IgG, 1:200. The biotin was localised using an avidin-biotin-HRP kit and DAB (see above). Cell shapes and projections were evaluated on an Olympus BH microscope under positive-low phase contrast optics, and drawn with the aid of a camera lucida drawing tube at 400⫻ or 1000⫻ magnification.
RESULTS We investigated the shapes and projections of 98 filled neurons from 44 preparations, each from a different guineapig. Four shapes could be distinguished amongst the neurons that were filled with biocytin. Further investigation of these neurons (described below) led to their designation as VIP phenotype, NPY phenotype, stellate phenotype, or Dogiel II phenotype (Fig. 1). Of the 98 neurons, 27 were large Dogiel type I (VIP phenotype) neurons, 17 were NPY neurons, 19 were stellate neurons, and 16 were Dogiel type II neurons. Nineteen neurons were inadequately filled and their morphologies were not identified.
Neurons With the VIP Phenotype Immunoreactivity for VIP was observed in a high proportion of submucosal neurons, and previous counts indi-
MORPHOLOGIES AND PROJECTIONS OF NEURONS
477
Fig. 1. Neurons of the four phenotypes encountered in the submusocal ganglia. A: A neuron of the VIP phenotype, the morphology of which was revealed by intracellular injection of biocytin. Arrow in this and other images indicates the axon. The silhouette of this neuron is shown in Figure 3. B: A neuron of the NPY phenotype (morphology revealed by intracellular injection of biocytin). These neurons typically had dendrites that were clustered on one side or at one end of the cell body. C: Three
neurons revealed by NPY immunoreactivity. The dendrites run toward the centre of the ganglion. D: A single neuron with NPY immunoreactivity. Dendrites are clustered at the pole opposite the axon (arrow, only partly in focus). E: A stellate neuron (morphology revealed by intracellular injection of biocytin). F: Dogiel II neuron (morphology revealed by intracellular injection of biocytin). The silhouette of this neuron is in Figure 7. Scale bar: 20 m.
cate that these comprise about 45% of the nerve cell bodies (Furness et al., 1984). VIP immunoreactivity was strongest around the nucleus in the cell bodies, as previously reported (see Introduction), and did not reveal the dendritic processes. Neurons that were filled with biocytin, and were shown by double staining to be immunoreactive for VIP, had complex irregular dendrites and a single axon (Figs. 2 and 3). The dendrites were a mixture of short processes (up to about 15 m long) that were often flattened in the plane of the ganglion (lamellar processes) and longer fine processes (about 20 –30 m in length), with irregular surfaces (Fig. 3). Axons often had irregularities in the first 20 – 40 m of their length. This morphology was not shared by neurons of the three other phenotypes. Neurons that were filled and had this morphology, in preparations that were not stained for VIP, were readily identified. The axons of neurons with the VIP phenotype could be traced for long distances (0.5–3 mm) within the submucosal plexus (Fig. 4). Along their course, the axons were sometimes varicose, but they passed through many submucosal ganglia without giving rise to networks of terminals or appearing to innervate nerve cell bodies. In rare cases, a small side-branch (no more than 15 m long) occurred. The axons sometimes came in close proximity to submucosal arterioles, and followed the arterioles without
branching over them. It was common for the axons to branch once or twice within the submucosal plexus (Fig. 4). The axons ended abruptly, often in a small swelling, which is interpreted to be an end-bulb where the living axon was broken when the adjacent layers were removed by dissection.
Neurons With the NPY Phenotype The morphologies of neurons of this phenotype could be recognised in preparations stained to reveal NPY immunoreactivity, or after dye-filling (Figs. 1 and 5). These neurons had long (20 – 40 m) tapering and often branching dendrites, and a single axon. They did not have the numerous short lamellar dendrites that were observed on neurons of the VIP phenotype. The dendrites were generally grouped, arising from one pole or side of the neuron (Fig. 5). The clustered dendrites were often directed toward the centre of the ganglia (Fig. 1). The extent and patterns of the dendrites were similar in terms of NPY immunoreactivity and dye filling (Fig. 5). The axons were generally smooth in the first 20 – 40 m after leaving the cell body, unlike axons of neurons of the VIP phenotype. Axons of neurons of the NPY phenotype could be traced for only short distances (⬍1 mm) in the submucosal plexus before they finished in end-bulbs, which, like those of the VIP neurons, were probably points where the axons had
478
FURNESS ET AL.
Fig. 2. A: A VIP neuron that was filled with biocytin, which is shown by binding of streptavidin-Texas Red. The varicose axon at the left is from a different neuron. B: Immunoreactivity for VIP exhibited by the filled neuron (arrow). C: The neuron drawn by camera lucida after the
biocytin was converted to a permanent deposit with diaminobenzidine. The neuron has a morphology typical of neurons with the VIP phenotype (see Fig. 3).
Fig. 3. Camera lucida images of neurons of the VIP phenotype. These neurons typically have a mixture of short lamellar dendrites (arrows indicate some examples of lamellar dendrites), longer tapering dendrites, and a single axon. The neurons marked 1A, 2C, and 4 are in the corresponding figures. The neurons with asterisks were shown to be immunoreactive for VIP. The unmarked neurons were not tested for VIP immunoreactivity.
been broken off during dissection. The axons often branched within the plexus. Where they passed through ganglia they did not give rise to side-branches. There were numerous NPY-immunoreactive axons around submucosal arterioles, but it was not possible to trace these back to immunoreactive cell bodies. Other investigations have shown that these are axons of NPY-immunoreactive sympathetic, noradrenergic neurons (Furness et al., 1983). Axons of biocytin-filled neurons of the NPY phenotype did not follow arterioles in the submucosa. During electrophysiological recording, axonal flow causes an accumulation of transported material at the broken end, resulting in the formation of a swelling—the end-bulb.
Fig. 4. Camera lucida drawing of a neuron of the VIP phenotype, showing the course of its axon, which runs through 10 ganglia (g), but does not provide any collaterals to innervate the ganglion cells it passes. Along its course, the axon divides. It finishes in end-bulbs (arrows), which are swellings where the axon broke when it entered other layers, and material has accumulated by axonal flow.
MORPHOLOGIES AND PROJECTIONS OF NEURONS
479
Fig. 6. Stellate neurons. These are camera lucida images, after the injected biocytin has been revealed by the diaminobenzidine reaction. The neurons have a single axon and tapering, branched dendrites arranged around the cell body. A photomicrograph of the cell marked 1E is included in Figure 1.
Fig. 5. Neurons of the NPY phenotype. A: Neuron shapes revealed after the neurons were filled with biocytin that was localised by a diaminobenzidine reaction product (see Fig. 1). A photograph of the cell marked 1B is shown in Figure 1. B: Neuron shapes revealed by immunoreactivity for NPY. Neurons of the NPY phenotype typically have a single axon, and the majority of their dendrites are clustered at one side or one pole.
Neurons With a Stellate Phenotype These neurons were identified only by dye filling. They had round cell bodies and tapering dendrites, which sometimes branched, and a single axon (Fig. 6). The dendrites were arranged around the cell bodies, unlike the clustered dendrites of neurons of the NPY phenotype. Moreover, unlike the neurons of the VIP phenotype, these neurons did not have short lamellar dendrites. The initial parts of the axons occasionally had swellings, but not as commonly as did the VIP neurons. Neurons of this type were tested for VIP or calretinin immunoreactivity after impalement. They were not immunoreactive for these markers, although neurons with these immunoreactivities were observed in the same ganglia. The axons of stellate neurons ran for short distances, sometimes through ganglia, but they did not provide sidebranches in the ganglia. After they ran for distances of about 0.5–1 mm, the axons terminated in end-bulbs. Axons sometimes followed arterioles, but individual axons did not create a perivascular plexus.
Neurons With a Dogiel II Phenotype These neurons were readily recognised after intracellular dye filling. They had generally smooth cell bodies and gave rise to several (usually two or three) long axonal processes (Fig. 7). Occasionally there was only one axonal process. They sometimes had a small number of short
Fig. 7. Dogiel II neurons in camera lucida images after intracellular dye filling. These neurons had several long processes (usually two or three) and a small number of dendrite-like processes that were not very elaborate. A photomicrograph of the cell marked 1F is included in Figure 1.
(5–15 m long) fine processes. The initial parts of the axons were usually smooth. The axons of Dogiel type II neurons provided varicose terminals in submucosal ganglia, generally in adjacent ganglia, but occasionally in the ganglion containing the cell body (Fig. 8). The branches were varicose and ap-
480
FURNESS ET AL.
et al., 1984; Brookes et al., 1991; Song et al., 1992; Evans et al., 1994; Quinson et al., 2001). The present work indicates that neurons of each of these types have an identifiable combination of morphology and projection pattern within the submucosal plexus (Table 1). The functions of the different types of submucosal neurons were deduced from previously published data regarding neurons, including projections to other targets, electrophysiological properties, and neurochemistries (Furness et al., 2000).
Secretomotor/Vasodilator Neurons of the VIP Phenotype
Fig. 8. A Dogiel II neuron shown in a camera lucida drawing. The cell body (arrow) gives rise to a number of processes that provide networks of fibres in the ganglion of the cell body and in other ganglia. A: Part of the area marked inset, shown at greater magnification. Camera lucida drawing. B: Photomicrograph of a portion of the terminal network shown in the inset and in part A. Not all fibres are in the plane of focus of the micrograph.
peared to envelope neuronal somata. Some axons finished as end-bulbs. There were no differences in axonal projection patterns between neurons with and without short processes.
Electrophysiological Characteristics The electrophysiological properties of the neurons were similar to those previously reported (Bornstein et al., 1989; Evans et al., 1994; Song et al., 1997). Action potentials in Dogiel type II neurons were followed by longlasting hyperpolarising potentials (AHPs), and fast excitatory postsynaptic potentials (EPSPs) were not evoked in these neurons by stimulation of the nerve fibre bundles connecting to the ganglia. Slow EPSPs in response to 10 Hz stimulation for 1–3 sec were observed. In contrast, nerve fibre stimulation caused fast EPSPs in VIP, NPY, and stellate neurons. Long-duration AHPs were not observed in these neurons.
DISCUSSION Previous studies of neurons with cell bodies in submucosal ganglia of the guinea-pig small intestine revealed four types, which were distinguished on the basis of their immunoreactivity for particular neurochemicals (Furness
Neurons with VIP immunoreactivity comprise about 45% of nerve cells in the submucosal ganglia (Furness et al., 1984). However, cell shapes are not revealed by immunoreactivity for VIP or other marker substances of these neurons, such as dynorphin (Steele and Costa, 1990) and galanin (Furness et al., 1987). These markers are localised to the perinuclear cytoplasm, and do not reveal dendrites. Therefore, with currently available technology, the morphologies of cell bodies of VIP phenotype neurons are best determined by intracellular injection of dye, and subsequent identification by immunohistochemistry. This was done by Evans et al. (1994), who identified VIP neurons as receiving inhibitory synaptic inputs, and having at least five (and often more) filamentous dendrites and a single long axon that sometimes branched and ran through many ganglia. They referred to the shape as Dogiel type III morphology. Song et al. (1997) also identified filled neurons in the submucosal ganglia by their VIP immunoreactivity. They described the neurons as having a single axon and filamentous dendrites, and did not distinguish their morphologies from the morphologies of NPY neurons. We identified VIP neurons with similar morphology, but were able to distinguish them from NPY neurons. We observed that the VIP neurons had short lamellar dendrites, in addition to the filamentous dendrites, and that they commonly had swellings or eccentric protuberances on the first part of their axons. These features were in fact illustrated in previous studies (Evans et al., 1994; Song et al., 1997). The shapes are more similar to those of Dogiel type I neurons than to Dogiel type III neurons, which have very long tapering dendrites (Dogiel, 1899; Stach, 1980, 1982a; Brehmer et al., 1999). The axons of the VIP neurons were traced through several submucosal ganglia and were found to travel greater distances in the plexus than the axons of other neuron types. Retrograde tracing also demonstrates that VIP neurons run for greater distances in the submucosa than do the other three types of neuron (Song et al., 1992). Sometimes the axons give rise to short side-branches in the ganglia, but generally they run through the ganglia without branching (Evans et al., 1994; Song et al., 1997; current results). The side-branches do not form pericellular baskets. In fact, all VIP-immunoreactive pericellular endings disappear from around the submucosal nerve cells following section of the fibres that run to the submucosa from the myenteric plexus (Costa and Furness, 1983). On the other hand, axons of VIP neurons can be directly traced to extensive arborisations in the mucosa, beneath the mucosal epithelium, and to arterioles (Song et al., 1997). Nerve lesion studies also demonstrate that VIP immunoreactive fibres innervating the mucosa and submucosal arterioles arise from submucosal nerve cells (Costa and Furness, 1983). Some collaterals of the axons
481
MORPHOLOGIES AND PROJECTIONS OF NEURONS
TABLE 1. The characteristics of the four types of neurons whose cell bodies are in submucosal ganglia of the guinea pig small intestine Neuronal phenotype Functional designation Shape Dendrites Axon length Endings in submucosal ganglia Targets of nerve terminals
Axon initial segment Principal chemical marker Other markers
VIP
NPY
Stellate
Non-cholinergic secretomotor/ vasodilator neuron Dogiel type I Mixed lamellar and filamentous Long within submucosa No
Cholinergic secretomotor neuron
Cholinergic secretomotor/ vasodilator neuron
Type IV (Stach) Filamentous: Tapering, branched, clustered Short within submucosa No
Stellate Filamentous: Tapering branched, not clustered Short within submucosa No
Mucosal epithelium, submucosal arterioles, myenteric ganglia
Mucosal epithelium
Mucosal epithelium, submucosal arterioles
Frequent spines VIP
Smooth NPY
Rare spines calretinin
Galanin, NMU, PHI, dynorphin
ChAT, CGRP, SOM
ChAT
of submucosal VIP neurons project to myenteric ganglia (Song et al., 1998).
Secretomotor Neurons of the NPY Phenotype The dendrites of these neurons, unlike those of other uniaxonal submucosal neurons, were readily revealed by immunohistochemistry for the chemical marker. The dendritic morphologies revealed by intracellularly injected dye and by NPY immunoreactivity were indistinguishable, indicating that the dendrites were fully stained by anti-NPY antibodies. It was previously pointed out that these neurons in the guinea-pig submucosal ganglia have type IV morphology, as originally defined by Stach (Brehmer et al., 1999). Neurons of this morphology have a single axon and asymmetrically arranged tapering filamentous dendrites (Stach, 1982b; Brehmer et al., 1999). As previously reported, these neurons did not provide terminals within submucosal ganglia. They provide innervation of the mucosa but not of submucosal arterioles, and are cholinergic secretomotor neurons (Furness et al., 1983, 1985, 2000).
Secretomotor Neurons of the Stellate Phenotype Neurons of this third group were distinguishable by shape from both the VIP and NPY neurons. The neurons were monoaxonal and had filamentous dendrites that, unlike the dendrites of NPY neurons, were not aggregated to one side. Unlike the VIP neurons, these neurons did not have lamellar dendrites, and, also unlike the VIP neurons, the axons ran only short distances in the submucosal plexus. Moreover, neurons of this type did not exhibit VIP immunoreactivity. In submucosal ganglia, monoaxonal neurons that are not NPY or VIP immunoreactive are calretinin immunoreactive cholinergic secretomotor/vasodilator neurons (Costa et al., 1996; Furness et al., 2000). However, the stellate neurons that were identified after dye filling were not calretinin-immunoreactive. Evans et
Dogiel II Intrinsic primary afferent neuron Dogiel type II Few, short, fine Short within submucosa Yes, pericellular networks Mucosal epithelium, submucosal and myenteric ganglia Smooth ChAT, TK ⫾ calbindin, cytoplasmic NeuN
al. (1994) also reported that neurons that were dye-filled were not immunoreactive for calretinin. We conclude that impalement of the neurons interferes with immunoreactivity of the neurons for calretinin, and that these are indeed neurons that normally contain calretinin. These neurons did not supply terminals in submucosal ganglia, which is consistent with the absence of calretinin immunoreactive nerve terminals from the ganglia (Brookes et al., 1991). Calretinin neurons are also immunoreactive for choline acetyltransferase, have axons that are immunoreactive for the vesicular acetylcholine transporter, and project to arterioles and the mucosa (Neild et al., 1990; Brookes et al., 1991; Li et al., 1998).
Dogiel Type II Neurons The fourth neuron type that was found was the Dogiel type II neuron. These neurons have been shown to have immunoreactivity for tachykinins and choline acetyltransferase, but neither of these markers adequately reveal the processes of the neurons (Furness et al., 1984; Song et al., 1992). Dye filling, correlated with electrophysiological characterisation, showed these to be (electrophysiologically) AH neurons (Iyer et al., 1988; Evans et al., 1994). By shape, they are smooth-surfaced neurons, with a small number of short processes and one to four long processes (Iyer et al., 1988; Evans et al., 1994). In one previous study in the guinea-pig ileum, their projections were followed after they were filled with marker dye that was converted to a fluorescent product (Evans et al., 1994). Axons were followed through ganglia, where they gave rise to small side-branches that did not form networks around nerve cell bodies. However, they were found to form networks around nerve cell bodies in a later investigation, in which DAB was used to provide a permanent staining of dyefilled axons (Reed and Vanner, 2001). In the present work, the biocytin was also converted to a permanent DAB deposit, which reveals more detail of the side-branches, in agreement with the observations of Reed and Vanner
482
FURNESS ET AL.
(2001). The side-branches formed extensive networks in the ganglia. A similar result was obtained in the distal colon, where Dogiel type II neurons also have AH electrophysiology and innervate other nerve cells (Lomax et al., 2001).
Reflex Pathways Secretomotor reflexes can be initiated by stimulation of the mucosa, and have, as their final motor neurons, secretomotor and secretomotor/vasodilator neurons with cell bodies in submucosal ganglia (Cooke and Reddix, 1994; Furness et al., 2000). Although there is evidence for secretomotor reflex pathways that pass via the myenteric plexus (Jodal et al., 1993), secretomotor reflexes can be evoked in segments of intestine in which the myenteric plexus is removed or ablated (Diener and Rummel, 1990; See et al., 1990). This implies that synaptic connections from intrinsic primary afferent neurons (IPANs) to interneurons or directly to motor neurons occur in submucosal ganglia. Consistent with these observations, stimuli applied to the mucosa in segments of intestine with the submucosa connected, but the external muscle and myenteric plexus removed, cause excitatory postsynaptic potentials (EPSPs) in submucosal nerve cells (Pan and Gershon, 2000). Fast EPSPs occurred in NPY, VIP, and calretinin neurons, that is, in each of the three classes of secretomotor neurons. Thus the present data, which demonstrate that the Dogiel type II neurons are the only submucosal neurons that provide substantial numbers of pericellular endings in submucosal ganglia, plus previous data indicating that these are intrinsic primary afferent neurons (Kirchgessner et al., 1992), suggest that there are monosynaptic reflex pathways in which IPANs connect directly with secretomotor and secretomotor/vasodilator neurons. These could be seen as a peripheral analogue of spinal monosynaptic reflex connections. Like the spinal reflexes, they are probably subjected to the modifying influences of other inputs—for example, from the myenteric plexus and from sympathetic pathways (Cooke and Reddix, 1994; Furness et al., 2000).
Comparisons With Other Species In larger species (e.g., pigs and humans) the submucosal plexus contains more neuron types, is more complex, and is directly involved in control of both secretion and motility (Timmermans et al., 2001). Therefore, although the functional types of neuron that occur in the submucosal ganglia of the guinea-pig small intestine likely to occur in larger mammals, there are other neuron types present, and it is not possible to extrapolate the relationships between morphology and function, as determined in the present work, to other species.
CONCLUSIONS We have discovered that each of the four types of neurons with cell bodies in the submucosal ganglia of the guinea-pig small intestine has a distinctive morphology. The Dogiel type II neurons, which are intrinsic primary afferent neurons, provide networks of terminals around the other types, which are secretomotor neurons. These new data, in combination with previously reported functional data, indicate the existence of simple secretomotor reflex pathways that consist of a primary afferent neuron connecting directly with a motor neuron.
LITERATURE CITED Bornstein JC, Furness JB, Costa M. 1989. An electrophysiological comparison of substance P-immunoreactive neurons with other neurons in the guinea-pig submucous plexus. J Auton Nerv Syst 26:113–120. Brehmer A, Schrodl F, Neuhuber W. 1999. Morphological classifications of enteric neurons—100 years after Dogiel. Anat Embryol 200:125–135. Brookes SJH, Steele PA, Costa M. 1991. Calretinin immunoreactivity in cholinergic motor neurones, interneurones and vasomotor neurones in the guinea-pig small intestine. Cell Tissue Res 263:471– 481. Brookes SJH. 2001. Classes of enteric nerve cells in the guinea-pig small intestine. Anat Rec 262:58 –70. Cooke HJ, Reddix RA. 1994. Neural regulation of intestinal electrolyte transport. In: Johnson LR, editor. Physiology of the gastrointestinal tract. New York: Raven Press. p 2083–2132. Costa M, Furness JB. 1983. The origins, pathways and terminations of neurons with VIP-like immunoreactivity in the guinea-pig small intestine. Neuroscience 8:665– 676. Costa M, Brookes SJH, Steele PA, Gibbins I, Burcher E, Kandiah CJ. 1996. Neurochemical classification of myenteric neurons in the guinea-pig ileum. Neuroscience 75:949 –967. Diener M, Rummel W. 1990. Distension-induced secretion in the rat colon: mediation by prostaglandins and submucosal neurons. Eur J Pharmacol 178:47–57. ¨ ber den Bau der Ganglien in den Geflechten des Dogiel AS. 1899. U Darmes und der Gallenblase des Menschen und der Sa¨ugetiere. Arch Anat Physiol Leipzig Anat Abt Jg 1899:130 –158. Evans RJ, Jiang MM, Surprenant A. 1994. Morphological properties and projections of electrophysiologically characterized neurons in the guinea-pig submucosal plexus. Neuroscience 59:1093–1110. Furness JB, Costa M, Walsh JH. 1981. Evidence for and significance of the projection of VIP neurons from the myenteric plexus to the taenia coli in the guinea-pig. Gastroenterology 80:1557–1561. Furness JB, Costa M, Emson PC, Håkanson R, Moghimzadeh E, Sundler F, Taylor IL, Chance RE. 1983. Distribution, pathways and reactions to drug treatment of nerves with neuropeptide Y- and pancreatic polypeptide-like immunoreactivity in the guinea-pig digestive tract. Cell Tissue Res 234:71–92. Furness JB, Costa M, Keast JR. 1984. Choline acetyltransferase and peptide immunoreactivity of submucous neurons in the small intestine of the guinea-pig. Cell Tissue Res 237:329 –336. Furness JB, Costa M, Gibbins IL, Llewellyn Smith IJ, Oliver JR. 1985. Neurochemically similar myenteric and submucous neurons directly traced to the mucosa of the small intestine. Cell Tissue Res 241:155–163. ˚ , McDonald TJ, Brooks BD. 1987. Furness JB, Costa M, Ro¨kaeus A Galanin-immunoreactive neurons in the guinea-pig small intestine: their projections and relationships to other enteric neurons. Cell Tissue Res 250:607– 615. Furness JB. 2000. Types of neurons in the enteric nervous system. J Autonom Nerv Syst 81:87–96. Furness JB, Clerc N, Gola M, Kunze WAA, Fletcher EL. 2000. Identification of component neurons and organisation of enteric nerve circuits. In: Krammer HJ, Singer MV, editors. Neurogastroenterology—from the basics to the clinics. Dordrecht: Kluwer Academic Publishers. p 134 –147. Iyer V, Bornstein JC, Costa M, Furness JB, Takahashi Y, Iwanaga T. 1988. Electrophysiology of guinea-pig myenteric neurons correlated with immunoreactivity for calcium binding proteins. J Auton Nerv Syst 22:141–150. Jodal M, Holmgren S, Lundgren O, Sjo¨qvist A. 1993. Involvement of the myenteric plexus in the cholera toxin-induced net fluid secretion in the rat small intestine. Gastroenterology 105:1286 –1293. Kirchgessner AL, Tamir H, Gershon MD. 1992. Identification and stimulation by serotonin of intrinsic sensory neurons of the submucosal plexus of the guinea pig gut: activity-induced expression of Fos immunoreactivity. J Neurosci 12:235–248.
MORPHOLOGIES AND PROJECTIONS OF NEURONS Li ZS, Fox Threlkeld JET, Furness JB. 1998. Innervation of intestinal arteries by axons with immunoreactivity for the vesicular acetylcholine transporter VAChT. J Anat 192:107–117. Lomax AE, Bertrand PP, Furness JB. 2001. Electrophysiological characteristics distinguish three classes of neuron in submucosal ganglia of the guinea-pig distal colon. Neuroscience 103:245–255. Neild TO, Shen KZ, Surprenant A. 1990. Vasodilatation of arterioles by acetylcholine released from single neurones in the guinea-pig submucosal plexus. J Physiol (Lond) 420:247–265. Pan H, Gershon MD. 2000. Activation of intrinsic afferent pathways in submucosal ganglia of the guinea pig small intestine. J Neurosci 20:3295–3309. Quinson N, Robbins HL, Clark MJ, Furness JB. 2001. Calbindin immunoreactivity of enteric neurons in the guinea-pig small intestine. Cell Tissue Res 305:3–9. Reed DE, Vanner SJ. 2001. Converging and diverging cholinergic inputs from submucosal neurons amplify activity of secretomotor neurons in guinea-pig ileal submucosa. Neuroscience 107:685– 696. See NA, Greenwood B, Bass P. 1990. Submucosal plexus alone integrates motor activity and epithelial transport in rat jejunum. Am J Physiol 259:G593–G598. Song ZM, Brookes SJH, Steele PA, Costa M. 1992. Projections and pathways of submucous neurons to the mucosa of the guinea-pig small intestine. Cell Tissue Res 269:87–98.
483
Song ZM, Brookes SJH, Neild TO, Costa M. 1997. Immunohistochemical and electrophysiological characterization of submucous neurons from the guinea-pig small intestine in organ culture. J Auton Nerv Syst 63:161–171. Song ZM, Costa M, Brookes SJH. 1998. Projections of submucous neurons to the myenteric plexus in the guinea pig small intestine. J Comp Neurol 399:255–265. Stach W. 1980. Zur neuronalen Organisation des Plexus myentericus (Auerbach) im Schweinedu¨nndarm I. Typ I-Neurone. Z Mikrosk Anat Forsch 94:833– 849. Stach W. 1982a. Zur neuronalen Organisation des Plexus myentericus (Auerbach) im Schweinedu¨nndarm III. Typ III-Neurone. Z Mikrosk Anat Forsch 96:497–516. Stach W. 1982b. Zur neuronalen Organisation des Plexus myentericus (Auerbach) im Schweinedu¨nndarm IV. Typ IV-Neurone. Z Mikrosk Anat Forsch 96:972–994. Steele PA, Costa M. 1990. Opioid-like immunoreactive neurons in secretomotor pathways of the guinea-pig ileum. Neuroscience 38: 771–786. Timmermans J-P, Hens J, Adriaensen D. 2001. Outer submucous plexus: an intrinsic nerve network involved in both secretory and motility processes in the intestine of large mammals and humans. Anat Rec 262:71–78.
