Dr S.Y.Yuan Department of Human Physiology and Centre for Neuroscience, Flinders University, GPO Box 2100, Adelaide, South Australia 5001, Australia. Tel.: + 61 8 8204 5006; fax: + 61 8 8204 5768; e-mail: firstname.lastname@example.org
The pyloric sphincter (PS) controls gastric emptying and prevents the reflux of duodenal content into the stomach. Neuronal pathways and reflexes controlling the guinea-pig PS were physiologically investigated in isolated preparations. Simultaneous intracellular or extracellular and tension recordings from PS circular muscle with electrical and stretch stimulation were used. Electrical stimulation evoked an initial small contraction followed by a relaxation with a corresponding inhibitory junction potential (IJP) then a second large contraction with a corresponding excitatory junction potential (EJP). Hyoscine (1 μmol L–1) blocked the first contraction, and reduced the second contraction and EJP by 52.5% and 61%, respectively. These responses were further reduced by the NK2 antagonist, MEN10627 (1 μmol L–1), and the NK1 antagonist, SR140333 (1 μmol L–1). N-nitro-L-arginine (100 μmol L–1) and apamin (0.5 μmol L–1) blocked the relaxation and the IJP. Duodenal electrical stimulation evoked an EJP, whereas antral stimulation evoked an IJP followed by a small EJP. All were blocked by hexamethonium (100 μmol L–1). Duodenal stretch evoked tetrodotoxin-sensitive reflex contractions and membrane depolarization with action potentials in the PS. Thus, PS enteric motor neurones receive inputs from the duodenum and the stomach. There are stretch-sensitive ascending excitatory reflex pathways from the duodenum to the PS.
The pyloric sphincter (PS) is a bundle of thickened circular smooth muscle located between the stomach and the duodenum, first described in humans and in other mammals (rabbit, cat, dog, pig, ox and horse) in 1942.1 A later study in smaller animals confirmed the existence of an anatomical sphincter.2 Its location indicates that this sphincter controls the emptying and retention of gastric contents.1 In the gastro-pyloro-duodenal region, sequential contractions of the antrum, pylorus, and proximal duodenum result in smooth gastric emptying and rapid removal of the chyme from the duodenum. If the contractions in the three areas are not temporally coordinated, the pylorus or duodenum, or both, may cause resistance to gastric emptying.3 Relaxation of the PS is produced by strong inhibitory inputs received during gastric emptying.4 When gastric contents reach the duodenum, it activates chemo-and mechano-sensory receptors that control a feedback mechanism to regulate the motility of the antroduodenal canal and adjust the rate of gastric emptying.3
Compared to neighbouring regions, the pylorus is supplied with a relatively high density of nerve fibres.2 It receives excitatory and inhibitory inputs from enteric motor neurones, which are influenced by both extrinsic nerves and hormonal factors3–6. The pylorus can be either excited or inhibited by stimulation of vagus nerves7,8 and is also inhibited by sympathetic stimulation.7 Nitric oxide synthase (NOS) and vasoactive intestinal peptide (VIP) immunoreactive nerve fibres have been found in the canine PS.9 The release of NO and VIP in the pylorus in response to antral and vagal electrical stimulation suggests that both transmitters may play a role in pyloric inhibition.10 Acetyl-choline and substance P have been suggested to mediate excitation of the pylorus.7,11
Polarized enteric reflex activities and underlying reflex pathways have been demonstrated in the stomach and the small and large intestines.12–18 Previous functional studies in dog, pig and cat have revealed the existence of descending inhibitory pathways from the gastric antrum to the pylorus, which influence the timing of pyloric closure.7,19,20 It has also been demonstrated that distension, or chemical or osmolar stimulation of the duodenum can activate ascending excitatory motor pathways to the antro-pyloric region and thereby slow gastric emptying.20–23 The function of feedback control via the intrinsic pathways can be abolished by duodenal transection.23 It has been shown that the enteric nervous system is continuous from the stomach, through the PS, to the duodenum.24 However, the underlying intrinsic pathways projecting from the stomach and the duodenum to the pyloric sphincter are still poorly understood. Work in our laboratory using retrograde tracing and immunohistochemistry revealed that the PS of the guinea-pig is innervated by enteric short choline acetyltransferase (ChAT)-immunoreactive excitatory and NOS-immunoreactive inhibitory motor neurones and by longer descending NOS-containing inhibitory motor neurones from the stomach.25 It is important to correlate morphological evidence with functional study in this species. In the present study, we used simultaneous intracellular, extracellular and tension recordings to investigate the enteric neuronal reflex pathways to the PS of the guinea-pig.
MATERIALS AND METHODS
Adult guinea-pigs of both sexes (250–350 g), were killed by cervical dislocation, followed by bleeding from the carotid arteries, a technique approved by the Flinders University Animal Ethics Committee. The stomach with the PS and 4 cm of duodenum attached was quickly removed and placed in a petri dish filled with modified Krebs solution (composition in mmol L–1: NaCl 118; KCl 4.7; MgSO4 1.2; NaHCO3 25; NaH2PO4 1.0; CaCl2 2.5; D-glucose 11). The gastric fundus was removed. The remaining stomach and duodenum were cut open by an incision along the lesser curvature, continuing through the PS and the duodenum segment. The tissue was washed clean with fresh Krebs solution and transferred to a 40-mL organ bath for simultaneous intracellular or extracellular and tension recordings. The preparation in the organ bath was superfused continuously with Krebs solution at 3.3 mL min–1. Drugs were administered into the organ bath by a peristaltic pump, from a reservoir containing the final concentration of drugs in the superfusate, bubbled with 95% O2/5% CO2 and maintained at 37 °C. The preparations in the organ bath were set up as described below and allowed to equilibrate for at least 30 min before recording. All data were recorded onto Maclab recording systems (ADI, Castle Hill, NSW, Australia) with sampling frequencies of 100 Hz for mechanical recordings and 1 kHz for electrophysiological recordings.
For the purpose of pharmacological investigation of the transmitters involved in neuromuscular transmission, PS preparations with circular muscle and myenteric plexus attached (2–2.5 mm longitudinally and full circumferential length) were isolated from the gastroduodenal junction after removal of mucosa and submucosa (isolated PS preparation). One end of the preparation was pinned down with entomological pins (diameter 200 μm), circular muscle upwards, and the other end was attached via an array of hooks to a force transducer for muscle tension recording. Junction potentials of PS circular muscle cells were recorded intracellularly from an area closely pinned by 30 pins, made from 50 μm diameter tungsten wire, on the fixed edge of the preparation. This area provided mechanical stability, while maintaining electrical continuity of the smooth muscle syncytium with the rest of the PS.16 In order to maintain the intracellular recording from the same muscle cell of the PS, exposure of the preparation to drugs was kept to the shortest possible duration consistent with stability of the maximal effects being observed.
Investigation of neuronal pathways
The gastro-pyloro-duodenal preparations were pinned flat with mucosa uppermost. The circular muscle of the PS was exposed after removal of the mucosa and submucosa from the PS region. One end of the PS was closely pinned, as described above, for intracellular recording. Ascending and descending neuronal pathways from the duodenum and the antrum to the PS were activated via two pairs of parallel field-stimulating electrodes (1 mm diameter silver wires, 2 mm apart, insulated apart from the last 200 μm of their tips). Each pair of stimulating electrodes was located separately on the mucosal surface of gastric antrum and duodenum, 5 mm from the PS.
Experiments on enteric reflexes
Gastro-pyloro-duodenal preparations were pinned down along one edge, with the mucosa uppermost. A peninsula (4 mm circumferentially × 2 mm longitudinally) of the PS was made by two short circumferential cuts from one end of the PS parallel to the circular muscle (Fig. 1). The mucosa and submucosa within the peninsula were removed to expose the circular muscle layer of the PS for extracellular and tension recordings. The use of these peninsulas in the recording area reduced mechanical disturbance caused by neighbouring muscle movement. The free end of the peninsula was attached via an array of hooks (2 mm wide), to a force transducer. The electrical activity of sphincteric circular muscle cells was recorded simultaneously with intracellular and extracellular suction electrodes located 1 mm circumferential to the array of hooks within the peninsula. To activate reflex pathways from the duodenum to the PS or antrum, the longitudinal edge of the tissue, 10 mm from the PS, was attached via an array of hooks (20 mm wide) to a purposed-designed ‘tissue stretcher’.26 This consisted of an isometric transducer mounted on a stepper motor-driven linear actuator for stretching the tissue circumferentially with controllable speed and distance (Fig. 1).
Activation of neuronal pathways
In all experiments, focal electrical stimulation on the PS (1, 5, 10, 20 Hz, 0.2 ms for 1 s from a Grass SD9 stimulator) with constant voltage (50 V) was applied through a pair of parallel focal-stimulating electrodes (0.12 mm diameter platinum wires, 0.3 mm apart, insulated to within 1 mm of their tips). These stimulating electrodes were located on the circular muscle of the PS, 2–6 mm circumferential to the recording electrode, to activate motor neurones and their axons in the PS muscle. In the pharmacological experiments, the PS was focally stimulated with electrical pulses at different frequencies (1, 5, 10, and 20 Hz) for 1 s at intervals of 20 s between the trains. In the reflex experiments, circumferential stretching was used as a stimulus. The preparation was circumferentially stretched by 2 mm in the duodenum and by 5 mm in the stomach using the tissue stretcher, at 8 mm s–1 held for 6 s, at 2-min intervals.
After the preparation had equilibrated for at least 30 min, intracellular recordings were made by inserting a glass microelectrode (filled with 2 mol L–1 KCl, resistance 30–80 ω) into an exposed circular muscle cell of the PS. Once a stable impalement was established, with a resting membrane potential more negative than –35 mV for more than 20 s, electrical or reflex (stretch) stimulation was applied. Extracellular recordings were made by attaching a flexible mounted suction electrode with 100–300 μm tip diameter to the exposed circular muscle of the PS. Suction was provided by a syringe via a three-way connector. In order to reveal spontaneous mechanical activity, the slack in the tissue was taken up until a resting tension of 2 mN was reached, which was defined as the resting length. At least three control responses to the electrical or stretch stimulation at each site were recorded, at intervals of 2 min, before applying any drugs.
Tetrodotoxin (TTX), hexamethonium chloride, Nω-nitro-L-arginine (L-NOARG) and apamin were obtained from Sigma Chemical Co. (St Louis, MO, USA); hyoscine hydrobromide was from David Bull Laboratories (Victoria, Australia). Hexamethonium, hyoscine and L-NOARG were directly dissolved in physiological saline and prepared on the day of the experiment. Apamin was dissolved in 0.1% trifluoroacetic acid at a concentration of 5 × 10–4 mol L–1 and stored in frozen aliquots. TTX was stored in aqueous solution at 10–2 mol L–1. SR140333 (1 μmol L–1) and MEN10627 (1 μmol L–1) were dissolved in 0.1% dimethylsulphoxide (DMSO) at concentrations of 10–4 and 10–3 mol L–1 and stored in frozen aliquots. DSMO, at a final concentration of 0.001%, did not have any effect on IJP or EJP elicited by the stimuli in the small intestine.27 Drugs were administered into the organ bath via the Krebs inflow.
In the experiments using intracellular recording, inhibitory junction potentials (IJP) or excitatory junction potentials (EJP) were analysed by measuring their maximum amplitudes and duration of half amplitude of the response. The duration of slow waves was measured from the beginning of the response to the apparent return to the resting membrane potential. The contraction and relaxation induced by direct electrical stimulation on the PS were also analysed by measuring their maximum amplitudes and duration. For extracellular recording, the number of action potentials (spikes) during the stretch stimulation was counted. The multiple peaks of tension response were analysed by measuring the area under the curve. Data are expressed as mean ± standard error from at least four experiments; n equals the number of animals. Differences were compared using Student’s paired t-test; differences were considered to be significant when P < 0.05.
After removal of the mucosa and submucosa, the PS appeared as a distinct thickened circular muscle band (1.5–2.5 mm wide in longitudinal axis of the preparation) at the gastroduodenal junction.
Intracellular recording from these muscle cells revealed that the resting membrane potential of the smooth muscle cells, measured between slow waves, was – 41.7 ± 0.5 mV (n=6, 16 cells). Slow waves with a frequency of 2.3 ± 0.1 cycles min–1 and amplitude of 2–7 mV were recorded in three of 12 isolated PS preparations, when the intracellular recording was carried out from the outer layer of the circular muscle cells near the serosal surface of the PS. Action potentials were observed on the rising phase of some of these slow waves. However, they could not be detected from the inner circular muscle layer. Spontaneous contractions of the PS were observed in all isolated PS preparations and occurred at 2.6 ± 0.2 cycles min–1 (range 1–5 cycles min–1; n=6). These contractions increased in amplitude in the presence of 0.5 μmol L–1 apamin but were not significantly affected by application of 0.6 μmol L–1 TTX (n=6). In three experiments, the slow waves, when detectable, occurred at the same frequency as the PS spontaneous contractions and each preceded a muscle contraction. In most cases, spontaneous contractions of the PS were recorded without slow waves in the impaled cell. Similar spontaneous contractions of the PS (mean, 2.1 ± 0.2 cycles min–1, ranging from 1–6 cycles min–1; n=4) were also observed in gastro-pyloro-duodenal preparations with the stomach and duodenum attached. In the majority of these preparations, spontaneous circular muscle contractions were also recorded in both gastric and duodenal regions, and tended to wax and wane in amplitude, as seen in the sphincter. Their frequencies ranged from 5–10 cycles min–1 (mean 6 ± 0.2 cycles min–1; n=4) in the stomach and from 3–21 cycles min–1 (mean 10 ± 0.9 cycles min–1; n=7) in the duodenum. In addition, spontaneous IJP, but not EJP, up to 9 mV were occasionally recorded in some of the PS preparations. No corresponding relaxation was observed.
Focal stimulation of the PS
Focal electrical stimulation (1, 5, 10 and 20 Hz, 50 V; 0.2 ms for 1 s) on the PS, 6 mm circumferential to the intracellular recording electrode, evoked a frequency-dependent, triphasic mechanical response (Fig. 2). This consisted of an initial, very small contraction immediately following the cessation of stimulation. This contraction was of low amplitude (1.6 ± 0.8 mN; range 0–5 mN at 20 Hz stimulation) and appeared to be abruptly truncated by a relaxation, which was followed by a large second contraction (16.7 ± 5.8 mN; range 0.4–34.8 mN at 20 Hz stimulation). A large IJP, up to 23.6 mV in amplitude, was associated with the relaxation and was followed by a small prolonged EJP, up to 9.5 mV (Fig. 2). Action potentials (3–5 spikes) on the top of the initial phase of EJP were often observed (not shown in figures). The mean amplitudes of these control responses for each stimulating frequency (1, 5, 10 and 20 Hz) are shown separately in Figs 3-5, (n=6). These triphasic responses were reproducible at intervals of 2 min. The amplitude of these responses increased with increasing frequencies of stimulation (Figs 3,4,5).
At 20 Hz stimulation, hyoscine (1 μmol L–1) blocked the first contraction (Figs 3,6) and enhanced both the relaxation and IJP by 61.8% and 14.4%, respectively (Figs 4,6). Hyoscine also greatly reduced the second contraction and EJP by 52.5% and 61%, respectively (Figs 5,6), confirming that excitatory neuromuscular transmission is mediated, in part, via muscarinic cholinergic receptors.28L-NOARG and apamin were then used to establish the involvement of NO and the role of an apamin-sensitive component in the inhibitory transmission.29,30 Application of the NOS inhibitor, L-NOARG (100 μmol L–1), greatly reduced not only the amplitudes of IJP and relaxation but also the duration measured at half peak amplitude of IJP, from 2.0 ± 0.1 s to 1.5 ± 0.1 s (P < 0.05; t=3.6; d.f.=5; n=6), and the duration of relaxation, from 6.5 ± 0.4 s to 3.1 ± 0.3 s (P < 0.05; t=13.8; d.f.=5; n=6). Further addition of apamin (0.5 μmol L–1) blocked both the relaxation and IJP (Figs 4,6), and greatly enhanced the amplitude of both the second contraction and the EJP, by 167.6% and 78.1%, respectively (Figs 5,6). SR140333 and MEN10627 were used to examine the involvement of tachykinins acting on NK1 and NK2 receptors in excitatory neuromuscular transmission.31,32. Addition of the NK2 receptor antagonist, MEN10627 (1 μmol L–1), greatly reduced the second contraction and the EJP, by 32.9% and 17%, respectively (Figs 5). However, at 5 and 10 Hz of stimulation, MEN10627 (1 μmol L–1) did not show similar attenuations of the EJP. Sequential addition of the NK1 receptor antagonist, SR140333 (1 μmol L–1) further reduced the contraction and EJP (Figs 5)). The small residual responses were abolished by 0.6 μmol L–1 TTX (Figs 5).
Electrical stimulation of the stomach and duodenum
The existence of neuronal pathways from the stomach and the duodenum to the PS was examined by applying electrical stimulation to the gastric and duodenal regions to activate the second-order neurones and their axons within these pathways. Electrical field stimulation (10 Hz, 50 V; 1 ms for 1 s) of the gastric antrum, 5 mm from the PS, evoked IJP up to 10 mV in amplitude in the smooth muscle cells of the PS, followed by small EJP up to 7 mV. The mean amplitudes of EJP and IJP were 4.3 ± 0.3 mV and 7 ± 1.4 mV, respectively (n=4) (Fig. 7). Similar stimulation applied to the duodenum, 5 mm from the PS, evoked EJP up to 7.5 mV (mean, 6.8 ± 0.4 mV) (Fig. 7) (n=4). These responses were reproducible at intervals of 2 min, and were considerably smaller than those evoked by electrical stimulation of motor neurone axons directly in the PS, where stimulation, 2 mm circumferential to the recording electrode with lower frequency (1 Hz), evoked larger EJP and IJP (mean, 7.5 ± 0.6 mV and 9 ± 1.1 mV; n=4). Hexamethonium (100 μmol L–1) abolished the EJP evoked by duodenal and antral stimulation and reduced IJP evoked by antral electrical stimulation by more than 80% (P < 0.03; t=4.1; d.f.=3), but did not significantly change responses evoked by stimulation on the PS. Residual inhibitory responses to antral stimulation were abolished by 0.6 μmol L–1 TTX (Fig. 7) (n=4).
Stretching the duodenum and the stomach
The existence of enteric neuronal reflexes from the stomach and duodenum to the PS was examined by mechanical activation (stretching) of enteric reflex pathways in the gastric and duodenal regions. Circumferential stretch (2 mm) of the duodenum in the flat sheet preparations (at 8 mm s–1 for 6 s; n=6) evoked transient multiple phasic contractions in the stretched region of the duodenum (Figs 8,9A), and also phasic ascending reflex contractions in the PS (Figs 8,9B). The area under the curve of response was 940 ± 236 mN.s (n=6) in the duodenum and 67 ± 34 mN.s (n=6) in the PS (Fig. 9). Simultaneously extracellular recording from the PS revealed a depolarization (mean, 195 ± 20 μV; n=6) with 3–11 superimposed action potentials (mean 6 ± 1 action potentials; n=6), which closely correlated with the contractions recorded from the PS (Figs 8,10). These mechanical and electrical responses recorded from the duodenum and the PS were reproducible at intervals of 2 min. With TTX (0.6 μmol L–1) to block neuronal activity, the pyloric reflex responses (contractions, depolarizations and action potentials) (Figs 8,9,10) and the duodenal multiple phasic contraction induced by duodenal stretch (Figs 8,9) were abolished, although small passive mechanical changes in duodenal tension still occurred during distension (n=6).
Circumferential stretch (5 mm) of the stomach, 10 mm oral to the PS (8 mm s–1 for 6 s; n=4), also evoked transient phasic contractions in the stretched region and contractions of the PS, preceded by a small relaxation. Simultaneous extracellular recording from the PS revealed a depolarization (range from 0.1 to 3 mV), in some cases preceded by a small hyperpolarization (range from 0.9 to 1.3 mV) (n=4). In addition, in two of four preparations, a hyperpolarization, up to 950 μV, was recorded from the PS during gastric stretch, and was correlated with the relaxation recorded from the PS. The responses evoked by gastric stretch were poorly reproducible at a fixed interval of 2–10 min, preventing detailed pharmacological analysis.
This work has demonstrated, for the first time, the existence of ascending and descending neuronal pathways and an ascending excitatory reflex to the PS of the guinea-pig. The results are consistent with a previous anatomical study using retrograde tracing and immuno-histochemistry.25 They are also consistent with previous findings from in vivo preparations in other species.10,19 ,21,22
The relatively positive resting membrane potentials of the smooth muscle cell of the PS (−41.7 mV), compared with more negative membrane potentials of smooth muscle in the stomach (about −50 mV in the gastric corpus)16 may contribute to the high resting tone observed in the sphincteric region.5 The underlying mechanisms are currently not clear. Slow waves were detected in the circular muscle cell of the outer, but not the inner layer of the PS in this study. This suggests that slow waves may decline in amplitude from the myenteric to submucosal border of the circular muscle layer, as reported previously in canine pylorus.33 The ICC near the myenteric plexus are the major source of circular muscle slow waves in the upper gastrointestinal tract.34 The results from a detailed study on the circular muscle of canine pylorus using electrical recording and morphological examination has suggested that the different sizes of PS muscle bundles between the myenteric and submucosal regions may contribute to the different cable properties for conducting slow waves.33 This may explain why slow waves were hardly detected in the inner layer of the PS muscle in the current study. Spontaneous contractions were recorded from the circular muscle in all three regions across the gastroduodenal junction; their frequency being highest in the duodenum (about 10 cycles min–1) and lowest in the PS (less than 3 cycles min–1). These frequencies are consistent with the frequency of slow waves recorded intracellularly from the PS in this study and recorded extracellularly from these regions in a previous study in the cat.20 The frequency recorded from the stomach (about 6 cycles min–1) is similar to the frequency of slow waves recorded previously from the circular muscle cells in the gastric corpus and the antrum of the guinea-pig16,35–37, and the mouse.38
Potential mediators for inhibitory transmission
Single-pulse focal electrical stimulation on the surface of the PS evoked fast IJP (>10 mV) similar to those previously reported in the circular muscle of small intestine.39 At high frequency stimulation, IJP were made up of both fast and slow components. The slow IJP and corresponding muscle relaxation induced by the PS stimulation were markedly reduced by an NOS inhibitor, indicating that NO is an important mediator of the slow IJP and muscle relaxation. The NOS immunoreactive fibres present in the PS are most likely to be the axons of the inhibitory motor neurones to the sphincter.9 A similar role for NO has been proposed in the human PS.40 The remaining fast inhibitory component was abolished by apamin. Since the early 1960s, accumulated evidence has suggested that the nonadrenergic, noncholinergic (NANC) inhibitory transmission, which is responsible for producing IJP and relaxation in gastrointestinal smooth muscle cells, is mediated by several putative neurotransmitters.41 In most preparations of different parts of the gastrointestinal tract in different species, inhibitory transmission also involves fast apamin-sensitive IJP and slow NO-mediated IJP, for example, in the guinea-pig small29,39,42, 43 and large intestine43 and human colon.44 There is evidence that a purine, but not pituitary adenylate cyclase-activating peptide, may mediate the fast apamin-sensitive IJP.45 In this study, the slow IJP are the major electrical changes for producing muscle relaxation in this preparation. However, the role of fast apamin-sensitive IJP in producing the change of mechanical activity (muscle relaxation or contraction) is still not clear. It has been suggested in a previous reflex study in the guinea-pig ileum that the fast apamin-sensitive IJP may play a role in introducing a latency in the descending contraction.46 In addition, vasoactive intestinal peptide, which is present in enteric inhibitory motor neurones,9 may also participate in inhibitory transmission.10,39 The current results are similar to findings in the guinea-pig lower oesophageal sphincter and stomach, in which both NO- and apamin-sensitive components are important inhibitory mechanisms to the circular muscle of both regions.16,47 Thus, the relative importance of mechanisms of inhibitory transmission to the muscle differs greatly between preparations. The functional significance of this remains to be determined.
Potential mediators for excitatory transmission
The initial small contraction of the PS, evoked by focal electrical stimulation, is mediated by acetylcholine acting on muscarinic receptors on the smooth muscle, as it was abolished by hyoscine. Hyoscine also reduced the second delayed large contraction and enhanced the amplitude of inhibitory responses (relaxations and IJP). This suggests that cholinergic transmission also contributed to the second large contraction associated with the EJP and spikes. This is consistent with previous reports on cholinergic transmission to the PS in other species7 and the existence of cholinergic motor neurones in the myenteric plexus of the guinea-pig PS.25 When all neurones were activated by electrical stimulation, the initial cholinergic contraction was interrupted by the inhibitory junction potential and relaxation, largely mediated by NO. An additional complication in interpreting the relative contribution of the multiple mechanisms of transmission is the possible interaction between transmitters. It has been reported that NO may modulate cholinergic excitatory transmission at a postjunctional site in smooth muscle48 and that NO may affect acetylcholine release prejunctionally.49 In addition, multiple transmission is also important in excitation of the PS because NK1 and NK2 receptor antagonists greatly reduced the excitatory responses remaining after hyoscine, suggesting the involvement of tachykinins.,11,31,32 The ChAT and substance P immunoreactive fibres present in the PS are probably be the axons of the excitatory motor neurones to the sphincter.25, 50,51
Neuronal pathways and reflexes to the PS
Our previous retrograde tracing study revealed that most motor neurones to the PS are short and are located within 3 mm of the PS. A few motor neurones with NOS immunoreactivity were found in the stomach, up to 29 mm oral to the PS, but very few motor neurones to the PS were located in the duodenum.25 In the present study, electrical stimulation of the stomach 5 mm oral to the PS evoked a large IJP followed by a small EJP. The EJP were abolished by hexamethonium, indicating that there are excitatory neuronal pathways from the stomach to the PS acting on the final motor neurones via nicotinic synapses. The EJP could be the result of antidromic activation of ascending interneurones with collaterals synapsing with the excitatory motor neurones. The IJP were greatly reduced by hexamethonium, indicating the presence of descending inhibitory pathways. The residual IJP, which were insensitive to hexamethonium, could be due to direct activation of descending inhibitory motor neurones in the stomach with long aboral projections. In addition, the existence of hexamethonium-insensitive synapses in the descending inhibitory pathways, as reported in the small intestine,12 cannot be excluded. Electrical stimulation of the duodenum, 5 mm aboral to the PS, evoked only EJP, which were abolished by hexamethonium, suggesting that there are also ascending cholinergic neurones in the duodenum, that synapse on enteric excitatory motor neurones to the PS via nicotinic synapses.
Excitatory responses induced by duodenal stretch were recorded both in the duodenum and in the PS. Duodenal stretch evoked neurogenic contractions in the stretched region, which were abolished by TTX. It is likely that these contractions correspond to the abrupt initiation of neuronal peristalsis which has been shown in the guinea-pig ileum.26 The remaining TTX-insensitive contractions in the duodenal region are presumably myogenic in origin. Duodenal stretch also evoked neurogenic depolarization and ascending reflex contractions in the PS. They were abolished by TTX, indicating that they were mediated by neuronal reflex pathways. Thus, this study demonstrates the existence of ascending excitatory reflex pathways from the duodenum to the PS, and suggests that intrinsic neuronal reflex pathways may contribute to feedback control of pyloric motor activity. In a similar in vitro study in the cat, PS reflex contraction could be elicited by duodenal acidification.20 It is likely a similar ascending excitatory reflex is present in other species. Similar reflexes activated by chemical or mechanical stimulation of the mucosa may also be present in this junctional region. The polarity of functional pathways revealed in this work and the immunohistochemical classes of enteric motor neurones appear to be similar along the entire gastrointestinal tract12,16, 18, 52 across species.7, 53 Thus, the classes of inhibitory and excitatory motor neurones supplying the guinea-pig PS are similar to the rest of the gastrointestinal tract and thus the PS is subject to comparable neuronal control.
This work has been funded by the National Health and Medical Research Council of Australia and Flinders Medical Centre Foundation. The authors wish to thank Dr V. Zagorodnyuk for his helpful comments on the manuscript.