Co-operation between neural and myogenic mechanisms in the control of distension-induced peristalsis in the mouse small intestine

Abstract

1. Myogenic and neural control of intestinal transit were investigated in a model of distension-induced peristalsis. A comparison was made between the electrical and mechanical activities and outflow of contents observed in control mice and in W/W ^v mice, which lack the interstitial cells of Cajal associated with Auerbach's plexus.

2. Distension caused a periodic appearance of increased motor activity due to stimulation of enteric nerves in both control and W/W ^v mice. Excitation was primarily delivered by cholinergic nerves, whereas periodic inhibition was mediated by neuronal nitric oxide.

3. In control mice, outflow was driven by propagating slow-wave activity and was only in the aboral direction. Outflow only occurred when slow waves carried sufficient action potentials to cause phasic intraluminal pressure increases of ≥ 1 cmH₂O through direct stimulation of the musculature or by distension-induced neurally mediated activation.

4. In W/W ^v mice, outflow was associated with propagating action potentials that occurred due to either neural stimulation or direct muscle stimulation. Action potential propagation and outflow occurred in both oral and aboral directions.

5. In summary, in both control and W/W ^v mice, distension induced periodic motor activity through stimulation of the enteric nervous system. Intraluminal contents were not moved in front of such motor activity. Rather, within such periods of activity that occurred concurrently throughout an entire segment, pulsatile outflow was directed by individual propagating slow waves with superimposed action potentials in control tissue, and by propagating action potentials in W/W ^v mice, which lack interstitial cells of Cajal.

Peristaltic activity allows intestinal contents to move anally and can occur through a variety of mechanisms. A local stimulus, such as a balloon, can evoke ascending excitation and descending inhibition, which creates peristaltic motor activity that moves the balloon. Bursts of motor activity in a certain section of the intestine can ‘move’ aborally, such as during migrating motor complexes (MMCs) (Szurszewski, 1981; Costa & Furness, 1982; Sarna, Northcott & Belbeck, 1982), which are prominent during fasting, or migrating action potential complexes (MAPCs) (Bueno, Fioramonti, Honde, Fargeas & Primi, 1985; Diamant & Scott, 1987), which are prominent postprandially. Movements of MMCs and MAPCs are directed by the enteric nervous system. Contractile activity can also propagate anally in association with propagating slow-wave activity (Szurszewski, 1981; Costa & Furness, 1982; Weisbrodt, 1987; Siegle, Buhner, Schemann, Schmid & Ehrlein, 1990; Sarna & Otterson, 1993.) The neural and myogenic control of peristaltic activity are often treated as independent phenomena and in many experimental situations either one mechanism or the other is put forward as underlying the peristaltic motor activity. In other cases, no distinction is made between the activating mechanism (which could be neural) and the mechanisms underlying the propulsive nature of the contractile activity (which could be myogenic). Renewed interest in the myogenic control of motor activity came first from the realization that myogenic pacemaker activity originates from specialized cells: the interstitial cells of Cajal (ICC; Thuneberg, 1982). Although speculation about this possibility dates back many years, recent structural and physiological evidence has linked the generation of gastrointestinal slow-wave activity with the presence of subpopulations of ICC (Ward, Burns, Torihashi & Sanders, 1994; Huizinga, Thuneberg, Kluppel, Malysz, Mikkelsen & Bernstein, 1995; Torihashi, Ward, Nishikawa, Nishi, Kobayashi & Sanders, 1995; Mikkelsen, Malysz, Huizinga & Thuneberg, 1997). Absence of ICC leads to intestinal motor abnormalities in mice (Maeda, Yamagata, Nishikawa, Yoshinaga, Kobayashi & Nishi, 1992; Huizinga, Malysz, Hagel & Ruo, 1995; Sato et al. 1996). Evidence has now also been presented that the absence of ICC may contribute to abnormalities in intestinal transit in humans, such as in infantile hypertrophic pyloric stenosis (Vanderwinden, Liu, De Laet & Vanderhaeghen, 1996) and Hirschsprung's disease (Vanderwinden, Rumessen, Liu, Descamps, De Laet & Vanderhaeghen, 1996). It is also suggested that abnormalities found in the ICC of patients with ulcerative colitis may be associated with inflammation-induced motor abnormalities (Rumessen, 1996), a hypothesis which has found support in animal studies (Der-Silaphet, Berezin, Collins & Huizinga, 1996).

Interstitial cells of Cajal are specialized smooth muscle cells. They form an electrical syncytium with the smooth muscle layers, and should, therefore, be considered to be part of the myogenic control system. In addition, although ICC were originally thought to be neural in origin (see Thuneberg, 1982), they are now known not to be derived from the neural crest but to develop independently from the enteric nervous system (Lecoin, Gabella & Le Douarin, 1996).

The fact that interstitial cells of Cajal are always very intimately associated with neural structures (Thuneberg, 1982; Berezin, Huizinga & Daniel, 1988; Faussone-Pellegrini, 1992) suggests a close co-operation between myogenic and neural control. Distension-induced peristalsis was chosen as a model to study such an interaction. Weems & Seygal (1981) developed a model in which a segment of cat ileum was subjected to intraluminal pressures from 5 to 10 cmH₂O, which led to periodic increases in motor activity in time intervals of ∼8 min. These bursts of motor activity led to outflow of the intestinal contents in an aboral direction. The bursts of motor activity were abolished by tetrodotoxin (TTX) and atropine, and were therefore thought to be due to the activity of intrinsic cholinergic neurons. These bursts of motor activity occurred concurrently across the segment under study and the organization of the motor activity during the bursts was not investigated. The present study was designed to characterize the electrical and associated motor activity during these periods of increased motor activity induced by distension. The underlying hypothesis was that within neurally evoked bursting activity, it is the propagating electrical and mechanical activity that underlies transit of contents. A comparison was made between W/W^v mutant mice and controls to evaluate the role of slow-wave activity, which is absent in W/W^v mice (Ward et al. 1994; Huizinga et al. 1995).

The general objective was to increase our understanding of the factors involved in distension-induced peristalsis through simultaneous measurement of electrical activity, intraluminal pressure changes and content outflow.

METHODS

Mice (+/+ and W/W^v from Jackson Laboratories, Bar Harbor, ME, USA) were killed by cervical dislocation. The small intestine was exposed by a mid-line abdominal incision, and a 6 cm tubular segment of intestine was taken 1 cm distal to the pyloric sphincter. The tissue was mounted in a chamber as outlined in Fig. 1 (based on Weems & Seygal, 1981) and equilibrated in 60 ml of 37-38°C Krebs solution saturated with 95 % O₂-5 % CO₂ for 1 h. Intraluminal pressure recordings were carried out through fluid-filled plastic open-ended tubes and amplified through a Grass ink writing amplifier-recorder (7 P122 D-7 PCM 12).

Figure 1. Schematic drawing of the set-ups.

A, positioning of electrodes on the proximal segment of the mouse small intestine, and positioning of pressure recording ports inside the segment. Intraluminal pressure can be varied by raising the level of Krebs solution in both the right- and left-hand columns. A pressure gradient is established by raising the level of Krebs solution in either the right- or the left-hand column. B, similar to A, except that outflow from the system is permitted. The large reservoir of buffer feeding the left-hand column can maintain a constant intraluminal pressure even when some outflow occurs. Under most experimental conditions the pressure is set such that the water level is ≈0.5 cm below the outflow level. In both set-ups, the intestinal segment and part of the pressure columns are placed in an organ bath filled with Krebs solution at 38 °C.

Suction electrodes were attached to the serosal side of the intestine. The recording and ground electrodes were silver chloride-coated silver wires of 0.06 mm diameter. The recording electrode was insulated with a flexible plastic tubing with an inner diameter of 0.2 mm and an outer diameter of 1.0 mm. Suction was provided by a mechanical pump.

Expelled fluid was collected in a narrow tube with a pressure port in the bottom of the tube. Outflow was monitored as increases in pressure and volume. This measurement was not sensitive enough to record individual flow pulses; these were monitored visually as outflow pulses (at the arrow in Fig. 1B).

Stock solutions of barium chloride (BaCl₂), atropine, tetrodotoxin (TTX), verapamil and l-NAME (all from Sigma) were prepared with Nanopure® filtered water and diluted to their final concentrations in Krebs solution.

Tissue was maintained in Krebs solution (pH 7.3-7.35) containing (mM): NaCl, 120.3; KCl, 5.9; CaCl₂, 2.5; MgCl₂, 1.2; NaHCO₃, 20.0; NaH₂PO₄, 1.2; and glucose, 11.5. Pressure was evoked by increasing the level of Krebs solution in both the left- and right-hand columns. Phenol Red was added to aid visualization of outflow. Animals were cared for in accordance with the principles and guidelines of the Canadian Council on Animal Care. Data are represented as means ±s.e.m.

RESULTS

Control mice

Correlation between pressure gradients and flow of intraluminal content

When an intraluminal pressure gradient of > 1 cmH₂O was established in a segment of proximal small intestine, i.e. Krebs solution was added to either the right- or left-hand column (Fig. 1A), flow occurred in a steady stream either aborally or orally, in the direction of lowest pressure. If intraluminal pressure was allowed to build up (by adding Krebs solution to both columns), but immediate outflow was prevented because of the height of the columns (Fig. 1A), propagating forces (described below) propelled the contents anally, building up a pressure gradient of 0.5-1 cmH₂O. The pressure did not increase to a higher level than this, and when the propulsive forces were periodic (as described below), the pressure gradient would fall to zero in periods of low motor activity in between propagating pressure waves. Hence, whether or not flow of the intraluminal contents occurred depended to a large extent on intraluminal pressure gradients.

Correlation between electrical activity, intraluminal pressure and outflow

The following experiments were carried out under conditions in which there was no pre-set intraluminal pressure gradient.

Slow-wave activity was associated with individual transient increases in intraluminal pressure on a one-to-one basis (Fig. 2). The propagation velocities of both phenomena, measured independently, were always identical. Outflow always occurred in a pulsatile manner in synchrony with a propagating slow wave (Fig. 3). However, slow-wave activity by itself did not lead to propulsive contractile activity. Outflow occurred when slow waves bore significant action potential activity associated with the development of phasic increases in intraluminal pressure of > 0.5-1.2 cmH₂O, each tissue segment studied having a specific value. If such slow wave-action potential complexes occurred sporadically, outflow occurred sporadically. If all slow waves were associated with sufficient action potentials, leading to high amplitude pressure waves, outflow pulses occurred in a pulsatile manner at the slow-wave frequency (Fig. 3B). The amplitude of intraluminal pressure waves was related to the intensity of action potential generation superimposed on slow waves. Such activity could easily be evoked by BaCl₂ (0.1-0.5 mM). In seven experiments, control activity consisted of slow waves with a frequency of 44.6 ± 3.0 cycles per minute (c.p.m.), propagating at 0.89 ± 0.40 cm s⁻¹. The amplitude of the associated pressure waves was 0.19 ± 0.10 cmH₂O. There was no outflow accompanying this activity. With the addition of BaCl₂ (0.5 mM), in the presence or absence of TTX, action potential activity increased, and the amplitude of the pressure waves increased to 1.03 ± 0.24 cmH₂O. Outflow occurred in a pulsatile manner at the unchanged slow-wave frequency, resulting in a flow rate of 0.69 ± 0.51 ml min⁻¹.

Figure 2. Propagating slow waves with superimposed action potentials were associated with propagating transient increases in intraluminal pressure.

A, electrical activity recorded simultaneously at 3 different sites, 1 cm apart (see Fig. 3A). In control mice, slow waves always occurred at identical frequencies at the 3 recording sites. In this experiment the propagation velocity was 1.0 cm s⁻¹. B, intraluminal pressure changes recorded at 3 sites inside the lumen (Fig. 1). The slow waves with superimposed action potentials were correlated 1:1 with increases in intraluminal pressure. The waves of intraluminal pressure increase propagated at 1.0 cm s⁻¹, in synchrony with the propagating slow waves.

Figure 3. Propagating intraluminal pressure waves and outflow are dependent on smooth muscle excitation, not necessarily through neural activity.

In the presence of BaCl₂ (0.5 mM) all slow waves bore action potentials, and propagating slow waves were associated with propagating pressure waves and outflow. A, simultaneous recordings of electrical activity taken at distances of 1 cm (seeFig. 1B a and b), intraluminal pressure changes underneath electrodes 1 and 3, respectively. This activity was evoked in the presence of TTX to prevent neural activity. Outflow (▪, recorded visually) occurred in a pulsatile manner related 1 : 1 to the pressure waves. Lines in A indicate propagation at a velocity of 0.5 cm s⁻¹, consistent with visually observed aboral propagation of ring contractions.

Distension-induced initiation of periodic activity of intrinsic nerves

Distending the small intestine by increasing the intraluminal pressure from 0 to 2-4 cmH₂O evoked burst-type activity. That is, in a periodic manner, action potential activity appeared on three to twenty consecutive slow waves (Figs 4, 5A, C and D, and 6A). The burst duration was 18.6 ± 13.1 s and the burst frequency was 2.5 ± 1.4 bursts min⁻¹ (n= 37). The slow-wave frequency was 45.8 ± 5.5 c.p.m. In 20 % of the experiments the bursts were evoked in response to an intraluminal pressure of ∼2 cmH₂O, but in 80 % of the experiments the pressure needed was 3 cmH₂O. The periods of occurrence of action potentials superimposed on slow waves were associated with periods of transient intraluminal pressure increases. When the intraluminal pressure was increased to 4.6 ± 0.6 cmH₂O, the burst-type activity changed to continuous action potential activity on all slow waves. Distension-evoked action potential generation was inhibited by atropine (1 μM; n= 7; Fig. 5).

Figure 4. Distension-induced periodic occurrence of action potentials superimposed on slow waves.

The intestinal segment was distended by setting the intraluminal pressure at 4 cmH₂O. A, action potentials occurred on 5-10 slow waves, followed by 2-6 slow waves without action potentials. The accompanying intraluminal pressure changes (B) were accordingly periodic. Outflow occurred with high-amplitude propagating pressure waves (▪). Lines indicate the propagation of slow waves, which occurred at a velocity of 1.0 cm s⁻¹ in this experiment, consistent with the velocity of visually observable propagating ring contractions.

Figure 5. Effect of atropine or blockade of NO synthesis on distension-induced activity.

A and B, atropine. The segment was distended by setting the intraluminal pressure to 4 cmH₂O. A, action potentials occurred on 12-15 slow waves, followed by 2-6 slow waves without action potentials. B, addition of atropine (1 μM) abolished action potential generation, and the intraluminal pressure changes became very small. Outflow stopped. C and D, L-NAME. Recordings were taken from 2 different intestinal segments, with the intraluminal pressure set at 3.5 and 4 cmH₂O, respectively. The top traces show the development of periods of increased action potential generation. In C, the periods of increased activity lasted 10-15 s, alternating with 2-10 s of relative quiescence. In D, the periods of increased action potential development lasted 40-60 s (1 period shown) alternating with ≈30 s of relative quiescence. The addition of L-NAME immediately abolished the quiescent periods and continuous action potential generation developed.

Figure 6. Distension-induced neural activity abolished by TTX.

The intestinal segment was distended by setting the intraluminal pressure to 4 cmH₂O. A, action potentials occurred on 3-5 slow waves, alternating with 3-5 slow waves without action potential activity. The propagation velocity was 0.7 cm s⁻¹. Outflow (▪) occurred only when slow waves had sufficient action potential activity. B, addition of TTX (0.3 μM) caused continuous action potential activity on all slow waves, apparently abolishing intrinsic periodic inhibitory activity. The propagation velocity was 0.7 cm s⁻¹. Under these conditions, outflow occurred with every propagating slow wave. C, enlargement of the boxed period of increased activity in A, with associated pressure changes and outflow. D, enlargement of the boxed area in B, with associated pressure changes and outflow.

The association between periods of increased motor activity and outflow was determined with the set-up shown in Fig. 1B. The periods of increased activity occurred at 1.9 ± 1.6 bursts min⁻¹, for a duration of 26.8 ± 20.2 s (n= 9). Within the periods of increased activity, slow-wave activity occurred at 48.4 ± 3.1 c.p.m. at an apparent propagation velocity of 0.87 ± 0.30 cm s⁻¹. It was not possible to determine propagation characteristics of bursts; they appeared to occur simultaneously throughout the segment. The exact time at which a period of increased activity started was, however, difficult to determine because of poor action potential development in the beginning of the period. The amplitude of the phasic pressure increases associated with the slow waves was 0.76 ± 0.34 cmH₂O in relatively electrically quiescent periods, and 4.8 ± 2.2 cmH₂O in periods of increased action potential activity. Outflow occurred only with higher-amplitude pressure waves and was pulsatile at the slow-wave frequency.

The periodicity of increased activity was abolished by TTX (0.3 μM). In the presence of TTX either (a) slow-wave activity occurred without spiking activity and there was no outflow (three out of four experiments), or (b) continuous action potential generation was observed superimposed on slow waves, with associated intraluminal pressure increases and outflow (Fig. 6B).

The periods of relative quiescence between bursts were abolished by L-NAME, which blocks nitric oxide (NO) synthesis. In the presence of L-NAME, continuous action potential generation was observed superimposed on slow waves (Fig. 5C and D). Outflow occurred in four out of five experiments, associated with propagating slow waves, and pressure waves of 0.89 ± 0.24 cmH₂O. The slow-wave frequency was 51 ± 6.9 c.p.m. In three experiments in which distension induced periodic activity at 1.3 ± 0.6 bursts min⁻¹, addition of L-NAME plus L-arginine prevented any change in activity. Atropine inhibited the generation of action potentials and reduced the amplitude of pressure waves to 0.52 ± 0.18 cmH₂O, and outflow stopped (n= 5) (Fig. 5B). Outflow could be restored by the addition of BaCl₂. BaCl₂ (0.5 mM) induced irregular spiking activity and outflow occurred when associated pressure waves reached > 0.6 cmH₂O.

W/W^v mutant mice

Correlation between pressure gradients and flow of intraluminal content

In W/W^v mice, without pressure- or drug-mediated stimulation, action potential activity occurred in an irregular manner, was not synchronized across the intestinal segment (comparison of three simultaneous recordings, 1 cm apart) and no outflow occurred (Fig. 7A). With stimulation (as described below), when intraluminal pressure was allowed to build up (through addition of Krebs solution to both columns; see Fig. 1A), but outflow was prevented (because of the height of the columns), action potentials propelled the contents orally or anally, building up a pressure gradient of 0.8-1.2 cmH₂O (n= 7). The pressure gradient fell back to zero in between action potentials.

Figure 7. Distension-induced periodic activity in the W/W ^v mouse.

A and B, top 3 traces: electrical activity recorded simultaneously from 3 different sites, 1 cm apart. Bottom 2 traces: intraluminal pressure changes recorded simultaneously at intraluminal sites of electrodes 1 and 3. A, without sufficient distension, the occurrence of action potentials at the 3 sites was unco-ordinated. Pressure changes were also unco-ordinated and no outflow occurred. B, distension by 3.5 cmH₂O of intraluminal pressure evoked periodic activity. Periods of action potential generation (≈10 s) alternated with periods of electrical quiescence. The periods of activity occurred simultaneously at the different recording sites. Individual action potentials, but not individual bursts, were related 1 : 1 to increases in intraluminal pressure and outflow pulses (▪). Hence, the outflow was pulsatile in nature, with each transient outflow associated with an individual propagating pressure wave of sufficient amplitude. C, enlargement of one of the periods of action potential generation, indicating the propagation of individual action potentials. The propagation velocity was 1.0 cm s⁻¹.

Distension-induced initiation of the periodic activity of intrinsic nerves in W/W ^v mice

The following experiments were carried out under conditions without a pre-set intraluminal pressure gradient.

Distending the small intestine by increasing the intraluminal pressure from 0 to 4 cmH₂O (by addition of Krebs solution to both columns) evoked burst-type activity. That is, action potentials appeared in bursts, alternating with periods of electrical quiescence (Figs 7A and 8A). Action potential bursts were associated with bursts of phasic intraluminal pressure increases, whereby each individual pressure increase was correlated one-to-one with the occurrence of an action potential. The bursts occurred at 2.5 ± 2.1 bursts min⁻¹, and the burst duration was 23.2 ± 16.4 min (n= 8).

Figure 8. Neurally mediated bursting in the W/W ^v mouse.

A-C: top traces, action potential generation; bottom traces, intraluminal pressure changes. A, distension to 4 cmH₂O induced bursting activity in both action potential generation and intraluminal pressure changes. Bursting activity was associated with pulsatile outflow with the highest pressure changes. B, addition of L-NAME (0.1 mM) induced continuous generation of action potentials without periods of electrical quiescence. In this particular experiment the amplitude of the pressure wave appeared just insufficient to cause outflow. C, addition of atropine (10 μM) strongly inhibited the generation of action potentials. No outflow occurred.

The association between burst-type activity and outflow was studied (Fig. 7) using the set-up described in Fig. 1B. Bursts occurred at 2.3 ± 1.5 bursts min⁻¹, and the burst duration was 27.1 ± 17.1 s (n= 9). Within the bursts, the action potential frequency was 41.1 ± 9.5 c.p.m. When the propagation of the action potentials within the bursts was in the aboral direction, the propagation velocity was 0.58 ± 0.21 cm s⁻¹. Such bursts started to appear at the oral end of the segment, and the front of the burst appeared at the distal end of the segment about 3 s later, so that a burst propagation velocity of 1.0 ± 0.43 cm s⁻¹ could be calculated. Since the burst duration lasted 5-50 s, for most of the time burst activity appeared simultaneously in all three recording sites, so that action potentials within the bursts propagated over the entire segment. The amplitude of the phasic pressure increases associated with these action potentials was 0.8 ± 0.5 cmH₂O in between the bursts and 5.7 ± 2.2 cmH₂O during the bursts. Outflow occurred only during bursting and was pulsatile at the action potential frequency. The outflow amounted to 0.36 ± 0.11 ml min⁻¹. No outflow occurred prior to the development of bursting activity.

In the presence of L-NAME (0.1 mM), burst-type activity disappeared and action potential activity became continuous (Fig. 8B). The action potential frequency was 45.8 ± 4.5 c.p.m. and the amplitude of pressure waves was 4.8 ± 1.1 cmH₂O. In three experiments in which distension induced burst-type activity at 2.0 ± 1.0 bursts min⁻¹, L-NAME was added in the presence of L-arginine (10 mm), which prevented any effect of L-NAME. A similar effect on burst-type activity was seen with TTX (0.3 μM; n= 6), with action potential activity becoming continuous. The action potential frequency was 39.1 ± 12.0 c.p.m. and outflow occurred with pressure waves of 3.9 ± 4.3 cmH₂O (four out of six experiments). However, when pressure waves were 1.5 ± 0.7 cmH₂O, no outflow occurred (two out of six experiments). The presence of atropine abolished most spiking activity and no outflow occurred (n= 6; Fig. 8C).

Occurrence of flow of contents in control and W/W ^v mice when segments of small intestine are reversed in the organ bath

Segments of intestine were reversed in the organ bath in the set-up described in Fig. 1B, with the oral end now facing the outflow port. The intraluminal pressure was set at 3.5 cmH₂O without creating a pressure gradient.

In control mice, slow waves with superimposed action potentials, in the presence of distension and/or BaCl₂, always propagated in an aboral direction. The slow-wave frequency was 44.5 ± 1.5 c.p.m. (n= 5). No outflow occurred at the oral end of the intestinal segment. The water column in the right-hand chamber (at the oral end of the tissue) dropped 1.1 ± 0.2 cmH₂O.

In contrast, in W/W ^v mice there was significant periodic oral outflow at times (n= 5). When action potentials propagated in the oral direction (Fig. 9), outflow occurred in the oral direction. In the presence of distension and TTX, electrical activity periodically became very regular and similar in frequency in all segments studied. This indicates that the musculature can be directly activated by distension without mediation of enteric nerves. Activity in the presence of TTX or atropine was higher in WW^v mice than in control mice, possibly because the musculature itself is depolarized and, hence, more excitable (Malysz, Thuneberg, Mikkelsen & Huizinga, 1996). With identical frequencies, the action potentials were synchronized and apparent propagation was observed (Fig. 9B and C). Outflow occurred when the action potentials propagated orally. When the action potential frequency changed in any of the recording sites, propagation and outflow both stopped. When action potential activity became synchronized again, outflow resumed (Fig. 9D).

Figure 9.

Reverse peristalsis in the W/W ^v mouseThis experiment was carried out on a reversed segment in the outflow apparatus. The aboral end of the segment was hooked up to the pressure reservoir (Fig. 1B and text) and outflow was allowed from the oral end. A, for most of the time, no regular propagation of action potentials across the whole segment could be identified, and no ring contraction was seen to propagate along the segment of intestine. No outflow occurred. B, the addition of atropine decreased action potential generation and revealed bursting activity. This shows that in the W/W ^v mouse small intestine, which is much more easily excited than control tissue, burst-type activity can occur without activity of cholinergic nerves. The periods of quiescence were abolished by L-NAME (not shown). Individual action potentials were seen to propagate aborally, and pressure decreased periodically in the right-hand column at the action potential frequency (Fig. 9B). The propagation velocity was 1.7 cm s⁻¹. C, the addition of TTX abolished the quiescent periods, and action potentials were generated continuously. However, only the action potentials at two sites were co-ordinated and no outflow occurred. D, a few minutes later, activity underneath all three electrodes became spontaneously co-ordinated, and outflow occurred with each propagating action potential and pressure wave at 1.2 cm s⁻¹. Ring contractions of the segment were seen to propagate at the same velocity.

DISCUSSION

The main determinants of flow in the proximal small intestine from control mice, as measured in an isolated segment under varying intraluminal pressures, were the existence of a pressure gradient or the occurrence of propagating slow waves with superimposed action potentials associated with relatively high intraluminal pressure waves. The latter was dependent on a stimulus delivered to the musculature pharmacologically or by distension, or through distension-induced activity of enteric nerves.

The existence of pressure gradients had a dominating influence on transit. A high pressure gradient allowed flow in a non-pulsatile manner that was independent of the characteristics or the degree of electrical activity. Also, when propagating electrical and mechanical activities were generated, this did not lead to flow of intraluminal contents if the flow was against a significant pressure (> 1 cmH₂O under our experimental conditions). Hence, strong propagating activity may not be associated with any net movement if it has to work against significant pressure. Backflux easily occurred when intraluminal pressure was higher at the aboral site.

Without a positive pressure gradient, when work must be generated to obtain outflow (see also Weems & Seygal, 1981), we observed that outflow was always associated with propagating slow waves with superimposed action potentials that are associated with intraluminal pressure waves of > 2-6 cmH₂O. The threshold for outflow was variable between segments, but had a specific value for each segment. A variety of stimuli can cause peristaltic propulsive motor activity. When significant action potential generation occurred on every slow wave, outflow occurred in a pulsatile manner associated with every propagating slow wave; from a volume of ∼0.36 ml in a 4 cm section of intestine, ∼0.07 ml flowed at the slow-wave frequency with each slow wave. Propagating intraluminal pressure waves are associated with slow waves possessing superimposed action potentials in a 1:1 ratio (this study; Grivel & Ruckebusch, 1972). In many studies, contractile activity is recorded without sufficient resolution and/or electrical activity is not recorded, and therefore the relationship between the two is difficult to appreciate. When transient pressure changes summate, which can occur when recording high amplitude pressures, pressure changes become broad and a relationship to slow waves may not be obvious.

When the outflowing fluid is allowed to rise in a vertical column (Fig. 1A), the pressure will increase with every propagating slow wave to ∼1 cmH₂O, but this pressure will fall back in the period between two propagating slow waves. Hence, the propagating slow waves will not build up a significant pressure and net outflow will not occur despite the fact that the motor pattern is peristaltic. When intraluminal pressure was allowed to rise to between 2 and 4 cmH₂O, a bursting pattern of action potential generation was elicited that was sensitive to TTX. Action potentials occurred on several sequential slow waves, alternating with slow waves without action potentials. This pattern of electrical activity was associated with bursts of phasic increases in intraluminal pressure. It is difficult to assess propagation of these bursts of activity in a relatively short segment and, for most of the time, the bursts occurred simultaneously at all three recording sites. During burst activity, slow waves with superimposed spikes propagated across the segment and caused outflow of fluid in a pulsatile manner, synchronous with the propagating slow waves. The direction of propagation was always aboral. When the tissue was mounted in the opposite direction, i.e. the aboral end fixed to the pressure column and the oral end free to allow outflow, no outflow at the oral end occurred. In fact, the pressure dropped in the oral column, identifying flow in the aboral direction, and a negative pressure gradient developed (although this was no more than ∼1 cmH₂O).

Under the conditions of the current set of experiments, the increase in intraluminal pressure activated cholinergic nerves to excite the musculature, with generation of action potentials as a consequence. Inhibitory innervation, mediated by NO, appeared essential for the periodic burst-type appearance of the action potentials, since blockade of NO synthesis caused the periods of electrical and mechanical quiescence to disappear. This was also observed in the mouse colon (Bywater, Small & Taylor, 1989; see also Waterman & Costa, 1994). The fact that this happened in both normal and W/W ^v mice suggests that pressure-induced activation of the enteric nervous system was normal in W/W ^v mice. As pointed out by Weems & Seygal (1981) and others (Wood, 1989; Costa & Brookes, 1994), general distension triggers programmed patterns of neural activity within the enteric nervous system. However, it is often assumed that this neural activity is the only determinant of propulsive contractile activity. In fact, there is little evidence for this (Waterman, Tonini & Costa, 1994), and little is known about the exact neural circuitry activated during general distension. It is noteworthy that the bursts of activity recorded at different sites of the intestine appear simultaneously, suggesting that during general distension, ‘neurohumoral conditions are optimal for spike potentials to occur over a length of the bowel’ (Weisbrodt, 1987). The circumstances of the activities evoked in the current set of experiments probably led to massive synchronous activation of motor neurons along a relatively long segment of intestine, as is predicted to occur in distension-induced peristalsis of the guinea-pig ileum (Tonini, Costa, Brookes & Humphreys, 1996). The latter study suggested that distension may not induce simple reflexes but a ‘complex motor pattern’. We believe we have described some characteristics of that motor pattern. The cellular basis of the neural mechanisms underlying the triggering of these motor patterns remains to be investigated. Obviously, this is just one of several patterns of neural activity that can be evoked (Costa & Furness, 1982; Costa & Brookes, 1994; Waterman et al. 1994).

In the presence of TTX, peristalsis can occur by direct muscle stimulation through distension, or by direct pharmacological muscle excitation. Evoking action potentials superimposed on the omnipresent propagating slow waves causes propulsion when the associated intraluminal pressure changes exceed threshold. In the presence of L-NAME, the dominant inhibitory innervation is abolished and neural excitation is dominant. In our experiments, the ensuing activity was not periodic but continuous, and outflow occurred when pressure waves were of sufficient magnitude.

In W/W ^v mice, slow-wave activity is not present (Ward et al. 1994; Huizinga et al. 1995). When intraluminal pressure was increased without additional stimulation in the presence of TTX, no outflow occurred and the action potentials occurring at different sites appeared independent, or at least were not sufficiently synchronized. With additional stimulation, such as the action of BaCl₂ on the muscle cells, action potentials increased in frequency and became similar at the different recording sites. Action potentials could then become synchronized (Daniel, Bardakjian, Huizinga & Diamant, 1994), which was clearly observed in the present experiments. Outflow occurred in association with every propagating action potential and propagating pressure wave. The electrical activity observed in W/W ^v mice can appear similar to slow-wave activity. However, the action potentials are easily distinguished pharmacologically. Action potentials are abolished by L-type calcium channel blockers (Ward & Sanders, 1992; Liu & Huizinga, 1993) and hyperpolarization (Malysz, Der-Silaphet, Lee, Thomsen, Das & Huizinga, 1996), whereas slow waves are relatively insensitive to these treatments.

In W/W ^v mice, distension induces periodic occurrence of action potentials associated with intraluminal pressure changes. There is a one-to-one relationship between action potentials and intraluminal pressure increases. The bursts of activity usually occurred simultaneously in the whole segment under study. Outflow occurred only during bursting activity and in a pulsatile manner, in synchrony with the occurrence of action potentials. It is clear from these experiments that the W/W ^v mouse small intestine has the mechanisms to produce propulsive contractile activity without the benefit of slow-wave activity. The occurrence of propagation, however, is not as regular or as predictable. An example of this is the activity in a reversed segment, with outflow allowed from the oral end. Whilst in the control intestine no outflow occurs (except in response to a pressure gradient), in the W/W ^v mouse outflow is regularly observed. Outflow occurs when the musculature is stimulated sufficiently either by distension-induced activation of nerves or by direct muscle stimulation. When action potentials obtain a similar frequency, and synchronization and propagation in the oral direction occur, flow in the oral direction also occurs. In the W/W ^v small intestine there is apparently no mechanism to restrict the propagation to an aboral direction, a restriction normally provided by the slow waves. The pattern of propagating action potentials was immediately disrupted whenever the action potential frequency at one of the sites changed, which occurred frequently. Such irregular occurrence of propagation does not take place in the presence of slow waves, since their frequency and propagation velocity are very constant. The present study provides additional evidence that the enteric neural circuitry is largely unaffected in W/W ^v mice (Maeda et al. 1992; Ward et al. 1994). It also provides the mechanism of peristalsis in the W/W ^v mouse. Despite the fact that peristalsis is possible in W/W ^v mice, it may not be as effective as in control mice (Huizinga et al. 1995), since the intestinal transit in W/W ^v mice is significantly impaired (Kamiya, Oku, Itayama & Ohbayashi, 1985).

The existence of multiple control levels for gastrointestinal motility has always been recognized. In 1899, Bayliss & Starling noticed myogenic waves of contractions with associated transit of content, in addition to their discovery of the law of the intestine, describing neural control (Bayliss & Starling, 1899). The characteristics of myogenic control have been further elucidated by many others, notably Diamant & Bortoff (1969) and Szurszewski, Elveback & Code (1970). A synthesis of myogenic and neural control mechanisms was given by Costa & Furness (1982) and Weisbrodt (1987). Despite this knowledge, the myogenic component in peristaltic activity is often overlooked or not investigated. This sometimes occurs because of lack of resolution in electrical and mechanical recordings (Weems & Weisbrodt, 1986). In addition, activity, when studied under many experimental conditions, is often abolished by TTX or atropine, suggesting a neural origin. However, the blocking of motor activity by TTX may only indicate that an essential stimulus was provided by the nervous system. It does not necessarily mean that the only operational control system for peristalsis was provided by enteric nerves. The present study illustrates an example of co-operation between the neural and myogenic control of intestinal motility to ensure proper transit.