Simultaneous intracellular recordings from longitudinal and circular muscle during the peristaltic reflex in guinea-pig distal colon

Authors


Corresponding author T. K. Smith: Department of Physiology and Cell Biology, University of Nevada School of Medicine, Reno, NV 89557, USA. Email: tks@physio.unr.edu

Abstract

  • 1Simultaneous intracellular recordings were made from longitudinal muscle (LM) and circular muscle (CM) cells of guinea-pig distal colon during the peristaltic reflex.
  • 2Spontaneous rhythmical depolarizations with superimposed action potentials (mean amplitude: 19 ± 2 mV) were regularly recorded from the LM (mean interval: 7 ± 1 s). In contrast, in the CM layer, spontaneous action potentials occurred with an irregular frequency. Although spontaneous action potentials in LM were rarely correlated in time with those in CM, spontaneous inhibitory junction potentials (sIJPs) were found to occur synchronously in both muscles (5 out of 27 animals; 19 %).
  • 3Graded inflation of an intra-luminal balloon or mucosal stimulation oral to the recording electrodes elicited gradeable compound IJPs synchronously in both LM (mean amplitude: 6 ± 1 mV) and CM (mean amplitude: 9 ± 1 mV) (descending inhibitory reflex). Evoked IJPs were often followed by action potentials in both muscle layers.
  • 4Mucosal stimuli applied anal to the recording electrodes elicited compound excitatory junction potentials (EJPs) synchronously in both muscles layers that were often associated with the generation of action potentials. In the LM, evoked EJP amplitudes ranged from 3 mV (subthreshold) to 31 mV (including the action potential) and in the CM from 4 mV (subthreshold) to 44 mV (including the action potential).
  • 5Apamin (500 nM) reduced the evoked IJP in the CM by 55 % (from 11 ± 2 to 5 ± 1 mV), but caused no significant reduction in the LM layer (from 8 ± 1 to 6 ± 1 mV). Apamin-resistant IJPs in both muscle layers were likely to be due to nitric oxide, since they were abolished by L-NA (100 μM).
  • 6Atropine (1 μM) abolished the ascending excitatory reflex in both muscles.
  • 7Injection of neurobiotin into the LM and CM confirmed that simultaneous intracellular recordings were made from different muscle layers.
  • 8In conclusion, during the peristaltic reflex, the LM and CM layers receive synchronous inhibitory neuromuscular inputs during descending inhibition and synchronous excitatory neuromuscular inputs during ascending excitation. No evidence was found to support reciprocal innervation.

Bayliss & Starling (1899, 1900) first described the peristaltic reflex in the small and large intestine. In response to local distension or during propulsion of a bolus, this reflex consisted of a synchronous contraction of both the longitudinal muscle (LM) and circular muscle (CM) layers oral to, and a synchronous relaxation of both muscles anal to, the point of stimulation. In subsequent years, however, two opposing schools of thought arose, regarding the movements of the two muscle layers in the intestine. One school supported the idea of Bayliss & Starling that the LM and CM contract synchronously and relax synchronously during propulsive movements (Cannon, 1912; Leusdan & Riesser, 1927; Bozler, 1949); while the other considered that the two muscles move out of phase with one another during propulsion (Henderson, 1928; Raiford & Mulinos, 1934; Alvarez, 1940; Kosterlitz & Lees, 1964). Other investigators have noted that in the human ileum, for example, the longitudinal and circular coats sometimes contracted together, while at times the longitudinal muscle contracted first (Forster & Hertzman, 1938). Also, Alvarez (1940) noted in rabbit intestine that ‘…although the two muscle layers often contracted together, they also contract independently’.

Later, Kottegoda (1969, 1970) formulated the now widely accepted idea that the muscles of the intestine have to move out of phase with one another during propulsion because they are reciprocally innervated. He wrote, ‘…there exists an arrangement of nerves which ensures that the two muscle coats of the intestine do not contact simultaneously but are activated reciprocally so that when one muscle layer contracts the other relaxes…’ (Kottegoda, 1969). The models of intestinal peristalsis put forward by Wood (1987; see Fig. 31) also support Kottegoda's view that the two muscles are reciprocally innervated. Wood (1999) has argued that the LM and CM layers cannot contract or relax together because ‘…the laws of geometry dictate that they are antagonist muscles (i.e., shortening of one opposes shortening of the other).’ Under certain circumstances, passive mechanical interactions can occur between the muscle layers, which can sometimes generate movements that appear to be reciprocal in nature (see Gregory & Bentley, 1968; Wood & Perkins, 1970; Hennig et al. 1999). However, when dissection techniques are employed to minimize these mechanical interactions, it has been demonstrated that both the LM and CM of the small (Spencer et al. 1999b) and large intestine (Smith & Robertson, 1998; Smith & McCarron, 1998; Spencer et al. 1999a) contract together during the ascending excitatory reflex and relax together during the descending inhibitory phase of the peristaltic reflex, as well as during extrinsic nerve stimulation (McKirdy, 1972; Spencer et al. 1999a), in agreement with the original work of Bayliss & Starling (1899).

The issue of reciprocal innervation has long been considered a subject of controversy, and a weakness in our understanding of the subject arises from an absence of simultaneous recordings of the neuromuscular inputs to the two muscle layers. Although a number of studies have recorded the intracellular electrical activity in the CM during reflex stimulation of the small intestine (Hirst & McKirdy, 1974; Smith et al. 1991; Brookes et al. 1999), the electrical activity in the LM during the peristaltic reflex is poorly understood. In fact, no studies that we are aware of have ever recorded the intracellular electrical activity of the LM during the peristaltic reflex in the colon. As a result, the relative similarities or differences in neuromuscular transmission to both the LM and CM during the peristaltic reflex are unclear.

In the current study, we have used simultaneous intracellular recording techniques from the LM and CM, to compare the nature of neuromuscular transmission to the two muscle layers of the colon during mucosal stimulation and balloon distension. Our major finding is that during the peristaltic reflex, the LM and CM receive simultaneous inhibitory neuromuscular inputs anal to, and simultaneous excitatory neuromuscular inputs oral to, the site of stimulation.

METHODS

Guinea-pigs weighing 200-350 g were killed by CO2 inhalation overdose, in accordance with the animal ethics committee of the University of Nevada School of Medicine. The abdominal cavity was opened and the terminal 10-15 cm of distal colon was removed, the mesenteric attachment trimmed away, the lumen flushed clean with Krebs solution, and the colon placed immediately into a modified Krebs solution (composition below).

Dissection procedure to identify longitudinal muscle from circular muscle cells

The distal colon was opened along the mesenteric attachment and the terminal distal region was pinned to the base of a Sylgard-lined (Dow Corning Corp., Midland, MI, USA) Petri dish, so that the mucosal surface faced uppermost. The mucosa and submucosa were then delicately removed from this opened region to expose the underlying CM. Strips of CM were then carefully removed from the colon, to expose the underlying myenteric plexus and LM layer (see Fig. 1). Therefore, these preparations included an island of LM and associated myenteric plexus that remained in neural continuity with the enteric plexuses of the remaining preparation of colon. This dissection procedure enabled us to clearly visualize the myenteric ganglia and underlying LM in one region, while also identifying the thicker CM in the same field of view. Therefore, the microelectrodes could readily be positioned so as to record from CM cells and at the same time the neighbouring underlying LM cells (see Fig. 1). In all experiments, preparations were pinned serosal side down, in a recording chamber whose base consisted of a microscope coverslip that was lightly coated with a fine layer of Sylgard silicon. Unambiguous identification of the LM layer from the CM layer was aided by the use of an inverted microscope (Olympus, CK2; Napa, CA, USA).

Figure 1.

Preparation used for simultaneous intracellular recording from longitudinal muscle (LM) and circular muscle (CM) during reflex activation of the colon

A, photograph of the preparation used to simultaneously record from LM and CM layers. Left side of photo shows the thicker undissected CM and the microelectrode (1) positioned above the CM cells. Right side of photo shows the myenteric plexus upon the underlying thinner LM following removal of a strip of CM. The position of the microelectrode (2) is located directly above the LM. B, a schematic representation of the colonic preparation. An intraluminal balloon was used for colonic distensions. Distension volume was monitored with a sliding potentiometer. A fine artists’ paint brush was also used for mucosal stimulation, applied oral or anal to the microelectrodes.

Confirmation that intracellular recordings were made from the LM and CM

There are three observations which confirmed that our independent microelectrode impalements were made from the LM and CM. Firstly, with our dissection techniques we could easily visualize the location of the microelectrodes in each muscle (Fig. 1A). Secondly, spontaneous depolarizations occurred only in the LM and not in the CM. Also, spontaneous action potentials occurred independently in the two muscles (see Results). Thirdly, perhaps most convincingly, injection of neurobiotin from the recording electrodes into cells in each muscle layer revealed that fluorescently labelled muscle cells were always at 90 deg to each other.

Electrical recording technique

Intracellular recordings were made from the LM and CM simultaneously, using independently mounted microelectrodes, whose fine positioning could be adjusted using two micromanipulators (model M3301L; World Precision Instruments, Sarasota, FL, USA). Electrodes were filled with 1.5 m KCl solution and had resistances of about 100 MΩ. Electrical signals were amplified using a dual input high impedance amplifier (Axoprobe 1A; Axon Instruments, Foster City, CA, USA), using two Axon HS-2 headstages (gain 0.1L). Output signals from the amplifier were digitized on an A/D converter, and filtering frequencies ranging from 0.66 to 1.5 kHz were used. Recordings were simultaneously visualized and recorded onto a PC running Axoscope (version 8.0; Axon Instruments) and also onto a digital four-channel oscilloscope (Gould 1604, Ilford, UK).

Since atropine or an L-type Ca2+ channel antagonist was not used during simultaneous recordings from the LM and CM, a major obstacle of the impalements in this study was muscle movement. We largely overcame this by extensively pinning the preparation using approximately 60-100 micro-pins (Ø= 25 μm) obtained from platinum-iridium wire and locally isolating a small region of the LM and CM for electrode placements. Since tissue was always contracting spontaneously the duration of impalements into both muscle layers was always short. Nevertheless, stimuli were only applied to the bowel once stable impalements had been established.

Resting membrane potentials (RMPs) of the LM and CM were obtained by the difference between the recorded intracellular value of RMP and the stable potential obtained after removing the electrode from the cell.

Immunohistochemistry

To confirm that intracellular recordings were made from different muscle layers in the colon, neurobiotin (2 % w/v) and 0.5 m KCl were injected into cells of either muscle layer. Preparations injected with neurobiotin were fixed overnight at 3-4 °C in 4 % w/v paraformaldehyde in 0.1 m phosphate buffer. The next morning, tissues were rinsed 4-6 times for approximately 20 min in 0.1 m phosphate buffered saline (PBS). Tissues were then incubated for 46 h at 3-4 °C with streptavidin-FITC (Vector Laboratories, Burlingame, CA, USA) that had been diluted 1:200 in PBS. Preparations were then rinsed 4-6 times for 20 min each wash again in PBS, then mounted in buffered glycerol, coverslipped and examined with a BioRad MRC 600 confocal microscope (Hercules, CA, USA). An excitation wavelength was used appropriate for FITC (488 nm). Confocal micrographs are digital composites of Z-series scans of 7-20 optical sections through a depth of 0.56-17 μm. Final images were constructed with BioRad Comos software.

Protocol for stimulation of mucosal and balloon distension reflexes

Oral and anal reflexes in the LM and CM were elicited by brush stroking the mucosa using one to five brush strokes, delivered via a fine artists’ paint brush (see Smith & Furness, 1988; Smith & McCarron, 1998; Spencer et al. 1999b). Physiological distension stimuli to mimic fecal pellets in the bowel were delivered via an embolectomy catheter using distension volumes ranging from 100 to 500 μl. To monitor when distension stimuli were delivered to the preparation, a potentiometer was used, which signalled the exact timing of the stimulus onset. During balloon distension and mucosal stimulation, reflexes were recorded in the muscle layers approximately 5-7 cm oral or anal to the stimulus. During double impalements, the two microelectrodes were positioned within 200 μm of each other in the longitudinal axis.

Drugs and solutions

The following drugs were used throughout the current study: atropine sulphate, apamin and Nω-nitro-l-arginine (l-NA) (all from Sigma Chemical Co., St Louis, MO, USA). The composition of the modified Krebs solution was (mm): NaCl, 120.4; KCl, 5.9; NaHCO3, 15.5; NaH2PO4, 1.2; MgSO4, 1.2; CaCl2, 2.5; and glucose, 11.5. The Krebs solution was gassed continuously with a mixture containing 3 % CO2-97 % O2 (v/v), pH 7.3-7.4.

Measurements and statistics

Students’ paired t tests were used where appropriate. A minimum significance level of P < 0.05 was used throughout. Data are presented as means ± standard error of the mean (s.e.m.); n is the number of animals on which observations were made. Measurements of amplitude and half-width and time to peak response were made using Axoscope 8.0 (Axon Instruments). The propagation velocity of action potentials was ascertained by dividing the distance the spike travelled by the time taken to travel this distance (measured on the rising phase of the spike at half-amplitude). Since spontaneous depolarizations and evoked EJPs usually generated smooth muscle action potentials, the amplitudes of these events were measured from the 10 % peak amplitude point (on the rising phase), to the peak of the potential change. That is, measurements included the action potential, as it was not possible to discern the amplitude of the EJPs or spontaneous depolarizations from their associated action potentials.

RESULTS

Spontaneous electrical activity in longitudinal and circular muscle layers

Simultaneous recordings were made from 48 pairs of longitudinal muscle (LM) and circular muscle (CM) cells (n= 27 animals). Overall, there was no significant difference in the resting membrane potentials between the LM (-35 ± 1 mV, range: −29 to −54 mV) and CM layers (-36 ± 2 mV, range: −25 to −58 mV; P > 0.05; 40 cells; n= 24).

During these recordings, a variety of different electrical events were noted. In all muscle cells from both muscle layers, membrane potentials were highly unstable, revealing spontaneous action potentials in both LM and CM, spontaneous depolarizations in LM, and spontaneous IJPs in both muscles.

Spontaneous depolarizations and action potentials in longitudinal muscle

In 23 out of 48 LM cell impalements, from 12 out of 27 animals tested, rhythmical spontaneous depolarizations (SDs) were recorded, which were usually associated with the generation of one or two action potentials in their plateau phase (Fig. 2 and Fig. 3). Spontaneous depolarizations in the LM had a mean peak amplitude of 19 ± 2 mV (23 cells, n= 12 animals), a mean half-duration of 400 ± 80 ms (range: 32-1800 ms) (23 cells, n= 12 animals), and a mean interval between SDs of 7 ± 1 s (range: 0.8-20 s) (measured from half-amplitude on the rising phase). Although SDs and action potentials were regularly recorded in the LM layer, clearly defined rhythmical SDs were not observed in the CM, which generally fired fewer ongoing spikes compared to the LM layer (Fig. 3). In fact, it was not uncommon to record quiescent cells in the CM. We found no evidence of electrotonic spread of SDs in LM to the neighbouring CM layer. Also, action potentials in one muscle layer usually discharged asynchronously with action potentials in the other layer, further suggesting the absence of electrotonic coupling under these conditions (Fig. 2).

Figure 2.

Spontaneous intracellular electrical activity recorded simultaneously from the LM and CM

In the LM, spontaneous rhythmic depolarizations are shown. In the CM, there is an irregular discharge of spontaneous action potentials. Note the absence of electrotonic conduction of action potentials, since activity in one muscle layer does not propagate into the neighbouring muscle layer.

Figure 3.

Spontaneous electrical activity when recorded simultaneously from the LM and CM layers

A, in the LM, spontaneous depolarizations occur with superimposed action potentials. In contrast, in the neighbouring CM, spontaneous IJPs occur, with no electrotonic conduction of action potentials from the LM to the CM. A single spontaneous action potential is shown in the CM. B, in another animal, spontaneous IJPs occur synchronously in both the LM and CM layers.

Spontaneous junction potentials in longitudinal and circular muscle

It was possible to record spontaneous junction potentials in the CM, while there were SDs and action potentials in the LM (Fig. 3A). It is particularly noteworthy that in 10 out of 48 pairs of recordings from the LM and CM (from 5 out of 27 animals tested), there were periods during which inhibitory junction potentials occurred synchronously in both muscle layers (sIJPs; up to 20 mV in amplitude) (Fig. 3B). In addition to sIJPs, spontaneous EJPs were also recorded in CM in 5 of 27 animals, which unlike SDs, were abolished by atropine. These occurred as transient depolarizations ranging up to 25 mV in amplitude and often elicited a single action potential on their plateau phase. Spontaneous EJPs may have also occurred in the LM, but these were more difficult to identify due to the domination of SDs and spike potentials in this layer (see below). A major identifying feature that impalements were made from different muscle layers was the observation that action potentials in one muscle layer always occurred out of phase with those in the neighbouring muscle layer.

Responses from longitudinal and circular muscles during the descending inhibitory reflex

We made simultaneous intracellular recordings from the LM and CM during balloon distension of the colon applied oral to the microelectrodes. When stable recordings were obtained from both muscle layers, it was found that distension evoked pronounced IJPs synchronously in both muscles anal to the stimulus (n= 15; Fig. 4). This descending inhibitory reflex was readily gradeable with changes in stimulus intensity (Fig. 5A and B). That is, as the balloon distension volume was increased, there was a corresponding increase in the amplitude of the evoked IJP in both muscle layers (Fig. 5A and B). Under no circumstances did we ever record any evidence of reciprocal neuromuscular inputs. Evoked IJPs in LM reached 6 ± 1 mV in amplitude (range: 2-10 mV), and had a duration (measured at half-maximal amplitude) of 2 ± 0.2 s (range: 0.6-3.7 s) (16 cells, n= 15 animals). In the CM, evoked IJPs had a mean peak amplitude of 9 ± 1 mV (range: 3-19 mV) and duration of 2 ± 0.3 s (range: 0.7-4.2 s) (16 cells, n= 15 animals). Evoked IJP amplitudes in the CM were significantly larger than those evoked in the LM (P < 0.05; paired t test). However, there was no significant difference in the half-durations of evoked IJPs in either muscle layer.

Figure 4.

Effects of balloon distension on intracellular electrical activity recorded simultaneously from the LM and CM layers

Upon distension of an intra-luminal balloon (500 μl; lower trace) applied oral to the recording electrodes, IJPs were evoked simultaneously in both muscle layers with an identical latency (descending inhibitory reflex). Following the evoked IJPs, there was a burst of action potentials in both muscles. Note, spontaneous and evoked action potentials in the LM do not propagate into the neighbouring CM and vice versa.

Figure 5.

Effects of grading the distension stimulus intensity on evoked inhibitory junction potentials in the LM and CM

A, graded increases in the balloon distension volume elicited similar graded increases in the amplitude of IJPs evoked in both the LM and CM. B, graphical representation of the effects of increasing distension stimuli on evoked IJP amplitudes in the LM (•) and CM layers (□) (from n= 5-9 animals). C, effects of mucosal stimulation on electrical activities of the LM and CM. A single brush stroke applied orally elicited synchronous IJPs in both LM and CM with an identical latency (see dotted line). In the LM layer, a single action potential occurred on repolarization of the IJP.

We also investigated the electrical reflex responses of both muscle layers in response to mucosal stimulation. During simultaneous recordings from six pairs of LM and CM cells, from six animals, we applied mucosal brush strokes (1-3 strokes in a series, delivered at approximately 1-2 Hz) to an orally situated region of mucosa. It was found that in all trials from all animals tested, mucosal stimuli elicited simultaneous IJPs in both muscle layers (Fig. 5C). Evoked IJPs in the LM had a mean peak amplitude of 8 ± 2 mV (range: 3-20 mV) and half-duration of 4 ± 1 s (range: 1-6 s). In the CM, evoked IJPs had a mean amplitude of 10 ± 2 mV (range: 6-18 mV) and half-duration of 4 ± 1 s (range: 3-5 s) (6 cells, n= 6 animals). The latencies of IJPs in both muscle layers (from the onset of the stimulus to the peak of the IJP amplitude was 2.3 ± 0.6 s in the LM and 2.3 ± 0.4 s in the CM. These latencies were not significantly different from one another (P > 0.05; paired t test).

Some evoked IJPs in both muscle layers were followed by short bursts of action potentials (1-7), which were not correlated between muscle layers (Fig. 4). These action potentials gave rise to strong contractions which usually dislodged at least one of the microelectrodes.

Conduction of spontaneous and evoked action potentials in circular muscle

We tested whether in pairs of circular muscle cells, action potentials would occur independently at close electrode separations (i.e. < 200 μm), as was found when recording simultaneously from the LM and CM layers, or whether spikes within the CM layer would be coordinated over these distances. Simultaneous intracellular recordings were made from pairs of circular muscle cells separated by 100 μm in the longitudinal axis. It was found in five pairs of CM cells (from n= 5 animals) that spontaneous action potentials consistently propagated over this distance, since activity at the two electrodes appeared to be phase locked. On an expanded time scale, it was noted that spikes propagated over 100 μm of CM in the longitudinal axis with a mean time of 11.0 ± 2.2 ms (range: 4-18 ms; n= 5), giving an apparent propagation velocity of 9 mm s−1 (n= 5). In the circumferential axis, action potentials were found to propagate dramatically faster than in the longitudinal axis, where action potentials occurred synchronously in the CM, even when the electrodes were separated by 5.5 mm circumferentially (see Fig. 6). In the circumferential axis, the mean time for action potentials to propagate over 1 cm of CM was 51 ± 4 ms (range: 40-52.2 ms), which gave an apparent propagation velocity of 192 mm s−1 (4 pairs of CM cells; n= 2).

Figure 6.

Effects of intra-luminal balloon distension on electrical activity recorded simultaneously from two CM cells

Despite the two electrodes being separated by 5.5 mm in the circumferential axis, spontaneous action potentials can be seen to occur synchronously. When the balloon is inflated oral to the two electrodes, IJPs are evoked synchronously at the two distant recording sites. Following the IJPs, there is a brief burst of action potentials, which are phase locked at the two recording sites. The lower trace is a monitor of the distension stimulus.

At these large electrode separations, it was found that when the colon was stimulated with either an intraluminal balloon or mucosal stimulation oral to the recording electrodes, IJPs were evoked synchronously at both recording electrodes (6 pairs of CM cells; n= 4) (Fig. 6). Also, the burst of action potentials following the evoked IJPs were, like spontaneous action potentials, well coordinated at the two sites (Fig. 6). The mean amplitude of action potentials in CM was 24 ± 2 mV (range: 13-35 mV; 6 cells; n= 5) with a mean half-duration of 50 ± 2 ms (6 cells, n= 5).

Effects of apamin and l-NA on evoked IJPs in longitudinal and circular muscles

The bee venom toxin apamin has been shown to block the evoked IJPs in the CM layer of guinea-pig ileum following electrical stimulation (Bywater et al. 1981) or balloon distension and mucosal stimulation (Smith & Furness, 1988). In light of this, we sought to investigate the effects of apamin on evoked IJPs in the LM and CM of the guinea-pig distal colon. In four animals tested, apamin (500 nm) was found to significantly attenuate evoked IJP amplitudes in the CM (control: 11 ± 2 mV; apamin: 5 ± 1 mV; P= 0.05), but not in the LM (control: 8 ± 1 mV; apamin: 6 ± 1 mV; P= 0.17) (Fig. 7). Overall, from five animals, apamin (500 nm) did not significantly modify the resting membrane potential of the LM (control: 32.8 ± 1.9 mV; apamin: 33.9 ± 5.4; P > 0.05), or CM (control: 37.6 ± 6.5 mV; apamin: 34.4 ± 3.4 mV; P > 0.05). To test whether apamin-resistant evoked IJPs in the LM and CM may have involved release of nitric oxide (NO), l-NA (100 μm) was further applied to preparations already perfused with apamin. In all animals tested (n= 4), application of l-NA abolished apamin-resistant inhibitory transmission to both muscle layers (Fig.7C). l-NA (100 μm) applied after apamin significantly depolarized the resting potentials of the CM (from 35.2 ± 2.7 to 30.2 ± 2.9 mV; P= 0.02; n= 5), but did not signficantly modify the membrane potentials of the LM (from 36.7 ± 4.3 to 31.1 ± 1.0 mV; P= 0.28; n= 4). To further test the possibility of an apamin-sensitive neurotransmitter to the LM, l-NA was applied first to the colon. In all of six animals tested, l-NA (100 μm) significantly reduced, but did not block the reflex-evoked IJP to the LM layer (amplitude, control: 14 ± 2 mV; l-NA: 8 ± 1 mV; P= 0.004; n= 6). In 4 of these 6 animals, apamin (500 nm) was further applied and was found to consistently abolish the l-NA-resistant evoked IJP to the LM. In the CM layer the evoked IJP was also incompletely blocked by l-NA (control: 7 ± 2 mV; l-NA: 4 ± 0.7 mV; P= 0.25), where l-NA-resistant evoked inhibitory transmission to the CM was also abolished by apamin.

Figure 7.

Effects of apamin and a nitric oxide synthesis inhibitor on the descending inhibitory reflex

A, control electrical responses in the LM and CM following intraluminal balloon distension. B, apamin (500 nm) attenuated the amplitude of the ‘fast’ IJP in the CM layer, without any pronounced effect on the LM layer. C, the ‘slow’ apamin-resistant IJP in both muscle layers is abolished by the further addition of l-NA (100 μm).

Responses of the longitudinal and circular muscle during an ascending excitatory reflex

To investigate whether stimuli applied anal to the microelectrodes would elicit an ascending excitatory reflex in both muscle layers, simultaneous recordings were made from both muscles and mucosal stimuli (1-3 brush strokes at 1-2 Hz) were applied anally. Brush stroking the mucosa anally elicited EJPs synchronously in both muscle layers (Fig. 8 and Fig. 9). Evoked excitatory potentials in both the LM and CM often reached threshold for action potential initiation (Fig. 8 and Fig. 9). In preparations where EJPs reached action potential threshold, a variable number of action potentials were elicited in each muscle and even during multiple trials within the same preparation (0-4 spikes). This always generated powerful contraction of the muscles and the impalements were usually lost. In the LM layer, the mean amplitude of the evoked EJP and the associated action potential was 17 ± 4 mV (range: 3 mV subthreshold to 33 mV suprathreshold), with a half-duration of 380 ± 160 ms (7 cells, n= 7 animals). In the CM, the mean amplitude of the evoked EJP and its action potential was 19 ± 6 mV (range: 4 mV subthreshold to 44 mV suprathreshold) with a half-duration of 1900 ± 250 ms (7 cells, n= 7 animals). There was no significant difference in the latencies of evoked EJPs in either muscle layer (LM: 1080 ± 250 ms; CM: 1160 ± 240 ms; P > 0.05; n= 7).

Figure 8.

Simultaneous intracellular recordings from LM and CM during the ascending excitatory reflex

A, mucosal stimuli (1 brush stroke, see arrow), applied anal to the electrodes, elicited synchronous EJPs and an accompanying action potential in both the LM and CM. B, in another preparation, single anal mucosal stimuli applied on four separate occasions (see arrows) elicited an EJP in the CM that did not reach action potential threshold, but in the LM an EJP and action potential were evoked.

Figure 9.

Effects of atropine on the ascending excitatory reflex

A, a single brush stroke applied to the mucosa (see arrow), anal to the recording electrodes, evoked a compound EJP and associated action potentials synchronously in the LM and CM layers. Spontaneous depolarizations and action potentials can be observed in the LM prior to and following the stimulus. B, in the presence of atropine (1 μm), the evoked EJP was abolished in both the LM and CM and a small IJP was evoked in both muscles synchronously.

Effects of atropine on the ascending excitatory reflex

We tested the involvement of cholinergic transmission underlying the evoked EJP in both LM and CM, by applying atropine to the colon. Atropine (1 μm) abolished the evoked EJP in both the LM and CM (n= 4; Fig. 9) and no evidence was found of any non-cholinergic excitatory neurotransmission to either the LM or CM. There was no significant membrane potential change following addition of atropine in the CM layer (control: −34.4 ± 5.4 mV; atropine: −36.5 ± 4.5 mV; P= 0.63; n= 3) or LM layer (control: −38.8 ± 0.7 mV; atropine: −37.1 ± 5.8 mV; P= 0.91; n= 3). It is noteworthy that in 2 of the 4 animals where atropine abolished the ascending excitatory reflex, it was possible to elicit IJPs (ascending inhibition) synchronously in the LM and CM (Fig. 9B), from anal mucosal stimuli once the evoked EJP had been blocked.

Identification of microelectrode impalements in different muscle layers

To confirm that intracellular recordings were made from different muscle layers, neurobiotin was injected into the smooth muscle cells. In 13 guinea-pigs, we made intracellular recordings from the LM and CM simultaneously and passed outward current pulses (from +10 pA to +0.2 nA) of 500 ms duration at 1 Hz for 2-7 min. This process was found to clearly identify seven pairs of LM and CM cells from five of the animals injected. In all cells, where neurobiotin was injected into different muscle layers, we were clearly able to identify that microelectrode impalements were in fact made from different muscle layers, since they were found to exist in different axes (Fig. 10). In the remaining animals, where longer duration injections were performed and many pairs of cells were injected within the same animal, it was found that the neurobiotin had leaked from the injected cell into neighbouring cells. Nevertheless, it was still possible to ascertain that the electrodes were in different muscle bundles from different muscle layers.

Figure 10.

Confocal micrograph (Z-series 15 stacks) showing neurobiotin injection into a single LM cell and CM cell

After the neurobiotin was conjugated with streptavidin FITC and viewed under the appropriate fluorescence filters, a single LM cell (B) and CM cell (A) were observed. LM cells could be clearly identified over CM cells, since they were at 90 deg to each other. The scale bar in each panel represents 50 μm.

DISCUSSION

It has been well documented that the central neural control of reflex pathways during voluntary movement involves the reciprocal activation of skeletal muscle, whereby one muscle (the flexor) contracts, while at the same time, the opposing muscle (the extensor) relaxes. This action permits the generation of free muscle movements (i.e. the contraction of the biceps and relaxation of the triceps muscle in man). Although this mechanism is well established in the generation of voluntary skeletal muscle movements of the body (see Jankowska et al. 1985; Enriquez-Denton et al. 2000), our data clearly suggest that a similar mechanism does not apply to the involuntary movements of the longitudinal and circular smooth muscle layers during the peristaltic reflex in the large intestine.

In this study, we were definitively able to test the hypothesis of reciprocal innervation by simultaneously recording the electrical activity from the LM and CM layers of the guinea-pig colon during the peristaltic reflex. During these recordings, the colon was stimulated by intraluminal balloon distension to mimic colonic distensions by fecal pellets, and by mucosal stimulation, to compare the reflex responses in both muscle layers. A major finding of the current study is that excitatory junction potentials were evoked synchronously in the LM and CM oral to, and inhibitory junction potentials anal to, a local physiological stimulus. This suggests a strong similarity in the timing of reflex activation of the inhibitory and excitatory motoneurons to both muscle layers. Also, these data suggest that both muscle layers of the distal colon have a functional innervation by both excitatory and inhibitory motoneurons.

Descending inhibitory reflex

With both stimulus regimens, the LM and CM received IJPs simultaneously anal to a local stimulus, with no difference in the latency of onset. This suggests that during reflex stimulation, the inhibitory motoneurons to the LM and CM must receive a synchronous burst of fast excitatory postsynaptic potentials from descending interneurons or sensory neurons directly. Further support for this was shown when spontaneous IJPs occurred synchronously in both muscle layers (Fig. 3B), suggesting that the different inhibitory motoneurons that innervate each muscle layer receive synchronous bursts of synaptic inputs from enteric interneurons.

Apamin reduced the evoked IJP amplitude in the CM by half, but had no significant effect on IJPs in the LM, when applied first to the colon, suggesting that ‘SK’ channels may be more important in inhibitory neurotransmission to the CM rather than LM layer. However, an apamin-sensitive IJP was recorded from the LM layer when l-NA was applied first to the colon. These effects of apamin are supported by the previous findings of Smith et al. (1992), where it was reported that apamin abolished the evoked fast IJP revealing an underlying slower component, when extracellular recordings were made from the entire muscularis externa of the guinea-pig colon. In our study, the apamin-resistant evoked IJP in both LM and CM was abolished by l-NA, suggesting that nitric oxide is likely to underlie the slow inhibitory response in both layers. Similar findings were obtained in the guinea-pig colon by Watson et al. (1997) in their study of the CM layer only. The apamin-sensitive transmitter(s) appears, then, to be involved in the ‘fast’ IJP in both guinea-pig colon and ileum (Bywater et al. 1981; Niel et al. 1983), whereas the slow apamin-insensitive IJP is mediated by NO (Niel et al. 1983; He & Goyal, 1993; Watson et al. 1997). In the mouse colon, the evoked IJP also consists of a fast apamin-sensitive component, and a fast and slow apamin-insensitive component (Spencer et al. 1998a). In the current study, the relative contributions of the apamin-sensitive transmitter and NO varied in each animal, yet a combination of l-NA and apamin was required to block evoked inhibitory neuromuscular transmission. It is our impression that in the guinea-pig distal colon, the apamin-sensitive IJP is most pronounced in the CM layer. However, although the nitrergic ‘slow’ IJP occurs in both muscle layers, it is more readily observed in the LM layer.

Ascending excitatory reflex

In contrast to the descending inhibitory reflex, mucosal stimulation applied anal to the microelectrodes elicited synchronous EJPs in both the LM and CM, which often discharged bursts of action potentials and generated powerful muscle contraction of both muscle layers (Smith & McCarron, 1998; Spencer et al. 1999a). It is particularly interesting that following application of atropine to abolish EJPs in both muscle layers, IJPs were evoked simultaneously in both muscles, oral to the point of stimulation (from 2 of 4 animals). Smith et al. (1992) also reported an ascending inhibitory reflex during recordings from the muscularis externa in this tissue. Yet, the exact role of such inhibition is unclear. One possibility is that the ascending inhibitory reflex acts to truncate the ascending excitatory reflex. In our hands, acetylcholine appears to underlie the major membrane potential changes associated with the evoked EJP in both the longitudinal and circular muscle layers, at least during reflex stimulation.

Previously, distension and mucosal stimulation have been shown to evoke polarized electrical reflex responses in the CM of the guinea-pig ileum (Smith et al. 1991). However, in the colon, electrical responses in the smooth muscle following mucosal stimulation have not been recorded. The current study has shown that mucosal stimulation alone is indeed a powerful stimulus to evoke polarized reflex responses in both the LM and CM layers of the colon.

Synchronous activation of the LM and CM during the peristaltic reflex

Our major findings using intracellular recordings from both muscle layers add further confirmation to our previous contractile (Smith & Robertson, 1998; Smith & McCarron, 1998; Spencer et al. 1999a,b) and calcium imaging (Stevens et al. 2000) studies that the LM and CM in both the small and large bowel are neurally connected to contract together and relax together during the peristaltic reflex. We have found no evidence to support the idea of reciprocal neuromuscular inputs during peristalsis, as suggested by Kottegoda (1969, 1970) and Wood (1987). We suspect that any apparent reciprocal movements reported by other investigators in vivo, or in vitro, are most likely to be due to passive mechanical interactions between the muscle layers (see Gregory & Bentley, 1968; Wood & Perkins, 1970).

Spontaneous electrical activity in the LM and CM

A complex variety of electrical events were recorded in both muscle layers of the colon. Spontaneous depolarizations (SDs) were commonly recorded from the LM at a frequency of about 9 min−1, which usually generated smooth muscle action potentials. The configuration, time course and frequency of the SDs in the LM were very similar to the electrical events reported in the LM layer of human jejunum (see Fig. 9B of Hara et al. 1987). In addition to SDs, we also recorded spontaneous action potentials, similar to the action potentials described in the smooth muscle of the taenia coli (Tomita, 1967). It is noteworthy that the resting potentials of the CM were low (≈35 mV) in our preparation since they were similar to those of the LM. Previous studies of the distal colon have reported RMPs of approximately −50 mV for the CM (see Watson et al. 1997). Two likely explanations are possible. (1) The pacemaker cells (interstitial cells of Cajal, ICC) may have been removed from the submucosal border when the submucosa was removed to facilitate microelectrode penetration into CM cells. Removal of submucosal ICC appears to lead to depolarization of the CM (Liu & Huizinga, 1993). (2) Alternatively, the depolarized resting membrane potentials of both muscle layers may have been due to stretch applied to the muscle during the pinning procedure of the tissue. Previously, stretch has been shown to depolarize smooth muscle (see Bülbring, 1955). The absence of clearly defined slow waves in the CM may be due to (1) removal of submucosal ICC during the dissection procedure (see Smith et al. 1987), (2) removal of circumferential integrity by cutting open the intestinal tube, which may uncouple slow wave pacemakers (Smith, 1989), or (3) the rareness or absence of slow waves in the guinea-pig colon as is the case in the mouse colon (Bywater et al. 1989; Spencer et al. 1998b). However, when slow waves have been recorded in this tissue with extracellular electrodes, the responses to reflex stimulation are similar to the responses described in this study (Smith et al. 1992). That is, stimulation of the ascending excitatory reflex elicits a suprathreshold, premature slow wave, whereas activation of the descending inhibitory reflex reduces the amplitude and delays the onset of slow waves.

Conclusions

Using simultaneous intracellular recordings from both the LM and CM layers during the peristaltic reflex, we have shown unequivocally that the two muscle layers of the distal colon receive simultaneous inhibitory neuromuscular inputs during descending inhibition, and simultaneous excitatory neuromuscular inputs during ascending excitation. That is, both the LM and CM of the guinea-pig distal colon exhibit similar polarized neural reflex pathways.

Acknowledgements

This study was supported by a grant from the National Institutes of Health (USA) (no. RO1 NIDDK 45713) awarded to T.K.S. We also wish to acknowledge Yulia Bayguinov and Sean Ward for their expert technical assistance with the confocal microscope and immunohistochemistry, supported by the core facilities of a program project grant PO1 DK41315.

Ancillary