Properties of unitary potentials generated by intramuscular interstitial cells of Cajal in the murine and guinea-pig gastric fundus

Abstract

Intracellular recordings were made from isolated bundles of the circular muscle layer of mouse and guinea-pig gastric fundus. These preparations displayed an ongoing discharge of membrane noise (unitary potentials), similar to that recorded from similar preparations made from the circular layer of the antrum. Bundles of muscle from the fundus of W/W^V mice, which lack intramuscular interstitial cells of Cajal (ICC_IM) lacked the discharge of membrane noise observed in wild-type tissues. When the membrane potential was changed by passing depolarizing or hyperpolarizing current pulses, the discharge of membrane noise was little changed. The membrane noise was unaffected by adding chloride channel blockers; however, agents which buffered the internal concentration of calcium ions reduced the discharge of membrane noise. Treatment of tissues with CCCP, which interferes with the uptake of calcium ions by mitochondria, also reduced the membrane noise and caused membrane hyperpolarization. Similar observations were made on bundles of tissue isolated from the circular layer of the guinea pig antrum. Together the observations indicate that membrane noise is generated by a pathway located in ICC_IM. The properties of this pathway appear to vary dramatically within a given organ. The lack of voltage sensitivity of the discharge of membrane noise in the fundus provides a possible explanation for the lack of rhythmic electrical activity in this region of the stomach.

Many regions of the gastrointestinal tract generate slow waves. In the small intestine and stomach, slow waves are initiated by a network of interstitial cells of Cajal located in the myenteric region (ICC_MY) and are absent in tissues which lack ICC_MY (Ward et al. 1994, 1997, 1999; Huizinga et al. 1995; Ordog et al. 1999). In the gastric antrum of guinea pigs and mice, ICC_MY generate large amplitude, long lasting pacemaker potentials which spread passively to the circular muscle layer (Dickens et al. 1999; Hirst & Edwards, 2001; Hirst et al. 2002a; Cousins et al. 2003). Pacemaker depolarizations are augmented in the circular muscle layer by unitary potentials, which can summate into regenerative potentials, a second component of slow waves, in response to pacemaker depolarization (see Ohba et al. 1975; Dickens et al. 1999; Hirst et al. 2002a). Unitary potentials, which sum to give a regenerative potential, are initiated by a second population of ICC, intramuscular interstitial cells (ICC_IM), which are distributed amongst the smooth muscle cells making up the circular layer of the gastric antrum (Burns et al. 1996, 1997). Unitary and regenerative potentials are absent in antral muscles devoid of ICC_IM (Dickens et al. 2001; Hirst et al. 2002a). Regenerative potentials can be initiated in bundles of circular muscle containing ICC_IM and smooth muscle cells by direct electrical stimulation (Suzuki & Hirst, 1999; Edwards et al. 1999; Van Helden et al. 2000; Fukuta et al. 2002; Kito et al. 2002). Isolated bundles of circular muscle, which lack ICC_MY, generate regenerative potentials spontaneously, but only if ICC_IM are present (Suzuki & Hirst, 1999; van Helden et al. 2000; Fukuta et al. 2002; Kito et al. 2002; Suzuki et al. 2003). When regenerative potentials occur spontaneously they do so at a lower frequency than the pacemaker potentials generated by ICC_MY. Hence, in gastric muscles, ICC_MY are the dominant pacemaker and ICC_IM make little contribution to the rate of slow wave generation.

Fundus muscles lack networks of ICC_MY, and these tissues do not generate spontaneous electrical activity (Burns et al. 1996; Ward et al. 2000). In wild-type animals both the longitudinal and circular layers of the fundus contain abundant populations of ICC_IM and generate spontaneous electrical activity in the form of a discharge of membrane noise: this is only the case if ICC_IM are present (Burns et al. 1996; Ward et al. 2000; Beckett et al. 2002). Thus in the fundus unitary potentials do not summate to produce regenerative potentials but the reasons why antral and fundus muscles do not share the ability to produce regenerative potentials are unclear.

In the present study we examined the properties of isolated bundles of fundus circular muscle from wild-type mice and guinea pigs containing ICC_IM and from W/W^V animals which lack ICC_IM. Neither preparation generated spontaneous regenerative potentials, but wild-type mice displayed an ongoing discharge of membrane noise. The discharge of membrane noise in wild-type mice and guinea pig fundus showed many temporal characteristics similar to the activity recorded in the antrum; however, several electrical characteristics and pharmacological sensitivities differed.

Our observations suggest that ICC_IM produce the membrane noise in electrical recordings from gastric fundus muscles but that their frequency cannot be changed by changing their membrane potential. The results provide a possible explanation for the lack of electrical rhythmicity in the fundus.

Methods

Experiments were carried out on age-matched C57BL/6 wild-type and W/W^V mutant mice obtained from Jackson Laboratory (Bar Harbour, ME, USA) and on guinea-pigs. Mice of either sex were killed by cervical dislocation and exsanguination. Guinea-pigs of either sex, weighing between 150 and 250 g were killed by isoflurane inhalation followed by exsanguination. The use and treatment of animals was approved by the Animal Use and Care Committee at the University of Nevada.

With either animal the stomach was exposed and transferred to a dissecting chamber filled with oxygenated (97% O₂–3% CO₂) physiological saline (composition; mm: NaCl, 120.7; NaHCO₃, 15.5; NaH₂PO₄, 1.2; KCl, 5.9; MgCl₂, 1.2; CaCl₂, 2.5; and dextrose, 11.5). The stomach was cut along the lower curvature and the mucosa dissected away from the fundus. Subsequently the preparation was re-pinned serosal surface uppermost and the longitudinal muscle layer was dissected away. Single bundles of circular muscle (diameter 50–150 μm, length 400–800 μm) were dissected free and pinned in a recording chamber (see Suzuki & Hirst, 1999). Intracellular recordings were made using sharp microelectrodes (90–150 MΩ) filled with 3 m KCl. In some experiments the preparations were impaled with two independently mounted electrodes; one was used to record membrane potential changes and the other to inject current so changing the membrane potential of the preparations. Signals were amplified with an Axoprobe amplifier, low pass filtered (cut-off frequency 1 kHz) digitized and stored on a computer for later analysis. Preparations were constantly perfused with physiological saline solution warmed to 37°C. In many experiments, nifedipine (1.0 μm), which has been shown not to significantly affect the waveform of antral regenerative potentials (Suzuki & Hirst, 1999), was added to the physiological saline to suppress muscle movements.

Atropine sulphate, apamin, N^ω-nitro-l-arginine (L-NA), anthracene-9-carboxylic acid (9-AC), acetoxy-methyl ester of bis-(2-amino-5-phenoxy)ethane-N,N, N′,N′-tetraacetic acid (BAPTA-AM), carbonyl cyanide m-chlorophenyl hydrazone (CCCP), diisothio-cyanatostilbene-2,2′-disulphonic acid (DIDS), and nifedipine (obtained from Sigma Chemical Co., St Louis, MO, USA) were used in these experiments.

To examine the distribution of ICC in bundles of circular muscle, immunohistochemistry was performed using an antibody against the Kit receptor (ACK2; Gibco-BRL, Gaithersburg, MD, USA), which has been shown previously to be specific for ICC in the gastrointestinal tract (Torihashi et al. 1995). ICC were labelled with ACK2 that was conjugated to the fluorescent dye Alexa Fluor 594 (AF 594) using the Alexa Fluor 594 Protein Labeling Kit (Molecular Probes, Eugene, OR, USA). Tissues were placed in cooled physiological saline (4°C) and refrigerated for 1 h to reduce non-specific binding. ACK2-AF 594 conjugates were diluted to 5 μg ml⁻¹ in cold saline and applied to tissues for 40 min at 4°C. Following antibody loading and a brief wash in cold saline, gastric tissues were fixed with paraformaldehyde solution (PFS; 4% w/v; 10 min) and washed in phosphate buffered saline (PBS) to remove excess fixative before mounting on glass slides. Tissues were examined with a Zeiss LSM 510 Meta confocal microscope (Carl Zeiss Inc., Germany) with an excitation wavelength appropriate for Alexa Fluor 594. ICC_IM distribution was determined by optically sectioning through the entire muscle bundle, followed by reconstruction of z-stacks. Confocal micrographs shown in this paper are digital composites of Z-series scans of 9–24 optical sections through a depth of 7.5 μm to 15.0 μm. Final images were constructed using Zeiss LSM 5 Image Examiner software.

All data are expressed as the mean ± standard error of the mean (s.e.m.). Student's t test was used to determine if data sets differed; P values of less than 0.05 were taken to indicate significant differences between sets of observations.

Results

General observations

Confocal micrographs of isolated bundles of circular muscle, dissected from either C57BL/6 wild-type or W/W^V mice and labelled with an antibody to the Kit receptor (ACK2), showed that ICC_IM were distributed throughout the wild-type muscle bundles and were observed running parallel to the circular muscle fibres (n = 6; Fig. 1A). The number of ICC_IM counted per cross-sectional area in each bundle ranged between 8 and 20 cells (average 13.7 ± 1.7, n = 6). ICC_IM were not detected in muscle bundles from W/W^V mice (n = 5; Fig. 1E). Kit immunohistochemistry also confirmed that fundus circular muscle bundles were devoid of ICC_MY.

Figure 1. Distribution of kit positive cells in bundles of circular muscle isolated from the fundus of C57BL/6 and W/W^V mice and the discharge of membrane noise.

A and E, the distribution of Kit-immunopositive ICC_IM in bundles of circular muscle dissected from the fundus of a C57BL/6 mouse (A) and a W/W^V mouse (E). Elongated, spindle-shaped ICC_IM running parallel to the circular muscle fibres were readily detected in preparations obtained from C57BL/6 mice (A) but not in W/W^V mice (E). A and E are composites of 9 μm × 0.8 μm optical sections through the circular muscle bundles. The white dotted lines denote the border of the muscle bundles in both A and E. The 50 μm scale bar in E is for both panels A and E. Samples of membrane noise recorded from a circular muscle bundle isolated from the fundus of a C57BL/6 mouse in control conditions, in the presence of nifedipine (1 μm) and in the presence of both nifedipine and apamin (0.2 μm) are shown in B, C and D, respectively. Recordings from a bundle isolated from W/W^V fundus in control conditions, in the presence of nifedipine (1 μm) and in the presence of both nifedipine and apamin (0.2 μm) are shown in F, G and H, respectively. The calibration bars shown in panel H also refer to panels B–D and F–G.

When intracellular recordings were made from muscle bundles from wild-type mice, resting potentials in the range of −35 to −45 mV were recorded (average −40.6 ± 0.9 mV, n = 14). When the preparations were impaled with two electrodes, with one being used to inject pulses of current, electronic potentials were detected with the second electrode. The muscle bundles had input resistances in the range 1.5–16.0 MΩ (6.0 ± 1.7 MΩ, n = 8) and membrane time constants in the range 28–71 ms (44.3 ± 5.0 ms, n = 8). An ongoing discharge of membrane noise (unitary potentials) was recorded in each of the wild-type circular muscle bundles. This activity never summated into regenerative potentials (n = 14; Fig. 1B) as previously observed in similar preparations of muscle bundles from the gastric antrum (see Suzuki et al. 2003). Occasionally, the membrane noise appeared to trigger action potentials (n = 3), and these events were abolished by nifedipine (1 μm), clearly distinguishing them from regenerative potentials.

When intracellular recordings were made from bundles from W/W^V fundus, resting potentials in the range −42 to −48 mV were recorded (average −45.0 ± 1.0 mV, n = 5). In all 5 W/W^V preparations, the membrane potential recordings displayed low frequency transient hyperpolarizing potentials (Fig. 1F). These events remained in the presence of TTX (0.5 μm) but were readily abolished by apamin (0.1 μm; Fig. 1H). The discharge of membrane noise, as recorded in wild-type muscles was not recorded in W/W^V tissues under control conditions (Fig. 1F), in the presence of nifedipine (Fig. 1G) or in the presence of nifedipine and apamin (Fig. 1H).

Spectral density curves were constructed from membrane noise recordings obtained from five wild-type and five W/W^V preparations (Fig. 2). These data were recorded in the presence of nifedipine (1 μm) and apamin (0.1 μm). Typical spectra are shown in Fig. 2B and D. Power spectral density curves from wild-type muscle recordings had characteristics and shapes similar to those calculated from recordings from the circular layer of guinea-pig antrum (Edwards et al. 1999). Energy increased from low values at frequencies above 10 Hz to reach a plateau at frequencies between 3 and 1 Hz. At frequencies above 10 Hz, the power spectral density curves were dominated by electrode recording noise (Fig. 2B). Power spectral density curves, determined from isolated bundles of antral circular muscle containing ICC_IM could be well described by assuming that the noise was made up of many unitary potentials (Edwards et al. 1999). When the same approach was applied to spectral density curves obtained from fundal preparations, similarly good fits were obtained (Fig. 2B). In this series of experiments constant A had values in the range 261–321 ms (294 ± 12 ms) and constant B had values in the range 33–42 ms (36.6 ± 1.5 ms; n = 5). In muscle bundles from W/W^V mutants a plateau between 3 and 1 Hz was not observed and was better fitted with a linear function. The electrode recording noise dominated the power spectra at frequencies above 10 Hz (Fig. 2D).

Figure 2. Power spectral density curves calculated from recordings made from wild-type and W/W^V fundus muscle bundles.

Samples of membrane noise recorded from a circular muscle bundle isolated from the fundus of a C57BL/6 mouse (A) and a W/W^V mouse (C) in the presence of nifedipine and apamin are shown in A and C, respectively. B and D, power spectral density curves derived from wild-type (A) and W/W^V (C) recordings, respectively.

Effect of changing membrane potential on discharge of membrane noise recorded from isolated circular bundles of mouse fundus

In some experiments bundles of wild-type fundus circular muscle were impaled with two recording electrodes. One electrode was used to inject current and the other recorded changes in membrane potential. In each of eight fundus preparations, periods of membrane depolarization failed to initiate regenerative potentials as activated by depolarization in bundles of antral muscle (Suzuki & Hirst, 1999; Hirst et al. 2002b; Kito et al. 2002; Teramoto & Hirst, 2003). Regenerative potentials were also not activated at the break of a period of membrane hyperpolarization (Fig. 3A). In the antrum, the discharge of membrane noise is abolished by making the membrane potential some 10–20 mV more negative for 10 s (Teramoto & Hirst, 2003). In each of eight wild-type fundus muscles, making the membrane potential by some 20 mV more negative failed to abolish the discharge of membrane noise (Fig. 4). The shape of the power spectral density curve was unchanged but the power was slightly increased (Fig. 4C and D). Together these observations show that the discharge of membrane noise by fundal ICC_IM is little affected by changes in membrane potential.

Figure 3. Electrical properties of a bundle of circular muscle isolated from a C57BL/6 wild-type fundus.

A, averaged electrotonic potentials recorded from a wild-type circular muscle bundle produced by injecting hyperpolarizing currents. Break of membrane hyperpolarization failed to initiate regenerative potentials. B, altering the membrane potential over a range of potentials using depolarizing and hyperpolarizing current pulses also failed to elicit regenerative potentials.

Figure 4. Effect of changing the membrane potential on discharge of membrane noise in a bundle of circular muscle isolated from wild-type fundus.

A, an averaged trace of 10 electrotonic potentials produced by injecting 1 nA of hyperpolarizing current into a circular muscle bundle. Three sample traces of individual electrotonic potentials are shown in B. Note that hyperpolarizing the bundles by over 20 mV failed to abolish the discharge of membrane noise. Hyperpolarized (a) and baseline regions of these traces (b) were used to construct the power spectral density curves shown in C and D (○, hyperpolarized region, a in panel C; •, baseline region, b in panel D). Note that hyperpolarization did not change the shape of the power spectral density curve but the power function was increased by approximately 100%.

Effect of chloride channel blockers on the discharge of membrane noise recorded from isolated circular bundles of mouse fundus

In the circular layer of the murine gastric antrum, it has previously been demonstrated that the discharge of membrane noise by ICC_IM is abolished by the chloride channel blockers DIDS and 9-AC (Suzuki et al. 2003). In contrast the discharge of membrane noise by ICC_IM in the murine fundus was unaffected by DIDS (100 μm; n = 5) or by 9-AC (500 μm; n = 4). An example of these experiments is shown in Fig. 5.

Figure 5. Lack of effect of chloride channel blocker on membrane noise recorded from wild-type circular muscle bundle.

Three traces of membrane noise recorded from a circular muscle bundle in control conditions (nifedipine; 1 μm) are shown in A and power spectra in B. The chloride channel blocker DIDS (100 μm) did not alter the discharge of membrane noise, demonstrated by representative traces (C) and the calculated power spectra (D).

Effect of BAPTA-AM and inhibitors of mitochondrial Ca²⁺ uptake on the discharge of membrane noise recorded from isolated circular bundles of mouse fundus

In the antrum the discharge of membrane noise results from the discharge of unitary potentials by ICC_IM. This also appeared to be the case with the discharge of membrane noise recorded from the mouse fundus. The effects of BAPTA-AM (20–30 μm) were examined in nine wild-type fundus bundles. In four preparations BAPTA-AM produced membrane hyperpolarization, ranging from 3.5 to 18.0 mV (average 8.4 ± 3.2 mV, n = 4). In three preparations BAPTA did not change the resting membrane potential and in the remaining two muscle bundles BAPTA caused slight membrane depolarization (of 3.0 and 4.5 mV). Regardless of the effect on resting membrane potential, in all preparations BAPTA produced a decrease in the discharge of membrane noise (n = 9; Fig. 6C and D).

Figure 6. Effect of buffering [Ca²⁺]_i on discharge of membrane noise in a bundle of circular muscle isolated from the mouse fundus.

A and B, membrane noise discharge recorded from a wild-type circular muscle bundle in control conditions and the calculated power spectra, respectively. BAPTA-AM (20 μm) dramatically reduced the frequency of membrane noise discharge until individual unitary potentials could be resolved (C). Power spectra curves calculated from BAPTA-AM traces demonstrated a substantial decrease in power (D). The power function had to be reduced by 50% of control to obtain an adequate fit in BAPTA-AM.

The mitochondrial uncoupler CCCP (10 μm) also reduced the discharge of membrane noise (Fig. 7C and D) and consistently produced membrane hyperpolarization ranging from 8 to 24 mV (average 14.1 ± 2.3 mV, n = 7).

Figure 7. Effect of inhibiting Ca²⁺ uptake by mitochondria on discharge of membrane noise in a bundle of circular muscle isolated from the mouse fundus.

Membrane noise recorded from a wild-type circular muscle bundle in control conditions and the corresponding power spectra curve are shown in A and B, respectively. C, the mitochondrial inhibitor CCCP (10 μm) abolished membrane noise activity. D, power spectra curve calculated from the traces recorded in the presence of CCCP.

Properties of membrane noise recorded from bundles of circular muscle taken from fundus of guinea pigs

All of the previous observations on the mouse fundus have suggested that membrane noise arises from a discharge of unitary potentials by ICC_IM but that the discharge is not affected by changes in membrane potential, nor does it involve the activation of chloride channels. This suggests that mouse antral and fundal ICC_IM although sharing some properties have a number of differences. To check whether this simply reflected a species variation, a brief series of experiments was carried out on bundles of circular muscle obtained from guinea pig fundus. In brief the observations made on the guinea pig were similar to those made on the mouse. The resting membrane potentials of the bundles ranged from −40 to −53 mV (−44.8 ± 1.2 mV), their input resistance ranged from 1.2 to 15.0 MΩ (4.9 ± 1.8 MΩ) and their membrane time constants ranged from 55 to 153 ms (115 ± 12 ms). Depolarizing or hyperpolarizing bundles of guinea pig fundus failed to evoke regenerative potentials (Fig. 8A; n = 7). The discharge of membrane noise was also unaffected by the chloride channel antagonist 9-AC (500 μm − 1 mm; n = 4). An experiment is illustrated in Fig. 8B, where the power spectral density curve determined in control is compared to that determined in the presence of 9-AC (1 mm). However, again the discharge of membrane noise was inhibited by BAPTA-AM (30 μm), which dramatically reduced the membrane noise, associated with a membrane hyperpolarization of 5–15 mV (9.4 ± 1.6 mV; n = 5; Fig. 8C).

Figure 8. Similarities between properties of membrane noise recorded from a circular muscle bundle isolated from guinea pig fundus and that recorded from a mouse circular muscle bundle.

A, current injection pulses were used to vary the membrane potential of a circular muscle bundle of guinea pig fundus over a wide range of potentials (from 15 mV more positive than resting membrane potential to 40 mV more negative than resting membrane potential). Neither membrane depolarization, nor the break after 5 s of hyperpolarization initiated a regenerative potential. B, power spectral density curves calculated from membrane noise recorded before (•) and after (○) the addition of 9-AC could be fitted by the same function. C, power spectral density curves calculated from membrane noise recorded in control conditions (•) and in the presence of 30 μm BAPTA-AM (○). After 15 min in BAPTA-AM membrane noise was effectively abolished.

Discussion

These experiments have shown that the properties of unitary potentials in the proximal and distal stomach vary profoundly. Discharges of unitary potentials were only recorded from fundus muscles that contained ICC_IM (Figs 1 and 2; and see Beckett et al. 2002). This has also been shown to be the case in the murine antrum (Suzuki et al. 2003). Thus, it is likely that the specialized mechanisms responsible for generation of unitary potentials are present in ICC_IM. In the mouse fundus the discharge of unitary potentials never summated to produce regenerative potentials, a typical feature of spontaneous activity of guinea pig and murine antral stomach muscle bundles dissected free of the dominant pacemaker ICC_MY (see Suzuki & Hirst, 1999; Suzuki et al. 2003). Summation of unitary potentials into a regenerative potential can also be induced by depolarization or at the break of hyperpolarization in the antrum. In the murine fundus it was not possible to cause summation of unitary potentials by depolarization or hyperpolarization. These observations suggest that the voltage sensor affecting fundus unitary potentials is either absent or considerably less sensitive than the voltage sensor in antral ICC_IM (Hirst & Ward, 2003). Since the presence of such a voltage sensor appears to be essential for the increased frequency and summation of unitary potentials that occurs during a regenerative potential (Hirst et al. 2002b), the present observations give an explanation for why the fundus fails to generate (and propagate) this type of activity. Further differences in unitary potentials recorded from fundus and antrum were found in their pharmacological properties; the discharge of unitary potentials by fundal ICC_IM was not affected by chloride channel blockers (Fig. 5) whereas these agents blocked unitary potentials in the antrum (Hirst et al. 2002b; Suzuki et al. 2003).

In the stomach both the antrum and pylorus generate an ongoing discharge of slow waves. In these regions ICC_MY generate pacemaker potentials which passively depolarize the circular muscle layer (Hirst & Ward, 2003; Cousins et al. 2003). The passive waves of pacemaker depolarization are then augmented by the voltage activation of ICC_IM (Hirst & Ward, 2003). In both regions when ICC_MY are dissected away, a low frequency discharge of rhythmical activity (regenerative potentials) is generated by ICC_IM (Suzuki & Hirst, 1999; Van Helden et al. 2000). The discharge of regenerative potentials appears to rely on the ability of ICC_IM to electrically entrain neighbouring ICC_IM so that many are activated together (Van Helden & Imtiaz, 2003). In the fundus, although there is no myenteric network of pacemaker ICC (ICC_MY; Burns et al. 1996), both the longitudinal and circular layers contain numerous ICC_IM that produce unitary potentials as in the antrum (Burns et al. 1996; Beckett et al. 2002). It was not clear why this population of ICC_IM could not sustain an irregular discharge of rhythmical activity analogous to that detected in isolated circular layers of the pylorus and antrum. The present experimental observations provide an explanation for this difference since fundal ICC_IM did not appear to be electrically excitable (Fig. 3). ICC_IM within other regions of the gastrointestinal tract may also be electrically unexcitable. For example, in the small intestine ICC_MY generate pacemaker potentials (Kito & Suzuki, 2003). However in the small intestine of mutant mice which lack ICC_MY but retain ICC_IM, rhythmical activity is not detected (Ward et al. 1994, 1995). This might suggest that the ICC distributed amongst the intestinal smooth muscle cells also lack voltage sensitivity. Similarly in the dog colon, although recordings have not been made from identified pacemaking ICC, the slow waves recorded from smooth muscle cells immediately adjacent to the pacemaking region bear a remarkable similarity to pacemaker potentials recorded from identified pacemaker ICC in other regions of the gastrointestinal tract (Ward et al. 1991; Dickens et al. 1999; Kito & Suzuki, 2003). Furthermore when colonic slow waves are recorded at sites distant from the pacemaker region there is a decrement in the amplitude of the signal (Ward et al. 1991), this would suggest that colonic ICC_IM also lack voltage sensitivity and are unable to augment the waves of pacemaker depolarization which passively propagate through the muscle layers. Finally it is of interest to note that the voltage sensitivity of antral ICC_IM can be abolished by treating preparations with n-ethylamide but when this is done a resting discharge of membrane noise continues to occur (Hirst et al. 2002b).

The discharge of membrane noise in antral muscles results from the random generation of unitary potentials by ICC_IM. These normally occur at a sufficiently high frequency so that individual unitary potentials cannot be resolved. However, when internal calcium ions are buffered by treating tissues with BAPTA-AM, the frequency of unitary potentials is reduced and individual events are readily resolved (Edwards et al. 1999). Similar observations were made on preparations of fundus muscle; buffering [Ca²⁺]_i resulted in a greatly reduced frequency of unitary potential discharge (Fig. 6). Given that this was associated with hyperpolarization in several preparations, the simplest explanation is that in the fundus the continuous discharge of unitary potentials by ICC_IM results in net inward current. The similar importance of mitochondrial function in the generation of unitary potentials in both antrum (Kito & Suzuki, 2003) and fundus was shown by the ability of mitochondrial Ca²⁺ uptake inhibitors to abolish the discharge of membrane noise (Fig. 7). Together these observations suggest that unitary potentials, and the discharge of membrane noise, reflect the internal handling of Ca²⁺ in both tissues. However it appears likely that the channels activated as a consequence of changes in [Ca²⁺]_i, differ between antral and fundal ICC_IM. Thus, whereas the discharge of membrane noise in antral ICC_IM is abolished by agents that block chloride channels (Hirst et al. 2002b), that generated by fundal ICC_IM is unchanged (Figs 5 and 8). However, as the power spectral density curves determined from both antral and fundal preparations have similar characteristics, it would appear likely that the time courses of individual unitary potentials in either tissue are very similar. Clearly the time courses of signals in both sets of ICC_IM are governed by similar rate limiting steps which are unrelated to the kinetics of the channels finally activated.

It seems that very similar spontaneous activity in the proximal and distal stomach subserves different functions. The ability of unitary potentials in the antrum to summate means that these events can amplify slow wave potentials generated by ICC_MY and potentially help to preserve the magnitude of these depolarization events as they conduct through the thickness of the muscle layers. The ability to summate also allows ICC_IM in the antrum to function as a secondary source of rhythmic activity. The role of unitary potentials in the fundus, which does not have the mechanism to produce slow waves and cannot summate unitary potentials, may be to regulate the level of membrane excitability. ICC_IM receive and transduce neural inputs in the stomach, and the ongoing discharge of unitary potentials in the fundus may precondition postjunctional cells to generate appropriate responses to excitatory and inhibitory neurotransmitters.

In summary these observations have shown that ICC_IM in the proximal and distal stomach have quite different properties. Those that are sensitive to voltage may contribute to rhythmical electrical activity either directly or by augmenting pacemaker signals generated by ICC_MY. Those that lack sensitivity to voltage presumably do not contribute to rhythmical electrical activity but continue to serve an essential function in the transmission of neural information from enteric motor nerves to the layers of smooth muscle making up the gut wall (Burns et al. 1996; Ward et al. 2000; Beckett et al. 2002, 2003; Suzuki et al. 2003).