Atypical slow waves generated in gastric corpus provide dominant pacemaker activity in guinea pig stomach

Abstract

When intracellular recordings were made from the circular layer of the intact muscular wall of the isolated guinea pig gastric corpus, an ongoing regular high frequency discharge of slow waves was detected even though this region lacked myenteric interstitial cells. When slow waves were recorded from preparations consisting of both the antrum and the corpus, slow waves of identical frequency, but with different shapes, were generated in the two regions. Corporal slow waves could be distinguished from antral slow waves by their time courses and amplitudes. Corporal slow waves, like antral slow waves, were abolished by buffering the internal concentration of calcium ions, [Ca²⁺]_i, to low levels, or by caffeine, 2-aminoethoxydiphenyl borate or the chloride channel blocker DIDS. Corporal preparations demonstrated an ongoing discharge of unitary potentials, as has been found in all other tissues containing interstitial cells. The experiments show that the corpus provides the dominant pacemaker activity which entrains activity in other regions of the stomach and it is suggested that this activity is generated by corporal intramuscular interstitial cells.

Gastrointestinal slow waves are generated by electrical activity arising in interstitial cells of Cajal (ICC) rather than by gastrointestinal smooth muscle cells (Sanders, 1996; Hirst & Ward, 2003). In most regions of the gastrointestinal tract, a myenteric network of ICC (ICC_MY) rhythmically generates large amplitude pacemaker potentials (Dickens et al. 1999; Kito & Suzuki, 2003) and slow waves are not detected when ICC_MY are absent (Ward et al. 1994; Ordog et al. 1999). In the small intestine and isolated gastric antrum, activity in ICC_MY often precedes activity in the nearby muscle layers (Yamazawa & Iino, 2002; Hennig et al. 2004). In the gastric antrum, pacemaker potentials passively depolarize the adjacent muscle layers (Cousins et al. 2003). These waves of depolarization are augmented by the secondary regenerative component in the slow wave, which is generated by intramuscular ICC (ICC_IM) (Dickens et al. 2001; Hirst et al. 2002a). Together these observations suggest that slow waves, throughout the gastrointestinal tract, are initiated by ICC_MY (Hirst & Ward, 2003). However, in the guinea pig, slow waves occur at lower frequencies in the isolated antrum (3 per minute, Tsugeno et al. 1995; Hirst & Edwards, 2001) than in the intact stomach (5 per minute, Hennig et al. 1999). Moreover in isolated antral preparations, successive pacemaker potentials are generated at irregular intervals (Hirst & Edwards, 2001) and radiate from different points in the network of ICC_MY (Hennig et al. 2004), so initiating slow waves at varying sites (Ward et al. 2004). In contrast, in the intact stomach, slow waves are initiated at regular intervals and propagate in an orderly anal direction from the corpus, along the antrum to the pylorus (Szurszewski, 1981; Sanders & Publicover, 1989).

Here we describe some of the properties of slow waves generated by the guinea pig corpus. Corporal slow waves occurred at regular intervals, with higher frequencies than those generated by the isolated antrum. Unlike other intact regions of the gastrointestinal tract, corporal slow waves were generated by tissues which lack ICC_MY. The observations support the view that the dominant pacemaker activity in the stomach originates in the corpus and suggest that it is generated by corporal ICC_IM rather than by gastric ICC_MY.

Methods

Electrophysiological methods

Procedures for the acquisition of physiological data from isolated tissues were approved by the Animal Experimentation Ethics Committee at the Australian National University. Guinea pigs of either sex were stunned, exsanguinated and the stomach removed. A region of stomach on either side of the greater curvature was isolated and immersed in oxygenated physiological saline (for composition see Suzuki & Hirst, 1999). In most experiments the fundus, antrum and pylorus were discarded; the mucosa and serosa were dissected away from the corpus. Half of the corpus, cut along the greater curvature, was pinned, serosal surface uppermost, in a recording chamber: the other half was fixed in ice-cold acetone and the distribution of ICC determined. Intracellular recordings were made as previously described (Hirst & Edwards, 2001). In some experiments, recordings were made from preparations containing the antrum and corpus; distributions of ICC were again determined in parallel preparations. The properties of corporal slow waves, generated in the circular layer, were analysed using isolated bundles of corporal circular muscle (see Suzuki & Hirst, 1999). Preparations were superfused with warmed physiological saline (35°C); nifedipine (1 μm) was added to reduce muscle movements.

Immunohistochemical methods

The distribution of ICC was determined in the paired preparations fixed in acetone for 15 min. Preparations were rinsed with phosphate-buffered saline (PBS) 3 times at 10 min intervals, left overnight in PBS at 4°C, washed with PBS containing 1% bovine serum albumin for 1 h at room temperature and then incubated with primary antibody (ACK2, Chemicon International rat anti-mouse CD117 monoclonal antibody, diluted 1 : 500 in PBS with 0.5% Triton X-100) at 4°C for 2 days. Preparations were washed for 24 h in PBS and incubated for 2 h with secondary antibody (Alexa fluor 488 goat anti-rat, Molecular Probes), diluted 1 : 500 in PBS. After rewashing overnight, preparations were mounted in a glycerol medium and viewed using a confocal microscope, illumination wavelength 488 nm, emission wavelength above 505 nm. Final images were prepared from z-stacks of part or the entire muscle wall.

All data are expressed as means ± standard error of the mean (s.e.m.). Student's t test was used to determine if data sets differed. Differences with a P value < 0.05 were taken to be significant. 2-Aminoethoxydiphenyl borate (2APB) (Calbiochem, San Diego, CA, USA), caffeine, the acetoxymethyl ester of bis-(2-amino-5-phenoxy)ethane-N,N,N′,N′-tetraacetic acid (MAPTA-AM), bis-(aminophenoxy)ethane-N,N,N,N′-tetraacetic acid tetra (acetoxymethyl)ester (BAPTA/AM), N-ethylmaleimide, 4,4′-diisothiocyano-2,2′-stilbene disulphonic acid (DIDS), and nifedipine (Sigma Chemical Co., St Louis, MO, USA) were used in these experiments.

Results

Properties of slow waves recorded from the isolated corpus

When intracellular recordings were made from isolated sheets of corporal muscle, slow waves were recorded (Fig. 1A). Corporal slow waves had peak amplitudes ranging from 5.3 to 26.8 mV (mean ± s.e.m., 16.1 ± 4.6 mV, n = 6, where each n value represents a measurement from a separate animal) with a maximum rate of rise, dV/dt_max, ranging from 5 to 16 mV s⁻¹ (9.8 ± 4.6 mV s⁻¹; Fig. 1B) and were superimposed on peak negative potentials ranging from −44 to −56 mV (−50.0 ± 1.9 mV). Slow waves had half-widths (time above 50% peak amplitude) ranging from 4.13 to 5.95 s (4.78 ± 0.26 s). The frequency of corporal slow waves ranged from 4.9 to 5.3 waves min⁻¹ (5.1 ± 0.2 waves min⁻¹; Fig. 1C): the intervals between successive corporal slow waves showed little variation (Fig. 1C). In the paired preparations stained for CD117, ICC_IM were detected in the longitudinal (Fig. 1Da) and circular layers (Fig. 1Db) but ICC_MY were not detected in any plane of focus (Fig. 1Dc). These observations show that the corpus, a region which lacks ICC_MY, generates a regular ongoing discharge of slow waves.

Figure 1. Properties of corporal slow waves.

The upper trace (A) shows recordings of corporal slow waves, which were superimposed on a peak negative potential of −53 mV. The upstrokes of corporal slow waves had low rates of membrane potential change with dV/dt_max being about 10 mV s⁻¹ (B). Corporal slow waves were generated at regular intervals at a frequency of about 5 waves min⁻¹ (C); time calibration bar applies to all traces. The micrographs show the distribution of interstitial cells. ICC_IM were detected in the longitudinal (Da) and circular layers (Db) but ICC_MY were not detected in the entire wall of the corpus (Dc); calibration bar in Dc applies to all micrographs.

Comparisons between slow waves generated in the antrum and corpus

Characteristic corporal slow waves were recorded from the corporal end of antral–corporal preparations (Fig. 2Aa) with peak amplitudes ranging from 6.1 to 20.1 mV (15.0 ± 5.4 mV, n = 5). Corporal slow waves lacked primary components (Fig. 2Ba), had a dV/dt_max ranging from 3 to 14 mV s⁻¹ (9.4 ± 1.9 mV s⁻¹; Fig. 2Ab and Bb) and were superimposed on peak negative potentials ranging from −40 to −56 mV (−51.0 ± 3.0 mV); the peak negative potential recorded between each corporal slow wave varied more than between antral slow waves (Fig. 2Aa and Ba). The frequency of corporal slow waves ranged from 4.1 to 5.0 waves s⁻¹ (4.6 ± 0.1 waves s⁻¹; Fig. 2D). When the corporal ends of the preparations were examined, again only ICC_IM were detected (Fig. 2C). Using unpaired t tests, the amplitudes, maximum rates of rise, peak negative potentials and rates of discharge of slow waves recorded from the corpus in preparations where the antrum remained attached were found not to differ significantly from those recorded in the isolated corpus.

Figure 2. Comparison between corporal and antral slow waves.

The upper left pair of traces shows slow waves (Aa) and their dV/dt (Ab) recorded at the corporal end of a corporal–antral preparation. The upper central pair of traces shows a slow wave (Ba) and associated dV/dt (Bb) displayed with expanded time base; the peak negative potential was −49 mV. The upper right micrograph shows a full wall thickness view of interstitial cells present at the corporal end (C). The lower left pair of traces show slow waves (Ea) and their dV/dt (Eb) recorded at the antral end of the same corporal–antral preparation. The lower central pair of traces shows one of these slow waves with initial component (Fa) and associated dV/dt (Fb) on an expanded time base; peak negative potential was −67 mV. The lower right micrograph shows a full wall thickness view of interstitial cells present at the antral end (G); note that ICC_MY were readily detected in the antrum. Slow waves were generated at the same rate (D) in the corpus (filled circles) and antrum (open circles). All observations were from the same preparation. The calibration bar in G applies to both micrographs.

Slow waves were also recorded in the circular layer at the antral end of antral–corporal preparations (Fig. 2Ea). Antral slow waves had peak amplitudes ranging from 24.9 to 33.1 mV (27.7 ± 1.7 mV, n = 5) and, when recorded near the greater curvature, had prominent initial components followed by secondary components (Fig. 2Fa). Antral slow waves had a dV/dt_max ranging from 18 to 25 mV s⁻¹ (20.1 ± 1.3 mV s⁻¹; Fig. 2Eb and Fb). Slow waves were superimposed on peak negative potentials ranging from −61 to −74 mV (−66.6 ± 2.4 mV) with frequencies ranging from 4.1 to 5.1 waves min⁻¹ (4.6 ± 0.2 waves min⁻¹; Fig. 2D). Micrographs showed that the antral end of the preparations contained a network of ICC_MY, with ICC_IM distributed in the circular layer (Fig. 2G). Using Student's paired t tests, the frequencies of slow waves at the antral and corporal ends were not found to be significantly different whereas the amplitudes, peak negative potentials and maximum rates of rise of antral and corporal slow waves all differed significantly. These observations show that antral and corporal slow waves occur at the same frequency when the corpus and antrum are attached but that corporal and antral slow waves differ.

Properties of slow waves recorded from bundles of circular muscle isolated from the corpus

The properties of corporal slow waves were examined using short segments of isolated circular layer. Segments generated an ongoing discharge of slow waves (Fig. 3). In the experiment illustrated in Fig. 3, recordings were made successively from an intact sheet of corpus (Fig. 3A) and from a bundle of circular muscle isolated from that sheet (Fig. 3B). Both preparations generated slow waves of similar amplitude but those recorded from the bundle were briefer and discharges of membrane noise were apparent (Fig. 3A and B). Slow waves, recorded from isolated bundles of the corporal circular layer, had frequencies ranging from 3.5 to 6.1 waves min⁻¹ (4.8 ± 0.2 waves min⁻¹; n = 16), peak amplitudes of 6.1 to 31.4 mV (15.9 ± 2.1 mV), half-widths ranging from 2.52 to 3.95 s (3.23 ± 0.14 s) and peak negative potentials of −42 to −58 mV (−49.8 ± 1.3 mV). Using Student's unpaired t tests, the frequencies, amplitudes and peak negative potentials of slow waves recorded from isolated circular bundles or sheets of corpus were not found to be significantly different whereas their half-widths differed significantly. This indicates that an isolated bundle of the circular layer of the corpus is capable of generating regular pacemaker activity, even in the absence of longitudinal muscle.

Figure 3. Properties of isolated bundles of corporal circular muscle.

Slow waves recorded from an intact sheet of corpus (A) are compared with those recorded from a bundle of circular muscle (B) isolated from the same preparation. Both preparations displayed ongoing discharges of slow waves which occurred with similar frequencies. The discharge of slow waves in the bundle (B) was abolished by applying 50 μm MAPTA-AM for 10 min (C). A power spectral density curve (open circles), calculated from the membrane potential recording obtained after treating with MAPTA-AM, shows the characteristic form (D), similar to those obtained from other tissues containing ICC_IM: the continuous line shows the best theoretical fit obtained when time constant A had a value of 425 ms and B had a value of 75 ms (Edwards et al. 1999). The peak negative potentials in A, B and C were −50, −48 and −48 mV, respectively. The time and voltage calibration bars apply to traces A–C.

Preliminary experiments on the pharmacological properties of corporal slow waves suggest that they share many of the properties of responses generated by ICC_IM in the circular layer of the antrum. These depend upon the release of Ca²⁺ from internal stores (Edwards et al. 1999). Corporal slow waves were abolished similarly by either BAPTA-AM (50 μm for 15 min; n = 4) or MAPTA-AM (50 μm for 15 min; n = 2; Fig. 3C). The membrane potential settled near the peak negative potential detected between slow waves and displayed discharges of membrane noise (Fig. 3C). Power spectral density curves, determined from the membrane potential recordings obtained either before or after treatment with a calcium-chelating agent, all had a characteristic shape. Power increased from low values at around 10 Hz and approached a plateau value near 1 Hz. Power spectral density curves determined from tissues containing ICC_IM can be described by assuming that membrane noise is made up of unitary potentials with the form of the curves being determined by the time courses of individual unitary potentials (Edwards et al. 1999). Using the same approach, good fits were obtained with constants A and B having values ranging from 385 to 560 ms (455 ± 30 ms; n = 5) and from 55 to 80 ms (70 ± 5 ms), respectively; similar values have been determined in other gastric preparations containing ICC_IM (Edwards et al. 1999; Dickens et al. 2001; Beckett et al. 2002). Corporal slow waves, like antral responses, were abolished by 1 mm caffeine (Fig. 4A and B; n = 5), 50 μm 2APB (Fig. 4C and D; n = 3) and 100 μm DIDS (Fig. 4E and F; n = 3); each of these effects was reversed by washing with drug-free solution.

Figure 4. Pharmacological properties of slow waves recorded from isolated bundles of the corpus circular layer.

Corporal slow waves were abolished by 1 mm caffeine (A and B), by 50 μm 2APB (C and D) or by 100 μm DIDS (E and F). The time and voltage calibration bars apply to all traces.

Discussion

These experiments show that the corpus, a region of the stomach which lacks ICC_MY, generates slow waves. The frequency of corporal slow waves was higher than that detected in the isolated guinea pig gastric antrum (Tsugeno et al. 1995; Hirst & Edwards, 2001) but the same as the frequency of gastric contractions detected in the intact guinea pig stomach after abolishing nerve activity (Hennig et al. 1999). When the corpus was left in continuity with the antrum, the frequency of slow waves generated in the antrum was the same as that of those generated in the corpus, indicating that the dominant pacemaker frequency of the intact stomach is set by the corpus. Clearly corporal pacemaker activity could be generated either by corporal smooth muscle cells or by corporal ICC_IM but the preliminary observations made on corporal slow waves suggest that the cellular mechanisms involved in their generation resemble those found to generate rhythmic activity in ICC elsewhere.

Isolated preparations of corpus generated slow waves with a frequency of about 5 waves min⁻¹ (Fig. 1). In the fundus, where ICC_MY are also absent, ICC_IM are present in both layers but, as these lack voltage sensitivity, spontaneous rhythmical activity is not detected (Beckett et al. 2004). Although the arrangement of corporal ICC resembles that of fundal ICC (Burns et al. 1997; Fig. 1), the corpus generates a regular discharge of slow waves, suggesting that the properties of fundal and corporal ICC_IM differ. Many previous reports have indicated that slow waves, recorded from the isolated antrum, occur at lower frequencies (Tsugeno et al. 1995; Hirst & Edwards, 2001) than those found in the corpus. Furthermore in the antrum, the interval between slow waves shows considerable variation (Hirst & Edwards, 2001); in the corpus such variation was not detected (Figs 1 and 2). When the corpus and antrum were left in continuity, corporal and antral slow waves occurred at the same frequency and the intervals between antral slow waves showed little variation (Fig. 2), indicating that corporal slow waves drive antral slow waves. When the shapes of antral and corporal slow waves were compared, antral slow waves had initial and secondary components whereas those generated in the corpus lacked an initial component (Fig. 2; see also Komori & Suzuki, 1986). The initial component of the antral slow wave results from passive depolarization of the antral circular layer by pacemaker potentials generated in ICC_MY and the secondary component results from the slow voltage-dependent activation of ICC_IM (Edwards & Hirst, 2005). The observation that antral slow waves, recorded from preparations attached to the corpus, have two components suggests that corporal slow waves trigger pacemaker potentials in antral ICC_MY, presumably at the interface between corpus and antrum, where ICC_MY are first detected in low numbers (authors' unpublished observations). The ability of altered patterns of activity in antral circular muscles to increase the frequency of antral pacemaker potentials has been demonstrated previously: cholinergic nerve stimulation increases the frequency of regenerative responses generated in the antral circular layer, by ICC_IM, such that dominant pacemaker activity shifts to that layer and entrains ICC_MY (Hirst et al. 2002c).

Corporal slow waves shared many of the properties of regenerative responses generated by antral ICC_IM except that they occurred with a higher frequency and greater regularity. In the isolated antral circular layer, regenerative responses are generated irregularly at a frequency of 1 wave min⁻¹ (Kito et al. 2002); in the corpus slow waves are generated regularly at about 5 waves min⁻¹ (Fig. 1). These observations provide a further illustration that the properties of gastric ICC_IM vary with location, with antral ICC_IM lacking the ability to generate regular high frequency discharges of coordinated activity. Nevertheless in both the antral and corporal circular layers, rhythmical activity was abolished by buffering [Ca²⁺]_i to low levels with calcium-chelating agents or by low concentrations of caffeine (Suzuki & Hirst, 1999; Figs 3 and 4). In both preparations discharges of unitary potentials, with similar power spectral density curves, are detected (Edwards et al. 1999; Fig. 3). Blocking chloride-selective channels abolishes antral regenerative responses (Hirst et al. 2002b): rhythmical activity in the corpus was also abolished by DIDS (Fig. 4). Antral slow waves involve the release of Ca²⁺ from inositol 1,4,5-trisphosphate-dependent stores (Suzuki et al. 2000) and are abolished by 2APB (Hirst & Edwards, 2001; Kito et al. 2002); similarly corporal slow waves were abolished by 2APB (Fig. 4). In the antrum, regenerative responses are only detected in tissues which contain ICC_IM (Dickens et al. 2001; Hirst et al. 2002a; Suzuki et al. 2003). Given the pharmacological similarities, it seems likely that corporal slow waves are generated by corporal ICC_IM rather than by corporal smooth muscle cells.

Many isolated regions of the stomach generate rhythmical electrical activity which occurs at different frequencies. The isolated corpus generates a regular discharge of slow waves with a frequency of 4.5 waves min⁻¹. The isolated antrum generates slow waves with a frequency of 3 waves min⁻¹ and the isolated circular layer of the antrum generates rhythmical activity with a frequency of 1–2 waves min⁻¹. Furthermore it has been shown that the ability of gastric ICC_MY to generate a dominant pacemaker frequency varies with region (Ordog et al. 2002). When all of these regions are coupled together in the intact stomach, activity in all regions occurs at the same higher frequency. Thus the corpus must entrain antral ICC_MY and these in turn entrain activity within antral ICC_IM. It is not clear why such a duplication of potential pacemaker sites exists within the stomach wall. The complexity and apparent duplication of primary, secondary and tertiary pacemaker sites within the stomach raises the possibility that in clinical disorders, if the dominant pacemaker site is shifted or damaged, organized movements of stomach contents may well be disrupted.