Spontaneous electrical activity and associated changes in calcium concentration in guinea-pig gastric smooth muscle

Abstract

Spontaneous electrical activity and internal Ca²⁺ concentration ([Ca²⁺]_i) were measured simultaneously using conventional microelectrodes and fura-2 fluorescence, respectively, in isolated circular smooth muscle bundles of the guinea-pig gastric antrum. The smooth muscle bundles generated periodic slow potentials with accompanying spike potentials and associated transient increases in [Ca²⁺]_i (Ca²⁺-transients). Nifedipine abolished the spike potentials but not the slow potentials, and reduced the amplitude of associated Ca²⁺-transients. Caffeine, in the absence or presence of ryanodine, reduced resting [Ca²⁺]_i levels and abolished the slow potentials and associated Ca²⁺-transients. Depolarization elevated and hyperpolarization reduced resting [Ca²⁺]_i levels with associated changes in the frequency of slow potentials. The amplitude of Ca²⁺-transients changed in a bell-shaped manner with the membrane potential change. Slow potentials and associated Ca²⁺-transients were abolished if [Ca²⁺]_i levels were reduced by BAPTA-AM or if the internal Ca²⁺ pump was inhibited by cyclopiazonic acid. 2-Aminoethoxy-diphenylborate (2-APB), a known inhibitor of inositol trisphosphate (IP₃)-mediated Ca²⁺ release, also blocked slow potentials and Ca²⁺-transients. Carbonyl cyanide m-chlorophenyl hydrazone (CCCP), a mitochondrial protonophore, depolarized the membrane, elevated [Ca²⁺]_i levels and abolished slow potentials and Ca²⁺-transients. Inhibition of mitochondrial ATP-sensitive K⁺ channels by glybenclamide and 5-hydroxydecanoic acid (5-HAD) abolished slow potentials and Ca²⁺-transients, without altering the smooth muscle [Ca²⁺]_i. It is concluded that in antrum circular muscles, the frequency of slow potentials is correlated with the level of [Ca²⁺]_i. The slow potential is coupled to release of Ca²⁺ from an internal store, possibly through the activation of IP₃ receptors; this may be initiated by the activation of ATP-sensitive K⁺ channels in mitochondria following Ca²⁺ handling by mitochondria.

Rhythmical mechanical movements of the stomach are associated with membrane potential changes in smooth muscle cells in the form of either slow waves, spike potentials or both (Tomita, 1981). Interstitial cells of Cajal (ICC), distributed in the myenteric region of the gastric wall, have been suggested to initiate pacemaker activity, with the electrical signals generated by ICC propagating passively to the smooth muscle cells through gap junctions (Thuneberg, 1982, 1989; Ward et al. 1994; Sanders, 1996; Huizinga et al. 1997). The functional connections between ICC and smooth muscle cells have been demonstrated in the gastric antrum of guinea-pig. There, three types of cells (circular and longitudinal smooth muscle cells and ICC) show synchronized periodical excitation (Dickens et al. 1999). In isolated circular muscles from the guinea-pig gastric antrum, however, spontaneous regenerative potentials with slow time course are also generated (Suzuki & Hirst, 1999). The regenerative potential can be inhibited by relatively low concentrations of caffeine, and it is suggested that this potential may be the second component of the slow wave proposed by Ohba et al. (1975) in the guinea-pig stomach (Dickens et al. 2001).

The ionic mechanisms underlying the generation of discharges of slow waves in gastric smooth muscles remain unclear. In canine stomach, slow waves are associated with an elevation of [Ca²⁺]_i, and nicardipine, which blocks voltage-gated L-type Ca²⁺ channels (Mori et al. 1996), inhibits the plateau component of slow waves and associated [Ca²⁺]_i elevation (Ozaki et al. 1991), suggesting that influx of Ca²⁺ through voltage-gated L-type Ca²⁺ channels forms the plateau potential. However in gastric muscles of many laboratory animals, organic Ca²⁺ antagonists such as verapamil (Golenhofen & Lammel, 1972), diltiazem (Ishikawa et al. 1985) and nifedipine (Dickens et al. 1999) inhibit the spike potentials but not slow waves. In circular smooth muscle tissues of the guinea-pig stomach, nifedipine inhibits spike potentials but not regenerative slow potentials (Suzuki & Hirst, 1999; Suzuki, 2000). This indicates that the ionic mechanisms responsible for the generation of spike potentials differ from those involved in the generation of slow waves or regenerative potentials, with only the former being produced by activation of voltage-gated L-type Ca²⁺ channels. Slow waves are sensitive to temperature changes and have a high Q₁₀ value (Tomita, 1981) and are inhibited by metabolic inhibitors (Nakayama et al. 1997). Both observations suggest that the membrane events underlying slow waves are in some way coupled to the metabolic activity of cells participating in their initiation. Although the frequency and amplitude of slow waves (Huang et al. 1999) or regenerative slow potentials (Nose et al. 2000) are voltage-dependent, the release of Ca²⁺ from the internal stores following the activation of inositol 1,4,5-trisphosphate (IP₃) receptors is also thought to be involved in their generation (Suzuki & Hirst, 1999; Edwards et al. 1999; Van Helden et al. 2000; Ward et al. 2000). This is indeed the case in the mouse where the stomachs of mutant mice lacking IP₃ receptors fail to generate slow waves (Suzuki et al. 2000).

The present experiments were carried out to investigate the relationship between regenerative slow potentials and changes in [Ca²⁺]_i in circular smooth muscles isolated from the antrum region of guinea-pig stomach, since changes in smooth muscle [Ca²⁺]_i may be one of the important factors to regulate spontaneous activity of gastric muscles, as in the case of mouse intestine (Ward et al. 2000). Simultaneous measurements of the changes in membrane potential and [Ca²⁺]_i were carried out using intracellular microelectrodes and fura-2 fluorescence, respectively. The results indicated that there were nifedipine-sensitive and -insensitive components in the regenerative slow potential-mediated increase in [Ca²⁺]_i. The level of [Ca²⁺]_i was considered to be related to the pacemaking mechanisms of the spontaneous activity, and a possible involvement of mitochondrial ATP-sensitive K⁺ channels linked with uptake of Ca²⁺ into mitochondria and the release of Ca²⁺ from the internal store through activation of IP₃ receptors is suggested. Some of these observations were presented at the 76^th Annual Meeting of the Japanese Physiological Society (Fukuta & Suzuki, 1999).

METHODS

Albino guinea-pigs of either sex, weighting 250–300 g, were anaesthetized with ether, and exsanguinated from the femoral artery. All animals were treated ethically according to the guiding principles for the care and use of animals in the field of physiological sciences, approved by the Physiological Society of Japan. The stomach was excised, and opened by cutting along the small curvature in Krebs solution. The mucosal layers were removed by cutting with fine scissors, and smooth muscle tissues were isolated from the antrum region. The circular tissue preparation (single bundle with about 150 μm width and 1–2 mm long) was prepared by mechanical removal of all of the longitudinal muscle layers with fine forceps, using a binocular microscope. Such preparations have previously been shown to contain intramuscular ICC (IC-IM) but be devoid of myenteric ICC (IC-MY) (Suzuki & Hirst, 1999). Preparations were pinned out (diameter of pins, 30 μm) on a Sylgard plate (silicone elastomer, Dow Corning Corporation, Midland, MI, USA) at the bottom of the recording chamber (volume, approximately 0.5 ml), usually with the mucosal side uppermost. The recording chamber was mounted on a stage of an inverted microscope (Olympus IX 70, Tokyo, Japan), and the preparation was superfused with warmed (35 °C) Krebs solution at a constant flow rate (2 ml min⁻¹).

After about 1 h incubation of the tissue with warmed (35 °C) Krebs solution, spontaneous activity was confirmed using an intracellular microelectrode. Active preparations were loaded with fluorescent dye (fura-2 AM, 10 μm) for 1 h at room temperature. After the loading, preparations were superfused again with dye-free, warmed (35 °C) Krebs solution for 30 min, at a constant flow (2 ml min⁻¹). Preparations loaded with fura-2 were illuminated with ultraviolet light, using two wavelengths of light − 340 nm and 380 nm - alternating at a frequency above 40 Hz. The ratio of the emission fluorescence (R_F340/F380) was measured from the central region of the tissue, in a square field fitted to the width of the preparation (approximately 150 × 150 μm), through a barrier filter of 510 nm, using the Arugus Hisca system (Hamamatsu Photonics, Hamamatsu, Japan). The intensity of fluorescence was expressed by this ratio. The concentration of Ca²⁺ was calibrated from the intensity of fluorescence according to the in vitro calibration methods using the Calcium Calibration Buffer Kits (Molecular Probes Inc., OR, USA). The determined [Ca²⁺]_i was 46.72 nm at R_F340/380 = 0.50 and 175.86 nm at R_F340/F380 = 0.80, and the relationship was linear between these ratios of fluorescence.

Electrical responses of circular smooth muscle cells were recorded using conventional microelectrodes filled with 0.5 m KCl (tip resistance, 150–250 MΩ). Cells were impaled in the area used to record fluorescence responses of the tissue. Membrane potential changes, recorded using a high input impedance amplifier (Axoclamp-2B, Axon Instruments, Inc., Foster City, CA, USA), were displayed on a cathode-ray oscilloscope (SS-7602, Iwatsu, Osaka, Japan) and stored on a personal computer for later analysis.

Smooth muscle tissues were stimulated extracellularly using a suction electrode method originally applied to small vessels (Zhang et al. 1989). Briefly, rectangular thin strips, 0.05–0.1 mm wide (which contained 1–2 circular muscle bundles) and about 5 mm long, were prepared from circular tissues and one end of the tissue was drawn into a glass capillary suction electrode (tip diameter, 0.08–0.1 mm). The tissues left outside of the electrode (1–1.5 mm long) were immobilized on the Sylgard plate using fine pins (diameter, 30 μm). Microelectrodes were inserted in the immobilized part of the tissue (0.2–0.3 mm distance from the edge of the electrode) to record membrane potential changes in smooth muscle cells. Fluorescence responses were also recorded from the immobilized part of the tissue. Rectangular electrical stimuli were applied through the suction electrode using an electric stimulator (SEN-8101, Nihon-Kohden, Tokyo, Japan). This method allowed the membrane potentials to be displaced by about ± 20 mV from the resting level during current injection (Nose et al. 2000).

The Krebs solution had the following ionic composition (mm): Na⁺ 137.4, K⁺ 5.9, HCO₃⁻ 11.5, Mg²⁺ 2.5, Ca²⁺ 2.5, H₂PO₄⁻ 1.2, Cl⁻ 134, glucose 11.5. The solution was aerated with O₂ containing 5 % CO₂, and the pH of the solution was maintained at 7.2–7.3. Krebs solutions with high potassium ion concentration (high-K⁺ solution) were prepared by replacing Na⁺ with K⁺.

Drugs used were caffeine, carbonyl cyanide m-chlorophenyl hydrazone (CCCP), cyclopiazonic acid (CPA), glibenclamide, 5-hydroxydecanoic acid (5-HDA), nifedipine (all from Sigma, St Louis, MO, USA), fura-2 AM (Dojindo, Osaka, Japan), ryanodine (Calbiochem, San Diego, CA, USA), 2-aminoethoxy-diphenylborate (2-APB) (Ono Pharmac. Co., Japan) and HMR-1098 (1-[[5-[2-(-5-chloro-o-anisamide)ethyl]-2-methoxyphenyl]sulphonyl]-3-methylthiourea sodium salt, Aventis Pharma Japan, Tokyo, Japan). Ryanodine and HMR-1098 were dissolved in distilled water, while 2-APB, CCCP, CPA, fura-2 and nifedipine were dissolved in dimethyl sulphoxide (DMSO), to make stock solutions, before adding to Krebs solution to make the desired concentrations. The final concentration of the solvents in Krebs solution did not exceed 1/1000. Caffeine was dissolved directly in Krebs solution. Addition of these chemicals to Krebs solution did not alter the pH of the solution. Preliminary experiments were carried out to check if these chemicals altered the fura-2 fluorescence; the observations showed that they produced negligible changes in fluorescence.

Experimental values measured were expressed as the mean ± standard deviation (s.d.), with the n value representing the number of tissues obtained from different animals. Statistical significance was tested using Student's paired t test, and probabilities less than 5 % (P < 0.05) were considered significant.

RESULTS

Spontaneous electrical and calcium responses in circular smooth muscle bundles

In single bundles of the circular smooth muscle, electrical responses and changes in [Ca²⁺]_i were measured simultaneously using intracellular microelectrodes and fura-2 fluorescence, respectively. Smooth muscle cells showed a periodical generation of regenerative slow potentials, with amplitudes of 20–38 mV (mean, 30.8 ± 5.4 mV, n = 45) at intervals between peaks of 21–61 s (mean, 30.8 ± 7.8 s, n = 45), as reported previously (Suzuki & Hirst, 1999; Nose et al. 2000). A burst of spike potentials was generated on top of each slow potential. The membrane potential between slow potentials was stable, apart from being interrupted by the random generation of unitary potentials (Edwards et al. 1999): the value varied between −59 mV and −70 mV (mean, −63.9 ± 4.2 mV, n = 45). The resting level of [Ca²⁺]_i was 0.72 ± 0.09 R_F340/F380 (n = 45), and each slow potential was accompanied by a transient elevation of [Ca²⁺]_i (Ca²⁺-transient) with the amplitude ranging between 0.075 and 0.297 R_F340/F380 (mean, 0.17 ± 0.02 R_F340/F380, n = 45) (Fig. 1A). Superimposed recordings (Fig. 1C) indicated that the slow potentials started prior to the elevation of [Ca²⁺]_i, with a variable delay in the range 140–980 ms (mean, 595 ± 95 ms, n = 45).

Figure 1. Spontaneous electrical and Ca²⁺ responses of antral circular muscle.

Electrical (A) and calcium responses (B) were recorded simultaneously from single circular muscle bundles of the guinea-pig gastric antrum, in the absence (a) and presence (b) of 1 μm nifedipine. C, superimposed fast traces of electrical (continuous line) and calcium responses (dotted line) recorded simultaneously from the same tissue. The resting membrane potential was: Aa, −65 mV; Ab, −65 mV. The resting [Ca²⁺]_i was: Ba, 0.60 R_F340/F380; Bb, 0.58 R_F340/F380. All responses were recorded from the same tissue.

In the presence of nifedipine (1 μm), the amplitude of Ca²⁺-transient was reduced from 0.181 ± 0.021 R_F340/F380 to 0.075 ± 0.028 R_F340/F380 (n = 30, P < 0.05), with an associated reduction in its duration by about one-third (width at half amplitude: control, 5.70 ± 1.21 s; in nifedipine, 3.68 ± 1.31 s; n = 30; P < 0.05). Nifedipine also reduced the resting [Ca²⁺]_i level from 0.75 ± 0.10 R_F340/F380 to 0.64 ± 0.10 R_F340/F380 (n = 30, P < 0.05), suggesting that some L-type Ca²⁺ channels are active at the resting levels. The reduction of the Ca²⁺-transient by nifedipine was associated with an abolition of spike potentials, but an ongoing generation of slow potentials with mean amplitude of 28.5 ± 4.8 mV continued (n = 30; Fig. 1B). In three experiments, increasing the concentration of nifedipine to 10 μm, after exposure of the tissue to 1 μm nifedipine for 2–3 h, did not result in a significant alteration of the resting [Ca²⁺]_i levels (in 1 μm nifedipine, 0.65 ± 0.05 R_F340F380; in 10 μm nifedipine, 0.64 ± 0.06 R_F340/F380; n = 3; P > 0.05) and amplitude of Ca²⁺-transients (relative amplitude, 90 ± 21 %). Electrical responses of the membrane (membrane potential, amplitude of slow potentials) also remained unaltered by this increase in concentration of nifedipine (data not shown). Thus, each Ca²⁺-transient consisted of nifedipine-sensitive and -insensitive components, which could be related to spike potentials and slow potentials, respectively.

Effects of caffeine on slow potentials and Ca²⁺-transients

Experiments were carried out to investigate the effects of 1 mm caffeine on regenerative slow potentials and Ca²⁺-transients in circular smooth muscle of the guinea-pig stomach antrum, in the presence of 1 μm nifedipine. Addition of caffeine to the Krebs solution hyperpolarized the membrane by 1–5 mV (mean, 3.1 ± 0.3 mV, n = 15, P < 0.05), reduced the resting [Ca²⁺]_i level from 0.67 ± 0.02 R_F340/F380 to 0.65 ± 0.02 R_F340/F380 (n = 15, P < 0.05) and abolished both slow potentials and Ca²⁺-transients (Fig. 2). The recovery from the inhibition by caffeine was very rapid, requiring some 2–3 min. Ryanodine (10 μm) did not alter the amplitude and frequency of slow potentials and Ca²⁺-transients, and caffeine applied in the presence of ryanodine again inhibited the electrical and calcium responses (Fig. 2A and B). In separate experiments, ryanodine was applied for up to 80 min, and no detectable alteration was observed in the amplitude and frequency of slow potentials, Ca²⁺-transients and the responses to caffeine (n = 3, data not shown). A higher concentration (30 μm) of ryanodine was also applied to two preparations. No significant effect was detected on the amplitude of slow potentials (control, 30.1 ± 3.5 mV; in ryanodine, 29.8 ± 3.3 mV; P > 0.05) and Ca²⁺-transients (control, 0.08 ± 0.03 R_F340/F380; in ryanodine, 0.08 ± 0.04 R_F340/F380; P > 0.05). These results indicate that the inhibition by caffeine of the slow potential and Ca²⁺-transient does not involve the release of Ca²⁺ from ryanodine-sensitive internal stores.

Figure 2. Effects of caffeine on electrical and Ca²⁺ responses.

Caffeine (1 mm) was applied for 3 min (shown by bar) during simultaneous recording of electrical (a) and calcium responses (b) from a single bundle of circular smooth muscle of the guinea-pig gastric antrum, in the absence (A) and presence (B) of 10 μm ryanodine. Thin horizontal lines indicate the resting membrane potential (Aa, −60 mV; Ba, −62 mV) and the resting [Ca²⁺]_i level (Ab, 0.60 R_F340/F380; Bb, 0.58 R_F340/F380). A and B were recorded from different tissues. Nifedipine (1 μm) was present throughout.

Intracellular Ca²⁺ level and slow potentials

Experiments were carried out to observe the effects of membrane potential and [Ca²⁺]_i on the frequency and amplitude of slow potentials and Ca²⁺-transients in circular smooth muscle of the guinea-pig stomach antrum. The membrane potential of smooth muscle cells was displaced to depolarized or hyperpolarized levels by electrical currents applied extracellularly using a suction electrode. The membrane was also depolarized by applying solutions containing elevated concentrations of potassium ions (high-K⁺ solution, [K⁺]_o = 8.3–24.7 mm). High-K⁺ solution containing 15.8 mm K⁺ depolarized the membrane by 15.0 ± 2.2 mV (n = 8), elevated the level of [Ca²⁺]_i by 0.03 ± 0.002 R_F340/F380 (n = 8) and reduced the time between peaks of slow potentials from 25.5 ± 6.5 s to 14.8 ± 3.8 s (n = 8; Fig. 3A). Similar changes in [Ca²⁺]_i were produced when the membrane was depolarized by 12 mV using a current injection (Fig. 3B). Hyperpolarizing the membrane by 10 mV with current injection, on the other hand, reduced the level of [Ca²⁺]_i by about 0.02 R_F340/F380. At the same time the amplitude of each Ca²⁺-transient was reduced by 30–60 % of its control value and the frequency of slow potentials and Ca²⁺-transients fell (Fig. 3C). Strong hyperpolarization (≥ −80 mV) reduced [Ca²⁺]_i below 0.55 R_F340/F380 and abolished slow potentials and Ca²⁺-transients (data not shown).

Figure 3. Electrical and Ca²⁺ responses and membrane potentials.

Electrical (a) and calcium responses (b) were recorded simultaneously from single bundles of circular smooth muscle of the guinea-pig gastric antrum, while the membrane was displaced to a 12 mV depolarized level by high-K⁺ solution ([K⁺]_o = 15.8 mm) (A) or to 12 mV depolarized (B) and 10 mV hyperpolarized levels (C) by current injections. Each set of responses was obtained from different tissues. In B and C, the upper straight lines indicate the current monitor, in which upwards and downwards deflections indicate depolarizing and hyperpolarizing current stimuli, respectively. The resting membrane potential was: Aa, −60 mV; Ba, −65 mV; Ca, −65 mV. The resting [Ca²⁺]_i level was: Ab, 0.62 R_F340/F380; Bb, 0.58 R_F340/F380; Cb, 0.60 R_F340/F380. Nifedipine (1 μm) was present throughout.

The relationship between amplitude of slow potentials and membrane potentials before the onset of each slow potential indicated that the amplitude increased and decreased with hyperpolarization and depolarization of the membrane, respectively (Fig. 4A). Membrane potentials were displaced to depolarized or hyperpolarized levels by current injection or to depolarized levels by applying various concentrations of high-K⁺ solution. The amplitude of slow potential was increased with hyperpolarization and decreased with depolarization, in a nearly linear manner. The calculated regression line using the least-square method was Y = −0.50X + 3.27 (Y, amplitude of slow potential; X, membrane potential). The time between the peaks of slow potentials increased with hyperpolarization and decreased with depolarization, and the relationship was again linear (data not shown), confirming the previous observation (Nose et al. 2000) that the frequency of the slow potential is a function of membrane potential.

Figure 4. The relationship between membrane potential, [Ca²⁺]_i level and amplitude of slow potential and Ca²⁺-transient.

A, the relationship between membrane potential (abscissa) and the amplitude of the slow potential (ordinate) recorded from single bundles of circular muscle of the guinea-pig gastric antrum. The membrane potentials were changed by high-K⁺ solution (•) or current injection (○). The regression line in the figure is given by Y = 3.27 − 0.50X (Y, amplitude of slow potential; X, membrane potential; r = 0.82, n = 265 from 14 tissues, P < 0.05). Nifedipine (1 μm) was present throughout. B, the relationship between amplitude of the Ca²⁺-transient (abscissa) and membrane potential (ordinate). The membrane potentials were changed by high-K⁺ solution (•) or current injection (○). The line in the figure was drawn by eye (n = 267 from 14 tissues). Nifedipine (1 μm) was present throughout. C, the relationship between [Ca²⁺]_i (abscissa) and membrane potential (ordinate). The membrane potentials were changed by high-K⁺ solution (•) or current injection (○). The regression line is given by Y = 0.68 + 0.002X (Y, [Ca²⁺]_i; X, membrane potential, r = 0.81, n = 269 from 14 tissues; P < 0.05). Nifedipine (1 μm) was present throughout. D, the relationship between [Ca²⁺]_i (ordinate) and the time (interval) between the peaks of slow potentials (abscissa). The levels of [Ca²⁺]_i were measured at different membrane potentials produced by high-K⁺ solutions (•) or current injections (○). The regression line was given by Y = 42.3 − 49.3X (Y, the time between the peak of slow potentials; X, [Ca²⁺]_i; r = 0.57, n = 429 from 14 tissues; P < 0.05). Nifedipine (1 μm) was present throughout. E, the relationships between [Ca²⁺]_i (ordinate) and the interval between the peaks of slow potentials produced by current injection measured from six different tissues (values obtained from individual tissues were identified by connected lines). F, the relationships between [Ca²⁺]_i (ordinate) and the interval between the peaks of slow potentials produced by increasing concentrations of potassium ions ([K⁺]_o = 5.9–18.8 mm) in eight different tissues (values obtained from individual tissues were connected by lines).

The relationship between the amplitudes of Ca²⁺-transients and membrane potential is shown in Fig. 4B. Depolarization of the membrane by either high-K⁺ solution or current injection reduced the amplitude of the Ca²⁺-transient. Hyperpolarization of the membrane by current injection again caused a reduction of the amplitude of Ca²⁺-transients. Thus the amplitude of Ca²⁺-transients was related in a bell-shaped way to membrane potential, with the largest Ca²⁺-transient being generated around the level of the resting membrane potential (−70 to −50 mV). When the resting level of [Ca²⁺]_i was plotted as a function of membrane potential, displaced by current injection, depolarization increased and hyperpolarization decreased [Ca²⁺]_i linearly. Similarly, depolarization of the membrane by high-K⁺ solutions also increased resting [Ca²⁺]_i linearly (Fig. 4C). The regression line of the relationship for current stimulation was Y = 0.68X + 0.002 (Y, [Ca²⁺]_i; X, membrane potential; r = 0.91), and this was similar to the relationship for high-K⁺ solution (Y = 0.70X + 0.006, r = 0.88). The time between the peak amplitude of slow potentials (i.e. interval) plotted as a function of resting [Ca²⁺]_i levels produced by high-K⁺ solution or current injection indicated a roughly reverse relationship, with a large diversity between 0.54 and 0.68 R_F340/F380 [Ca²⁺]_i (r = −0.57, P < 0.05; Fig. 4D). Attempts were made to plot the relationship between [Ca²⁺]_i and the time between peaks of slow waves (interval of slow potentials) in individual tissues, and the relationship was successfully measured at more than four different levels of [Ca²⁺]_i in six and eight tissues for current injection and high-K⁺ solutions, respectively. Displacement of membrane potentials by current injections clearly showed that the interval of slow potentials was decreased in parallel with the increase in [Ca²⁺]_i (Fig. 4E). Increase in [Ca²⁺]_i by high-K⁺ solutions consistently decreased the interval between slow potentials (Fig. 4F). Thus, the results showed that the frequency of regenerative slow potentials is increased by the elevation of [Ca²⁺]_i levels.

Chelating intracellular Ca²⁺ with BAPTA-AM inhibits slow potentials in circular smooth muscle of the antrum (Suzuki & Hirst, 1999). Experiments were carried out to observe the changes in electrical responses and [Ca²⁺]_i in the presence of 10 μm BAPTA-AM. Application of BAPTA-AM reduced [Ca²⁺]_i gradually with an associated depolarization of the membrane, which reached a stable value some 10–15 mV positive to the resting level after 10 min (control, −63.5 ± 3.5 mV; in BAPTA, −54.6 ± 2.1 mV; n = 9; P < 0.05). At this time the resting [Ca²⁺]_i was reduced from 0.70 ± 0.06 R_F340/F380 to 0.63 ± 0.05 R_F340/F380 (n = 9, P < 0.05). Slow potentials and Ca²⁺-transients were abolished by BAPTA-AM (Fig. 5).

Figure 5. Effects of BAPTA on electrical and Ca²⁺ responses.

Electrical (A) and calcium responses (B) were recorded simultaneously, during application of 10 μm BAPTA-AM (shown by the bar), from a single bundle of circular smooth muscle of the guinea-pig gastric antrum. Nifedipine (1 μm) was present throughout. The dotted lines indicate the resting membrane potential (−65 mV) and the resting [Ca²⁺]_i level (0.65 R_F340/F380).

Intracellular Ca²⁺ stores and slow potentials

The role of internal Ca²⁺ stores in the generation of regenerative slow potentials was examined by using cyclopiazonic acid (CPA), an agent which depletes internal Ca²⁺ stores by inhibiting Ca²⁺-ATPase on the membrane of the sarcoplasmic reticulum (Uyama et al. 1992), and 2-aminoethoxydiphenyl borate (2-APB), a known inhibitor of the inositol 1,4,5-trisphosphate (IP₃)-mediated Ca²⁺ release from internal stores (Cui & Kanno, 1997; Maruyama et al. 1997). All experiments were carried out in the presence of 1 μm nifedipine.

CPA (10 μm) depolarized the membrane and increased the resting [Ca²⁺]_i level. The effects developed slowly and reached stable values within 5–8 min. In the continued presence of CPA for over 10 min, the membrane stayed at depolarized level of 6–12 mV above the resting level (−55.8 ± 3.3 mV, n = 8, P < 0.05), and the slow potential was abolished (Fig. 6Aa). The level of [Ca²⁺]_i was also increased to a stable level of 0.67 ± 0.04 R_F340/F380 (n = 8, P < 0.05), within 10–15 min (Fig. 6Ab), and the Ca²⁺-transient was abolished. These inhibitory effects of CPA were reversible, requiring over 30 min washing for recovery (n = 5, data not shown). 2-APB (10 μm) first reduced the amplitude of slow potentials, and after 2–3 min abolished them, with no significant change in the resting membrane potential (control, −59.8 ± 3.9 mV, n = 8; in 2-APB, −61.9 ± 2.9 mV, n = 8; P > 0.05; Fig. 6Ba). 2-APB transiently elevated the level of [Ca²⁺]_i and abolished the Ca²⁺-transients (Fig. 6Bb). These inhibitory effects of 2-APB were reversible, with both electrical and [Ca²⁺]_i responses recovering within 5–7 min. These results suggest that the release of Ca²⁺ from an internal store is coupled to the generation of regenerative slow potentials and Ca²⁺-transients, and possible involvement of the activation of IP₃ receptors is also suggested.

Figure 6. Effects of CPA or 2-APB on electrical and Ca²⁺ responses.

CPA (10 μm, A) and 2-APB (10 μm, B) were applied while recording electrical (a) and calcium responses (b) simultaneously from single bundles of circular smooth muscle of the guinea-pig gastric antrum. Nifedipine (1 μm) was present throughout. The horizontal lines beneath the traces indicate the resting membrane potential (Aa, −62 mV; Ba, −65 mV) and the resting [Ca²⁺]_i level (Ab, 0.61 R_F340/F380; Bb, 0.60 R_F340/F380). A and B were recorded from different tissues.

Mitochondrial functions and slow potentials

Slow waves are sensitive to metabolic poisons (Nakayama et al. 1997) and the Q₁₀ values for the frequency of slow waves are high, indicating that the generation of slow waves may be linked to the metabolic activity of cells (Tomita, 1981). It is therefore reasonable to speculate that there is an involvement of mitochondrial functions in the spontaneous activity of gastric smooth muscles (see Ward et al. 2000). Attempts have been made to inhibit the electrochemical gradient across the mitochondrial inner membrane by blocking Ca²⁺ uptake with the application of CCCP, a protonophore in mitochondria (Duchen, 1999).

In circular smooth muscle of the gastric antrum, CCCP (1 μm) depolarized the membrane by 8–11 mV (from −64.0 ± 1.9 mV to −52.9 ± 6.8 mV, n = 6, P < 0.05), increased the resting level of [Ca²⁺]_i by 0.07 ± 0.01 R_F340/F380 (n = 6), and abolished both slow potentials and Ca²⁺-transients (Fig. 7A). These effects of CCCP were reversible but superfusion of tissues with CCCP-free solution for over 1 h was required for complete recovery (n = 3, data not shown).

Figure 7. Effects of mitochondrial protonophore and K_ATP channel inhibitors on electrical and Ca²⁺ responses.

Electrical (a) and calcium responses (b) were recorded simultaneously, while 1 μm CCCP (A), 30 μm glybenclamide (B), 3 mm 5-HDA (C), 10 μm HMR-1098 (D) and 10 μm HMR-1098 followed by 30 μm glybenclamide (E) were applied (shown by the bar), in single bundles of circular smooth muscle of the guinea-pig gastric antrum. Nifedipine (1 μm) was present throughout. All sets of responses were obtained from different tissues. The dotted lines indicate the resting membrane potential (Aa, −62 mV; Ba, −65 mV; Ca, −62 mV; Da, −65 mV; Ea, −65 mV) and the resting [Ca²⁺]_i level (Ab, 0.61 R_F340/F380; Bb, 0.60 R_F340/F380; Cb, 0.58 R_F340/F380; Db, 0.60 R_F340/F380; Eb, 0.60 R_F340/F380).

The effects of glybenclamide on slow potentials and changes in [Ca²⁺]_i were also observed in antrum circular smooth muscles, since in cardiac muscles this chemical reportedly inhibited the depolarization of mitochondrial membrane produced by K⁺ channel openers and resulting release of Ca²⁺ from the mitochondria (Holmuhamedov et al. 1998). Glybenclamide (over 10 μm) inhibited the generation of slow potentials, with no significant alteration of the resting membrane potential. In the cell shown in Fig. 7B, 30 μm glybenclamide inhibited the generation of slow potentials, with no alteration of the resting membrane potential (control, −62.1 ± 2.7 mV, n = 8; in glybenclamide, −59.5 ± 4.6 mV, n = 8; P > 0.05). Application of glybenclamide transiently increased the amplitude of Ca²⁺-transients and then prevented the generation of slow potentials and Ca²⁺-transients, with some reduction of the level of [Ca²⁺]_i by 0.01–0.05 R_F340/F380 (mean, 0.04 ± 0.005 R_F340/F380, n = 8). 5-HDA, an inhibitor of the mitochondorial ATP-sensitive K⁺ channels (Grover & Garlid, 2000), at 3 mm concentration, also abolished slow potentials and Ca²⁺-transients (Fig. 7C). The inhibitory effects of 5-HDA were much weaker than glybenclamide, and the amplitudes of slow potentials and Ca²⁺-transients were both reduced slowly with associated reduction in the frequency, and finally disappeared within 6–8 min, with no marked change in the resting membrane potential (control, −63.5 ± 2.5 mV; in 5-HDA, −62.5 ± 3.5 mV; n = 8; P > 0.05) and [Ca²⁺]_i level (control, 0.65 ± 0.08 R_F340/F380; in 5-HDA, 0.66 ± 0.10 R_F340/F380; n = 8; P > 0.05). However, HMR-1098 (10 μm), a selective inhibitor of sarcolemmal ATP-sensitive K⁺ channels in cardiac muscles (Sato et al. 2000), did not alter the amplitude of slow potentials (96.8 ± 1.75 % of control, n = 5, P > 0.05) and Ca²⁺-transients (93.8 ± 7.5 %, n = 5, P > 0.05) (Fig. 7D). In the presence of HMR-1098, glybenclamide again inhibited slow potentials and Ca²⁺-transients (Fig. 7E). These results suggest that the activation of ATP-sensitive K⁺ channels at the mitochondrial membrane, but not sarcolemmal membrane, is involved in the initiation of regenerative slow potentials.

DISCUSSION

When the relationship between spontaneous electrical activity and [Ca²⁺]_i response was determined in canine gastric muscles, the initial upstroke phase of the slow wave was accompanied by a nicardipine-insensitive Ca²⁺-transient, whereas the plateau phase of the slow wave was accompanied by a nicardipine-sensitive Ca²⁺-elevation (Ozaki et al. 1991). The present experiments also indicated that in circular smooth muscle of the guinea-pig gastric antrum, spontaneous electrical activity is accompanied by a Ca²⁺-transient consisting of a nifedipine-sensitive and a nifedipine-insensitive component. The nifedipine-sensitive component presumably results from an influx of Ca²⁺ through L-type Ca²⁺ channels; this component accompanies spike potentials and contributes more than half of the Ca²⁺-transient. This differs from canine gastric muscles, which generate plateau potentials by activation of nicardipine-sensitive Ca²⁺ channels. Thus, the contribution of L-type Ca²⁺ channels to spontaneous activity differs between these animal species. Nevertheless, in both species the Ca²⁺-transient associated with electrical activity is elicited spontaneously in the presence of Ca²⁺ antagonists, indicating that this Ca²⁺ component may be produced by either an influx of Ca²⁺ from the external medium through non-L-type Ca²⁺ channels or the release of Ca²⁺ from the internal stores, or both. In the guinea-pig stomach antrum, the electrical and Ca²⁺ responses were abolished by inhibiting the release of Ca²⁺ from internal stores by CPA, suggesting that the nifedipine-resistant component of the Ca²⁺-transient is produced mainly by the release of Ca²⁺ from the internal store. The inhibition of slow potentials and Ca²⁺-transients by 2-APB further suggests that the release of Ca²⁺ from the internal store involves the activation of IP₃ receptors, as in the case of the mouse small intestine (Malysz et al. 2001) and guinea-pig stomach (Van Helden et al. 2000; Hirst & Edwards, 2001). However, 2-APB has other actions, including the inhibition of conductances linked to capacitative Ca²⁺ entry (Prakriya & Lewis, 2001). This effect is unlikely to contribute to the inhibition of regenerative slow potentials as our preliminary studies (Suzuki et al. 2001) have found that inhibition of capacitative conductances with agents such as SKF-96365 (30 μm) or La³⁺ (10 μm) does not modulate slow potentials. However, additional experiments are required to confirm that IP₃ receptors are indeed involved in the generation of slow potentials. In many smooth muscles, internal Ca²⁺ stores have IP₃ receptor-sensitive and ryanodine receptor-sensitive components (Iino, 1989). The lack of effect of ryanodine on the spontaneous generation of slow potentials and Ca²⁺-transients supports the idea that the IP₃ receptor-mediated Ca²⁺ pool is the main source of the nifedipine-insensitive component of the Ca²⁺-transient. These suggestions on the nature and role of the internal Ca²⁺ store required for the generation of the slow potential are quite similar to those suggested to be involved in the ICC, which are thought to be pacemaker cells in the gastrointestinal tract (Ward et al. 2000).

Caffeine allows a differentiation between regenerative slow potentials and slow waves in gastric smooth muscle of the guinea-pig, in that the entire spontaneous electrical responses of circular muscle, but not longitudinal muscle and myenteric ICC (IC-MY), are inhibited by 1 mm caffeine (Dickens et al. 1999; Suzuki & Hirst, 1999; Suzuki, 2000). Similar differences in the sensitivity to caffeine are also noted in electrical responses recorded from the mouse stomach (Dickens et al. 2001). The present experiments show that the Ca²⁺-transient, along with the slow potential, is abolished in the presence of caffeine. It is unlikely that slow potentials are produced by signals conducted from IC-MY as these are removed in the dissection (see Suzuki & Hirst, 1999). Furthermore, when IC-MY are allowed to remain attached, their responses are not abolished by caffeine (1 mm; see Suzuki & Hirst, 1999). This suggests that the regenerative slow potentials result from myogenic activity of antral circular smooth muscles (Nose et al. 2000). Alternatively, the slow potential may be triggered by other types of interstitial cells such as IC-IM, which can be identified immunohistochemically in gastric smooth muscle tissues (Thuneberg, 1989; Komuro et al. 1996; Burns et al. 1997). Recently, gastric smooth muscles of the W/W^V mutant mouse, which are devoid of IC-IM, have been shown to lack the caffeine (1 mm) -sensitive component of slow waves (Dickens et al. 2001). This suggests that the generation of the caffeine-sensitive component of slow waves, i.e. regenerative slow potentials, requires IC-IM.

The cellular mechanisms of the inhibition of spontaneous activity by caffeine, however, remain unclear. Caffeine has multiple actions on smooth muscle, such as an enhanced release of Ca²⁺ from the internal stores (Iino, 1989; Somlyo & Somlyo, 1994), an inhibition of IP₃ receptors distributed on internal membranes (Berridge, 1993), and an inhibition of cyclic nucleotide phosphodiesterase resulting in the accumulation of cyclic AMP (Arnaud, 1987). The present experiments show that in antrum circular muscles, caffeine in concentrations sufficient to block the slow potential (≥ 1 mm) hyperpolarizes the membrane and reduces the level of [Ca²⁺]_i. The reduction of [Ca²⁺]_i may be an important factor in the inhibition of slow potentials. However, the level of [Ca²⁺]_i in the presence of caffeine (equal to about 0.65 R_F340/F380) was much higher than the threshold level for generation of spontaneous activity in gastric muscle (about 0.50 R_F340/F380), suggesting that other factors may be involved in the caffeine-induced inhibition of the slow potential. The actions of caffeine were very rapid in both onset and recovery, and in the guinea-pig intestine caffeine (up to 30 mm) does not alter the concentration of cyclic AMP (Prestwich & Bolton, 1995). These observations suggest that the elevation of cyclic AMP may not be involved in the caffeine-induced inhibition of slow potentials. The main site of action of caffeine may be on the internal membranes of smooth muscle cells (Iino, 1989). An inhibition of the agonist-induced production of IP₃ has been described in mouse pancreatic acinar cells (Toescu et al. 1992) and in human gastric mucus-secreting cells (Hamada et al. 1997). One possibility is that in gastric smooth muscle cells, caffeine inhibits IP₃ production, which may be accelerated by membrane depolarization (Suzuki, 2000).

The importance of mitochondrial Ca²⁺ handling for the generation of spontaneous activity in the stomach muscle was first highlighted by studies on ICC. Here, inhibition of Ca²⁺ uptake into mitochondria by reducing the electrochemical gradient across the mitochondrial membrane abolished spontaneous activity (Ward et al. 2000). The present experiments examined the effects of CCCP, which uncouples the proton transporter in the mitochondrial membrane (Duchen, 1999), on regenerative slow potentials. CCCP depolarized the membrane, increased the level of [Ca²⁺]_i and abolished the slow potential. Elevation of [Ca²⁺]_i by CCCP appears to be due to the disruption of uptake of Ca²⁺ into mitochondria, rather than the influx of Ca²⁺ through voltage-sensitive L-type Ca²⁺ channels, since all experiments were carried out in the presence of nifedipine. Thus, the inhibition of slow potentials during an increase in [Ca²⁺]_i induced by CCCP may be causally related to the functional disorder of mitochondria. Dysfunction of mitochondria by inhibition of the ATP-sensitive K⁺ channels at the mitochondrial membrane with glybenclamide or 5-HDA (Holmuhamedov et al. 1998), but not by HMR-1098 (Sato et al. 2000), also prevented the generation of slow potentials. Again this suggests that Ca²⁺ uptake into mitochondria may be important for the initiation of spontaneous activity in circular smooth muscle of the stomach. However, the ability of glybenclamide to inhibit slow potentials was much greater than that of 5-HDA, even though concentrations some 100 times higher were used. This may be partly related to the differences in tissues, since 5-HDA is selective in cardiac mitochondria, and the type of ATP-sensitive K⁺ channels distributed in smooth muscle mitochondria may differ from those in cardiac tissue (Grover & Garlid, 2000). The actions of CCCP differed from the inhibitors of ATP-sensitive K⁺ channels, in that the former inhibited slow potentials with an associated elevation of [Ca²⁺]_i, while the inhibition of the activity by the later was not accompanied by an elevation of [Ca²⁺]_i. Thus, disordered mitochondrial Ca²⁺ handling may prevent the rhythmic activity of smooth muscles, irrespective of the level of [Ca²⁺]_i.

The inhibition of regenerative slow potentials was associated with an elevation of [Ca²⁺]_i with CPA, 2-APB and CCCP, a reduction of [Ca²⁺]_i with caffeine, BAPTA and glybenclamide, and a stable level of [Ca²⁺]_i with 5-HDA. These data suggest that change in the level of [Ca²⁺]_i is not the only factor involved in inhibition of slow potentials. However, parallel changes in the frequency of slow potentials and the level of [Ca²⁺]_i were detected when the membrane potential was depolarized or hyperpolarized using current injection. Thus the voltage dependence of the frequency of slow potentials (Nose et al. 2000) may be related to [Ca²⁺]_i in smooth muscle cells. It is generally accepted that IC-MY are the pacemakers of gastric activity (Sanders, 1996; Huizinga et al. 1997; Dickens et al. 1999). If this is the case, it suggests that the concentration of intracellular Ca²⁺ in smooth muscle may be tightly related to the level of [Ca²⁺]_i in the ICC, since the voltage dependence of the frequency of spontaneous activity appears in the presence or absence of functional IC-MY (Nose et al. 2000). Alternatively, elevation of [Ca²⁺]_i in smooth muscle may modulate pacemaker mechanisms generated by IC-IM. The behaviour of [Ca²⁺]_i in gastric smooth muscle resembles that of ICC (Ward et al. 2000), suggesting the possibility that changes in [Ca²⁺]_i in smooth muscle cells reflect the changes in Ca²⁺ concentrations in pacemaker cells. Moreover it has been pointed out that IC-IM and IC-MY have many properties in common (Hirst & Edwards, 2001).

The present experiments allow speculation about the steps for generation of regenerative slow potentials in circular smooth muscle bundles of the guinea-pig gastric antrum, as follows: (1) an initial signal may appear in mitochondria as a metabolically coupled event, possibly associated with an elevation of Ca²⁺. (2) The signals generated in mitochondria may be transferred to the surface membrane to increase ionic conductance, thus eliciting depolarization of the membrane, such as the spontaneously occurring unitary potentials (Edwards et al. 1999). (3) These local depolarizations are signalled to internal Ca²⁺ store by an unidentified messenger. The possible messenger is IP₃, as suggested previously (Suzuki & Hirst, 1999; Suzuki, 2000; Dickens et al. 2001). The facilitated production of IP₃ during depolarization in smooth muscles (Ganitkevitch & Isenberg, 1993) may support this concept. The unknown process is the instantaneous release of Ca²⁺ from the internal store. A known mechanism for the instantaneous release of Ca²⁺ from the internal store of smooth muscle cells is the Ca²⁺-induced Ca²⁺ release mechanism mediated through ryanodine receptors (Iino, 1989), but as Ca²⁺-transients are not altered by ryanodine, the involvement of this mechanism is unlikely. (4) The elevated [Ca²⁺]_i may then activate a set of ion channels at the surface membrane to generate regenerative slow potentials. In the IP₃ receptor mutated mouse, gastric muscles fail to rhythmically generate slow waves but produce irregular spikes (Suzuki et al. 2000). These alterations may be explained by assuming that the generation of slow waves requires an instantaneous release of Ca²⁺ from the internal store, whereas the depolarization of the plasma membrane by the mitochondrial signal could activate L-type Ca²⁺ channels to elicit spike potentials.