Ca²⁺ transients in ICC-MY define the basis for the dominance of the corpus in gastric pacemaking

Abstract

Myenteric interstitial cells of Cajal (ICC-MY) generate and actively propagate electrical slow waves in the stomach. Slow wave generation and propagation are altered in gastric motor disorders, such as gastroparesis, and the mechanism for the gradient in slow wave frequency that facilitates proximal to distal propagation of slow waves and normal gastric peristalsis is poorly understood. Slow waves depend upon Ca²⁺-activated Cl⁻ channels (encoded by Ano1). We characterized Ca²⁺ signaling in ICC-MY in situ using mice engineered to have cell-specific expression of GCaMP6f in ICC. Ca²⁺ signaling differed in ICC-MY in corpus and antrum. Localized Ca²⁺ transients were generated from multiple firing sites and were organized into Ca²⁺ transient clusters (CTCs). Ca²⁺ transient refractory periods occurred upon cessation of CTCs, but a relatively higher frequency of Ca²⁺ transients persisted during the inter-CTC interval in corpus than in antrum ICC-MY. The onset of Ca²⁺ transients after the refractory period was associated with initiation of the next CTC. Thus, CTCs were initiated at higher frequencies in corpus than in antrum ICC-MY. Initiation and propagation of CTCs (and electrical slow waves) depends upon T-type Ca²⁺ channels, and durations of CTCs relied upon L-type Ca²⁺ channels. The durations of CTCs mirrored the durations of slow waves. CTCs and Ca²⁺ transients between CTCs resulted from release of Ca²⁺ from intracellular stores and were maintained, in part, by store-operated Ca²⁺ entry. Our data suggest that Ca²⁺ release and activation of Ano1 channels both initiate and contribute to the plateau phase of slow waves.

1. Introduction

Gastric motility is characterized by different patterns of contractile activity in the proximal and distal stomach. Contractile tone decreases in the proximal stomach during ingestion of food to accommodate the increase in volume without extreme increases in luminal pressure, and gradual restoration of tone moves contents on to the distal stomach [1, 2]. The distal stomach (i.e. corpus and antrum) generates peristaltic contractions that originate in the proximal corpus [3, 4]. Peristaltic contractions result from rhythmic depolarizations of smooth muscle cells (SMCs), known as slow waves, and excitation-contraction coupling due to activation of voltage-dependent Ca²⁺ channels expressed by SMCs [4–7]. The dominant pacemaker area in the stomach resides typically in the corpus along the greater curvature of the stomach, and slow waves generated by this pacemaker propagate distally to the pyloric sphincter [3].

Gastric slow waves originate from and propagate actively within networks of interstitial cells of Cajal that lie between the circular (CM) and longitudinal muscle (LM) layers in the plane of the myenteric plexus (i.e. ICC-MY) [8, 9]. ICC-MY are multi-polar cells with processes that form gap junctions with other ICC and less frequently with SMCs [10, 11]. Thus, SMCs and ICC-MY are electrically coupled, forming along with another type of interstitial cell, known as PDGFRα⁺ cells [12], the SIP syncytium [13]. The inward currents that cause slow wave depolarization conduct passively to SMCs, as these cells lack the mechanism required for slow wave regeneration [14].

Much about the molecular and functional characteristics of ICC has been revealed during the past 3 decades, but the pacemaker mechanism is not fully understood. Loss of ICC or the functions of ICC are a major factor in functional motility disorders in the proximal GI tract [15–17], so better understanding of the pacemaker mechanism may provide insights into new therapeutic approaches for treating gastroparesis, functional dyspepsia and other motility disorders. ICC express the receptor tyrosine kinase, c-Kit, and depriving ICC precursors of signaling via c-Kit causes defects in ICC [18, 19] or lesions in ICC networks in adult animals [8]. ICC express Ano1, that encodes a Ca²⁺-activated Cl⁻ channel (CaCC), that is involved in the initiation of slow waves and also contributes to the plateau phase of slow waves [20–23]. The majority of prior studies were performed on ICC of the small intestine, an often-used model for pacemaker activity and function of ICC-MY. However, gastric slow waves are fundamentally different from the events in the small intestine in that gastric slow waves are longer in duration and generate peristaltic rather than segmental motility patterns. A proximal-to-distal frequency gradient exists in gastric slow wave activity in animal models, and active propagation of slow waves occurs over many centimeters in larger animals [24–26]. Reasons for the long durations of gastric slow wave depolarizations and why ICC-MY in the corpus generate the dominant slow wave frequency in gastric muscles are unknown, but these questions are important because gastric arrhythmias, aberrant slow wave activity and disrupted peristalsis are associated with gastric emptying disorders in diabetic and idiopathic forms of gastroparesis [15, 27–31].

Ubiquitous high expression of Ano1 (aka Tmem16a) in ICC [20, 32, 33] suggests that regulation of intracellular Ca²⁺ ([Ca²⁺]_i) is a fundamental process in these cells. Indeed, changes in [Ca²⁺]_i have been observed in studies of ICC and different patterns of Ca²⁺ dynamics have been noted in ICC with pacemaker functions and ICC thought to be involved in transduction of inputs from motor neurons [21–23, 34–37]. Release of Ca²⁺ activates Ano1 channels in the plasma membranes of ICC, generating spontaneous transient inward currents (STICs) [38]. STICs cause spontaneous transient depolarizations (STDs) that can bring membrane potential to the threshold for generation of a slow wave. We hypothesized that dynamic transitions in [Ca²⁺]_i are fundamental for pacemaker activity in gastric ICC-MY: Ca²⁺ transients initiate slow waves and persistent Ca²⁺ release events maintain the long duration of depolarization that constitutes the plateau phase of gastric slow waves. We also hypothesized that differences in Ca²⁺ dynamics are responsible for the dominance of corpus ICC as gastric pacemakers. Using a novel mouse engineered with constitutive expression of the Ca²⁺ sensor GCaMP6f has allowed cell-specific, high resolution video imaging of Ca²⁺ transients in ICC in situ.

2. Materials and methods

2.1. Animals

Kit^+/copGFP mice (B6.129S7-^{Kittm1Rosay/J}; 5–8 wk old) were bred in house (Ro et al. 2010).

Kit-KI-GCaMP6f mice, a new transgenic mouse that expresses GCaMP6f driven by the endogenous Kit promotor, were generated and maintained in our animal facility [39]. GCaMP6f-floxed mice (Ai95 (RCL-GCaMP6f)-D) were purchased from Jackson Laboratories (Bar Harbor, MN, USA). Kit-iCre mice (c-Kit^+/Cre-ERT2) were gifted from Dr. Dieter Saur (Technical University Munich, Munich, Germany). Indicible Cre mice were injected with tamoxifen (TAM; Intraperitoneal injection; IP) at 7–8 weeks of age (2 mg of TAM for three consecutive days), as described previously [40], to induce activation of the Cre recombinase and expression of the GCaMP6f sensor. Both strains of mice expressing the Ca²⁺ sensor, TAM injected Kit-iCre-GCaMP6f and the Kit-KI-GCaMP6f, rely on GCaMP6f to report Ca²⁺ dynamics. Both strains had similar Ca²⁺ signal kinetics in ICC.

Mice were anaesthetized by inhalation of isoflurane (Baxter, Deer-field, IL, USA) and killed by cervical dislocation before removal of gastric tissues. All animals and procedures performed in this study were approved by the Institutional Animal Use and Care Committee at the University of Nevada, Reno and in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

2.2. Tissue preparation

The abdomen was opened and the entire stomach was removed from mice and placed in oxygenated Krebs-Ringer bicarbonate solution (KRB). The stomachs were cut along the lesser curvature and intraluminal contents were washed away with KRB. The corpus and antrum regions of the stomach were isolated from the fundus and pyloric regions by surgical incisions.

The corpus and antrum were distinguished and separated from each other by an incision across the incisura angularis and the mucosal layers were carefully removed by sharp dissection. Corpus and antrum segments were isolated from the greater curvature of the stomach.

2.3. Cell sorting and quantitative PCR

Kit^+/copGFP mice (B6.129S7-^{Kittm1Rosay/J}; 6–8 wks old) were used for evaluations of gene expression in antrum and corpus ICC. After enzymatic dispersion of muscles, ICC were sorted by fluorescence-activated cell sorting (FACS) and assessed for purity as previously described [34]. Total RNA was isolated using an Illustra RNAspin Mini RNA Isolation Kit (GE Healthcare). qScript cDNA SuperMix (Quanta Biosciences) was used according to the manufacturer’s instructions to synthesize first-strand cDNA. Quantitative PCR (qPCR) was performed using Fast Sybr Green chemistry on the 7900HT Fast Real-Time PCR System (Applied Biosystems) and gene-specific primers (Supplemental Table 1). Regression analysis was performed to generate standard curves from the mean values of technical triplicate qPCRs of log10 diluted cDNA samples. Evaluation of gene expression in ICC was compared with expression in the unsorted cells from antrum and corpus muscles of Kit^+/copGFP mice.

2.4. Electrophysiological experiments

Corpus and antrum tissues (2 × 6 mm), cut parallel to the circular layer from the greater curvature, were placed in a Sylgard coated (Dow Corning, Midland, MI) recording chamber with the serosal aspect of the muscle facing upward. One end of the muscle strip, corresponding to the greater curvature was pinned to the floor of the chamber with tungsten pins (50 μm; Goodfellow, Huntingdon, England); the other end of the muscle strip, corresponding to the lesser curvature, was attached to a Gould isometric strain gage (Gould UC3; Gould Instruments, OH, USA) with suture thread. A resting force of 5 mN was applied, which was previously shown to set the muscles at optimum length [41]. The muscles were allowed to equilibrate for 1 hour while being perfused with oxygenated 37 °C KRB. Intracellular microelectrode recordings were performed, as previously described [41]. Briefly, circular muscle cells near the greater curvature were impaled with glass microelectrodes having resistances of 80–120 MΩ. Transmembrane potentials were recorded with a high impedance electrometer (Axon Instruments, Axon Instruments, Union City, CA, USA). Data were recorded on a computer running AxoScope 10 data acquisition software (Axon Instruments) and figures were made using Clampfit analysis software (Axon Instruments).

2.5. Ca²⁺ imaging of ICC-MY in situ

Isolated muscles from the corpus and antrum regions were pinned in imaging chambers and equilibrated for 60 min while perfusing the chamber with 37 °C KRB. A spinning-disk confocal system (CSU-W1; spinning disk, Yokogawa Electric, Tokyo, Japan) attached to an upright Nikon Eclipse FN1 microscope was used for Ca²⁺ imaging. The system was equipped with two solid-state laser lines of 488 nm and 561 nm. The laser lines are combined with a borealis system (ANDOR Technology, Belfast, UK). Two high-speed electron multiplying charged coupled devices (EMCCD) cameras (Andor iXon-Ultra 897 EMCCD Cameras; ANDOR Technology, Belfast, UK) are mounted to the system to maintain fast speed acquisition at full frame of 512 × 512 active pixels, as previously described [42]. Images sequences were acquired using water immersion Nikon CFI Fluor lenses (10 × 0.3 NA, 20 × 0.5 NA, 40 × 0.8 NA, 60 × 0.8 NA and 100 × 1.1 NA) (Nikon Instruments, New York, USA) at 33 – 50 fps and MetaMorph software (MetaMorph Inc., TN, USA). In experiments with pharmacological agents, control images were collected (30 s) and then after responses to drugs (20 min). It has been previously demonstrated that slow waves in the gastric antrum are not inhibited by nifedipine [43, 44]. Therefore, experiments were performed in the presence of nicardipine (1 μm) to reduce movements and facilitate high-resolution imaging of cells in situ.

2.6. Analysis of Ca²⁺ imaging experiments

Image sequences of Ca²⁺ transients in ICC-MY (stacks of TIFF images) were exported and transferred to an analysis workstation computer for preprocessing and analysis, as previously described [45]. Briefly, Fiji/Image J (National Institutes of Health, MD, USA, http://rsbweb.nih.gov/ij), Automated Spatio Temporal Map analysis plugin (STMapAuto), https://github.com/gdelvalle99/STMapAuto [46] and custom build software (Volumetry G8d) were used. Movies of Ca²⁺ transients were motion stabilized, background subtracted and smoothed (Gaussian filter: 1.5 × 1.5 μm, StdDev 1.0). A particle analysis was employed using a flood-fill algorithm to enhance Ca²⁺ firing site activity detection. The areas of Ca²⁺transients in cells were saved as Ca²⁺ particles (PTCLs), and the combined areas and total number of PTCLs were calculated. To show the overall regions in fields of view (FOV) where Ca²⁺ transients occurred, PTCLs were summed throughout the video to map their occurrence. The spatial information in the Ca²⁺ occurrence maps data is an indication of each firing site that gives information on their temporal activation.

2.7. Drugs and solutions

KRB solution containing (mmol/L): NaCl, 5.9; NaHCO₃, 120.35; KCl, 1.2; MgCl₂, 15.5; NaH₂PO_4,1.2; CaCl₂, 2.5; and glucose, 11.5 was used to maintain tissues physiological conditions. The KRB solution was warmed to a physiological temperature of 37 ± 0.3 °C and bubbled with a mixture of 97% O₂ – 3% CO₂. For experiments utilizing external solutions with 0 [Ca²⁺]_o, CaCl₂ was omitted and 0.5 mM ethylene glycolbis (b-aminoethyl ether)-N, N, N’, N’–tetraacetic acid (EGTA) was added to the solution. Nicardipine, pinacidil was purchased from Millipore-Sigma (St. Louis, Missouri, USA). NNC 55–0396 were purchased from Alomone Labs (Jerusalem, Israel). Thapsigargin, isradipine and CPA were purchased from Tocris Bioscience (Ellisville, Missouri, USA). GSK 7975A was purchased from Aobious (Aobious INC, MA, USA).

2.8. Statistical analysis

Statistical analyses were performed using either a Students t-test or one-way ANOVA with a Tukey post hoc test where appropriate. Data is presented as the mean ± standard error unless otherwise stated. In all tests, P < 0.05 was considered significant. When describing data, n refers to the number of animals used in a dataset. Probabilities < 0.05 are represented by a single asterisk (*), probabilities < 0.01 are represented by two asterisks (**), probabilities < 0.001 are represented by three asterisks (***) and probabilities < 0.0001 are represented by four asterisks (****). Statistical tests were performed on original datasets. All statistical tests were performed using GraphPad Prism 8.0.1 (San Diego, CA).

3. Results

3.1. Electrical and mechanical behaviors in corpus and antrum muscles

ICC generate electrical slow waves that are conducted to SMCs in GI muscles to produce motility behaviors. In the stomach slow waves occur at higher intrinsic frequencies in the corpus than in the antrum, and we sought to characterize these differences in the mouse stomach.

Circular muscle cells in strips of gastric corpus were impaled and displayed inter-slow wave membrane potentials (e.g. RMP at the most negative level) averaging −58 ± 1 mV. Slow waves occurred spontaneously at 7.1 ± 0.4 cycles min⁻¹ and averaged 15.1 ± 1.7 mV in amplitude and ½ maximal duration of 3.6 ± 0.3 s occurred (Fig. 1A; n = 11). Corpus slow waves consisted of an initial slow depolarization, often with spontaneous transient depolarizations (STDs) superimposed, followed by a sinusoidal wave with superimposed STDs (Fig. 1A&E). Each slow wave was associated with a phasic contraction of the muscle (i.e. 7.1 ± 0.4 cycles min⁻¹) averaging 0.34 ± 0.05 mN in amplitude (Fig. 1B).

Fig. 1. Electrical and mechanical activities of gastric corpus and antrum.

A&B show simultaneous intracellular electrical and isometric force recordings from the gastric corpus. C&D show similar recordings of electrical and mechanical activity from the gastric antrum. The gastric corpus is more depolarized and slow waves are at a greater frequency than antrum. Corpus slow waves are also smaller in amplitude and do not produce as forceful contraction as antrum. E shows gastric corpus slow waves at a faster sweep speed. Corpus slow waves consisted of an initial slow depolarization with spontaneous transient depolarizations (STDs; *) followed by a sinusoidal wave that also had superimposed STDs. F shows gastric antrum slow waves at a faster sweep speed. Antrum slow waves consisted of an upstroke depolarization, partial repolarization, and a plateau phase that was sustained for several seconds before repolarization to a diastolic RMP. An inflection (arrows and inset) often separated the upstroke phase into two discrete components. G–J Summarized bar graphs illustrating the differences in RMP and electrical slow wave parameters between corpus and antrum. **** P < 0.0001. All data graphed as mean ± SEM. n = 11.

Circular muscle cells in the gastric antrum, investigated in the same manner, displayed more polarized RMPs compared to the corpus, averaging −71 ± 1 mV (Fig. 1G; n = 11). Slow waves recorded from antrum cells were more robust and amplitude compared to the corpus that averaged 33 ± 3 mV (Fig. 1H; n = 11) and ½ maximal duration of 6.5 ± 0.5 secs (Fig. 1I; n = 11). However, the frequency of antrum slow waves was less (3.4 ± 0.3 cycles min⁻¹; Fig. 1J; n = 11) than the frequency of these events in the corpus. Antrum slow waves typically consisted of an upstroke depolarization, partial repolarization, and a plateau phase that was sustained for several seconds before repolarization to the RMP. In recordings from some cells the upstroke depolarization displayed an inflection that tended to separate this phase into two discrete components (see Fig. 1F and inset in Fig. 1F). As with corpus muscles, each antrum slow wave was associated with a phasic contraction (i.e. 3.4 ± 0.3 min⁻¹; Fig. 1D; n = 6) that averaged 0.46 ± 0.07 mN in force. A summary of the differences in RMP and electrical slow wave parameters between gastric corpus and antrum is summarized in (Fig. 1G–J).

Gastric muscles undergo spontaneous phasic contractions which make post-acquisition analysis of imaging data difficult due to movement artifacts. Therefore, in imaging experiments discussed below we suppressed the contractile activities if the muscles by addition of nicardipine or nifedipine in all experiments. We first tested the effects of the dihydropyridine, nifedipine, on electrical activity to confirm that this class of drugs did not disrupt or block slow waves (Fig. 2; n = 10). In these expeirments the RMP of antrum muscles averaged −72 ± 1.6 mV (Fig. 2 Ca; n = 10), and slow waves were observed with an average amplitude of 38 ± 2.7 mV (Fig. 2 Cb; n = 10), a ½ maximal duration of 6.2 ± 0.4 secs (Fig. 2 Cc; n = 10) and frequency of 3.3 ± 0.2 cycles min⁻¹ (Fig. 2 Cd; n = 10). Nifedipine (1μM) caused a small depolarization in RMP to −69 ± 1.5 mV and reduced the ½ maximal duration of slow waves to 4.1 ± 0.5, other electrical parameters were not changed from control (Fig. 2 Cb&Cd; n = 10). The effects of nifedipine on electrical activity are summarized in (Fig. 2 Ca–Cd).

Fig. 2. Effect of l-type Ca²⁺ channel inhibition on antrum slow waves.

A&B Electrical slow waves of the gastric antrum under control conditions (i.e. no drugs; A) and after the addition of nifedipine (1μM; B). Note the slight membrane depolarization and reduced ½ maximal duration. C Summarized data of changes in RMP Ca, slow wave amplitude Cb, ½ maximal duration of slow waves Cc and slow wave frequency Cd. Only RMP and ½ maximal duration of slow waves was significantly affected by nifedipine. All data graphed as mean ± SEM. n = 10, * P < 0.05; ** P < 0.01.

3.2. Ca²⁺ dynamics in ICC-MY of the corpus and antrum

ICC-MY have been reported to be the pacemaker cells responsible for generating slow waves in gastric muscles [8, 9], and Ca²⁺-activated Cl⁻ channels encoded by Ano1 are required for slow wave activity [20, 47]. Therefore, we imaged Ca²⁺ dynamics as displayed by GCaMP6f expressed in ICC-MY in isolated corpus and antrum muscles at 60–100x. This approach allowed visualization of Ca²⁺ transients with high spatial resolution. Subcellular Ca²⁺ transients originated from a multitude of distinct firing sites in ICC-MY networks (Fig. 3A–H; n = 10). Activity plots of Ca²⁺ transients during 30 s of recording from gastric corpus (Fig. 3C) and antrum (Fig. 3G and supplemental movie 1) showed that firing of Ca²⁺ transients is organized into temporal clusters (Ca²⁺ transient clusters or CTCs). The avarge frequency of CTCs in the corpus were 6.9 ± 0.3 cpm, and in the antrum 4.2 ± 0.4 cpm (n = 10) which is consistant with gradient in electrical slow waves in corpus and antrum muscles. Each distinct firing site was color-coded in corpus and antrum ICC-MY (Fig. 3B&F) and plotted as a function of time as occurrence maps (Fig. 3D&H) to visualize and quantify the origin and activity of Ca²⁺ transients. Occurrence maps constructed from the Ca²⁺ transients in corpus ICC-MY (Fig. 3D) and antrum ICC-MY (Fig. 3H) showed that firing sites could fire once or multiple times during CTCs. The number of firing sites in a single ICC-MY in the corpus ranged from 2 to 6 sites, and in the antrum ranged from 3 to 8 sites (Fig. 3I). Ca²⁺ firing sites were most active during the first 600 ms after initiation of a CTC, and then activity waned with time in corpus and antrum ICC-MY (Fig. 3I; n = 10). The average number of firing sites discharging at various times during CTCs is illustrated in the distribution histogram in Fig. 3I. The initial high rate of Ca²⁺ transients and decay as a function of time would suggest that Ca²⁺ influx could be an important factor for initiation of Ca²⁺ transients and organizing the Ca²⁺ transients into CTCs.

Fig. 3. Differences in Ca²⁺ transients firing in ICC-MY between the corpus and antrum.

A Image of an ICC-MY network from gastric corpus of a Kit-GCaMP6 mouse at 60 × magnification. B Image showing individually color-coded Ca²⁺ firing sites in the FOV shown in A. C Plot of total Ca²⁺ transients PTCLs activity from all ICC-MY Ca²⁺ firing sites within the FOV in the corpus. D The temporal characteristics of each individual, color-coded firing site is displayed as an occurrence map, with each “lane” representing the occurrence of firing PTCLs within each firing site. Note that multiple corpus ICC-MY Ca²⁺ sites fires during the intra-wave period. E Representative image of an ICC-MY network from gastric Antrum at 60 × magnification. F Image showing individually color-coded Ca²⁺ firing sites in the FOV shown in E (see Supplemental Movie 1). G Total Ca²⁺ transients activity plot from all ICC-MY Ca²⁺ firing sites within the FOV in the antrum. H an occurrence map, with each “lane” representing the occurrence of firing PTCLs within each firing site. Note that limited number of Ca²⁺ sites firing during the intra-wave period in antrum ICC-MY. I Distribution plot showing averages of firing sites number during a Ca²⁺ wave in ICC-MY corpus and antrum. Values are calculated for 5 s and plotted in 297 ms bins (n = 10). J Summary graph show average number of PTCL Ca²⁺ firing sites in ICC-MY during the intra-wave period. K & L Summary graphs show average PTCL areas and counts for Ca²⁺ firing sites in ICC-MY. ** = P < 0.01, n = 6. All data graphed as mean ± SEM.

A major difference noted between the patterns of Ca²⁺ transients in corpus and antrum ICC-MY was firing of Ca²⁺ transients between CTCs. In the antrum CTCs were followed by a nearly quiescent refractory period upon completion of a CTC. Several seconds passed before low levels of activity were observed (Fig. 3H). This low level activity preceeded the firing of the next CTC. In contrast, the period between CTCs in corpus ICC-MY clearly displayed lower levels of Ca²⁺ transient firing than during CTCs, but multiple events occurred with repetitive Ca²⁺ transients firing from single firing sites. Thus, Ca²⁺ signaling was more dynamic during the intervals between CTCs in corpus than in antrum ICC-MY, as tabulated in Fig. 3J (n = 10). An average of 12.2 ± 1.4 sites were active between CTCs in corpus vs. 1.9 ± 0.5 sites in antrum calculated for 2 s before the onset of CTC). This observation suggests that the mechanism responsible for Ca²⁺ transients is more excitable in corpus than in antum ICC-MY. Although the number of firing sites per cell during CTCs was higher in antrum than in corpus, the area and total number of Ca²⁺ PTCLs within the FOV was not significantly different in ICC-MY of the two regions (Fig. 3K&L; n = 10).

3.3. Initiation of clustered Ca²⁺ transients requires a Ca²⁺ influx mechanism

The effects of reducing extracellular Ca²⁺ ([Ca²⁺]_o) on Ca²⁺ transients in antrum ICC-MY were examined. Removal of [Ca²⁺]_o (Ca²⁺-free KRB solution containing 0.5 mM EGTA) abolished Ca²⁺ transients within 15 min (Fig. 4; n = 6). Reduction in [Ca²⁺]_o from 2.5 mM to 0 mM decreased firing of Ca²⁺ transients in a concentration-dependent manner (Fig. 4A–K; n = 6). ICC-MY Ca²⁺ PTCL area, number of firing sites and CTC frequency were reduced in response to reducing [Ca²⁺] _o (Fig. 4F–H; n = 6). When [Ca²⁺]_o was reduced from 2.5 mM (control conditions) to 1 mM, the average of Ca²⁺ PTCL area was reduced to 55.8 ± 4.8% of control (Fig. 4I; n = 6), PTCL count to 44.9 ± 7% (Fig. 4J; n = 6). The durations of CTCs were also reduced to 69.8 ± 3.2% in 1 mM [Ca²⁺]_o, but CTC frequency increased slightly, but not significantly, to 108.6 ± 5.4% (Fig. 4K; n = 6). Further reduction in [Ca²⁺] _o to 0.5 mM reduced Ca²⁺ transient parameters to 22.9 ± 4.6% PTCL area (Fig. 4I; n = 6), 23.8 ± 5.1% PTCL count (Fig. 4J; n = 6). The duration of CTCs were reduced to 24.2 ± 4.8%, and CTC frequency was reduced to 27.3 ± 6.2% (Fig. 4K; n = 6). Removal of Ca²⁺ from the extracellular solution reduced Ca²⁺ transient PTCL area to 0.8 ± 0.5% (Fig. 4I; n = 6) and PTCL count was reduced to 1.3 ± 0.8% (Fig. 4J; n = 6). CTCs were abolished in 0 [Ca²⁺]_o (Fig. 4K; n = 6). These experiments highlight the importance of the transmembrane Ca²⁺ gradient for maintenance of intracellular Ca²⁺ signaling in ICC-MY.

Fig. 4. Effects of extracellular Ca²⁺ on ICC-MY Ca²⁺ transients.

A Colored heat-map image of total Ca²⁺ transients of antrum ICC-MY under control conditions with [Ca²⁺]_o = 2.5 mM and B after Ca²⁺ removal from the extracellular solution ([Ca²⁺]_o = 0 mM and 0.5 mM EGTA). Ca²⁺ activity is color-coded with warm areas (white, red) representing bright areas of Ca²⁺ fluorescence and cold colors (purple, black) representing dim areas of Ca²⁺ fluorescence. Scale bar is 50 mm in both A & B. C Ca²⁺ transients of firing sites in ICC-MY were color-coded and plotted as an occurrence map under control conditions with [Ca²⁺]_o = 2.5 mM. D showing the effects of reducing [Ca²⁺]_o to 0.5 mM Ca²⁺. E showing the effects of removal of [Ca²⁺]_o (final solution contain 0 mM Ca²⁺ and buffered with 0.5 mM EGTA). F–H Traces of Ca²⁺ PTCL activity in ICC-MY (PTCL area, blue and PTCL count, green) under control conditions F, in presence of 0.5 mM Ca²⁺ G and after removal of [Ca²⁺]_o as shown in H. Summary graphs of Ca²⁺ PTCLs in ICC-MY under control conditions and with reduced [Ca²⁺]_o to 1 mM, 0.5 mM and removal of [Ca²⁺]_o is plotted in I (PTCL area); J (PTCL count) and K (CTC frequency). Data were normalized to controls and expressed as percentages (%). Significance was determined using one-way ANOVA, **** = P < 0.0001, n = 6. All data graphed as mean ± SEM.

3.4. Expression of Ca²⁺entry channels in gastric ICC

As demonstrated by the previous experiments the Ca²⁺ gradient is essential to maintain active Ca²⁺ transients in gastric ICC-MY. We examined the expression plasma membrane Ca²⁺ transporters that might be responsible for Ca²⁺ influx in ICC. ICC expressing the fluorescent reporter copGFP were dispersed enzymatically from corpus and antrum muscles, and ICC were sorted to purity by fluorescence activated cell-sorting (FACS), as previously described [45]. The purity of FACS-sorted antrum and corpus ICC was evaluated using cell-specific markers (Fig. 5A; n = 4). Throughout the GI tract, c-Kit receptors and ANO1 channels are considered signatures for ICC. Cells sorted by FACS displayed significant enrichment in Kit and Ano1 transcripts vs. unsorted cells (Fig. 5A; Table 1). Myh11 (smooth muscle cell marker), pdgfra (platelet-derived growth factor receptor α cell marker), and Uch11 (pan neuronal marker encoding PGP9.5) expression levels were minimal (Fig. 5A; Table 1), suggesting that ICC were enriched in the population of cells sorted by FACS. Expression of voltage-dependent Ca²⁺ channels and Na⁺/Ca²⁺ exchanger isoforms was assayed by qPCR (Fig. 5B&C).

Fig. 5. Molecular expression of Ca²⁺ influx channels in ICC.

A Relative expression of cellular-specific biomarker genes in ICC (sorted to purity by FACS) and compared with unsorted cells dispersed from gastric antrum and corpus tissues obtained from Kit^+/copGFP mice. Relative expression was determined by qPCR and normalized to Gapdh expression. Kit (tyrosine kinase receptor, found in ICC), Ano1 (Ca²⁺-activated Cl⁻ channel), Myh11 (smooth muscle myosin). PDGFRa (platelet-derived growth factor receptor α cell marker) and Uch11 (neural marker encoding PGP 9.5) B Relative expression of major voltage dependent Ca²⁺ entry channels, l-Type Ca²⁺ channels (Cacna1c and Cacna1d) and T-type Ca²⁺ channels (Cacna1g, Cacna1h and Cacna1i). C Relative expression of the Na⁺/Ca²⁺ exchanger (NCX) isoforms (Slc8a1, Slc8a2 and Slc8a3) in gastric antrum and corpus ICC. All data graphed as mean ± SEM (n = 4).

Table 1.

Molecular expression quantification of important Ca²⁺ signaling genes in gastric ICC.

	Antrum Unsorted cells	ICC	P	Corpus Unsorted cells	ICC	P
Kit	0.024 ± 0.0001	0.849 ± 0.0057	****	0.069 ± 0.0019	0.979 ± 0.0097	****
Ano1	0.064 ± 0.0008	1.506 ± 0.0286	****	0.212 ± 0.0012	1.806 ± 0.0408	****
Myh11	0.640 ± 0.0079	0.280 ± 0.0058	–	1.527 ± 0.0176	0.275 ± 0.0054	–
PDGFRa	0.080 ± 0.0016	0.011 ± 0.0002	–	0.125 ± 0.0012	0.014 ± 0.0021	–
Uchl1	0.024±0.0004	0.017±0.0008	–	0.072±0.0012	0.011±0.0008	–
Cacna1c	0.005 ± 0.0003	0.008 ± 0.0003	****	0.007 ± 0.0006	0.010 ± 0.0006	–
Cacna1d	0.014 ± 0.0004	0.047 ± 0.0004	****	0.019 ± 0.0001	0.065 ± 0.0012	****
Cacna1g	0.001 ± 0.0001	0.004 ± 0.0001	****	0.004 ± 0.0001	0.005 ± 0.0003	****
Cacna1h	0.004 ± 0.0005	0.042 ± 0.001	****	0.004 ± 0.0006	0.061 ± 0.0012	****
Cacna1i	undetected	undetected		undetected	undetected
Slc8a1	0.001 ± 0.00003	0.018 ± 0.0001	****	0.005 ± 0.0003	0.021 ± 0.0009	****
Slc8a2	0.001 ± 0.00004	0.017 ± 0.0005	****	0.005 ± 0.0004	0.016 ± 0.0002	****
Slc8a3	0.005 ± 0.0001	0.005 ± 0.0001	–	0.006 ± 0.0004	0.005 ± 0.0002	–
Atp2a1	undetected	undetected		undetected	undetected	–
Atp2a2	0.139 ± 0.0021	1.469 ± 0.0128	****	0.372 ± 0.0040	1.584 ± 0.0099	****
Atp2a3	0.031 ± 0.0006	0.622 ± 0.0062	****	0.094 ± 0.0003	0.676 ± 0.0067	****
Itpr1	0.029 ± 0.0002	0.079 ± 0.0010	****	0.076 ± 0.0033	0.116 ± 0.0005	–
Itpr2	0.012 ± 0.0004	0.007 ± 0.0005	–	0.014 ± 0.0003	0.013 ± 0.0002	NS
Itpr3	0.010 ± 0.00002	0.002 ± 0.00007	–	0.017 ± 0.0009	0.001 ± 0.0001	–
Ryr1	0.0004 ± 0.0001	0.001 ± 0.0001	****	0.0002 ± 0.0001	0.0004 ± 0.0001	*
Ryr2	0.005 ± 0.001	0.012 ± 0.0006	****	0.006 ± 0.0002	0.013 ± 0.0011	****
Ryr3	0.008 ± 0.0002	0.001 ± 0.0005	–	0.019 ± 0.0012	0.001 ± 0.0001	–
Orai1	0.028 ± 0.0003	0.060 ± 0.0043	****	0.052 ± 0.0006	0.069 ± 0.0025	****
Orai2	0.010 ± 0.0001	0.003 ± 0.0001	–	0.015 ± 0.0002	0.005 ± 0.0001	–
Orai3	0.013 ± 0.0003	0.042 ± 0.0010	****	0.038 ± 0.0001	0.047 ± 0.0014	****
Stim1	0.061 ± 0.0009	0.104 ± 0.0020	****	0.089 ± 0.0030	0.119 ± 0.0029	****
Stim2	0.032 ± 0.0012	0.155 ± 0.0055	****	0.058 ± 0.0009	0.176 ± 0.0041	****

ANOVA post-test used to derive P values. P values indicated as not significant NS, P < 0.1 *, P < 0.01 **, P < 0.001 *** and P < 0.0001 ****. – indicate no significant gene enrichment in ICC compared to unsorted cell sample. n = 4.

L-type voltage-dependent Ca²⁺ channels (Ca_V1.2 and Ca_V1.3), encoded by Cacna1c and Cacna1d, are expressed in ICC, but transcripts for Cacna1d were more abundant in both antrum and corpus ICC (Fig. 5B; Table 1). T-type voltage-dependent Ca²⁺ channels, encoded by Cacna1g (Ca_V 3.1) and Cacna1h (Ca_V 3.2), transcripts were also expressed in antrum and corpus ICC, but Cacna1i (Ca_V 3.3) transcripts were not resolved (Fig. 5B; n = 4; Table 1). Ca²⁺ entry via the Na⁺/Ca²⁺ exchanger (NCX) was proposed as a mechanism to sustain ANO1 channel activation in ICC during the plateau phase of slow waves in small intestinal ICC-MY [48]. Expression of NCX isoforms (Slc8a1, Slc8a2 and Slc8a3) also occurs in gastric ICC. Slc8a1 and Slc8a2 showed highest expression in antrum and corpus ICC, as compared to Slc8a3 (Fig. 5C; n = 4; Table 1).

3.5. L-type and T-type Ca²⁺ channels contributions to Ca²⁺ transients in ICC-MY

Voltage-dependent Ca²⁺ channel (VDCC) transcripts for L-type Ca²⁺ channels (Ca_V1.3, Ca_V1.2) and T-type Ca²⁺ channels (Ca_V3.1, Ca_V3.2) were expressed in gastric ICC (Fig. 5B). As described above (Fig. 2) and previously, electrical slow waves persist in the stomach in the presence of dihydropyridines, such as nicardipine and nifedipine [43, 49, 50]. Therefore, we tested the effects of isradipine (an atagonist for both Ca_V1.3 and Ca_V1.2) to evaluate the contribution of these channels to Ca²⁺ transients and CTCs. Isradipine (1μM) reduced Ca²⁺ transients in ICC-MY (Fig. 6; n = 6), as shown by the reduction in firing sites evident in occurrence maps and Ca²⁺ PTCL plots (Fig. 6C–F). Ca²⁺ PTCL area was reduced to 76.2 ± 3.3% (Fig. 6G; n = 6), and PTCL count was reduced to 75.9 ± 3.5% (Fig. 6H; n = 6). The number of firing sites was reduced to 78.9 ± 2.9% by isradipine (Fig. 6I; n = 6). Isradipine also significantly reduced the durations of CTCs to 70.7 ± 5.7% (Fig. 6J; n = 6) and increased the frequency of CTCs to 107.2 ± 2.9% (n = 6). These data show that l-type Ca²⁺ channels, most likely Ca_V1.3 because Ca_V1.2 channels were pre-blocked with nicardipine, contribute to Ca²⁺ entry into ICC-MY and have a role in sustaining the durations of CTCs.

Fig. 6. L-type Ca²⁺ channel antagonist, isradipine effects on ICC-MY Ca²⁺ transients.

A&B Representative heat-map images of an antrum ICC-MY network showing active Ca²⁺ PTCLs under control conditions and in the presence of isradipine (1mM). Ca²⁺ activity is color-coded with warm areas (white, red) representing bright areas of Ca²⁺ fluorescence and cold colors (purple, black) representing dim areas of Ca²⁺ fluorescence. Scale bar is 50 mm in both A & B. C & D Ca²⁺ activity in ICC-MY showing color-coded Firing sites plotted as an occurrence map under control conditions C and in the presence of isradipine (1mM) D. Traces of firing sites showing PTCL area (E; blue) and PTCL count (E; green) under control conditions and in the presence of isradipine; PTCL area (F; blue) and PTCL count (F; green). Summary graphs of Ca²⁺ PTCL activity in ICC-MY before and in the presence of isradipine are shown in G (PTCL area/frame), H (PTCL count/frame), I the number of PTCL active sites. J Summary graph of Ca²⁺ transient clusters (CTCs) duration. Data were normalized to controls and expressed as percentages (%). Significance determined using unpaired t-test, ** = P < 0.01, *** = P < 0.001, n = 6. All data graphed as mean ± SEM.

Expression of T-type Ca²⁺ channels (Cacna1h) was also observed in gastric ICC (Fig. 5B), and previous studies suggested a role for this conductance in the upstroke depolarization and active propagation of slow waves [49, 50]. We evaluated the role of T-type Ca²⁺channels in modulating Ca²⁺signaling in ICC-MY using a specific T-type channel antagonist, NNC 55–0396 (10 μM). NNC 55–0396 reduced Ca²⁺ transient firing (Fig. 7; n = 6), Ca²⁺ transients firing sites (Fig. 7C&D), Ca²⁺ PTCL areas and PTCL counts (Fig. 7E&F). PTCL area was reduced to 28.3 ± 6.1% (Fig. 7G; n = 6) and PTCL count was reduced to 35.9 ± 5.8% (Fig. 7H; n = 6). The number of firing sites was reduced to 33.4 ± 9% (Fig. 7I; n = 6) by NNC 55–0396. NNC 55–0396 reduced CTC duration to 53.2 ± 4.5% (Fig. 7J; n = 6) and CTC frequency were also reduced to 68.2 ± 13.1% (n = 6). The data suggest that T-type Ca²⁺ channels are fundamental to the initiation and organization of Ca²⁺ transients into CTCs in ICC-MY.

Fig. 7. Effects of T-type Ca²⁺ channel antagonist, NNC 55–0396 on antrum ICC-MY Ca²⁺ transients.

A&B Ca²⁺ transient particles heat-map images in ICC-MY under control conditions A and in the presence of NNC 55–0396 (10 μM) B. Active firing sites were color-coded and plotted as an occurrence maps in the ICC-MY network under control C and in the presence of NNC 55–0396 D. Trace plots of Ca²⁺ transient PTCLs activity of ICC-MY in control conditions showing PTCL area (blue) and PTCL count (green) E and in the presence of NNC 55–0396 F. Summary graphs of average percentage changes in PTCL area G, PTCL count H, the number of PTCL active sites I. J Average percentage changes of Ca²⁺ transient clusters (CTCs) duration. Data were normalized to controls and expressed as percentages (%). Significance determined using unpaired t-test, *** = P < 0.001, **** = P < 0.0001, n = 6. All data graphed as mean ± SEM.

Corpus ICC-MY Ca²⁺ transients were also reduced by NNC 55–0396 (Fig. 8; n = 5), Ca²⁺ transient firing sites were significantly reduced. Ca²⁺ PTCL area was reduced to 36.4 ± 4.9% (Fig. 8G; n = 5), and PTCL count was reduced to 40.2 ± 5.1% (Fig. 8H; n = 5). The number of firing sites was reduced by NNC 55–0396 to 32.9 ± 2.5% (Fig. 8I; n = 5). NNC 55–0396 also reduced CTC duration to 50.1 ± 3.1% (Fig. 8J; n = 5) and CTC frequency to 60.7 ± 6.3% (n = 5). An interesting observation was that the T-type Ca²⁺ antagonist disrupted the organization of CTCs in antrum and corpus ICC-MY. While NNC 55–0396 did not entirely abolish CTCs, there was a less well-defined wave front of initiation in the presence of the drug (Figs. 7D and 8 D), and there was an increase in stochastic firing of Ca²⁺ transients during inter-CTC intervals in antrum ICC-MY (Fig. 7C&D).

Fig. 8. Effects of T-type Ca²⁺ channel antagonist, NNC 55–0396 on corpus ICC-MY Ca²⁺ transients.

A&B Ca²⁺ transient particles heat-map images in ICC-MY under control conditions A and in the presence of NNC 55–0396 (10 μM) B. Ca²⁺ activity is color-coded with warm areas (white, red) representing bright areas of Ca²⁺ fluorescence and cold colors (purple, black) representing dim areas of Ca²⁺ fluorescence. Scale bar is 30 mm in both A & B. Active firing sites were color-coded and plotted as an occurrence maps in the ICC-MY network under control C and in the presence of NNC 55–0396 D. Trace plots of Ca²⁺ transient PTCL activity of ICC-MY in control conditions showing PTCL area (blue) and PTCL count (green) E and in the presence of NNC 55–0396 F. Summary graphs of average percentage changes in Ca²⁺ PTCL area G, PTCL count H, the number of PTCL active sites I. J Average percentage changes of Ca²⁺ transient clusters (CTCs) duration. Data were normalized to controls and expressed as percentages (%). Significance determined using unpaired t-test, **** = P < 0.0001, n = 5. All data graphed as mean ± SEM.

3.6. Effects membrane hyperpolarization on ICC-MY Ca²⁺ transients

T-type and L-type Ca²⁺ channels are both voltage-dependent. Thus, transmembrane potentials may have a significant effect on generation and propagation of CTCs. We tested the effects of hyperpolarization on CTCs using pinacidil (10 μM, a selective K_ATP channel agonist) that hyperpolarizes murine gastric muscles [51]. Pinacidil (10 μM) reduced the number of Ca²⁺ transient PTCLs to 82.6 ± 4.7% (Fig. 9H; n = 5), reduced the duration of CTCs to 74.4 ± 2.4% (Fig. 9J; n = 5) and increased CTC firing frequency to 116.1 ± 5.7% of control (n = 5). Other Ca²⁺ transient firing parameters were unchanged.

Fig. 9. membrane hyperpolarization effects on ICC-MY Ca²⁺ transients.

A Ca²⁺ Firing sites in ICC-MY are color-coded and plotted in a heat map under control conditions and in the presence of pinacidil (10 μM) B. Occurrence maps showing active firing sites under control conditions C and in the presence of pinacidil D. Trace activity of firing sites PTCL area (blue) and PTCL count (green) under each condition are shown in E&F. Summary graphs of Ca²⁺ PTCL activity in ICC-MY in the presence of v pinacidil are shown in G (PTCL area) and H (PTCL count). I The number of active Ca²⁺ PTCL and total Ca²⁺ transient cluster (CTC) duration in J. Data were normalized to controls and expressed as percentages (%). Significance determined using unpaired t-test, ** = P < 0.01, **** = P < 0.0001, n = 5. All data graphed as mean ± SEM.

3.7. Role of intracellular Ca²⁺ stores in Ca²⁺ transients and CTCs

Ca²⁺ store filling mechanisms provided by sarco/endoplasmic reticulum Ca²⁺ATPases (SERCA) pump are important regulators of Ca²⁺homeostasis and signaling in ICC GI organs [38, 52–55]. Therefore, we examined the molecular expression of the SERCA pump isoforms in gastric ICC. Atp2a2 and Atp2a3 were both are enriched in ICC, but transcripts of Atp2a1 were not resolved in ICC from the antrum and corpus (Fig. 10A; Table 1). ER Ca²⁺ channels (RyR and InsP₃R) are expressed in extracts of GI muscles [56–59]. We evaluated the expression of RyR and InsP₃R isoforms in gastric ICC. Abundant transcripts of Iptr1 were found in antrum and corpus ICC, as compared to Iptr2 and Iptr3 isoforms (Fig. 10B; Table 1). Ryr2 transcripts were more highly expressed than Ryr1 and Ryr3 (Fig. 10B; Table 1).

Fig. 10. Molecular expression of SERCA, RyR and InsP₃R transcripts.

A Relative expression of the SERCA pump isoforms (Atp2a1, Atp2a2 and Atp2a3) in gastric ICC and compared with unsorted cells dispersed from gastric antrum and corpus tissues obtained from Kit^+/copGFP mice. B Relative expression of ER Ca²⁺ channels (RyR and InsP₃R) in antrum and corpus ICC. InsP₃Rs encoded by (Iptr1, Iptr2 and Iptr3) and RyRs isoforms (Ryr1, Ryr2 and Ryr3). Expression was determined by qPCR and the relative expression of each gene was normalized to the house-keeping gene, Gapdh. All data graphed as mean ± SEM (n = 4).

Ca²⁺ release from intracellular stores is involved in generation of pacemaker currents and initiation of slow waves, as determined by electorphysiological methods [50, 55, 60]. We tested the effects of the SERCA pump antagonist, cyclopiazonic (CPA) on antrum slow waves to confirm the role of Ca²⁺ stores in murine gastric slow waves. Under the control conditions used in imaging studies (i.e. in the presence of nifedipine; 1 μM) gastric muscles displayed an average RMP of −69 ± 1.5 mV and generated spontaneous slow waves 38 ± 2.7 mV in amplitude, ½ maximal duration of 6.2 ± 0.4 s and frequency of 3.3 ± 0.2 cycles min⁻¹ (Fig. 11A; n = 10). CPA (10 μM; 20 min) depolarized RMP to −59 ± 3 mV, decreased slow wave amplitude to 14 ± 5.6 mV, reduced ½ maximal duration of slow waves to 1.0 ± 0.3 s and reduced frequency to 1.6 ± 0.4 cycles min⁻¹ (Fig. 11D–F; n = 5).

Fig. 11. Effect of cyclopiazonic acid on gastric antrum slow waves.

A Slow waves recorded under control conditions (in the presence of nifedipine (1μM) and after the addition of cyclopiazonic acid (CPA; 10 μM). B–E Summarized data of the effects of CPA (10 μM) on RMP (B), slow wave amplitude (C), slow wave frequency (D) and ½ maximal duration (E). ** P = <0.01; *** P = <0.001.

We examined whether Ca²⁺ release from intracellular stores affects Ca²⁺ signaling in ICC-MY. Thapsigargin (1 μM; A SERCA pump antagonist; 20 min) reduced but did not block, Ca²⁺ transient firing in antrum ICC-MY (Fig. 12A–H). Ca²⁺ PTCL area was reduced to 37.1 ± 9% (Fig. 12E; n = 6) and PTCL count was reduced to 33.8 ± 5.9% (Fig. 12F; n = 6). The number of firing sites was reduced by thapsigargin to 53.8 ± 4.9% (Fig. 12G; n = 6). Thapsigargin (1 μM) significantly reduced the CTCs duration to 21.4 ± 2.5% (Fig. 12H; n = 6) and the frequency of CTCs to 71.1 ± 11.2% (n = 6). After exposure for 40 min, thapsigargin (1 μM) further reduced Ca²⁺ transients in ICC-MY. PTCL area 4.9 ± 3% (Fig. 12E; n = 6), PTCL count 4.1 ± 2.4% (Fig. 12F; n = 6), number of firing sites 3.8 ± 2.3% (Fig. 12G; n = 6) and the CTCs duration were significantly reduced to 3.1 ± 1.6% (Fig. 12H; n = 6). Cyclopiazonic acid (CPA, 10 μM) also reduced Ca²⁺ transient firing (Fig. 12I–P). Ca²⁺ PTCL area was reduced to 16.8 ± 4% (Fig. 12M; n = 5) and PTCL count was reduced to 21.6 ± 2.9% (Fig. 12N; n = 5). The number of firing sites was reduced by CPA to 32.9 ± 2.8% (Fig. 12O; n = 5). CPA significantly reduced CTC duration to 20.1 ± 2.6% (Fig. 12P; n = 5) and frequency to 60 ± 5.8% (n = 5).

Fig. 12. Ca²⁺ stores contributions to Ca²⁺ transients in antrum ICC-MY.

A ICC-MY Ca²⁺ firing site activity in antrum ICC-MY are color-coded and plotted in occurrence maps under control conditions and in the presence of thapsigargin (1 μM) B. Plot traces of firing sites PTCL area (C; blue) and PTCL count (C; green) under control conditions and in the presence of thapsigargin PTCL area (D; blue) and PTCL count (D; green). Summary graphs of Ca²⁺ PTCL activity in ICC-MY in the presence of thapsigargin after 20 min and 40 min incubation periods are shown in E (PTCL area), F (PTCL count), G the number of PTCL active sites and H Average percentage changes of Ca²⁺ transient clusters (CTCs) duration (n = 6). CPA (SERCA pump inhibitor; 10 μM) reduced transients compared to control as shown in occurrence maps of firing sites I&J and Ca²⁺ activity traces K&L. Summary graphs of Ca²⁺ PTCL activity in ICC-MY in the presence of CPA are shown in M (PTCL area), N (PTCL count), O the number of PTCL active sites. P Total CTC duration (n = 5). Data were normalized to controls and expressed as percentages (%). Significance determined using unpaired t-test, **** = P < 0.0001. All data graphed as mean ± SEM.

Ca²⁺ transient firing in corpus ICC-MY was also reduced by thapsigargin (1 μM; 20 min) in a time dependent manner similar to what was observed in antrum ICC-MY (Fig. 13A–D). Ca²⁺ PTCL area was reduced to 27.8 ± 4.4% (Fig. 13E; n = 5), PTCL count was reduced to 28.6 ± 5% (Fig. 13F; n = 5) and the number of firing sites to 69.5 ± 3.8% (Fig. 13G; n = 5). CTC duration was reduced to 29.7 ± 5.5% by thapsigargin (Fig. 13H; n = 5). Ca²⁺ transients in corpus ICC-MY were further reduced by thapsigargin after 40 min. PTCL area 1.6 ± 1% (Fig. 13E; n = 5), PTCL count 1.4 ± 1.2% (Fig. 13F; n = 5), number of firing sites 0.8 ± 0.6% (Fig. 13G; n = 5) and the CTCs duration were significantly reduced to 3.8 ± 2.3% (Fig. 13H; n = 5).

Fig. 13. Ca²⁺ stores contributions to Ca²⁺ transients in corpus ICC-MY.

A ICC-MY Ca²⁺ firing site activity are color-coded and plotted in occurrence maps under control conditions and in the presence of thapsigargin (1 μM) B. Plot traces of firing sites PTCL area (C; blue) and PTCL count (C; green) under control conditions and in the presence of thapsigargin PTCL area (D; blue) and PTCL count (D; green). Summary graphs of Ca²⁺ PTCL activity in ICC-MY in the presence of thapsigargin after 20 min and 40 min incubation periods are shown in E (PTCL area), F (PTCL count), G the number of PTCL active sites and H Average percentage changes of Ca²⁺ transient clusters (CTCs) duration (n = 5). Data were normalized to controls and expressed as percentages (%). Significance determined using unpaired t-test, **** = P < 0.0001. All data graphed as mean ± SEM.

Store-operated calcium entry (SOCE) is a mechanism for maintenance of Ca²⁺ stores to sustain ER Ca²⁺ release. Ca²⁺ depletion from the ER causes STIM proteins to translocate to ER-plasma membrane (PM) junctions and bind to and activate Orai channels [45, 61–66]. Gastric antrum and corpus ICC showed enrichment in Orai1 and Orai3, and Orai2 showed a low level of expression (Fig. 14A; Table 1; n = 4). Transcripts of stromal interaction molecules, Stim1 and Stim2, were abundant in antrum and corpus ICC (Fig. 14A; Table 1; n = 4).

Fig. 14. Role of SOCE in maintaining ICC-MY Ca²⁺ transients.

A Relative expression of store-operated Ca²⁺ entry (SOCE) channels (Orai1, Orai2 and Orai3) and stromal interaction molecules STIM1 and STIM2 in gastric ICC and compared with unsorted cells dispersed from gastric antrum and corpus tissues obtained from Kit^+/copGFP mice. B Heat-map images of an antrum ICC-MY networks showing total active Ca²⁺ PTCLs under control conditions and in the presence of GSK-7975A (C, 10 μM, for 20 min). D & E occurrence maps of color-coded Ca²⁺ firing sites showing the effect of the SOCE channel antagonist, GSK-7975A (10 μM) on ICC-MY Ca²⁺ transients. Traces of PTCL area (F; blue) and PTCL count (F; green) under control conditions and in the presence of GSK-7975A, PTCL area (G; blue) and PTCL count (G; green). Summary graphs of Ca²⁺ PTCL activity in ICC-MY in the presence of GSK-7975A are shown in H (PTCL area), I (PTCL count), J the number of PTCL active sites and K Average percentage changes of Ca²⁺ transient clusters (CTCs) duration (n = 6). Significance determined using unpaired t-test, **** = P < 0.0001. All data graphed as mean ± SEM.

We examined the role of SOCE in maintenance Ca²⁺ transients in ICC-MY using an Orai antagonist: GSK 7975A (10 μM). GSK 7975A reduced the frequency of CTCs (Fig. 14B–G). Firing site occurrence (Fig. 14D&E) and Ca²⁺PTCL areas and counts were also reduced (Fig. 14F&G). Ca²⁺PTCL area was reduced to 52.1 ± 4.1% (Fig. 14H; n = 6), and the PTCL count was reduced to 55 ± 4.3% (Fig. 14I; n = 6). GSK 7975A also inhibited the number of firing sites to 70.5 ± 2.7% (Fig. 14J; n = 6). GSK 7975A also significantly reduced the CTC duration to 70.5 ± 4.8% (Fig. 14K; n = 6) and frequency to 73.5 ± 8.6% (n = 6). The data indicate that Ca²⁺influx via SOCE provides a mechanism to maintain Ca²⁺stores in ICC-MY.

4. Discussion

Slow wave activity is a fundamental behavior of gastric muscles and responsible for the peristaltic contractions that triturate solid foods into small particles to facilitate gastric emptying and efficient digestion and absorption of nutrients in the small intestine [4]. This study provides new insights into ICC mechanisms that lead to pacemaker dominance of the corpus, propagation of slow waves within ICC networks and why the plateau phase of gastric slow waves, that regulates excitation-contraction coupling and the force and duration of contractions [5], persists for seconds. We found that ICC-MY generate clusters of localized Ca²⁺ transients (CTCs) that correspond to the frequency of electrical slow waves. CTCs are initiated and organized by Ca²⁺ entry. Reduction in extracellular Ca²⁺ and a T-type Ca²⁺ channel antagonist reduced or blocked the rhythmic CTCs, as shown previously for slow waves [50]. When Ca²⁺ entry was blocked, Ca²⁺ transient activity resorted to stochastic activity similar to that observed in ICC lacking voltage-dependent Ca²⁺ entry mechanisms (e.g. ICC-DMP in the small intestine and ICC-IM in the colon) [34, 67]. Ca²⁺ transients in gastric ICC-MY resulted from release from intracellular stores, as indicated by the effects of SERCA pump antagonists. Analogous effects on electrical activity were observed with CPA, although it is realized that the effects of SERCA pump inhibitors might have complex effects in a syncytial tissue composed of multiple cell types. A role for store release of Ca²⁺ in ICC-MY was further supported by the effects of an ORAI channel antagonist, GSK-7975A, and suggested that stores are maintained, at least in part, by SOCE.

Electrophysiology experiments performed on gastric muscles from a variety of mammalian species have shown that isolated corpus muscles generate slow waves at a higher rate than isolated antrrum muscles [24, 26]. The frequency gradient leads to pacemaker dominance of the corpus in the intact stomach and facilitates spread of slow waves and contractions from the corpus through the antrum to the pyloric sphincter [68]. The reasons for dominance of the corpus vs. antrrum pacemaker cells was not determined by prior studies. An observation of major importance in the current study is that the pattern of Ca²⁺ transient firing differs dramatically in antrum and corpus ICC-MY. CTCs occur in ICC-MY in both regions of the stomach, but in antrum CTCs were followed by a dramatic refractory period, during which few Ca²⁺ transients occurred. However with time, refractoriness ebbed and sporadic Ca²⁺ transients developed. The occurrence of these events immediately preceded the onset the next CTC. This demonstrates the intrinsic pacemaker activity of the antrum, and shows it occurs at a lower frequency than in corpus. The pattern of Ca²⁺ transients was quite different in ICC-MY of the corpus. CTCs were followed by a definite reduction in firing frequency, but numerous Ca²⁺ transients persisted during the inter-CTC interval. Ca²⁺ release from stores activates Ano1 channels in ICC [38]. Ano1 currents depolarize ICC and bring the cells to the threshold for generation of slow wave currents [69]. The abundance of Ca²⁺ transients in corpus ICC-MY during the inter-CTC intervals increases the probability of STICs and STDs, the likelihood of summation of STDs and the probability of activating voltage-dependent Ca²⁺ conductances that initiate the slow wave upstroke potential and facilitate cell-to-cell active propagation [50]. These observations suggest that Ca²⁺ transients generated during the inter-CTC interval are the fundamental pacemaker events that initiate CTCs. The fact that inter-CTC Ca²⁺ transients are more prevalent in corpus ICC-MY is likely to explain why corpus pacemakers generate slow waves at higher frequencies and serve as the dominant pacemaker in the stomach.

Equivalent characteristics are also seen in electrophysiological recordings of slow waves from intact corpus and antrum muscles. Inter-slow wave depolarization of membrane potential can be observed in some cells in corpus muscles [24, 26, 70], and the inter-slow wave interval can include oscillations in membrane potential (like STDs) that precede the initiation of slow wave upstroke potentials (see Fig. 1). During the interval between slow waves, membrane potentials of antrum cells tend to be stable, and slow wave upstroke potentials in antrum muscles, when electrically coupled to the corpus, take off from an exponential ‘foot’ depolarization, indicating they are propagated events (see Fig. 1).

The refractory period for Ca²⁺ release following CTCs is a novel observation that is likely to be the key mechanism for regulation of slow wave frequency. Cessation of Ca²⁺ release events may be due to unloading of Ca²⁺ stores during CTCs. A similar refractory period was observed for generation of STDs (also termed ‘unitary potentials’) in which the frequency of STDs decreased after a slow wave and then gradually increased before initiation of the next slow wave [71]. These authors suggested that STDs occurring during the interval between slow waves triggers the upstroke of the next slow wave, a conclusion consistent with the findings of the current study. In canine and guinea pig gastric muscles the refractory period for the upstroke is relatively short, but several seconds are required to fully reset the plateau phase [72, 73]. Reloading of stores is consistent with the relatively long refractory periods for restoration of the plateau potential, as this phase depends upon Ca²⁺ release from stores and was greatly reduced in duration by SERCA pump antagonists. The long refractory periods for the plateau phase of gastric slow waves (~8–12 s) are consistent with the time required to refill Ca²⁺ stores after unloading episodes [74]. The long refractory periods of slow waves are unlikely to be a function of voltage dependent Ca²⁺ channels, as the activation and inactivation kinetics and resetting of these channels upon restoration of negative membrane potentials occur more rapidly than the long refractory periods of slow waves [75]. We speculate that reloading of Ca²⁺ stores restores the excitability of Ca²⁺ release channels in the ER and this is the primary determinant of slow wave frequency.

Our study showed how CTC duration can affect the frequency of these events and therefore slow waves. Blocking l-type Ca²⁺ channels (most likely Ca_V1.3) and modestly reducing extracellular Ca²⁺ (to 1 mM) reduced the durations of CTCs and increased their frequency. Reduced CTC duration was associated with fewer Ca²⁺ release events and therefore reduced store unloading. This might reduce the duration of the refractory period and explain the increase in CTC frequency. Future investigation of how pacing frequency affects store loading and the refractory periods of Ca²⁺ transients and CTCs will be needed to more fully understand how slow wave frequencies are regulated by neurotransmitters and hormones that display chronotropic effects.

Once initiated, slow waves spread actively through networks of ICC [21, 50, 76], but these events conduct passively into SMCs and through regions of smooth muscle devoid of ICC [8, 9]. Because antrum slow waves are so important for the peristaltic contractions of the distal. stomach and eventual emptying of solids [4], we devoted considerable attention to the underlying Ca²⁺ dynamics in antrum ICC-MY. There was a impressive clustering of discrete Ca²⁺ transients during each slow wave cycle (see Fig. 3H and supplemental Video 1). Ca²⁺ transients originated from a multitude of Ca²⁺ release sites in the processes and soma of ICC-MY. Clusters were initiated by an initial rise in [Ca²⁺]_i during the first 600 msec. This sharp transition in [Ca²⁺]_i likely represents Ca²⁺ entry as it persisted after addition of thapsigargin. After the initial rise in [Ca²⁺]_i due to entry, Ca²⁺ release sites were unleashed, causing a dynamic array of Ca²⁺ transients that was sustained for at least 4 s. Multiple release events were observed from most sites during each CTC cycle.

The extended period of Ca²⁺ release during CTCs correspond to the long durations of depolarization during the plateau phase of slow waves. Sustaining the plateau depolarization likely results from an extended period of activation of Ano1 channels. Direct recordings from ICC have shown that the plateau phase reaches a maximum depolarization to approximately −10 mV [9, 51, 77–80]. This potential approximates the reversal potential for STICs that are due to activation of Ano1 channels [81] and thus approximates the transmembrane equilibrium potential for Cl⁻ ions (E_Cl) in ICC.

Outward current mechanisms, via activation of voltage- or Ca²⁺ dependent K⁺ conductances, that might be involved in or required for slow wave repolarization have not been resolved in ICC [69]. This is a distinguishing feature of the excitable events generated by ICC and different from other excitable cells. The lack of an outward current conductance is likely to be a necessary feature for preserving the long durations of the slow wave plateau phase, because if a voltage- or Ca²⁺-dependent outward current developed upon depolarization or a rise in [Ca²⁺]_i, then it would compete with the dominance of the Cl⁻ conductance that tends to ‘clamp’ membrane potential near E_Cl during the plateau phase. When Ca²⁺ transients cease, the open probabilities of Ano1 channels are reduced and repolarization of slow waves occurs. In other words, membrane potentials of ICC-MY are bistable: one level (RMP or the inter-slow wave potential) is determined by dominant K⁺ conductances, and the other level (plateau phase of slow waves) is determined by the dominance of the Cl⁻ conductance (i.e. Ano1). Jumping between levels requires initial depolarization (provided by STICs and activation of a T-type Ca²⁺ conductance), Ca²⁺ entry and activation of intracellular Ca²⁺transients (as characterized in the current study), and activation of Ano1 channels. The plateau phase is terminated by cessation of Ca²⁺ transients and restoration of low open probability for Ano1 channels. Our results suggest that the dynamic changes in Ca²⁺ release provide the mechanism for activation and deactivation of Ano1 currents during the slow wave cycle.

The question of what sustains Ca²⁺ release events for several seconds during the plateau phase is important to consider. Ca²⁺ entry through a T-type conductance is likely to be very brief, yet some mechanism sustains the occurrence of Ca²⁺ transients, and these events can fire repetitively during a single CTC (Fig. 15). The plateau appears to be dependent upon sustained Ca²⁺ entry, as reducing extracellular Ca²⁺ and antagonists of L-type Ca²⁺ channels reduce the amplitude and duration of the plateau phase [5, 24, 50, 80, 82]. Genes encoding Ca_V1.2 and Ca_V1.3 (i.e. Cacna1c and Cacna1d) L-type channels were expressed in gastric ICC, and this family of channels displays window currents (i.e. sustained low level channel activation due to incomplete inactivation) [83] in the range of potentials experienced during the plateau phase [9, 80]. In the present study we found that isradipine, an equipotent antagonist of Ca_V1.2 (containing α1C) and Ca_V1.3 (containing α1D) channels [84, 85], reduced the durations of CTCs. Studies on small intestinal ICC-MY suggest that the Na⁺/Ca²⁺ exchanger (NCX) can flip into Ca²⁺ entry mode during the plateau depolarization due to the entry of Na⁺ resulting from the recovery of Cl⁻ via the Na⁺K⁺Cl⁻ exchanger (NKCC1) [48, 81]. Genes for NCX1 and NCX2 are expressed in gastric ICC, but this mechanism has not yet been investigated in gastric ICC. Entry of Ca²⁺ through ORAI channels, expected to be activated by Ca²⁺ release and store depletion, could also be a contributing factor. However, SOCE would not seem to be a major mechanism for sustaining the durations of Ca²⁺ transient clusters, as this would represent positive feedback that would tend to prevent repolarization from the plateau phase.

Fig. 15. Ca²⁺ signaling drives pacemaker activity in gastric ICC-MY.

Steps in Ca²⁺ signaling in relation to generation of slow waves in gastric muscles are shown. For clarity the various proteins functioning during the slow wave cycle appear in panels at the stages where they become functional. 1. During the inter-slow wave interval, Ca²⁺ release from ER causes Ca²⁺ transients that activate Ano1 channels in the plasma membrane (PM). Activation of Ano1 causes development of spontaneous transient inward currents (STICs) which have depolarizing influence and generate spontaneous transient depolarizations (STDs). 2. Slow wave upstroke. The depolarization from STDs activates T-type Ca²⁺ current (VDCC1) that rapidly depolarize ICC-MY close to 0 mV. Influx of Ca²⁺ contributes to activation of Ano1 channels. 3. Plateau phase. Depolarization caused by the slow wave upstroke activates L-type Ca²⁺ current (VDCC2). Ca²⁺ entry causes localized Ca²⁺ induced Ca²⁺ release, and a multitude of Ca²⁺ release sites leads to development of CTCs. Ca²⁺ transient during CTCs activate Ano1 channels and maintain the depolarized state during the plateau phase. 4. Repolarization. When stores are depleted, Ca²⁺ transients cease and the open probability of Ano1 channels decreases to low levels causing repolarization. Repolarization also causes deactivation of VDCCs. SOCE (not shown) and SERCA pumps restore store Ca²⁺ to reset the mechanism for the next slow wave cycle. Corpus and antrum ICC-MY both manifest intrinsic pacemaker activity, however the frequency of pacemaking in the corpus is higher than in antrum. The present study suggests that the major difference between the pacemakers in the corpus and antrum is that the probability of Ca²⁺ transient firing is higher in corpus ICC-MY than in antrum. It should also be noted that scheme illustrated in the figure is relevant only to Ca²⁺ transients and activation of conductances in ICC. In intact muscles ICC are electrically coupled to SMCs. Electrical activity is recorded typically from SMCs, so additional conductances contribute to the shaping of the waveforms of slow waves. For example, the rapid repolarization following the initial upstroke (see Fig. 1C) is due to activation of an A-type current in SMCs [86, 87].

In summary the long duration plateau phase in gastric muscles is mirrored by the long period of Ca²⁺ release events in ICC-MY. Ca²⁺ release in ICC activates Ano1 channels that push membrane potential toward E_Cl. Ca²⁺ transients are organized into CTCs by voltage-dependent Ca²⁺ entry through a T-type Ca²⁺ cconductance. CTCs appear to be initiated by Ca²⁺-induced Ca²⁺ release. CTCs are sustained for seconds by continuous Ca²⁺ entry through L-type Ca²⁺ channels, most likely Ca_V1.3 operating within the voltage range for window currents. CTCs are followed by refractory periods during which the occurrence of Ca²⁺ transients is reduced. Initiation of CTCs corresponds to waning of the refractory period, increase in spontaneous Ca²⁺ transients, summation of Ca²⁺ transients, summation of Ano1 currents and depolarization of ICC-MY to the threshold for activation of T-type Ca²⁺ channels (Fig. 15). The occurrence of Ca²⁺ transients in corpus ICC-MY during the inter-CTC interval is likely to explain why the frequency of corpus slow waves is higher than in antrum and why the corpus ICC-MY serve as the dominant pacemaker for slow waves in the stomach.

7. Data and materials availability

All data are available in the main text or the supplementary materials