Muscarinic activation of Ca2+-activated Cl current in interstitial cells of Cajal


  • M. H. Zhu and I. K. Sung contributed equally to this work.

Kenton M. Sanders: Department of Physiology and Cell Biology, University of Nevada School of Medicine, Reno, NV 89557, USA.  Email:


Non-technical summary  Interstitial cells of Cajal (ICC) are tightly associated with excitatory and inhibitory motor neurons in the gastrointestinal tract, and these cells are also connected electrically to smooth muscle cells. We have suggested that ICC participate in responses to neurotransmitters released from neurons that drive motility and help move nutrients and wastes through the gut. We studied responses of isolated ICC to cholinergic neurotransmitter and found that a Ca2+-activated Cl current is activated in ICC in response to cholinergic stimulation. Such a current would result in depolarization that could be conducted to surrounding smooth muscle cells by the electrical connections. Exciting ICC would cause generalized excitation of the smooth muscle tissue. A different conductance is activated in smooth muscle cells by cholinergic stimulation. We tested drugs that blocked the Cl current in ICC and found that responses to nerve stimulation in intact intestinal muscles were blocked by these drugs. This suggests that ICC mediate electrical responses to cholinergic nerve stimulation. In some human gastrointestinal motility disorders, ICC are damaged or lost. If these cells provide responses to neurotransmitters, this might provide an explanation for motor dysfunction in the gut.


Abstract  Interstitial cells of Cajal (ICC) provide pacemaker activity and functional bridges between enteric motor nerve terminals and gastrointestinal smooth muscle cells. The ionic conductance(s) in ICC that are activated by excitatory neural inputs are unknown. Transgenic mice (KitcopGFP/+) with constitutive expression of a bright green fluorescent protein were used to investigate cellular responses of ICC to cholinergic stimulation. ICC displayed spontaneous transient inward currents (STICs) under voltage clamp that corresponded to spontaneous transient depolarizations (STDs) under current clamp. STICs reversed at 0 mV when ECl= 0 mV and at –40 mV when ECl was –40 mV, suggesting the STICs were due to a chloride conductance. Carbachol (CCh, 100 nm and 1 μm) induced a sustained inward current (depolarization in current clamp) and increased the amplitude and frequency of STICs and STDs. CCh responses were blocked by atropine (10 μm) or 4-DAMP (100 nm), an M3 receptor antagonist. STDs were blocked by niflumic acid and 5-nitro-2-(3-phenylpropylamino)-benzoic acid (both 100 μm), and CCh had no effect in the presence of these drugs. The responses of intact circular muscles to CCh and stimulation of intrinsic excitatory nerves by electrical field stimulation (EFS) were also compared. CCh (1 μm) caused atropine-sensitive depolarization and increased the maximum depolarization of slow waves. Similar atropine-sensitive responses were elicited by stimulation of intrinsic excitatory neurons. Niflumic acid (100 μm) blocked responses to EFS but had minor effect on responses to exogenous CCh. These data suggest that different ionic conductances are responsible for electrical responses elicited by bath-applied CCh and cholinergic nerve stimulation.




Ca2+-containing physiological salt solution




4-diphenylacetoxy-N-(2-chloroethyl) piperidine hydrochloride


deep muscular plexus


green fluorescent protein


Discosoma sp. red fluorescent protein


electrical field stimulation




interstitial cells of Cajal




NG-nitro-l-arginine; l-NG-nitroarginine




niflumic acid


5-nitro-2-(3-phenylpropylamino)-benzoic acid


resting membrane potential


spontaneous transient depolarizations


spontaneous transient inward currents


canonical transient receptor potential




Spontaneous electrical slow waves time phasic contractions in gastrointestinal (GI) smooth muscles that result in peristaltic contractions and segmentation (Szurszewski, 1987). The force of contractions, and therefore the propulsive potential of contractions in intact organs, is determined, in part, by the degree to which slow waves couple to activation of voltage-dependent, L-type Ca2+ channels in smooth muscle cells (Ozaki et al. 1991). Activation of Ca2+ channels is determined by the level of depolarization during slow waves, and in small intestinal muscles slow waves can initiate smooth muscle action potentials (Szurszweski, 1987). Slow waves are regulated on a moment-to-moment basis by excitatory and inhibitory enteric motoneurons that innervate the tunica muscularis. In the case of excitatory motoneurons, major neurotransmitters include acetylcholine (ACh) and neurokinins (substance P and neurokinin A; Shuttleworth & Keef, 1995). However, responses to excitatory nerve stimulation at frequencies less than 10 Hz are typically blocked by muscarinic antagonists (e.g. Forrest et al. 2006). Therefore, post-junctional mechanisms coupled to ACh binding to muscarinic receptors (M2 and/or M3 receptors; Zholos & Bolton, 1997; Tobin et al. 2009) are important factors in neural regulation of GI motor patterns.

Smooth muscle cells of the GI tract express muscarinic receptors and binding of ACh activates non-selective cation channels in these cells (Benham et al. 1985; Inoue & Isenberg, 1990). These channels are encoded by canonical transient receptor potential (TRPC) channels, types 4 and 6 (i.e. TRPC4 and TRPC6), as shown by recent experiments using selective gene deactivation (Tsvilovskyy et al. 2009). Inward currents through these channels depolarize smooth muscle cells and couple to the increased open probability of voltage-dependent Ca2+ channels (Unno et al. 2003). This sequence of events explains the major excitatory effect of exogenous ACh and other muscarinic agonists on GI smooth muscles. Other studies have suggested that excitatory motoneurons innervate interstitial cells of Cajal (ICC) in GI muscles, and muscarinic effects are greatly diminished when these cells are absent (Ward et al. 2000). ICC of the small intestine have also been shown to express a variety of receptors for motor neurotransmitters, including M2 and M3 receptors (Chen et al. 2007). Thus, motor responses to neurotransmitter released from motoneurons may differ from responses to exogenous agonists, because a significant portion of responses to transmitters released from neurons may be mediated by ICC. At present we know very little about responses elicited in ICC by muscarinic agonists.

In a previous study, we investigated effects of ACh on Kit+ cells (an immunological marker for ICC; Ward et al. 1994) grown in culture for several days after enzymatic isolation from the tunica muscularis (Kim et al. 2003). ACh activated a non-selective cation conductance in these cells. In later studies it was found that significant phenotypic changes occur in Kit+ cells in cell culture. Native ICC express a Ca2+-activated Cl conductance that is fundamental to the pacemaker function of ICC, but this conductance is lost after a few days in culture (Hwang et al. 2009; Zhu et al. 2009). Inward current in ICC results from robust expression of ANO1 (encoded by mTmem16a) (Gomez-Pinilla et al. 2009; Hwang et al. 2009), which generates Ca2+-activated Cl channels (Caputo et al. 2008; Schroeder et al. 2008; Yang et al. 2008). In the present study we investigated whether cholinergic agonists activate Ca2+-activated Cl channels in ICC of the murine small intestine. ICC were identified after enzymatic dispersion by the constitutive expression of a bright green fluorescent protein (copGFP; Zhu et al. 2009; Ro et al. 2010). Finding differences in the ion channels that are activated by muscarinic agonists may provide a means to assess the relative contributions of ICC and smooth muscles to cholinergic excitatory neurotransmission.



C57BL/6 (2 and 3 months old, Charles River Laboratories, Wilmington, MA, USA) and KitcopGFP/+ mice (7 and 12 days old), as described previously (Zhu et al. 2009), were used for these experiments. Animals were anaesthetized with isoflurane (Bar Harbor, MN, USA) prior to decapitation, and small intestines were removed. The institutional Animal Use and Care Committee at the University of Nevada approved all procedures used in the breeding and killing of animals.

Intracellular recording

Intracellular recordings of membrane potential were made from jejunal muscle strips (C57BL/6) pinned to the Sylgard (Dow Corning Corp, Midland, MI, USA) floor of a recording chamber with the circular muscle layer facing upward. The recording chamber was constantly perfused with Krebs Ringer buffer (KRB) containing (mm): NaCl 120; KCl 5.9; MgCl2 1.2; NaHCO3 15.5; NaH2PO4 1.2; dextrose 11.5; and CaCl2 2.5. The pH of KRB was 7.3–7.4 when bubbled with 97% O2–3% CO2 at 37.0 ± 0.5°C (Hwang et al. 2009). Circular muscle cells were impaled with glass microelectrodes filled with 3 m KCl and having tip resistances of 70–100 MΩ. Muscles were stimulated by drugs dissolved in the KRB perfusate or by electrical field stimulation (EFS; 0.3 ms, 5–10 Hz, 150 V, 10 s train duration) delivered by two platinum electrodes placed on either side of the muscle strip and supplied by a physiological stimulator (S48 stimulator, Grass Technologies, West Warwick, RI, USA). Responses to EFS were due to activation of intrinsic nerves and blocked by tetrodotoxin (1 μm).

Smooth muscle cell transmembrane potentials were measured with a high-input impedance amplifier (Duo 773; WPI, Sarasota, FL, USA), digitized using an analog-to-digital converter (Digidata 1300 series; Molecular Devices, Inc., Sunnyvale, CA, USA) and stored on a computer running Axoscope (version 10.0, Molecular Devices). All data were analysed using clampfit (version 10.0, Molecular Devices) and Graphpad Prism (version 3.0, Graphpad Software Inc., San Diego, CA, USA) software.

Isolation of cells for patch-clamp studies

Muscle strips of murine small intestine were equilibrated in Ca2+-free Hanks’ solution for 20 min. Cells were dispersed from these strips with an enzyme solution containing: collagenase (Worthington Type II, 1.3 mg ml−1), bovine serum albumin (Sigma, St Louis, MO, USA, 2 mg ml−1), trypsin inhibitor (Sigma, 2 mg ml−1) and ATP (Sigma, 0.27 mg ml−1). Cells were plated onto sterile glass coverslips coated with murine collagen (2.5 mg ml−1, Falcon/BD) in 35 mm culture dishes. Membranes of freshly dispersed cells were extremely fragile and giga seals broke down rapidly when these cells were used immediately after dispersion. Therefore, the cells were allowed to stabilize overnight at 37°C in a 95% O2–5% CO2 incubator in smooth muscle growth medium (Clonetics, San Diego, CA, USA) supplemented with 2% antibiotic–antimycotic (Gibco, Grand Island, NY, USA) and stem cell factor (5 ng ml−1, Sigma). Ca2+-activated Cl currents in ICC were preserved after this short stabilization period, as shown previously (Zhu et al. 2009).

Homologous expression of mTMEM16a and muscarinic recptors (Chrm3) in HEK 293 cells

Tmem16a (GenBank: NM_178642) in pYX-ASC vector (IMAGE: 30547439) and Chrm3 (GenBank: NM_033269) in pCR-Blunt II-TOPO vector (IMAGE: 40111499) were obtained from Open Biosystems (Huntsville, AL, USA). We previously reported that mouse GI tissues predominantly express the Tmem16a‘c’ variant which contains the alternative exon 13 coding for four amino acid residues (Hwang et al. 2009). Thus, we generated the Tmem16a(c) variant by inserting the 12 base pair ‘c’ segment into Tmem16a coding region using the QuickChange XL site-directed mutagenesis kit (Agilent Technologies, Santa Clara, CA, USA). For expression studies, Tmem16a(c) transcript was cloned into pcDNA4/TO/c-myc-HIS vector (Invitrogen) as described previously (O’Driscoll et al. 2008). Green fluorescent protein (GFP) coding sequence was inserted at EcoR1 and Not1 restriction sites to generate a GFP–Tmem16a(c) fusion protein. Chrm3 transcript was inserted into pDsRed-Express vector (Clontech). All plasmids were sequenced at the Nevada Genomics Centre.

Functional expression of Tmem16a(c) alone or in combination with Chrm3 was carried out in HEK 293 cells. HEK 293 cells were cultured in DMEM (Gibco) medium supplemented with FBS (10%, v/v), penicillin–streptomycin (1%, v/v) and Glutamax (1%, v/v; Invitrogen). Cells were plated onto 35 mm culture dishes and transfected with 1 mg plasmid DNA using FuGene6 (Roche) following the manufacturer's protocol. Transfected cells were seeded at low density on glass coverslips 24 h later and patch-clamp recordings were performed 24–48 h later. Expression of Tmem16a and Chrm3 were monitored by GFP and DsRed fluorescent reporter molecule expression, respectively.

Patch-clamp experiments

Green fluorescent protein-positive cells (i.e. ICC and HEK 293 cells transfected with mTMEM16a cDNA) were identified using inverted fluorescence microscopes. Conventional dialysed whole-cell patch-clamp configuration was used to record membrane currents (voltage-clamp mode) and membrane potentials (current-clamp, I= 0, mode) of ICC and HEK 293 cells. In single channel recordings, cell-attached and inside-out configurations were used. The pipette tip resistance ranged between 3 and 6 MΩ for whole-cell recordings and 5–10 MΩ for single channel recordings. Whole-cell and single channel experiments were conducted at 30°C, and temperature was maintained with a CL-100 bath heater (Warner Instruments; Hamden, CT, USA) and Thermoclamp-1 (Automate Scientific; Berkeley, CA, USA). Membrane currents or transmembrane potentials were recorded with an Axopatch 200B patch-clamp amplifier (Molecular Devices) and digitized with a 16-bit analog-to-digital converter (Digidata 1322A and 1440A, Molecular Devices) and stored directly on-line using pCLAMP software (version 9.2 and 10.0, Molecular Devices). Data were sampled at 4 kHz and filtered at 2 kHz for whole-cell experiments and 10 kHz sampling with 1 kHz filtration for single channel experiments. MiniDigi with Axoscope (version 9.2, Molecular Devices) was used for monitoring the changes in holding currents (basal currents) throughout experiments. All data were analysed using clampfit (version10.0, Molecular Devices) and Graphpad Prism (version 3.0, Graphpad Software) software.

Solutions for patch-clamp experiments

External solution for whole-cell recordings was a Ca2+-containing physiological salt solution (CaPSS) containing (mm): 135 NaCl, 5 KCl, 2 CaCl2, 1.2 MgCl2, 10 glucose and 10 Hepes adjusted to pH 7·4 with Tris. Table 1 (solution I–III) shows the composition of the pipette solutions used for the whole-cell recordings in ICC. For cell-attached configuration of single channel recordings, external solution contained (in mm): 140 KCl, 1 EGTA, 10 Hepes and 0.88 CaCl2, adjusted to pH 7.2 with Tris (HK solution). Pipette solution contained (mm): 140 NMDG (N-methyl-d-glucamine)-Cl, 10 EGTA and 10 Hepes, adjusted pH 7.2 with Tris. This pipette solution was used for external solution for inside-out patches. Free Ca2+ concentrations were calculated by Maxchelator software ( A fast solution exchange system (AutoMate Scientific, Inc., Berkeley, CA, USA) was used to test the effects of drugs on ICC. In whole-cell recordings from HEK 293 cells, external and internal solution were solution IV (see Table 1). Free [Ca2+] in the internal solutions was calculated by adding the amount of Ca2+ as determined by Maxchelator (see above). The recording chamber was perfused by external solution at room temperature (20–21°C).

Table 1.  The composition of pipette solutions
Solutions (mm)IIIIIIIV
  1. CPD, creatine phosphate disodium. Pipette solutions were adjusted to pH 7.2 with Tris.

Potassium aspartate0000
Caesium aspartate001100
Junction potential (mV)5.35.314.6 


Nicardipine, NG-nitro-l-arginine; l-NG-nitroarginine (l-NNA), atropine, niflumic acid (NFA), 4-diphenylacetoxy-N-methylpiperidine methiodide (4-DAMP) and 5-nitro-2-(3-phenylpropylamino)-benzoic acid (NPPB) were purchased from Sigma. 2—Iodo-N6-methyl-(N)-methanocarba-2-deoxyadenosine-3′,5′-bisphosphate (MRS2500) and tetrodotoxin (TTX) were purchased from Tocris (Ellisville, MO, USA). To make stock solutions, drugs were dissolved in water, ethanol or dimethyl sulphoxide as per manufacturer's recommendations. Stock solutions were made by dissolving l-NNA (100 μm), TTX (1 μm), MRS2500 (1 μm) and 4-DAMP (100 μm) in water, atropine (1–10 μm) in ethanol, and nicardipine (1 μm), NFA (100 μm) and NPPB (100 μm) in dimethyl sulphoxide (DMSO). Final dilutions were made by adding stock solutions to the bath or extracellular solutions used for specific experiments, as specified in the text and Table 1. Final concentrations of ethanol and diemthyl sulphoxide were was less than 0.1%.

Statistical analysis

Data were expressed as means ± SEM. The n values given represent the number of tissues in conventional microelectrode recordings and mechanical experiments and the number of cells in patch-clamp experiments. Differences between data sets were evaluated by Student's paired t test and considered significantly different when P < 0.05.


Effects of exogenous carbachol and cholinergic neurotransmission on small intestinal muscles

We first documented electrical responses of mouse jejunum circular muscle strips to excitatory nerve stimulation and exogenous carbachol (CCh). Electrophysiological experiments were performed in the presence of nicardipine (1 μm) to reduce contractions and allow maintenance of intracellular impalements. Muscles were stimulated with electrical field stimulation (EFS, 5 and 10 Hz, 0.3 ms duration pulses for 10 s). Under control conditions EFS (5 and 10 Hz) elicited hyperpolarization and reduced the maximal degree of depolarization during slow waves (Fig. 1A and D), indicating the dominance of inhibitory neurons with generalized nerve stimulation. l-NNA (100 μm) and MRS2500 (1 μm), inhibitors of nitrergic and purinergic (i.e. P2Y1 receptors) post-junctional pathways (Mutafova-Yambolieva et al 2007), respectively, were added to block major inhibitory neurotransmitters, and under these conditions, EFS elicited depolarization, increased the maximum level of depolarization during slow waves, and increased the duration of slow waves (Fig. 1B and E). Atropine (1 μm), added 15 min before EFS, blocked the ‘on-response’ during stimulation (Fig. 1C and F), but as shown in Fig. 1A and D, a post-stimulus depolarization was unmasked by atropine. The nature of the post-stimulus response was not further investigated. Fig. 1G–N show summaries of the effects of excitatory nerve stimulation on resting membrane potential (RMP; most negative level between slow waves), slow wave peak-to-peak interval, maximal level of depolarization attained by slow waves during stimulation, and the duration of slow waves (duration at half-amplitude).

Figure 1.

Effects of electrical field stimulation (EFS) on membrane potentials and electrical slow waves in circular muscle cells of the murine small intestine
A and D, EFS (5 and 10 Hz) (black bars) under control (muscle perfused with KRB) conditions induced hyperpolarization and increased the duration between slow waves (P-P), indicating a slight reduction in frequency. B and E, pre-treatment with MRS2500 (MRS, 1 μm) and l-NNA (100 μm), during the same impalement, depolarized cells, and EFS (5 and 10 Hz) caused further depolarization, increased P-P, maximum depolarization during slow waves and the duration of slow waves. C and F, pre-treatment with atropine (Atr, 1 μm) in the continued presence of l-NNA and MRS2500 abolished the depolarization during EFS (5 and 10 Hz) and blocked other changes to slow waves. A post-stimulus depolarization was unmasked by atropine (see text). G–N, summarized responses to EFS, showing effects on data on membrane potentials (RMP; G and K), peak-to-peak intervals between slow waves (P-P; H and L), maximum depolarization achieved during slow waves (Max Depol; I and M) and duration of slow wave (SW) (half-duration (T1/2) of SWs; J and N) in unstimulated (Cont) and during EFS (5 and 10 Hz as indicated). *P < 0.05, **P < 0.01. Control solution, KRB (Krebs Ringer buffer); M+L, MRS2500+l-NNA; M+L+A, MRS2500+l-NNA+atropine at concentrations given above.

Slow waves are generated by activation of Ca2+-activated Cl currents (Hwang et al. 2009), so we tested the effects of niflumic acid (NFA) on responses to EFS. After addition of MRS2500 and l-NNA to block inhibitory neural responses (Fig. 2A and D), EFS caused depolarization and increased the maximal depolarization attained during slow waves (Fig. 2B and E; as above). Pre-treatment with NFA (100 μm) reduced the amplitudes of slow waves from 20 ± 4.2 mV to 8 ± 4.4 mV (P < 0.0019) and largely blocked responses to nerve stimulation (Fig. 2C and F). Figure 2G–L summarize the effects of EFS (5 and 10 Hz) on RMP, slow wave frequency and the maximal depolarization achieved during slow waves before and after NFA in five muscles.

Figure 2.

Effects of niflumic acid on responses to EFS
A and D, EFS (5 and 10 Hz) under control (KRB) conditions induced hyperpolarization as in Fig. 1. B and E, pre-treatment with MRS2500 (MRS, 1 μm) and l-NNA (100 μm) depolarized cells a few millivolts, and caused further depolarization during EFS (5 and 10 Hz). C and F, pre-treatment with niflumic acid (NFA 100 μm), in the continued presence of l-NNA and MRS2500, abolished responses during EFS (5 and 10 Hz). G–I, summarized data on effects of EFS (5 and 10 Hz) on resting membrane potentials (RMP), peak-to-peak interval (P-P) and maximum depolarization (Max Depol) in unstimulated (control values before stimulation were in the presence of M+L and NFA) and during EFS (5 Hz). Effects of EFS were blocked by NFA.

We also tested the effects of exogenous CCh on slow waves (Fig. 3A). CCh (in the presence of 1 μm TTX) caused depolarization, reduced the frequency of slow waves, enhanced the maximal level of depolarization during slow waves, and increased the duration of slow waves (Fig. 3A and B). These responses were blocked by atropine (1 μm). Figure 3C–F summarizes the effects of CCh on electrical activity. The effects of CCh were also tested after pre-treatment with NFA. As above, NFA (100 μm) reduced slow wave amplitude, but in contrast to the effects noted on responses to EFS, NFA had little effect on responses to CCh (Fig. 4A and B). Note particularly that the depolarization in RMP caused by CCh was unchanged by NFA. Data for experiments of this type on five muscles are summarized in Fig. 4C–E.

Figure 3.

Effects of an exogenous muscarinic agonist, carbachol (CCh), on membrane potentials and slow waves of small intestine smooth muscle cells
A, representative long-term recording during a single impalement showing that carbachol (CCh, 1 μm) depolarized jejunal smooth muscle cells, and pre-treatment of atropine inhibited effects of CCh (in the presence of tetrodotoxin (TTX; 1 μm) to block activation of enteric motoneurons). Dotted line denotes the control resting membrane potential. Sweep speed in A is too slow to resolve individual slow waves. Ba–c show expanded time scales and slow wave details from the record in A. Note that CCh caused depolarization of cells and increased the maximal depolarization achieved during slow waves. C–F, summarized data showing effects of CCh (and the inhibitory effects of atropine (Atr) on responses to CCh) on resting membrane potential (RMP), peak-to-peak intervals between slow waves (P-P; **P= 0.009, *P= 0.016), maximum depolarization achieved during slow waves (Max Depol), and half-durations of slow wave (T1/2 of SW). In E and F, *P < 0.05.

Figure 4.

Effects of niflumic acid on responses to carbachol in murine jejunal smooth muscle cells
A, long-term recordings showing typical responses of jejunal cells to carbachol (CCh, 1 μm) and after pre-treatment with niflumic acid (NFA, 100 μm). Dotted line denotes the control resting membrane potential. Ba–d, expanded time-scale traces showing details of slow waves at times designated in A. C–E, summarized data showing effects of CCh (and effects of CCh in the presence of NFA) on membrane potentials (RMP), peak-to-peak intervals between slow waves (P-P), and maximum level of depolarization achieved during slow waves (Max Depol) in control (Cont), CCh, NFA, and NFA with CCh (NFA+CCh). *P < 0.05, **P < 0.01, ***P < 0.001. Note that NFA had little effect on responses to CCh.

Effects of carbachol on isolated ICC

We tested the effects of CCh on membrane potential and spontaneous transient depolarizations (STDs) under current clamp (I= 0) and on membrane currents and spontaneous transient inward currents (STICs) under whole-cell voltage clamp. For current-clamp experiments, external (CaPSS) and pipette solution 1 (Table 1) were used. Resting membrane potential (RMP; most negative potentials between STDs) in ICC averaged –58 ± 1.9 mV, STDs depolarized to a maximum level of –7 ± 0.8 mV and occurred at an average frequency of 18.3 ± 0.5 cycles min−1 (n= 7) during control conditions. CCh (100 nm) depolarized ICC to –45 ± 2 mV (n= 7, P < 0.01) and STDs reached a maximum depolarization of –5 ± 0.9 mV (P < 0.01). The frequency of STDs also increased (to 26.3 ± 0.5 cycles min−1, n= 7, P < 0.001; Fig. 5A, C and E–G). A higher concentration of CCh (1 μm) further depolarized ICC (to –28 ± 1.8 mV; n= 6, P < 0.01 in comparison to control, Fig. 5B, D and E). The frequency and maximal depolarization of STDs during exposure to 1 μm carbachol were difficult to determine because of the development of irregular membrane potential oscillations (n= 6, Fig. 5B and D).

Figure 5.

Effects of carbachol (CCh) on the resting membrane potential (RMP) and spontaneous transient depolarizations (STDs) of ICC in current-clamp (I= 0) mode
A and B, application of CCh (100 nm in A and 1 μm in B) depolarized cells and increased frequency of STDs in ICC from KitcopGFP/+ mice. External and internal solutions were CaPSS and solution I (see Table 1), respectively. C and D, traces with expanded time scales from A (times denoted by a and b) and B (times denoted by c and d) showed details of the effects of CCh. EG, summarized data showing responses to CCh on resting membrane potential (RMP), maximum depolarization achieved during STDs (Max Depol) and frequency of STDs, respectively. cpm, cycles per minute. **P < 0.01, ***P < 0.001.

ICC of murine small intestine express ANO1 (encoded by mTmem16a; Zhu et al. 2009), so we tested the effects of NFA and NPPB on STDs. In experiments with NFA, RMP averaged –65 ± 3.3 mV and STDs occurred at an average frequency of 15.3 ± 1.2 cycles min−1 (n= 7). NFA (100 μm) hyperpolarized cells to –74 ± 2.1 mV and inhibited STDs. Responses to CCh (1 μm) were blocked by NFA (Fig. 6A–C). In the experiments in which we used NPPB, control RMP averaged –69 ± 3.6 mV and STDs occurred at an average frequency of 16.8 ± 0.4 cycles min−1 (n= 5). NPPB (100 μm) hyperpolarized cells to –81 ± 0.7 mV, inhibited STDs, and blocked responses to CCh (1 μm) (Fig. 6D–F).

Figure 6.

Effects of Cl channel-blocking drugs (NFA and NPPB) on responses of ICC to CCh in current-clamp (I= 0) mode
A, NFA (100 μm) induced hyperpolarization and decreased and eventually blocked STDs. CCh (1 μm) in the presence of NFA had no effect on RMP or STD generation. External and internal solutions were CaPSS and solution I (see Table 1), respectively. B and C, summarized data in control, niflumic acid (NFA) and NFA+CCh on RMP and STD frequency, respectively. D, NPPB (100 μm) induced hyperpolarization and blocked STDs. CCh (1 μm) in the presence of NPPB had no effect on RMP or STD generation. External and internal solutions were CaPSS and solution I (see Table 1), respectively. E and F, summarized data from 5 cells in control, NPPB, and NPPB+CCh on RMP and STD frequency, respectively. *P < 0.05, **P < 0.01, ***P < 0.001.

Effects of carbachol on holding current and STICs in ICC

ICC were next studied under voltage clamp to investigate the nature of the conductances responsible for membrane potential effects. In these experiments we used Cs+-rich pipette solutions (solutions II and III, Table 1), to prevent contamination from K+ conductances, and CaPSS as the external solution. Cells were held at –80 mV and stepped to various potentials to measure steady-state responses. Spontaneous transient inward currents (STICs) were noted in steady-state recordings, and these events reversed at 0 mV when cells were dialysed with solution II (ECl= 0; Fig. 7A; n= 6). STICs reversed at approximately –30 mV (i.e. before junction potential correction, see Table 1), when cells were dialysed with solution IV (ECl=–40 mV; n= 4; Fig. 7B).

Figure 7.

Current–voltage relationship of holding current (HC) and spontaneous transient inward currents (STICs) in ICC
A, STICs and HC were recorded while stepping cell to potentials from –60 to +20 mV under whole-cell voltage clamp (external solution was CaPSS and pipette solution II, Table 1, ECl= 0 mV). HC and STICs were reversed at 0 mV. B, STICs and HC were recorded while stepping cell to potentials from –80 to +10 mV under whole-cell voltage clamp using internal solution III (Table 1). When ECl was adjusted to –40 mV in this experiment, HC and STICs reversed at about –30 mV (before correction of junction potential). Dotted line denotes 0 pA. a and b, expanded time scale from portions of records in A and B denoted by arrows.

CCh (100 nm and 1 μm) increased the amplitude and frequency of STICs (Fig. 8A and B). For example, STICs increased in amplitude from 34.3 ± 1.9 to 48.3 ± 3.7 pA (P < 0.05) and in frequency from 27.2 ± 2.7 to 39.8 ± 3.1 cycles min−1 in response to 100 nm CCh (n= 4, P < 0.01) at a holding potential of –80 mV. We used steady-state changes in holding potential to investigate reversal potentials of holding current (HC) and STICs activated by CCh. In these experiments, HC averaged –29.7 ± 4.1 pA at a holding potential of –80 mV during control conditions. CCh (1 μm) increased HC to –83.9 ± 13.3 (n= 4, P < 0.05) in cells dialysed with solution III (ECl=–40 mV). Both HC and STICs reversed at about –30 mV (Fig. 8C). Figure 8D–F summarizes the effects of CCh on HC and STICs.

Figure 8.

Effects of CCh on holding current (HC) and STICs
A and B, CCh (100 nm and 1 μm) increased the amplitude and frequency of STICs and increased HC (holding potential, –80 mV) using CaPSS as the external solution and pipette solution II (Table 1). C, HC and STICs, enhanced by CCh (1 μm), reversed at about –30 mV (before junction potential correction) when ECl=–40 mV (solution III; Table 1). D–F, summarized data showing effects of CCh (100 nm and 1 μm) on holding current (D), STIC amplitude (E) and STIC frequency (F). *P < 0.05, **P < 0.01, ***P < 0.001.

Effects of muscarnic antagonists on responses to carbachol

CCh generates excitatory motor responses in the GI smooth muscle cells via muscarinic (M2 and M3) receptors (Zholos & Bolton, 1997). We tested the effects of CCh on the membrane potentials before and after atropine under current clamp (I= 0). External and internal solutions were CaPSS and solution I, respectively (see Table 1 and Methods). The RMP and frequency of STDs were –59 ± 2.4 mV and 16 ± 2 cycles min−1, respectively (n= 5). After pre-treatment with atropine (10 μm), CCh (1 μm) had no significant effect on RMP or STDs (Fig. 9A; P > 0.05). Since M3 receptors are highly expressed in ICC (Chen et al. 2007), we also tested the effects of an M3 receptor blocker, 4-DAMP, on CCh responses under current clamp (I= 0). Pre-treatment with 4-DAMP (100 nm) did not affect RMP or the STDs, but blocked responses to CCh (Fig. 9B; both P > 0.37).

Figure 9.

Effects of muscarinic receptor antagonists in responses to CCh
A and B, muscarinic antagonists atropine (A) and 4-DAMP (B) were tested on responses to CCh under current clamp (I= 0) with external solution (CaPSS) and pipette solution I (Table 1). Pre-treatment with atropine (10 μm) or 4-DAMP (100 nm) blocked CCh effects on the RMP and STDs. C and D, pre-treatment with atropine (10 μm) or 4-DAMP (100 nm) antagonized CCh-induced increase in HC and STICs at a holding potential of –80 mV (solution II, Table 1).

We also examined the effects of atropine and 4-DAMP under voltage-clamp conditions (external solution was CaPSS and pipette solution III). ICC were held at –80 mV. The HC and frequency of STICs were –30.4 ± 6.5 pA and 35.4 ± 3.6 cycles min−1, respectively (n= 5). Pre-treatment with atropine (10 μm) or 4-DAMP (100 nm) did not affect HC and STICs, but these compounds blocked the effects of CCh on HC and STICs (Fig. 9C and D; all greater than P > 0.08).

Currents activated by CCh in on-cell and excised patches

On-cell and excised patch (inside-out configuration) recordings were performed to determine whether the conductance activated by CCh is due to the Ca2+-activated Cl channels previously described in these cells. In the on-cell configuration, CCh (10 μm) activated inward current at a holding potential of –80 mV. Tonic inward currents were observed in 20 of 25 patches (Fig. 10A), and repetitive transient currents (i.e. STIC-like) were observed in the remaining five patches (Fig. 10C). In each case, many channels opened simultaneously, making it impossible to resolve single channel events. The average tonic inward current activated by CCh in on-cell recordings was –72 ± 12 pA (n= 20). After activating responses to CCh under on-cell configuration, the patches were excised, creating inside-out patches in a symmetrical Cl gradient (see Methods). Changing [Ca2+] from 100 nm to 1 nm at the intracellular surface of the patch caused a dramatic decrease in channel openings (Fig. 10B and D.), demonstrating the Ca2+ sensitivity of the channels in the patches that were activated by CCh. Channels in patches with tonic and phasic channel openings were both sensitive to changes in [Ca2+].

Figure 10.

Effects of CCh on Ca2+-activated Cl currents in ICC
A and C, CCh activated inward currents at a holding potential of –60 mV in on-cell configuration (O-C). External and internal solutions were 140 mm KCl and 140 mm NMDG-Cl (see Methods), respectively. CCh (10 μm) induced a tonic increase in current in the example shown in A and oscillatory inward currents in the example in C (see text for details). B and D, after observing currents activated in on-cell configuration, patches were excised (inside-out configuration; I-O), revealing significant Ca2+-sensitive currents when the bath solution (cytoplasmic surface) was switched to one containing 1 nm Ca2+. With inside-out recordings, external and internal solutions were symmetrical NMDG-Cl (140/140 mm) and holding potential was –60 mV.

Two populations of ICC

ICC isolated from the small intestines of wild-type mice consist of two morphologically and anatomically distinct populations of cells (e.g. ICC in the region of the myenteric plexus (ICC-MY) and ICC in the region of the deep muscular plexus (ICC-DMP). Our electrophysiological data generally fell into two groups of behaviour: (i) cells with STICs but no voltage-activated slow wave currents (Zhu et al. 2009); and (ii) cells with STICs and depolarization-activated slow wave currents. To more carefully identify the nature of the cells in the present study, we also characterized currents in cells obtained from a novel cross of WV/+ mice with KitcopGFP/+ (i.e. WcopGFP/+) mice to obtain WcopGFP/WV mice in which ICC expressed copGFP. As in W/WV mice (Ward et al. 1994), only ICC-DMP developed in significant numbers in WcopGFP/WV mice, so cells with copGFP fluorescence dispersed from small intestinal muscles were assumed to be ICC-DMP. ICC-DMP (n= 5) displayed STIC activity (Fig. 11A), but neither step depolarization nor ramp depolarization evoked slow wave currents in these cells (Fig. 11B and D). Thus, cells with STICs but no voltage-activated slow wave currents are likely to be ICC-DMP in the population of cells isolated from KitcopGFP/+ mice, which were used for the bulk of the experiments in this study.

Figure 11.

ICC STICs (ICC-DMP) of WcopGFP/WV murine small intestine
A, ICC-DMP of WcopGFP/WV murine small intestine displayed STIC activity at a holding potential of –80 mV (external solution was CaPSS and pipette solution II, Table 1). Dotted line denotes 0 pA. B and C, voltage-activated ‘slow wave currents’ could not be evoked in these cells either by ramp potentials (–80 to +80 mV; B) or step depolarizations to –35 mV (C).

Effects of carbachol on slow wave currents in ICC

We characterized IClCa in ICC previously and showed large inward currents (slow wave currents due to IClCa) activated by depolarization. We also studied the effects of CCh on depolarization-activated IClCa. Cells were held at –80 mV and stepped repeatedly to –35 mV (see inset Fig. 12) with CaPSS as the external solution and solution II as the internal solution. Inward currents, averaging –1378 ± 148 pA, were evoked by stepping to –35 mV. CCh (100 nm) did not affect the peak amplitude of the evoked current significantly (i.e. −1229 ± 85 pA in control and −1247 ± 77 pA in the presence of CCh, n= 5; P= 0.13, see arrows in Fig. 12A). CCh (1 μm) also did not affect the peak amplitude of the evoked current significantly (i.e. −1260 ± 56 pA in control and −1268 ± 56 pA in the presence of CCh, n= 5; P= 0.65, see arrows in Fig. 12C), but greatly decreased the rates of current relaxations (from 0.5 ± 0.1 s to 0.86 ± 0.1 s in 100 nm CCh, calculated as the time to 50% recovery; n= 5, P < 0.01, Fig. 12B). A higher concentration of CCh (1 μm) further decreased the rate of current relaxation to 1.1 ± 0.2 s (P < 0.01 compared to control, n= 5, Fig. 12D). The effects of CCh on current relaxations were blocked by 4-DAMP (n= 5, P > 0.59; Fig. 12E and F).

Figure 12.

Effects of CCh on ‘slow wave currents’ in ICC
A and C, large inward currents (i.e. ‘slow wave current’ due to IClCa; Zhu et al. 2009) were evoked by step depolarization from –80 to –35 mV in a population of KitcopGFP/+ cells. CCh (100 nm, A) or (1 μm, C) did not significantly affect the peak amplitude of the evoked current during depolarization. However, CCh decreased the rate of current relaxation. B and D, summarized data showing half-time of current relaxation in control and CCh (100 nm and 1 μm). E, 4-DAMP (100 nm) pre-treatment abolished effects of CCh on slow wave current relaxation. F, summarized data showing half-time of current relaxation in control, 4-DAMP and 4-DAMP+CCh. **P < 0.01.

Coupling of muscarinic receptor stimulation to activation of mTmem16a

Tmem16a is highly expressed in ICC and is a molecular candidate for IClCa in these cells. We expressed mTmem16a in HEK 293 cells to investigate the CCh stimulation of mTmem16a in a model system. Firstly, we examined the Ca2+ sensitivity of mTmem16a channels using solution IV (see Table 1) as both the pipette solution and the bath solution. Experiments on HEK 293 cells were performed at room temperature. Cells dialysed with [Ca2+]i at 10−8m (n= 3), 10−7m (n= 6), 5 × 10−7m (n= 9) or 10−6m (n= 6) were held at –80 mV and stepped to +70 mV. Increasing [Ca2+]i increased current density (Fig. 13A).

Figure 13.

Effects of CCh on mTmem16a currents in HEK 293 cells
mTmem16a was expressed in HEK 293 cells. A shows currents in cells dialysed with different concentrations of [Ca2+]i (i.e. < 10−8m, a; 10−7m, b; 5 × 10−7m, c; or 10−6m, d). Arrow denotes peak current used for summarized data in graph at right of raw data traces, plotting averaged currents at +70 mV during steps from a holding potential of –80 mV (as in left panel). B, as 10−7m[Ca2+]i yielded maximal currents, the effects of CCh were tested under these conditions. CCh (up to 10 μm) had no effect on mTmem16a currents expressed in HEK 293 cells. C, CCh (10 μm) enhanced Tmem16a currents in cells in which M3 muscarinic receptors (M3R) were co-expressed. Right panel in C shows currents a and b before and in the presence of CCh at an expanded time-scale The dotted line indicates the zero current level. D, summarized data showing effect of CCh as a function of time in cells with mTmem16a expressed (open circles) or with M3R and mTmem16a co-expressed (filled circles).

The effects of CCh on mTmem16a currents were examined using symmetrical NMDGCl/NMDGCl (solution IV, Table 1) with internal free [Ca2+]i at 100 nm. CCh (10 μm) had no direct effect when only mTmem16a was expressed in HEK 293 cells (n= 5, Fig. 13B). Experiments in native cells, however, suggested that regulation of Ca2+-activated Cl currents by CCh was coupled through binding to M3 receptors (Fig. 9). Therefore, we co-expressed M3 receptors with mTmem16a. CCh activated mTmem16a currents in these cells (Fig. 13C). Changes in peak currents at +70 mV (activation) and –80 mV (deactivation, see arrows in Fig. 13C, right panel) in cells with mTmem16a only and co-expression of mTmem16a and M3 receptors are summarized in Fig. 13D (n= 6). In non-transfected cells, CCh failed to evoke current responses at potentials between −80 mV and +70 mV (n= 5, not shown).


Acetylcholine is the predominant excitatory neurotransmitter in the GI tract and, thus, understanding its actions on post-junctional cells is an important step in understanding neural control of GI motility. The majority of studies in the past have focused on cell signalling mechanisms in smooth muscle cells, based on the assumption that these cells mediate transduction of cholinergic motor stimuli. This approach has generated important information, but electron microscopy has shown that at least three types of cell (smooth muscle cells, ICC and PDGFRα+ cells) commonly lie in close proximity to enteric nerve terminals in the tunica muscularis of the GI tract. In the case of ICC, there are very close, synaptic-like contacts with nerve varicosities (Daniel & Posey-Daniel, 1984; Rumessen et al. 1992; Horiguchi & Komuro, 1998; Wang et al. 1999; Mitsui & Komuro, 2002; Horiguchi et al. 2003); however, close contacts between varicosities and smooth muscle cells have also been noted in muscles of some species (Daniel & Posey-Daniel, 1984; Mitsui & Komuro, 2002). In addition, ICC and PDGFRα+ cells form gap junctions with smooth muscle cells (Torihashi et al. 1993; Horiguchi & Komuro, 1998, 2000). These morphological observations suggest that all three cell types could be targets for motor neurotransmitters, and evidence has been presented that both ICC and PDGFRα+ cells mediate portions of responses to motor neurotransmitters (Burns et al. 1996; Ward et al. 2000; Kurahashi et al. 2011). A distributed post-junctional response has been challenged (Goyal & Chaudhury, 2010), however, suggesting that additional experiments are needed to clarify the cells, receptors, transduction mechanisms and effectors that are integral to motor neurotransmission.

In the present study we investigated the ionic conductances in ICC that were activated by cholinergic stimulation. Cholinergic responses in intact muscle were mainly characterized by membrane depolarization and elevation in the maximal level of depolarization during slow waves. Exogenous CCh and cholinergic neurotransmission elicited similar responses (see Figs 1 and 3). Many previous studies have shown that cholinergic (muscarinic) stimuli increase inward current carried by G-protein-regulated, non-selective cation conductances consisting of TRPC4 and TRPC6 channels in mammalian GI smooth muscle cells (Benham et al. 1985; Inoue & Isenberg, 1990; Unno et al. 2006; Tsvilovskyy et al. 2009). Our data show that isolated ICC are also responsive to cholinergic stimulation. Exposure of ICC to CCh resulted in enhanced frequency and amplitude of spontaneous transient inward currents (STICs) and development of sustained inward current. Equivalent voltage responses (STDs) were elicited by CCh under current-clamp conditions in ICC, and STDs were blocked by NFA and NPPB. Previously, ICC were shown to express ANO1 (encoded by Tmem16a), specifically in the tunica muscularis (Gomez-Pinilla et al. 2009; Hwang et al. 2009). In the current study we found that CCh activated Cl currents in ICC and excised patch experiments showed that this conductance is Ca2+ sensitive. Thus, part of the cholinergic response in small intestinal muscles seems to be due to activation of Ca2+-activated Cl channels in ICC. Responses of intact muscles to cholinergic nerve stimulation were also reduced by NFA, but responses to exogenous CCh were not affected. These data suggest that post-junctional responses to excitatory motor neuro-transmission were dependent, in part, upon transduction of muscarinic stimuli by ICC. A caveat is that previous studies have shown that NFA has non-specific blocking effects on currents carried by TRPC4 channels (Walker et al. 2002); thus, a simple interpretation of results with NFA in intact muscles is problematic.

A recent series of experiments using mice with genetic deactivation of Trpc4, Trpc6 or both showed that the channels encoded by these genes are required for cholinergic responses in smooth muscle cells (Tsvilovskyy et al. 2009). In mice with both genes deactivated, responses of intact ileal longitudinal muscles to electrical field stimulation were reduced, suggesting that knock-out of these channels reduces post-junctional cholinergic responses. These findings, while different from the results of the current study, do not contradict our interpretation of our results. Different cells appear to be targets of innervation in circular and longitudinal muscle layers. For example, the longitudinal muscle layer in the small intestine of the mouse lacks intramuscular ICC (ICC-IM; Torihashi et al. 1995), and there are few muscle motoneurons innervating the longitudinal muscle. In other small laboratory animals with thin longitudinal muscle layers, it appears that smooth muscle cells are directly innervated at or near the myenteric surface by neurons of the tertiary plexus (Richardson, 1958; Llewellyn-Smith et al. 1993). Therefore, motor innervation of the longitudinal muscle layer of the small intestine does not appear to involve ICC in small laboratory animals. In circular muscle, however, processes of motoneurons are densely targeted toward the deep muscular plexus (DMP), near the inner aspect of the circular muscle layer (Sang & Young, 1998), and this is also the site where ICC-IM (termed ICC-DMP in the small intestine) are concentrated. This is also the site of very close contacts between varicose nerve terminals and ICC-DMP (Rumessen et al. 1992; Torihashi et al. 1993; Horiguchi & Komuro, 1998; Wang et al. 1999). Based on our current results and a previous study involving development of motor innervation in the small intestine (Ward et al. 2006), ICC-DMP may mediate a significant portion of cholinergic responses in the circular muscle layer.

ICC are electrically coupled to smooth muscle cells (Torihashi et al. 1993; Horiguchi & Komuro, 1998); thus, conductance changes in ICC and resulting currents influence the behaviour of the syncytium of cells. ICC from the murine small intestine were responsive to cholinergic stimulation via activation of muscarinic receptors. STICs and holding currents were enhanced by CCh under voltage clamp, and the reversal potentials of the conductance(s) and single channel conductance activated by CCh suggest that the inward currents were due to activation of Cl channels. STICs, due to periodic activation of Cl channels, would tend to summate in intact small intestinal muscles and produce a net depolarizing influence on the ICC–smooth muscle cell network. Thus, both the increase in STIC frequency and the net increase in inward holding current evoked in ICC by cholinergic stimulation might result in a depolarizing influence in intact muscles. This may be an important means of transduction of excitatory neural inputs in the small intestine, and this concept is consistent with previous morphological and functional data (Ward et al. 2000, 2006; Wang et al. 2003) and responses of intact muscles to EFS shown in the present study.

We also found that CCh enhanced the duration of voltage-activated slow wave currents in ICC. From studies of cells from WcopGFP/WV mice, we deduced that cells with voltage-activated slow wave currents are ICC-MY, and this is consistent with the pacemaker function of this sub-population of ICC in generating slow waves. CCh caused no increase in the amplitude of the slow wave currents evoked by depolarization in ICC-MY, but relaxation of slow wave currents was slowed. This response might translate into longer duration slow waves in intact muscles stimulated by cholinergic agonists, and we noted that slow wave duration increased in muscles stimulated by exogenous CCh.

In summary, CCh elicits inward currents in at least two of the cell types in close apposition to varicose nerve terminals in GI muscles. In smooth muscle cells, CCh activates non-selective cation current (as shown in many previous studies), but here we have shown that CCh enhances Cl currents, most probably due to the expression of Tmem16a. Immunohistochemical studies have demonstrated that Tmem16a expression is restricted to ICC in the GI tract (Hwang et al. 2009). Niflumic acid blocked electrical responses to CCh in ICC and in intact muscles in response to stimulation of cholinergic nerves. Niflumic acid had little effect on responses to exogenous CCh. Thus, our data are consistent with the concept that ICC, via activation of channels encoded by Tmem16a, are responsible for a significant portion of the post-junctional electrical response to excitatory motoneurons in the murine small intestine.


Author contributions

M.H.Z. performed whole-cell patch-clamp experiments, and collected and analysed data. I.K.S. performed intracellular recordings on intact muscles, and collected and analysed data. H.Z. performed experiments in which single channel currents were measured, and collected and analysed data. T.S.S. performed electrophysiological experiments on HEK 293 cells, and collected and analysed data. F.C.B, and K.O’D. generated molecular constructs and expressed mTmem16a and muscarinic receptors in HEK 293 cells. S.D.K. and K.M.S. conceived of project, designed experiments, helped with analysis and interpretation of data, and wrote manuscript. All authors approved the final manuscript.


This project was supported by the National Institute of Diabetes and Digestive and Kidney Diseases throught grants P01 DK41315 and R37 DK40569. Expression of mTmem16a was performed in the Molecular Core of P01 DK41315. The authors are grateful to Drs Heather Young and John Furness for comments about the innervation of murine small intestinal muscle.