Hyperpolarization‐activated cyclic nucleotide‐gated channels working as pacemaker channels in colonic interstitial cells of Cajal

Abstract Hyperpolarization‐activated cyclic nucleotide‐gated (HCN) channels function as pacemaker channels in spontaneously active cells. We studied the existence of HCN channels and their functional roles in the interstitial cells of Cajal (ICC) from the mouse colon using electrophysiological, immunohistochemical and molecular techniques. HCN1 and HCN3 channels were detected in anoctamin‐1 (Ca2+‐activated Cl− channel; ANO1)‐positive cells within the muscular and myenteric layers in colonic tissues. The mRNA transcripts of HCN1 and HCN3 channels were expressed in ANO1‐positive ICC. In the deletion of HCN1 and HCN3 channels in colonic ICC, the pacemaking potential frequency was reduced. Basal cellular adenylate cyclase activity was decreased by adenylate cyclase inhibitor in colonic ICC, whereas cAMP‐specific phosphodiesterase inhibitors increased it. 8‐Bromo‐cyclic AMP and rolipram increased spontaneous intracellular Ca2+ oscillations. In addition, Ca2+‐dependent adenylate cyclase 1 (AC1) mRNA was detected in colonic ICC. Sulprostone, a PGE2‐EP3 agonist, increased the pacemaking potential frequency, maximum rate of rise of resting membrane in pacemaker potentials and basal cellular adenylate cyclase activity in colonic ICC. These results indicate that HCN channels exist in colonic ICC and participate in generating pacemaking potentials. Thus, HCN channels may be therapeutic targets in disturbed colonic motility disorders.

spontaneous electrical activity, known as a pacemaker activity, of ICC is directly transmitted to smooth muscles and causes slow waves that determine smooth muscle contractility. 3 In ICC, a variety of ion channels have been reported, including transient receptor potential (TRP) channels, 4 Ca 2+ -activated Cl − channels, 5 voltage-dependent Ca 2+ channels, 6,7 voltage-dependent K + channels, 8 ERG channels, 9 Na + channels 10 and ATP-sensitive K + channels. 11,12 Among these channels, TRP channels and Ca 2+ -activated Cl − channels (Anoctamin-1; ANO1) have been suggested as pacemaker channels in ICC. 13,14 Particularly, ANO1 is a strong candidate as a pacemaker channel because of the expression of ANO1, failing of slow waves in ANO1 knockout mouse, blocking of slow waves by Ca 2+ -activated Cl − channel blockers in intact mouse and human intestine and single-channel recording of Ca 2+activated Cl − channels in ICC. 14,15 Hyperpolarization-activated cyclic nucleotide (HCN) channels are pacemaking channels that cause diastolic depolarization to initiate rhythmic activity in cardiac and neuronal cells. [16][17][18] Previously, we reported that HCN channels might be involved in the generation of pacemaker activity in colonic ICC but not intestinal ICC of mice. 19 HCN channels are gated by intracellular cAMP directly. 20 In colonic ICC, adenylate cyclase inhibitors reduced pacemaking potential frequency while cAMP-specific phosphodiesterase inhibitors enhanced pacemaking potential frequency. In addition, HCN channel blockers inhibited the pacemaker activity, and mRNA transcripts of HCN channels were expressed in colonic ICC. 19 Sulprostone, a PGE 2 -EP 3 agonist, also enhanced the pacemaking potential frequency of colonic ICC, and the enhanced effects were blocked by HCN channel blockers. 21 Thus, we suggest that HCN channels are candidate pacemaker channels modulated by endogenous substances in colonic ICC. However, our previous results were mainly obtained using pharmacological studies. Thus, to confirm the existence and functional role of HCN channels in colonic ICC, we further performed electrophysiology, immunohistochemistry and molecular studies.

| Preparation of cells
Experimental and animal care protocols used in the study were all in accordance with the guiding principles approved by the Ethics Committee of Chosun University and the National Institutes of Health Guide, South Korea, for the Care and Use of Laboratory Animals. Mice had free access to water and were fed a standard mouse diet until the day of the experiment. Balb/C mice (5-8 days old) of either sex were anaesthetized with ether and euthanized by cervical dislocation. The small intestines from 1 cm below the pyloric ring to the cecum were removed and colon from below the cecum to the rectum was removed, and the middle portion of the colon was used. The small intestine and colon were opened along the mesenteric border. The luminal contents were washed away with Krebs-Ringer bicarbonate solution. Tissues were pinned to the base of a Sylgard dish, and the mucosa was removed by dissection. Small strips of small intestinal and colonic muscle were equilibrated in Ca 2+ -free Hank's solution containing: KCl 5.36 mM, NaCl 125 mM, NaOH 0.34 mM, Na 2 HCO 3 0.44 mM, glucose 10 mM, sucrose 2.9 mM and HEPES 11 mM for 30 min. The cells were dispersed with an enzyme solution containing 1.3 mg/ml collagenase (Worthington Biochemical Co), 2 mg/ml bovine serum albumin (Sigma), 2 mg/ml trypsin inhibitor (Sigma) and 0.27 mg/ml ATP. Cells were plated onto sterile glass coverslips coated with murine collagen (2.5 μg/ml Falcon/BD) in 35-mm culture dishes. The cells were then cultured in smooth muscle growth medium (SMGM; Clonetics Corp.) supplemented with 2% antibiotics/antimycotics (Gibco) and 5 ng/ml urine stem cell factor (SCF, Sigma) at 37°C and 5% CO 2 .

| Patch-clamp recording
The current clamp mode of patch clamp was applied to record pacemaking potentials in colonic ICC-like cells that showed a network-like structure in cultures after 2-3 days. The cell culture dishes were mounted on a model TE-2000s inverted microscope (Nikon). The bath solution used was 3 ml/min, and all experiments were performed at 30°C. Pacemaker potentials were amplified using Axopatch 200 B (Axon Instruments). Data were filtered at 5 kHz and displayed on a computer monitor. Results were later analysed using pClamp and GraphPad Prism version 5.0 (GraphPad Software Inc.).

| Collection of ICC and reverse transcription chain reaction (RT-PCR)
Total cultured cells or picked cells (below 10 cells) with typical ICC morphology (i.e., triangular or spindle-shaped with several branches) were sucked into a recording pipette under negative pressure. After this, the samples were expelled into phosphate-buffered saline by applying positive pressure. After picking, the cells were centrifuged at 9800 g at 4°C for 8 min before lysis. Total RNA was isolated using

| Immunohistochemistry
Adult C57BL/6 mice were transcardially perfused using 100 mM PBS followed by 4% paraformaldehyde in PBS. The colons were removed and post-fixed in 4% paraformaldehyde in PBS for one day and then dehydrated in 30% sucrose solution for 48 h. Dehydrated colons were embedded with optimal cutting temperature compound (OCT compound) prior to frozen sectioning and then were sectioned using cryo-microtome (Leica) at 12 μM thickness. Colon tissue sections were rinsed in phosphate-buffered saline (PBS).
Tissue sections were also permeabilized with 0.1% Triton X-100 diluted in PBS containing 5% normal horse serum for 2 h at room temperature. Afterwards, the sections were incubated for 24 h at 4°C in the primary antibody (1:50 dilution for each HCN subunit; Alomone Labs, and ANO1; Santa Cruz Biotechnology) diluted in PBS containing 5% normal horse serum. After rinsing them in PBS three times at 10 min each, the sections were incubated with an Alexa ing 5% normal horse serum for 2 h at room temperature. Images were obtained using a Zeiss LSM5 confocal microscope using 5× and 20× objective lenses. For each antigen, all colon sections were processed identically, and images were obtained using the same microscope parameters (pinhole;159). The relative intensity of immunopositive signals for HCN1, HCN2, HCN3 and HCN4 subunits in the colon was measured using ImageJ software. 22 HCN immunolabelling of ANO1positive cells was also assessed using confocal microscope imaging.

| Measurement of adenylate cyclase activity
Adenylate cyclase activity was measured in cultured ICC. ICC were rinsed twice with ice-cold PBS and collected in PBS. ICC were homogenized with five strokes of a homogenizer. The homogenate was then centrifuged at 1000 g for 5 min at 4°C. The supernatant was then transferred into a centrifuge tube and centrifuged at 5000 g for 10 min. The pellet was suspended in buffer (30 mM Na-HEPES, 5 mM MgCl 2 and 2 mM DTT, pH 7.5) to attain a total protein concentration of approximately 1 mg/ml before being added into tubes containing the drug and adenylate cyclase assay buffer (30 mM Na-HEPES, 100 mM NaCl, 1 mM EGTA, 10 mM MgCl 2 , 1 mM isobutylmethylxanthine, 1 mM ATP, 10 mM phosphocreatine, 5 μM GTP, 60 U/ml creatine phosphokinase and 0.1% bovine serum albumin, pH 7.5). Adenylate cyclase activity assays were performed using a mouse adenylate cyclase type 6 ELISA kit (MyBiosource Company).

| Transient transfection with small interfering RNAs (siRNAs)
In adherence to the manufacturer's instructions for the use of the transfection reagent, ICCs were plated with 60%-70% confluency and transfected into groups. Before transfection, ICCs were starved Gene Sequences Accession No

Size (bp)
Ano-1 (F) AGG CCA AGT ACA GCA TGG GTA TCA NM_178642 213  8-bromo-cAMP and sulprostone. Solvent for All drugs were DMSO and all drugs were purchased from Sigma-Aldrich.

| Statistical analysis
Data are expressed as mean ± SEM. Student's t-test for paired data was used to assess the significance of any differences observed.
Statistical significance was set at p < 0.05. The n-values reported in the text refer to the number of cells used in the patch-clamp experiments.

| Expression of HCN channels in interstitial cells of Cajal
The expression of HCN channels was investigated using immunohistochemistry studies. Double staining with anti-ANO1 and HCN antibodies showed HCN1 immunoreactivity in ANO1-positive ICC of muscular and myenteric layers from mouse colon ( Figure 1A, upper panels; Figure S1A, white arrowheads). Lower panels in Figure 1A represent the HCN1 signals in the ANO1-immunopositive cells of myenteric layers ( Figure 1A, lower panels). However, HCN3 immunoreactivity was only seen in the myenteric region in ANO1-positive ICC ( Figure 1C, upper and lower panels; Figure S1C, white arrowhead). Only a low level of HCN2 immunoreactivity was expressed in the longitudinal layer ( Figure 1B; Figure S1B), and HCN4 was expressed in neither the myenteric nor muscle layers ( Figure 1D; Figure S1D). The summarized distribution of the HCN channels is shown in Figure 1E. To support this, RT-PCR was performed to detect the expression of HCN channels at the cellular level.
RT-PCR analysis revealed the mRNA transcripts for all four HCN channel subtypes in whole mounted, cultured colonic cells. However, the mRNA transcripts for only HCN1 and HCN3 channels were detected in ANO1-positive ICC by picking ( Figure 1F). These results suggest that HCN channels exist in colonic ICC.

| HCN1 and HCN3 channels mediate the generation of pacemaking activity in cultured colonic ICC
To test whether HCN1 and HCN3 channels function as pacemaker channels, we transfected siRNA vectors of HCN1 and HCN3 channels into cultured ICC after 1 day in culture. In the control group, ICC that transfected Lipofectamine only showed 11.9 ± 2.46 cy-

| Basal cAMP involved in generating pacemaker potentials in colonic ICC
It is well known that intracellular cAMP directly activates HCN channels. Thus, we examined intracellular adenylate cyclase (AC) activity using an AC inhibitor and a cAMP-specific phosphodiesterase inhibitor in ANO1-positive ICC. Basal AC activity of colonic ICC in the control group was higher than that in the small intestinal ICC. Furthermore, treatment with rolipram (10 μM), a cAMP-specific phosphodiesterase inhibitor, showed increased AC activity, whereas SQ22536 (10 μM), an AC inhibitor, decreased AC activity in colonic ICC ( Figure 4A, B, n = 4-7). However, rolipram or SQ22536 did not

| 5 cAMP increases intracellular Ca 2+ [Ca 2+ ] i oscillations in colonic ICC
The activation of pacemaker channels is closely coupled with spon-   Figure 6D.

| DISCUSS ION
In the present study, we confirmed the presence of HCN channels that act as pacemaker channels and contribute to the generation of spontaneous pacemaking activity in colonic ICC. HCN channels are comprised of four isoforms (HCN1-HCN4) that show different patterns of localization in the CNS and heart. 23,24 Therefore, it seems that HCN isoforms play different physiological roles. In the GI tract, HCN channels are distributed in enteric neurons and express four channels in the myenteric plexus of the mouse colon. 25 Yang et al. 26 Reported that HCN2 channels are distributed only in myenteric neurons of the mouse GI tract but not necessarily in ICC. HCN2-positive cells were labelled as cholinergic neurons via immunohistochemical studies. Thus, the authors suggest that HCN2 channels may be related to the acetylcholine release of cholinergic neurons. 26 Recently, HCN2, HCN3 and HCN4 channels were reported in enteric neurons and ICC of normal human colons. 27 They also found that the expression of HCN3 channels is markedly reduced not only in the aganglionic bowel but also in the ganglionic bowel of patients with Hirschsprung's disease (HSCR). In addition, they found decreased expression of ICC both in the aganglionic bowel and ganglionic bowel in patients. 28 These findings, therefore, suggest that decreased expression of HCN3 channels and ICC may contribute to motility dysfunction in HSCR. We also reported that RT-PCR using mouse HCN primers revealed mRNA for HCN1 and HCN3 channels in cultured ANO1-positive ICC from mouse colon, suggesting that these channels may participate in generating pacemaker activity. 19 However, our previous experiments were mainly pharmacological in nature. Thus, to confirm the presence of HCN channels and the to confirm that a change in basal intracellular cAMP level is coupled by HCN channel activation, we measured intracellular AC activity in colonic ICC. Our previous data showed that rolipram, a cAMPspecific phosphodiesterase, increased pacemaking potential frequency, whereas SQ22536, an AC inhibitor, decreased pacemaking potential frequency in colonic ICCs but not in small intestinal ICC.
The present study showed increased AC activity due to rolipram and decreased AC activity due to SQ22536 in colonic ICC. As the generation of pacemaker potentials was blocked by HCN channel blockers in colonic ICC, our data suggest that periodic activation of pacemaker potentials is linked to intracellular cAMP concentration oscillations. In the generation of pacemaker potentials in the sinoatrial node, HCN channels are activated at resting membrane potentials, decreasing maximal diastolic potential; therefore, they are involved in periodic diastolic depolarization. 31 Maximal diastolic potential is increased and the slope of diastolic depolarization is decreased in HCN1 knockout mice, which thereby reduces the heartbeat.
The slope of diastolic depolarization in the sinoatrial node was increased by β-adrenoceptor activation and decreased by muscarinic activation. 32  with ANO1. Additionally, tonic activation of HCN channels at resting membrane potential induces inward currents and leads to depolarization of the membrane. This is followed by the activation of T-type Ca 2+ channels, initiating pacemaker potentials in colonic ICC.
In conclusion, the results presented here confirm our previous hypothesis that HCN channels are present in colonic ICC and are involved in generation of spontaneous pacemaker activity. Therefore, therapeutic targeting of HCN channels may be effective in decreasing colonic motility disorders.

CO N FLI C T O F I NTE R E S T
The authors confirm that this article and its authors have no conflicts of interest to declare.

Data openly available in a public repository that issues datasets with
DOIs.