Inhibition of TRPC4 channel activity in colonic myocytes by tricyclic antidepressants disrupts colonic motility causing constipation

Abstract Tricyclic antidepressants (TCAs) have been used to treat depression and were recently approved for treating irritable bowel syndrome (IBS) patients with severe or refractory IBS symptoms. However, the molecular mechanism of TCA action in the gastrointestinal (GI) tract remains poorly understood. Transient receptor potential channel canonical type 4 (TRPC4), which is a Ca2+‐permeable nonselective cation channel, is a critical regulator of GI excitability. Herein, we investigated whether TCA modulates TRPC4 channel activity and which mechanism in colonic myocytes consequently causes constipation. To prove the clinical benefit in patients with diarrhoea caused by TCA treatment, we performed mechanical tension recording of repetitive motor pattern (RMP) in segment, electric field stimulation (EFS)‐induced and spontaneous contractions in isolated muscle strips. From these recordings, we observed that all TCA compounds significantly inhibited contractions of colonic motility in human. To determine the contribution of TRPC4 to colonic motility, we measured the electrical activity of heterologous or endogenous TRPC4 by TCAs using the patch clamp technique in HEK293 cells and murine colonic myocytes. In TRPC4‐overexpressed HEK cells, we observed TCA‐evoked direct inhibition of TRPC4. Compared with TRPC4‐knockout mice, we identified that muscarinic cationic current (mI cat) was suppressed through TRPC4 inhibition by TCA in isolated murine colonic myocytes. Collectively, we suggest that TCA action is responsible for the inhibition of TRPC4 channels in colonic myocytes, ultimately causing constipation. These findings provide clinical insights into abnormal intestinal motility and medical interventions aimed at IBS therapy.


| INTRODUC TI ON
The function of the gastrointestinal (GI) tract is to ingest, digest, and absorb nutrients and eliminate waste. GI motor patterns in mammalian are largely composed of colonic motor complexes (CMCs), a mass movement, and spontaneous phasic contractions called ripples. 1,2 In order to generate these repetitive propagating sequences in GI motility, motor complexes in smooth muscle layer require harmonized coordination between enteric neurons, interstitial cells of Cajal (ICCs) and smooth muscle cells (SMCs). 3,4 The enteric nervous system (ENS) includes intrinsic neural plexuses and autonomic extrinsic neural pathways, which are of fundamental importance for generating major motor patterns and regulating the amplitudes of contractions. 5,6 The polarized enteric circuits conducted to intramuscular ICC (ICC-IM) and SMCs generate CMCs. 3,7 And spontaneous electrical activity generated by myenteric ICC (ICC-MY) is conducted to SMCs which consequently leads to slow wave and spontaneous muscle contractions. 3,8 The mechanical role of spontaneous contractions is controversial. Ehrlein et al. 9 suggested that they were mainly involved in mixing rather than propulsion. On the contrary, it has been argued that spontaneous contractions are not effective contractions for mixing and propulsion by themselves, but they cause larger contractions when they occur concurrently with other contractions. 2,10 Repetitive motor patterns (RMPs), defined only in human GI motility, are complex propulsive contractions triggered by enteric nerve stimulation or spontaneous. 2 The smooth muscle response to external signals ultimately depends on the excitability of the syncytium consisting of ICCs and SMCs. 11,12 Precise and balanced regulation of parasympathetic signalling is essential for GI motility; however, some medications, particularly antidepressants, can produce anticholinergic effects with constipation. 13,14 Antidepressants, which are the standard drug therapy used in the treatment of depression, typically act to restore the balance of neurotransmitter levels. Commonly prescribed antidepressants, such as tricyclic antidepressants (TCAs), selective serotonin reuptake inhibitors (SSRIs) and serotonin-norepinephrine reuptake inhibitors, have different mechanisms of action and consequent pleiotropic effects.
TCAs, which consist of a 3-ring in their core chemical structure, include amitriptyline (AMI; Elavil), imipramine (IMI; Tofrnail), doxepin (Adapin) and desipramine (DES; Norpramin). Previously, most evidences have indicated that TCA antagonizes histamine H1, adrenoreceptor and muscarinic acetylcholine receptor, directly producing a cluster of symptoms called anticholinergic side effects, which include sedation, drowsiness, postural hypotension, blurred vision, dry mouth and constipation. 15 Although SSRIs are a first-line antidepressant because they have potentially fewer side effects, 16,17 the clinical value of TCAs still exists not only in neuropathic pain 18,19 and Parkinson's disease (PD) 20 but also in irritable bowel syndrome (IBS). 21,22 One of the functional GI disorders, IBS, is accompanied by abdominal pain and abnormal stool form. IBS is classified into predominant stool patterns, such as IBS with diarrhoea (IBS-D), IBS with constipation (IBS-C) and mixed IBS (IBS-M). 23 Treatment is aimed at relieving pain and improving bowel problems, but addressing each individual's specific symptoms depending on the severity of symptoms is more important. The first-line medical therapies for IBS are those that use laxative, antidiarrhoeal, and antispasmodic agents, but most randomized controlled trials could hamper personalized treatment based on the predominant symptoms. 24 When the report that antidepressants had potential as a treatment for IBS was first presented three decades ago, SSRIs and TCAs were prescribed for the purpose of stabilizing the central nervous system (CNS). Since then, clinical evaluations of TCA have revealed not only central neuromodulation but also reduced GI motility. In addition, Siproudhis and colleagues demonstrated that AMI could be used to reduce the pressure of defecation by relaxing the anal sphincter muscle. 25 Based on these findings, TCA prescription for patients with IBS is considered suitable, especially for patients with predominant pain and diarrhoea (IBS-D).
Although, to date, there are several preclinical experimental studies that have investigated the efficacy of TCAs, the identification of on-and off-targets is needed to overcome some of the potential limitations of TCAs. Duncan et al. 26 found that doxepin, similar to other TCAs, caused long QT interval prolongation by inhibiting hERK channels of ventricular myocytes from rabbits. Dennis et al. 27 identified that, in addition to blockade of the hERG current, trafficking inhibition and degradation of hERG are responsible for the cardiotoxicity of TCAs. Hamaguchi et al. 28,29 reported that the inhibitory effect of IMI on the TRPM-like channels contributes to vascular smooth muscle homeostasis by reducing Mg 2+ influx in the porcine carotid artery. On the basis of a wide area of pain and symptoms caused by TCAs, we focus on the potential role of transient receptor potential canonical (TRPC) proteins, ubiquitously expressed in the nervous, digestive and reproductive systems. 30,31 In intestinal smooth muscle cells isolated from knockout mice, TRPC4 and TRPC6 channels, gated by muscarinic receptors, are already well known to be responsible for muscarinic cationic current (mI cat ). 32 Based on pharmacological intervention that inhibits large intestinal motility to alleviate abdominal symptoms accompanied by diarrhoea, 33 we evaluated whether the TRPC4 channel has potential as a clinical candidate.
In this study, we report on the inhibitory effect of the TRPC4 current by TCAs underlying the causality of colonic motility for TCAinduced constipation using human and murine colons. Within the estimated serum concentration range of TCAs, 34

| Isolation of murine colonic myocytes
The animal experiments were approved by the ethics committee of

| Mechanical tension recordings
Whole colonic segments with intact mucosal layers were dissected parallel to the longitudinal muscle using a pair of scissors. 35,36 To mimic natural colonic segment, a flat dissected colon was reformed to tubular shape. Circular muscle tension of each segment (5 cm in length and 2 cm in width) was recorded at three sites (proximal, middle and distal sites, 2 cm apart) via perpendicular traction using sutures placed at each site. Sutured muscle was connected to an isometric force transducer (Biopac Systems) using threaded stainless steel micro serrefines (Fine Science Tools).
Colonic segments were equilibrated for at least 2 h before experiments under a resting force of 1 g. Prewarmed (36.5 ± 0.5°C) and preoxygenated KRB solution was perfused continuously into the tissue chamber. The AUC for 10 min was analysed for RMPs before and after the application of drugs. The mechanical responses were recorded and digitized using Acknowledge software (Biopac Systems). Data were analysed offline using Clampfit (version 10.7.

Molecular devices).
For recording spontaneous contractions, colonic muscle strips (6 mm in length and 2 mm in width) without the mucosal layers were dissected parallel to the circular muscle layer using a knife consisting of double parallel scalpel blades set 1.5 mm apart. 36 The remnant muscle strips were connected to an isometric force transducer (Biopac Systems) and suspended in a

| Surface biotinylation
The cells were lysed to extract the proteins using lysis buffer (1% Triton X-100, 150 mM NaCl, 50 mM HEPES, 2 mM MgCl 2 , 2 mM EDTA, pH 7.4 and protease inhibitor cocktail (Roche)). After incubation at 4°C for 30 min, each sample was centrifuged at 13,000 g at 4°C for 15 min. Supernatants were collected, and protein concentration was measured at 750 nm using DC Protein Assay (BioRad).
A defined quantity of total protein was electrophoresed on an 8% SDS-polyacrylamide gel, and then subsequently transferred onto a nitrocellulose membrane. Each membrane was blocked using 5% BSA prepared in TBST. Each membrane was incubated at 4°C overnight with primary antibody diluted in 5% BSA prepared in TBST buffer. The proteins were probed with GFP (Life technology). β-Tubulin (Sigma) antibodies were used as housekeeping proteins.
Each membrane was incubated with secondary antibody for 1 h 30 min.
For surface biotinylation, washed twice with prechilled PBS cells were incubated in 0.5 mg/ml sulfo-NHS-SS-biotin (Thermo Scientific) in PBS for 30 min on ice. Afterwards, the free biotin was quenched by the addition of 100 mM glycine in PBS. The cells were then processed as described above to make cell extract. 40 μl of 50% aqueous slurry of immobilized avidin beads (Thermo Scientific) was added to 400 μl of cell lysates (0.5-1 mg protein). After incubation with gentle shaking for 1 h at room temperature, beads were washed three times with 0.5% Triton-X-100 in PBS, and proteins were extracted in 5×sample buffer. Collected proteins were then analysed by Western blot. Na + /K + -ATPase (NKA) was used as control of membrane protein detected by alpha 1 Sodium Potassium ATPase antibody (Abcam).

| TCA strongly inhibits the mechanical activity of human colonic smooth muscle
The prescription of TCAs with high anticholinergic activity exhibits a number of anticholinergic signs and symptoms, such as dry mouth, blurred vision, urinary retention, constipation and hallucinations. 38,39 In addition, in TCA overdose, there are likely symptoms that raise suspicion for cardiovascular toxicity, such as arrhythmias and refractory hypotension. 40 Nevertheless, it has been shown to be extraordinarily beneficial to improve global IBS symptoms. 21,41 However, in the action of TCA for treating IBS by inhibiting GI motility, molecular candidates and the exact mechanism have not been clearly established.
To experimentally prove the clinical benefit in patients with diarrhoea by TCA treatment, we isolated normal colonic specimens from patients with colorectal cancer. As in our previous studies using mechanical tension recording, 35 Figure 2B). In pretreatment with 1 μM TTX to block neural stimulation input, spontaneous contractions before and after AMI treatment showed a similar pattern (28.30 ± 10.63%, Figure 2D). As shown in Figure S2C

| TRPC4 channels closely contribute to the regulation of human colonic muscle contractions
We next attempted to identify the molecular candidate of TCA that inhibits colonic motility. Numerous studies have confirmed that TRPC4 channels in intestinal SMCs are gated by muscarinic receptors, 43 and approximately 80% of mI cat are mediated by TRPC4 activity. 32 To determine the functional role of TRPC4 in GI motility using its pharmacological agonist or antagonist, we investigated RMP and EFS-induced contractile activity as observed above. As shown in Figure 3A, 100 nM Englerin A (EA), a potent and selective activator of TRPC4, significantly increased repetitive motor activities. These sustained (tonic) contractions could result from smooth muscle, 44 suggesting that TRPC4 has considerable potential for the depolarization of colonic myocytes. Tsvilovskyy et al. 32 previously suggested that TRPC4 is indirectly activated by acetylcholine involved in neurogenic contraction. To rule out a contribution of TRPC4 to the neurogenic contraction, EFS-induced contraction was compared in the absence or presence of Pico145 (a remarkable inhibitor of TRPC4). Pico145 (100 nM) was slightly suppressed by 64.64 ± 10.71% at the highest (16 Hz) frequencies ( Figure 3B). Additionally, in the circular smooth muscle strip, Pico145 caused a substantial decrease in the AUC of spontaneous contraction rather than an amplitude of nearly half ( Figure S3A). Conversely, EA dramatically increased the amplitude only of spontaneous contractions with tone ( Figure 3C). EA-enhanced spontaneous contractions were not altered by TTX ( Figure 3D). These results indicated that blockade of the TRPC4 channel induces atrophy not only in ENS-mediated contractions but also in smooth muscle activation. Functional role of TRPC4 in neurogenic contraction should not be overlooked, but given that the dominant role of TRPC4 in the reports thus far is considered primarily to activate depolarization of intestinal myocytes, it is, therefore, considered to predominantly act to activate SMCs. These findings and suggestions indicated that TRPC4 is an essential determinant of colonic myocyte contraction causing intestinal motility. Thus, TRPC4 seems to be a reasonable candidate as a molecular target of TCA-induced constipation and IBS treatment with TCA.

| TCA evokes direct extracellular inhibition of TRPC4 channel activity
To investigate the electrical properties of the TRPC4 channel induced by TCA, we conducted patch clamp recordings in TRPC4overexpressing HEK293 cells. As mentioned above about the relevance of TRPC4 to altered electromechanical activity in colonic contraction induced by TCA, we expected that TCA inhibits TRPC4 channel activity. Since the stimulation of the muscarinic acetylcholine receptor elicits mI cat for initiating cholinergic contraction, we measured the TRPC4 current by coexpression with muscarinic acetylcholine receptor type 2 (M 2 R) and type 3 (M 3 R), which are mainly expressed in smooth muscle. 45,46 As the Gα q -PLC pathway is a primary activation of the TRPC4 channel, CCh stimulates M 3 R, 47,48 apparently showing a typical doubly rectifying TRPC4 current by These results indicated that TCA evokes direct extracellular inhibition of the TRPC4 current without changing TRPC4 expression. Therefore, TCA compounds absorbed into the gut have negative potential that is sufficient to broadly block TRPC4 functions in intestinal smooth muscle.

| TCA remarkably suppresses the mI cat formed by TRPC4 in isolated murine colonic myocytes
It is well defined that mI cat , observed in murine myocytes, is prominently elicited by a TRPC4-mediated cationic current. 32,51 To further clarify whether TCA blocks the mI cat of the colonic myocyte response to CCh, we prepared myocytes from murine sigmoid colon tissue by enzymatic isolation following our previous procedure. 52 Under the optimized conditions of TRPC4 recording similar to that of TRPC4-overexpressing HEK cells, the mI cat from a single  Figure 5C).
The following experiment was designed to evaluate whether TCA-induced myocyte inactivation could be improved by modulating TRPC4 activity as a therapeutic approach for constipation. As shown in Figure 6A, potent inhibition of RMPs with reduced amplitude and frequency was rescued by TRPC4 activation with 10 nM EA. The higher concentration of 100 nM EA led to tonic contractions with a cumulative response in partial frequency recovery of the proximal (21.20 ± 5.38%), middle (25.81 ± 10.22%) and distal (32.62 ± 8.14%). Even under TTX pretreatment, the amplitude of spontaneous contractions, which was reduced to 15.28 ± 2.58% in AMI, was restored to 49.03 ± 8.84% by EA ( Figure 6B). In addition,  Figure 6C,D).
These results of Figure 5 indicate that the mI cat suppressed by TCA is ultimately responsible for the inhibition of TRPC4 channels expressed in colonic myocytes. The results of Figure 6 mimicking the therapeutic evaluation of TCA-induced constipation and IBS-D indicate that colonic motility atrophied by TCA was improved by the restoration of TRPC4 activity.

| DISCUSS ION
The GI tract is made up of four layers: the innermost layer is the mucosa, underneath this is the submucosa, followed by the muscularis propria (muscular layer) and finally, the outermost layer-the adventitia (or serosa). The muscular layer is made up of two layers of smooth muscle, the inner, circular layer, and the outer, longitudinal layer. The rhythmicity of segmentation requires a myogenic pacemaker that is evoked by synchronized reconstitution of the enteric motor neurons and ICC-IM. In contrast to myogenic slow waves, whose frequency of depolarization is determined by ICC, mass peristalsis comprises an extensive region of sustained and coordinated contraction of proximal and distal smooth muscle segments that propagates rapidly by neurogenic contractions. RMPs recorded in colonic segment are propulsive contraction with high amplitude and low frequency.
These are neurogenic contractions which are mediated by enteric neuron. Although our observations on TCA-induced disturbed motility were limited to only the human sigmoid colon (Figures 1-3), this TCA action is ultimately responsible for the inhibition of TRPC4 channels expressed in colonic myocytes (Figures 4 and 5). Constipation is one of the most commonly reported adverse symptoms with many medications (anticholinergics, antihypertensives, antidepressants, iron supplements, narcotic analgesics and calcium channel blockers). Constipation is characterized by digested food wastes that absorb too much water to create a dry solid matter called stool or prolonged transit time of stool that moves slowly through the digestive tract due to poor GI motility. 55 In various regulatory aspects of GI tract, it has been demonstrated through an amount of reports that the TRP channels have important functions. 57,58 Several TRP channels are involved in ENS signalling and are known to contribute to visceral sensitivity and hypersensitivity (TRPA1, V1 and V4). [59][60][61] In addition, TRPM7 expressed in ICC is expected to contribute to intestinal pacemaking. 62 In particular, TRPV1 antagonists are currently used as therapeutic agents for heartburn and visceral hypersensitivity, but various TRP channels are likely to be proposed as targets for intestinal diseases. 63 In complete contrast to constipation, diarrhoea is the primary symptom in patients with IBS-D, characterized by sudden urges to have bowel movements along with loose stools, frequent stools, abdominal pain and discomfort gas. Considering this, a therapeutic approach targeting TRPC4 can be more effective in ameliorating diarrhoea, such as IBS-D symptoms. TRPC6 can also be a target in terms of improving colonic motility through inhibition of smooth muscle contraction, but its contribution in depolarization of myocyte is relatively lacking compared with TRPC4, 32 and our results also showed an almost complete decrease in current in TRPC4 KO mice and blocker ( Figure 5).
Therefore, the TRPC4 channel should be considered a reasonable candidate as a molecular target of TCA-induced constipation and IBS treatment with TCA. Although we believe that TRPC4 channel of colonic myocytes has functional potential as an alternative molecular target to treat IBS with TCA, it is worthy of further study using a humanized mouse model of IBS.
Given the psychiatric range used for antidepressant (100-200 mg/day) treatment, 34 the dosage of TCAs used for IBS (25-125 mg/day) 21 is considered to be low. Moreover, the concentrations we applied in the suppression of spontaneous RMP and myocyte activities by gut motility are likely to be higher than the estimated serum concentration (100-300 ng/ml) of TCA. 34 Nevertheless, the pharmacological properties of TCAs can produce unintended biological activities via potential off-target effects. In addition to TRPC4, previous studies have reported that various ion channels are inhibited by TCA, and in particular, ATP-dependent K + channels 52 and Ltype calcium channels 64 are importantly involved in ICC activation and myocyte contraction in the GI tract, so it is difficult to exclude the effect of TCA. As shown in the EFS-induced contractile activity of supplementary Figure 3C,D, even after preinhibition of TRPC4 with Pico145, the amplitude that is partially suppressed by TCA remained. On the contrary, pretreatment with AMI was completely blocked even at higher frequencies. Likewise, we cannot rule out the possibility that it accounts for another target of AMI together with TRPC4 despite the apparent absence of TRPC4 activity by Pico145.
To date, the treatment of diarrhoea with antidepressants has relied on clinical statistics, and although the physiological mechanisms are not clearly understood, these findings conclude that TRPC4 is a critical regulator of the suppression of intestinal motility by TCA.
Taken together, our new target, TRPC4, will provide clinical insights into medical interventions aimed at IBS, as well as expanding the understanding of various adverse effects of TCA.