Cholinergic regulation of epithelial ion transport in the mammalian intestine

Abstract

Acetylcholine (ACh) is critical in controlling epithelial ion transport and hence water movements for gut hydration. Here we review the mechanism of cholinergic control of epithelial ion transport across the mammalian intestine. The cholinergic nervous system affects basal ion flux and can evoke increased active ion transport events. Most studies rely on measuring increases in short-circuit current (I_SC = active ion transport) evoked by adding ACh or cholinomimetics to intestinal tissue mounted in Ussing chambers. Despite subtle species and gut regional differences, most data indicate that, under normal circumstances, the effect of ACh on intestinal ion transport is mainly an increase in Cl^- secretion due to interaction with epithelial M₃ muscarinic ACh receptors (mAChRs) and, to a lesser extent, neuronal M₁ mAChRs; however, AChR pharmacology has been plagued by a lack of good receptor subtype-selective compounds. Mice lacking M₃ mAChRs display intact cholinergically-mediated intestinal ion transport, suggesting a possible compensatory mechanism. Inflamed tissues often display perturbations in the enteric cholinergic system and reduced intestinal ion transport responses to cholinomimetics. The mechanism(s) underlying this hyporesponsiveness are not fully defined. Inflammation-evoked loss of mAChR-mediated control of epithelial ion transport in the mouse reveals a role for neuronal nicotinic AChRs, representing a hitherto unappreciated braking system to limit ACh-evoked Cl^- secretion. We suggest that: i) pharmacological analyses should be supported by the use of more selective compounds and supplemented with molecular biology techniques targeting specific ACh receptors and signalling molecules, and ii) assessment of ion transport in normal tissue must be complemented with investigations of tissues from patients or animals with intestinal disease to reveal control mechanisms that may go undetected by focusing on healthy tissue only.

Introduction

The epithelial lining of the intestine separates two biological compartments: the gut lumen (i.e. the external environment) from the interstitium. Under normal circumstances, the polarized epithelial cells impede the movement of luminal material into the body and regulate the passage of a variety of substances – nutrients, electrolytes and water – from lumen to body and vice versa. The ability to secrete and absorb electrolytes and fluids is critical to maintain proper hydration of the organism – the intestine and the organs that drain into them (e.g. salivary glands) secrete ∼9 l of fluid into the lumen daily, most of which (∼8 l) is reabsorbed before it reaches the rectum (Keely and Barrett, 2000). Proper hydration provides the aqueous environment necessary for the processes of contact digestion and nutrient absorption and provides surface lubrication to propel intestinal contents aborally. A properly hydrated gut provides the medium for activity of antimicrobial peptides and movement of water into the lumen contributes to host defence through its ability to flush noxious substances and harmful organisms (e.g. parasites and pathogenic bacteria), thus minimizing their contact with the gut wall. Therefore, regulation of water movement is critical from both physiological (i.e. nutrition) and immunological (i.e. host defence) perspectives. If the process of intestinal hydration becomes dysregulated, as during acute intestinal infections or in the context of intestinal inflammation, diarrhoea or constipation can result.

It is not surprising then that water movement into and out of the intestine is highly regulated. There are three known mechanisms by which water can cross the epithelial barrier: (1) absorption along with sodium and glucose via sodium-coupled glucose transporters in the small intestine; (2) specific molecular water channels known as aquaporins; (3) passive osmosis driven by the directed transport of ions across epithelial cells. The expression and function of aquaporins are well described in the kidney, lung, eye and brain (reviewed in King et al., 2004), and their expression and function in the intestinal epithelium is currently the subject of much research (Mobasheri et al., 2004; Laforenza et al., 2005a, 2005b). Sodium-coupled glucose transporters and aquaporins are not discussed here, and we refer the reader to recent reviews on these subjects (Loo et al., 2002; Masyuk et al., 2002; Shachar-Hill and Hill, 2002; Matsuzaki et al., 2004). Here, we review control of enteric epithelial ion transport by acetylcholine, drawing attention to species-specific differences, differences in tissues from naïve animals compared to those with disease and comment on putative research directions.

Intestinal ion transport

Ion transport is an active process driven by electrochemical gradients that are established by the Na⁺/K⁺-ATPase pump located in the basolateral membrane of polarized epithelial cells. The magnitude, rate and direction of ion flux are dependent on the integrated activity of specific channels, transporters and exchangers asymmetrically expressed on the apical and basolateral cell membranes (Figure 1). The activity of the ion transport machinery is regulated by a variety of compounds signalling through membrane-bound or intracellular receptors to alter levels of intracellular messengers, such as Ca²⁺ and cyclic nucleotides.

Figure 1.

Schematic representation of the intestinal epithelial Cl⁻ secretory pathway that is upregulated by ACh. The electrochemical gradient is established by the basolateral Na/K-ATPase pump; this allows entry of Na⁺, Cl⁻ and K⁺ through the basolateral membrane via the NKCC1 cotransporter; K⁺ is recycled through basolateral K⁺ channels, one of which is controlled by [Ca²⁺]_i, likely the intermediate conductance K⁺ channel (IK), and Cl⁻ is shunted to the apical membrane, where it exits into the lumen via the apical CFTR Cl⁻ channel and possibly also through an apical Ca²⁺-activated Cl⁻ channel (CaCC, designated Cl_Ca). For additional details, refer to Barrett and Keely (2000).

Study of intestinal ion transport has been greatly facilitated by the development of the Ussing chamber (Ussing and Zerahn, 1951). With this technique, passive flow of ions across a tissue or epithelial cell layer is eliminated by balancing electrical, osmotic, hydrostatic and chemical gradients across the preparation, such that only active ion transport is measured. In the Ussing chamber, electrodes are placed close to each side of the tissue to allow detection of the spontaneous potential difference (PD) across the epithelium, generated as a consequence of active ion transport. The voltage-clamp apparatus matched to the Ussing chamber allows for the PD to be held at a predetermined value by injecting current via two additional silver–silver chloride electrodes, creating a short-circuit current (I_SC). Typically, the PD is clamped at zero volts and the amount of injected current required to maintain this reflects the net active ion transport across the preparation. Using the I_SC and the PD values (obtained by switching intermittently to an open certain mode), tissue conductance, which indicates passive ion flow, can be calculated using Ohm's law. By substituting specific ions (e.g. isethionate and acetate for Cl⁻) or using radiolabelled ions in the bathing buffers, by selectively blocking or removing specific receptors, channels, exchangers, pumps and transporters, and by blocking non-epithelial cell signalling (e.g. nerve block by tetrodotoxin (TTX), which blocks fast Na⁺ channels on most, but not all, enteric nerves without affecting epithelial Na⁺ channels), the contributions of these components to active electrolyte transport is determined. In the same manner, the effects of nerve stimulation and biological agents (e.g. biogenic amines, bacterial toxins) on active intestinal ion transport can be assessed in tissues or epithelial preparations mounted in Ussing chambers.

Acetylcholine

One of the most important biological regulators of intestinal ion transport is the major parasympathetic mediator acetylcholine (ACh), which is synthesized by the catalytic conversion of acetyl-CoA and choline to CoA and ACh by choline acetyltransferase (ChAT) (Hebb, 1954; Hebb and Whittaker, 1958). ACh was described as a classical neurotransmitter in the early 1920s (reviewed in Eiden, 1998). More recently, a distinction has been made between ACh as a neurotransmitter and ACh as a signalling molecule in non-neuronal tissue (e.g. blood) and in organisms that lack a nervous system (e.g. protozoa, algae, bacteria, primitive plants) (Grando et al., 2003; Wessler et al., 2003). As a classical neurotransmitter, ACh synthesis is followed by its transport via the vesicular ACh transporter (VAChT) to synaptic vesicles, from which it is released en masse from nerve terminals as a result of membrane action potentials. However, in non-neuronal cells ACh is not stored in vesicles and its release, which is likely facilitated by organic cation transporters, is continuous and much slower compared to ACh released from nerves (Wessler et al., 2003). In all tissues, ACh is rapidly and specifically degraded by acetylcholinesterase (AChE), although in mammals a second, nonspecific cholinesterase, designated butyrylcholinesterase, which is abundant in the intestine, may also contribute to the degradation of ACh (Massoulié et al., 1993; Li et al., 2000).

As with other neurotransmitters, direct measurement of ACh can be accomplished through microdialysis followed by high-performance liquid chromatography: this method has been applied in the intestine (Furuichi et al., 2001). However, owing to the difficulty and invasiveness of this technique and because ACh does not lend itself to reliable fixation for detection in preserved tissues, the presence of ACh is often inferred based on expression of the enzymatic components of the cholinergic pathway (Eckenstein and Thoenen, 1982). Analyses of ChAT and VAChT are more accurate than AChE for determining the presence of ACh, as the breakdown of ACh need not occur in the same location as its production. Additionally, VAChT is only relevant to neuronal ACh, leaving ChAT as the universal surrogate for ACh. ChAT activity can be determined radiochemically in tissue homogenates and the protein localized by immunohistochemistry (IHC) with commercial antibodies (Fonnum, 1975; Eckenstein and Thoenen, 1982; Sharkey et al., 1998).

Neuronal cholinergic networks, as defined by IHC detection of ChAT, relating to the control of intestinal motility and, to a lesser extent, ion secretion have been extensively studied in the guinea-pig (Costa et al., 2000; Furness, 2000; Brookes, 2001; Bornstein et al., 2004). Although enteric nerve signalling is well defined in the guinea-pig (Brown and Timmermans, 2004), it is important that this information be complemented by delineation of cholinergic pathways in mammalian species commonly used as models of human intestinal (patho)physiology/disease. Table 1 details the enteric localization of ACh in several mammalian species, including humans. The divide between ACh as a neurotransmitter and ACh as a signalling molecule is especially apparent in the intestine and cholinergic pathways exhibit regional differences along the length of the intestine. Also, while several cholinergic pathways are conserved across species, there are differences between species, one notable example being the colocalization of ChAT and vasoactive intestinal polypeptide (VIP) in secretomotor neurons of the submucosal plexus (SMP) in humans (Schneider et al., 2001), a pattern not observed in guinea-pigs at least. Although the enteric nervous system can be considered a discrete entity, cholinergic input from extrinsic nerves that project fibres into the gut (i.e. vagus and pelvic splanchnic nerves) should be recognized as a component in the bigger picture perspective of cholinergic control of physiological events in the gut.

Table 1.

Localization of ACh production in the mammalian gut

Species	Tissue	ChAT/VAChT-positive cell typesa	References
Guinea-pig	Small intestine	MP: Excitatory muscle motor neurons, Ascending interneurons to excitatory circular muscle motor neurons of MP, descending interneurons (three types that all are ChAT-IR – those containing ChAT/5-HT are involved in secretomotor reflex), IPANs (mucosal varicose processes likely release ACh), intestinofugal nerves to prevertebral gangliaSMP: Secretomotor/vasodilator neurons, secretomotor (non-vasodilator) neurons, IPANs (mucosal varicose processes likely release Ach)	McMurray et al. (1993), Neunlist et al. (1999), Costa et al. (2000), Furness (2000), Brookes (2001), Furness et al. (2004), Harrington et al. (2005)
	Large intestine	MP: Ascending neurons, IPANs, excitatory muscle motor neuronsSMP: Ascending secretomotor/vasodilator neurons, secretomotor neurons EC cells of small intestine
Mouse	Small intestine	MP: ∼60–65% of neurons; including Excitatory muscle motor neurons, Descending interneuronsSMP: ∼40% of neurons; includes likely secretomotor neurons, secretomotor/vasodilator neurons	Sang and Young (1998)
	Large intestine	MP: ∼55% of neurons; includes excitatory muscle motor neurons, descending interneuronsSMP: ∼20% of neurons; includes likely secretomotor neurons, secretomotor/vasodilator neurons
		Note: ChAT-IR EC cells observed in both SI and LI; ChAT- and VAChT-IR nerve terminals observed throughout MP, SMP, muscle layers, at the base of glands and associated with submucous blood vessels in both SI and LI
Rat	Small and large intestine	MP: Nerve cell bodies and fibres (unidentified subtypes)SMP: Nerve cell bodies and fibres (unidentified subtypes) EC cells of small intestine	McMurray et al. (1993)
Pig	Small intestine	MP: ∼80% of total traced neurons; includes IPANs, excitatory muscle motor neurons, secretomotor neurons, descending interneurons, intestinofugal neuronsOSP: ∼85% of total traced neurons; includes IPANs, excitatory muscle motor neurons, secretomotor neurons, descending interneurons, intestinofugal neuronsISP: ∼25% of total traced neurons; includes IPANs and secretomotor neurons Nerve fibres within circular muscle layer; cell bodies and fibres at the margins of PP follicles; occasional fibres within PP follicles and domes and adjacent to submucosal blood vessels; neurons of interfollicular ganglia, Note: minor discrepancies between jejunum and ileum	Kulkarni-Narla et al. (1999), Hens et al. (2000), Brown and Timmermans (2004)
Human	Small and large intestine	MP: Excitatory muscle motor neurons, descending interneurons, ascending interneurons, IPANsSMP: Secretomotor neurons EC cells, T cells	Porter et al. (1996), Kawashima and Fujii (2000), Porter et al. (2002), Schemann and Neunlist (2004)

Abbreviations: EC, enterochromaffin; IPANs, intrinsic primary afferent neurons; IR, immunoreactive; ISP, inner submucosal plexus; LI, large intestine; MP, myenteric plexus; OSP, outer submucosal plexus (ISP and OSP are subdivisions of SMP found in larger mammals, i.e. pig); PP, Peyer's patch; SI, small intestine; SMP, submucosal plexus.

^aBold text indicates neurons believed to be involved in intestinal secretion and secretory reflexes in the guinea-pig.

Acetylcholine receptors and pharmacology

ACh receptors (AChRs) fall into two categories: metabotropic G-protein-coupled muscarinic receptors, of which there are five known subtypes, M₁–M₅, and ionotropic nicotinic receptors, which are made up of five subunits that may be in homomeric or heteromeric combinations. Muscarinic (mAChRs) and nicotinic AChRs (nAChRs) are expressed on neuronal as well as on non-neuronal cells in the gut (Tables 2 and 3). Signalling pathways and functions associated with mAChRs have been reviewed in detail elsewhere (Caulfield, 1993; Lanzafame et al., 2003).

Table 2.

Non-pharmacological demonstration of muscarinic ACh receptors in the mammalian gut

Species	Tissue	Receptor subtype detected	Cell type/tissue region	Method	References
Rabbit	Ileum	M1 (3%)a, M2 (69%), M3 (4%), M4 (12%), M5 (0%)	ND	IP	Dorje et al. (1991a)
Mouse	Ileum	M2, M4	Circular and longitudinal smooth muscle (M2 only); M2 and M4 colocalization in cell bodies and fibres of ChAT-positive myenteric neurons	IF	Takeuchi et al. (2005), Haberberger et al. (2006)
	Colon	M1	Epithelium	RT-PCR
Rat	Ileum	M1 (17%), M2 69%), M3 (7%)	Epithelial cells (adult and fetal), including endocrine L cells	IF; IP	Wall et al. (1991), Levey (1993), Anini et al. (2002)
	Colon	M4	Nerve fibres (likely presynaptic)	IHC
Sheep	Ileum	M1, M2, M3 M4, M5	Adult smooth muscle	IHC	Oyachi et al. (2003)
Pig	Ileum	M2, M3	Smooth muscle	NB	Maeda et al. (1988)
Human	Ileum	M3	Mucosa (adjacent to epithelium), muscularis mucosa, submucosal blood vessels	IF	Sato et al. (1999), Oue et al. (2000), Lindqvist et al. (2002), Anini and Brubaker (2003), Banks et al. (2004)
	Colon	M1, M2, M3	Epithelial cells, endocrine L cells	IF
	(juvenile)	M2, M3	Mucosal layer, smooth muscle, cell bodies and fibres of myenteric plexus neurons	IHC; ISH
	Colon	M1	Mucosal layer	IHC
	Blood	M3	Crypt epithelial cells	IF
	Cultured cell lines	M1, M2, M3, M4, M5	MNL; Note: variable expression of M1–M3 among subjects	RT-PCR
		M1, M2, M3	NCI-H716 endocrine L cells	WB; RT-PCR
		M1, M2, M3, M4, M5	CEM, MOLT-3, HUT-78, HPB-ALL and Jurkat T cells; Note: variable expression of M1–M3 among cell lines	RT-PCR
		M1, M2, M3, M4, M5	BALL-1, Daudi and NALM-6 B cells; Note: variable expression of M1–M3 among cell lines	RT-PCR

Abbreviations: IF, immunofluorescence; IHC, immunohistochemistry; IP, immunoprecipitation; ISH, in situ hybridization; MNL, mononuclear leukocytes; NB, Northern blot; ND, not determined; RT-PCR, reverse transcription-polymerase chain reaction; WB, Western blot.

^aPercentage of total receptors precipitated (IP experiments performed in whole tissue).

Table 3.

Non-pharmacological demonstration of nicotinic ACh receptors in the mammalian gut

Species	Tissue	Receptor subunits detected	Cell type/tissue region	Method	References
Guinea-pig	Ileum and Colon	α3, α5, β4 (?)	Nerve cell bodies and fibres – MP (more abundant staining in colon): inhibitory and excitatory motor neurons, secretomotor neurons, afferent and/or ascending interneurons; SMP: secretomotor neurons, interneurons	IF; IHC	Kirchgessner and Liu (1998), Obaid et al. (1999), Zhou et al. (2002), Glushakov et al. (2004)
		β2	EC cells	IF; IHC
		α7	SMP neurons of small intestine Nerve cell bodies and fibres of MP and SMP; especially abundant in SMP	IF; IHC
	Cultured cells	α3, α5, β4 (?), β2, α7	Primary cultured MP neurons of small intestine (only a few neurons positive for α7); Note: colocalization of nAChR subunits with cholinergic and non-cholinergic neurons	IF
Human	Ileum and Colon	α3, (α5, β4 ?)	MP and SMP neurons and fibres traversing smooth muscle; Epithelium; Lymphoid tissue; Note: variable distribution among subjects	ISH; IHC	Hiemke et al. (1996), Battaglioli et al. (1998), Sato et al. (1999), Richardson et al. (2001), Richardson et al. (2003), Summers et al. (2003)
	Blood	α3, α4	CD4+ Lymphocytes; Note: variable expression among subjects	ISH; IHC
		α2, α5, α6, α7, β2	MNL; Note: variable expression of α6 and β2 among subjects	RT-PCR
	Cultured cell lines	α4, α5, α7, β1	HT29 colonic epithelial cell line	RT-PCR
		α3	Jurkatt and MOLT-4T cells	NB
		α2, α3, α5, α6, α7, β4	CEM, MOLT-3, HUT-78, HPB-ALL and Jurkat T cells	RT-PCR
		α2, α5, α6, α7, β2, β4	BALL-1, Daudi and NALM-6 B cells	RT-PCR

Abbreviations: EC, enterochromaffin; IHC, immunohistochemistry; IF, immunofluorescence; ISH, in situ hybridization; MNL, mononuclear leukocytes; MP, myenteric plexus; NB, Northern blot; SMP, submucosal plexus.

Briefly, odd numbered mAChRs couple to G_q/11 α subunits to activate phospholipase C (PLC), whereas even numbered mAChRs couple to the G_i α subunit to inhibit adenylate cyclase. Activation of the PLC pathway leads to degradation of membrane-bound phosphatidylinositol 4,5-bisphosphate (PIP₂) into diacyl glycerol (DAG) and inositol-1,4,5 triphosphate (IP₃), which, together, lead to activation of protein kinase C (PKC) and increased levels of intracellular Ca²⁺ as the main outcome. nAChRs are allosteric ion channel proteins that are permeable to monovalent cations and, to a lesser extent, Ca²⁺. Binding of an agonist (i.e. ACh) to nAChRs preferentially stabilizes it in an open or active conformational state. Subsequent entry of cations through the open channel depolarizes the cell membrane, thus contributing, in the case of a neuron, to the generation of an action potential. The structure and function of nAChRs has been described in detail in the context of the central nervous system and the neuromuscular junction (Hogg et al., 2003; Gotti and Clementi, 2004).

Several agonists of AChRs mimic the effects of ACh and are often used in its place experimentally because they are less susceptible than ACh to degradation by AChE. However, mAChR agonists, including ACh, are not highly subtype selective and this needs to be appreciated when evaluating the data from studies relying on these reagents (Hulme et al., 1990). Additionally, some mAChR agonists, such as pilocarpine, McN A-343 and, surprisingly, even ACh and bethanechol (BCh), may behave as partial, rather than full, agonists in some experimental systems (Yule et al., 1993; Sharif et al., 1995). Receptor subtype selectivity has also been a problem with commonly used mAChR antagonists (Hulme et al., 1990). However, the characterization of toxins isolated from the venom of mamba snakes has provided several compounds that are highly selective toward the M₁ and M₄ subtypes of mAChRs (see Table 4; reviewed in Karlsson et al., 2000). For an extensive discussion of mAChR pharmacology and associated physiology, refer to Caulfield (1993), Eglen et al. (1996), Caulfield and Birdsall (1998) and Eglen (2005).

Table 4.

Affinity profiles of AChR agonists and antagonists for known AChR subtypes

Drug	Ki (nM)	References
mAChR agonist
Carbachol	205 (M1), 93.4 (M3), 35.7 (M5)a	Kukkonen et al. (1996)
mAChR antagonist
AF-DX 116	1300 (M1), 186 (M2), 838 (M3), 2800 (M5)b 344.6 (M1), 32.4 (M2), 198.6 (M3), 123.5 (M4)b	Buckley et al. (1989), Billard et al. (1995)
AF-DX 384	30.9 (M1), 6.03 (M2), 66.1 (M3), 10.0 (M4), 537.0 (M5)b	Dorje et al. (1991b)
Atropine	0.21 (M1), 1.5 (M2), 0.15 (M3), 0.21 (M5)b 0.28 (M1), 0.76 (M2), 0.19 (M3), 0.13 (M4), 0.24 (M5)a	Buckley et al. (1989), Moriya et al. (1999)
4-DAMP, 4-diphenylacetoxyl-N- methylpiperidine methiodide	0.58 (M1), 3.8 (M2), 0.52 (M3), 1.27 (M4), 1.05 (M5)b 0.57 (M1), 7.3 (M2), 0.37 (M3), 0.72 (M4), 0.55 (M5)a	Dorje et al. (1991b), Moriya et al. (1999)
Hexahydro-sila-difenidol	22.4 (M1), 131.8 (M2), 15.5 (M3), 31.6 (M4), 93.3 (M5)b	Dorje et al. (1991b)
Hexocyclium	2.3 (M1), 23 (M2), 1.4 (M3), 5.5 (M4), 3.7 (M5)b	Buckley et al. (1989)
Himbacine	107.2 (M1), 10.0 (M2), 93.3 (M3), 11.0 (M4), 489.8 (M5)b 49.2 (M1), 4.4 (M2), 29.6 (M3), 5.5 (M4)b	Dorje et al. (1991b), Billard et al. (1995)
Methoctramine	16 (M1), 3.6 (M2), 118 (M3), 57 (M5)b 50.1 (M1), 13.2 (M2), 213.8 (M3), 31.6 (M4), 134.9 (M5)b	Buckley et al. (1989), Dorje et al. (1991b)
M1-toxins:		Jerusalinsky et al. (2000), Karlsson et al. (2000)
MT7	0.1 (M1), >2000 (M2), >2000 (M3), >2000 (M4), >2000 (M5)b
MT1	22–49 (M1), >1000 (M2), >1000 (M3), 29-58 (M4), >1000 (M5)b
MT2	630 (M1), >2000 (M2), >2000 (M3), 1900 (M4), >2000 (M5)b
MT4	62 (M1), >1000 (M2), >1000 (M3), 87 (M4), >1000 (M5)b
MT5	180 (M1), >1000 (M2), >1000 (M3), 540 (M4), >1000 (M5)b
M4-toxins:		Jerusalinsky et al. (2000), Karlsson et al. (2000)
MT3	78–1100 (M1), >1000 (M2), >1000 (M3), 1.4-2 (M4), >1000 (M5)b
MT6	190 (M1), >425 (M2), >425 (M3), 3.6 (M4), >425 (M5)b
MTα	23 (M1), 44 (M2), 3 (M3), 5 (M4), 8 (M5)b
MTβ	>1000 (M1), >2000 (M2), 140 (M3), 120 (M4), 350 (M5)b
Oxybutinin	0.66 (M1), 13 (M2), 0.72 (M3), 0.54 (M4), 7.4 (M5)a	Moriya et al. (1999)
Pirenzepine	16 (M1), 906 (M2), 180 (M3)b 6.3 (M1), 223.9 (M2), 138.0 (M3), 37.2 (M4), 89.1 (M5)b 4.6 (M1), 211.8 (M2), 37.8 (M3), 21.1 (M4)b 11.4 (M1), 373 (M3), 69.3 (M5)a 7.1 (M1), 303 (M2), 75 (M3), 17 (M4), 66 (M5)a	Buckley et al. (1989), Dorje et al. (1991b), Billard et al. (1995), Kukkonen et al. (1996), Moriya et al. (1999)
Scopolamine	0.085 (M1), 0.88 (M2), 0.063 (M3), 0.1 (M4)b	Billard et al. (1995)
Silahexacyclium	2.0 (M1), 35 (M2), 1.2 (M3), 3.2 (M4), 2.0 (M5)b	Buckley et al. (1989)
Tiquizium	4.1 (M1), 4.0 (M2), 2.8 (M3), 3.6 (M4), 8.2 (M5)a	Moriya et al. (1999)
[R]-Trihexyphenidyl	0.37 (M1), 7.08 (M2), 2.45 (M3), 0.83 (M4), 5.01 (M5)b	Dorje et al. (1991b)
Tripitramine	1.6 (M1), 0.3 (M2), 38 (M3), 6 (M4), 3.4 (M5)b	Maggio et al. (1994)
nAChR agonist
Anabaseine	58 (Rat α7)	Kem et al. (1997)
Carbachol	4130 (Bovine α3β4)c	Free et al. (2003)
Cytisine	517 (Bovine α3β4)c	Free et al. (2003)
1,1-Dimethyl-4-phenylpiperazinium	469 (Bovine α3β4)c	Free et al. (2003)
(−)-Epibatidine	230 (Rat α7) 0.15 (Human α3β2)d, 0.031 (Human α3)e, 3.1 (Human α7)e	Gerzanich et al. (1995), Bunnelle et al. (2004)
(+)-Epibatidine	0.23 (Human α3β2)d, 0.00059 (Human α3)e, 9.8 (Human α7)e	Gerzanich et al. (1995)
(±)-Epibatidine	0.3 (Bovine α3β4)c 233.6 (Rat α7)	Davies et al. (1999), Free et al. (2003)
Nicotine	400–14 000 (Rat α7) 517 (Bovine α3β4)c	Kem et al. (1997), Davies et al. (1999), Free et al. (2003), Bunnelle et al. (2004)
RJR-2403	36 000 (Rat α7)	Bunnelle et al. (2004)
nAChR antagonist
α-Bungarotoxin	0.7 (Rat α7)	Davies et al. (1999)
Methyllycaconitine	1.15 (α7)f 1.1–4.3 (Rat α7) 420 (Bovine α3β4)c	Alkondon et al. (1992), Maggi et al. (1999), Davies et al. (1999), Free et al. (2003)
MG 624 (4-oxystilbene derivative)	106 (α7)f	Maggi et al. (1999)
D-Tubocurarine	1570 (α7)f 105 (Bovine α3β4)c	Maggi et al. (1999), Free et al. (2003)

^aStably transfected Sf9 insect cells.

^bStably transfected CHO-K1 cells.

^cStably expressed in human embryonic kidney (HEK) 293 cells.

^dStably expressed in Xenopus oocytes.

^eImmuno-isolated from SH-SY5Y neuroblastoma cells.

^fImmuno-immobilized from chick optic lobe.

Note: When affinity values were reported as pK_i, the negative antilogarithm was calculated to determine K_i value; unless otherwise specified, affinity for α7 subunits is determined by displacement of ³[H]-α-bungarotoxin binding.

nAChR pharmacology is complicated by the fact that, while at least 11 genes have been identified for mammalian non-muscle nAChR subunits (α2–α7, α9–α10, β2–β4), not all of the subunit combinations natively expressed in different tissues and species are known (Nicke et al., 2004). Generally, the α4, β2 and α7 subunits predominate in the central nervous system, whereas α3, β4 and α7 subunits are common in the peripheral nervous system. α7 subunit-containing receptors occur in homomeric combination, whereas α4, α3, β2 and β4 subunits occur in heteromeric combinations of both α and β subunits (Gotti and Clementi, 2004). A few compounds are able to distinguish between homomeric α7 nAChRs and non-α7-containing nAChRs, most notably the snake venom component α-bungarotoxin and methyllycaconitine. Both compounds competitively and selectively antagonize homomeric α7 nAChRs (Davies et al., 1999). Conus snail venom has also provided several potentially competitive nAChR antagonists, the α-conotoxins, which show subtype selectivity for specific known subtype combinations, although it is unknown if this subtype selectivity applies to all species (Nicke et al., 2004). Non-competitive antagonists of nAChRs include the archetypal ganglionic blockers mecamylamine and hexamethonium (HEX), which demonstrate relatively low affinity for nAChRs (K_i values in the mM range) (Macallan et al., 1988; Davies et al., 1999).

Table 4 lists the affinities of several major AChR agonists and antagonists for specific AChR subtypes. Table 5 lists the affinities of mAChR agonists and antagonists that have been determined in intestinal tissue from a few different species. These drugs have been instrumental in profiling mAChR subtypes involved in intestinal ion transport. Subtype characterization of nAChRs involved in intestinal ion transport has been hindered by a lack of knowledge of the precise nAChR subtypes expressed in the gut. Details of nAChR involvement in intestinal ion transport requires better characterization of the receptor subtypes expressed in the gut before pharmacology can be fully validated. It should be cautioned that, when dealing with promiscuous compounds (i.e. the majority of mAChR ligands), studies using only one compound to implicate the involvement of a specific receptor subtype are not reliable – AChR subtype characterization requires a diverse array of pharmacological tools. Thus, ACh receptor subtype involvement inferred from the use of poorly selective antagonists and/or partial agonists should be verified using newer, more selective drugs and with molecular biology techniques, such as gene targeting strategies to delete specific AChR subtypes (as reviewed in Wess (2004) for mAChRs).

Table 5.

Affinity profiles of mAChR agonists and antagonists in intestinal tissue

Drug	Tissue preparation	Ki (nM)	References
mAChR agonist
Bethanechol	Guinea-pig ileum – mucosal scrapings	53 000	Carey et al. (1987), Kuwahara et al. (1987b)
	Guinea-pig proximal colon – mucosal scrapings	53 000
	Guinea-pig distal colon – mucosal scrapings	35 000
Carbachol	Guinea-pig ileum – mucosal scrapings	8500	Carey et al. (1987), Kuwahara et al. (1987b)
	Guinea-pig proximal colon – mucosal scrapings	9800
	Guinea-pig distal colon – mucosal scrapings	5500
McN A343	Guinea-pig ileum – mucosal scrapings	180	Carey et al. (1987)
MAChR Antagonist
AF-DX 116	Canine ileum – purified synaptosomal fraction from deep muscular nerve plexus	6200	Chandan et al. (1991a, 1991b), Kostka et al. (1992), O'Malley et al. (1995)
	Pig jejunum – mucosa-submucosa	300
	Rat colon – mucosa-submucosa	2189
Atropine	Pig jejunum – mucosa-submucosa	0.4	Chandan et al. (1991b), O'Malley et al. (1995)
	Rat colon – mucosa-submucosa	0.87
4-DAMP	Canine ileum – purified synaptosomal fraction from deep muscular nerve plexus	39	Carey et al. (1987), Kuwahara et al. (1987b), Chandan et al. (1991a, 1991b),
	Guinea-pig ileum – mucosal scrapings	1.8	Kostka et al. (1992), O'Malley et al. (1995)
	Pig jejunum – mucosa–submucosa	2.5
	Rat colon – mucosa–submucosa	5.2
	Guinea-pig proximal colon – mucosal scrapings	3.0
	Guinea-pig distal colon – mucosal scrapings	2.2
Methoctramine	Canine ileum – purified synaptosomal fraction from deep muscular nerve plexus	2250	Kostka et al. (1992)
Pirenzepine	Canine ileum – purified synaptosomal fraction from deep muscular nerve plexus	2540	Carey et al. (1987), Kuwahara et al. (1987b), Chandan et al. (1991a, 1991b), Kostka et al. (1992), O'Malley et al. (1995)
	Guinea-pig ileum – mucosal scrapings	35
	Guinea-pig proximal colon – mucosal scrapings	100
	Guinea-pig distal colon – mucosal scrapings	170
	Porcine jejunum – mucosa–submucosa	87.5
	Rat colon – mucosa–submucosa	155.0

Note: When affinity values were reported as pK_i, the negative antilogarithm was calculated to determine K_i value.

Cholinergic control of enteric epithelial ion transport

Baseline ion transport

Typically, the baseline active ion transport of normal mammalian intestinal tissue, as demonstrated by I_SC, reflects a negative charge on the luminal side and positive charge on the serosal side of the tissue (Field, 1974). Absorption of Na⁺ and Cl⁻ typically accounts for most of this baseline ion transport; the relative contributions of K⁺, HCO₃⁻ and H⁺ vary according to intestinal region and species. Numerous studies indicate that baseline epithelial transport is in part due to tonic nerve activity, and this is via neurons of the SMP (Furness, 2000; Vanner and MacNaughton, 2004), with a small contribution from secretomotor neurons that project from the myenteric plexus into the mucosa (Costa et al., 2000). This concept is supported by the fact that gut preparations stripped of the outer muscle layers and attendant myenteric plexuses still display positive baseline I_SC values.

Additional evidence in support of neural control of baseline I_SC is provided by consideration of the effect of the neuronal blocker TTX, which can elicit a drop in baseline I_SC. For instance, TTX applied to muscle-stripped segments of porcine distal jejunum produces a dramatic drop in baseline I_SC, although this effect is less pronounced in the proximal jejunum (Rangachari and McWade, 1986; Hildebrand and Brown, 1990; Chandan et al., 1991a, 1991b). Similarly, TTX treatment reduces the baseline I_SC across muscle-stripped dog, rabbit and porcine colon and full-thickness mouse jejunum and colon (Sheldon et al., 1989; Biagi et al., 1990; Traynor et al., 1991; Sayer et al., 2002). Data in favour of, and refuting, the ability of TTX to lower baseline I_SC in muscle-stripped human left/sigmoid colon has been presented (Hubel et al., 1987; Kuwahara et al., 1989), and, in support of the latter, TTX treatment does not alter baseline I_SC in the rat colon (Zimmerman and Binder, 1983). It is difficult to determine the source of variability in these studies: differences in tissue preparation, subtle region-specific differences or environmental variability (i.e. diet, resident intestinal flora) are possibilities. Moreover, use of TTX gives no indication of the phenotype of neuron, or the possibility of signalling within the ENS via interneurons that impacts upon baseline I_SC. Also, analyses of the role of neurons in the control of epithelial ion transport should, for example, be complemented by identification of the ion fluxes responsible for the TTX effect.

Although technically demanding, some enteric tissues (e.g. dog colon) can be stripped of both the outer muscle layers and the submucosa, leaving sheets of epithelial crypts and a supporting basement membrane. [It should be noted that the procedure to remove the muscle and mounting in the Ussing chamber have the potential to activate stretch receptors that could elicit intracellular and intercellular signalling cascades that could affect active ion transport across the enterocyte.] These preparations maintain active ion transport that is unaffected by TTX, although baseline I_SC values are lower than in preparations containing an intact submucosa (Rangachari and McWade, 1986; Diener et al., 1989). Monolayers of human colon-derived epithelial cell lines (i.e. T84 cells, HT-29 cells, Caco-2 cells) also exhibit baseline ion transport in the absence of any neural input, demonstrating that whereas enteric nerves modulate epithelial baseline active ion transport, the enterocyte itself is primarily responsible for the generation of driving forces to support vectorial ion movements. Thus, baseline production of intracellular ATP allows for energy-dependent ionic gradients to be established while other mediators, such as cAMP, positively regulate the opening of apical and basolateral ion channels. Other non-neuronal mediators present in the mucosa (e.g. prostaglandins (PGs); Craven and DeRubertis, 1983) probably also contribute to baseline levels of active ion transport, although this has mostly been studied in the context of stimulated ion transport (see following sections).

Using atropine (ATR) as the gold standard mAChR antagonist has revealed contrasting data in terms of cholinergic input to tonic enteric epithelial I_SC. Thus, full thickness mouse colon, muscle-stripped rat colon and porcine jejunum respond to ATR treatment with a decrease in I_SC (Diener et al., 1989; Chandan et al., 1991b; Sayer et al., 2002). Contrarily, treatment with ATR does not affect baseline I_SC in guinea-pig small bowel, porcine jejunum, and rabbit and human colon (Cooke, 1984; Keast et al., 1985; Hubel et al., 1987; Kuwahara et al., 1989; Biagi et al., 1990; Hildebrand and Brown, 1990; Chandan et al., 1991b). Few studies have assessed the impact of nAChR antagonists on baseline I_SC: porcine jejunum, mouse mid-colon and rat distal colon all display small but consistent reductions in baseline I_SC following treatment with HEX (Hildebrand and Brown, 1990; Sun et al., 2000; Sayer et al., 2002), although this effect requires high (100 μM) concentrations of HEX in the porcine jejunum (Hildebrand and Brown, 1990; Chandan et al., 1991a).

Collectively, the data support the contention that the enteric nervous system contributes to baseline epithelial ion transport, albeit to varying degrees between species and gut regions, and on the balance of the data it seems that ACh plays a minor role. Perplexingly, however, there is little consensus on the role of specific AChRs, with studies suggesting involvement of mAChRs but not nAChRs, nAChRs but not mAChRs, both mAChRs and nAChRs, or neither receptor type. Much of this variation is attributed to species and regional differences, although tissue preparation and environmental variability may also play a role. Additionally, some of the vagaries in the field could be reconciled by a consideration of the expression profiles of specific mAChR and nAChR subtypes and the location of these receptors (i.e. on stimulatory verses inhibitory neurons).

Stimulated ion transport

The majority of studies of cholinergic regulation of ion transport have focused on stimulated ion transport, either through electrical field stimulation (EFS – causes release of ACh and other neurotransmitters) or through application of a cholinergic compound to tissues in Ussing chambers. Thus, it has long been known that ACh and cholinergic agonists stimulate fluid and Cl⁻ secretion in the mammalian (including human) gut (Tidball, 1961; Hubel, 1976; Isaacs et al., 1976). It is also well established that the addition of carbachol (CCh) or BCh to muscle-stripped intestine (small or large bowel) results in a rapid onset, but transient (∼10–15 min duration), increase in I_SC that is due mainly to luminally directed Cl⁻ efflux across the epithelium (Cooke, 1984; Chandan et al., 1991b; Chough et al., 1993; Strabel and Diener, 1995). Additionally, cholinergic stimulation promotes HCO₃⁻ secretion and may inhibit Na⁺ absorption in some tissues (Browning et al., 1978; Chandan et al., 1991c; Geibel et al., 2000).

However, species- and tissue-specific differences are commonplace when considering: (1) the contribution of cholinergic nerves to the transient (1–5 min duration) increase in I_SC evoked by EFS, (2) the contribution of neurons to I_SC responses evoked by CCh or BCh and (3) involvement of mAChR and nAChR in cholinergic control of increases in I_SC. For example, CCh-induced ΔI_SC in guinea-pig ileum is TTX insensitive (Cooke, 1984), whereas colonic responses to CCh are reduced by TTX (Kuwahara et al., 1987b). In porcine proximal, but not distal jejunum, TTX pretreatment actually augments the increase in I_SC evoked by low concentrations of CCh (0.1–3 μM) (Chandan et al., 1991a, 1991b). Also, CCh- and BCh-evoked ΔI_SC in the rat small intestine are TTX insensitive (Przyborski and Levin, 1997), whereas, in the rat colon, I_SC responses to several AChR agonists, including ACh and CCh, are reduced by TTX (O'Malley et al., 1995). Along with the species and regional differences in TTX sensitivity to cholinergically mediated ion transport, there is further evidence that should caution against making broad conclusions based on individual studies: Prior et al. (2004) showed that colonic tissue from two lines of the Flinders strain of rats have different TTX sensitivities to CCh-induced increases in I_SC.

Although the relative contribution of nerves to the ΔI_SC evoked by CCh or BCh can be debated, there is little doubt that EFS elicits increases in I_SC that are due in large part to ACh release. Thus, EFS-induced ΔI_SC are typically reduced by ∼50% in the presence of ATR (Cooke et al., 1983a, 1983b; Cooke, 1984; Kuwahara et al., 1987a, 1987b; Hildebrand and Brown, 1990; Chandan et al., 1991a, 1991b; Javed and Cooke, 1992). Some investigators, based on the use of HEX, have implicated a small role for nAChR in the EFS response in whole thickness tissue (Hildebrand and Brown, 1990; Chandan et al., 1991a, 1991b). Thus, the majority of reports indicate that changes in epithelial ion transport as a consequence of ACh release (mimicked in vitro by EFS or application of exogenous cholinomimetics) are via mAChRs and pharmacological analyses do, in general, support the dogma that this is via epithelial M₃ and neuronal M₁ mAChRs (Cooke et al., 1983a; Cooke, 1984; Carey et al., 1987; Kuwahara et al., 1987a, 1987b; Diener et al., 1989; Traynor et al., 1991; Chandan et al., 1991a, 1991b; Javed and Cooke, 1992; O'Malley et al., 1995; Przyborski and Levin, 1997; Townsend et al., 2005). Stimulation of nAChRs does affect intestinal I_SC, typically in full-thickness tissue (i.e. with an intact myenteric plexus), but the magnitude of the ΔI_SC is significantly smaller than that evoked by CCh or BCh (Zimmerman and Binder, 1983; Diener et al., 1989; Hildebrand and Brown, 1990; Chandan et al., 1991a, 1991b; O'Malley et al., 1995; Sayer et al., 2002). This does not trivialize the importance of nAChR input to cholinergic control of epithelial ion transport, rather it highlights the fact that high concentrations of CCh, with its mixed mAChR and nAChR effects, are unlikely to reveal nAChR effects because these are over-run by the effects of mAChR ligation (Tapper et al., 1978; Sayer et al., 2002).

Surprisingly, cholinergically mediated ion transport is poorly understood in the mouse. In full-thickness jejunal tissue from suckling and adult mice (it is extremely difficult to strip away the outer muscle layers from the mouse gut), I_SC responses to EFS are partially mediated by mAChRs and CCh-induced ΔI_SC are TTX insensitive (Carey and Cooke, 1989; Sheldon et al., 1989). However, I_SC responses following application of the Na⁺ channel activator veratridine, which stimulates enteric nerves pharmacologically, rather than electrically, do not involve muscarinic receptors (Sheldon et al., 1990). Dimethylphenylpiperazinium (DMPP) (i.e. nAChR agonist)-induced increases in I_SC in the jejunum are, in accordance with other species, neuronally driven (Sheldon et al., 1989). CCh- and BCh-evoked increases in I_SC in the mouse colon are mediated in part by nerves (Sagmanligil and Levin, 1993; Carew and Thorn, 2000), although subsequent studies failed to confirm this (Sayer et al., 2002). The latter study found that HEX did not significantly affect ΔI_SC evoked by CCh, but that nicotine evoked a drop in I_SC that was inhibited by TTX pretreatment. Furthermore, as a striking addition to the field, mice lacking M₃ mAChRs produce I_SC responses to CCh or BCh of normal magnitudes, with M₁ mAChRs implicated as the compensating mechanism for the lack of the M₃ receptor subtype (Haberberger et al., 2006; Hirota and McKay, 2006).

There is a dearth of information regarding cholinergic regulation of ion transport in the human intestine. In muscle-stripped segments of left colon, 25–50% of the ΔI_SC evoked by EFS is due to cholinergic input (Hubel et al., 1987; Kuwahara et al., 1989) and responses to CCh are completely abolished by ATR (Kuwahara et al., 1989).

Clearly, the lack of data obtained from human tissue needs to be addressed, as does a comparison of the applicability of findings from mouse tissue, particularly as the mouse is the species of choice to model human intestinal disease. Additionally, whereas cholinergic control of enteric epithelial ion transport is, under normal circumstances, dominated by M₃ and M₁ mAChRs (lack of involvement of M₂ and M₄ mAChR subtypes is generally inferred from pharmacological data, but has never been directly proven, and M₅ mAChR involvement has not been addressed), the modulating effect of nAChRs should be better characterized, as should cholinergic neuronal interactions within the submucosa and between the submucosal and myenteric plexuses. Finally, in segments of rabbit small intestine containing Peyers's patches, CCh evoked a decrease rather than an increase in I_SC (Brayden and Baird, 1994) and greater involvement of nAChR has been implicated in colonic tissue from young compared to mature guinea-pigs (Powell and Reddix, 2000); thus, whereas the majority of our understanding of the control of enteric epithelial ion transport has been gleaned from analyses of tissues from adult animals, there is value in considering tissue from infants/young animals and from those with an enhanced complement of immune cells (see below).

Intracellular epithelial effects of cholinergic agonists: influence on ion transport

Assessment of rat intestine has demonstrated that ACh-driven epithelial Cl⁻ secretion is Ca²⁺ dependent (Hardcastle et al., 1984), activating a basolateral K⁺ efflux from the enterocyte, which is required for the subsequent increase in I_SC (Hardcastle and Hardcastle, 1986). Whole-cell patch-clamp recordings from rat and human colonic epithelial cells showed that ACh (and CCh) act through mAChRs to evoke membrane hyperpolarization, an effect that is dependent on the presence of Ca²⁺ (Yada and Okada, 1984; Yada et al., 1989; Devor et al., 1990; Bohme et al., 1991). This hyperpolarization is attributed to K⁺ efflux from the basolateral cell membrane, which, in the human T84 colonic epithelial cell line, occurs independently of cAMP-stimulated K⁺ efflux, indicating a separate Ca²⁺-activated K⁺ channel in these cells (McRoberts et al., 1985; Dharmsathaphorn and Pandol, 1986) (Figure 1). Furthermore, microelectrode recordings in the human colon-derived HT-29 epithelial cell line show that CCh-stimulated K⁺ efflux immediately follows, but is not dependent on, epithelial cell depolarization, a phenomenon attributed to Cl⁻ efflux from both apical and basolateral membranes (Bajnath et al., 1992a, 1992b), although the functional significance of this finding is unclear. CCh-induced increases in I_SC have consistently been associated with serosal-to-mucosal movement of Cl⁻ across epithelial cell monolayers (Dharmsathaphorn and Pandol, 1986), and most, but not all, of the studies to date favour a CCh-driven apical Cl⁻ conductance in T84 cells (frequently used as a model of secretory epithelia) (Dharmsathaphorn and Pandol, 1986; Cliff and Frizzell, 1990; Devor et al., 1990; Venglarik et al., 1990).

cAMP-dependent secretagogues (e.g. VIP, PGE₂ and forskolin) increase apical Cl⁻ conductance via the cystic fibrosis transmembrane conductance regulator (CFTR) (Mandel et al., 1986) and potentiate the I_SC response to CCh in human colonic epithelial cell lines and freshly isolated human colonic tissue (Dharmsathaphorn and Pandol, 1986; Mall et al., 1998). The Cl⁻ conductances evoked by cAMP versus Ca²⁺ are not identical, thus suggesting the existence of two distinct Cl⁻ channels in human intestinal epithelium (Cliff and Frizzell, 1990; Vaandrager et al., 1991; Bajnath et al., 1992a), a postulate supported by molecular analyses (Mohammad-Panah et al., 2001; Evans et al., 2004). Collectively, these studies indicate that activation of basolateral M₃ mAChRs raises intracellular Ca²⁺, resulting in Ca²⁺-dependent basolateral K⁺ efflux, which creates the driving force for flux of Cl⁻ into the gut lumen, observed in the Ussing chambers as an increase in I_SC (Kachintorn et al., 1993b) (Figure 1).

In human colonic epithelial cell monolayers, CCh stimulation induces turnover of membrane phospholipids (Kopp et al., 1989; Dickinson et al., 1992). Indeed, it has been elegantly demonstrated that inositol 3,4,5,6-tetrakisphosphate (IP₄) generation, subsequent to IP₃ production, contributes to the transient nature of the epithelial response to CCh, serving to turn off Cl⁻ secretion (Kachintorn et al., 1993a; Vajanaphanich et al., 1994). PKC appears to have a similar role in limiting the Cl⁻ secretory actions of CCh (Cohn, 1990; Kachintorn et al., 1992), although there is also evidence that at least one PKC isoform, PKCα, contributes initially to the stimulation of Cl⁻ secretion, while acting to inhibit this response in the longer run, possibly through inhibition of the basolateral Ca²⁺-activated K⁺ channel (Bajnath et al., 1992b; van den Berghe et al., 1992) (Figure 2). Analyses with the T84 cell line suggest that phosphorylation of the MAP kinases, extracellular signal-regulated kinase and p38, constitute additional ‘off' signals for CCh-stimulated epithelial chloride secretion (Keely et al., 1998; Keely and Barrett, 2003). This occurs via transactivation of the epidermal growth factor receptor (EGFr) and requires release of transforming growth factor-α (TGF-α), elevations in intracellular Ca²⁺, activation of calmodulin, and activation and association of Src kinase and PYK-2 with the EGFr (Uribe et al., 1996a; Keely et al., 2000; McCole et al., 2002). Interestingly, phospholipid production has also been linked to epidermal growth factor (EGF)-induced inhibition of Cl⁻ secretion (Uribe et al., 1996b). However, the phospholipid-generating effects of EGF are dependent on phosphatidylinositol 3-kinase, whereas those of CCh are not, demonstrating that mobilization of distinct intracellular signalling pathways can culminate in a similar physiological end point.

Figure 2.

Intracellular regulation of intestinal epithelial Cl⁻ secretion by ACh (based on the secretory pathway depicted in Figure 1). G_q/11 α subunits (coupled to odd-numbered mAChRs) stimulate membrane phospholipid turnover by PLC. Inositol 1,4,5-trisphosphate (I[1,4,5]P₃) stimulates a rise in [Ca²⁺]_i; eventual conversion of IP₃ products to IP₄ by phosphatidylinositol kinases turns off Cl⁻ secretion, possibly by inhibiting Ca²⁺-activated Cl⁻ channels (Cl_Ca). DAG activates classical (e.g. PKCα) and novel (e.g. PKCɛ) isoforms of PKC; PKCα and/or PKCɛ may be required for initial stimulation of apical Cl⁻ secretion, but ultimately serve to downregulate this process through effects on ion transport molecules. Refer to the text for further details.

Cholinergic–non-cholinergic interactions: influence on ion transport

ACh does not exist in isolation and so it is not surprising that other signalling molecules from nerves, stromal or immune cells influence cholinergic regulation of epithelial ion transport. The opportunities for such interaction are immense given the diversity of biogenic amines, neuropeptides and neurotransmitters that exist in the gut (at least 25 in the enteric nervous system alone; McConalogue and Furness, 1994). As examples: in the guinea-pig small intestine, noradrenaline inhibits the secretory responses to cholinergic nerve stimulation and somatostatin inhibits ΔI_SC evoked by CCh and DMPP (Keast et al., 1986); opioid analogues can inhibit tonic ACh release from the rabbit ileal SMP (Hautefeuille et al., 1985); neuropeptide Y and the structurally related peptide YY reduce baseline I_SC and diminish BCh-induced ΔI_SC in mouse jejunum (Riviere et al., 1993; Nakanishi et al., 1996; McKay et al., 1996a); increases in I_SC evoked by substance P (SP) in guinea-pig jejunum are reduced in the presence of ATR, implying SP-induced release of ACh as a significant part of the Cl⁻ secretory response (Perdue et al., 1987; Goldhill and Angel, 1998). Products of arachidonic acid metabolism also influence cholinergic responses. Thus, the prosecretory effects of leukotrienes in the guinea-pig colon are partially dependent on the activity of mAChRs (Hammerbeck and Brown, 1993), PGD₂ antagonizes the effects of BCh and CCh on canine proximal colon (Rangachari and Betti, 1993), and blockade of PG synthesis with indomethacin reduces the I_SC response evoked by CCh by 70% in human descending colon (Mall et al., 1998).

The observation that other neurotransmitters modulate the cholinergic control of enteric epithelial ion transport is intuitive, especially considering that most cholinergic nerves also synthesize and release other neurotransmitters/neuropeptides. Indeed, the implications of this cotransmission, occurring in the context of endogenous eicosanoids and cytokines (see below), for modulation of the cholinergic control of epithelial ion transport is immense, not to mention the effects of the temporal and stoichiometric relationship between the release of multiple mediators. Clearly, many outstanding issues remain: Do specific neurotransmitters simply oppose/antagonize the effect of ACh, and if so, how (e.g. activation of inhibitory neurons)? Are other neurotransmitters influencing the release of ACh under normal or pathophysiological conditions? Are specific neurotransmitters influencing expression or activity of ACh receptors or of subsequent intracellular signalling cascades? The answers to these questions will significantly enhance our understanding of the complexities of the cross-talk between the cholinergic system and other mediators in the control of enteric ion transport in health and disease.

Alterations to cholinergically mediated intestinal ion transport during pathology

Intestinal infection and inflammation profoundly effect water balance in the gut, leading to either secretory or malabsorptive diarrhoea or to constipation. However, and seemingly paradoxically, gut dysfunction is often characterized by a general hyporesponsiveness to prosecretory agents, including diminution of ΔI_SC evocable by cholinomimetics. Thus, tissue from diabetic rats (Perdue and Davison, 1988), rats exposed to 10 Gy radiation (Francois et al., 1998), rats infected with nematode parasites (Masson et al., 1996; Venkova and Greenwood-van Meerveld, 2006), mice with T-cell-driven enteropathies (McKay et al., 1999; Radojevic et al., 1999), aganglionic colon from children with Hirschsprung's disease (Hardy et al., 1993) and tissues from virtually every animal model of colitis (Goldhill et al., 1993; Kachur et al., 1995; Asfaha et al., 1999; Diaz-Granados et al., 2000; Miceli et al., 2002; Sanchez de Medina et al., 2002; Perez-Navarro et al., 2005a, 2005b) display reduced I_SC responses to cholinomimetics. In some instances, such as pathologies accompanied by severe ulceration, such hyporesponsiveness, could simply reflect a lack of intact epithelium, although this is not an adequate explanation for most intestinal pathologies. Indeed, there is significant evidence implicating several specific mechanisms in the cholinergic hyporesponsiveness associated with intestinal infection and inflammation. In some instances, this involves altered ACh metabolism; for example, increased AChE expression and activity have been noted in the jejunal epithelium of rats infected with Nippostrongylus brasiliensis (Russell et al., 2000) and in aganglionic segments of rectosigmoid colon from children with Hirschsprung's disease (characterized by a lack of development of cholinergic enteric nerves in the hindgut) (Hardy et al., 1993). Tri-nitrobenzene sulphonic acid (TNBS)-induced colonic inflammation and infection with Trichenella spiralis both affect ACh metabolism in smooth muscle preparations as well, where decreased packaging and exocytosis of ACh from the myenteric plexus is observed (Main et al., 1993; Davis et al., 1998; Poli et al., 2001). In contrast, others have found either no change or an increase in intestinal ChAT expression (Palmer and Koch, 1995; Green et al., 2004), whereas increased levels of non-neuronal ACh have been reported in other (non-intestinal) inflammatory diseases (Wessler et al., 2003). It is possible that this hyporesponsiveness represents a means to limit ongoing Cl⁻ secretion during disease, but the full significance of these data is not clear. Moreover, should some commonality be identified in the mechanism responsible for the epithelial hyporesponsiveness to cholinomimetics; this could represent a significant advance in the development of therapies to regulate enteric water balance.

We have reported that I_SC responses to CCh in colonic tissue from mice with dextran sodium-sulphate (DSS)-induced colitis are not only attenuated, but reversed in direction (Sayer et al., 2002). The CCh-induced decrease in I_SC was abolished by pretreatment with TTX or inhibitors of inducible nitric oxide (NO) synthase and partially reversed by pretreatment with HEX; overall, the findings were compatible with a novel means of controlling epithelial ion transport via activation of neuronal nAChRs in the myenteric plexus leading to NO liberation from glial cells (Green et al., 2004). Although NO release can explain the CCh-induced drop in I_SC, it does not account for the lack of a CCh-induced increase in I_SC in colonic tissue from DSS-treated mice. We speculate that this deficiency is due to loss of mAChRs or inflammation-induced changes in epithelial Ca²⁺-dependent secretory pathways, although proinflammatory cytokines such as interferon-γ (IFN-γ) and tumour necrosis factor-α (TNF-α) may actually increase expression of AChRs (Poea-Guyon et al., 2005) (see below).

The issue of direct regulation of cholinergic effects by cytokines has been taken up in vitro using CCh treatment of human colon-derived epithelial monolayers. Exposure to activated immune cells or recombinant IFN-γ results in diminished responses to CCh (Holmgren et al., 1989; Madara and Stafford, 1989; McKay et al., 1996b; Madsen et al., 1997; McKay and Singh, 1997) that have been linked to reduced expression of the Na⁺/K⁺-ATPase pump and the Na⁺/K⁺/2Cl⁻ co-transporter (Sugi et al., 2001; Bertelsen et al., 2004), whereas TNF-α treatment appears to enhance the effect of CCh (Oprins et al., 2000, 2002). Treatment of colonic epithelial cells with IFN-γ leads to an increased production of TGF-α, which, through activation of the EGFr, as mentioned earlier, has been shown to inhibit Ca²⁺-dependent Cl⁻ secretion (Uribe et al., 1996a, 2002). These studies only hint at the full influence inflammatory mediators undoubtly have on cholinergically mediated epithelial ion transport; the potential of other cytokines to alter cholinergic responses remains to be defined, and the number of cytokines/growth factors that occur constitutively or sporadically in the gut translates into a daunting myriad of possibilities. This area is especially fascinating in light of the reverse relationship, that is, the ability of the cholinergic pathway to alter cytokine responses. This relationship is designated as the cholinergic anti-inflammatory pathway, through which stimulation of the vagus nerve suppresses endotoxin-induced release of inflammatory mediators, including TNF-α, from immune cells, an effect that is mediated by α7 nAChRs (Pavlov et al., 2003; Wang et al., 2003).

In vitro infection with the bacterial pathogen enterohemorrhagic Escherichia coli reduces CCh-induced secretory responses in T84 cell monolayers (Hecht and Koutsouris, 1999). Intriguingly, application of fatty acids has been postulated as another means of limiting CCh-induced Cl⁻ secretion (Schultheiss et al., 2001) and the gut commensal flora is the predominant source of fatty acids. Additionally, both viral and helminthic infections have been shown to reduce CCh-evoked ΔI_SC (Argenzio et al., 1996; Reardon et al., 2001). Finally, and of major health concern, diarrhoea resulting from Vibrio cholerae infection is potentiated by cholinergically mediated secretory events (Banks et al., 2004) that can be mobilized by serotonin release from enterochromaffin cells (Lundgren, 1998), which, in turn, at least in the duodenum, evokes ACh release (Tuo and Isenberg, 2003). Infections can have long-lasting effects on the host (e.g. post-infectious irritable bowel syndrome) (Farthing, 2005), and so awareness of any cholinergic contribution to the altered intestinal fluid status elicited by infections (Jodal et al., 1993) can be of value in not only providing a means of symptom relief/disease management but also in the evaluation of disease aetiology. These isolated examples simply illustrate the general lack of data on the impact of one of the major endogenous homeostatic control systems (i.e. the cholinergic system) on the duration of infection (and vice versa) and the physiological perturbations that they evoke.

Conclusion

ACh is an important regulator of the electrogenic ion fluxes that govern directional water movements in the intestine. Identification of ACh receptor subtypes involved in this process has not always been straightforward and this is partially attributable to a lack of pharmacological agonists and antagonists with high specificity for receptor subtypes. For the most part, however, it is generally accepted that cholinergic control of intestinal ion transport is accomplished predominantly via activation of M₁ and M₃ AChR on neurons and enterocytes, respectively. These fundamental pharmacological data can now be confirmed or refuted and supplemented by a molecular knockout strategy that has already been applied to the mouse and has revealed some intriguing and unexpected data (Wess, 2004; Haberberger et al., 2006; Hirota and McKay, 2006). Also, data from human and mouse tissue is noticeably scant; a particular concern in the context of modelling and treating human disease. Moreover, it is clear that infection and inflammation result in an altered enteric cholinergic system, a generalized hyporesponsiveness of the epithelium to cholinergic stimulation and unmasking of regulatory pathways that are either not operational, or go undetected, when tissues from healthy individuals are assessed. Furthermore, there is intriguing, but fragmented, data on how disease can uncouple mAChRs from intracellular signalling cascades and we have only a rudimentary knowledge of how other neuroimmune (and stromal)-derived factors influence the overall activity of cholinergic signalling in the gut and produce changes in active epithelial ion transport. The cholinergic system is a principal player in enteric homeostasis, and we suggest that the outstanding challenges in this field lie in unraveling the communication within the neuronal circuitry and the role of specific AChRs therein, in defining the mechanism(s) by which immune responses (i.e. disease) modulate the cholinergic system, and in assessing cholinergic signalling in the setting of a dynamic tissue in which other signalling molecules are released, either tonically or sporadically.

Abbreviations

AChR

ACh receptor

ATR

atropine

BCh

bethanechol

CCh

carbachol

DMPP

dimethylphenylpiperazinium

EFS

electrical field stimulation

EGF

epidermal growth factor

HEX

hexamethonium

IFN-γ

interferon-γ

IHC

immunohistochemistry

mAChR

muscarinic acetylcholine receptor

MP

myenteric plexus

nAChR

nicotinic acetylcholine receptor

PD

potential difference

PG

prostaglandin

I_SC

short-circuit current

SI

small intestine

SMP

submucosal plexus

SP

substance P

TTX

tetrodotoxin

TNF

tumour necrosis factor

VIP

vasoactive intestinal polypeptide

VAChT

vesicular acetylcholine transporter

Conflict of Interests

The authors state no conflict of interests.