Address for correspondence, K. M. Sanders PhD, Department of Physiology and Cell Biology, University of Nevada School of Medicine, Reno, NV 89557, USA., Tel: (775) 784 6908; fax: (775) 784 6903; ude.adaven.enicidem@srednask
The publisher's final edited version of this article is available at Neurogastroenterol MotilSmooth muscle cells (SMC) make up the muscular portion of the gastrointestinal (GI) tract from the distal oesophagus to the internal anal sphincter. Coordinated contractions of these cells produce the motor patterns of GI motility. Considerable progress was made during the last 20 years to understand the basic mechanisms controlling excitation-contraction (E-C) coupling. The smooth muscle motor is now understood in great molecular detail, and much has been learned about the mechanisms that deliver and recover Ca 2+ during contractions. The majority of Ca 2+ that initiates contractions comes from the external solution and is supplied by voltage-dependent Ca 2+ channels (VDCC). VDCC are regulated largely by the effects of K + and non-selective cation conductances (NSCC) on cell membrane potential and excitability. Ca 2+ entry is supplemented by release of Ca 2+ from IP3 receptor-operated stores and by mechanisms that alter the sensitivity of the contractile apparatus to changes in cytoplasmic Ca 2+ . Molecular studies of the regulation of smooth muscle have been complicated by the plasticity of SMC and difficulties in culturing these cells without dramatic phenotypic changes. Major questions remain to be resolved regarding the details of E-C coupling in human GI smooth muscles. New discoveries regarding molecular expression that give GI smooth muscle their unique properties, the phenotypic changes that occur in SMC in GI motor disorders, tissue engineering approaches to repair or replace defective muscular regions, and molecular manipulations of GI smooth muscles in animals models and in cell culture will be topics for exciting investigations in the future.
Keywords: Ca 2+ sensitization, calcium channels, excitation-contraction coupling, IP3 receptor, potassium channels
The gut is a muscular tube that ingests food, assimilates nutrients and eliminates waste. Processing of nutrients requires movement of materials through and between organs, and this is accomplished by coordinated movements of the tunica muscularis. Muscle cells are the final effectors of force development and work; however, they receive regulatory input from several control systems, including interstitial cells of Cajal (ICC), motor neurones, hormones, paracrine substances and inflammatory mediators. At many times, but particularly during the postprandial period, these regulatory systems supply parallel inputs to gastrointestinal (GI) smooth muscles. Thus, the motor outputs of the gut are highly integrated responses. The most important issues for gastroenterology are the patterns of force development and the coordination of contractions between regions of gut. These are the main physiological responses (indeed the purpose!) of the tunica muscularis.
Smooth muscles of the GI tract exhibit autonomous behaviour, and there are intrinsic regulatory pathways in smooth muscle cells (SMC) that can amplify or defeat signalling from higher regulatory systems [e.g. enteric nervous system (ENS) or hormones]. For example, normal signals from the ENS or from digestive hormones may fail to produce desired contractile behaviours if excitation-contraction (E-C) coupling mechanisms in SMC are inactivated. Thus, it is important to understand the complex mechanisms that regulate E-C coupling in SMC. The intrinsic (i.e. non-neural, non-hormonal) regulation of gut motility has been referred to in the literature as myogenic, but it is now recognized that the traditional use of this term includes, besides smooth muscle mechanisms, the behaviour of ICC. ICC are electrically coupled to SMC, and affect resting membrane potentials of the smooth muscle/ICC syncytium and impose pacemaker activity, sensitivity to stretch and responsiveness to neurotransmitters 1 on the behaviour of SMC. For further discussion of ICC in the behaviour of the tunica muscularis, the reader is referred to Gianrico Farrugia in this edition of Neurogastroenterology and Motility. 2
In the past there was debate about the relative importance of myogenic vs neurogenic mechanisms in regulating GI motility patterns. Now we understand the fundamental importance of the ENS in generating the major motor patterns, particularly of the intestine and colon, and regulating the amplitudes of contractions, but we also recognize that muscle responses to neural inputs ultimately depend upon the relative excitability of the ICC-smooth muscle syncytium. Much has been learned about the excitability mechanisms in GI SMC during the past 2 decades, but major questions remain. Most information about GI muscles comes from studies of five, non-primate animal species, and we lack comparative information about the control of human smooth muscle excitability and contractility. There are dramatic differences in cellular phenotypes in different smooth muscles within a single species and across species. Thus, the excitability and contractile mechanisms of human GI smooth muscles must be studied in detail, and we must be efficient about this task as the availability of human tissues for in vitro investigation may be brief. Characterization of human GI muscles will be guided by the groundwork provided by studies of animal models.
Development and differentiation of SMC results in the transcriptional activation of genes that determine the structural and functional features of mature SMC. Several genes indicative of the SMC phenotype have been identified, including caldesmon, smooth muscle myosin heavy chain, γ-smooth muscle actin, calponin, SM22, α- and β-tropomyosins, and α1 integrin. Genes encoding these proteins are transcriptionally regulated, and the expression is controlled by binding of serum response factor (SRF). SRF, a serum response element that binds to a cis-regulatory element known as a CArG box (CCAAAAAAGG), is responsible for transcriptional activation of smooth muscle-specific genes. 3
Inactivation of SRF during development is fatal, making it difficult to assess the role of SRF in development and maintenance of SMC. 4 Recent studies have compensated for this problem by utilizing SRF conditional knock-outs and tamoxifen-inducible CRE recombinase expressed specifically in SMC. 5,6 Inactivation of SRF after birth causes severe dilation of the GI tract, impaired contraction of smooth muscles, defective peristalsis and death within 2 weeks of administration of tamoxifen. Colonic SMC from SRF-deficient mice grew in culture but displayed structural defects, decreased expression of smooth muscle-specific genes (e.g. smooth muscle actin, heavy chain of smooth muscle myosin and smoothelin), and increased senescence. 7 The technique of spatial and temporal activation of CRE has great potential for future studies of regulation of phenotype in smooth muscles.
Serum response factor is a ubiquitous transcription factor that is not unique to smooth muscles, so regulation of the smooth muscle phenotype cannot be explained by SRF–CArG interactions alone. Myocardin (Myocd), a transcriptional coactivator of SRF, may be a key regulator in SRF-dependent expression of smooth muscle genes. 7 Myocardin is expressed in embryonic and adult cardiac and smooth muscles. 7–9 Forced expression of myocardin in fibroblasts induces a smooth muscle-like phenotype. 10,11 Myocardin induces acetylation of nucleosomal histones near SRF-binding sites that control expression of smooth muscle-specific genes. 12
Although progress is being made to determine the factors controlling the development and maintenance of the SMC phenotype, this process is highly complex, and multiple mechanisms may contribute. For example, differentiation of stem cells into SMC-like derivatives can bypass myocardin, suggesting the existence of parallel regulatory pathways. 13 In addition to regulation of gene expression, smooth muscle-specific isoforms of several proteins, including caldesmon, smooth muscle myosin heavy chain, α-tropomyosin and vinculin/metavinculin are generated by alternative splicing in a smooth muscle-specific manner. 3 Knowledge about the regulation of the SMC phenotype comes from studies of the cardiovascular system. There are likely to be unique regulatory pathways controlling gene expression in GI muscles, and discovery and understanding of these pathways in development and pathophysiology are in need of further investigation.
Smooth muscle contraction is regulated by phosphorylation of the 20 kDa myosin light chain subunits of myosin (MLC20; Fig. 1 ). 14 This can occur by the Ca 2+ /calmodulin-dependent actions of myosin light chain kinase (MLCK) 15 or by the Ca 2+ -independent actions of several additional kinases, including Rho-kinase, integrin-linked kinase (ILK) and zipper-interacting protein kinase (ZIPK) (see Fig. 1 ). 16,17 Contraction is initiated primarily by activation of MLCK under physiological conditions, and thus increased cytoplasmic Ca 2+ [Ca 2+ ]i is the primary driving force for contraction, as in other muscles. Phosphorylation of MLC20 facilitates actin binding to myosin and initiates cross-bridge cycling.
Schematic representation of major pathways involved in excitation-contraction coupling in gastrointestinal (GI) smooth muscles (Drawing inspired from organization of molecular details in Ref. 94). Most of the Ca 2+ required for activation of the contractile apparatus enters cells via VDCC. The major species of VDCC is CaV1.2 channels, but other VDCC are present to a varying extent in GI muscle cells. The open probability of VDCC is regulated negatively by a variety of species of K + channels expressed by smooth muscle cells. K + channels are activated by a number of inhibitory agonists (see text for details). The VDCC are regulated positively by receptor-operated (ROC) and stretch-activated (SAC), non-selective cation channels that depolarize smooth muscle cells. ROCs and SACs can also contribute varying amounts to Ca 2+ entry, depending upon the molecular species and relative permeability to Na + and Ca 2+ . Ca 2+ entry can also be supplemented to a varying extent by release of Ca 2+ from IP3 receptor-operated Ca 2+ channels in the sarcoplasmic reticulum (SR) membrane. IP3 levels are enhanced by synthesis stimulated by agonists binding to G-protein-coupled receptors coupled through Gαq/11 and activation of phospholipase Cβ (PLCβ). A rise in cytoplasmic Ca 2+ ([Ca 2+ ]i) binds to the Ca 2+ binding protein, calmodulin, and activates MLCK. Myosin light chain kinase phosphorylates the 20 kDa light chain of myosin (MLC20) and facilitates cross-bridge cycling. There are also Ca 2+ -independent kinases present in smooth muscles that can activate MLCK in a Ca 2+ -independent manner, but under physiological circumstances, excitation-contraction coupling is largely due to the Ca 2+ -dependent pathway. Phosphorylation of MLC20 is balanced by MLCP. Dephosphorylation of MLC20 reduces cross-bridge cycling and leads to muscle relaxation. The activity of MLCP is regulated by pathways that in effect regulate the Ca 2+ sensitivity of the contractile apparatus. One of these pathways is regulated by G protein regulation of GDP-GTP exchange factor (Rho-GEF), RhoA and activation of RhoK (representing ROCK1 and ROCK2 isoforms). Rho kinase and protein kinase C (PKC) can phosphorylate PKC-potentiated inhibitory protein (CPI-17) at T38 which in turn inhibits the catalytic subunit of MLCP (PP1c). Rho kinase can also phosphorylate the regulatory subunit of MLCP (MYPT) at T696 and T853. Phosphorylation of MYPT decreases the activity of MLCP and its ability to target dephosphorylation of MLC20. Other kinases including zipper-interacting kinase (ZIPK) can also phosphorylate CPI-17 and MYPT. Red arrows depict pathways leading to enhanced contraction; blue arrows indicate pathways leading to reduced contraction.
Phosphorylation of MLC20 and the contractile state is balanced by the actions myosin light chain phosphatase (MLCP; Fig. 1 ). MLCP is composed of three subunits. 18 One subunit anchors MLCP to phosphorylated MLC20, thus targeting a second subunit, the 37 kDa catalytic subunit (type 1 serine/threonine phosphatase, PP1c), to myosin. The function of the third 20 kDa subunit is unknown. The targeting subunit is called the myosin phosphatase target subunit (MYPT) and several isoforms exist. MYPT1 is the major subunit expressed in smooth muscles. MYPT also regulates the enzymatic activity of MLCP.
Patterned firing of excitatory and inhibitory motor neurones varies the force of contractions and determines the motor patterns of the GI tract; however, neural control is superimposed upon myogenic mechanisms that regulate the excitability and contractility of SMC. Electrophysiological properties of the SMC/ICC syncytium largely control the access of Ca 2+ to the contractile proteins of SMC. Electrical mechanisms also provide an important level of coordination that contributes to the motor patterns of GI organs. Features of myogenic electrical coordination include: setting of membrane potentials and SMC excitability, intrinsic pacemaker activity, initiation of Ca 2+ action potentials and control of Ca 2+ entry, and regenerative propagation of excitable events linking the excitability of large groups of SMC.
Smooth muscle cells have negative resting membrane potentials determined by the relative permeabilities of the plasma membrane to physiological ionic species. Smooth muscle cells demonstrate dominant permeability to K + ions, but significant contributions from non-selective cation conductances (NSCC) decrease membrane potential substantially from the equilibrium potential of K + ions (approximately −90 mV). The sodium pump, which is electrogenic, also contributes several mV to resting potential. Regulation of resting membrane potential is an important aspect controlling excitability of SMC. Resting membrane potentials of GI SMC in different regions of the gut vary widely (e.g. from −85 to −40 mV). The wide range in resting potentials is due to the differential expression of ion channels by SMC and differences in the relative open probabilities of ion channels. Regulation of membrane potential preconditions responses of SMC to depolarizations from pacemaker cells and agonists. There is an incomplete picture of the factors involved in generating resting potentials in most GI SMC, and we know little about how this important determinant of excitability is regulated by hormones or adversely affected by paracrine substances or inflammatory mediators in pathophysiological states.
There are three main patterns of electrical activity in GI muscles. In the fundus and some sphincters, electrical activity of most species appears to be characterized by slow tonic changes in membrane potential. 19 Membrane potentials are relatively depolarized in these muscles and lie within a range where a sustained, low level of activation of voltage-dependent Ca 2+ channels (VDCC) results in continuous influx of Ca 2+ (see below). Tone in other regions, such as the band-like longitudinal muscles of the colon and caecum and some sphincters is due to firing of Ca 2+ action potentials and a tetanus-like elevation in cytoplasmic Ca 2+ . Most other regions of the GI tract have phasic electrical activity due to the pacemaker activity of ICC (see Ref. 1). ICC generate pacemaker activity, in the form of electrical slow waves, and this activity propagates actively through networks of ICC. 1 Slow waves conduct passively into SMC where depolarization increases the open probability of VDCC. This can result in the generation of Ca 2+ action potentials. Action potentials do not occur between slow waves, when the open probability for Ca 2+ channels is low. As Ca 2+ influx is the main impetus for E-C coupling in GI smooth muscles, contractions occur in response to the periods of enhanced Ca 2+ channel open probability and muscles relax between slow waves. This activity naturally organizes the pattern of contractions into phasic contractions timed by the slow waves and forms the basis for segmental contractions and gastric peristalsis.
Confusion about the relative role of electrical slow waves and Ca 2+ action potentials (or spikes) in generation of contractile activity in SMC goes back to whole animal studies where strain gauges were used to measure contractions in organs. This method was not necessarily sensitive enough to record small contractile events, and the recordings were prone to movement artefacts. Recordings from this era led to terminology suggesting that contractions did not occur in response to slow waves, which were called electrical control activity. Excitation-contraction coupling occurred only when Ca 2+ action potentials (electrical response activity) occurred. 20 This terminology does not accurately describe E-C coupling in GI muscles and has fallen from usage. In vitro recordings show that basal slow wave activity causes E-C coupling, because slow waves depolarize SMC to potentials where the open probability of VDCC is increased and Ca 2+ entry via these channels activates the contractile apparatus. 21,22 In the small bowel and colon depolarization from slow waves reaches the action potential threshold in SMC. Neural and hormonal control in these organs modulates membrane potential of the ICC/smooth muscle syncytium upon which slow waves are superimposed. Activation of inward currents by excitatory agonists depolarizes membrane potential and increases the peak depolarization during slow waves. Enhanced depolarization increases activation of VDCC and generation of Ca 2+ action potentials. Ca 2+ influx during action potentials initiates large amplitude contractions, and propulsive contractions in the small and large intestines. Activation of outward currents, or suppression of tonic inward currents, stabilizes membrane potential or causes hyperpolarization. Slow waves during periods of hyperpolarization achieve lower levels of depolarization, fewer Ca 2+ action potentials, less Ca 2+ entry and less forceful contractions.
The ICC and SMC are coupled electrically, forming a multicellular syncytium. Activation of depolarizing or hyperpolarizing ionic conductances in either cell type affects the total input resistance and excitability of the syncytium. For example, activation of K + channels in ICC reduces excitability of coupled SMC and reduces the likelihood of reaching the action potential threshold. This is an important concept because much of the excitatory and inhibitory neural innervation of GI muscles occurs via ICC (e.g. 23 , 24 ). Ultimately, the integrated preconditioning of the excitability apparatus in SMC (i.e. the degree to which VDCC are activated during slow waves and whether threshold for Ca 2+ action potentials is reached) determines contractile force and patterns in response to inputs from the ENS. Responses to other stimuli, such as hormones and paracrine substances, are likely to target both ICC and SMC, depending upon the expression of appropriate receptors and second messenger pathways.
Electrical activity in the body and antrum of the stomach is different from the small bowel and colon. In the stomach, modulation of the amplitude of slow waves explains the range of contractile amplitudes 25 because Ca 2+ reaches significant levels during slow waves and couples these events to contraction. 21,22 Recordings from several species, including humans, have shown that excitatory agonists increase the amplitude of slow waves in the stomach and the long durations of gastric slow waves cause sustained Ca 2+ entry. 21 Inhibitory agonists have the opposite effect, reducing slow wave amplitude, reducing the period and magnitude of Ca 2+ entry and reducing the force of contractions. 26
There are important regional differences in ion channel expression in SMC and differences between SMC and ICC that explain the diversity of electrical behaviours in gut smooth muscles. 27 For example, the broad range in resting potentials and electrical patterns of GI muscles is partly a function of the variable expression of K + channels in SMC; however, the expression of a large variety of non-selective cation channels is also likely to contribute to the diversity in electrical activities. At least 20 species of K + channels are expressed by SMC of the GI tract (see table 2 in Ref. 27), and more are likely to be described. A complete description of K + channel expression and function in human GI SMC has not been accomplished. Because there are significant species variations noted in animal studies, it will be imperative to perform detailed experiments on human GI SMC to understand excitability mechanisms in the human GI tract.
Smooth muscle cells from possibly every region of the GI tract and in every species studied express VDCC, delayed rectifier K + channels and large conductance Ca 2+ -activated K + (BK) channels. These conductances are also present in the human oesophagus, jejunum and colon (cf. Refs 28 – 33 ). Species and organto-organ variations in channel family members and current density (resulting from channel copy numbers) have been reported, but it is difficult to make highly quantitative comparisons between studies because of differences in cell preparations, regions and muscle layers from which cells are obtained, and experimental conditions (e.g. dialysed vs perforated patch recordings; physiological vs isometric ionic gradients, temperature etc.). All GI SMC express voltage-dependent, dihydropyridine-sensitive Ca 2+ channels (CaV1.2; 31 , 34 ), and these channels are the backbone of E-C coupling in the gut. Treatment with dihydropyridines greatly reduces muscle contraction, whether spontaneous, agonist activated or neurally regulated ( Fig. 2 ). Thus, depolarization and activation of Ca 2+ entry via Ca 2+ channels is a fundamental mechanism in the physiological activation of GI muscles. At the depolarized potentials of some GI SMC, continuous, low probability opening of Ca 2+ channels (window current) occurs, producing continuous low level influx of Ca 2+ . This Ca 2+ influx via VDCC contributes to tone and possibly other Ca 2+ sensitive cell signalling pathways.
Contraction of circular and longitudinal gastrointestinal (GI) muscles depends upon Ca 2+ entry through VDCC. These traces show contractile responses of the canine gastric antrum, as an example, to carbachol (CCh, 10 −7 mol L −1 ; black bar indicates addition in each panel). Addition of CCh enhanced the amplitude of phasic contraction. Application of CCh within a few minutes after reducing extracellular Ca 2+ (to 0.1 mmol L −1 ) or after addition of nicardipine (10 −6 mol L −1 ) caused much smaller contractile responses. The presence of a phasic contractile response in the presence of nicardipine, albeit of greatly reduced amplitude, probably indicates contributions of other VDCC or ROCs to Ca 2+ entry.
While CaV1.2 channels are extremely important in E-C coupling, GI SMC express additional types of VDCC. For example, T-like currents (CaV3 channels) have been reported in GI SMC, 33 and these could be a factor in delivery of Ca 2+ to the contractile apparatus in some cells. We have also found CaV1.3 channel expression in GI muscles (S. Ro and K.M. Sanders, unpublished observations), and these channels, while still L-type, are relatively resistant to dihydropyridine block. The expression of additional VDCC means that the effectiveness of dihydropyridines in blocking contractions varies between muscles.
There is lingering controversy in the literature about the role of Ca 2+ channels and voltage-dependent Ca 2+ entry in GI smooth muscle contractions. For example, in a recent review, the following statement was made:
‘The mechanisms of Ca 2+ mobilization in smooth muscle cells of the circular and longitudinal muscle layers vary. In muscle cells from both layers, G protein-coupled agonists initiate contraction by increasing cytosolic Ca 2+ , or [Ca 2+ ]i. Initial contraction of smooth muscle cells and the increase in [Ca 2+ ]i are not affected by Ca 2+ channel blockers or by withdrawal of extracellular Ca 2+ in circular smooth muscle cells but are abolished in longitudinal smooth muscle cells. 35
This idea seems to have developed from shortening studies of isolated, unloaded SMC in which voltage-dependent Ca 2+ currents were not measured, 36 and it is possible that CaV1.2 channels in these cells were inactivated by depolarization or damage during enzymatic dispersion. Ca 2+ release from cellular stores or Ca 2+ sensitization mechanisms cannot account for the level of cell-to-cell coordination that results from regenerative propagation of electrical activity over many centimetres of tissue. In fact, VDCC are expressed by both circular and longitudinal muscle cells (see Ref. 27), and it is clear from contractile studies that spontaneous contractions and responses to agonists and neurotransmitters of both muscle layers are greatly and rapidly reduced (i.e. this effect is not due to washing Ca 2+ from internal stores) by dihydropyridines or by reduced extracellular Ca + (e.g. Fig. 2 ). More detailed studies, in which Ca 2+ currents and intracellular Ca 2+ transients have been measured, demonstrate voltage-dependent Ca 2+ influx in GI circular SMC that is sufficient to account for the rise in Ca 2+ in response to depolarization. 22,37 Studies using conditional inactivation of the CaV1.2 gene confirm the importance of Ca 2+ entry in contractions of GI muscles. After inactivation of CaV1.2, mice showed reduced excretion of faeces, loss of rhythmic contractions and paralytic ileus. 38 Thus CaV1.2 channels are essential for spontaneous and evoked E-C coupling in GI muscles.
Many K + channels are also expressed in GI SMC including, delayed rectifier K + channels (KV1 family channels, usually KV1.5 and KV2 family channels); 39–42 large-conductance Ca 2+ -activated K + channels (BK), 43,44 small conductance Ca 2+ -activated K + channels (SK) 45,46 and ATP-dependent K + channels (KATP) 47 . All these channels are regulated by physiological agonists such as those coupled by cAMP-mediated effects (BK and KATP) or release/entry of Ca 2+ (BK and SK). For example, SK channels contribute to the tonic inhibition of GI muscles because of sustained release of inhibitory neurotransmitters. 48 Apamin, a toxin that inhibits SK channels, depolarizes and increases the excitability of many GI muscles. 49
Additional species of K + channels are present in GI SMC, including. two-pore K + channels, 50–52 inward rectifiers (including KATP), 53,54 ether-a-go-go related gene channels, 55,56 M-current channels, 57,58 MinK channels, 58 and intermediate conductance Ca 2+ -activated K + channels. 45 Some of these channels, such as TREK channels of the two-pore K + family, are likely to mediate responses to neurotransmitters and other biological agonists. 59,60 The broader importance of these conductances in GI motility, particularly in humans, is not fully appreciated.
There are also a variety of ion channels in SMC that are activated by binding of agonists and by physical stimuli such as stretch. These include, at a minimum, the NSCC activated by muscarinic stimulation (i.e. IACh), 61–65 a variety of peptide and amine agonists (see Refs. 66 – 68 ) and purine receptors (i.e. members of the P2X family of ion channels). 69 Activation of these channels provides excitatory input to SMC mainly by producing inward current and depolarization. The direct contribution of receptor-operated ion channels or stretch-activated cation channels to Ca 2+ influx varies depending upon the relative permeabilities of these channels to Na + and Ca 2+ . This topic has not been studied adequately to ascribe relative contributions of these channels to Ca 2+ entry during agonist responses.
Logic would suggest that excitatory agonists might activate CaV1.2 channels as this is the main pathway for E-C coupling, but regulation of CaV1.2 currents is circuitous. For example, in human oesophageal SMC, acetylcholine (ACh) is a potent stimulant of contractions and responses to ACh are inhibited 90% by nifedipine, but ACh inhibits CaV1.2 currents. 31 Thus, the contractile response to ACh is largely due to Ca 2+ entry via CaV1.2 channels, but not due to direct activation of Ca 2+ channels by muscarinic agonists. Equally contradictory regulation occurs with some inhibitory substances. Agonists coupled through GaS, production of cAMP and activation of cAMP-dependent protein kinase are inhibitory to GI contractions, but this pathway stimulates CaV1.2 currents, 70 as in the heart (cf. Ref. 71). These apparently contradictory observations are explained by the fact that CaV1.2 channels in GI SMC are mainly regulated by membrane potential: excitatory agonist activation of NSCC and suppression of K + conductances leads to depolarization and activation of CaV1.2 current (and Ca 2+ influx); inhibitory substances, linked to elevations in cAMP or cGMP or enhancement of localized Ca 2+ release events (Ca 2+ sparks or puffs, see below), are coupled to activation of K + channels. As above, activation of K + channels hyperpolarizes membrane potential and impedes opening of CaV1.2 Ca 2+ channels.
Co-expression of many ion channels in GI muscles and the heart or other organs creates problems with exploiting channel-active substances for therapeutic regulation of GI motility. As in the dramatic case of cisapride, a drug with unacceptable cardiac side effects, many small soluble molecules with motility effects due to effects on ion channels, are doomed as therapeutic agents because of cardiac side effects. The function of ion channels is significantly affected by accessory subunits and interactions with regulatory proteins that often have greater tissue or cell-specific expression patterns, so rather than designing drugs that block pore-forming regions of ion channels (that are highly conserved in ion channel families), it might be possible to manipulate GI motor function with compounds that affect interactions with subunits or cytoskeletal proteins.
Ca 2+ entry from the extracellular fluid can be supplemented under various circumstances by release of Ca 2+ from intracellular stores ( Fig. 1 ). Paradoxically, Ca 2+ release can lead either to stimulation of the contractile apparatus or activation of K + channels which hyperpolarizes cells and reduces Ca 2+ entry, as described above. Ca 2+ stores are provided by the sarcoplasmic reticulum (SR) and mitochondria in smooth mguscles. 72–75
There are two major mechanisms for release of Ca 2+ from the SR, ryanodine receptors (RyR) and inositol (1,4,5)-trisphosphate (IP3) receptor-operated Ca 2+ channels (IP3R). Measurements of IP3R and RyR-binding sites in GI muscle cells suggest that the ratio of these channels is 10:1 (i.e. IP3R:RyR), 76 and studies of permeabilized SMC have shown that some IP3 receptor-operated stores may be entirely devoid of RyR. 77,78 Ca 2+ is a natural ligand for RyR; however, Ca 2+ entry does not result in significant Ca 2+ -induced Ca 2+ release in GI SMC. This may be due to the fact that the Ca 2+ sensitivity of RyR in GI smooth muscles is too low to play a physiological role in Ca 2+ mobility. 78 Cyclic ADP ribose (cADPR) is another ligand for RyR; however, the literature on this mechanism is controversial. Some investigators have reported that cADPR facilitates Ca 2+ -induced Ca 2+ release in GI SMC (e.g. Ref. 79), but other investigators have rejected the idea that cADPR is important as a means of releasing Ca 2+ from RyR in SMC. For example, [Ca 2+ ]i was unchanged when cADPR was liberated from caged cADPR in voltage-clamped colonic myocytes. However, caffeine induced substantial Ca 2+ transients in these cells, indicating the integrity of RYR-activated stores. 80
The release of Ca 2+ from IP3R is regulated by the production of IP3 resulting from agonist binding to a variety of G-protein-coupled receptors expressed by GI SMC. Cytoplasmic Ca 2+ ([Ca 2+ ]i) also affects the open probability of IP3R, and this relationship is bell shaped such that low concentrations of Ca 2+ facilitate channel opening, but elevated [Ca 2+ ]i becomes inhibitory. 81–83 Thus, there is a natural feedback mechanism controlling Ca 2+ release from IP3 receptors.
Peripheral SR is closely associated with the plasma membrane in SMC. Gaps of only 10 nm occur between these structures. 84 Close coupling between SR and the plasma membrane defines a microdomain (PM) that can display Ca 2+ dynamics that are distinct from changes in cytoplasmic Ca 2+ (i.e. [Ca 2+ ]PM is not equivalent to [Ca 2+ ]i). The very small volumes of these microdomains mean that small numbers of Ca 2+ ions can have large effects of [Ca 2+ ]PM. Changes in [Ca 2+ ]PM are extremely difficult to measure because the events are fast and the small spaces are below levels resolvable by fluorescent Ca 2+ indicators that distribute equally through the cytoplasm. Attempts to measure [Ca 2+ ]PM have suggested that Ca 2+ may rise to >10 μmol L −1 in response to Ca 2+ entry, 85,86 or Ca 2+ release from the SR. 87,88
Ca 2+ -sensitive ion channels are present in the plasma membranes of PM, and Ca 2+ release from the SR can regulate the open probabilities of these channels. Ca 2+ release events occur from either IP3 receptor-operated stores (Ca 2+ puffs), 89 RyR (Ca 2+ sparks), 90 or from both types of receptors due to Ca 2+ -sensitive properties of the channels. Ca 2+ release events can also develop, via the Ca 2+ -induced, Ca 2+ release properties of SR Ca 2+ channels, into Ca 2+ waves that travel along the interior aspect of the plasma membrane. Ca 2+ waves, which are common in many SMC, occur without contraction, so they must be due to localized increase in [Ca 2+ ]PM rather than a generalized increase in [Ca 2+ ]i. Ca 2+ puffs and sparks initiate spontaneous transient outward currents in GI smooth muscles (STOCs), 91,92 affecting membrane potential and excitability. Both SK and BK channels contribute to STOCs (e.g. Ref. 89). The role of Ca 2+ sparks, puffs and waves in human GI muscles has not been characterized.
Studies with Ca 2+ -sensitive indicators have shown that different agonists evoke markedly different contractile responses with equivalent changes in [Ca 2+ ]i. 93 Thus, the Ca 2+ –force relationship shifts right or left depending upon the stimulus. For example, a given increase in [Ca 2+ ]i in response to depolarization with elevated external K + causes less forceful contractions than an equivalent change in [Ca 2+ ]i caused by muscarinic stimulation. This shows that the Ca 2+ sensitivity of the contractile apparatus is regulated in SMC, and the basis for this phenomenon has been investigated heavily (for a review, see Ref. 14). Ca 2+ sensitivity can be increased by excitatory agonists causing more forceful contractions as a function of [Ca 2+ ]i or decreased by inhibitory agonists to mute the effectiveness of Ca 2+ transients. Here we discuss mechanisms to increase sensitivity; a more thorough review of this topic can be found in Somlyo and Somlyo. 14
Ca 2+ sensitivity is regulated by changing the activity of MLCP ( Fig. 1 ). If the activity of MLCP is reduced, then the regulatory subunit of myosin (MLC20) will tend to remain phosphory-lated and this will increase cross-bridge cycling and force generation. Phosphorylation of MYPT1, the regulatory subunit of MLCP, decreases the ability of MLCP to dephosphorylate MLC20 and increases the Ca 2+ sensitivity of the contractile apparatus. 94 MYPT1 can be phosphorylated at several sites including T696 and T853. Phosphorylation at either site inhibits the ability of MLCP to dephosphorylate MLC20. Rho kinase (RhoK) was the first kinase identified to phosphorylate MYPT1, but other kinases can also phosphorylate MYPT1 ( Fig. 1 ).
Another mechanism of regulation of MLCP is through CPI-17 ( Fig. 1 ), a small (17 kDa), PKC-potentiated inhibitory protein of protein phosphatase (PP1). CPI-17 inhibits the catalytic subunit of MYCP, and does not act via MYPT1. Phosphorylation of CPI-17 at T38 activates its inhibitory function. 95 In addition to PKC, the same kinases that phosphorylate MYPT1 (e.g. PKC, RhoK, ZIPK and ILK) can also phosphorylate CPI17. Expression levels of RhoK, MYPT1 and CPI-17 are important factors in determining the importance of these pathways in a given type of smooth muscle. MYPT1 is ubiquitously expressed, but levels of RhoK and CPI-17 vary in different muscles. Several studies have assessed the role of Ca 2+ sensitization in GI smooth muscles by assaying levels of RhoK, CPI-17, phosphorylated CPI-17 and phosphorylated MYPT1 (cf. Refs 96 , 97 ), and there are interesting links emerging between pathophysiological responses of smooth muscles and expression of signalling molecules involved in Ca 2+ sensitization (cf. Refs 98 – 100 ).
The extent to which Ca 2+ sensitization mechanisms are activated in SMC by neurotransmitter substances released from enteric motor neurones is an unresolved question. Smooth muscle cells express muscarinic receptors and intracellular signalling pathways linked to activation of receptor-operated cation channels and Ca 2+ sensitization mechanisms, but little ACh released by enteric motor neurones appears to reach SMC. In Ca 2+ sensitization studies, muscarinic agonists have been added to solutions bathing muscles. This method of delivering agonists may result in activation of a totally different population of receptors and signalling pathways than release of neurotransmitter from nerve terminals. This is because enteric motor neurones rather selectively innervate intermuscular ICC, and rarely make synaptic contacts with SMC. 23,101–103 Loss of ICC blocks nearly all cholinergic postjunctional responses. 23 It appears that rapid breakdown by acet-ylcholinesterases prevents ACh released from motor nerve terminals to spill out of synaptic clefts onto SMC.
There is currently a strong desire to apply techniques to manipulate gene expression to studies of GI SMC because of the experimental power of this approach. Unfortunately, gene manipulation in SMC has been relatively difficult because of the properties and plasticity of these cells.
At the start of molecular studies, it would seem important to establish the baseline molecular phenotypes of physiological GI SMC. Attempts to determine the unique molecular features of SMC have been confounded by the fact that gene and protein expression studies have mainly been performed on smooth muscle tissues, not on SMC. There are many cell types present in the tunica muscularis, and it is risky to make conclusions about specific gene expression or changes in gene expression from analyses of whole tissue homogenates or cell cultures from heterogeneous cell populations. Basic questions, such as what differences explain the unique characteristics of circular and longitudinal muscle cells and what is the basis for differences in behaviour of tonic and phasic GI muscles have not been answered. There are no suitable baseline standards for evaluation of the changes in smooth muscle phenotype that might occur in pathophysiological responses such as hypertrophy, hyperplasia, atrophy or inflammation. Studies of this type have concentrated on only a few genes at a time. Therefore, an important goal for the next several years will be to develop large-scale cell isolation and purification techniques to refine molecular analyses of SMC and to determine the molecular basis for differences in GI muscles (e.g. regional differences, tonic vs phasic muscles, etc.) and consequences of GI diseases on SMC.
Cellular localizations of some proteins and dramatic changes in protein expression have been evaluated by immunohistochemical techniques. Automated proteomic techniques, however, have largely not come of age for studies of GI SMC. Again, the main problem with using this technology is the purity of the starting material. One might argue that as SMC are the major source of protein in the tunica muscularis, protein analysis can be performed on whole tissue homogenates. This might be true for studies of abundant proteins, but evaluating interesting proteins with low copy numbers is problematic if the cellular source of the protein is not certified.
There seems to be a growing tendency to ignore the function of SMC in many molecular studies of signalling pathways. One should be cautious of studies conducted on primary cultures and passaged SMC. Smooth muscle tissues are mixed populations of cells that define their phenotypes in relation to diverse and highly specialized molecular microenvironments. Reconstitution of the microenvironment has been largely disappointing in culture conditions. Smooth muscle cells rapidly redifferentiate in culture, entering a so-called proliferative or synthetic state characterized by reduced expression of contractile proteins, changes in cytoskeletal architecture, loss of dense bodies and disappearance of excitability mechanisms. One questions whether cells so radically altered should be called SMC and whether the signalling pathways studied in these cells are relevant in physiological SMC. Smooth muscle disorders in other organs, such as asthma and atherosclerosis, might result from phenotypic transformations of SMC. 104 Thus, it is important to study the mechanisms responsible for phenotypic transformation. The extent to which phenotypic transformation contributes to the pathophysiology of GI motility disorders is not well understood, but inflammatory mediators, ageing and other factors have been shown to modify SMC phenotype in the gut. 105 It is, however, totally unresolved as to whether the phenotypic changes observed in pathophysiological conditions are recapitulated in cultures of SMC. Frankly, the use of cultured SMC at the present time appears to have more to do with obtaining a cellular preparation upon which molecular manipulations can be performed than about studying the mechanisms of smooth muscle function.
Some investigators have utilized serum-free medium in an attempt to reduce of SMC dedifferentiation in culture. A rigorous description of the fidelity of the SMC phenotype in these cultures has not been provided. Typically, authors provide evidence that smooth muscle actin or another smooth muscle-specific protein continues to be expressed in cultures as evidence of fidelity. However, more work needs to be done to characterize the molecular profiles of cultured cells. The power of modern analytical techniques (e.g. gene arrays, proteomics, etc.) has not been sufficiently trained upon this question. It would be useful to rigorously compare the expression profiles of native smooth muscle tissues with mixed cultures or purified SMC with cell cultures derived from enriched SMC. The results, at best, would verify the usefulness of cultured cells as a model of smooth muscle pathways and function. At worst, such analysis would reveal that the phenotype of cultured SMC differs radically from native cells; changes may occur in the expression of thousands of genes.
Rapid progress is being made using tissue engineering approaches. In these experiments SMC are grown on specialized matrices that better mimic the microenvironments of SMC in vivo. 106 These techniques are promising and might eventually lead to functional smooth muscle cultures or tissues that preserve functional elements of the native phenotype (e.g. Ref. 107).
A specialized gene array for identifying the smooth muscle phenotype would be a useful tool for assaying relative changes in key smooth muscle genes in cultured cells and pathophysiological models. At a minimum, such an array should include major genes known to be involved in various aspects of smooth muscle function and characteristic genes of contaminating cell types that should be assayed to determine the relative purity of cell preparations or cultures (e.g. Ref. 108). A smooth muscle phenotype array should include, at a minimum: known plasma membrane genes, including ion channels, transporters and receptors, caveolar genes, SR transporters, Ca 2+ sequestration, SR ion channels, major intracellular signalling molecules, and smooth muscle-specific components of the cytoskeleton and contractile apparatus. Secondly, a collection of genes for determining the relative contamination from other cell types should be included in any analysis. There are a variety of cell-specific genes that can identify whether a given population of cells has increased or decreased relative to other cells. Providing more rigorous background information about the relative changes in the SMC phenotype should become standard practice in studies of cultured cells and pathophysiological models and should be insisted upon in reports of these studies.
Several studies have used antisense or siRNA techniques to manipulate gene expression in GI SMC. These techniques are difficult to apply to intact tissues or organ culture preparations because penetration of plasmids is generally insufficient, although some new studies have reported relatively high efficiency in reducing gene expression (cf. Ref. 100). Because of problems in applying these techniques to intact muscles, most investigations have used cultured cells for studies involving acute knock-down techniques. 109,110
Transgenic approaches remain the best means of manipulating genes in SMC. The major problems with this approach are the time and monetary investment to create transgenic animals, and these issues are compounded by the possibilities that: (i) a given mutation will be fatal before animals are born or reach an appropriate age for study, or (ii) a mutation will not result in a demonstrable phenotype. The latter does not necessarily disprove a hypothesis as altered phenotypes can result from developmental compensation for inactivated or overexpressed genes. Fortunately, the technology and cost-effectiveness of transgenic techniques are constantly improving. Techniques, such as manipulation of transgenes bordered by loxP sites with CRE recombinase, make it possible to reduce lethality and other undesirable effects common with universal inactivation or overexpression of genes. Cell-specific transgenics, however, still suffer from issues related to compensation as genes targeted with loxP are affected as soon as CRE recombinase is expressed during development (E10–E15 for SMC-specific genes). Secondly, there are no completely unique genes for SMC that have been identified, except possibly the heavy chain of smooth muscle myosin. Virtually, all other smooth muscle-specific genes, including smooth muscle actin, SM22α, smoothelin, caldesmon, calponin and metavinculin, are transiently expressed in other cell types at various stages of development. 111 Thus, CRE recombinase driven by most promoters is likely to be expressed in cells other than SMC, and whole animal phenotypes might therefore be contaminated by defects in addition to SMC.
Ideally, gene targeting should be cell selective and timed such that the targeted gene can be activated or inactivated at any point during an animals life. To accomplish this, chimeric proteins have been used in which CRE recombinase is fused to modified hormone binding domains (HBD) of oestrogen or progesterone receptors. These proteins regulate the expression of CRE, which is silent until a ligand (such as tamoxifen for oestrogen receptors or RU-486 for progesterone receptors) binds to the receptor portion of the fusion protein. Use of this approach requires crossing mice with the cell-specific CRE recombinase-HBD fusion protein and mice with a gene targeted with two loxP sites arranged to either inactivate or activate the gene. Although reports have appeared using tamoxifen-inducible CRE to inactivate genes in GI SMC (see Refs 5 , 6 , 38 , 112 ), others have expressed difficulty obtaining sufficient CRE activity to inactivate genes targeted with loxP (see Ref. 113; J. Miano, personal communication). Thus, temporal control of transgenes has not been ideal in SMC in spite of utilization of specific smooth muscle promoters to drive CRE expression (e.g. SM22α; see Ref. 112). Differences in recombination efficiency might also be due to the fact that different floxed genes have different sensitivities to CRE recombinase. This might be due to the relative chromatin state of a particular gene (open/active state vs closed/ condensed state; see Ref. 113).
Another problem with conditional SMC knock-outs stems from the plasticity of these cells. Changes in phenotype during pathophysiological conditions can be accompanied by down-regulation of SMC-specific genes. 114 Thus, cell-specific promoters chosen to drive CRE expression in SMC may be silenced by the processes leading to a change in phenotype. Identification and application of efficient and stable promoters that target conditional expression of CRE recombinase in SMC, or other techniques to produce reliable gene inactivation in a cell- and time-specific manner would be a very important advance for smooth muscle biology.
Several issues are in need of experimental attention to increase our understanding of GI smooth muscle function and the processes resulting in human GI motility diseases. As described, a significant problem facing smooth muscle biologists is the ability to take full advantage of modern molecular techniques to study SMC function. There is a need for animals, tissues and cells in which gene expression can be reliably manipulated while these preparations retain the major functions and phenotype of SMC and tissues. Considerable work is needed to characterize the expression of genes and proteins in SMC and to determine the features which give muscles of the GI tract their unique properties and regionally different behaviours. We also need this information to provide a baseline for studies in which the phenotype of SMC changes during pathological conditions.
We need to determine which mechanisms worked out in animal models are involved in E-C coupling in human GI smooth muscles. Species differences, abundantly apparent from studies of animal models, suggest that the mechanisms responsible for E-C coupling in human muscles might be quite different from the mechanisms characterized in laboratory animals. Work on this topic should progress soon and rapidly while human muscles are still available from surgeries for cancer and other GI diseases.
The relative contributions of the many signalling pathways involved in regulation of contraction need to be evaluated in terms of physiological bioavailability of agonists. For example, studies of Ca 2+ sensitization mechanisms have used application of neurotransmitters to solutions bathing cells or muscles. Now we know that neurotransmitters have restricted access to smooth muscle receptors. Thus, different populations of receptors and signalling pathways might be accessed by transmitter substances released from nerve terminals than with bath application.
Finally, the responses of SMC are highly integrated. Many stimuli are present in the environment at a given moment in time, and this stimulus barrage is complicated by the addition of inflammatory mediators and paracrine substances during pathophysiological conditions. As we discover molecular mechanisms responsible for E-C coupling, a systems biology approach using modelling techniques should be implemented to supplement empirical findings. Such models, if based on real molecular events and mechanisms and verified by simulations of physiological and pathophysiological behaviours, might improve the efficiency of experiments, help predict efficacy of therapeutic approaches or even suggest hypotheses for new therapies.
The author is grateful to Orline Bayguinov for the experiment shown in Fig. 2 and for comments from Sang Don Koh and Brian Perrino about the content.
