Abstract
Asthma is a disease characterised by reversible contraction of airway smooth muscle. Many signalling pathways are now known to underlie that contraction, almost all of which revolve around Ca2+ handling. Ca2+ homeostasis in turn is governed by a wide variety of ionic mechanisms, which are still poorly understood. The present review will briefly summarise those mechanisms that have been recognised for decades, but will then devote considerable attention to several novel ionic signalling mechanisms such as capacitative Ca2+ entry, the reverse mode of the Na+/Ca2+ exchanger, the role of Cl- channels in the release of internal Ca2+ and that of ryanodine receptors in the refilling of the sarcoplasmic reticulum, as well as the regulation of the monomeric G-protein Rho by ionic mechanisms. Lastly, evidence will be provided that Ca2+-dependent contraction may be driven by spatial and temporal heterogeneities in the intracellular Ca2+ concentration (i.e. Ca2+ waves/oscillations) rather than by an increase in the global steady state intracellular Ca2+ concentration.
Asthma is characterised by variable increases in airway resistance. Irrespective of the aetiology of this disease, inappropriate contraction of the smooth muscle of the airways is a major factor that contributes to the increase in resistance. Thus, understanding the mechanisms surrounding force generation in airway smooth muscle (ASM) is paramount in the quest for novel treatments for this and many other respiratory diseases.
Advances in the understanding of excitation–contraction (EC) coupling in ASM have borrowed from work carried out using other muscle preparations (usually vascular smooth muscle or even skeletal/cardiac muscles). In some respects this is justified, given the similarities between these different cell types. However, a growing body of literature attests to the marked differences between ASM and those other cell types, as previously reviewed by Janssen 1. Thus, in order to accurately understand ASM physiology, it is important to synthesise and integrate data collected specifically from ASM and not to rely on findings made in cardiac or vascular smooth muscle, as so often seems to be the case.
The present review will begin with a brief presentation of the generally accepted working knowledge of EC coupling in ASM, focusing on Ca2+-dependent contractions that are triggered following agonist stimulation. In addition to the traditional mechanisms governing Ca2+ handling in ASM, evidence of other events which have been uncovered more recently in ASM will be summarised, such as capacitative Ca2+ entry and the reverse mode of the Na+/Ca2+ exchanger (NCX), which may contribute to agonist-induced elevations in intracellular Ca2+ concentration ([Ca2+]i). Lastly, evidence will be provided that Ca2+-dependent contraction may be driven by spatial and temporal heterogeneities in [Ca2+]i (i.e. Ca2+ waves/oscillations) rather than an increase in the global steady state [Ca2+]i.
OVERVIEW OF EC COUPLING IN ASM
The contractile apparatus consists of actin and myosin filaments and accessory and regulatory proteins. During excitation, phosphorylation of the light chain of smooth muscle myosin by myosin light chain kinase (MLCK) leads to enhancement of its intrinsic adenosine triphosphatase (ATPase) activity and subsequent mechanical interactions with actin. This results in the sliding of the myosin molecule along the actin filament. The details of this complex interaction are beyond the scope of the current discussion, but have been reviewed elsewhere 2. MLCK is activated by the Ca2+/calmodulin complex: thus, Ca2+ plays a key and central role in EC coupling in ASM (fig. 1⇓). A great many ionic events within the ASM cell culminate in changes in [Ca2+]i, as outlined below. It will first be described how [Ca2+]i is maintained at, or restored to, resting levels through the actions of several Ca2+ pumps and exchangers (see Ca2+ homeostasis: Ca2+ pumps and exchangers section). Next, it will be examined how [Ca2+]i is elevated through the release of internally sequestered Ca2+ (see Release of internally sequestered Ca2+ section) as well as the influx of external Ca2+ across the membrane (see Voltage-dependent influx of external Ca2+ and Voltage-independent influx of Ca2+ sections); the latter will necessitate a brief diversion in order to explore a variety of other ionic events and mechanisms that contribute to Ca2+ influx via voltage-gated Ca2+ channels (VGCC; see Voltage-dependent Ca2+ influx section). Finally, some very recent work will be summarised describing how these mechanisms have been assembled and integrated into very elaborate yet fundamental cell functions (see Novel mechanisms section).
Several other signalling pathways involved in mediating contraction will not be discussed in the present review. Some of these involve a change in Ca2+ sensitivity of the contractile apparatus via a suppression of myosin light chain phosphatase (MLCP) activity, mediated by RhoA-associated kinase or by protein kinase (PK)C. Excellent reviews have been provided by Somlyo and Somlyo 3. Alternatively, others have described contractions that are Ca2+ independent, and indeed that appear to be independent of ionic mechanisms in general, and these will also not be considered in detail in the present review, although excellent reviews have been given by Gunst and co-workers 4, 5.
Ca2+ HOMEOSTASIS: Ca2+ PUMPS AND EXCHANGERS
In general, [Ca2+]i is maintained at low levels by: 1) extrusion of Ca2+ from the cytoplasm via a Ca2+ ATPase and the NCX on the plasmalemma; 2) sequestration of Ca2+ into intracellular organelles; and 3) buffering of [Ca2+]i by various cytosolic Ca2+-binding proteins (fig. 1⇑).
Several organelles are known to sequester Ca2+, including the sarcoplasmic reticulum (SR), mitochondria and the nuclear envelope 6, 7. Of these, the SR is widely agreed to play the major role in regulating contractile function in ASM. The mitochondria seem to play a supportive or modulatory role in the regulation of [Ca2+]i in vascular smooth muscle, but this has not been adequately investigated in ASM 6, 8. There is widespread agreement that sarco-endoplasmic Ca2+ ATPase (SERCA) plays a key role in maintaining resting [Ca2+]i, because blockers purported to be selective for this enzyme, such as cyclopiazonic acid (CPA) or thapsigargin (TG), typically evoke a rise in [Ca2+]i and contractions. This suggests that there is a constant “leak” of Ca2+ out of the SR that is driven by the tremendous concentration gradient across the SR membrane (10,000-fold, comparable to that existing across the plasmalemma). Inhibition of SERCA leads to functional depletion of the SR, indicated by loss of the Ca2+-transient or Ca2+-dependent membrane currents in response to cholinergic stimulation 9, 10. This Ca2+ leak may account, in part, for the spontaneous Cl- currents (spontaneous transient inward currents) and K+ currents (spontaneous transient outward currents (STOCs)) recorded in ASM 11–15.
Some extrusion pathway must also be involved in Ca2+ homeostasis; otherwise, the internal Ca2+ pool would eventually become overloaded. The ubiquitous plasmalemmal pump, termed the plasma membrane Ca2+ ATPase (PMCA), is most probably involved, although there are no selective blockers or knock-out models available to test this hypothesis. Ca2+ can also be extruded via the NCX, which uses the energy resident within the Na+ gradient to move Ca2+ against its electrochemical gradient, extruding one Ca2+ ion in exchange for three Na+. NCX appears to be important in ASM of some species such as the cow 16, pig 17 and guinea pig 18, but apparently not at all in that of the dog 19. The ability of NCX to operate in a capacity of Ca2+ extrusion is widely and easily accepted; however, under certain conditions it can equally operate in the reverse, Ca2+-influx, mode (see NCX and Ca2+ store refilling section).
RELEASE OF INTERNALLY SEQUESTERED Ca2+
Activation of the contractile apparatus in ASM is initiated by agonist-induced mobilisation of internally sequestered Ca2+ from the SR. G-protein-coupled receptor (GPCR)-mediated activation of phospholipase C triggers the cleavage of membrane-bound phosphatidylinositol-4,5-bisphosphate, liberating inositol-1,4,5-trisphosphate (IP3), which in turn acts on specific ligand-activated Ca2+ channels on the SR (IP3 receptors). There are three subtypes of IP3 receptor which differ in their gating kinetics: although there have been many descriptions of IP3-induced release of Ca2+ in ASM 20–23, the subtype(s) present and/or involved in ASM are as yet unclear. IP3 receptors are also regulated by autacoids other than IP3. When [Ca2+] at the cytosolic face of the IP3-receptor rises into the micromolar range, the channel enters an inactivated state 24, 25, a property which is generally deemed to be important in explaining repetitive Ca2+ waves (see Ca2+ oscillation frequency determines contractile function section). Its activity is also modulated by phosphorylation of the receptor/channel by several different kinases, including PKA and PKC. As in other cell types, these IP3 receptors are inhibited by xestospongins, 2-aminoethoxydiphenyl-borate (2-APB) or heparin.
The SR also expresses another class of Ca2+ channel, referred to as the ryanodine receptor (RyR) because of its sensitivity to certain plant alkaloids. Once again, there are three subtypes of RyR. Using the experimental tools currently available, it would appear that human ASM expresses only RyR3 and not the other two subtypes 26, while RyR1 and RyR2 were found in rat ASM 27 and all three subtypes were found in murine ASM 28. Even more interestingly, in the latter study, RyR1 were localised towards the periphery of the cells whereas RyR3 were more centrally located around the nucleus 28. Such regiospecific heterogeneity would hint at physiologically important mechanisms (discussed later on and in the Transition from pharmacomechanical coupling to electromechanical coupling section). The endogenous ligand for the RyR is debated, with many proposing it to be cyclic ADP ribose 29–33, perhaps acting through FK506 binding protein of 12.6 kDa 31. The RyR also shows substantial sensitivity to [Ca2+]i, being enhanced by a range of [Ca2+]i above baseline (a phenomenon referred to as Ca2+-induced Ca2+ release, or CICR) and then suppressed as [Ca2+]i rises into potentially cytotoxic ranges 7, 34–36. Caffeine, a pharmacological tool often used to evoke Ca2+ release, acts by enhancing the Ca2+ sensitivity of the RyR such that basal levels of [Ca2+]i are sufficient to trigger CICR 36. Ryanodine, on the other hand, has several binding sites on the RyR with widely ranging affinities: relatively low concentrations of ryanodine bind to high-affinity sites which lock the RyRs in a subconductance state (leading to Ca2+ release), whereas higher concentrations bind to other sites and completely block Ca2+ conduction 36. RyRs are also regulated by endogenous signalling pathways, generally via phosphorylation by Ca2+/calmodulin-dependent kinase II or PKA 35, 36.
In many nonairway smooth muscle cell types, RyRs contribute to contraction through CICR 7, or to relaxation of vascular smooth muscle through activation of Ca2+-dependent K+ channels 37. The potential role of the RyR in ASM, on the other hand, is less clear. They are certainly functionally active: the RyR agonist caffeine typically evokes reproducible Ca2+ transients which are sensitive to SERCA inhibitors as well as to high concentrations of ryanodine 9–11, 26, 38, 39 or anaesthetics 26. However, agonist-evoked contractile responses in ASM show differential sensitivity to high concentrations of ryanodine: those in the airways of the mouse 28, 38, 39, rat 27 and dog 40 are suppressed, while in the airways of the human 26 and cow (unpublished data) they are not affected. This species-related difference may be due to the differential expression patterns (subtypes and regional heterogeneity) mentioned above. Conversely, relaxations in ASM evoked by β-agonists or nitric oxide donors are not inhibited by ryanodine (unpublished data). The current authors have hypothesised that RyRs function to discharge an overloaded Ca2+ store (see Superficial buffer barrier section and fig. 1⇑). This could account not only for the spontaneous transient currents mentioned above, but also for the observation that interleukin-4 inhibits cholinergic responses by activating RyRs and depleting the SR 41.
VOLTAGE-DEPENDENT INFLUX OF EXTERNAL Ca2+
The role for VGCCs in ASM is puzzling. Several lines of evidence suggest that they play a similar role in ASM as they do in other cell types. For example, many types of ion channels are involved in regulating membrane potential (Vm) at rest (see Resting membrane potential section) and in the generation of electrical slow waves (see Voltage-dependent Ca2+ influx section); also, many bronchoconstrictors trigger membrane depolarisation (see Agonist-evoked membrane depolarisation section), while bronchodilators can cause membrane hyperpolarisation. Thus, it would seem self-evident that Ca2+ channels play an important role in EC coupling in ASM, just as they do in other smooth muscle cell types. While this may be true under certain experimental conditions (see Transition from pharmacomechanical coupling to electromechanical coupling section), it is important to note that this is generally not true under normal physiological conditions: release of internally sequestered Ca2+ alone is sufficient to mediate the contractile response, as attested to by the persistence of contraction during exposure to blockers of the Ca2+ channels, removal of external Ca2+, or during voltage-clamp at Vm at which the Ca2+ channels are not active 1, 9, 42–44. Conversely, the present authors’ group and others have shown dihydropyridine-sensitive Ca2+ channels to play an important role in the refilling of the internal Ca2+ store 9, 45–48.
Resting membrane potential
Resting membrane potential (VR) in ASM does not correspond to the equilibrium potential of any single ion species. For example, VR ranges from -70– -30 mV, while the equilibrium potential for potassium (EK) is ∼ −80 mV, that for chloride (ECl) ranges from -33– -6 mV 49, 50, and those for Na+ and Ca2+ are in the very positive range (fig. 2⇓). This would suggest that VR is determined by membrane permeability to more than one ion. The fact that VR is intermediate between EK and the equilibrium potentials for all other ions suggests that K+ conductances must contribute to VR. Consistent with this, blockade of K+ channels using nonspecific blockers such as tetraethylammonium (TEA), Ba2+, or Cs+ leads to membrane depolarisation and contraction 51. The specific type of K+ channel involved in setting VR is largely the voltage-dependent “delayed rectifier” subtype 52–54. Maxi-K channels are unlikely to be active at VR and resting levels of [Ca2+]i. Most studies show that blockade of large-conductance Ca2+-sensitive K+ (KCa) channels (e.g. using charybdotoxin or iberiotoxin) has no effect on VR or resting tone 55–57, but there are exceptions 58. In addition, STOCs are not recorded until Vm exceeds -30– -40 mV in the smooth muscle cells of canine 13, guinea pig 13, 15 and porcine 59, 60 airways. Small-conductance KCa channels have been shown to be active at VR and may account for the depolarisation evoked by high concentrations of TEA; however, some find apamin has no effect on VR 61. Adenosine triphosphate-sensitive K+ channels are clearly not involved in setting VR, since resting potential and tone are unaffected by glibenclamide 52, 62, 63.
With respect to the contribution to VR made by other ion conductances, recent evidence of a tonic Cl- conductance in ASM has been presented. For example, in isolated strips of canine tracheal ASM, replacement of external Cl- with the impermeant anion isethionate (which would shift ECl in the positive direction) elicits membrane depolarisation and increases electrotonic potentials 64. The current authors have described spontaneous Cl- currents at Vm ranging -100– +50 mV 11, 13, 14. Thus, VR should be found somewhere between ECl (in smooth muscle, ECl ranges from -33– -6 mV; 49, 50) and EK (∼ −80 mV).
With a 10,000-fold concentration gradient tending to move Ca2+ into the cell, even a minute permeability to Ca2+ can lead to a significant Ca2+ influx, which will tend to depolarise the membrane. Many studies have indeed provided evidence of a persistent influx of Ca2+ across the membrane: for example, decreasing external [Ca2+] causes [Ca2+]i to drop 65. The voltage-activation and -inactivation properties of VGCCs in some ASM preparations are such that there can be a persistent Ca2+ influx or “window current” at membrane potentials ranging from -40–0 mV (see Voltage-dependent Ca2+ influx section). The possible contribution of voltage-independent Ca2+ channels will also be discussed in the Voltage-independent influx of Ca2+ section.
Agonist-evoked membrane depolarisation
Many agonists that evoke bronchoconstriction also depolarise the ASM cell membrane into the range of potentials at which VGCCs begin to activate (see Voltage-dependent Ca2+ influx section). In human, canine, guinea-pig and equine ASM, a wide variety of spasmogens evoke inward current of up to several thousand pA at VR 9, 11, 12, 66, 67; this current reaches a peak within a few seconds, then decays back to baseline levels in the continued presence of the agonist and is sometimes followed by a series of smaller secondary inward currents 11, 13. A variety of pharmacological tools and electrophysiological strategies have since revealed that this net inward current represents a mixture of conductance changes. Initially, there is a net decrease in membrane resistance 11, suggesting opening of ion channels: in all ASM preparations studied to date, these largely comprise Ca2+-dependent Cl- channels 11, 12, 59, 60, 66–68. In some species, this is accompanied by a more sustained activation of nonselective cation channels (NSCCs) 11, 69. These changes are then followed by a prolonged suppression of outward K+ currents 11, 56, 70, 71. Altogether, the transient activation of Cl- and NSCC currents leads to a displacement of the Vm towards the reversal potentials of those two currents (-30 and 0 mV, respectively), while the decreased K+ conductance decreases the strong hyperpolarising influence of EK (EK = -80 mV). During maximal excitation, the sum total of these changes is a depolarisation to ∼-40 mV, often with minor oscillations or “slow waves” around that level (described in more detail in the Electrical slow waves section). The current authors would argue that the role of this net change in membrane conductance is not to depolarise the Vm into a range that maximally stimulates VGCCs (+20 mV; see Voltage-dependent Ca2+ influx section), but rather to effectively maintain the Vm around -40 mV, which is optimal for a persistent low level of voltage-dependent Ca2+ influx (“window current”; see Voltage-dependent Ca2+ influx section) and for Ca2+ entry via the reverse-mode NCX (as argued in subsection entitled NCX and Ca2+ store refilling and in figures 2⇑, 3⇓ and 4⇓).
Cl- current activation appears to be a direct result of elevation of [Ca2+]i, because it can be triggered by agonists that release internally sequestered Ca2+ in a G-protein-independent fashion (e.g. caffeine) 9, 11 or by voltage-dependent Ca2+ influx 14. However, these channels quickly undergo a rapid phosphorylation by calcium/calmodulin-dependent protein kinase, which uncouples their activity from [Ca2+]i 72. The channels involved in mediating these currents exhibit a very small unitary conductance (<20 pA) and voltage-dependent inactivation 13. There have been no further characterisations of the Cl- currents in ASM since the mid-1990s, even though their molecular identity is still unknown: in other smooth muscle preparations, these may include bestrophins 73 or several different members of the Ca2+-activated Cl- channel (CLCA) family of Cl- channels 74.
The NSCCs activated by GPCRs in ASM were first studied in detail by Kotlikoff and co-workers 22, 69, 75. These are not Ca2+-activated since they are not triggered by caffeine, but are nonetheless Ca2+-dependent and coupled to plasmalemmal receptors via Gi/o proteins. Moreover, they were determined to be Ca2+ permeant (the fraction of current carried by Ca2+ estimated to be 14% at -60 mV) and blocked by Ni2+. The current authors have also recently described a NSCC which is activated by depletion of the internal Ca2+ store (described in detail in the section entitled Voltage-independent influx of Ca2+). It is unclear whether these channels are of a different subtype than the ones described by Kotlikoff and co-workers 22, 69, 75.
The K+ channels suppressed by spasmogenic stimulation appear to include both the Ca2+-dependent and voltage-dependent varieties. Suppression of the Ca2+-dependent current has been shown to be mediated by G-proteins 70, 76, but may also involve agonist-induced depletion of the internal Ca2+ pool with consequent decrease in Ca2+ release directed at the plasmalemma. Suppression of the voltage-dependent current, however, involves a shift in the voltage-dependence of activation to more positive potentials 70.
Voltage-dependent Ca2+ influx
Membrane depolarisation can result in opening of VGCCs and influx of external Ca2+. These currents have been characterised in the trachealis of the dog 77, cow 78, horse 79 and guinea pig 80 as well as the bronchi of the human 81 and dog 82. By and large, these currents exhibit properties typical of “L-type” Ca2+ channels, including: 1) threshold potential for activation of -45– -25 mV, and maximal activation at 0– +20 mV; 2) inactivation which is relatively slow (with a time constant of 0≈5 ms) and voltage-dependent (being half maximal at -25– -30 mV); and 3) blockade by dihydropyridines (e.g. nifedipine, nicardipine and nitrendipine) or phenylalkylamines (e.g. verapamil and D-600). These Ca2+ currents are also blocked by certain inorganic cations such as Cd2+, Co2+, Mn2+ and Ni2+, while divalents such as Ba2+ and Sr2+ permeate the channels. It is important to note that voltage-dependent inactivation of the channels is incomplete over the range of potentials at which the channels begin to activate (-40–0 mV); in other words, there is a range of potentials within which there is a window current, or persistent noninactivating influx of Ca2+ 83. Such a window current may be responsible for the Ca2+ influx that maintains slow wave activity in intact airway tissues (see Electrical slow waves section)
While the whole-cell studies summarised above indicate that the Ca2+ currents in tracheal smooth muscle (TSM) are predominantly of the dihydropyridine-sensitive or L-type, single-channel studies indicate the presence of two types of Ca2+ channel, with unitary conductances of 21–26 pS and 10 pS, respectively 84. The current authors have obtained evidence of both T- and L-type Ca2+ currents in canine bronchial smooth muscle 82. That is, depolarisation evokes a “fast” inward current (peak activation within 20 ms followed by rapid inactivation) and a “slow” inward current (peak activation at 50 ms and a much slower rate of inactivation); the overall Ca2+ current is approximately a 50:50 mixture of these two types. The fast current exhibits threshold and maximal inward current at -50 and -10 mV, respectively, while these values are -40 and +10 mV for the slow current. More importantly, the two currents differ markedly with respect to the voltage-dependence of inactivation: inactivation of the fast current is first noted at -80 mV, is half-maximal at -60 mV, and is complete at -40 mV, compared with values of -60, -40 and -10 mV, respectively for the L-type Ca2+ current. A Ca2+ current with these voltage activation and inactivation properties has not been reported in the trachealis of any of the species in which the L-type current has been carefully characterised. T-type Ca2+ currents are often important for pacemaking in other cell types: their role in canine bronchial smooth muscle is as yet unclear.
Electrical slow waves
Human 85 and guinea pig 86 ASM exhibit spontaneous mechanical and electrical activities, referred to as “basal tone” and “slow waves”, respectively. Electrical slow waves generally comprise oscillations in Vm centred around -40– -30 mV with amplitudes of up to 25 mV and a frequency of 1 Hz. These are unaffected by neuronal blockers (e.g. tetrodotoxin or removal of external Na+) or by antagonists of various receptors (e.g. antimuscarinics or antihistamines), suggesting that this activity is myogenic in origin. These spontaneous activities seem to be dependent, in part, on a constitutive metabolism of arachidonic acid, although the enzymatic pathway involved is apparently species-dependent. For example, basal tone and slow waves in guinea pig ASM are reduced by inhibitors of cyclo-oxygenase but are relatively unaffected by inhibitors of lipoxygenase and thromboxane synthase 86–88. Spontaneous activity in human ASM, on the other hand, is not reduced by inhibition of cyclo-oxygenase but is abolished by inhibition of lipoxygenase 85, 89–91. These observations suggest that spontaneous activity is mediated by constitutive generation of prostaglandins in guinea pig ASM, but of leukotrienes in human ASM.
The ASM of the dog 51, 92, 93, cat 94, ferret 95, cow 96 and horse 97 at rest is mechanically and electrically quiescent. However, spasmogens can evoke changes in mechanical and electrical activity that resemble the spontaneous activities described above. For example, while canine ASM is electrically and mechanically quiescent at rest, slow waves identical to those recorded from guinea pig and human ASM are evoked by cholinergic stimulation 51, the thromboxane analogue U46619 98, leukotrienes 95, 99, or K+ channel blockers 51. Similarly, histamine-induced stimulation of bovine TSM evokes phasic contractions and slow waves 96. In addition, indomethacin evokes contractions (apparently mediated by leukotrienes) in equine ASM tissues that were previously devoid of mechanical activity 97.
Whether recorded from tissues at rest (e.g. human and guinea pig) or from tissues stimulated by excitatory agonists (e.g. canine, bovine and equine), slow-wave activity is influenced by conditions which modulate voltage-dependent Ca2+ currents. It is reduced by removing external Ca2+ or by Ca2+ channel blockers 51, 100, suggesting it is initiated and maintained by a persistent influx of Ca2+ through those channels. In this context, it is noteworthy that the threshold potentials for slow-wave activity and for activation of L-type VGCCs in ASM are both ∼-40– -30 mV (fig. 2⇑). The repolarising phase of these oscillations in ASM seems to involve activation of Ca2+-dependent Cl- channels 101.
Thus, ASM, which is electrically quiescent under “basal” conditions, is capable of exhibiting slow-wave activity identical to that recorded from spontaneously active tissues if appropriately provoked. This suggests that a myogenic oscillatory mechanism resides in all ASM tissues and is invoked by excitatory stimulation. It is tremendously interesting that these electrical slow waves oscillate up and down through the range of potentials in which the window current of VGCCs exists, as well as in the vicinity in which the NCX might transition into reverse mode under certain conditions of excitatory stimulation (see NCX and Ca2+ store refilling section). This has enormous implications for EC coupling in ASM.
VOLTAGE-INDEPENDENT INFLUX OF Ca2+
Transient receptor potential ion channels
Transient receptor potential (TRP) ion channels were first described in the Drosophila visual system and are named after the role they play in phototransduction. This seminal discovery has since been followed by the identification of seven TRP homologues in mammals, namely the classical or canonical TRP (TRPC) family, the vanilloid TRP family, the melastatin TRP family, the mucolipin TRP family, the polycystin TRP family, the antigen-repeat TRP family, of which there is only one mammalian member, and no mechanoreceptor potential C channel, which does not appear to occur in the mammalian genome (for extensive reviews, see 102 and 103). In many nonexcitable cells, including mast cells, basophils, T-cells and megakaryocytes, TRP proteins are thought to form inwardly rectifying highly Ca2+-selective channels also referred to as Ca2+ release-activated Ca2+ (CRAC) channels 104. These channels mediate Ca2+ influx upon depletion of IP3-sensitive intracellular Ca2+ stores; a phenomenon that has been coined capacitative Ca2+ entry (CCE) 105. However, in excitable cells such as smooth muscle, TRP proteins form NSCCs that allow for the permeation of both mono- and divalent cations 106. As a consequence of their ionic promiscuity, NSCCs exhibit a characteristically linear current–voltage (I–V) relationship, with a reversal potential around 0 mV, when measured in near-physiological ionic conditions. Although NSCC proteins share a structural theme reminiscent of VGCCs, they are only distantly related. Indeed, NSCCs differ from their voltage-gated cousins in that they lack an inherent voltage sensor. Consequently, NSCCs activate in a voltage-independent fashion 104.
It has been proposed that TRPC proteins are largely responsible for the formation of NSCCs expressed in mammalian smooth muscle cells including rabbit cerebral 107 and human 108 and rat 109 pulmonary arterial myocytes. In ASM, there is evidence to support the expression of TRPC1, -3, -4, -5 and -6 in human 110, rat 110, guinea pig 111 and porcine 112 airways. However, there is little knowledge of the relative levels of expression of these subunits in the native cell, or of the functional significance of their expression. Increased expression of TRPC1 110 and TRPC3 110 has been observed in proliferating ASM cells cultured under normal conditions or in the presence of inflammatory mediators (e.g. tumour necrosis factor (TNF)-α), respectively. Despite minimal biophysical and electrophysiological data regarding the properties of TRP proteins in ASM, a number of recent reviews have broadly implicated TRP channels in the pathogenesis of several airway diseases including asthma, chronic obstructive pulmonary disease, cystic fibrosis and emphysema 113–115. Below a body of evidence is discussed that supports the present authors’ current understanding of the properties and physiological implications of NSCC activation by GPCR stimulation or intracellular Ca2+ store depletion in ASM.
GPCR signalling activates nonselective cation channels
Many agonists exert their contractile effects on ASM by signalling through GPCRs (e.g. acetylcholine, histamine and leukotrienes). GPCR stimulation not only leads to a transient IP3-dependent release of internally sequestered Ca2+ from the SR, but also results in the recruitment of NSCCs with a low current amplitude that exhibit a significant Ca2+ permeability under “physiological” conditions (representing ∼14% of the inward current at -60 mV; 75). It has been suggested that this current is responsible for the sustained elevation in [Ca2+]i observed during prolonged agonist stimulation 69, 75, 116. In contrast, patch-clamp studies reveal that Na+ ions contribute the majority of the inward current conducted by NSCCs, generating speculation with respect to the role of this conductance in agonist-evoked responses. Some have proposed that the influx of Na+ through NSCCs is important in effecting a substantial depolarisation, resulting in the recruitment of VGCCs to sustain or augment Ca2+ influx, thereby contributing to contraction. Alternatively, it has been proposed that the substantial Na+ permeation mediated by NSCCs may function to drive NCX in reverse mode, favouring Ca2+ influx (see NCX and Ca2+ store filling section). This model is supported by a recent report that TRP proteins are co-localised with NCX1 117. Furthermore, Dai et al. 17 reported that sustained contractions required Ca2+ influx mediated by the NSCCs and the reverse mode of the NCX. Indeed, findings from the current authors’ laboratory support a similar mechanism underlying sustained agonist-evoked contractions and SR Ca2+ refilling in ASM (see NCX and Ca2+ store filling section) 16.
In both equine 69 and porcine 116 trachealis, activation of NSCCs by histamine and/or acetylcholine relies exclusively upon the activities of Gi/Go proteins coupled to H1 histamine and M2 muscarinic receptors, respectively. Observations by Wang and Kotlikoff 69 suggest that GPCRs may signal through convergent pathways to recruit a common species of NSCCs in equine trachealis; a phenomenon evidenced by the equivalent I–V relationships measured during acetylcholine or histamine stimulation and by the lack of an additive nature of these currents upon concurrent stimulation with both agonists. However, this is most likely a species-dependent phenomenon, as acetylcholine-activated NSCC currents in porcine TSM could not be duplicated with histamine or leukotriene (LT)D4 116.
While it is apparent that an elevation in [Ca2+]i alone is insufficient to activate NSCCs in several ASM preparations 69, 75, 116, both histamine- and acetylcholine-induced NSCC currents do not activate unless GPCR stimulation is accompanied by an elevation in [Ca2+]i. For example, acetylcholine-activated NSCC currents in porcine trachealis do not manifest at [Ca2+]i <50 nM 116. Furthermore, in equine tracheal myocytes, the inhibition of histamine-induced cation current observed in the presence of U73122 (phospholipase C inhibitor) can be restored by simultaneous application of histamine and caffeine 116.
Unfortunately, the molecular identities of the proteins responsible for NSCCs in ASM are still unclear. Work in this area has been greatly hindered by the lack of specific inhibitors for TRP channels. Nevertheless, in isolated cultured human ASM cells, Ca2+ influx induced by acetylcholine or bradykinin is insensitive to organic Ca2+ channel inhibitors including nifedipine, nisoldipine or diltiazem, but can be completely inhibited by several polyvalent cations including Co2+, Mn2+, Ni2+, Cd2+ or La3+ 116, 118. Similarly, acetylcholine-activated NSCCs in equine tracheal myocytes cannot be blocked by nisoldipine, yet are rapidly inhibited by 10 mM Ni2+ 75. Although nifedipine reduces LTD4-induced contraction of human small bronchioles by nearly 40%, these responses are completely abolished by application of La3+, Gd3+ and SKF 96365, suggesting that NSCCs mediate the majority of Ca2+ influx observed in this preparation 119, 120. While these observations alone are insufficient to unequivocally implicate specific TRP protein(s) in forming NSCCs in ASM, the pharmacological and biophysical properties of the NSCCs examined thus far indicate that members of the TRPC family may be involved.
Nonselective cation channels mediate CCE
In ASM cells, depletion of intracellular Ca2+ stores activates a Ca2+ influx pathway which in human 110, rat 110, guinea pig 121, porcine 65, 112 and bovine 65 ASM is mediated by a channel that is sensitive to Ni2+, La3+ and SKF 96365, but not to nifedipine. Although it is unclear whether this pathway is modulated by release of Ca2+ through IP3 receptors, it does appear to be modulated by release of Ca2+ from RyRs 112. Indeed, Ay et al. 112 demonstrated that CCE can be induced in porcine TSM cells by depleting the internal Ca2+ store using CPA or prolonged exposure to caffeine. While concurrent administration of acetylcholine significantly augmented CCE, this enhancement was found to be due to the recruitment of a separate group of NSCCs whose function specifically relies on release of SR Ca2+ via IP3 receptors (i.e. they are blocked by xestospongin D) but did not appear to involve diacylglycerol or activation of PKC 112. It is still unclear which TRP proteins mediate these responses in porcine ASM; however, in cultured human ASM there is compelling evidence that a preferential increase in the expression of TRPC3 is responsible for mediating elevated resting [Ca2+]i and enhanced CCE in the presence of the pro-inflammatory cytokine TNF-α 122. The latter study reported that CPA-induced CCE was enhanced by TNF-α, but inhibited by Ni2+ and La3+. Furthermore, these responses could be blocked using small interfering RNA directed toward TRPC3. Interestingly, in TNF-α treated tissue, CCE responses measured in the presence of acetylcholine were insensitive to atropine (muscarinic receptor antagonist), whereas responses in nontreated cells were reduced by blockade of the muscarinic receptors. This suggests that TNF-α preferentially upregulates the expression of TRPC proteins involved in the formation of “store-operated” NSCCs 122.
While there is a substantial knowledge base regarding the electrophysiological properties of the NSCCs involved in mediating CCE in vascular smooth muscle preparations 106, there remains a relative dearth of information in the ASM field. Several studies have examined the effects of store depletion on Ca2+ influx and the regulation of [Ca2+]i in ASM (see above), but only three reports to date have utilised patch-clamp techniques to measure directly the resultant membrane currents 65, 110, 123. In cultured human and rat bronchial smooth muscle cells, Sweeney et al. 110 demonstrated that the current density of a Ni2+-sensitive NSCC activated by passive store depletion was markedly increased in proliferating versus growth-arrested cells. This increase in membrane current density was found to be associated with enhanced CCE accompanied by a complementary increase in TRPC1 mRNA expression; unfortunately Western blots were not performed to confirm that TRPC1 protein levels were similarly increased, and other TRPC family members were not examined. The current authors’ group has recently provided a detailed description of the electrophysiological properties of a CPA-evoked NSCC conductance in porcine and bovine trachealis 65. It was described how internal Ca2+ store depletion augments a basal NSCC conductance in TSM cells which is sensitive to micromolar concentrations of La3+ and Gd3+, and exhibits a marked permeability to Na+ at physiologically relevant Vm 65. Indeed, the current authors demonstrated that while the bulk of the current mediated by this NSCC conductance is carried by Na+, it too possesses a significant Ca2+ permeability that is capable of sustaining elevated [Ca2+]i responses in the presence of extracellular Ca2+. Interestingly, this conductance appeared to be constitutively active under “resting” conditions, and was likely to be responsible for the basal, nifedipine-insensitive, Ca2+ influx pathway observed in this preparation 65. Others have similarly reported the existence of a basal Ca2+ influx pathway operating in ASM 121, 122, 124, 125, speculating that this plasmalemmal Ca2+ “leak” contributes to regulation of basal [Ca2+]i 126.
Although it is clear that SR Ca2+ depletion is sufficient to activate NSCC in ASM, it is not clear whether the NSCC conductance described by the present authors’ group 65 and others 110 is in turn regulated by [Ca2+]i, as appears to be the case for GPCR-activated NSCCs in other ASM preparations 69, 75, 116. This uncertainty stems from the fact that CPA is known to induce a biphasic change in [Ca2+]i 65, 127: this change includes an initial transient increase in [Ca2+]i (due to direct antagonism of SERCA-mediated sequestration of Ca2+ and an unmasking of SR Ca2+ leak via RyRs) followed by a sustained decline in [Ca2+]i (due to extrusion of Ca2+ from the cell) 127. To resolve this question, the current authors showed that loading the cells with BAPTA, which depletes the SR without stimulating an elevation in [Ca2+]i, is sufficient to activate NSCCs having pharmacological and electrophysiological properties identical to the CPA-evoked NSCC currents (unpublished observations). Although the pharmacological profile of store-depletion activated currents in ASM is similar to the GPCR-activated currents described above (see GPCR signalling activates nonselective cation channels section), only the former can be activated in the absence of an elevation in [Ca2+]i, suggesting that several functionally distinct species of NSCCs may be expressed in ASM.
The mechanisms responsible for conveying SR Ca2+ status to the plasmalemmal ion channels which mediate CCE have been studied for decades but remained frustratingly elusive. However, recent work in several nonmuscle cells point to the involvement of stromal interacting molecule (STIM), initially described as a candidate tumour suppressor gene almost a decade ago 128, and its human homologue STIM1. In particular, STIM has been found to be an essential component of TG-induced CCE and CRAC currents in Drosophila S2 cells, rat basophilic leukaemia (RBL) cells, Jurkat T-cells and HEK 293 cells 129–132. This type I transmembrane protein contains an EF-hand motif located near its amino-terminus which is thought to function as a Ca2+ sensor of the endoplasmic reticulum (ER)/SR Ca2+ store 133, 134. In resting Jurkat T-cells, RBL cells, human T-cells and pheochromacytoma cells, STIM1 resides in ER-like structures, co-localised with SERCA2, but translocates to the plasmalemma upon Ca2+ store depletion 134. In cells made to express EF-hand mutants of STIM1 with compromised Ca2+ binding ability, STIM1 localises to the plasmalemma but not the ER, and the cells exhibit elevated resting [Ca2+]i and persistent TG-independent Ca2+ influx, both of which are inhibited by 2-APB, SKF 96365 and Gd3+. Altogether, the data suggest that STIM1 translocates to the plasma membrane when it is no longer able to bind ER/SR Ca2+ (e.g. when the Ca2+ store is depleted) and directly activates CRAC channels in those cells. The question then becomes whether this mechanism operates in ASM.
Only one publication to date has addressed this question in ASM, finding both CCE and store depletion-evoked membrane currents in human bronchial smooth muscle cells to be dependent upon STIM1 but not STIM2 (a related protein) 123. Interestingly, targeted silencing of STIM1 inhibited both CPA- and histamine-induced Ca2+ influx and CCE, while bradykinin responses were largely unaffected. Furthermore, CPA-induced NSCC currents were almost completely abolished by targeted suppression of STIM1 (but not STIM2). Altogether these observations have several profound implications. First, STIM1 may indeed be mechanistically involved in transmitting SR Ca2+ store status to store-operated NSCCs in ASM. Secondly, the study provides evidence of agonist-specific differences in the ability to activate store-operated NSCCs, which may reflect variations in the level of PLC activation and subsequent IP3-mediated SR Ca2+ release (i.e. the relative magnitude of SR depletion) associated with each agonist 123. In addition, the study also highlights the difficulty of ascertaining the relative contribution of GPCR-activated and store-operated NSCCs in mediating agonist-evoked membrane currents and/or Ca2+ influx.
Regardless of the mechanism by which NSCCs in ASM are activated (i.e. GPCR stimulation or depletion of intracellular Ca2+ stores), they clearly contribute to the regulation of [Ca2+]i. In some cases, NSCCs are involved in CCE-evoked contraction of human 110, 119, 120 and rat 110 bronchial and guinea pig 121 TSM. In addition, there is mounting evidence to suggest that NSCCs may facilitate Ca2+ entry indirectly by depolarising the membrane (thereby activating L-type Ca2+ channels) and/or promoting reverse-mode operation of NCX. Unfortunately, it is still largely unclear whether these NSCCs actively participate in the refilling of the SR, a function that is entirely consistent with the nature of the stimulus responsible for activating these channels. Further clarification of these questions may require the use of small interfering RNA or other genomic techniques (e.g. gene knock-out models) in order to determine which TRP proteins are functionally involved in the regulation of ASM contraction and airway calibre.
NOVEL MECHANISMS
NCX and Ca2+ store refilling
Following agonist-induced elevation of [Ca2+]i, conditions are returned to normal through the extrusion of Ca2+ to the extracellular domain and/or re-sequestration into the SR. Ca2+ extrusion must be carefully matched to net Ca2+-influx: otherwise, there will be a deficit in the amount of intracellular Ca2+ available to refill the SR. Thus, influx pathways must exist to account for the loss of extruded Ca2+. The role of NSCCs in this function has been discussed previously in the Voltage-independent influx of Ca2+ section, while a possible role for voltage-dependent Ca2+ channels in this capacity will be considered later in the Transition from pharmacomechanical coupling to electromechanical coupling section. Recently, Dai et al. 17 introduced a third pathway contributing to Ca2+ influx during agonist-induced contraction of porcine ASM (Ca2+ influx mediated by the reverse mode of the NCX) as has been described in other nonairway preparations 17, 135–138. The current authors later showed the importance of this mechanism in SR refilling in bovine ASM 16, as have others in murine ASM 139. This Ca2+-influx pathway appears to be dependent on the availability of extracellular Na+ 16. Interestingly, Rosker et al. 140 have described an intimate relationship between the NCX and TRP proteins, building on an earlier report by Moore et al. 141 that suggested that NCX and the Na+/K+ pump are co-localised in smooth muscle caveolae in close apposition to the SR. Thus, it is possible that the reverse mode of the NCX, driven by membrane depolarisation and elevated [Na+]i in the region between the SR and the plasmalemma, may contribute to the Ca2+ influx pathways required to replenish the smooth muscle cell of Ca2+ and allow for maximal refilling of the SR.
The direction of ionic flux through the NCX can be predicted by calculating its reversal potential (ENCX) under varying concentrations of [Ca2+]i and [Na+]i, and comparing it to the Vm under the same conditions. That is, given the stoichiometry of the NCX (three Na+ ions transported for every Ca2+), and the relative valencies of these cations, ENCX is defined with reference to the equilibrium potentials for sodium (ENa) and calcium (ECa) thus 85:
ENCX = 3ENa–2ECa (1)
Since ENa and ECa are not static, ENCX is a dynamic value that will change depending on the intracellular environment (i.e. changes in [Ca2+]i and [Na+]i; figs 3⇑ and 4⇑). As such, when Vm is more negative than ENCX, the NCX will act in the forward mode, extruding Ca2+ from the cell; conversely, if Vm is more positive than ENCX, the NCX will act in the reverse mode, leading to Ca2+ influx. During agonist-induced release of internally sequestered Ca2+, there is membrane depolarisation and significant inward Na+ current (via activation of NSCC and/or TRP channels) which elevates [Na+]i in the domain immediately adjacent to the plasma membrane. Increased [Na+]i, in conjunction with low-to-moderate [Ca2+]i, brings ENCX into a range of potentials more negative than that of the depolarised membrane (i.e. <-30 mV; fig. 4⇑), thereby forcing the NCX into reverse-mode operation. Ca2+ entering the cell may be directly shuttled into the depleted SR, keeping the [Ca2+]i in this subplasmalemmal region relatively low until Ca2+ uptake slows (i.e. because the SR is full), at which point Ca2+ accumulates and NCX flips back into the forward mode.
Some previous studies, including those of the current authors, have concluded a central role for L-type Ca2+ channels in store-refilling on the basis of data obtained using techniques aimed at manipulating Vm (e.g. KCl-induced depolarisation or cromakalim-induced Vm clamping) 47, 142, even though inhibition of L-type Ca2+ channels did not completely abolish store-refilling 47, 48. Now recognising reverse-mode NCX as another voltage-dependent Ca2+ influx pathway, the current authors would re-evaluate those earlier data; future studies examining Ca2+ handling and EC coupling should not overlook the potential for Ca2+ influx via the NCX.
Cl- and Ca2+ handling
As mentioned in much of the present review, Ca2+-dependent contraction in ASM is driven primarily by the release of Ca2+ from the SR. In addition to activating the contractile apparatus, Ca2+ release from the SR can trigger membrane currents, including Ca2+-dependent Cl- currents. Cl--mediated depolarisation drives Ca2+-dependent contraction in many nonairway smooth muscle types through the activation of L-type Ca2+ channels. However the present authors, and others, have reported that inhibition of Cl--mediated depolarisation using selective Cl- channel blockers (niflumic acid and 5-nitro-2-(3-phenylpropylamino)benzoic acid) has little effect on the cholinergic concentration–response relationship in standard organ bath experiments. This poses the question: “what role does Cl- play in ASM?” A growing number of reports suggest that intracellular Cl- (Cl-i ) may be important in Ca2+ handling in smooth muscle.
The movement of Ca2+ across the SR membrane is an electrogenic processes (fig. 5⇓). Since the efflux of Ca2+ through RyR and IP3-receptors is entirely a passive process, accumulation of charge across the SR membrane could electrostatically hinder the sustained movement of Ca2+ into or out of the SR. However, compensatory ion fluxes across the SR membrane have been found in other nonairway preparations which are thought to neutralise any charge accumulation due to Ca2+ movement 143, 144. Cl- seems to be particularly important in this respect. For example, replacement of Cl- with large impermeant anions reduced the early phasic portion of agonist-induced contraction of ileal longitudinal smooth muscle in parallel with a reduction in cellular 45Ca2+ uptake 145, 146. Similarly, inhibition of SR Cl- channels reduced Ca2+ sequestration into saponin-permeabilised gastrointestinal smooth muscle cells 144, and depletion of Cl-i reduced angiotensin II- and norepinephrine-induced contractions of vascular smooth muscle 147. Recently, the current authors reported that the depletion of Cl-i through prolonged bathing in Cl--free bathing solution significantly reduced successive agonist-induced contractions in ASM 148; additionally, the rates of contraction and relaxation (following agonist removal) were greatly reduced when Cl-i was reduced, suggesting the release and reuptake of Ca2+ were both hindered by depletion of Cl-i. In contrast to those effects of chronic Cl-i depletion, acute removal of Cl- from the bathing solution accelerated contractile responses (but slowed relaxations). An interpretation of this is that sudden removal of external Cl- would augment Cl- efflux from the cell during agonist stimulation, resulting in greater reduction in [Cl-]i in the region adjacent to the SR and allowing for more Cl- to leave the SR, which in turn would augment the rate of Ca2+ release, as evidenced by a greater rate of contraction. Since the removal of extracellular Cl- increases Cl- efflux, the cell experiences a short-term depletion of Cl-i, resulting in a reduced ability to reuptake Ca2+ into the SR, as seen in the decreased rate of relaxation.
Altogether, then, there is a growing body of evidence to suggest that the burst of Ca2+-release leads to a rapid activation of Cl- currents which in turn boosts further Ca2+ release (by promoting efflux of Cl- from the SR to compensate for the charge build-up). The rapid inactivation of those Cl- channels by calcium/calmodulin-dependent protein kinase (described in the Agonist-evoked membrane depolarisation section) may represent an important timing mechanism, allowing for “resetting” of the Cl- and Ca2+ gradients by SR and plasmalemmal pumps, and thus setting the stage for subsequent bursts of Ca2+ (i.e. repetitive Ca2+ waves).
Regulation of Rho/ROCK by Ca2+, Cl- and Vm
The current authors were the first to show in ASM that RhoA is activated by 60 mM KCl 149, contrary to initial expectations. Follow-up work showed that this is directly related to elevated [Ca2+]i, although membrane depolarisation per se may also be involved 150. Changes in Rho-associated kinase (ROCK) activity paralleled those in RhoA, suggesting KCl does not exert an additional effect on ROCK (i.e. is only stimulating RhoA). It will be important, then, to ascertain how Ca2+ and Vm stimulate RhoA activity.
Rho activation itself may ultimately prove to be Ca2+ dependent: the current authors have already ruled out roles for Ca2+/calmodulin-dependent kinase II and for PKC (which is also Ca2+ dependent) in activation of Rho 150; others have shown focal adhesion kinase (FAK) phosphorylation following GPCR stimulation to be Ca2+ dependent 151, 152. Another possibility is that Ca2+, being a divalent cation, influences the interactions between G12,13, RhoGDI, RhoGEF, RhoGAP and/or RhoA by interfering with Mg2+-dependent activation of RhoA (Mg2+ influences guanosine triphosphate/guanosine diphosphate exchange activity 3, 153–156).
Conversely, Vm per se could influence Rho activity, both directly and indirectly. Directly, Rho is a charged molecule and would be expected to be influenced by the transmembrane voltage gradient as it approaches and inserts into the membrane (both steps are prerequisite for its activation). Indirectly, voltage-sensitive channels may signal to Rho upon membrane depolarisation: there is a growing literature describing direct physical interactions between various enzymes and ion channels, including L-type Ca2+ channels 157, 158. As such, the voltage-dependence of Rho-activation could be conferred by the Ca2+ channels. Such a direct channel–enzyme interaction does not necessarily imply that Ca2+-channel blockers should prevent Rho activation, because these only prevent ion conduction through the channels and do not prevent the depolarisation-induced conformational changes that might stimulate Rho activity.
Ca2+-dependent Cl- channels may also transduce the signal between the plasmalemma and Rho. There are two pieces of evidence that suggest that Cl-i modulates RhoA/ROCK signalling in ASM. First, while characterising the agonist-evoked Cl--currents in ASM 11, 12, 66, 67, the current authors noted anecdotally that cells perfused internally with a Cl--deficient electrode solution quickly lost the ability to contract to acetylcholine, even though the membrane current responses persisted. In contrast, control cells perfused internally with a 140 mM Cl- electrode solution could contract repeatedly for a number of hours 9; more interestingly, those contractions could be evoked during voltage-clamp at -60 mV and in the presence of CPA, indicating they are independent of both voltage-dependent Ca2+ influx and release of internal Ca2+. At that time, the reason for the discrepancy between high and low Cl- conditions was unknown. Additionally, the current authors have more recently found that the Cl- channel blocker niflumic acid markedly suppresses cholinergically-induced RhoA-activation (unpublished data). These observations lead to a suggestion that Cl-i may modulate Rho/ROCK activity. Changes in subplasmalemmal [Cl-] might facilitate translocation of RhoA to the membrane, or enhance interactions between the different components of this signalling cascade: others have shown G-protein activity to be modulated by Cl-i 159.
Ca2+ oscillation frequency determines contractile function
The traditional approach to Ca2+-dependent contraction in smooth muscle considers only global and sustained elevations of [Ca2+]i. However, there is increasing evidence for considerable spatial and temporal heterogeneities in [Ca2+]i elevation 17, 21, 39, 160, 161. In particular, agonists evoke a large Ca2+ wave which sweeps across the length of the cell, followed by a series of subsequent Ca2+ waves at periodic intervals with a frequency which seems to correlate with the degree of tone 39, 161. Ca2+ waves in neighbouring cells are generally asynchronous. In some cases, unitary Ca2+ events referred to as Ca2+ sparks have been described 15.
The mechanisms underlying Ca2+-wave generation and propagation have been studied for over a decade, yet are still poorly understood. Agonist-induced oscillations in [Ca2+]i have been proposed to be triggered by IP3. However, Dai et al. 17 reported the generation of oscillations in the presence of the IP3-receptor antagonists xestospongin C and 2-APB. Continuous Ca2+ oscillations can be inhibited by caffeine and ryanodine 17, 39, 162, reinforcing the central role of the SR and possibly suggesting that CICR is involved. The importance of SR Ca2+ is also highlighted by the sensitivity of Ca2+ oscillations to SERCA inhibitors or chelation of intracellular Ca2+ 17, 39. In addition to Ca2+ released from the SR, Ca2+ influx pathways play a crucial role in the continuous generation of agonist-induced asynchronous Ca2+ oscillations. Removal of extracellular Ca2+ from bathing solutions results in a gradual run-down of oscillations. Early studies revealed that agonist-induced oscillations were sensitive to verapamil and enhanced by BAY K-8644, suggesting a role for L-type Ca2+ channels 60. It is unlikely that the Ca2+ waves are triggered by electrical slow waves, given that the frequency of the former is roughly 1–2 orders of magnitude lower than that of the latter. Conversely, inhibitors of NSCCs (e.g. SKF 96365 and Ni2+) and of reverse-mode NCX inhibit oscillations and the corresponding force generation in responses to cholinergic stimulation 17, suggesting involvement of those other pathways as well. The contributions of many other molecular and cellular entities and signalling events have also been examined, including: mitochondria; cytosolic Ca2+-buffering proteins; Ca2+uptake and extrusion rates; and the rates of activation and inactivation of IP3-receptors and RyR, etc. The integration of these various parameters into a mathematical model is beyond the scope of the present manuscript, but has already been carried out elsewhere 163–165.
Very little is known about how agonist-induced Ca2+ oscillations in ASM determine the level of sustained force generation: how are these Ca2+ signals integrated and transduced at the level of the contractile apparatus? It is evident from the study of genetically hyper- and hyporesponsive mice that Ca2+ waves of similar frequency can result in very different contractile responses 160, and comparisons of agonist-evoked constriction of airways versus pulmonary vasculature reveal very different frequency-response relationships 21, 166–168. Thus, it appears that the Ca2+-integrative ability of the contractile apparatus in smooth muscle can vary considerably. However, the details of this frequency-sensitive integration in Ca2+-dependent contraction is lacking. One hypothesis relates this integration to the activity of MLCP. MLCK activity is directly proportional to [Ca2+]i, whereas MLCP activity is increased by accumulation of its substrate: phosphorylated myosin. As such, sustained contraction would require continuous stimulation of MLCK activity, and the net level of tone would be determined by the relative levels of MLCK and MLCP activities: in other words, the stimulus to contract would be encoded within the absolute magnitude of a tonic change in [Ca2+]i (fig. 6a⇓). Such a mechanism requiring continuous myosin phosphorylation/dephosphorylation and Ca2+ release/uptake would be very expensive energetically. Instead, if MLCP were briefly but reversibly suppressed by Ca2+, a Ca2+ wave would be sufficient to stimulate MLCK and evoke a contraction that would eventually resolve (be reversed by MLCP) unless/until another Ca2+ wave were triggered to reinvigorate MLCK and keep MLCP in check: in other words, net myosin phosphorylation and excitation could be encoded within the frequency of Ca2+ spiking (fig. 6c⇓) in a much more economical fashion. Other possible targets that might be regulated by Ca2+ wave frequency include MLCK, RhoA or PKC, all of which can regulate the activity of the contractile apparatus.
Superficial buffer barrier
Electron microscopy of vascular smooth muscle cells shows the SR to form sheets around the internal periphery of the cell 169, thereby dividing the cytosol into two spaces in which [Ca2+]i could be regulated independently. Ion channels and plasmalemmal-associated enzymes would sense changes in [Ca2+]i in the peripheral space immediately underneath the plasmalemma, while the contractile apparatus would sense [Ca2+]i in the deep cytosolic space underneath the SR. Another recent study suggests that the same anatomical arrangement can be found in ASM 170. Such an arrangement has important physiological consequences, as it may allow the cell to effectively dissociate the influences of [Ca2+]i on mechanical and electrical activities. The current authors have also hypothesised that RyR may be preferentially located on the subplasmalemmal face of the SR and play an important role in directing Ca2+ from an overloaded SR towards Ca2+-extrusion mechanisms on the plasmalemma (e.g. PMCA; NCX), whereas IP3 receptors play the primary role in EC coupling. Others have provided functional evidence for this. One group described the differential distribution of RyR subtypes at the periphery and central regions of ASM cells 28. An earlier study showed the β-agonist isoproterenol to simultaneously decrease [Ca2+]i in the deep cytosol while elevating [Ca2+]i in the periphery of the cell 171: presumably, the goal of these changes is to move Ca2+ away from the contractile apparatus toward the membrane for subsequent extrusion out of the cell. This close apposition of the SR and plasmalemmal membranes, with juxtaposition of RyR on the SR and various Ca2+-handling entities on the plasmalemma (PMCA; NCX), is a sine qua non for a peculiar phenomenon described in the next section in which Ca2+ flows in the opposite direction, from the extracellular space directly into the SR, bypassing SERCA.
The division of the cytosol into two compartments may not only be relevant to Ca2+ signalling. A recent study of Rho/ROCK-signalling in ASM found two distinct pools of Rho, one near the membrane and one in the deep cytosol, with different time-courses of activation following agonist stimulation and different roles in regulation of myosin light chain phosphorylation 172.
Transition from pharmacomechanical coupling to electromechanical coupling
Under normal physiological conditions, agonist-evoked contractions in ASM are sustained and largely insensitive to classical Ca2+-channel blockers. However, many groups have shown that under conditions in which handling of Ca2+ by the SR is disrupted, those responses become acutely dependent upon voltage-dependent Ca2+-influx. For example, following functional depletion of the SR using CPA, cholinergic stimulation of ASM leads to substantial phasic contractile responses which are precipitously abrogated by subsequent exposure to dihydropyridine blockers of L-type Ca2+ channels 46, 173–176; in this respect ASM acts much like vascular and gastrointestinal smooth muscle preparations. Similar observations are made in ASM tissues challenged with cholinergic agonists in the presence of ryanodine 177–179. Although the conditions surrounding these responses are totally artificial, such experiments may reveal something important and fundamental about EC coupling in ASM.
Data from several groups, including that of the current authors, suggest a direct coupling between L-type Ca2+ channels and the SR, allowing Ca2+ to move directly from the extracellular space into the SR without the involvement of SERCA activity. This mechanism was first proposed on the basis of mechanical data (contractions evoked by repeated cholinergic stimulation) as an indirect measure of Ca2+ handling 45–48. More compelling, though, was the monitoring of Ca2+ handling using membrane Cl- currents, which are directly Ca2+ dependent (in contrast to mechanical responses, which are quite heterogeneous with respect to Ca2+ dependence): using this approach, the present authors demonstrated SR depletion (reflected in disappearance of the Cl- currents) and then refilling using voltage pulses (reflected in reappearance of those currents) in voltage-clamped TSM pretreated with CPA 9. There have been innumerable studies of Ca2+ uptake in a wide variety of cell types: all the evidence suggests SERCA is the only type of Ca2+ pump on the SR, and no study has identified a SERCA that is insensitive to CPA. As such, refilling in the presence of CPA implies that voltage-dependent Ca2+ influx is somehow directed into the SR, but not necessarily via SERCA. Elsewhere, a model has been proposed which describes one such alternative refilling pathway 180–185. Briefly, agonist-induced depletion of the internal store triggers activation of protein tyrosine kinases (PTK), Ras and reorganisation of the cytoskeleton in such a way as to directly couple IP3 receptors on the SR with Ca2+ channels on the plasmalemma. Several observations made in ASM are consistent with such a mechanism: 1) spasmogenic stimulation of ASM is accompanied by activation of PTKs 186, 187 and Ras/Rho 188–191, as well as cytoskeletal rearrangement 189, 191, 192; 2) inhibition of PTK compromises SR refilling 193; 3) ASM depleted of FAK, which regulates cytoskeleton stability, shows marked suppression of cholinergic Ca2+ transients and contractions as well as changes in voltage-dependent Ca2+ channel function, without any disruptive changes in the contractile apparatus per se when assessed by addition of Ca2+ to permeabilised muscle strips 194; 4) in the rat ASM, RyR1 on the SR co-localise with voltage-dependent Ca2+ channels on the plasmalemma 27. However, the possible role for this novel SR refilling pathway has not yet been tested in ASM: the use of inhibitors of cytoskeletal organisation (e.g. cytochalasin D, jasplakinolide, colchicine and vinblastine) would be invaluable in this respect.
It is still quite unclear how Ca2+ enters the SR, if not via SERCA. Although SERCA appears to be the only pathway for active uptake into the SR, this does not mean it is the only possible route for Ca2+ entry. Under normal physiological conditions, IP3 receptors and RyRs allow Ca2+ to leave the SR: however, this movement is entirely passive, with Ca2+ moving down its concentration gradient. It is conceivable, then, that if the driving force on Ca2+ were reversed, these channels could allow Ca2+ back into the SR: detailed electrophysiological studies show these channels do not exhibit current rectification 195, 196. Ca2+ influx through L-type Ca2+ channels, TRP channels and/or reverse-mode NCX could allow [Ca2+]i to accumulate within the “superficial buffer barrier” formed by the close apposition of the plasmalemmal and SR membranes: as such, under conditions in which the SR were depleted (e.g. by agonist stimulation or by CPA), opening of those Ca2+-permeable channels on the SR would allow refilling. Given that Ca2+ release through RyR and IP3 receptors is inhibited by Ca2+ at micromolar concentrations via a negative feedback mechanism 24, 25 it might be reasonable to assume that retrograde flux back into the SR might likewise be susceptible to higher [Ca2+] in the subplasmalemmal space. However, fluorimetric measurements of [Ca2+] during agonist stimulation show that, while the initial Ca2+ spike may rise into the micromolar range, the sustained (or plateau) portion is well below that level. Also, it is possible that the cytoskeletal rearrangements which bring the plasmalemmal and SR-associated ion channels into close proximity are accompanied by (or cause) changes in the properties of the RyR and/or IP3-receptors (e.g. obstruction of the Ca2+ binding site which mediates channel inactivation).
Ca2+ which enters the depleted SR in this way could continue down its electrochemical gradient out of the other side of the SR and contribute to contraction. The contractions evoked under these conditions would necessarily be completely sensitive to L-type Ca2+-channel blockers, consistent with the otherwise paradoxical observations summarised at the beginning of this section.
UNRESOLVED QUESTIONS: DIRECTIONS FOR FUTURE RESEARCH
The exact distribution of the various ionic channels, pumps and exchangers needs to be ascertained with a high degree of precision. For example, are RyR and IP3 receptors uniformly distributed across the SR, or are they clustered heterogeneously in ways which have important functional implications? That is, are these SR-associated entities brought into close proximity with others on the plasmalemma, such as Ca2+-channels, NCX and PMCA (as demanded by the models proposed in the Superficial buffer barrier and Transition from pharmacomechanical coupling to electromechanical coupling sections)? Are NSCCs co-localised with NCX (as implied in the NCX and Ca2+ store refilling section)? Do VGCCs or Ca2+-dependent Cl- channels interact with Rho (see Regulation of Rho/ROCK by Ca2+, Cl- and Vm section)?
The mechanisms underlying discharge and refilling of the internal Ca2+ pool need to be better understood. What molecular entities are involved in producing the periodicity of the Ca2+ waves, and how are these waves transduced by the contractile apparatus? What might be the relationship between electrical slow waves and Ca2+ waves? The former would be expected to influence the driving forces on the cations which permeate NSCCs and the “window current” of the VGCCs, which in turn would influence the driving forces on NCX (as would the slow waves themselves); is there then crosstalk between the electrical slow waves and Ca2+ waves? Likewise, how might Cl- currents contribute to emptying/refilling of the Ca2+ store? What refilling pathways are operative other than SERCA: can there be reverse flow through RyR and/or IP3-receptors?
Another important area for further investigation is the regulation of Rho/ROCK by Ca2+, Cl- or Vm. Does that enzyme interact with certain ion channels at the membrane? Is there crosstalk between Rho/ROCK activities and the Ca2+ waves?
The molecular identities of the plasmalemmal and SR Cl- and NSCC channels remain unknown. This is in stark contrast to our relatively specific knowledge of the subtypes of voltage-dependent K+ and Ca2+ channels in ASM 52, 53. Is the Cl- conductance a member of the CLCA family 74, 197 or a bestrophin 73? Which of the many TRP proteins make up the NSCCs in ASM?
Much of the work summarised in the present review was obtained using tracheal preparations, even though there is a growing body of literature attesting to the important differences between tracheal and bronchial smooth muscle. These express different ion channels 82, 198 and appear to rely differently upon various EC coupling mechanisms 175. This is an urgent matter, since the bronchi play a far more important role than the trachea in determining resistance to airflow, and are the locus of inflammation and other functional/structural changes associated with asthma. Thus, it will be crucial to examine these mechanisms in bronchial preparations, and to rely less heavily on tracheal ones. Likewise, there are far too many examples of species-related differences to list here: this fact demands that there be a greater emphasis on human-derived preparations rather than those from other animals. Cultured human cells are not the best answer to this problem, given the many examples of phenotype transitions which occur in those preparations.
CONCLUSION
Perhaps the most fundamental conclusion that can be drawn is that a simplistic view of airway smooth muscle can no longer be held, i.e. that airway smooth muscle has: relatively uniform cytosolic ionic concentrations; homogeneous distributions of ion channels and pumps; simple fluxes of Ca2+; and smooth transitions from the resting state to the excited state. The more that is known about ionic mechanisms in airway smooth muscle, the more these are found to be exceedingly complex: nothing is static, very little is homogeneous. There are temporal and spatial heterogeneities in the intracellular concentrations of Ca2+, Na+ and Cl-, and perhaps also for other ions; even enzyme activities (e.g. Rho, Rho-associated kinase, protein kinase C, protein kinase A, adenylate and guanylate cyclases) are now found to exhibit marked spatial and temporal heterogeneities. There are oscillations in membrane potential and in the intracellular Ca2+ concentration. Not only can the Na+/Ca2+ exchanger operate normally to reduce the intracellular Ca2+ concentration, but it can also operate in the reverse mode to bring Ca2+ into the cell. Relaxants can simultaneously stimulate uptake of Ca2+ from the deep cytosol (via sarco-endoplasmic Ca2+ adenosine triphosphatase) and Ca2+ release into the subplasmalemmal space (via ryanodine receptor) for subsequent extrusion out of the cell (via plasma membrane Ca2+ adenosine triphosphatase). Perhaps more surprisingly, Ca2+ may also flow backwards through that same pathway, entering through various inward conductances (voltage-dependent Ca2+ channels, nonselective cation channels, Na+/Ca2+ exchanger) to accumulate in the subplasmalemmal space and then enter the sarcoplasmic reticulum (through ryanodine receptor), only to be re-released towards the contractile apparatus. Electrophysiologists are often focusing on the “wrong” end of the range of membrane potential: the important physiological events are not occurring at the positive potentials at which Ca2+ channels are maximally activated, nor at 0 mV at which we typically artificially set the equilibrium potential of Cl- and at which nonselective cation channels are obviated, but instead at ∼ -40 mV at which the physiological equilibrium potential of Cl-, inward nonselective cation channels, slow waves, window current and reverse-mode Na+/Ca2+ exchanger are found (fig. 2⇑).
- Received November 14, 2006.
- Accepted December 13, 2006.
- © ERS Journals Ltd
References
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵
- ↵