Molecular and Functional Significance of Ca 2+ -Activated Cl − Channels in Pulmonary Arterial Smooth Muscle

Abstract

Increased peripheral resistance of small distal pulmonary arteries is a hallmark signature of pulmonary hypertension (PH) and is believed to be the consequence of enhanced vasoconstriction to agonists, thickening of the arterial wall due to remodeling, and increased thrombosis. The elevation in arterial tone in PH is attributable, at least in part, to smooth muscle cells of PH patients being more depolarized and displaying higher intracellular Ca²⁺ levels than cells from normal subjects. It is now clear that downregulation of voltage-dependent K⁺ channels (e.g., Kv1.5) and increased expression and activity of voltage-dependent (Cav1.2) and voltage-independent (e.g., canonical and vanilloid transient receptor potential [TRPC and TRPV]) Ca²⁺ channels play an important role in the functional remodeling of pulmonary arteries in PH. This review focuses on an anion-permeable channel that is now considered a novel excitatory mechanism in the systemic and pulmonary circulations. It is permeable to Cl⁻ and is activated by a rise in intracellular Ca²⁺ concentration (Ca²⁺-activated Cl⁻ channel, or CaCC). The first section outlines the biophysical and pharmacological properties of the channel and ends with a description of the molecular candidate genes postulated to encode for CaCCs, with particular emphasis on the bestrophin and the newly discovered TMEM16 and anoctamin families of genes. The second section provides a review of the various sources of Ca²⁺ activating CaCCs, which include stimulation by mobilization from intracellular Ca²⁺ stores and Ca²⁺ entry through voltage-dependent and voltage-independent Ca²⁺ channels. The third and final section summarizes recent findings that suggest a potentially important role for CaCCs and the gene TMEM16A in PH.

Ca²⁺-activated Cl⁻ channels (Cl_Ca or CaCCs) are Cl⁻-permeable channels activated by a rise in intracellular Ca²⁺. They are widely distributed small-conductance channels located in the plasma membrane of many cell types and are hypothesized to serve important functions in sperm maturation, fluid transport across epithelial tissues, cardiac repolarization, olfactory transduction, photoreceptor stimulation, neuronal and skeletal muscle excitation, cardiac repolarization, and vascular smooth muscle excitability and contraction. For more information on the general properties and role of CaCCs in various cell types and tissues, the reader is invited to consult excellent reviews on this topic.^1–11

Byrne and Large ¹² were the first to report the existence of a Ca²⁺-activated Cl⁻ conductance in smooth muscle by describing its general properties in rat anococcygeus smooth muscle. Since this initial discovery, that group and others have described this conductance in a variety of smooth muscle cells, including vascular smooth muscle cells (VSMCs).^1,4,7 The identification in 2008 by three independent groups of the TMEM16, or anoctamin, family of genes as novel candidates for CaCCs^13–15 sparked a renewed interest in refining our understanding about their function in various tissues and whether they could serve as novel therapeutic targets. VSMCs, and pulmonary arterial smooth muscle cells (PASMCs) in particular, exhibit a significant “classical” CaCC conductance that now appears to be encoded by TMEM16A (anoctamin-1, or ANO1) and has been suggested to play a major role in vascular tone. After a description of the biophysical, regulation, and molecular properties of CaCCs in PASMCs and other vascular myocytes, this review focuses on the various sources of Ca²⁺ triggering CaCCs and how they might affect the control of membrane potential and pulmonary arterial tone under physiological and pathophysiological conditions.

BIOPHYSICAL PROPERTIES OF CACCS IN PASMCS

General macroscopic properties of Ca2+-activated Cl⁻ currents

CaCCs exhibit complex biophysical properties dictated by the interplay of Ca²⁺-, voltage-, and time-dependent gating mechanisms. Adding to this complexity are recent findings demonstrating that the channels in VSMCs are also regulated by phosphorylation involving both Ca²⁺-dependent and Ca²⁺-independent kinases and phosphatases. Initial studies describing the properties of CaCCs in smooth muscle and other cell types investigated with the whole-cell recording variant of the patch-clamp technique used transient sources of Ca²⁺ to activate the channels (e.g., Ca²⁺ entry via Ca²⁺ channels or Ca²⁺ release from internal stores). Although such methods suggested that Ca²⁺ plays an obligatory role in activating CaCCs, they provided little information on the biophysical properties of CaCCs, because only one of the three main variables controlling CaCC gating, voltage, was clamped, while internal Ca²⁺ levels and time-dependent kinetics were continuously changing. Setting intracellular [Ca²⁺] ([Ca²⁺]_i) to a known fixed value by using a Ca²⁺ chelating agent such as EGTA or BAPTA in the pipette solution allowed for studying Ca²⁺-activated Cl⁻ currents (I_Cl(Ca)) under well-controlled conditions, such that time-dependent gating mechanisms could be investigated in detail under voltage-clamp conditions.^7,16–18

Figure 1A shows a typical family of I_Cl(Ca) recorded in a rabbit PASMC dialyzed with a Cs⁺-containing pipette solution consisting of 10 mM BAPTA and 7.3 mM CaCl₂ to clamp [Ca²⁺]_i to 500 nM free Ca²⁺, with free [Mg²⁺] set to 1 mM. The elicited current reversed near the expected equilibrium potential for Cl⁻ of ˜0 mV (Fig. 1B) under the conditions of this experiment and displayed pronounced outward rectification (Fig. 1B), meaning that the underlying open channels carried more current in the outward direction (>0 mV; Cl⁻ influx) than in the inward direction (<0 mV; Cl⁻ efflux). In addition, the current exhibited typical time-dependent activation kinetics at positive membrane potentials and slow deactivation upon returning to the negative holding potential (Fig. 1A). Figures 1C and 1D illustrate another example, showing that the outward rectification is not a property of the pore but is rather dictated by voltage-dependent gating that results in channel closure at negative potentials. In this experiment, the rabbit PASMC was also dialyzed with a pipette solution containing 500 nM [Ca²⁺]_i. The protocol necessitated activating I_Cl(Ca) with an initial constant step to +100 mV from a holding potential of −50 mV, followed by a second step to a variable voltage to determine the magnitude and polarity of the I_Cl(Ca) tail current (Fig. 1C). The tail current measured immediately after repolarization reflects the behavior of the channel opened to a constant and stable open probability and thus that of the open pore. Plotting the magnitude of the tail current as a function of the voltage during the second step revealed a linear voltage relationship (Fig. 1D), as opposed to the strong outward rectification seen under near-steady-state conditions (Fig. 1B). This and many similar experiments by our group and other investigators confirmed that the open pore of CaCC in vascular myocytes conducts inward and outward currents equally.

Figure 1.

Typical macroscopic Ca²⁺-activated Cl⁻ currents (I_Cl(Ca)) recorded in a rabbit pulmonary artery smooth muscle cell with the whole-cell variant of the patch-clamp technique. The cell was dialyzed with a pipette solution containing CsCl and tetraethylammonium chloride with Ca²⁺ concentration set to 500 nM with the inclusion of the Ca²⁺ chelator BAPTA and an appropriate concentration of CaCl₂. ¹⁸ A, The top traces represent a representative family of I_Cl(Ca) elicited by the voltage-clamp protocol displayed below the traces. B, Current-voltage relationships for the currents in A measured immediately after the capacitative current surge (instantaneous I_Cl(Ca), labeled 1 in A; black squares) and at the end of 1-s pulses (late I_Cl(Ca), labeled 2 in A; red circles). The experiment shows the slight outward rectification of instantaneous I_Cl(Ca), while the quasi-steady-state I_Cl(Ca) displays pronounced outward rectification due to voltage-dependent gating. C, Double-pulse protocol examining the properties of the fully activated I_Cl(Ca). Currents (top traces) were evoked by the voltage-clamp protocol depicted below the traces. D, Amplitude of the tail current in C measured immediately after the capacitative currents during the second step (arrow in C) as a function of voltage, showing that the fully activated current is linear. This suggests that the open pore of the underlying Ca²⁺-activated Cl⁻channel carries inward and outward Cl⁻ currents equally well.

Unitary currents through native CaCCs

Single-channel experiments in smooth muscle cells (SMCs) from human mesenteric artery, ¹⁹ mouse, rat, and rabbit aorta,^20–22 and rabbit pulmonary artery^22–24 established that CaCCs are small-conductance anion channels with a single-channel conductance ranging from 1.2 to 3.5 pS. In agreement with studies describing macroscopic I_Cl(Ca), excised inside-out patch experiments in mouse aortic SMCs showed that single CaCCs are quickly activated by an elevation of Ca²⁺ on the cytoplasmic side of the membrane (Fig. 2A–2D). In addition, it was demonstrated in rabbit PASMCs that the unitary currents produced by these channels display ohmic behavior in the open state (Fig. 2E–2I), as well as voltage dependence (Fig. 2E, 2F, 2H). Another interesting property was the observation that single CaCCs display multiple conductance states, a behavior noted in both aortic^20,21 and pulmonary artery^23,24 SMCs (Fig. 2H–2J). The possible implication of this channel property for modes of activation by Ca²⁺ and voltage is discussed further below.

Figure 2.

Complex behavior of unitary currents through single Ca²⁺-activated Cl⁻ channels (CaCCs) recorded in mouse aortic and rabbit pulmonary artery myocytes with the inside-out configuration of the patch-clamp technique. A–D, Inside-out patch recordings showing the Ca²⁺ dependence of single CaCCs recorded from mouse aortic smooth muscle cells (SMCs). The bar above each plot illustrates the change in [Ca²⁺] in the bath. Each plot reflects open probability (NPo) as a function of time after patch excision (indicated by a vertical bar labeled “Excise”). Below each NPo-versus-time plot are selected segments of original single-channel recording traces. E–J, Inside-out patch recordings from rabbit pulmonary artery SMCs (PASMCs), illustrating the Ca²⁺ and voltage dependence of single CaCCs as well as the existence of multiple subconductance states. E–H show sample recordings obtained at −100 mV (bottom) and +100 mV (top). E, F, and H were respectively obtained with 50 nM, 250 nM, and 1 μM Ca²⁺ in the bath. G and I show expanded traces from F and H, respectively, demarcated by horizontal bracketed bars. J, Mean current-voltage relationships of unitary CaCCs for the fully conducting and two subconductance states. A–D were reproduced from Hirakawa et al. ²¹ with permission from the American Physiological Society. E–J were reproduced from Piper and Large ²³ with permission from the Physiological Society.

Selectivity of CaCCs

Replacement of Na⁺, the main cation in the extracellular medium, with less permeable cations such as N-methyl-D-glucamine had little effect on the magnitude and reversal potential of macroscopic or unitary I_Cl(Ca) in vascular myocytes,^1,7,19–21 indicating that CaCCs are much more permeable to anions than to cations. Anion replacement experiments showed that I_Cl(Ca) in VSMCs exhibit an anion permeability sequence consistent with permeation driven primarily by the hydration energy of individual ions in a conduction pathway that exhibits a low electric field strength, with SCN⁻ ≫ I⁻ > NO₃⁻ > Br⁻ > Cl⁻ ≫ aspartate ≅ glutamate.^{1,4,7,18,23,25–27}

Kinetic models predicting the Ca2+ and voltage dependence of native CaCCs

Quantitative analysis of the Ca²⁺ and voltage dependence of macroscopic and unitary I_Cl(Ca) in parotid acinar cells, ¹⁶ Xenopus oocytes, ¹⁷ and VSMCs^23,28 showed that the affinity of the Ca²⁺ binding site located on the internal side of the membrane is voltage dependent, with membrane depolarization increasing the efficacy of Ca²⁺ in activating the channel. On the basis of a Hill coefficient consistently higher than 1, the above studies also proposed that a minimum of 2 Ca²⁺ are required to cooperatively open the channel. Arreola et al. ¹⁶ were the first to propose a kinetic model describing the Ca²⁺ and voltage dependence of CaCCs recorded in parotid acinar cells:

$C_1\underset{{\beta }_1}{\overset{{\alpha }_{1}Ca}{\rightleftarrows }}{C}_2^{.Ca}\underset{{\beta }_1}{\overset{{\alpha }_{1}Ca}{\rightleftarrows }}{C}_3:\begin{array}{c}Ca\\ Ca\end{array}\underset{{\beta }_2}{\overset{{\alpha }_2}{\rightleftarrows }}O.$

In this 4-state model, 2 calcium ions bind sequentially, which then leads to channel opening from a closed, fully bound state. The Ca²⁺-binding steps are defined by two reversible equilibrium reactions, each defined by an equilibrium constant K₁ that is equal to β₁/α₁, where α₁ is the first-order binding-rate constant and β₁ is the unbinding-rate constant. The equilibrium transition from C₃.2Ca to O was defined by the equilibrium constant K₂ being equal to β₂/α₂, where α₂ is the opening-rate constant and β₂ is the closing-rate constant. In this scheme, both K₁ and K₂ are sensitive to voltage, each defined by a monotonic exponential function. The direct implication of this model is that Ca²⁺ binding is voltage dependent and therefore that the binding sites would lie within the transmembrane electric field (˜13% distance from the inside when the Wood-hull model ¹⁶ is applied). Although the kinetic model reasonably predicted the steady-state properties and kinetics of I_Cl(Ca) at low to intermediate [Ca²⁺]_i (100–250 nM), the voltage dependence of the kinetics of the simulated currents significantly deviated from actual data at higher [Ca²⁺]_i (1,000 nM).

Kuruma and Hartzell ¹⁷ performed a similar biophysical analysis of I_Cl(Ca) in Xenopus oocytes and proposed a radically different gating scheme based on an important observation also described by our group for I_Cl(Ca) in PASMCs. ²⁸ Figure 3 illustrates this observation, using the same data from the typical experiments shown in Figure 1. The two families of traces from Figure 1A and 1C were reproduced in Figure 3, showing only areas of interest. Time-dependent I_Cl(Ca) elicited by depolarizing steps more positive than +30 mV (right-hand set of traces) and deactivating I_Cl(Ca) tail currents (left-hand set of traces) were fitted with a single exponential function overlaid on the traces as a red or blue solid line. The estimated time constants of activation (τ_on; blue) and deactivation (τ_off; red) were plotted as functions of voltage in the plot below the traces. The plot clearly shows that τ_on is not particularly sensitive to voltage, whereas τ_off declines exponentially with membrane hyperpolarization. On the basis of very similar data and in contrast to the model proposed by Arreola et al., ¹⁶ Kuruma and Hartzell ¹⁷ reasoned that voltage-dependent gating of CaCCs must involve a mechanism whereby channel closure, rather than channel opening, confers voltage sensitivity, and they proposed the following kinetic model:

Figure 3.

Voltage dependence of activation and deactivation kinetics of whole-cell Ca²⁺-activated Cl⁻ currents, recorded from a rabbit pulmonary artery myocytes. The traces above the plot were reproduced from those shown in Figure 1. The corresponding voltage step protocols that elicited these currents are shown below the traces. Superimposed on the black original current traces are single exponential fits to the data delimited by the arrows. The plot below the traces shows the derived time constants of activation (τ_on; blue) and deactivation (τ_off; red) plotted as functions of voltage. The line passing through the red data points represents a single exponential fit, whereas the line passing through the blue data points represents a linear fit whose slope is not significantly different from 0 (P > 0.05).

This 7-state model comprises 3 Ca²⁺ binding steps, each of which can lead to independent channel opening. While the binding-rate constants k₁, k₂, and k₃ and the unbinding-rate constants k_–1 k_–2, and k_–3 were set to identical values, the opening-rate constants α₁, α₂, and α₃ were given progressively larger values. In this model, voltage sensitivity is governed by the closure-rate constants β(V)₁, β(V)₂, and β(V)₃, each defined by a specific exponential function related to voltage. This model predicted well the Ca²⁺ and voltage dependence and accounted for most kinetic aspects of I_Cl(Ca) in Xenopus oocytes. ¹⁷ Moreover, a separate analysis in rabbit pulmonary artery (PA) myocytes conformed to this model. ²⁸ In their detailed analysis of the gating properties of single CaCCs recorded from rabbit PASMCs, Piper and Large ²³ identified at least 3 subconductance state levels, ranging from ˜1.2 to 3.5 pS. Interestingly, the probability of detecting the lower subconductance states of 1.2 and 1.8 pS was high when the internal face of the membrane was exposed to 1 μM Ca²⁺, with no evidence of the fully conducting state at 3.5 pS. In contrast, only the fully conducting state of ˜3.5 pS and a subconductance level of ˜1.8 pS were apparent with 50 nM Ca²⁺ on the cytoplasmic side. Analysis of single-channel behavior with 250 or 500 nM Ca²⁺ revealed intermediate behaviors. Piper and Large ²³ speculated that the multiple subconductance states might correspond to the 3 open states in the kinetic model proposed by Kuruma and Hartzell. ¹⁷ Sequential binding of multiple calcium ions would not only enhance the open probability of the channel but would also partially obstruct the permeation pathway leading to subconductance state levels.

Pharmacology of CaCCs in vascular myocytes

The stilbene derivatives 4,4′-diisothiocyano-2,2′-stilbene (DIDS) and 4-acetamido-4′-isothiocyano-2,2′-stilbene (SITS), fenamates such as niflumic (NFA) and flufenamic acids, anthracene-9-carboxylic acid (A9C), diphenylamine-2-carboxylate, 5-Nitro-2-(3-phenylpropylamino) benzoic acid (NPPB), ethacrynic acid, frusemide, and indanyloxyacetic acid 94 (IAA-94) all inhibit CaCCs in VSMCs. However, the common denominator is that they inhibit the channels with low affinity (IC₅₀ in the range of ˜10⁻⁶ to 10⁻³ M) and with relatively poor selectivity in discriminating the contribution of CaCCs from that of other types of Cl⁻ channels (e.g., volume- or swelling-activated Cl⁻ channels [I_Cl-Swell]), and many of these compounds also target cation channels. There are a number of excellent reviews and original articles that summarize the older literature on this topic.^{1,4,6,7,29–31}

Of this “first generation” of CaCC inhibitors, NFA is the most potent blocker of these channels and the compound most frequently used to investigate the physiological role of CaCCs on vascular smooth muscle tone.^1,4,7 In portal vein SMCs, NFA was reasonably more effective at blocking I_Cl(Ca) than I_Cl-Swell. ³² The block produced by NFA is moderately voltage dependent, with membrane depolarization increasing its potency, possibly because of a preferential interaction of NFA with the open state. In several types of VSMCs, including PASMCs, NFA slows the deactivation of CaCCs at negative potentials,^{24,30,33–37} an observation in agreement with the idea that NFA preferentially binds the channel in the open state and remains bound to the channel for some time, delaying channel closure. ¹ Except for one report in rat aorta, ³⁸ NFA (≤100 μM) did not attenuate the dihydropyridine-sensitive contraction produced by KCl-induced depolarization,^39–43 nor did it inhibit voltage-gated Ca²⁺ channels directly.^34,41,44,45 However, NFA and other fenamates exert nonspecific effects that bear major implications for interpretation of physiological data, depending on the cell type in which they are tested. NFA is a nonsteroidal anti-inflammatory agent that could indirectly influence ion channels and vascular tone by altering the release of vasoactive by-products of arachidonic metabolism (e.g., prostaglandins) by the cyclo-oxygenase pathway. Fenamates were shown to stimulate the human slow delayed rectifier K⁺ channel (I_sK) expressed in Xenopus oocytes, ⁴⁶ a delayed rectifier K⁺ current in canine jejunal SMCs, ⁴⁷ and large-conductance Ca²⁺-activated K⁺ channels (BK_Ca) in VSMCs.^33,35,48 One study also reported the inhibition of a nonselective cation current in rat exocrine pancreatic cells. ⁴⁹ Cruickshank et al. ⁵⁰ demonstrated that NFA elicits Ca²⁺ release from internal stores in rat PASMCs, but this was not confirmed in two independent studies in the same preparation.^41,43 Kato et al. ⁵¹ described experiments showing that the relaxation by NFA or NPPB of rat pulmonary arteries precontracted with endothelin-1 (ET-1) was independent of their ability to inhibit CaCCs. Finally, two studies from our group reported that NFA exerts a paradoxical inhibition and stimulation of CaCCs in rabbit pulmonary and coronary artery myocytes. In these studies, the stimulation was more easily visualized after a rapid washout of the drug, although in PASMCs stimulation of I_Cl(Ca) was clearly evident at negative potentials in the presence of NFA. These experiments suggested the existence of multiple binding sites for NFA affecting both the pore and gating of the channel. A9C was also shown to produce peculiar bimodal effects on I_Cl(Ca) in rabbit PASMCs, inhibiting the steady-state current in a voltage-dependent manner while potently stimulating the inward tail current during repolarization. Interestingly, NFA and A9C produced dual effects on expressed TMEM16A CaCCs, ⁵² the gene speculated to encode for CaCCs in VSMCs,^53–55 very similar to those reported for I_Cl(Ca) in vascular myocytes. Taken together, this explains why investigators have to be extremely cautious in their interpretation of the effects of NFA and all other CaCC blockers when examining their effects on membrane potential, [Ca⁺⁺]_i, and force in intact arteries.

The recent discovery of the TMEM16, or anoctamin, family of CaCCs by three independent laboratories,^13–15 combined with the development of high-throughput screening assays, allowed for the identification of novel CaCC modulators displaying greater potency and specificity. Screening of medium- to large-scale synthetic and natural compound libraries led to the discovery of small-molecule inhibitors^56,57 and activators ⁵⁸ of CaCCs. Among the identified inhibitors, T16A_Inh-A01 and CACC_Inh-A01 were found to be the most potent and selective substances, with IC₅₀ ≤ 10 μM^52,56,57 and a block that was insensitive to membrane potential, ⁵² a useful property for functional studies in cells and tissues. T16A_Inh-A01 relaxed precontracted arteries, in line with a role in vascular contraction, as outlined below. Of the activators discovered, E_act (activator) and F_act (potentiator) were shown to stimulate native and ANO1-induced CaCC activity (EC₅₀ ≈ 3–6 μM) by respectively activating the channels in the absence of intracellular Ca²⁺ or enhancing their Ca²⁺ sensitivity. Tannins, which are particularly concentrated in green tea and red wine, were identified as good inhibitors of MEM16A or ANO1, ⁵⁹ tannic acid being the most potent tannin, with an IC₅₀ of ˜6 μM. Recently, these “second-generation” CaCC modulators were successfully used in physiological experiments examining the potential role of CaCCs on agonist-mediated tone, and they are discussed below.

Regulation of CaCCs in vascular myocytes

One important observation made at an early stage was that CaCCs in many, but not all, cell types display significant loss of activity or rundown after patch excision in the inside-out configuration or after achieving whole-cell recording mode. This led to the speculation that CaCCs may be subjected to tight regulation by various mechanisms, including phosphorylation. Wang and Kotlikoff ⁶⁰ were the first to demonstrate that CaCCs are inactivated by Ca²⁺/calmodulin-dependent protein kinase II (CaMKII)-mediated phosphorylation. These authors convincingly showed that in tracheal SMCs, caffeine-induced CaCC current declined more quickly than the underlying Ca²⁺ transient (measured with Fura-2/AM), an effect attributed to CaMKII-induced phosphorylation of CaCC or an associated regulatory subunit. A study from our group similarly indicated that CaCC activity is also downregulated by CaMKII-induced phosphorylation in rabbit PASMCs and coronary artery SMCs, through a process involving a decrease in voltage sensitivity. ¹⁸ Subsequent reports showed that the effect of CaMKII on CaCCs is antagonized by the serine-threonine phosphatase calcineurin (CaN) in rabbit coronary ⁶¹ and pulmonary ²⁷ artery myocytes. More recently, our group showed that in rabbit PASMCs, I_Cl(Ca) are also regulated by the Ca²⁺-independent protein phosphatases 1 (PP1) and 2A (PP2A). ⁶² Ayon et al. ⁶² proposed that CaN Aα and, most likely, PP1 act in a common upstream-downstream signaling pathway to regulate CaCCs. Clearly, there are many missing pieces to this puzzle. It is unknown whether CaMKII, CaN, and PP1/PP2A regulate phosphotransferase activity directly on the pore-forming subunit of the CaCC or on an accessory subunit. Indirect evidence suggests that CaMKII is likely not the only kinase involved. In rabbit PASMCs, I_Cl(Ca) were shown to run down ˜70%–80% within 5 minutes of cell dialysis with 3–5 mM adenosine triphosphate (ATP), and this rundown was obliterated by replacing internal ATP with the nonhydrolyzable ATP analog adenylyl imidodiphosphate (AMP-PNP), by omitting ATP^28,62,63 or by dialyzing the cells with excess CaN Aα. ²⁷ In contrast, specific inhibition of CaMKII by KN-93 increased I_Cl(Ca) that had already run down to a steady-state level only by a factor of <2. ¹⁸ Phosphorylation was suggested to inhibit I_Cl(Ca) by promoting state-dependent channel closure, by shifting the steady-state voltage dependence toward positive potentials. ²⁸ Although the physiological significance of the negative regulation by phosphorylation of CaCCs in PA myocytes remains unclear, we may speculate that because the anion channels offer a powerful depolarizing stimulus yielding sustained Ca²⁺ influx through L-type Ca²⁺ channels (Ca_L) and therefore reinforcement of CaCC activation, Ca²⁺-dependent downregulation by phosphorylation may provide a means to attenuate the impact of the positive-feedback loop between Ca_L and CaCCs. The phosphorylation-mediated CaCC inactivation was also shown to reduce the affinity of NFA to inhibit I_Cl(Ca) in PA myocytes, an observation consistent with the known open-state interaction of this compound with the channel, ⁶³ which may explain, at least in part, the discrepancy in IC₅₀ values measured in different VSMCs by different groups. ³⁰

Very recently, our group reported that phosphatidylinositol 4,5-bisphosphate, or PIP₂, inhibits native I_Cl(Ca) in rat PASMCs and TMEM16A-encoded I_Cl(Ca). ⁶⁴ Although the physiological relevance of this important observation remains to be determined, it highlights another potential mechanism of activation of the channel after stimulation of receptors coupled to phospholipase C (PLC)–mediated membrane phospholipid breakdown.

Molecular candidates proposed to encode for CaCCs in VSMCs

For several decades after the initial discovery of CACCs, research efforts to determine their physiological role in various tissues was hampered by the need to use pharmacological agents exhibiting poor potency and selectivity, by the existence of several types of CaCCs displaying distinct biophysical properties and modes of activation by intracellular Ca²⁺, and, more significantly, by the lack of a valid molecular candidate. Over the past 20 years, 5 structurally unrelated families of Cl⁻ channel genes have been postulated to form the molecular basis of CaCCs: the CLCA family;^65–69 the long human isoform variant of CLC-3, a member of the voltage-gated Cl⁻ channel superfamily of Cl⁻ genes that requires CaMKII for activation;^6,70,71 the human gene related to the Drosophila flightless locus called tweety;^72,73 the vitelliform macular dystrophy protein bestrophin;^74,75 and the recently identified TMEM16 or anoctamin gene family.^13–15,76 Among these genes, the last two families have recently emerged as better molecular candidates in many cell types.

Bestrophin was identified as the gene product of the vitelliform macular dystrophy type 2 gene (VMD2). ⁷⁵ Several mutations of this gene were associated with the vision disorder vitelliform macular dystrophy, or Best's disease, an autosomal dominant abnormality characterized by an accumulation of lipofuscin-like material in the macular area of the eye. In 2002, Sun et al. ⁷⁴ were the first to report that the heterologous expression of the several highly conserved orthologs and paralogs of bestrophin genes in human (hBest1, hBest2), Caenorhabditis elegans (ceBest1), and Drosophila (dmBest1) produced I_Cl(Ca). Subsequent studies revealed the existence of 4 bestrophin genes in humans (hBest1–hBest4) ⁷⁷ and 3 in mouse (where Best4 is a pseudogene). ⁷⁸ Conserved orthologs have also been found in Xenopus. ⁷⁹ Bestrophins are ubiquitously expressed in excitable tissues and various epithelia. ⁷⁵ The membrane topology of bestrophins is currently unclear, as 4^74,80,81 or 6^79,82 putative hydrophobic transmembrane-spanning domains (TMDs) have been suggested on the basis of hydropathy plot analysis. Moreover, Tsunenari et al. ⁸³ performed a detailed structural analysis of hBest1 by site-mapping and cysteine-scanning experiments and alternatively proposed that the channel comprises 4 putative transmembrane domains (TMD1, TMD2, TMD4, and TMD6) and an additional hydrophobic loop (TMD5) partially reentering the membrane from the extracellular side. Finally, Milenkovic et al. ⁸¹ suggested an alternative membrane topology for hBest1 whereby TMD1, TMD2, TMD5, and TMD6 would cross the membrane and TMD3 and TMD4 would be part of a long intracellular hydrophobic loop between TMD2 and TMD5. All proposed structural models predict that the N- and C-terminal ends lie in the cytoplasm. ⁷⁵

Although currents elicited by the expression of different bestrophin genes displayed variable rectification, with the exception of ceBest1 nearly all bestrophin products were shown to lack voltage-and time-dependent gating properties.^74,75 Similar to native CaCCs in VSMCs, bestrophins are activated by physiological intracellular Ca²⁺ levels with an apparent dissociation constant K_d ≈ 200–400 nM, exhibit an anion permeability sequence of I⁻ > Br⁻ > Cl⁻ > F⁻, and are inhibited by classical CaCC blockers such as DIDS, SITS, and NFA.^{74,75,79,84–86} A detailed mutational analysis of hBest1 identified a potential EF-hand Ca²⁺-sensing domain in the C-terminus of the protein that is analogous to those found on calmodulin (CaM). ⁸⁷ The single-channel conductance of endogenous and expressed bestrophins was reported to be similar to CaCCs in vascular myocytes^{19–21,23,24} laying in the range of 0.26 to ˜2 pS.^88,89

In 2008, three groups of investigators independently showed that two members of the TMEM16 gene family, TMEM16A and TMEM16B, or anoctamin-1 and anoctamin-2 (ANO1, ANO2), encoded for CaCCs.^{8,13–15,76,90} ANO1 and ANO2 belong to a mammalian family of genes comprising 10 paralogs. Of these 10 members, only ANO1, ANO2, and ANO6 (TMEM16F) are Ca²⁺-activated ion channels. Although ANO1 and ANO2 are confirmed CaCCs, ANO6 was described as encoding for a CaCC^91,92 or a nonselective Ca²⁺-activated cation channel permeable to Ca²⁺. ⁹³ ANO3, ANO4, ANO6, ANO7, and ANO10 have Ca²⁺-dependent phospholipid scramblase activity. ⁹⁴

Importantly, the biophysical properties and pharmacology of expressed ANO1 and ANO2 are very similar to those of CaCC currents recorded from native cells: (1) activation by physiological levels of intracellular Ca²⁺ (˜200–1,000 nM), (2) voltage and time dependence and outward rectification due to gating and an apparent K_d for Ca²⁺ that is voltage sensitive, (3) sensitivity to blocking by NFA, DIDS, and NPPB, (4) a relatively small single-channel conductance (8.3 pS, compared to ˜1–3 pS for native channels in VSMCs), and (5) an anion permeability sequence of SCN⁻ > NO₃⁻ > I⁻ > Br⁻ > Cl⁻ > F⁻ > gluconate.^{13–15,53,54,59,95} ANO1 messenger RNA (mRNA) transcripts were shown to be ubiquitously expressed in cell types previously reported to exhibit CaCC activity, such as exocrine secretory epithelia (e.g., pancreatic acinar cells, salivary and parotid gland cells), several types of retinal cells, sensory neurons, liver, lung, placenta, trachea, vomeronasal organ, uterus, interstitial cells of Cajal in the gastrointestinal tract, sperm cells, heart, skeletal, and smooth muscles.^{13–15,76,90,96–102} ANO1 is robustly expressed at the mRNA and protein levels in VSMCs, ⁵³ and its knockdown by silencing RNA (siRNA) resulted in a reduced CaCC conductance in rat pulmonary ⁵⁴ and cerebral artery ⁵⁵ myocytes. Besides ANO1, high transcript levels were found for ANO6 and ANO10 in rat pulmonary arteries, ⁵⁴ but their role in this tissue remains undefined.

Anoctamins are predicted to exhibit 8 transmembrane domains with both N- and C-termini facing the cytoplasm (Fig. 4A).^13,15,76,96 In this model, a portion of the large extracellular loop between TMD5 and TMD6 is proposed to reinsert in the membrane to form the pore, or P loop, of the channel. Mutations of several basic amino acid residues in this region altered the conductance and selectivity of the induced current ¹⁵ (reviewed by Hartzell et al. ⁷⁶ ). However, this topology was recently challenged on the basis of site mapping by site-directed mutagenesis and accessibility of antibodies and sulfhydryl reagents on specific residues or domains of the protein. ¹⁰³ In this revised model (Fig. 4B), the stretch of residues thought to form an extracellular loop between TMD5 and TMD6 in the original model would instead be located in the cytoplasm, with the pore domain lying between TMD6 and TMD7. It also unveiled a new potential Ca²⁺ binding site, which is discussed further below. As for the quaternary structure of the protein, two recent studies provided convincing evidence for ANO1 forming a stable homodimer whose assembly is maintained by noncovalent interactions.^108,109 Tien et al. ¹¹⁰ confirmed this hypothesis and identified a stretch of 19 amino acid residues in the N-terminal domain located upstream of the first transmembrane domain that are critically involved in homodimerization of the protein and its functional expression. There is also evidence for the speculated pore loop region of ANO1 to play an important role in the trafficking of the channel to the plasma membrane and determining channel density. ¹¹¹

Figure 4.

Proposed secondary structures for anoctamin-1 (ANO1), or TMEM16A. A, Original model, based on predicted hydropathy profiles, describing the basic membrane topology of the protein, which comprises 8 transmembrane domains (TMDs), C- and N-terminal ends located in the cytoplasm, and a pore loop located between TMD5 and TMD6.^13,15 B, Revised model proposing that the extracellular loop following TMD5 in the original model may instead face the cytoplasm. ¹⁰³ C, More detailed map of the revised membrane topology model of ANO1, revealing the locations and amino acid sequences of the 4 known spliced variants (a–d), the identification of 2 glutamate residues at positions 702 and 705 in mouse as potential Ca²⁺ binding sites, ¹⁰³ the location and hypothetical function of 3 classes of calmodulin (CaM)-binding domains (CaM-BD1 and CaM-BD2, ¹⁰⁴ CMB1 and CMB2, ¹⁰⁵ and regulatory CaM-binding motif [RCMB] ¹⁰⁶ ), and at least one N-glycosylation site (N-Glycos). The structure also exposes the location of 4 consecutive glutamate residues immediately proximal to spliced variant c, which were shown to be critical for Ca²⁺- and voltage-dependent gating of the channel. ¹⁰⁷

At least 4 exons of ANO1, but likely more, ¹¹² are alternatively spliced (identified as “a,” “b,” “c,” and “d“), producing different protein variants (Fig. 4C)^13,76,98 that bear major significance for expression and channel function. ¹³ Spliced variants b and d are alternatively spliced in pulmonary arterial smooth muscle, whereas spliced variant c is constitutively expressed.^53,54,113 The a variant, which initiates the protein at the N-terminal end of the protein, is generally thought to be constitutively expressed; however, a minimal variant lacking this spliced exon was found in some tissues. It is also known that this particular exon is under the control of an alternative promoter. Ohshiro et al. ¹¹⁴ recently confirmed that murine portal vein myocytes express heterodimers composed of distinct ANO1 spliced variants, further extending the level of functional complexity of this protein. A “minimal” isoform of ANO1 lacking all 4 spliced exons retained normal Ca²⁺ sensitivity but lacked voltage and time dependence. ¹¹⁵ However, a recent study by the same group challenged that initial assertion by demonstrating that protein translation may be initiated at a noncanonical starting codon (non-ATG), which suggests that a significant fraction of the N-terminus is important for protein trafficking and insertion into the membrane as well as for channel function. ¹¹⁶ A detailed analysis of spliced variants b (exon 6b in mouse; 22 amino acids) and c (exon 13 in mouse; 4 amino acids) suggested that the expression of the former reduces Ca²⁺ sensitivity by ˜4–fold at +80 mV, whereas omission of variant c (EAVK) attenuated the time dependence seen at positive potential, a process attributed to voltage-dependent gating. ⁹⁵ Another group interested in identifying the protein domains involved in the Ca²⁺ and voltage dependence of TMEM16A (see text below) revisited this question by analyzing the stretch of amino acids preceding and spanning variant c. ¹⁰⁷ While initial reports showed that spliced variant d (exon 15 in mouse; 26 amino acids) produced no detectable effect on the activity of expressed ANO1,^13,95 a recent study showed that its inclusion slowed time-dependent activation and deactivation kinetics. ¹¹⁷ Taken together, these observations highlight the profound influence of alternative splicing of ANO1 in different cell types and how the controlled expression of these variants may result in fine-tuning the activity of CaCCs and their role in cellular functions.

Heterologous expression studies of ANO1 have confirmed that channel gating is physiologically triggered by an elevation of Ca²⁺ on the cytoplasmic side, although the channel can also be opened by extreme positive potentials under conditions of strong Ca²⁺ buffering.^106,107 Other divalent cations, such as Ba²⁺,^107,118 Sr²⁺,^118,119 and Ni²⁺, ¹¹⁹ can also activate ANO1, albeit with a lower affinity (Ca²⁺ > Sr²⁺ ≈ Ni²⁺ > Ba²⁺); Mg²⁺, which competes with Ca²⁺, does not activate the channel, ¹¹⁸ whereas Zn²⁺ blocks the channel. ¹¹⁹ Whether ANO1 is directly gated by intracellular Ca²⁺ or has an obligatory requirement for an accessory Ca²⁺-binding protein, such as CaM or another unknown protein, is controversial. Neither a canonical EF-hand Ca²⁺-binding motif nor a CaM-binding domain (IQ) exists in ANO1. Analysis of the amino acid sequence revealed the existence of a stretch of 5 consecutive glutamate residues in the first intracellular loop between TMD1 and TMD2 (Fig. 4C). ¹⁰⁷ These residues were initially hypothesized to serve as a potential Ca²⁺-sensing domain, or “Ca²⁺ bowl,” by analogy to a series of aspartate residues identified in BK_Ca. ¹²⁰ The fifth glutamate residue is actually encoded by the first 3 base pairs of exon 13 in the mouse, a very short alternatively spliced exon (12 base pairs encoding for the c variant EAVK) that produces significant changes in the biophysical properties of the protein. Deletion of EAVK led to a profound decrease in the Ca²⁺ sensitivity of human ANO1, whereas mutating the 4 glutamate residues preceding EAVK to alanine residues produced anion currents lacking intrinsic voltage dependence. The same group also showed that voltage-dependent gating of ANO1 was allosterically facilitated by anions displaying a higher permeability than Cl⁻ (SCN⁻, NO₃⁻) or by high external Cl⁻concentration. ¹⁰⁷ Although this region is important in determining the Ca²⁺ and voltage dependence of ANO1, it is likely not the Ca²⁺ binding and trigger site for opening the channel. The most compelling evidence for the existence of a novel Ca²⁺-sensing domain came from a study by Yu et al. ¹⁰³ In addition to proposing a revised model of the secondary structure of ANO1, they showed that 2 glutamate residues at positions 702 and 705 of mouse ANO1 (Fig. 4C) were important for Ca²⁺ activation. Mutating both residues to glutamine led to a reduction in Ca²⁺ sensitivity of ˜2 orders of magnitude. In support of direct activation by Ca²⁺ is the recent demonstration that normal CaCC activity could be elicited by Ca²⁺ applied on the cytoplasmic side of purified ANO1 reconstituted in liposomes and that this process was lost by mutation of the same glutamate residues to glutamine, as discussed above, ¹²¹ leading these investigators to suggest that ANO1 is necessary and sufficient to recapitulate the biophysical properties of CaCCs.

Tian et al. ¹⁰⁴ proposed an alternative and indirect mechanism of activation of TMEM16A by internal Ca²⁺. In this model, which is analogous to small-conductance Ca²⁺-activated K⁺ channels, ¹²⁰ CaM physically interacts with ANO1, and this interaction is indispensable for channel activation in the presence of Ca²⁺. One problem with this finding is that the first of 2 putative CaM-binding domains (CaM-BD1) identified by bioinformatics in the N-terminus of ANO1 and shown to bind CaM overlaps spliced variant b (Fig. 4C). Yet robust Ca²⁺- and voltage-sensitive I_Cl(Ca) can be elicited by expression of TMEM16A lacking spliced variant b (Fig. 1).^{53,95,105–107,122} Moreover, inclusion of spliced variant b reduced the Ca²⁺ sensitivity of ANO1. ⁹⁵ In contrast, CaM-BD2 did not bind CaM and was found to play no role in channel activity. ¹⁰⁴ A potential role for activation of ANO1 by a phosphorylation step involving CaMKII or other kinases was also discarded,^104,122 although one study showed that ATP promoted channel activation by an unknown mechanism. ¹⁰⁴

More recently, Vocke et al. ¹⁰⁶ identified a different regulatory CaM-binding motif (RCMB) in the N-terminus of ANO1 and ANO2 (Fig. 4C) that is distinct from CaM-BD1 and CaM-BD2. They suggested that RCMB is involved in activation of ANO1 and ANO2 by Ca²⁺ and in Ca²⁺-dependent inactivation of ANO2. Finally, a separate report published in 2013 suggested that CaM does not activate ANO1 but does regulate it by increasing its HCO₃⁻ permeability relative to Cl⁻, ¹⁰⁵ a mechanism suggested to be important in fluid secretion in submandibular acinar gland cells. In the latter study, specific peptides corresponding to 2 regions, one located immediately proximal to the first TMD on the N-terminal segment (CBM1) and the other within the intracellular loop between TMD6 and TMD7 (CBM2), were able to reduce the association of CaM to ANO1 and attenuate its effect on HCO₃⁻ permeability (Fig. 4C). However, this hypothesis will require further investigation, as a preliminary report from another group recently showed that CaM did not alter anion permeability under well-controlled voltage-clamp conditions. ¹²³ The potential role of CaM in activation of ANO1 was revisited recently by Yu et al., ¹²² who presented an elegant series of experiments arguing against a prerequisite role of CaM for activation of ANO1 by Ca²⁺. Although this group found that CaM could bind to ANO1, the interaction was very weak. Among several convincing arguments presented supporting direct activation of ANO1 by Ca²⁺ was the fact that Ba²⁺, which is unable to bind to CaM and trigger a conformational change, ¹²⁴ was able to activate ANO1, a result consistent with previous findings.^107,118 Taken together, the above studies indicate that we still have a poor understanding of the molecular mechanisms involved in the activation of ANO1 by Ca²⁺ and voltage and of how phosphorylation regulates its activity. Future studies providing structural information will help clarify the mechanisms involved in gating.

Physiological role of bestrophin and anoctamin CaCCs in PASMCs

Very little is known about the functional role of bestrophins in VSMCs. Best1, Best2, and Best3 (Best4 is a pseudogene in mouse but not in human ⁷⁵ ) are expressed in the A7r5 SMC line; rat aorta, mesenteric arteries, and PASMCs; ¹²⁵ and the rabbit PA. ⁷ The I_Cl(Ca) recorded from PASMCs dialyzed with a solution designed to clamp internal Ca²⁺ at a fixed elevated level (e.g., 500 nM) exhibit both instantaneous and time-dependent components following voltage-clamp steps to positive potentials (Fig. 1). It is generally interpreted that the instantaneous jump in membrane current reflects the activity of channels that are readily available at the holding potential before a depolarizing step eliciting the instantaneous current and that the channels responsible for such behavior correspond to a single population of channels.^{7,17,18,28,61} However, it is possible that a portion of the instantaneous component might be attributed, at least in part, to one or more bestrophin gene products, since the latter are voltage and time independent. Similar to bestrophin-induced currents, the current-voltage relationship of the instantaneous component of I_Cl(Ca) in PA myocytes is almost linear and is less sensitive to block by the putative CaCC inhibitor NFA. ¹⁸

Two independent groups identified a Ca²⁺-activated Cl⁻ conductance activated by the second messenger cyclic guanosine monophosphate and Ca²⁺ (I_cGMP.Ca) in rat mesenteric arterial myocytes.^126–128 Subsequent studies showed that Best3 encodes for this conductance and appears to play a major role in the rhythmic vasomotion but not in the tonic contraction of mesenteric arteries induced by vasoconstricting agonists.^125,129 Interestingly, I_cGMP.Ca is time and voltage independent and is relatively insensitive to NFA but can be blocked by Zn²⁺ (10 μM) applied externally.^125–127 The potential role of Best3 in the pulmonary arterial circulation is unclear, as the expression of the protein is low. ¹²⁵

Because of similarities with I_Cl(Ca) in vascular myocytes and other cell types expressing CaCCs, anoctamins rapidly became the more promising candidate gene family suspected to encode for the still elusive native channel. Of the 10 anoctamin paralogs, ANO1 and ANO2 are confirmed CaCCs, while the evidence that the other family members display Ca²⁺-sensitive Cl⁻ channel activity is unclear.^99,130 ANO3–ANO7 appear to be distributed in the cytoplasm, most likely in the endoplasmic reticulum (ER). ¹³⁰

Anoctamins are expressed at significant levels in several types of smooth muscle, including the intestinal tract,^{97,98,131,132} the urethra, ¹⁰¹ the oviduct, ¹³³ airways,^102,134,135 and vasculature.^{22,43,53–55,113,114,136–140} The role of CaCCs and anoctamins in determining pulmonary arterial tone is still to be evaluated. While ANO2 knockout mice develop and breed normally, exhibiting an apparently normal phenotype, ¹⁴¹ ANO1 knockout mice die prematurely after birth because of malformation of their airways during embryonic development. ¹⁴² This poses a serious challenge in attempting to determine the role of ANO1 in electromechanical coupling of blood vessels from immature animals, especially in resistance vessels, which will require the development of conditional and inducible genetically modified animal models. Very recently, Heinze et al. ¹⁴⁰ presented convincing data obtained with a novel conditional smooth muscle–specific and inducible ANO1 knockout mouse and suggested that the channel plays an important role in the regulation of peripheral resistance and systemic blood pressure in the angiotensin II–induced hypertensive model.

The few studies carried out thus far that have measured the level of expression of anoctamins in VSMCs, including PASMCs, have all reported much higher transcript levels of ANO1 than of ANO2.^53–55 This is consistent with the fact that I_Cl(Ca) in vascular myocytes display a Ca²⁺ sensitivity that is similar to that of expressed ANO1^7,13–15,28 and is much higher than that reported for ANO2. ¹⁴³ Experiments with siRNA confirmed that ANO1 is the main pore-forming channel subunit composing the CaCC of rat pulmonary ⁵⁴ and cerebral^55,144,145 artery myocytes and that it contributes significantly to agonist-induced tone in small rat mesenteric arteries. ¹³⁹ ANO1 was recently proposed to be a key component involved in the depolarization and contraction associated with the myogenic response of resistance cerebral arteries. ¹⁴⁴ In cerebral artery myocytes and HEK-293 (human embryonic kidney) cells overexpressing ANO1, exposure of the cell to hypo-osmotic medium to trigger the regulatory volume decrease response or stretching the cell membrane by the application of negative pressure activated a Cl⁻ conductance that was inhibited by ANO1 antibodies or selective knockdown of ANO1 expression by siRNA. ¹⁴⁴ This idea stemmed from prior data showing that, under certain experimental conditions, ANO1 could be activated by cell swelling and act as a volume-regulated Cl⁻ channel. ¹⁴⁶ The contribution of ANO1 to the myogenic tone of cerebral arteries may be unique to this blood vessel, as our group previously showed that NFA (100 μM) had no effect on the basal myogenic response of rabbit mesenteric small arteries, while it potently attenuated the vasoconstriction triggered by phenylephrine. ⁴² The following sections discuss our current state of knowledge in regard to the contribution of various Ca²⁺ sources activating CaCCs in VSMCs and how these signaling interactions control the development of force and blood pressure.

CA²⁺ SOURCES TRIGGERING CACCS IN PASMCS

Activation of CaCCs by Ca2+ mobilization from inositol trisphosphate–sensitive stores

It was recognized very early that Ca²⁺ release from the sarcoplasmic reticulum (SR) represented an important source of Ca²⁺ stimulating CaCCs during agonist-induced vasoconstriction (original contributions are summarized in several reviews^1,4,7). Vasoconstricting neurotransmitters and hormones binding to receptors coupled to G_q and other G-coupled receptors activate PLC, a membrane-bound enzyme that cleaves PIP₂ into the second-messenger signaling molecules diacylglycerol (DAG) and inositol trisphosphate (InsP₃; Fig. 5A). DAG stimulates different classes of protein kinase C isoforms, which phosphorylate a large number of targets that stimulate smooth muscle contraction and promote cell proliferation. InsP₃ diffuses from the plasma membrane to the nearby SR, where it binds to one of three classes of InsP₃ receptors identified in VSMCs (InsP₃R1–R3), which are ligand-activated Ca²⁺ release channels. The binding of InsP₃ to one of these receptors triggers the opening of the channel and passive Ca²⁺ diffusion caused by the large electrochemical gradient between the inside of the SR and the cytoplasm (˜5,000–10,000-fold). The large transient buildup of Ca²⁺ in the cytoplasm evoked by InsP₃ is generally sufficient to elicit a contraction. The transient nature of the elevation in Ca²⁺ concentration is dictated by several factors, including the rate of biochemical synthesis and breakdown of InsP₃, the contribution from Ca²⁺ extrusion and reuptake pathways, and the complex mechanisms controlling the gating of InsP₃ receptors, which involve the dynamic interplay between InsP₃, cytoplasmic and intraluminar Ca⁺⁺, and ATP.^147–149

Figure 5.

Various sources of Ca²⁺ activating the Ca²⁺-activated Cl⁻channel (CaCC) ANO1 in vascular smooth muscle cells. A, The classical mode of activation of CaCCs by stimulation of a G_q-coupled receptor (G_qPCR) with a constricting agonist (e.g., 5-hydroxytryptamine, angiotensin II, endothelin), leading to an elevation of intracellular Ca²⁺ concentration and contraction, is illustrated in this panel by the circled numbers: (1) binding of an agonist to a G_qPCR; (2) breakdown of membrane-bound phosphatidyl-inositol by G_q-stimulated phospholipase C (PLC), leading to the production of the second messengers inositol trisphosphate (InsP₃) and diacylglycerol (DAG); (3) while DAG stimulates protein kinase C (PKC), which targets many proteins involved in contraction and cell proliferation, InsP₃ diffuses toward the sarcoplasmic reticulum (SR), where it binds to and activates a Ca²⁺-permeable receptor-operated channel (InsP₃R), triggering Ca²⁺ release in the cytoplasm; (4) Ca²⁺ diffuses in the cytoplasm and activates contraction and ANO1; (5) opening of ANO1 leads to Cl⁻ efflux and membrane depolarization due to an outwardly directed electrochemical gradient for Cl⁻, which stimulates voltage-gated L-type Ca²⁺ channels encoded by the gene Ca_V1.2; (6) Ca²⁺ entry through Ca_V1.2 would in turn promote actomyosin bridge cycling and further stimulate ANO1, establishing a positive-feedback loop sustaining membrane depolarization and Ca²⁺ entry; (7) there is evidence for stimulation of ANO1 in some vascular myocytes by Ca²⁺ entry through Ca_V1.2 triggering SR Ca²⁺ release from ryanodine receptors (RyR) in a process called Ca²⁺-induced Ca²⁺ release (CICR), which can provide an additional stimulus for ANO1 activation by Ca_V1.2; (8) spontaneous Ca²⁺ release by RyR in the SR, giving rise to “spontaneous transient inward currents,” or STICs, produced by transient openings of ANO1. B, Activation of ANO1 by voltage-independent Ca²⁺ entry pathways: (1) Ca²⁺ unbinding from the SR Ca²⁺ sensor protein STIM1 (stromal interacting molecule 1) by a decrease in the concentration of Ca²⁺ in the SR due to spontaneous Ca²⁺ leakage or mediated by an agonist leads to clustering of portions of the SR with the plasma membrane; (2) STIM1 then physically interacts with a complex composed of one or more members of the canonical transient receptor potential (TRPC) and/or Orai families of cation channels to trigger store-operated Ca²⁺ entry (SOCE); (3) Ca²⁺ entry via SOCE can then stimulate ANO1; (4) ANO1 can also be stimulated by Ca²⁺ influx through a receptor-operated channel stimulated by an agonist (ROCE); (5) although hypothetical, it is possible that ANO1 could be stimulated locally by Ca²⁺ influx through TRPV4 (vanilloid transient receptor potential) channels physically interacting with the anion channel.

Evidence from many laboratories has suggested that CaCCs may be an important and prime target in several types of VSMCs, since the inward current elicited by InsP₃ signaling is mainly carried by Cl⁻, displays kinetics of activation that are consistent with those of CaCCs, and is sensitive to putative CaCC blockers such as NFA, A9C, NPPB, and DIDS.^1,4,7 In PASMCs, constricting agonists binding to receptors that lead to activation of PLC, such as phenylephrine, an α₁-adrenergic receptor agonist, and angiotensin II, have been shown to induce intracellular Ca²⁺ oscillations.^150,151 Oscillatory membrane I_Cl(Ca) in the physiological range of membrane potentials were similarly triggered in PASMCs from several species by histamine, norepinephrine, ¹⁵² ET-1, ¹⁵³ ATP, ¹⁵⁴ and angiotensin II. ¹⁵⁴ In some cases, the inward Cl⁻ current cohabited with a transient outward Ca²⁺-activated K⁺ current (BK_Ca).^152,153 Oscillatory Ca²⁺ transients and I_Cl(Ca) were abolished by caffeine or thapsigargin, which depleted Ca²⁺ stores and inhibited PLC. These transients were also abolished by an intracellular application of heparin, which inhibits InsP₃ receptors. These studies confirmed the important role played by InsP₃ signaling in activating CaCCs and driving membrane potential in PA myocytes; however, several questions remain unanswered. Is activation of CaCCs in PASMCs by InsP₃-mediated Ca²⁺ release essential for eliciting the sustained depolarization triggered by agonists, or does it act only as a depolarizing “primer“? In addition to CaCCs, Ca²⁺ release from the SR may also influence directly or indirectly (e.g., store depletion, phosphorylation by Ca²⁺-dependent kinases and phosphatases) several types of plasma membrane ion channels (BK_Ca, voltage-dependent K⁺ channels, store- and receptor-operated channels, transient receptor potential [TRP] channels) that control membrane potential and Ca²⁺ entry into the cell. What is the relative contribution of these ion channels in the transient depolarization that is generally ascribed to InsP₃ signaling? How does oscillatory Ca²⁺ signaling triggered by constricting agonists and the subsequent activation of plasma membrane channels, including CaCCs, interplay in determining force development in the intact pressurized pulmonary arterial tissue (e.g., intercellular coupling, synchronous vs. asynchronous events)?

Activation of CaCCs by ryanodine-sensitive Ca2+ stores

Benham and Bolton ¹⁵⁵ were the first to speculate about the existence of focal Ca²⁺-release events leading to the activation of “spontaneous transient outward currents,” or STOCs, in several types of VSMCs and visceral SMCs, which were manifest of the stimulation of clusters of BK_Ca. Nelson et al. ¹⁵⁶ later confirmed the existence of local Ca²⁺-release events, or “Ca²⁺ sparks,” produced by the opening of clusters of ryanodine receptors located in portions of the SR making close contact with the plasma membrane (˜20 nm). For a comprehensive review of the properties and regulation of Ca²⁺ sparks and their role in triggering BK_Ca and STOCs and regulating resting membrane potential (E_m) and tone in intact arteries, the reader is invited to peruse excellent reviews on the topic.^157–159

By analogy to STOCs, a similar type of electrical event called “spontaneous transient inward currents,” or STICs, was discovered a few years later in airway^160–162 and vascular^{33,36,163–167} SMCs. Ion-replacement experiments confirmed that STICs originate from the coordinated opening of a group of Cl⁻ channels triggered by a transient elevation of Ca²⁺ on the cytoplasmic side of the membrane that arises from ryanodine-sensitive Ca²⁺ stores (Fig. 5A). STOCs and STICs were shown to coexist in some SMCs, producing events called “STOICs.”^168,169 STOICs were generally the product of an initial faster BK_Ca-mediated STOC followed by the slower CaCC-induced STIC. In a detailed analysis of the spatiotemporal relationship between Ca²⁺ sparks and STICs in mouse airway SMCs, Bao et al. ¹⁷⁰ suggested that the majority of CaCCs are arranged in clusters optimally located in the vicinity of groups of ryanodine receptors producing Ca²⁺ sparks to provide a functional unit tuning local Ca²⁺ signaling and cell excitability.

In the pulmonary circulation, STICs were first described in rabbit PASMCs. ¹⁶⁴ STICs in these cells were recorded simultaneously with STOCs, with a sequence of activation similar to that of STOICs described above, with the outward BK_Ca current generally preceding the activation of CaCCs. Consistent with the existence of STOCs and STICs was the later demonstration of the existence of ryanodine-dependent Ca²⁺ sparks in rat intralobar PA myocytes. ¹⁷¹ Interestingly, STICs were more frequently observed than STOCs near the E_m of PASMCs (approx. −50 mV). ¹⁶⁴ All 3 ryanodine receptor subtypes (RyR1–RyR3) are expressed in PASMCs,^172–174 but they are differentially distributed within the cell, with RyR1 and RyR2 predominantly expressed near the plasma membrane and RyR1 and RyR3 exhibiting enhanced staining in the perinuclear region. ¹⁷² Ca²⁺ sparks measured in these two regions displayed very distinct characteristics, possibly conferring unique profiles of Ca²⁺ regulation tuned for specific cellular functions. Thus, similar to those in airway SMCs, ¹⁷⁵ Ca²⁺ sparks in PASMCs appear to be generated primarily by the spontaneous opening of RyR1 and RyR2 Ca²⁺-release channels. ¹⁷²

Compared to Ca²⁺ sparks in cardiac myocytes, those recorded in PASMCs occurred less frequently and displayed lower amplitude but a similar time course and spatial spread. Interestingly, a low concentration of caffeine (0.5 mM) was shown to activate Ca²⁺ sparks and cause membrane depolarization in PASMCs, whereas the same treatment hyperpolarized SMCs from systemic arteries. ¹⁷¹ This suggests that Ca²⁺ sparks preferentially activate STICs in adult PA myocytes, while they primarily target BK_Ca and STOCs in cells from systemic arteries. It is also consistent with the developmental downregulation of BK_Ca in SMCs from distal pulmonary arteries.^176,177 Consistent with this profile, the E_m of PASMCs^178–180 and the basal tone of intact pulmonary arteries ¹⁸¹ were unaffected by BK_Ca blockers such as tetraethylammonium chloride or charybdotoxin, again emphasizing that the role of BK_Ca in PASMCs, relative to that of CaCCs, is at best a minor one.

Do Ca²⁺ sparks and STICs influence pulmonary arterial membrane potential and tone under physiological conditions? It seems unlikely that STICs would depolarize E_m under basal conditions, as both the probability of observing Ca²⁺ sparks and their frequency when detected (˜0.3 s⁻¹) are low. ¹⁷¹ However, their contribution could become significant during stimulation by certain agonists. ET-1, but not norepinephrine, enhanced Ca²⁺ spark frequency ˜3–4-fold, ¹⁷¹ conceivably activating STICs and CaCCs sufficiently to depolarize E_m and promote Ca²⁺ entry through voltage-gated Ca²⁺ channels. A subsequent study from the same group showed that the ET-1-mediated increase in Ca²⁺ spark frequency, amplitude, and duration involved the coordination of both InsP₃ and ryanodine receptors. ¹⁸² Crosstalk between ryanodine- and InsP₃-sensitive Ca²⁺ stores in PA myocytes was confirmed by several groups.^173,183 In summary, spontaneous and transient activation of CaCCs by Ca²⁺ sparks elicited by ryanodine-dependent and, in some cases, InsP₃-dependent Ca²⁺ releasable pools may serve as an excitatory mechanism in PASMCs. The membrane depolarization induced by STICs, especially in the presence of agonists, may in turn enhance the open probability of voltage-gated Ca²⁺ channels, leading to increased Ca²⁺ entry and vasoconstriction (see next section).

Activation of CaCCs by Ca2+ entry through voltage-dependent Ca2+ channels

It was recognized early that Ca²⁺ entry through dihydropyridine-sensitive voltage-gated Ca_L represented an important source of Ca²⁺ triggering CaCCs in SMCs and other cell types (Fig. 5A). Evidence for a tight coupling between Ca²⁺ influx through Ca_L (encoded by Ca_V1.2) and CaCCs was reported in SMCs isolated from rat,^{44,45,184,185} rabbit,^18,36,186 and mouse ¹⁸⁷ portal vein, rabbit esophagus, ¹⁸⁸ rabbit coronary artery, ³⁴ canine ¹⁶² and guinea pig ¹⁸⁹ trachea, and rat ⁴¹ and rabbit ²⁷ PA.

Figure 6 shows the results of typical whole-cell patch-clamp experiments, carried out in rat PASMCs by Yuan, ⁴¹ illustrating the tight relationship between the two channels. In Figure 6A, step depolarizations from a holding potential of −70 mV to potentials ranging from −60 to +60 mV elicited a family of currents consisting, for most voltage steps, of a small early transient inward Ca²⁺ current (I_Ca(L)) followed by a large secondary inward or outward I_Cl(Ca), depending on membrane potential, that significantly overlapped the time course of inactivation of I_Ca(L). Because of the superimposed I_Cl(Ca) (in these experiments, the equilibrium potential for Cl⁻ [E_cl] = 0 mV), I_Ca(L) reversed near +25 mV (Fig. 6B, left [I_tr]), whereas the current measured at the end of the step reversed at a potential below 0 mV (Fig. 6B, middle [I_td]). Figure 6A also shows that upon return to the holding potential, a large, persisting inward tail current is apparent and reflects slow I_Cl(Ca) closure due to a mixture of voltage-dependent channel deactivation and Ca²⁺ reuptake and extrusion mechanisms. The classical slow I_Cl(Ca) tail current was virtually uncontaminated by I_Ca(L) because the latter deactivates very quickly after repolarization and has thus been used as a useful index of I_Cl(Ca) conductance. Analysis of the voltage dependence of the tail current shows that activation of I_Cl(Ca) occurred near the threshold for activation of I_Ca(L), peaked around +20 mV, and progressively declined with stronger membrane depolarizations, consistent with the reduced driving force for Ca²⁺ (Fig. 6B, right [I_tail]). This and many other studies in SMCs confirmed that the inward tail current was carried by Cl⁻ as it reversed near E_Cl and was not influenced by omitting permeable cations in the buffer. Whereas only the delayed current during the step and the inward tail current following repolarization were attenuated by the CaCC blockers NFA, DIDS, SITS, and A9C or by inclusion of a Ca²⁺ chelator such as EGTA or BAPTA in the pipette solution, both the early inward I_Ca(L) and delayed I_Cl(Ca) and the tail current were abolished by Ca²⁺ channel blockers such as nifedipine, nicardipine, Co²⁺, and Cd²⁺. Predictably, enhancing I_Ca(L) with the Ca²⁺ channel agonist Bay K 8644 enhanced I_Ca(L)-induced delayed I_Cl(Ca) and tail current. Finally, external Ca²⁺ replacement with Ba²⁺, which readily permeates Ca_L, abolished I_Cl(Ca), indicating that Ba²⁺ is unable to substitute for Ca²⁺ in activating CaCCs.

Figure 6.

Ca²⁺-activated Cl⁻ currents triggered by Ca²⁺ entry through voltage-gated Ca²⁺ channels in rat pulmonary artery smooth muscle cells (PASMCs). A, Typical family of whole-cell currents (top traces) recorded from a single PASMC with the voltage-clamp protocol shown below the traces. The cell was bathed in a solution containing 1.8 mM Ca²⁺. B, Mean current-voltage relationships for currents measured as indicated in A and labeled as the early transient current (I_tr) measured between 50 and 90 ms, the time-dependent current (I_td) measured at the end of the depolarizing step (between 750 and 790 ms), and the tail current (I_tail) measured immediately after repolarization to the holding potential of −70 mV (between 900 and 940 ms). Each data point represents a mean ± SEM (n = 13). Reproduced from Yuan ⁴¹ with permission from the American Physiological Society.

In some SMCs and other cell types exhibiting CaCC activity, activation of I_Cl(Ca) in response to Ca_L opening appears to involve Ca²⁺-induced Ca²⁺ release (CICR) and implicates mostly ryanodine (Fig. 5A), but in some cases InsP₃ receptors, as perturbation of Ca²⁺ release or reuptake with agents such as caffeine, ryanodine, cyclopiazonic piazonic acid (CPA), 2-aminoethoxydiphenyl borate (2-APB), or tetracaine reduced the magnitude of I_Ca(L)-induced I_Cl(Ca).^34,187,190 The possible contribution of CICR in stimulating I_Cl(Ca) in PASMCs is unknown.

Regardless of whether CaCCs are activated locally by Ca²⁺ sparks or by a rise in global [Ca²⁺]_i, it is clear that E_Cl in VSMCs is physiologically set to optimally stimulate sustained Ca²⁺ influx through Ca_L. In smooth muscle, Cl⁻ is actively accumulated inside the cell by the complex interplay between the HCO₃-Cl⁻ exchanger, the Na⁺-K⁺-2Cl⁻ cotransporter, and an obscure transporter called “Pump III.” ¹⁹¹ In rabbit PASMCs, ion flux experiments established an intracellular Cl⁻ concentration of 51 mM, yielding a calculated equilibrium potential for Cl⁻ of −26 mV, ¹⁹² well above the E_m measured in these cells^41,192,193 and similar to that measured in other smooth muscles (for a review, see Wilson and Leblanc ¹⁹⁴ ). Thus, because E_Cl lies near the peak of the window Ca²⁺ current (approx. −20 mV), the depolarization evoked by activation of CaCCs favors sustained Ca²⁺ entry into the cell and contraction. This establishes a powerful positive-feedback loop, which is possibly counteracted by CaMKII-dependent inactivation of CaCC activity by phosphorylation. Support for the positive-feedback-loop hypothesis (Fig. 5A) came from indirect experiments examining the effects of Cl⁻ channel blockers such as NFA, DIDS, and A9C on the vasoconstriction elicited by constricting agonists. Most physiological studies with arterial smooth muscles reported a significant attenuation of agonist-induced tone by CaCC blockers that was generally similar to or smaller than that produced by direct inhibition of Ca_L (e.g., nifedipine).^{1,4,7,38–42} The main issue that has hampered any definitive conclusion in regard to this proposed mechanism concerns the relatively low efficacy of the blockers to detect vasorelaxation and the fact that they have possible confounding effects on other processes affecting excitation-contraction coupling. Detailed reviews on the potentially detrimental side effects of CaCC blockers are available.^1,4,7,30 For the majority of the studies supporting a role for CaCCs in electromechanical coupling, the key supporting argument was based on the observation that the blockers generally did not affect the contraction evoked by KCl-induced membrane depolarization. This implied that the vasorelaxation of precontracted arteries elicited by such inhibitors was not caused by a direct effect on the contractile machinery or by inhibition of Ca²⁺ entry through Ca_L. NFA (10–50 μM) caused relaxation of rat PAs precontracted with 5-hydroxytryptamine (5-HT) or phenylephrine and reversed the depolarization induced by 5-HT, but it produced no effect on intracellular Ca²⁺ release elicited by 5-HT. ⁴¹ The recent discovery of TMEM16A permitted the development of high-throughput screening of small molecules targeting CaCCs with greater affinity and potency.^57–59,195 One of these molecules, labeled T16A_Inh-A01, was shown to cause relaxation of several types of rat and mouse blood vessels precontracted with various agonists.^22,43,113 Again, because ANO1 knockout mice die prematurely after birth, unequivocal testing of this hypothesis in pulmonary arteries, in particular in small-intrapulmonary-resistance vessels, will require the use of transgenic mice, whereby ANO1 can be downregulated or overexpressed in adult animals using inducible and conditional genetic strategies, as recently shown in a study examining the role of ANO1 in systemic hypertension. ¹⁴⁰

Activation of CaCCs by store- and receptor-operated Ca2+ entry

In many cell types, including VSMCs, depletion of intracellular Ca²⁺ stores leads to activation of a plasma membrane voltage-independent Ca²⁺ entry pathway, a process called “store-operated Ca²⁺ entry” (SOCE) or “capacitative Ca²⁺ entry.” ¹⁹⁶ SOCE appears to be prominent in the pulmonary circulation, having been detected in both PASMCs^{136,138,197–201} and venous SMCs. ²⁰¹ The major physiological functions of this mechanism are to replenish the stores after their emptying by agonist-mediated receptor signaling, leading to intracellular Ca²⁺ release, and to serve as a significant Ca²⁺ influx pathway during agonist-induced vasoconstriction.^202–204 Experimentally, SOCE can be evoked by blocking Ca²⁺ reuptake into the SR with agents such as thapsigargin or CPA, both specific inhibitors of the SR Ca²⁺-ATPase (SERCA), in a cell exposed to a Ca²⁺-free buffer, which then leads to SR Ca²⁺ depletion. SOCE is unmasked by the reintroduction of external Ca²⁺ in the presence of the SERCA inhibitor and an L-type Ca²⁺ channel blocker such as nifedipine. Another method is to buffer intracellular Ca²⁺ with a chelating agent such as BAPTA or EGTA, which also causes SR Ca²⁺ depletion and activation of SOCE.

The molecular architecture of SOCE in smooth muscle is still unclear but has recently been the subject of more attention, with the discovery of two novel gene families directly involved in SOCE in T-lymphocytes, where “Ca²⁺ release-activated channels” (CRAC or CRACM channels) had been first characterized: ²⁰⁵ (1) stromal interacting molecules 1 and 2 (STIM1 and STIM2)^206–210 and (2) Orai1, Orai2, and Orai3.^211–215 Strong evidence supports STIM1 and STIM2 as Ca²⁺-binding proteins acting as the sensors of Ca²⁺ depletion in the ER, which in turn interact with plasma membrane CRACs encoded by Orai genes. This process involves a substantial reorganization of the cytoskeleton, leading to translocation and clustering of STIMs, where it is believed to make close contact with the areas of the plasma membrane where SOCE is taking place. Significant progress has been made in the past few years, providing a conceptual framework defining the structure and protein domains involved in ER Ca²⁺ sensing, the interaction of STIMs with Orai, and the role of caveolae and molecular regulation of this process; the interested reader should consult excellent reviews available on these topics.^216–221

In VSMCs, the situation is hazier, because even though store depletion clearly stimulates SOCE, the identity of the plasma membrane ion channels is unclear. Whereas the pore of CRAC and Orai channels is highly selective for Ca²⁺, the cation current activated by store depletion in VSMCs has been consistently shown to be a nonselective cation conduction pathway. Before the discovery of STIM and Orai proteins, investigators speculated that members of the TRP superfamily of ion channel genes, in particular those of the canonical subclass (TRPC), could be valid candidates, as many are nonselective channels displaying variable permeability to Ca²⁺.^{203,204,222,223} Studies have provided evidence for TRPC1, TRPC4, TRPC5, and/or TRPC6 as a part of the complex responsible for SOCE in VSMCs.^{203,222–226} More recent reports indicate that STIM1 and Orai1 are expressed in airway SMCs ²²⁷ and VSMCs,^{200,227–231} where they were shown to play a significant role in SOCE, vascular tone, and cell proliferation. In addition to stimulating Orai1, STIM1 also interacts with TRPC1.^232–234 Moreover, it has been speculated that after store depletion, STIM1, Orai1, and TRPC1 may form a ternary complex mediating voltage-independent SOCE (Fig. 5B).^235–238

Although the molecular composition of the SOCE machinery and the nature of the divalent cation conduction pathway in PASMCs still are to be determined, it is clear that this Ca²⁺ entry pathway may represent an important additional trigger for activation of CaCCs in vascular myocytes. Yuan ⁴¹ provided the first indirect evidence of coupling between SOCE and I_Cl(Ca) in rat PA myocytes. In the absence of external Ca²⁺, application of CPA to block the SERCA pump elicited transient I_Cl(Ca), consistent with Ca²⁺ first leaking from the SR stores and then being extruded, presumably through the plasma membrane Ca²⁺-ATPase and the Na⁺/Ca²⁺ exchanger. Readmission of external Ca²⁺ in the presence of CPA stimulated a transient I_Cl(Ca), although it is unclear whether a store-operated nonselective cation current was superimposed on the Cl⁻ current. More recently, our group specifically addressed this question in rabbit PASMCs. In the presence of nifedipine to inhibit Ca_L, blocking the SERCA pump with thapsigargin ¹³⁶ or CPA^136,138 activated a nonselective cation conductance displaying slight inward rectification, a response typical of that previously described in VSMCs by other investigators. A double-pulse protocol was used to analyze the relationship between SOCE and I_Cl(Ca). The traces in Figure 7Ab (Fig. 7 was reproduced from Angermann et al. ¹³⁸ with permission) show that in the presence of nifedipine, CPA elicited a membrane current from a holding potential of 0 mV that was mostly time independent during an initial step (1 s) ranging from −100 to +60 mV. A second pulse to +90 mV revealed the appearance of a slow, time-dependent outward current of which the magnitude grew with the level of hyperpolarization achieved during the first step. A slow inward tail current relaxing toward baseline was apparent upon return to −80 mV. The time-dependent currents at both +90 and −80 mV exhibited properties consistent with I_Cl(Ca) in PASMCs. Application of the SOCE blocker SKF-96365 abolished both the time-independent and the time-dependent currents evoked by CPA (Fig. 7Ac), whose magnitudes were similar to those recorded in the presence of nifedipine alone (Fig. 7Aa). Figure 7B, from the same study, showed that the voltage dependence of the current recorded at +90 mV with CPA and nifedipine displayed an inverse relationship to the voltage during the initial step, which was flattened by SKF-96365. Other series of experiments confirmed that the time-independent conductance is a nonselective cation current mainly permeable to Na⁺ and is blocked by SKF-96365 but is insensitive to NFA, whereas the time-dependent current components were abolished by the CaCC blocker. On the basis of these results, we hypothesized that CPA activated a store-operated nonselective cation conductance permeable to Ca²⁺ or indirectly permitting Ca²⁺ entry via reverse-mode Na⁺-Ca²⁺ exchange (NCX). Stronger hyperpolarizing steps would enhance Ca²⁺ entry through this pathway by increasing the driving force for Ca²⁺ and resulting in greater activation of I_Cl(Ca) during the subsequent step to +90 mV. Figure 8 (also from Angermann et al. ¹³⁸ ) shows the results of dual patch-clamp and Ca²⁺ fluorescence (Fluo-4) experiments confirming that in the presence of CPA and nifedipine, stronger and longer (3 s instead of 1 s) hyperpolarizing steps evoked larger, slowly developing Ca²⁺ transients that induced slow inward I_Cl(Ca) at very negative positive potentials and robust outward I_Cl(Ca) during the subsequent step to +90 mV. We concluded that in these recording conditions, reverse-mode NCX was not required for the SOCE-CaCC interaction, since (1) stepping to a positive potential (+90 mV) did not elicit a Ca²⁺ transient, as would be expected if sufficient Na⁺ had entered the cell via the store-operated nonselective cation current (I_SOC; Fig. 8A, 8B), and (2) the activation of I_Cl(Ca) stimulated by CPA was unaffected by the NCX inhibitor KB-R7943 (Fig. 8D). These experiments highlighted the tight relationship between SOCE and CaCC activity in PASMCs (Fig. 8C). Although an indirect mode of activation of CaCCs by reverse-mode NCX was not demonstrated, our experiments did not rule out a potential role of the exchanger in this interaction under more physiological conditions, as the cells were dialyzed with a Na-free pipette solution.

Figure 7.

Store depletion activates Ca²⁺-activated Cl⁻ currents in rabbit pulmonary artery smooth muscle cells. A, Three families of membrane currents recorded from the same smooth muscle cell and elicited by the voltage-clamp protocol displayed below the traces. The traces in a were recorded in the presence of 1 μM nifedipine (Nif). In the continued presence of Nif, the cell was then superfused with 30 μM cyclopiazonic acid (CPA) for 5 minutes (b) to deplete the stores. Subpanel c shows a family of traces recorded after application with the nonselective cation channel inhibitor SKF-96365 (SKF; 50 μM), with CPA and Nif still present. B, Mean current-voltage relationships for current density (pA/pF), measured at the end of the second step to +90 mV, as a function of the voltage during the preceding step. Each data point represents the mean ± SEM current recorded in the presence of 1 μM Nif alone (open circles; n = 5), with 30 μM CPA and Nif (filled circles; n = 12), or with Nif, CPA, and 50 μM SKF (filled squares; n = 4). There were no significant differences between the Nif-alone and Nif + CPA + SKF groups. All data points in the Nif + CPA group were significantly different from those of the other two groups (with a maximum P value of <0.05). The 3 lines passing through the data points represent linear regression fits. Reproduced from Angermann et al. ¹³⁸ with permission from the National Research Council of Canada.

Figure 8.

Voltage dependence of Ca²⁺ transients and Ca²⁺-activated Cl⁻ currents evoked by store-operated Ca²⁺ entry (SOCE) in rabbit pulmonary artery smooth muscle cells. A, Family of membrane currents and intracellular Ca²⁺ transients measured simultaneously with the Ca²⁺ indicator Fluo-4 (expressed as F/F₀) elicited by the protocol shown below the traces. The traces were recorded after 5 minutes of incubation with 1 μM nifedipine (Nif) and 30 μM cyclopiazonic acid (CPA) in the presence of 1.8 mM Ca²⁺ to induce SOCE. B, Mean current-voltage relationships for the instantaneous (filled circles) and late (open circles) membrane currents recorded in the presence of 1.8 mM Ca²⁺ with Nif and CPA for experiments similar to the one described in A. Each data point represents the mean ± SEM (n = 3). C, Dependence of normalized peak current at +90 mV during P2 (see protocol in A) on the magnitude of the Fluo-4 signal measured at the end of P1 (see protocol in A). Each data point represents the mean ± SEM of normalized peak current during P2 and Fluo-4 fluorescence at the end of P1 (n = 3). The solid line represents a second-degree polynomial fit. D, Peak current at +90 mV during P2 (see protocol in A) as a function of voltage during P1. Currents were recorded in the presence of Nif and CPA (filled circles) before the addition of 1 μM KB-R7943 (open circles), a Na⁺/Ca²⁺ exchange blocker. Each data point represents the mean ± SEM (n = 9). The two solid lines represent least squares exponential fits. Reproduced from Angermann et al. ¹³⁸ with permission from the National Research Council of Canada.

Further evidence in support of this interaction was also provided by Yamamura et al., ²³⁹ who showed that CaCC blockers such as NFA, flufenamic acid, and DIDS inhibited the transient and sustained phases of Ca²⁺ transients induced by angiotensin II and by store depletion in human PASMCs, which suggests that CaCCs and TMEM16A can both be activated by SOCE and receptor-operated Ca²⁺ entry (ROCE; Fig. 5B). Similarly to SOCE, TRPC channels are also thought to be the pore-forming subunits of ROCE in VSMCs.^240–242

One conceptual problem in associating SOCE/ROCE with CaCCs is that E_Cl is more positive than E_m. After stimulation with 5-HT in rat PA, inhibition of CaCCs by NFA caused membrane hyperpolarization, presumably by shifting E_m toward E_K (the equilibrium potential for K). Since Ca²⁺ influx during SOCE or ROCE would be expected to be increased by hyperpolarization due to an increased driving force for Ca²⁺, the hyperpolarization observed after the blocking of CaCCs by NFA should then enhance, rather than decrease, Ca²⁺ influx and intracellular Ca²⁺ levels. Consistent with the effects of NFA on SOCE and ROCE in human PA myocytes, ²³⁹ we also observed a significant relaxation by the TMEM16A inhibitor T16A_Inh-A01 of small intralobar rat pulmonary arteries exposed to CPA and nifedipine. This then leads to the following questions: Is E_Cl stable and lying where expected (approx. −25 mV), or is it somehow changing because of the experimental conditions? What is the level of E_m when CPA or thapsigargin is used to elicit SOCE, as opposed to a natural vasoconstrictor? Is the SOCE molecular complex physically and functionally linked to TMEM16A or to one of its regulatory subunits? More comprehensive research is required to answer these questions; intracellular membrane potential-recording experiments under very well controlled experimental conditions should help in elucidating this puzzle.

POTENTIAL ROLE OF CACCS AND ANO1 IN PULMONARY HYPERTENSION

Pulmonary hypertension (PH) is a rare chronic human disease exhibiting high morbidity and mortality rates. An elevation in pulmonary vascular resistance (PVR) is the direct consequence of PH, which enhances right ventricular (RV) afterload and in time leads to RV failure.^243–247 The World Health Organization has recently reclassified the various forms of the disease into 5 categories: (1) pulmonary arterial hypertension (PAH), a progressive and severe syndrome displaying a low prognosis for survival, (2) PH associated with left heart disease, such as congenital heart disease, (3) PH associated with lung disease and/or hypoxia (e.g., chronic obstructive pulmonary disease), (4) thromboembolic PH, and (5) miscellaneous.^244–247 PAH is defined by a pulmonary artery pressure (PAP) exceeding 25 mmHg at rest and 30 mmHg during exercise, a capillary wedge pressure ≤ 15 mmHg, and PVR ≥ 2–3 Wood units.^244–247 Three major factors contribute to elevating PAP in PAH patients: (1) enhanced vasoconstriction, (2) partial or complete luminal obstruction due to extensive and complex remodeling of the intimal, medial, and adventitial layers of proximal and distal pulmonary arteries, characterized by enhanced proliferation, reduced apoptosis, and infiltration of inflammatory and progenitor cells, and (3) a propensity for in situ thrombosis.^243–247

There is convincing evidence that PASMCs from animal models of PAH and humans with PAH are more depolarized and display higher [Ca²⁺]_i than cells from healthy subjects and that this, at least initially, contributes to the observed changes in responsiveness to endogenous vasoconstrictors and remodeling.^247–249 Consistent with this observation was the demonstration of a decreased voltage-dependent K⁺ current (I_K) accompanied by a reduced expression of several members of the Kv superfamily of voltage-dependent K⁺ channels, including Kv1.5, Kv2.1, Kv4.3, and Kv9.3.^250–259 It is thought that the depolarization caused by the reduced expression of K⁺ channel genes increases intracellular Ca²⁺ levels by activating Ca_L, which then triggers gene transcription, leading to reduced apoptosis and cell proliferation.^{248,256,260–263} One recent study revealed an upregulation of Ca_L activity and protein levels of the α_1C pore-forming subunit of the channel in the chronic hypoxic neonatal piglet model. ²⁶⁴ There is also evidence supporting the idea that other Ca²⁺ entry pathways might be altered in PAH.^{248,249,258,265} Several groups reported augmented ROCE and SOCE and in some cases increased expression of several members of the TRPC gene family, namely, TRPC1, TRPC3, TRPC4, and/or TRPC6, in pulmonary arteries from animal models of PH and in patients diagnosed with idiopathic PAH.^{198,225,265–269} Such changes may be important, because TRPC channels have been linked to PA cell proliferation.^225,270 Recent studies also reported increased expression of one member of the “vallinoid” subfamily of TRP channels, TRPV4, in PASMCs from chronic hypoxic rats and suggested that the osmo-mechanosensitive channel could participate in the increase in myogenic and agonist-mediated tone of small pulmonary arteries in PH.^271–273 Interestingly, TRPV4 and ANO1 were recently shown to physically interact and regulate water efflux through aquaporin channels in choroid plexus epithelial cells. ²⁷⁴ It is thus tantalizing to speculate that a similar interaction could exist in PASMCs, whereby Ca²⁺ entry through TRPV4 could activate ANO1 and regulate tone in health and disease (Fig. 5B). Finally, the CRAC channel Orai2 and the SOCE Ca²⁺ sensor STIM2 (but not STIM1) were shown to be upregulated and were suggested to account for the enhancement of SOCE in PASMCs from patients with idiopathic PAH and chronic hypoxic rats. ²⁶⁸

Very recently, two independent studies provided evidence for increased CaCC activity and expression of ANO1 in PA from two models of PH in the rat, monocrotaline-induced PH ¹¹³ and PH caused by chronic hypoxia. ⁴³ In both animal models, CaCC current density in PASMCs from pulmonary hypertensive rats was elevated by ˜ 2-fold relative to that in control animals, and this observation was correlated by similar increases in ANO1 expression at the mRNA and protein levels, which is similar to that observed in a chronic mouse model of asthma ¹³⁵ and some forms of cancer.^76,275–279 Consistent with these observations, the CaCC blockers NFA and T16A_Inh-A01 exerted more potent vasorelaxation of pulmonary arteries precontracted with 5-HT in pulmonary hypertensive than in control animals. Taken together, these studies suggest that CaCCs and ANO1 might be another important ionic conductance influencing pulmonary arterial resistance and blood flow in PH. This is further reinforced by the fact that crosstalk exists between CaCCs and SOCE/ROCE, and the proteins involved in these processes are also augmented in conditions favoring SMC proliferation and remodeling of the pulmonary arterial wall in PH. Indeed, ANO1 and ANO6 have been respectively shown to influence the direction and the speed of migration in Ehrlich-Lettre ascites cells, a cell model commonly used in cancer studies. ²⁸⁰ Specific knockdown of ANO1 by siRNA transfection of rat mesenteric arteries was recently demonstrated to abrogate both “classical” I_Cl(Ca) and I_Cl.cGMP and to inhibit spontaneous cyclical vasomotion as well as agonist- and KCl-induced contractions of small arteries. Surprisingly, downregulation of ANO1 also reduced the expression at the mRNA level of bestrophins, previously shown to be involved in the generation of I_Cl.cGMP and the elicitation of rhythmic vasoconstrictions in this vascular bed,^125,129 and the expression of Ca_V1.2 mRNA. The above study suggested that ANO1 plays an important function in regulating peripheral systemic vascular resistance. An earlier report by Wang et al. ¹⁴⁵ showed that in the 2-kidney, 2-clip systemic hypertensive rat model, ANO1 was downregulated and that this effect was responsible for the increased proliferation of basilar arterial SMCs. In this model, ANO1 was proposed to negatively regulate cell proliferation by altering the expression of cell cycle proteins and to essentially exert an antihypertensive effect. However, a more recent report showed that knockdown of ANO1 in a conditional smooth muscle-specific and inducible mouse model lowered systemic blood pressure in both untreated animals and mice rendered hypertensive by a chronic infusion of angiotensin II but had no effect on blood pressure of mice made hypertensive by a high-salt diet. ¹⁴⁰ Whether CaCCs and ANO1 play a similar role in the remodeling of the pulmonary arterial wall is unknown and will require investigation using a similar model.

SUMMARY ANDCONCLUDING REMARKS

Progress made in the past 2 decades has shed new light on the functional and molecular properties of CaCCs in VSMCs. It is now known that CaCCs in PASMCs are small-conductance Ca²⁺-dependent anion channels modulated by voltage that are subjected to tight regulation by phosphorylation involving CaMKII, CaN, PP1, and PIP₂ and that the pore-forming subunit is most likely encoded by ANO1. In pulmonary arterial smooth muscle, CaCCs can be activated by (1) mobilization of intracellular Ca²⁺ stores involving both InsP₃- and ryanodine-sensitive Ca²⁺ stores triggered spontaneously by localized Ca²⁺ sparks or by G-protein coupled receptor stimulation and (2) Ca²⁺ entry via voltage-gated Ca²⁺ channels, SOCE, or ROCE.

There remain many intriguing and pressing questions about the properties of CaCCs and their interaction with other signaling molecules. Are CaCCs encoded only by ANO1, or is there a role for other anoctamins, perhaps ANO2 and/or ANO6, or bestrophins in forming the native channel in PASMCs? How are native CaCCs regulated by phosphorylation at the single-channel level, and is ANO1 similarly and directly regulated, or is one of its unidentified accessory subunits targeted by kinase activity? How does the regulation of CaCCs by phosphorylation influence membrane potential, intracellular Ca²⁺ dynamics, and arterial tone? Under physiological conditions, activation of CaCCs is hypothesized to increase SMC excitability by promoting membrane depolarization, activation of voltage-dependent Ca²⁺ channels, increased Ca²⁺ influx, and vasomotor tone. Support for this hypothesis is based on the ability of pharmacological agents to reverse the depolarization and contraction induced by vasoconstrictors. Unequivocal testing of this hypothesis will require a demonstration that the depolarization and contraction to vasoconstrictors are impaired in intact arteries from ANO1 conditional-knockout mice or following ANO1 knockdown by RNA silencing in vitro, as was recently done in cerebral ¹⁴⁴ and mesenteric ¹³⁹ arteries. Moreover, is ANO1 (and its spliced variants) differentially expressed in the pulmonary arterial tree, as recently shown in systemic arteries? ¹⁴⁰ What is the functional impact of the interaction of CaCCs with SOCE and ROCE under physiological conditions, as well as in PH, where these signaling pathways are upregulated? Is there a role for CaCCs and anoctamins in hypoxic pulmonary vasoconstriction? Is the increased expression of the CaCC/ANO1 channel seen in animal models of PH also present in human PH? In addition to promoting vasoconstriction, does enhanced expression of the channel influence remodeling of the pulmonary arterial wall in PH, and if so, by what mechanism? How is ANO1 expression genetically regulated in health and pulmonary arterial diseases? There is no doubt that the discovery of anoctamins has resurrected an interest in this field and identified a new therapeutic target in the battle against the deleterious consequences of PAH, which will lead to exciting research in the near future.