Engagement of the EP2 prostanoid receptor closes the K+ channel KCa3.1 in human lung mast cells and attenuates their migration

Human lung mast cells (HLMC) express the Ca2+-activated K+ channel KCa3.1, which plays a crucial role in their migration to a variety of diverse chemotactic stimuli. KCa3.1 activation is attenuated by the β2-adrenoceptor and the adenosine A2A receptor through a Gs-coupled mechanism independent of cyclic AMP. Prostaglandin E2 promotes degranulation and migration of mouse bone marrow-derived mast cells through the Gi-coupled EP3 prostanoid receptor, and induces LTC4 and cytokine secretion from human cord blood-derived mast cells. However, PGE2 binding to the Gs-coupled EP2 receptor on HLMC inhibits their degranulation. We show that EP2 receptor engagement closes KCa3.1 in HLMC. The EP2 receptor-specific agonist butaprost was more potent than PGE2 in this respect, and the effects of both agonists were reversed by the EP2 receptor antagonist AH6809. Butaprost markedly inhibited HLMC migration induced by chemokine-rich airway smooth muscle-conditioned media. Interestingly, PGE2 alone was chemotactic for HLMC at high concentrations (1 µM), but was a more potent chemoattractant for HLMC following EP2 receptor blockade. Therefore, the Gs-coupled EP2 receptor closes KCa3.1 in HLMC and attenuates both chemokine- and PGE2-dependent HLMC migration. EP2 receptor agonists with KCa3.1 modulating function may be useful for the treatment of mast cell-mediated disease.


Introduction
Mast cells are tissue-dwelling cells derived from bone marrow progenitors. They are present in all organs throughout the human body, both at mucosal surfaces and within connective tissues. Mast cells play a major role in tissue homeostasis, host defence and the pathophysiology of many diverse diseases [1]. These include pulmonary fibrosis, rheumatoid disease and atherosclerosis, but they are most commonly associated with allergic disease due to their activation by allergen [2]. In many diseases mast cells re-locate to specific compartments within tissue. This is typified in asthma where mast cells migrate into the airway epithelium [3], airway smooth muscle (ASM) [4] and submucosal glands [5]. This places activated mast cells in direct contact with these dysfunctional airway elements, allowing the specific delivery of detrimental cell-cell signals. Drugs that inhibit this tissue relocation by preventing mast cell migration may prove particularly effective in the treatment of mast cell-mediated disease.
Human mast cells express the intermediate conductance Ca 21 -activated K 1 channel K Ca 3.1, which plays a critical role in their migration to diverse chemotactic stimuli [6], and to a lesser extent in their degranulation [7,8]. Drugs that directly block this channel or which close it indirectly therefore have potential as novel therapies for mast cell-dependent disease. K Ca 3.1 in human lung mast cells (HLMC) is closed by both salbutamol and adenosine via the b 2 -adrenoceptor and A 2A adenosine receptors, respectively [9,10], which are both G s -coupled G protein-coupled receptors (GPCR). In keeping with a critical role for K Ca 3.1 in HLMC migration, adenosine also inhibits HLMC chemotaxis via the A 2A receptor [10]. PGE 2 is a prostanoid with four specific GPCR, designated EP [1][2][3][4] . EP 2 and EP 4 couple to G s , EP 3 couples predominantly to G i although two isoforms also couple to G s , while EP 1 receptors mobilise intracellular Ca 21 , probably through G q [11]. PGE 2 therefore has diverse biological activities depending on the receptors it interacts with and the cells expressing them. With respect to mast cells, mouse mast cells express the EP 3 receptor whose activation induces both degranulation and chemotaxis [12,13]. Human cord blood-derived mast cells, express both EP 2 and EP 3 receptors. In these cells, PGE 2 enhances degranulation and PGD 2 production in cells primed by IL-4, an effect mediated through the EP 3 receptor. However, following FceRI-dependent activation PGE 2 inhibits human cord blood-derived mast cell LTC 4 , PGD 2 , IL-5 and TNF-a production via the EP 2 receptor but has no effect on degranulation [14]. In contrast, several studies have shown that PGE 2 consistently inhibits degranulation and eicosanoid production by HLMC, an effect mediated via the EP 2 receptor [15][16][17]. This receptor also appears to dominate PGE 2dependent signalling on mast cells within the human asthmatic lung as PGE 2 markedly attenuates both the early and late phase airway response to allergen challenge [18].
Since K Ca 3.1 in HLMC is closed by G s -coupled b 2 -adrenoceptors and A 2A adenosine receptors, we hypothesised that activation of the G s -coupled EP 2 receptor would also close K Ca 3.1 in these cells. Furthermore, if PGE 2 was to close K Ca 3.1, then it should inhibit HLMC chemotaxis. To test this hypothesis we have used the patchclamp technique to investigate the effects of PGE 2 on HLMC ion channel function, and investigated the effect of PGE 2 on HLMC migration induced by asthmatic ASM-conditioned medium.

PGE 2 alone does not open K Ca 3.1
We initially examined whether PGE 2 opens K Ca 3.1 in HLMC under resting baseline conditions. HLMC are typically electrically silent at rest and it is possible that PGE 2 could open either K Ca 3.1 or another channel as reported previously for adenosine [10]. No significant change in current amplitude was observed in seven cells following the addition of 10 À5 M PGE 2 (current changed from 1.2270.54 to 1.4270.85 pA at 140 mV; p 5 0.678). PGE 2 closes K Ca 3.1 in the presence of the K Ca 3.1 opener 1-EBIO We then examined whether PGE 2 closes K Ca 3.1. Because PGE 2 might potentially inhibit many IgE-dependent cell activation pathways that could reduce cytosolic-free Ca 21 , and thus reduce K Ca 3.1 activity indirectly, we concentrated on studying the effects of PGE 2 on K Ca 3.1 currents that were induced by the K Ca 3.1 opener 1-EBIO [8][9][10]. This compound opens K Ca 3.1 with a halfmaximal value of about 30 mM for heterologously expressed K Ca 3.1, with a maximal effect at about 300 mM [19]. 1-EBIO is specific for K Ca 3.1 in HLMC and opens it by enhancing the channels sensitivity to [Ca 21 ] i [19]. Thus at 100 mM EBIO, maximal K 1 currents are achieved in the presence of 100 nM free Ca 21 , which is below the resting [Ca 21 ] i of most cell types including HLMC [8].

K Ca 3.1 modulation by PGE 2 is mediated via EP 2 receptors
To examine whether the effects of PGE 2 were mediated via EP 2 prostanoid receptors we examined the effects of EP 2 receptor agonists/antagonists. First, we examined the effects of the selective EP 2 receptor agonist butaprost. Butaprost mimicked the effects of native PGE 2 in a dose-dependent manner ( Fig. 2A-D). At a concentration of 10 À5 M, butaprost reduced the K Ca 3.1 current from 146.0718.3 pA to 61.276.1 pA (p 5 0.00006, n 5 20) (Fig. 2B) with a corresponding shift in reversal from À67.571.4 to À54.473.2 mV (p 5 0.00006) (Fig. 2C). Half maximal suppression (IC 50 ) of K Ca 3.1 by butaprost occurred at approximately 2.1 Â 10 À7 M (n 5 6 cells) (Fig. 2D).

PGE 2 suppresses HLMC migration through the EP 2 receptor
Conditioned medium from asthmatic ASM, which has been activated with TNFa, IFNg and IL-1b, mediates HLMC chemotaxis predominantly via the CXCL10/CXCR3 pathway with additional contributions from ligands for CXCR1 and CXCR3 [20]. Inhibition of K Ca 3.1 by channel blockers markedly suppresses this HLMC chemotaxis [6], as do molecules that close K Ca 3.1 such as adenosine [10]. Migration of HLMC using conditioned medium from asthmatic airway smooth muscle was 2.870.9 fold that of medium control (n 5 4, p 5 0.046) (Fig. 6A) and this was not inhibited significantly by PGE 2 (Fig. 6B). However, the selective EP 2 agonist butaprost produced marked inhibition of HLMC migration to ASM-conditioned medium (Fig. 6B). Interestingly, PGE 2 was chemotactic on its own, but much less potent than in mouse bone marrow-derived mast cells [13] (Fig. 6C). This HLMC chemotactic activity of PGE 2 was markedly increased in the presence of EP 1/2 blockade by AH6809 10 À5 M (Fig. 6D). HLMC migration to both ASM-conditioned medium and PGE 2 itself is therefore attenuated by the EP 2 receptor. PGE 2 attenuates histamine release from cultured HLMC PGE 2 inhibits the degranulation of HLMC freshly isolated from lung tissue. Because the cells used in this study were cultured in stem cell factor (SCF), IL-6 and IL-10, we investigated whether PGE 2 also inhibits release from these HLMC. In keeping with observations in freshly isolated HLMC, PGE 2 inhibited histamine release dose dependently over the concentration range 10 À9 -10 À6 M (control mean7s.e.m. net histamine release 19.873.7% versus 7.573.3% with 10 À6 PGE 2 , n 5 4 donors, p 5 0.046).

K Ca 3.1 is not modulated by G i or G q -coupled receptors
We have now identified three distinct G s -coupled receptors that close K Ca 3.1 in HLMC when exposed to the relevant ligands, and in the case of the b 2 -adrenoceptor, the inverse agonist ICI 118551 opens the channel [9]. In contrast, the G i -coupled CXCR3 receptor does not couple directly to K Ca 3.1 [6]. We have therefore investigated further G i and G q agonists known to have biological effects on HLMC. Platelet activating factor (PAF, 10 À7 M) (G i ), lysophosphatidic acid (LPA, 10 À5 M) (G q,12/13,i ) and UTP (10 À4 M)(G q ) did not open K Ca 3.1 in resting HLMC (n 5 Z6 cells for each agonist) and did not close K Ca 3.1 that had been opened by 1-EBIO (n 5 Z5 cells for each agonist)(data not shown).

Discussion
In this study we have examined the effects of EP 2 receptor activation on K Ca 3.1 ion channel function in HLMC. In keeping with the known effects of the b 2 -adrenoceptor and A 2A adenosine receptor [9,10], the G s -coupled EP 2 receptor also closes this channel reversibly. Consistent with this effect on the channel, EP 2 receptor activation attenuates HLMC migration, and in conse-  quence masks the potential chemotactic activity of PGE 2 on HLMC, which is particularly evident when the EP 2 receptor is blocked pharmacologically. PGE 2 closed K Ca 3.1 reversibly following IgE-dependent mast cell activation demonstrating physiological relevance. PGE 2 also closed K Ca 3.1 channels that had been activated by the K Ca 3.1 opener 1-EBIO. The inhibition of K Ca 3.1 by PGE 2 was reversed both by removing it from the recording solution, and by the addition of the competitive EP 1 and EP 2 receptor antagonist AH6809 indicating a receptor-mediated mechanism. Since the effects of PGE 2 were mimicked by the specific EP 2 receptor agonist butaprost but not the EP 1 agonist 17-phenyl trinor PGE 2 , and no effects of native PGE 2 were seen in the presence of the EP 1/2 receptor antagonist AH6809, we have firm evidence that the suppression of K Ca 3.1 by PGE 2 is mediated via the EP 2 receptor.
Because 1-EBIO opens K Ca 3.1 directly in resting cells by increasing its affinity for Ca 21 [19], it suggests that there is tight coupling between the EP 2 receptor and the channel rather than modulation of intracellular signalling pathways. The ability of PGE 2 to close K Ca 3.1 is in keeping with our previous observations that this channel is closed by G s -coupled b 2 -adrenoceptors [9] and G s -coupled adenosine A 2A receptors [10]. These effects on K Ca 3.1 are not mimicked by cAMP analogues or the activator of adenylate cyclase forskolin [9], and considering they are seen in whole-cell configuration of the patch-clamp technique, indicate that the most likely mechanism of action is membrane-delimited involving the G as or bg subunits of these GPCR. This view is further supported by the observation that the b 2 -adrenoceptor inverse agonist ICI-118551 actually opens K Ca 3.1 [9], but G i agonists such as CXCL10 and PAF, which lower intracellular cAMP, do not directly activate this channel [6].
The ability of GPCR to modify K Ca 3.1 appears to be limited to G s -coupled receptors as we have not found any evidence that G i , G q , or G 12/13 -coupled receptors modify K Ca 3.1 activity. In particular blockade of EP 1 and EP 2 in this study did not uncover any G i -coupled EP 3 effects, and GPCR agonists active on human mast cells such as CXCL10 (G ai ) [6], PAF (G i ), LPA (G q, À12/13 and i ), and UTP (G aq P2Y2 receptor) do not open or close K Ca 3.1 in HLMC. It is well established using cell attached, inside-out and outside patch-clamp recording that most classes of GPCR including G as couple directly to ion channels through membranedelimited mechanisms [21,22]. This may lead to either channel opening or channel closing depending on the channel and receptor, and may utilise either the G a component or specific combinations of bg subunits [21,22]. A further level of specificity between receptor and channel is likely to be achieved by the close approximation of these proteins in tight membrane-restricted signalling complexes. For example, the b 2 adrenoceptor that can modify both K Ca 1.1 and voltage-gated Ca 21 channel gating is associated with these two channels in a macromolecular complex held together by A-kinase-anchoring proteins [23]. It is therefore interesting to speculate that K Ca 3.1 also localises with the various G as receptors that modify its function. This and the exact mechanism by which G as modifies K Ca 3.1 function will be an important area for future research.
Several molecules that attenuate HLMC secretion including PGE 2 , adenosine and b 2 -adrenoceptor agonists increase intracellular cAMP [24]. The generally held view is that this increase in intracellular cAMP couples to inhibition of secretion, supported by the observation that cAMP analogues and non-specific inhibitors of adenylate cyclase can also attenuate secretion from HLMC [24]. However, no mechanism has been identified in mast cells that explains how increases in cAMP inhibit the secretory pathway, and the exclusive role of cAMP in the inhibition of other systems such as smooth muscle relaxation has been challenged [25]. Opening of K Ca 3.1 enhances IgE-dependent Ca 21 influx and degranulation to a submaximal stimulus [8], and its blockade by charybdotoxin attenuates this [7]. Thus, while cAMP plays some role in the inhibition of mast cell mediator release, the demonstration that PGE 2 also closes K Ca 3.1 supports the view that cAMP-independent, K Ca 3.1-dependent mechanisms also contribute in part to EP 2 -dependent inhibition of HLMC mediator release.
In many diseases the recruitment of mast cells to key tissue structures appears critical for their pathophysiological effects [4,26,27]. For example, the contribution of mast cells to the disordered airway physiology of asthma is undoubtedly facilitated by their migration into the airway epithelium [3], submucosal glands [5] and ASM [4]. Inhibition of their migration and subsequent microlocalisation within these structures might therefore offer a novel approach to therapy. Blockade of K Ca 3.1 markedly inhibits HLMC migration in response to a number of diverse chemotactic stimuli including conditioned medium from activated asthmatic ASM [6]. The ability of PGE 2 to close K Ca 3.1 suggested that it should also inhibit HLMC migration. However, its ability to inhibit HLMC migration in response to ASM-conditioned medium was variable and did not reach statistical significance. In contrast, the selective EP 2 receptor agonist butaprost, which was more potent at closing K Ca 3.1 than PGE 2 , produced a marked and consistent inhibition of HLMC migration. PGE 2 was also chemotactic at high concentrations when used alone, but was more potent as a chemoattractant when EP 2 receptors were blocked. This indicates that there are competing pro-migratory and anti-migratory signals when PGE 2 is present, mediated by the EP 2 receptor (inhibitory) and most probably the EP 3 receptor (pro-migratory). These findings are in marked contrast to those in mouse bone marrowderived mast cells in which PGE 2 in isolation is a potent chemoattractant. This chemotactic activity is mediated through the EP 3 receptor, and the lack of inhibition can be attributed to the absence of EP 2 receptors in mouse bone marrow-derived mast cells [13]. In human cord blood-derived mast cells, which express both EP 2 and EP 3 receptors, PGE 2 is not a chemoattractant [13], compatible with an inhibitory role for the EP 2 receptor. In human lung therefore, PGE 2 is not likely to act as a HLMC chemoattractant, in keeping with its anti-inflammatory activities in this tissue [18].
In humans PGE 2 inhibits the early and late asthmatic airway response to allergen challenge [18]. The ability of PGE 2 to close K Ca 3.1 provides a mechanism through which it is able to achieve these effects. It will be of great interest to investigate whether PGE 2 has similar effects on lung T-cell K Ca 3.1 function, cytokine secretion and migration, and whether it can inhibit K Ca 3.1dependent ASM proliferation [28]. Drugs that block K Ca 3.1 are also in development as anti-inflammatory treatments [29]. Our demonstration that the EP 2 receptor closes K Ca 3.1 in HLMC provides further support for the development of specific EP 2 receptor agonists for the treatment of mast cell-mediated pulmonary disease.

Human mast cell purification and culture
All human subjects gave written informed consent, and the study was approved by the Leicestershire Research Ethics Committee. HLMC were dispersed and purified from macroscopically normal lung (n 5 14 donors) obtained within 1 h of resection for lung cancer using immunomagnetic affinity selection as described previously [30]. Final mast cell purity determined by Kimura stain was 499% and viability determined by Trypan blue was 497%. HLMC were cultured in DMEM/ glutamax/HEPES containing antibiotic/antimycotic solution, non-essential amino acids, 10% fetal calf serum, 100 ng/mL SCF, 50 ng/mL IL-6, and 10 ng/mL IL-10 for up to 10 wk as described previously [8,31].

Electrophysiology
The whole-cell variant of the patch-clamp technique was used [7,32]. Patch pipettes were made from borosilicate fibre-containing glass (Clark Electromedical Instruments, Reading, UK), and their tips were heat polished, typically resulting in resistances of 4-6 MO. The standard pipette solution contained (in mM) KCl, 140; MgCl 2 , 2; HEPES, 10; Na 1 -ATP, 2; GTP, 0.1 (pH 7.3). The standard external solution contained (in mM) NaCl, 140; KCl, 5, CaCl 2 , 2; MgCl 2 ,1; HEPES, 10 (pH 7.3). For recording, mast cells were placed in 35-mm dishes containing standard external solution. Whole-cell currents were recorded using an Axoclamp 200A amplifier (Axon Instruments, Foster City, CA, USA), and currents were evoked by applying voltage commands to a range of potentials in 10 mV steps from a holding potential of À20 mV. The currents were digitised (sampled at a frequency of 10 kHz), stored on computer, and subsequently analysed using pClamp software (Axon Instruments). Capacitance transients were minimised using the capacitance neutralisation circuits on the amplifier. Correction for series resistance was not routinely applied. Experiments were performed at 271C, and the temperature controlled by a Peltier device. Experiments were performed with a perfusion system (Automate Scientific, San Francisco, CA) to allow solution changes, although drugs were added directly to the recording chamber. PGE 2 was dissolved in ethanol to give a stock solution of 10 À2 M. Thus, at the maximal concentration PGE 2 used (10 À5 M), the concentration of ethanol in the recording chamber was 0.1%. This concentration of ethanol did not affect K Ca 3.1 currents when tested in isolation.

HLMC chemotaxis
HLMC chemotaxis assays were performed using the Transwell system (BD Biosciences, Oxford, UK) with 24-well plates as described previously [6,20]. Conditioned medium from asthmatic ASM that had been activated with TNFa, IL-1b and IFNg was placed in the lower wells as described previously [20], with appropriate cytokine containing medium in the negative control. PGE 2 was added to the bottom wells in the concentration range 10 À6 -10 À3 M. A total of 1 Â 10 5 HLMC in 100 mL were added to the top well. After incubating the cells for 3 h at 371C, we counted the number of HLMC in the bottom well using Kimura stain in a haemocytometer. HLMC migration was calculated as the fold increase of migrated cells in the test wells compared with the negative control (no chemoattractant in the lower well) as described previously [6,20].

Data presentation and statistical analysis
Data are expressed as mean7SEM. unless otherwise stated. Differences between groups of data were explored using Student's paired or unpaired t-test (two-tailed) as appropriate.