Abstract
The G-protein–coupled protease-activated receptor 2 (PAR2) plays an important role in the pathogenesis of various inflammatory and auto-immune disorders. In airway epithelial cells (AECs), stimulation of PAR2 by allergens and proteases triggers the release of a host of inflammatory mediators to regulate bronchomotor tone and immune cell recruitment. Activation of PAR2 turns on several cell signaling pathways of which the mobilization of cytosolic Ca2+ is likely a critical but poorly understood event. In this study, we show that Ca2+ release-activated Ca2+ (CRAC) channels encoded by stromal interaction molecule 1 and Orai1 are a major route of Ca2+ entry in primary human AECs and drive the Ca2+ elevations seen in response to PAR2 activation. Activation of CRAC channels induces the production of several key inflammatory mediators from AECs including thymic stromal lymphopoietin, IL-6, and PGE2, in part through stimulation of gene expression via nuclear factor of activated T cells (NFAT). Furthermore, PAR2 stimulation induces the production of many key inflammatory mediators including PGE2, IL-6, IL-8, and GM-CSF in a CRAC channel–dependent manner. These findings indicate that CRAC channels are the primary mechanism for Ca2+ influx in AECs and a vital checkpoint for the induction of PAR2-induced proinflammatory cytokines.
Introduction
The epithelial cells of the lung are directly exposed to inhaled air and form the first line of defense against environmental hazards (1–3). In addition to serving as a physical barrier to protect the lung, airway epithelial cells (AECs) play an active role in orchestrating inflammatory effector responses to inhaled substances through the production of a wide array of secreted cytokines and through their interactions with many immune cells (3, 4). Effector responses in AECs are coordinated through a multitude of interactions between extrinsic signaling molecules and intrinsic signal transduction programs activated within the AECs (1, 3). In this signaling repertoire, protease-activated receptor 2 (PAR2) is of particular importance in regulating allergic inflammatory responses that are characteristic of diseases like asthma. PAR2 receptors belong to a family of seven-transmembrane G-protein–coupled receptors (GPCRs) that are widely expressed in a variety of cell types and are activated by cleavage of the extracellular N terminus through the serine protease activity of PAR2 proteolytic agonists. In the airway epithelium, PAR2 receptors are activated by several types of allergens derived from dust mites, cockroach, and fungi, all well-known triggers of asthma, and also by endogenous protease molecules such as trypsin and mast cell tryptase (5–7). PAR2 activation in AECs stimulates the production of several proinflammatory cytokines (IL-6, GM-CSF, and thymic stromal lymphopoietin [TSLP]) and chemokines (IL-8 and eotaxin) (8–10). Moreover, asthmatic patients show increased expression of PAR2 receptors in their airway epithelium, and PAR2-lacking mice show reduced eosinophilic infiltration and airway hyperresponsiveness (11, 12). These findings underscore the importance of PAR2 proteins in mediating allergic inflammatory responses in the airway.
Despite the well-defined importance of PAR2 receptors in driving inflammatory responses, the signal transduction mechanisms involved in PAR2-mediated effector responses are not well understood. PAR2 activation stimulates a multicomponent signal transduction cascade within which the mobilization of Ca2+ by phospholipase Cβ (PLCβ) activation and subsequent inositol 1,4,5-triphosphate (IP3)–mediated release of Ca2+ from endoplasmic reticulum (ER) Ca2+ stores is a key signaling process (13, 14). As a multifunctional second messenger, Ca2+ activates distinct genetic programs and enzymatic cascades to regulate many processes in the immune system including lymphocyte activation, mast-cell degranulation, and neutrophil-mediated bacterial killing (15–18). There is growing interest in the role of cellular Ca2+ as a key second messenger regulating effector responses in the airway (19–21). Yet the functional architecture of the Ca2+ signaling network—the molecular entities and their organization and how this machinery regulates Ca2+ signaling and PAR2-evoked effector responses—remains poorly understood in airway epithelial cells.
In many nonexcitable cells, mobilization of cellular Ca2+ signaling occurs through the opening of store-operated Ca2+ release-activated Ca2+ (CRAC) channels (17, 18). These highly Ca2+ selective ion channels are encoded by the Orai genes (Orai1–3) and activated through direct physical interactions with the ER Ca2+ sensors stromal interaction molecule (STIM) 1 and STIM2 (22). Mechanistically, it is now known that STIM1 and STIM2 sense the [Ca2+]ER and, in response to ER Ca2+ store depletion, translocate to the junctional ER to interact with Orai channels (22, 23). In immune cells, previous studies have established that CRAC channels encoded by STIM1/Orai1 proteins play a central role in driving Ca2+ signaling that controls the function of T cell, mast cells, B cells, and neutrophils (15–18). However, the role of CRAC channels in regulating immune functions of the airway epithelium and, in particular, their contributions to the production and release of inflammatory mediators are unknown.
In this study, we show that primary human AECs show robust store-operated Ca2+ entry (SOCE) mediated by the CRAC channel proteins STIM1 and Orai1. Activation of CRAC channels stimulates several critical effector functions in AECs including gene transcription and production of a range of inflammatory modulators. We further find that stimulation of PAR2 receptors produces intracellular Ca2+ elevations that are critically dependent on CRAC channel activity. In turn, CRAC channel-mediated Ca2+ signals induce the production of several proinflammatory mediators IL-8, IL-6, GM-CSF, and PGE2 in part through stimulation of gene transcription via the nuclear factor of activated T cells (NFAT)–calcineurin signaling axis. These results demonstrate that CRAC channels are a major route of Ca2+ entry in AECs and serve as a key checkpoint for PAR2-mediated generation of inflammatory mediators.
Materials and Methods
Cells and media
Normal human bronchial epithelial (NHBE) cells were purchased from Lonza (Walkersville, MD) and grown in bronchial epithelial cell growth medium containing various growth factors. BEAS-2B, a bronchial epithelial cell line was a kind gift from Curtis Harris (National Cancer Institute) and were cultured in DMEM/F12 medium containing 5% FBS. A549, an alveolar epithelial cell line, was a kind gift from Dr. Jacob Sznajder (Northwestern University). 1Haeo− cells, a bronchial epithelial cell line, was a kind gift from Dr. Alice Prince (Columbia University). Both A549 and 1Haeo− cells were cultured in DMEM containing 10% FBS. All the cells were maintained at 37°C and in 5% CO2.
Plasmids and transfections
The plasmids and RNA interference (RNAi) constructs used in this study were transfected into primary cells and cell lines using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions. E106A Orai1-YFP plasmid has been described previously (24). Human NFATc3-GFP construct was a kind gift of Dr. Anjana Rao (University of California San Diego). NFAT-luciferase and Renilla luciferase plasmids were obtained from Dr. Richard Lewis (Stanford University). RNAi constructs used to downregulate Stim1 and Orai1 protein expression as well as scrambled small interfering RNA (siRNA) control was purchased from Ambion, Life Technologies (SilencerSelect predesigned siRNA). Transfected cells were used for experiments 24–36 h after transfection for cDNA constructs and 48–72 h for RNAi-treated cells.
Reagents and chemicals
The standard extracellular Ringer’s solution had the following composition (in mmol): 150 NaCl, 4.5 KCl, 10 d-glucose, 1 MgCl2, 2 CaCl2, and 5 Na-HEPES. pH was adjusted to 7.4 using NaOH. For the Ca2+-free Ringers solution, CaCl2 was excluded from the above composition and MgCl2 was increased to 3 mmol. The 20 mmol Ca2+ solution used for electrophysiological recordings contained 20 mmol CaCl2 and 130 mmol NaCl; other components of this solution were identical to the standard Ringer’s solution. The divalent-free Ringer’s solution contained (in mmol): 150 NaCl, 10 HEDTA, 1 EDTA, and 10 HEPES (pH 7.4). Stock solutions of 2-aminoethoxydiphenylborane (2-APB), 3,5-bis(trifluoromethyl)pyrazole (BTP2), thapsigargin (TG), RO2959, cyclosporin A (CsA), and FK-506 were dissolved in DMSO. PAR2 agonists including type IX trypsin (Sigma-Aldrich) as well as the agonistic peptides SLIGRL and SLIGKV, control peptide LRGILS, and the PAR1 peptide TFLLR (all from Tocris Bioscience) were constituted in water. 2-APB, CsA, TG, BTP2, and U73122 were from Sigma-Aldrich. RO2959 [difluoro-N-(5-(4-methyl-1-(5-methyl-thiazol-2-yl)-1,2,5,6-tetrahydro-pyridin-3-yl)-pyrazin-2-yl)-benzamide] was from Synta Pharmaceuticals.
Intracellular Ca2+ measurements
NHBE cells and the indicated cell lines were grown on poly-l-lysine–coated glass-bottom dishes (MatTek). Cells were loaded with 2.5 μmol Fura-2 AM (Invitrogen) in the dark for 40 min at room temperature in the appropriate culture medium containing 5–10% FBS. After washing away excess dye, cells were incubated in media for an additional 10 min before initiating Ca2+ imaging. Single-cell [Ca2+]cyt measurements were done according to the protocol described previously (25). Image acquisition and analysis was performed using IPLab (Scanalytics, Rockville, MD) and Slidebook (Denver, CO). For data analysis, regions of interest were drawn around single cells, background subtracted, and the F340/F380 intensity ratios were determined for each time point. The F340/F380 intensity ratios were converted to [Ca2+]cyt using the formula:
where R is the F340/F380 fluorescence intensity ratio and Rmax (= 9.645) and Rmin (= 0.268) were determined by in vitro calibration of Fura-2 pentapotassium salt. β (= 20.236) was determined from the Fmin/Fmax ratio at 380 nm, and Kd is the apparent dissociation constant of Fura-2 binding to Ca2+ (135 nmol). Where applicable, rate of SOCE, “Δ[Ca2+ ]cyt/Δt” was calculated as the slope of a line fitted to the first three points following addition of 2 mmol Ca2+ in the Ca2+ imaging trace. Slope was averaged over several cells in the imaging field to generate the relevant bar graphs.
Western blots
NHBE and BEAS-2B cells were cultured in six-well plates. At ∼70–80% confluency, cells were washed with cold PBS, following which cell lysis and Western blotting was done using protocols described previously (25). Orai1 and STIM1 proteins were detected using affinity-purified polyclonal Abs and peroxidase-labeled secondary Abs (26, 27).
Patch clamp measurements
Patch clamp recordings and analysis of CRAC currents were performed using an Axopatch 200B amplifier (Molecular Devices, Foster City, CA) using standard methods as previously described (28). The holding potential was +30 mV. The standard voltage stimulus consisted of a 100-ms step to −100 mV followed by a 100-ms ramp from −100 to +100 mV applied at 1-s intervals. Cells were pretreated with 1 μmol TG prior to establishment of the seal to deplete stores and activate CRAC channels. Data were leak-subtracted using currents collected in 50 μM La3+ to block CRAC channels.
NFAT-luciferase assay
Endogenous NFAT activity in cells was measured using the PGL3–NFAT luciferase reporter construct (29). NFAT activity was normalized to the activity of Renilla luciferase (pRL-Tk-luc). Plasmids were transfected into AECs in a ratio of 20:1 (NFAT/Renilla). At 24 h after transfection, cells were pretreated with the CRAC channel inhibitors BTP-2 and RO2959 for 1 h followed by treatment with 0.5 μmol TG and 50 nmol phorbol 12,13-dibutyrate (PDBu) for 8–10 h. Cells were lysed, and the luciferase activity was determined using the Dual-luciferase Reporter Assay Kit (Promega) and a single-tube luminometer (Berthold Instruments, Germany).
NFAT translocation assay
BEAS-2B cells were transfected with NFATc3-GFP construct and imaged for GFP 24–36 h later. Images were taken in resting 2 mmol Ca2+ solution and after depleting ER Ca2+ stores with TG (1 μmol, 20 min). The percentage of cells showing NFAT localization to the nucleus was calculated, averaged between several fields, and reported as a bar graph.
Analysis of cytokine secretion
AECs (NHBE, BEAS, A549, and 1Haeo− cells) were cultured in 24-well plates. Twelve hours before stimulation, the cell culture medium was changed to one containing 2 mmol Ca2+. Cells were treated with specific CRAC channel–activating stimuli (TG and PDBu or PAR2 agonists SLIGRL and SLIGKV), and the levels of various inflammatory mediators were measured using custom ELISA kits (RayBiotech for TNF-α, IL-6, IL-1α, and GM-CSF; Cayman Chemical for PGE2; R&D Systems for TSLP; and Life Technologies for IL-8 and IL-1β).
Data analysis
All bar graphs summarizing the data are reported as mean ± SEM. For data sets involving more than two groups, initial statistical analysis was performed using ANOVA with a confidence interval of 5%. This was followed by two-tailed paired Student t test for comparing different treatment conditions within the set. For data sets with only two groups, two-tailed paired Student t test was used to compare between control and test conditions.
Results
A screen for GPCR agonists identifies PAR2 receptors as activators of SOCE in the airway epithelium
In what context would CRAC channels be relevant for Ca2+ signaling in AECs? CRAC channels are voltage independent and require depletion of ER Ca2+ stores for activation (17). There are numerous extracellular stimuli (e.g., growth factors, pathogens, GPCR ligands, etc.) that act in this manner and could be regarded as potential candidates that stimulate CRAC channels in AECs. To examine this, we screened several ligands that are coupled to the generation of IP3 in NHBE cells. Although several receptors coupled to PLC–IP3 signaling are thought to be expressed in AECs (30), this screen revealed that only a subset of ligands, including agonists for PAR2 and purinergic P2Y receptors (and to a smaller extent, thrombin and bradykinin), evoked Ca2+ signals consistent with SOCE (Table I). PAR2 receptors are important for the ability of AECs to sense allergens and endogenous proteases and for mediating the subsequent inflammatory responses in the lung (31). We therefore sought to define the mechanisms and functional consequences of CRAC channel signaling in AECs following PAR2 activation.
Ligand . | Known Receptor . | Receptor Type . | SOCE Activation? . |
---|---|---|---|
PARs | |||
Type IX trypsin | PAR2 | GPCR | ++ |
SLIGKV, SLIGRL | PAR2 | GPCR | ++ |
TFLLR | PAR1 | GPCR | − |
Thrombin | PAR1, 3, 4 | GPCR | + |
Pathogens | |||
PamC3, LTA | TLR2 | TLR | − |
pI:C | TLR3 | TLR | − |
LPS | TLR4 | TLR | − |
Purinergic agonists | |||
ATP | P2X, P2Y | GPCR | ++ |
UTP | P2Y2, P2Y4 | GPCR | ++ |
UDP-glucose | P2Y14 | GPCR | − |
Other ligands | |||
Bradykinin | BK1, 2 | GPCR | + |
PGE2 | EP1 | GPCR | − |
Platelet-activating factor | PAFR | GPCR | − |
Epidermal growth factor | EGFR | GPCR | − |
Salbutamol | β2 adrenergic | GPCR | − |
Denatonium | Bitter taste receptors | GPCR | − |
tBHP (oxidative stress) | — | STIM1, IP3 | − |
Ligand . | Known Receptor . | Receptor Type . | SOCE Activation? . |
---|---|---|---|
PARs | |||
Type IX trypsin | PAR2 | GPCR | ++ |
SLIGKV, SLIGRL | PAR2 | GPCR | ++ |
TFLLR | PAR1 | GPCR | − |
Thrombin | PAR1, 3, 4 | GPCR | + |
Pathogens | |||
PamC3, LTA | TLR2 | TLR | − |
pI:C | TLR3 | TLR | − |
LPS | TLR4 | TLR | − |
Purinergic agonists | |||
ATP | P2X, P2Y | GPCR | ++ |
UTP | P2Y2, P2Y4 | GPCR | ++ |
UDP-glucose | P2Y14 | GPCR | − |
Other ligands | |||
Bradykinin | BK1, 2 | GPCR | + |
PGE2 | EP1 | GPCR | − |
Platelet-activating factor | PAFR | GPCR | − |
Epidermal growth factor | EGFR | GPCR | − |
Salbutamol | β2 adrenergic | GPCR | − |
Denatonium | Bitter taste receptors | GPCR | − |
tBHP (oxidative stress) | — | STIM1, IP3 | − |
NHBE cells were treated with the indicated ligands in a Ca2+-free Ringer’s solution. Extracellular Ca2+ (2 mmol) was then restored and the amplitude of the [Ca2+]cyt rise was examined. Cells were considered responders if the [Ca2+]cyt elevation was >2× SEM above the resting [Ca2+]cyt.
−, no response; +, response 2–3-fold above resting [Ca2+]i; ++, response was >3-fold above resting [Ca2+]i; LTA, lipoteichoic acid; PamC3, synthetic triacylated lipopeptide; pI:C, polyinosinic-polycytidylic acid; tBHP, tert-butyl hydroperoxide.
PAR2 activation stimulates Ca2+ signaling in AECs by activating CRAC channels
PAR2 activation initiates many cell signaling cascades including IP3-mediated mobilization of [Ca2+]cyt (13). However, the pathways involved in generating PAR2 mediated [Ca2+]cyt elevations, and the functional significance of these Ca2+ elevations in AECs are unclear. To address the potential role of store-operated CRAC channels in generating PAR2-mediated Ca2+ signals in AECs, we examined the ability of PAR2 ligands to elicit SOCE by imaging cytosolic Ca2+ transients using the Ca2+ indicator, Fura-2. Activation of PAR2 by the PAR2-activating peptides SLIGRL and SLIGKV caused a biphasic [Ca2+]cyt elevation that included a rapid initial rise in [Ca2+]cyt followed by sustained Ca2+ oscillations in NHBE cells (Fig. 1A, 1E and data not shown). The rise in [Ca2+]cyt following add-back of extracellular Ca2+ indicates that Ca2+ entry across the plasma membrane is necessary for sustaining PAR2-mediated Ca2+ elevations (Fig. 1A). The Ca2+ response was specific to PAR2 activation, because administration of control peptide LRGILS or PAR1-specific peptide TFLLR failed to elicit Ca2+ signaling (Fig. 1E). Response to SLIGKV was also abolished when cells were pretreated with the PLC inhibitor U73122, consistent with PAR2 activating a Gq-mediated PLC signaling pathway (Fig. 1E). Importantly, the CRAC channel inhibitors BTP2 and RO2959 (32, 33) attenuated both the number of oscillations and the amplitude of the Ca2+ response to PAR2 peptides (Fig. 1B–E). These results indicate that PAR2 stimulation evokes a biphasic Ca2+ signal the sustained component of which arises from CRAC channel activation.
Store-operated CRAC channels regulate Ca2+ signals induced by PAR2 activation. (A) The PAR2 agonist SLIGRL induces oscillatory Ca2+ signals due to SOCE. NHBE cells were treated with 100 μmol SLIGRL in a nominally Ca2+-free solution to deplete internal stores. Total of 2 mmol Ca2+ was readded to the external bath solution to activate SOCE. The Ca2+ traces show responses of individual cells in the imaging field. (B) Pretreatment with the CRAC channel inhibitor BTP2 does not affect store release but abrogates SOCE. Cells were treated with 500 nmol BTP2 for 1.5 h (A and B; data are mean ± SEM of n = 18–21 cells; representative of three independent experiments). (C) Bar graph summarizing the percentage of cells showing Ca2+ responses after readdition of external Ca2+. (D) Summary of the amplitude of Ca2+ response at 840 s (data are mean ± SEM of n = 38–45 cells; three independent experiments). (E) Summary of the average elevation in [Ca2+]cyt in response to the PAR2-activating peptide SLIGKV in Ca2+ and nominally Ca2+-free solutions, control peptide LRGILS, and PAR1-activating peptide TFLLR. The effects of CRAC channel inhibitor RO2959 and PLC inhibitor U73122 on SLIGKV-mediated Ca2+ elevation are also indicated. The [Ca2+]cyt was measured 156–192 s after addition of various peptides to NHBE cells with 2 mmol Ca2+ in the external bath (data are mean ± SEM of n = 12–23 cells; three independent experiments). (F) Effects of siRNA knockdown of STIM1 and Orai1 on SLIGKV-induced Ca2+ signals in BEAS-2B cells (data are mean ± SEM of n = 24–31 cells; representative of three independent experiments). The traces reflect the mean Ca2+ concentration of all cells in the imaging field. (G) Summary of mean [Ca2+]cyt levels 748 s after addition of ligand. Statistical significance was determined by unpaired two-sample Student t test). The following standardized notation for statistical significance (p values) is used in all the figures: *p < 0.05, **p < 0.01, ***p < 0.001. Con, control.
Store-operated CRAC channels regulate Ca2+ signals induced by PAR2 activation. (A) The PAR2 agonist SLIGRL induces oscillatory Ca2+ signals due to SOCE. NHBE cells were treated with 100 μmol SLIGRL in a nominally Ca2+-free solution to deplete internal stores. Total of 2 mmol Ca2+ was readded to the external bath solution to activate SOCE. The Ca2+ traces show responses of individual cells in the imaging field. (B) Pretreatment with the CRAC channel inhibitor BTP2 does not affect store release but abrogates SOCE. Cells were treated with 500 nmol BTP2 for 1.5 h (A and B; data are mean ± SEM of n = 18–21 cells; representative of three independent experiments). (C) Bar graph summarizing the percentage of cells showing Ca2+ responses after readdition of external Ca2+. (D) Summary of the amplitude of Ca2+ response at 840 s (data are mean ± SEM of n = 38–45 cells; three independent experiments). (E) Summary of the average elevation in [Ca2+]cyt in response to the PAR2-activating peptide SLIGKV in Ca2+ and nominally Ca2+-free solutions, control peptide LRGILS, and PAR1-activating peptide TFLLR. The effects of CRAC channel inhibitor RO2959 and PLC inhibitor U73122 on SLIGKV-mediated Ca2+ elevation are also indicated. The [Ca2+]cyt was measured 156–192 s after addition of various peptides to NHBE cells with 2 mmol Ca2+ in the external bath (data are mean ± SEM of n = 12–23 cells; three independent experiments). (F) Effects of siRNA knockdown of STIM1 and Orai1 on SLIGKV-induced Ca2+ signals in BEAS-2B cells (data are mean ± SEM of n = 24–31 cells; representative of three independent experiments). The traces reflect the mean Ca2+ concentration of all cells in the imaging field. (G) Summary of mean [Ca2+]cyt levels 748 s after addition of ligand. Statistical significance was determined by unpaired two-sample Student t test). The following standardized notation for statistical significance (p values) is used in all the figures: *p < 0.05, **p < 0.01, ***p < 0.001. Con, control.
Previous evidence indicates that the CRAC channel proteins Orai1 and STIM1 are expressed in the lung (21, 34), raising the possibility that these proteins may contribute to Ca2+ signaling in AECs. We therefore sought to investigate the contribution of STIM1 and Orai1 to PAR2 evoked Ca2+ signals. siRNA-mediated knockdown of STIM1 and Orai1 proteins significantly reduced their expression and inhibited the sustained Ca2+ signals in response to the PAR2-activating peptide SLIGKV in the bronchial cell line BEAS-2B (Fig. 1F, 1G, Supplemental Fig. 1A). Similarly, overexpression of a dominant-negative pore mutant of CRAC channels, E106A Orai1 (35), significantly reduced SLIGKV-activated Ca2+ response (data not shown). These results indicate that CRAC channels encoded by STIM1 and Orai1 are essential for mobilizing PAR2-evoked Ca2+ signals.
PAR2 activation by the serine protease trypsin stimulates CRAC channel-mediated Ca2+ signals in AECs
In the airway epithelium, PAR2 is activated by endogenous proteases including trypsin, mast cell tryptase, and human airway trypsin-like proteases that are released during inflammation and injury (31). We therefore tested the effects of trypsin on Ca2+ signaling in AECs. Administration of type IX trypsin produced a sustained Ca2+ signal only when Ca2+ was present in the external bath solution. The sustained component of the [Ca2+]cyt elevation was abrogated by the CRAC channel inhibitor BTP2 (Fig. 2A, 2B) without affecting store release. Likewise, overexpression of E106A Orai1, which functions as a dominant-negative inhibitor of CRAC channels (35) as well as the knockdown of CRAC channel proteins STIM1 and Orai1, attenuated trypsin-induced Ca2+ signal (Fig. 2C, 2D). Taken together, these results show that PAR2 activation in response to both synthetic peptides and enzymatic cleavage by trypsin mobilizes Ca2+ signaling in primary human AECs through activation of CRAC channels.
The PAR2 agonist, type IX trypsin, evokes Ca2+ elevations by activating CRAC channels. (A) Type IX trypsin (100 nmol) activates SOCE in NHBE cells. SOCE was activated using the same protocol as in Fig. 1A. The CRAC channel inhibitor BTP2 abrogates trypsin-induced Ca2+ elevations in NHBE cells (mean ± SEM of n = 17–25 cells, three independent experiments). Slope of the line (used to calculate rate of SOCE) following addition of 2 mmol Ca2+ is indicated for the control condition. (B) Summary of the rate of Ca2+ influx for SOCE shown in (A). (C) siRNA knockdown of STIM1 or Orai1 inhibits trypsin-mediated SOCE in BEAS-2B cells. (D) Summary of the rates of trypsin-mediated Ca2+ influx shown in (C) and in cells transfected with dominant-negative E106A Orai1 (mean ± SEM of n = 16–26 cells; three independent experiments). **p < 0.01, ***p < 0.001. Con, control.
The PAR2 agonist, type IX trypsin, evokes Ca2+ elevations by activating CRAC channels. (A) Type IX trypsin (100 nmol) activates SOCE in NHBE cells. SOCE was activated using the same protocol as in Fig. 1A. The CRAC channel inhibitor BTP2 abrogates trypsin-induced Ca2+ elevations in NHBE cells (mean ± SEM of n = 17–25 cells, three independent experiments). Slope of the line (used to calculate rate of SOCE) following addition of 2 mmol Ca2+ is indicated for the control condition. (B) Summary of the rate of Ca2+ influx for SOCE shown in (A). (C) siRNA knockdown of STIM1 or Orai1 inhibits trypsin-mediated SOCE in BEAS-2B cells. (D) Summary of the rates of trypsin-mediated Ca2+ influx shown in (C) and in cells transfected with dominant-negative E106A Orai1 (mean ± SEM of n = 16–26 cells; three independent experiments). **p < 0.01, ***p < 0.001. Con, control.
STIM1 and Orai1 contribute to SOCE in human AECs
The ability of CRAC channel antagonists to suppress PAR2-evoked Ca2+ signals implies that human AECs express the machinery for CRAC channels. To directly address this issue, we next examined the pharmacology, biophysical characteristics, and molecular basis of CRAC channels in AECs. Following the depletion of ER Ca2+ stores with the SERCA pump inhibitor TG in a nominally Ca2+-free solution, robust Ca2+ influx was seen when Ca2+ was added to the external bath in primary bronchial (NHBE) cells (Fig. 3A). SOCE in NHBE cells exhibited pharmacological properties consistent with CRAC channels including acute inhibition by 2 μmol La3+, potentiation by 5 μmol 2-APB, and inhibition by a higher dose of 2-APB (50 μmol) (Fig. 3A–C). Moreover, SOCE was significantly inhibited by the CRAC channel antagonists BTP2 and RO2959 (Fig. 3D, 3F). Similar results were obtained in the bronchial cell lines BEAS-2B and 1Haeo− as well as the alveolar epithelial cell line A549, suggesting that SOCE is widespread in lower AECs (Fig. 3I, Supplemental Fig. 1B, 1C). The rates of Ca2+ influx following Ca2+ add-back reflect store-operated channel activity and are summarized in Fig. 3G. These findings are consistent with recent reports indicating presence of CRAC channels in the bronchial cell lines 16HBE and CFBE41o− (21, 34).
Characterization of store-operated CRAC channels in human AECs. (A–C) ER Ca2+ stores were depleted with 1 μmol TG in a nominally Ca2+-free solution. Readdition of 2 mmol Ca2+ to the external bath solution evoked SOCE. (A) SOCE is acutely blocked by 2 μmol La3+. (B) A low dose of 2-APB (5 μmol) facilitated SOCE. (C) A high dose of 2-APB (50 μmol) inhibited SOCE. (D) Pretreatment with the CRAC channel inhibitor BTP2 (500 nmol) abolished SOCE in NHBE cells [(A–D); mean ± SEM of n = 18–29; representative of three independent experiments]. (E) Knockdown of STIM1 and Orai1 expression in NHBE cells significantly inhibited SOCE (n = 21–26 cells; three independent experiments). Summary of the rates of SOCE in NHBE cells following treatment with the CRAC channel inhibitors BTP2 (500 nmol) and RO2959 (500 nmol) and expression of dominant-negative E106A Orai1 (n = 21–42 cells; three experiments) (F) and following siRNA-mediated knockdown of STIM1-2, Orai 1–3 (mean ± SEM of n = 22–34; three experiments) (G). (H) Western blots showing expression of Orai1 and STIM1 in NHBE cells. (I) Summary of rates of SOCE in BEAS-2B cells (n = 31–48 cells; three independent experiments). (J) Whole-cell patch-clamp recordings of ICRAC in BEAS-2B cells. Cells were pretreated with TG to deplete ER Ca2+ stores. The external solution was periodically switched between 20 mmol Ca2+ and a divalent-free (DVF solution). In 20 mmol Ca2+o, the current exhibits a current-voltage (I-V) relationship with strong inward rectification and positive reversal potential (>+60 mV), consistent with the known biophysical hallmarks of CRAC channels (37). *p < 0.05, **p < 0.01, ***p < 0.001. Con, control.
Characterization of store-operated CRAC channels in human AECs. (A–C) ER Ca2+ stores were depleted with 1 μmol TG in a nominally Ca2+-free solution. Readdition of 2 mmol Ca2+ to the external bath solution evoked SOCE. (A) SOCE is acutely blocked by 2 μmol La3+. (B) A low dose of 2-APB (5 μmol) facilitated SOCE. (C) A high dose of 2-APB (50 μmol) inhibited SOCE. (D) Pretreatment with the CRAC channel inhibitor BTP2 (500 nmol) abolished SOCE in NHBE cells [(A–D); mean ± SEM of n = 18–29; representative of three independent experiments]. (E) Knockdown of STIM1 and Orai1 expression in NHBE cells significantly inhibited SOCE (n = 21–26 cells; three independent experiments). Summary of the rates of SOCE in NHBE cells following treatment with the CRAC channel inhibitors BTP2 (500 nmol) and RO2959 (500 nmol) and expression of dominant-negative E106A Orai1 (n = 21–42 cells; three experiments) (F) and following siRNA-mediated knockdown of STIM1-2, Orai 1–3 (mean ± SEM of n = 22–34; three experiments) (G). (H) Western blots showing expression of Orai1 and STIM1 in NHBE cells. (I) Summary of rates of SOCE in BEAS-2B cells (n = 31–48 cells; three independent experiments). (J) Whole-cell patch-clamp recordings of ICRAC in BEAS-2B cells. Cells were pretreated with TG to deplete ER Ca2+ stores. The external solution was periodically switched between 20 mmol Ca2+ and a divalent-free (DVF solution). In 20 mmol Ca2+o, the current exhibits a current-voltage (I-V) relationship with strong inward rectification and positive reversal potential (>+60 mV), consistent with the known biophysical hallmarks of CRAC channels (37). *p < 0.05, **p < 0.01, ***p < 0.001. Con, control.
We next sought to resolve the molecular machinery of CRAC channels in AECs. CRAC channels are activated by STIM proteins (STIM1 and -2) and encoded by the Orai proteins (Orai1, -2, and -3) (36). Western blots showed the expression of both STIM1 and Orai1 proteins in NHBE and BEAS-2B cells (Fig. 3H, Supplemental Fig. 1A). Overexpression of dominant-negative E106A Orai1 significantly abrogated SOCE in various AECs (Fig. 3F, 3I). Moreover, knockdown of STIM1 and Orai1 using siRNA significantly reduced the protein expression and inhibited both the rate and the amplitude of SOCE in NHBE cells and various cell lines (Fig. 3E, 3G, 3I, Supplemental Fig. 1B, 1C, and data not shown). siRNA-mediated knockdown of Orai2, Orai3, and STIM2 did not significantly affect SOCE, indicating that these isoforms do not appreciably contribute to SOCE in human AECs (Fig. 3G). We were unable to directly study the expression of Orai2 and Orai3 isoforms due to nonspecificity of the commercially available Abs we tested (data not shown). Nevertheless, together with the functional evidence presented in the previous section, these results indicate that the CRAC channel proteins STIM1 and Orai1 are essential for conferring SOCE in human AECs. We therefore focused our studies on the functional roles of STIM1 and Orai1 in the airway epithelium.
CRAC channels have a distinct electrophysiological profile characterized by an inwardly rectifying current-voltage (I-V) relationship, low permeability to large monovalent cations such as Cs+, and depotentiation in divalent-free solutions (37). Patch-clamp recordings of store-operated currents in BEAS-2B cells following treatment with TG to deplete ER Ca2+ stores showed the presence of a cation current consistent with these well-known properties of CRAC channels (17, 25) (Fig. 3J). In 20 mmol Ca2+o, the I-V relationship showed strong inward rectification and an extremely positive reversal potential. In divalent-free solution, the I-V relationship revealed a reversal potential of ∼+50 mV, indicating low permeability to internal Cs+ (Fig. 3J). Similar results were seen in the airway epithelial cell line A549 (data not shown). These results indicate that SOCE in airway cells exhibits the canonical biophysical properties of CRAC channels encoded by STIM1 and Orai1.
CRAC channel activation stimulates cytokine and chemokine production in AECs
AECs play a vital role in host defense by producing inflammatory mediators such as TSLP, IL-6, PGE2, IL-33, IL-8, and RANTES in response to a variety of stimuli including infectious agents, allergens, and other bioactive molecules (38–41). Although Ca2+ signals have been implicated in the generation of inflammatory mediators from AECs, the Ca2+ signaling pathways that mediate this process remain poorly understood (7, 20, 42).
To address a potential role for CRAC channels in the production of inflammatory mediators from AECs, we initially performed a multiplex cytokine screen in NHBE cells following activation of CRAC channels by the SERCA inhibitor TG (Supplemental Fig. 2). Previous studies have shown that Ca2+ signaling pathways in immune cells including T cells often work in synergy with protein kinase C (PKC)–dependent signaling to induce cytokines (43). Therefore, in these experiments, we also costimulated PKC by administering the phorbol ester, PDBu. Key hits from this screen were further examined by individual ELISA kits. These tests revealed that several key inflammatory mediators are regulated by CRAC channel activation in NHBE cells, including TSLP, IL-6, TNF-α, IL-1β, and GM-CSF (Fig. 4A, 4C–F). CRAC channel activation also induced the production of the arachidonic acid metabolite PGE2 from AECs (Fig. 4B). We also confirmed IL-6, TNF-α, and PGE2 production in various AEC lines (Supplemental Fig. 3A–C). The production of these inflammatory mediators was prevented in Ca2+-free medium and by the CRAC channel inhibitor BTP2 (Fig. 4A–F). Furthermore, knockdown of STIM1 and Orai1 inhibited CRAC channel–induced production of TNF-α and PGE2 (Fig. 4B, 4C). Thus, these results implicate a causal role for Ca2+ influx through CRAC channels in the production of inflammatory cytokines from AECs. Interestingly, whereas maximal induction of TNF-α and IL-6 required costimulation of NHBE cells with both TG and PDBu, TSLP and PGE2 were induced maximally by TG alone, suggesting that Ca2+ signals differentially regulate cytokine production depending on the presence of specific costimulatory signals.
CRAC channel activation induces production of inflammatory mediators in human AECs. (A–H) Production of cytokines and chemokines as detected by ELISA. AECs were stimulated with either TG alone (0.25 μmol) or TG (0.25 μmol) and PDBu (50 nmol), in the absence or presence of the CRAC channel inhibitor BTP2 (250 nmol), and in nominally Ca2+-free media. Cells treated with solvent alone (DMSO) were used as control. Cell-culture supernatants were collected 8–12 h after treatment for all cytokines and after 48 h for TSLP (n = 3 to 4 replicates; representative of three independent experiments). Unless otherwise noted, all experiments were in NHBE cells. *p < 0.05, **p < 0.01, ***p < 0.001. siCon, small interfering control.
CRAC channel activation induces production of inflammatory mediators in human AECs. (A–H) Production of cytokines and chemokines as detected by ELISA. AECs were stimulated with either TG alone (0.25 μmol) or TG (0.25 μmol) and PDBu (50 nmol), in the absence or presence of the CRAC channel inhibitor BTP2 (250 nmol), and in nominally Ca2+-free media. Cells treated with solvent alone (DMSO) were used as control. Cell-culture supernatants were collected 8–12 h after treatment for all cytokines and after 48 h for TSLP (n = 3 to 4 replicates; representative of three independent experiments). Unless otherwise noted, all experiments were in NHBE cells. *p < 0.05, **p < 0.01, ***p < 0.001. siCon, small interfering control.
In addition to the cytokines, CRAC channel activation also caused significant increase in the production of chemokines IL-8 and RANTES, which was blocked by the CRAC channel inhibitors BTP2 and RO2959 (Fig. 4G, 4H). The multiplex cytokine screen further revealed that although AECs also produce IFN-γ and CXCL1, these factors are not induced by CRAC channels, indicating that only a subset of inflammatory mediators produced by AECs are regulated by CRAC channels. Collectively, these results establish a key role for CRAC channels as an important route of Ca2+ entry for the production of cytokines and chemokines in the airway epithelium.
PAR2 agonists activate CRAC channels to induce cytokine production in AECs
Previous studies have established that PAR2 activation in AECs induces the production of a wide range of cytokines including IL-6, GM-CSF, TSLP, and chemokines IL-8 and eotaxin, the growth factor PDGF, arachidonic metabolites PGE2 and PGD2, as well as the enzyme MMP-9 (8–10, 40, 44). The dual findings that PAR2 activation mobilizes Ca2+ signals through CRAC channels (Figs. 1, 2) and that CRAC channel activation with TG stimulates cytokine production (Figs. 4, 5) therefore led us to consider a role for CRAC channels in the production of inflammatory mediators by PAR2 agonists. These tests revealed that administration of the PAR2 agonist peptide SLIGRL resulted in a significant increase in the production of PGE2, IL-6, GM-CSF, and IL-8 (Fig. 5A, 5C, 5E, 5F). PGE2 was induced in 2 h, whereas induction of IL-6, GM-CSF, and IL-8 occurred after 8–24 h of PAR2 activation. The PAR2-mediated production of these factors was significantly inhibited by the CRAC channel antagonist BTP2 (Fig. 5A, 5C, 5E, 5F). Interestingly, PAR2 activation by SLIGRL did not evoke TNF-α induction (not shown). This is in contrast to cell stimulation with TG/PDBu, which was very effective in evoking TNF-α production (Fig. 4D). The reasons for this difference need to be further investigated, but could be because the amplitude and duration of Ca2+ elevations in response to PAR2 activation are insufficient to induce TNF-α. PGE2 and IL-6 induction in response to SLIGRL were significantly inhibited by combined knockdown of STIM1 and Orai1 (Fig. 5B, 5D). Knockdown of STIM1 or Orai1 alone, however, did not significantly diminish the production of these cytokines (data not shown). We suspect that this is due to the low knockdown efficiency of protein expression in NHBE cells (∼50% for both proteins as seen by Western blot, Fig. 3H), which would be expected to only partially inhibit SOCE. Similarly, trypsin produced a significant increase in PGE2 production that was attenuated by the CRAC channel blocker BTP2 (Fig. 5A). Thus, these findings strongly suggest that CRAC channels contribute to the production and release of inflammatory modulators in response to PAR2 activation.
CRAC channels regulate generation of inflammatory mediators in response to PAR2 activation in human AECs. NHBE cells were stimulated with the PAR2-activating peptide SLIGRL (100 μmol) in the absence or presence of BTP2 or following knockdown of both STIM1 and Orai1. Cell-culture supernatants were collected after 2 h (for PGE2) or 24 h (for IL-6, IL-8, and GM-CSF measurements), and levels of PGE2, IL-6, IL-8, and GM-CSF were determined by cytokine-specific ELISA kits. (A) PGE2 induction by SLIGRL and type IX trypsin (100 nmol) is inhibited by BTP2 and U73122. A scrambled peptide (LRGILS) and a PAR1-specific peptide (TFLLR) had no effect on PGE2 induction. (B) Inhibition of PGE2 secretion by knockdown of STIM1 and Orai1. Inhibition of IL-6 secretion by BTP2 (C) and siRNA-mediated knockdown of STIM1 and Orai1 (D). (E and F) Effect of BTP2 on IL-8 and GM-CSF induction (n = 3 replicates, three independent experiments). *p < 0.05, **p < 0.01. siCon, small interfering control.
CRAC channels regulate generation of inflammatory mediators in response to PAR2 activation in human AECs. NHBE cells were stimulated with the PAR2-activating peptide SLIGRL (100 μmol) in the absence or presence of BTP2 or following knockdown of both STIM1 and Orai1. Cell-culture supernatants were collected after 2 h (for PGE2) or 24 h (for IL-6, IL-8, and GM-CSF measurements), and levels of PGE2, IL-6, IL-8, and GM-CSF were determined by cytokine-specific ELISA kits. (A) PGE2 induction by SLIGRL and type IX trypsin (100 nmol) is inhibited by BTP2 and U73122. A scrambled peptide (LRGILS) and a PAR1-specific peptide (TFLLR) had no effect on PGE2 induction. (B) Inhibition of PGE2 secretion by knockdown of STIM1 and Orai1. Inhibition of IL-6 secretion by BTP2 (C) and siRNA-mediated knockdown of STIM1 and Orai1 (D). (E and F) Effect of BTP2 on IL-8 and GM-CSF induction (n = 3 replicates, three independent experiments). *p < 0.05, **p < 0.01. siCon, small interfering control.
PAR2 receptor stimulation results in IL-6 and IL-8 production through the activation of NFAT
By what mechanism does Ca2+ influx through CRAC channels regulate PAR2-mediated production of cytokines? In many immune cells, cytokine induction is regulated transcriptionally by the transcription factor NFAT (45, 46). Ca2+ elevations activate NFAT through the protein phosphatase calcineurin, which then causes NFAT to move into the nucleus and bind to target DNA sequences (43). Although AECs express NFAT proteins (47), the specific targets of NFAT and the physiological relevance of this signaling remains unknown. To explore whether induction of cytokines by PAR2 activation occurs through the calcineurin–NFAT pathway, we measured PAR2-induced cytokine levels in the presence of the calcineurin inhibitor CsA, a widely used small-molecule inhibitor of calcineurin/NFAT signaling (48). Induction of both IL-6 and IL-8 was significantly inhibited by CsA (Fig. 6A). By contrast, the NF-κB inhibitor caffeic acid had no effect on induction of IL-6 or IL-8, indicating that under these experimental conditions, the induction of the cytokines primarily occurs through NFAT-dependent gene transcription (Supplemental Fig. 3D). Moreover, induction of PGE2 and GM-CSF was not affected by CsA, suggesting that the induction of these cytokines in response to PAR2 activation is not significantly regulated by calcineurin/NFAT signaling. These findings highlight a novel role for NFAT-dependent gene expression in the induction of specific cytokines and chemokines in response to PAR2 signaling.
Calcineurin–NFAT signaling regulates PAR2 and CRAC channel–mediated induction of inflammatory mediators from AECs. (A) Summary of the effect of calcineurin inhibitor CsA (500 nmol) on IL-6, IL-8, PGE2, and GM-CSF induction in response to the PAR2 agonistic peptide SLIGRL (100 μmol) in NHBE cells. Cells were stimulated for 2 h for PGE2 induction, 8 h for IL-6 and IL-8 induction, and 24 h for GM-CSF induction. Data are shown as fold change in cytokine production (actual concentrations: IL-6: 18–70 pg/ml; IL-8: 55–200 pg/ml; PGE2: 2300–4500 pg/ml; GM-CSF: 75–220 pg/ml). (B) Summary of the change in cytokine levels in NHBE cells (12 h stimulation) evoked by CRAC channel activation. NHBE cells were treated with TG (250 nmol) and PDBu (50 nmol) in the absence and presence of CsA (250 nmol). n = 3 replicates, three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001.
Calcineurin–NFAT signaling regulates PAR2 and CRAC channel–mediated induction of inflammatory mediators from AECs. (A) Summary of the effect of calcineurin inhibitor CsA (500 nmol) on IL-6, IL-8, PGE2, and GM-CSF induction in response to the PAR2 agonistic peptide SLIGRL (100 μmol) in NHBE cells. Cells were stimulated for 2 h for PGE2 induction, 8 h for IL-6 and IL-8 induction, and 24 h for GM-CSF induction. Data are shown as fold change in cytokine production (actual concentrations: IL-6: 18–70 pg/ml; IL-8: 55–200 pg/ml; PGE2: 2300–4500 pg/ml; GM-CSF: 75–220 pg/ml). (B) Summary of the change in cytokine levels in NHBE cells (12 h stimulation) evoked by CRAC channel activation. NHBE cells were treated with TG (250 nmol) and PDBu (50 nmol) in the absence and presence of CsA (250 nmol). n = 3 replicates, three independent experiments. *p < 0.05, **p < 0.01, ***p < 0.001.
NFAT regulates cytokine and chemokine production in response to CRAC channel activation
To explore the role of NFAT in CRAC channel–mediated cytokine production more directly, we stimulated NHBE cells with TG and PDBu for 12 h in the presence or absence of CsA and evaluated inflammatory mediator production by ELISA. CsA caused an almost complete inhibition of TNF-α and IL-6 production, implicating a key role for NFAT in induction of these cytokines (Fig. 6B). CsA also partially inhibited production of PGE2 and IL-8, suggesting that NFAT contributes to the production of these mediators (Fig. 6B). The partial inhibition of PGE2 induction in this case stands in contrast to the lack of CsA effect seen following 2 h of PAR2 stimulation (Fig. 6A), suggesting that at longer time points, PGE2 production is influenced by NFAT-mediated gene expression, possibly through transcriptional regulation of the enzymes mediating PGE2 generation (49). Interestingly, although the promoter elements of GM-CSF are believed to exhibit NFAT binding, the induction of this cytokine in response to both PAR2 and TG/PDBu was unaffected by CsA, suggesting that not all cytokines induced by CRAC channels are regulated through the calcineurin–NFAT pathway (Fig. 6B).
NFAT is normally phosphorylated at rest and resides in the cytoplasm in the inactive state. Ca2+ elevations mobilize NFAT activation through dephosphorylation of conserved serine/threonine residues by the Ca2+-activated protein phosphatase calcineurin, which causes NFAT to move into the nucleus and bind to target DNA sequences (43). Consistent with this property of NFAT, examination of a fluorescently tagged NFAT isoform (GFP-NFATc3) revealed that activation of CRAC channels by TG resulted in robust NFAT import into the nucleus (Fig. 7A, 7B). The import was blocked in a nominally Ca2+-free solution and by BTP2, indicating that Ca2+ entry through CRAC channels was required to trigger nuclear translocation of NFAT (Fig. 7A, 7B). Moreover, examination of NFAT-dependent gene transcription using a luciferase reporter assay indicated that CRAC channel activation stimulated NFAT-dependent luciferase expression in NHBE cells (Fig. 7C). Luciferase activity was abrogated in the absence of external Ca2+ and by the CRAC channel antagonists BTP2 and RO2959 (Fig. 7C) without affecting cell viability (Supplemental Fig. 4). Importantly, siRNA-mediated suppression of STIM1 and Orai1 expression significantly inhibited NFAT-dependent gene transcription (Fig. 7D). The calcineurin inhibitors CsA and FK506 also strongly inhibited induction of NFAT-luciferase activity consistent with the well-known role of calcineurin in controlling NFAT activation (Fig. 7C). Thus, these results highlight a broader role for NFAT in the generation of inflammatory mediators from the airway epithelium (Fig. 8).
CRAC channels activate NFAT-dependent gene expression in AECs. (A) Wide-field images showing the localization of NFATc3-GFP protein under resting conditions and after CRAC channel activation with TG (1 μmol, 20 min) in the presence or absence of Ca2+ in the external bath and in the presence of BTP2 (500 nmol). In each case, the pre- and post-TG images show the same cells. (B) Summary of the percentage of cells showing NFAT translocation in BEAS-2B cells following CRAC channel activation (mean ± SEM of n = 12–15 cells from three separate coverslips for each condition). (C) Summary of NFAT-luciferase activity in the presence and absence of CRAC channel (BTP2 and RO2959: 500 nmol) and calcineurin inhibitors (CsA 500 nmol, FK506 1 μmol). Data are mean ± SEM of four replicates; three independent experiments in NHBE cells. NFAT-luciferase activity was measured using a pGL3 NFAT luciferase reporter construct (29). Cells were stimulated with TG and PdBu to activate CRAC channels and PKC, respectively. (D) siRNA knockdown of STIM1 and Orai1 significantly inhibits NFAT luciferase activity in BEAS-2B cells (mean ± SEM of three replicates; three independent experiments). NFAT luciferase activity normalized to Renilla luciferase to correct for variations in cell density as described in the 2Materials and Methods. *p < 0.05, **p < 0.01. siCon, small interfering control.
CRAC channels activate NFAT-dependent gene expression in AECs. (A) Wide-field images showing the localization of NFATc3-GFP protein under resting conditions and after CRAC channel activation with TG (1 μmol, 20 min) in the presence or absence of Ca2+ in the external bath and in the presence of BTP2 (500 nmol). In each case, the pre- and post-TG images show the same cells. (B) Summary of the percentage of cells showing NFAT translocation in BEAS-2B cells following CRAC channel activation (mean ± SEM of n = 12–15 cells from three separate coverslips for each condition). (C) Summary of NFAT-luciferase activity in the presence and absence of CRAC channel (BTP2 and RO2959: 500 nmol) and calcineurin inhibitors (CsA 500 nmol, FK506 1 μmol). Data are mean ± SEM of four replicates; three independent experiments in NHBE cells. NFAT-luciferase activity was measured using a pGL3 NFAT luciferase reporter construct (29). Cells were stimulated with TG and PdBu to activate CRAC channels and PKC, respectively. (D) siRNA knockdown of STIM1 and Orai1 significantly inhibits NFAT luciferase activity in BEAS-2B cells (mean ± SEM of three replicates; three independent experiments). NFAT luciferase activity normalized to Renilla luciferase to correct for variations in cell density as described in the 2Materials and Methods. *p < 0.05, **p < 0.01. siCon, small interfering control.
Model for the Ca2+-dependent signaling pathway activated by PAR2. Activation of PAR2 receptors by serine proteases (trypsin) and peptide agonists (SLIGRL and SLIGKV) evokes PLC stimulation, leading to IP3 production and ER store release through the opening of IP3 receptors. Release of Ca2+ from intracellular stores leads to activation of STIM1, which opens CRAC channels comprised of Orai1 subunits. Ca2+ influx through CRAC channels activates calcineurin–NFAT signaling to transcriptionally induce IL-6 and IL-8 production, likely in conjunction with other pathways. BTP2 and RO2959, CRAC channel inhibitors; U73122, PLCβ inhibitor.
Model for the Ca2+-dependent signaling pathway activated by PAR2. Activation of PAR2 receptors by serine proteases (trypsin) and peptide agonists (SLIGRL and SLIGKV) evokes PLC stimulation, leading to IP3 production and ER store release through the opening of IP3 receptors. Release of Ca2+ from intracellular stores leads to activation of STIM1, which opens CRAC channels comprised of Orai1 subunits. Ca2+ influx through CRAC channels activates calcineurin–NFAT signaling to transcriptionally induce IL-6 and IL-8 production, likely in conjunction with other pathways. BTP2 and RO2959, CRAC channel inhibitors; U73122, PLCβ inhibitor.
Discussion
Ca2+ signals regulate many key physiological processes in AECs, including activation of Ca2+ activated Cl− conductances, ciliary beat frequency, and mucous and surfactant production (19, 50, 51). There is now growing evidence that Ca2+ signals also play an important role in the induction of inflammatory mediators from the airway epithelium in response to pathogens and allergens and may be involved in orchestrating inflammatory responses in several airway diseases including cystic fibrosis, asthma, and acute lung injury (7, 20, 52). Despite these observations, our understanding of how Ca2+ signals are generated in human AECs and how they are linked to effector functions remains incomplete. Knowledge of molecules that comprise the functional architecture of the Ca2+ signaling network and the mechanisms by which these proteins shape Ca2+ signals is important to gain a full understanding of how Ca2+ signals modulate downstream inflammatory responses in human AECs. In this study, we show that CRAC channels encoded by Orai1 and STIM1 mediate a critical role in generating Ca2+ signals in response to PAR2 activation and regulate the production of several key inflammatory mediators in the airway epithelium, in part by activating NFAT-dependent gene expression. These results highlight CRAC channels as a key checkpoint for transducing responses from proteases and for the generation of inflammatory mediators from the airway epithelium and identify CRAC channels as a potential target for therapeutic treatment of allergic and inflammatory airway diseases.
CRAC channels are a major route of Ca2+ entry in AECs
We show that primary human AECs exhibit SOCE and store-operated currents that share the pharmacological and biophysical hallmarks of CRAC channels (37). These features include block by low concentrations of (1 to 2 μmol) La3+, modulation by 2-APB, inhibition by the CRAC channel inhibitors BTP2 and RO2959, an inwardly rectifying I-V, fast inactivation, and depotentiation of the Na+ CRAC current (Fig. 3). Analysis of the underlying molecular machinery further reveals that the CRAC channel proteins STIM1 and Orai1 make essential contributions to SOCE in primary human AECs (Fig. 3E–G). Both bronchial and alveolar epithelial cells exhibited STIM1-Orai1–dependent SOCE, suggesting that this Ca2+ influx pathway is broadly conserved throughout the lower airway epithelium (Fig. 3, Supplemental Fig. 1). These findings confirm and broaden observations from past studies that have described STIM1- and Orai1-mediated SOCE in various AEC lines and corroborate the emerging viewpoint that CRAC channels are a well-conserved mechanism for Ca2+ influx in the lower airways (21, 34, 53). Our data, however, do not rule out a role for other CRAC channel isoforms. Indeed, SOCE was not fully abolished by STIM1 and Orai1 knockdown. Moreover, STIM2 and Orai3 knockdown resulted in a modest inhibition of SOCE (Fig. 3G), raising the possibility that these proteins may be expressed and make additional contributions to SOCE in AECs. More studies using isoform-specific Abs and knockout mice would be needed to test this issue. These issues notwithstanding, our findings unequivocally show that STIM1 and Orai1 are essential for conferring SOCE in human AECs.
CRAC channels mediate the downstream effects of PAR2 signaling
A screen for ligands that activate SOCE revealed that only a very specific set of receptors stimulate CRAC channel activation in AECs (Table I). A prominent hit in this screen was PAR2, an important sensor of allergens and endogenous proteases in the airway epithelium (5, 6, 9). Although many elements of the PAR2 signaling pathway have been studied in AECs, it is unclear whether PAR2 activation evokes CRAC channel-mediated Ca2+ signals. In this study, we find that Ca2+ influx through CRAC channels encoded by STIM1 and Orai1 is needed to maintain the sustained Ca2+ signal that arises from PAR2 activation (Figs. 1, 2). Furthermore, we find that PAR2 activation induces production of PGE2, IL-6, IL-8, and GM-CSF in AECs in a CRAC channel–dependent manner (Fig. 5). These findings highlight a novel role for CRAC channels as a key control element in the PAR2 signaling response. It is worth noting that although PGE2 production is seen following brief activation of PAR2 receptors (2 h), IL-6, IL-8, and GM-CSF production required prolonged stimulation of PAR2 (8–24 h). Because PGE2 evokes bronchodilation through effects on the airway smooth muscle, these results are consistent with the emerging idea that acute responses to PAR2 tend to be protective largely, whereas more sustained responses to PAR2 activation result in induction of cytokines that are proinflammatory, thereby contributing to the pathogenesis of diseases like asthma (12, 54).
CRAC channels induce the production of inflammatory mediators from AECs
In the repertoire of inflammatory mediators secreted by AECs, several, including PGE2, TSLP, IL-6, IL-8, and GM-CSF, are thought to be induced by cytosolic Ca2+ elevations (7, 55, 56). However, the source of Ca2+ signals driving the induction of these proinflammatory mediators remains unclear. We find that CRAC channel activation induces the production of a host of inflammatory mediators from primary bronchial cells including the cytokines IL-6, TSLP, TNF-α, and IL-1β, the arachidonic acid metabolite PGE2, the chemokines IL-8 and RANTES, and the growth factor GM-CSF (Fig. 4). Further examination shows that TNF-α, IL-6, and GM-CSF are specifically secreted only by costimulation of CRAC channels with PKC activation, whereas TSLP, PGE2, and IL-1β are maximally induced by CRAC channel stimulation alone (Fig. 4). These results indicate that cross-talk between signaling pathways can change the specific repertoire of inflammatory mediators released from AECs.
Among the inflammatory mediators that regulate airway inflammation, TSLP and PGE2 are produced primarily by the airway epithelium. TSLP plays a vital role in effecting a Th2-type airway inflammatory response that is characteristic of allergic diseases like asthma and is known to be induced by various bacterial, viral, and fungal products (1). In contrast, the arachidonic acid metabolite PGE2 has immunoprotective effects in the airway that include attenuation of bronchoconstriction in exercise or allergen-induced asthma and inhibition of lymphocytic proliferation (57, 58). In this study, we find that CRAC channel activation induces both TSLP and PGE2 in AECs but at very different time scales: PGE2 is induced in <2 h, whereas TSLP induction requires at least 48 h of AEC stimulation. These temporal differences indicate that CRAC channels regulate multiple stages of the inflammatory response in AECs, and the net consequences for airway function will depend on balance between the ensuing pro- and anti-inflammatory effector responses. Curiously, whereas CRAC channel activation produced PGE2, we saw no induction of the leukotrienes (LTB4, LTC4-E4), which are proinflammatory products of the arachidonic acid pathway (Figs. 4, 5, and data not shown). This is in contrast to findings in mast cells, where CRAC channel activation is a potent and specific trigger for the induction of LTC4 production (46). CRAC channels may thus activate different arachidonic acid metabolites and produce distinct biological responses depending on the cell type.
CRAC channels activate NFAT-dependent gene expression in AECs
Induction of IL-6, TNF-α, PGE2, and IL-8 in response to CRAC channel activation was suppressed by the calcineurin antagonist, CsA (Fig. 6), indicating that these cytokines are transcriptionally regulated by the calcineurin–NFAT pathway. This is consistent with observations that the promoter regions of a number of cytokine genes including IL-6, IL-8, and TNF-α harbor NFAT binding sites (43) and reveals an important role for NFAT as a CRAC channel–dependent regulator of cytokine production in AECs. In line with this interpretation, PAR2-mediated induction of IL-6 and IL-8 was also significantly inhibited by blocking the NFAT signaling pathway. Interestingly, whereas CsA inhibited IL-8 production in response to both PAR2 agonists and TG/PDBu to the same extent, the induction of IL-6 by PAR2 activation was only partially inhibited by CsA. We speculate that this difference may reflect the possibility that additional NFAT-independent mechanisms activated by PAR2 signaling come into play to drive IL-6 induction. Moreover, not all inflammatory mediators activated by PAR2 signaling require NFAT gene expression (e.g., PGE2 and GM-CSF; Fig. 6A), suggesting that the induction of these mediators occurs through alternate Ca2+-regulated pathways. One possibility is that Ca2+-mediated stimulation of calcium-dependent phospholipase A2 may drive the induction of PGE2 (59), something that remains to be tested. Taken together, our results highlight a hitherto unappreciated role for NFAT as an important mediator of downstream PAR2 signaling in the airway epithelium.
Based on these results, we propose a model for PAR2 receptor signaling in the airway epithelium in which activation of these receptors, either by peptides or endogenous proteases, causes Gq-PLCβ-IP3–mediated store release, leading to the generation of long-lasting Ca2+ signals through the activation of store-operated CRAC channels. The ensuing Ca2+ influx triggers the activation of transcription factor NFAT, which then translocates to the nucleus and drives expression of IL-6 and IL-8 genes in conjunction with other signaling mechanisms (Fig. 8). CRAC channel activation thus constitutes an important regulatory checkpoint for the induction of inflammatory mediators from the airway epithelium.
Acknowledgements
The authors thank members of the laboratory for advice and Anna Toth for helpful comments on the manuscript.
Footnotes
This work was supported by the National Institutes of Health (Grant NS057499) and a grant from the Skin Disease Research Center of Northwestern University.
The online version of this article contains supplemental material.
Abbreviations used in this article:
- AEC
airway epithelial cell
- 2-APB
2-aminoethoxydiphenylborane
- BTP2
3,5-bis(trifluoromethyl)pyrazole
- CRAC
Ca2+ release-activated Ca2+
- CsA
cyclosporin A
- ER
endoplasmic reticulum
- GPCR
G-protein–coupled receptor
- IP3
inositol 1,4,5-triphosphate
- I-V
current voltage
- NFAT
nuclear factor of activated T cells
- NHBE
normal human bronchial epithelial
- PAR2
protease-activated receptor 2
- PDBu
phorbol 12,13-dibutyrate
- PKC
protein kinase C
- PLC
phospholipase C
- RNAi
RNA interference
- siRNA
small interfering RNA
- SOCE
store-operated Ca2+ entry
- STIM
stromal interaction molecule
- TG
thapsigargin
- TSLP
thymic stromal lymphopoietin.
References
Disclosures
The authors have no financial conflicts of interest.