We examined the stimulus-secretion pathways whereby proteinase-activated receptor 2 (PAR-2) stimulates Cl− secretion in intestinal epithelial cells. SCBN and T84 epithelial monolayers grown on Snapwell supports and mounted in modified Ussing chambers were activated by the PAR-2-activating peptides SLIGRL-NH2 and 2-furoyl-LIGRLO-NH2. Short-circuit current (Isc) was used as a measure of net electrogenic ion transport. Basolateral, but not apical, application of SLIGRL-NH2 or 2-furoyl-LIGRLO-NH2 caused a concentration-dependent change in Isc that was significantly reduced in Cl−-free buffer and by the intracellular Ca2+ blockers thapsigargin and BAPTA-AM, but not by the Ca2+ channel blocker verapamil. Inhibitors of PKA (H-89) and CFTR (glibenclamide) also significantly reduced PAR-2-stimulated Cl− transport. PAR-2 activation was associated with increases in cAMP and intracellular Ca2+. Immunoblot analysis revealed increases in phosphorylation of epidermal growth factor (EGF) receptor (EGFR) tyrosine kinase, Src, Pyk2, cRaf, and ERK1/2 in response to PAR-2 activation. Pretreatment with inhibitors of cyclooxygenases (indomethacin), tyrosine kinases (genistein), EGFR (PD-153035), MEK (PD-98059 or U-0126), and Src (PP1) inhibited SLIGRL-NH2-induced increases in Isc. Inhibition of Src, but not matrix metalloproteinases, reduced EGFR phosphorylation. Reduced EGFR phosphorylation paralleled the reduction in PAR-2-stimulated Isc. We conclude that activation of basolateral, but not apical, PAR-2 induces epithelial Cl− secretion via cAMP- and Ca2+-dependent mechanisms. The secretory effect involves EGFR transactivation by Src, leading to subsequent ERK1/2 activation and increased cyclooxygenase activity.
- ion transport
- signal transduction
- protein kinase
- mitogen-activated protein kinase
- extracellular signal-regulated kinase 1/2
the intestinal epithelial barrier is a critical component of the defense function of the gut, preventing the translocation of bacteria and other pathogens from the intestinal lumen to the mucosa. Epithelial barrier function is multifactorial, with one component being the secretion of water by a subpopulation of crypt enterocytes (20). Loss of secretory function leads to an increase in bacterial translocation (4). Loss of barrier function has important implications in Crohn's disease, which is thought to be precipitated by a loss of immune tolerance to commensal bacteria crossing a defective epithelial barrier (19, 45). Since Cl− secretion is the dominant force behind luminal fluid secretion (5, 25), determining the mechanisms underlying epithelial Cl− transport is important for understanding epithelial barrier function in health and in inflammatory diseases of the gut.
Serine proteinases exert many of their physiological and pathophysiological effects by acting on G protein-coupled proteinase-activated receptors (PARs) (13, 27, 49, 53). The proteinases cleave a specific amino acid sequence on the extracellular NH2 terminus of the receptor, revealing a new NH2 terminus that acts as a “tethered” ligand, interacting with other extracellular domains of the receptor to initiate signaling. Four different receptor subtypes have been identified: PAR-1, PAR-2, PAR-3, and PAR-4 (14, 27, 38). PAR-2 can be activated by a number of serine proteinases, including trypsin, mast cell tryptase, kallikreins, and the tissue factor VIIa-Xa complex (48, 49). Many of these proteinases may be present in the setting of intestinal damage or immune activation and, therefore, may be available to activate PAR-2. Experimentally, PAR-2 can be activated by receptor-selective PAR-activating peptides such as SLIGRL-NH2 and 2-furoyl-LIGRLO-NH2 (31, 44), which mimic the amino acid sequence of the tethered ligand and activate the receptor without requiring NH2-terminal cleavage. PAR-2 is expressed on the apical and basolateral membranes of enterocytes within the gastrointestinal (GI) tract (16) and has a comparable distribution within the stomach, small intestine, and colon of humans and mice (7, 47).
Given the abundance of pancreatic trypsin in the intestinal lumen and of mast cell tryptase in the mucosa, epithelial expression of PAR-2 could have important implications for the secretory component of epithelial barrier function. Indeed, PAR-2 activation has been shown to stimulate Cl− secretion in intestinal epithelial cells (6, 8, 9, 15, 55). Although pharmacological studies have suggested that PAR-2-induced ion transport is, in part, dependent on cyclooxygenase (COX) activity and intracellular Ca2+ (24, 34, 35, 50), the ion channels responsible for PAR-2-induced electrolyte transport are not known, nor are the signaling pathways that couple PAR-2 activation with Cl− secretion. We explored this issue using the SCBN and T84 intestinal epithelial cell lines and demonstrated that activation of basolateral, but not apical, PAR-2 results in cAMP- and Ca2+-dependent Cl− secretion, which requires Src-dependent epidermal growth factor (EGF) receptor (EGFR) transactivation, MAP kinase activation, and COX activity.
MATERIALS AND METHODS
SCBN cells, a nontransformed intestinal epithelial cell line that we previously used for studies of ion transport, were cultured according to previously described methods (8, 9). Cells between passages 22 and 30 were grown to confluence (∼5 days) in 75-cm2 flasks or on Snapwell semipermeable supports (Costar, Nepean, ON, Canada). Cells in flasks were passaged using 1.5× trypsin and every 2 days were fed DMEM supplemented with 5% FCS, l-glutamine, streptomycin, plasmocin, and tylosin. Some experiments were repeated in T84 cells, a human colon adenocarcinoma cell line, cultured as previously described (18). Cells between passages 60 and 70 were grown to confluence (∼8 days) on Snapwell semipermeable supports and every 2 days were fed DMEM-Ham's F-12 medium supplemented with 10% FCS, l-glutamine, tylosin, streptomycin, and plasmocin. Cells grown in Snapwell supports were fed every 2nd day, and confluence was determined by the increase in transepithelial electrical resistance measured using an electrovoltohmmeter (World Precision Instruments, Sarasota, FL). Only monolayers with a resistance >500 Ω·cm2 were used.
SCBN and T84 cells were grown to confluence in 75-cm2 flasks, and PAR-2 mRNA expression was assessed by RT-PCR, with α-actin mRNA expression used as an internal control. PAR-2 primers consisted of 5′-TTCTGCTGGCGGCCGCT-3′ (sense) and 5′-CCTCAGGCAAGACATCATGG-3′ (antisense), and actin primers consisted of 5′-CGTGGGCCGCCCTAGGCACCA-3′ (sense) and 5′-TTGGCCTTAGGGTTCAGGGGG-3′ (antisense); both primers were synthesized at the University of Calgary Core DNA Services. Negative controls consisted of no cDNA and no primer.
To determine the cellular localization of PAR-2, immunofluorescence confocal microscopy was conducted as previously described (2). SCBN and T84 cells were grown to confluence on semipermeable membrane supports and fixed with 100% ice-cold methanol for 20 min. Cells were incubated with 10% BSA to prevent nonspecific binding. To delineate apical and basolateral PAR-2 immunoreactivity, cells were also stained for the tight junction protein occludin (1:250 dilution, monoclonal mouse anti-occludin). For PAR-2 staining, the A5 antibody (polyclonal rabbit anti-PAR-2) was used as previously described (34), with subsequent exposure to an FITC-conjugated anti-mouse (occludin) or a Cy3-conjugated anti-rabbit (PAR-2) secondary antibody.
SCBN cells were grown to confluence on Snapwell supports, mounted in modified Ussing chambers, and bathed with Krebs buffer (mM: 115 NaCl, 2.0 KH2PO4, 2.4 MgCl2, 1.3 CaCl2, 25.0 NaHCO3, and 8.0 KCl) containing 10 mM mannitol on the apical side and with Krebs buffer containing 10 mM glucose on the basolateral side. Buffers were gassed with 5% CO2-95% O2 and maintained at 37°C and pH 7.4. A voltage-clamp apparatus (VCC MC8, Physiologic Instruments, San Diego, CA) was used to clamp the transepithelial potential difference to 0 V. Changes in net electrogenic ion transport were determined on the basis of changes in short-circuit current (Isc). Isc was recorded with a digital data acquisition system (MP100, BioPac, Goleta, CA) and analyzed using AcqKnowledge software (version 3.5.7, BioPac). Concentration-response curves were determined for noncumulative basolateral application of the PAR-2-activating peptides SLIGRL-NH2 and 2-furoyl-LIGRLO-NH2. On return of the PAR-2-evoked Isc response to baseline, 10 μM forskolin (FSK) was added to the basolateral side to confirm cell responsiveness. Similarly, T84 cells grown to confluence on Snapwell supports were mounted in Ussing chambers. A concentration-response curve to SLIGRL-NH2 was constructed, and the functional polarity of the receptor was determined. Subsequent experiments were conducted using SCBN cells unless otherwise noted.
To determine whether PAR-2-induced changes in Isc were due to Cl− transport, experiments were repeated in Cl−-free Krebs buffer (117 mM sodium isethionate, 2 mM KH2PO4, 2.4 mM hemimagnesium gluconate, 650 μM calcium gluconate, 25 mM NaHCO3, and 8 mM potassium gluconate). After 10 min of incubation in Cl−-free buffer, monolayers were exposed basolaterally to 50 μM SLIGRL-NH2 or 10 μM FSK. At the end of some experiments, Cl−-free Krebs buffer was replaced with regular Krebs buffer, and the cell monolayers were exposed to FSK to verify cell viability. Responses were compared with those of cells exposed to regular, Cl−-containing Krebs buffer.
To determine the Ca2+ and PKA dependence of PAR-2-induced changes in Isc, SCBN monolayers mounted in Ussing chambers were pretreated with various Ca2+ inhibitors before basolateral exposure to 50 μM SLIGRL-NH2. Thapsigargin (250 nM), which blocks Ca2+-ATPase on the endoplasmic reticulum, was used to deplete Ca2+ stores. Verapamil (30 μM), which blocks L-type Ca2+ channels, was used to prevent extracellular Ca2+ uptake. To further substantiate our findings, cells were treated for 45 min with the Ca2+ chelator BAPTA-AM (30 μM) before exposure to SLIGRL-NH2. The vehicle DMSO was used as a control. H-89 (5 μM) was used as an inhibitor of PKA, and the combination of H-89 and thapsigargin was used to determine the dual role of Ca2+ and PKA in the Cl− secretory response to PAR-2 activation.
The potential role of the PKA-dependent CFTR channel was determined using a selective CFTR inhibitor (glibenclamide added apically, 100 μM). The role of bicarbonate secretion was determined using 300 μM DIDS. Monolayers were incubated with inhibitors for 10 min before exposure to 50 μM SLIGRL-NH2. The vehicle DMSO was used as a control.
To assess the role of various kinases and signaling molecules in the PAR-2-induced change in Isc, SCBN monolayers were pretreated with specific inhibitors of tyrosine kinases (100 μM genistein), EGFR tyrosine kinase (1 μM PD-153035), the ERK1/2 activator MEK (50 μM PD-98059 and 25 μM U-0126), the non-receptor tyrosine kinase Src (1 μM PP1), matrix metalloproteinases (MMPs; 20 μM GM-6001), and PKC [1 μM GF-109203X (GFX)]. After 10 min of incubation with each inhibitor, monolayers were exposed basolaterally to 50 μM SLIGRL-NH2, and the change in Isc was measured.
The production of cAMP following PAR-2 stimulation by 50 μM SLIGRL-NH2 was determined as previously described (9). Briefly, SCBN cells were grown to confluence in 10-cm dishes, serum starved for 2 h, and then resuspended in serum-free medium. Cell suspensions were pretreated with the phosphodiesterase inhibitor B-8279 for 20 min before exposure to the vehicle, 50 μM SLIGRL-NH2, or 10 μM FSK. After 5 min, the cells were immersed in liquid nitrogen and then in a 40°C water bath. This freeze-thaw process was repeated three times to lyse the cells and whole cell lysates. The concentration of cAMP was determined by ELISA (Cayman Chemical, Ann Arbor, MI) according to the manufacturer's instructions.
To determine increases in Ca2+ following PAR-2 activation by 50 μM SLIGRL-NH2, a Ca2+ assay was conducted as previously described (2). Briefly, cells were lifted using an enzyme-free cell dissociation buffer (HyQtase), pelleted, and resuspended in FBS-free DMEM. The cells were then loaded with the intracellular Ca2+ indicator fluo 3-AM (Molecular Probes, Eugene, OR) at a final concentration of 22 μM (25 μg/ml). Cells were incubated with fluo 3 for 25 min at room temperature in the presence of 0.25 mM sulfinpyrazone, washed with PBS, and resuspended in a Ca2+ buffer (mM: 150 NaCl, 3 KCl, 1.5 CaCl2, 20 HEPES, 10 glucose, and 0.25 sulfinpyrazone). A luminescence spectrometer (Series 2, Aminco Bowman, Rochester, NY) was used to measure fluorescence changes (at 530 nm) elicited by 50 μM SLIGRL-NH2. The Ca2+ ionophore A-23187 (2 μM) was used as a positive control, and responses to SLIGRL-NH2 were expressed as a percentage of the response to ionophore.
To confirm the physiological results obtained with the various pharmacological inhibitors, immunoblots were conducted to determine activation of EGFR, Src, Pyk2, cRaf, and ERK1/2 after exposure of SCBN cells to SLIGRL-NH2. SCBN cells were grown to confluence in 75-cm2 flasks, serum starved for 2 h, and then resuspended in serum-free medium. Cell suspensions were treated with the inhibitors described above for 20 min. In addition, some cell suspensions were treated with the pan-PKC inhibitor GFX (1 μM). Cells were then exposed to SLIGRL-NH2 (50 μM). After 5 min, whole cell lysates were collected, and proteins were resolved by a 10% SDS-PAGE gradient gel and transferred to a polyvinylidene difluoride membrane. After overnight exposure to the primary antibody at 4°C, blots were treated with horseradish peroxidase (HRP)-conjugated secondary antibody at room temperature for 1 h. Bands were then visualized with enhanced chemiluminescence reagent (Amersham Pharmacia, Baie d'Urfé, QC, Canada) and quantified by densitometry.
Phosphorylation of Src was determined using a rabbit polyclonal antibody specific to phosphorylated (Tyr416) Src (1:1,000 dilution) and an HRP-conjugated sheep anti-rabbit secondary antibody (1:10,000 dilution). Similarly, for Pyk2, immunoblotting was conducted using a rabbit polyclonal antibody specific to phosphorylated (Tyr402) Pyk2 (1:1,000 dilution) and an HRP-conjugated secondary antibody (1:10,000 dilution). Activation of EGFR was determined with a mouse monoclonal antibody specific to phosphorylated (Tyr1173) EGFR (2 μg/ml) and an HRP-conjugated rabbit anti-mouse secondary antibody (1:5,000 dilution).
To determine the possible interaction of Src and Pyk2 with the EGFR, coimmunoprecipitation experiments were conducted in which immunoprecipitation of the EGFR was performed using an antibody specific to EGFR (mouse monoclonal antibody, 4 μg/ml). The total protein level of Src (rabbit monoclonal antibody, 1:1,000 dilution) and Pyk2 (goat polyclonal antibody, 1:1,000 dilution) was then determined via an immunoblot, thereby allowing us to determine whether Src or Pyk2 formed a complex with the EGFR.
To determine cRaf activation, a rabbit monoclonal antibody specific to phosphorylated (Ser338) cRaf (1:500 dilution) and an HRP-conjugated goat anti-rabbit secondary antibody (1:5,000 dilution) were used. For detection of activated ERK1/2, a mouse monoclonal antibody specific to phosphorylated (Thr202/Tyr204) ERK1/2 (1:2,000 dilution) and an HRP-conjugated rabbit anti-mouse secondary antibody (1:10,000 dilution) were used. In each case, membranes were reprobed for total amounts of each protein with antibodies that recognized the phosphorylated and unphosphorylated forms of the proteins.
DMEM, antibiotic solutions, thapsigargin, verapamil, BAPTA-AM, H-89, PP1, PD-153035, PD-98059, glibenclamide, and U-0126 were obtained from Sigma-Aldrich (Oakville, ON, Canada); GM-6001, DIDS, and GFX from Calbiochem (Mississauga, ON, Canada); cell dissociation buffer (HyQtase) and FBS from HyClone (Logan, UT); and antibodies for EGFR, ERK, Src, cRaf, and Pyk2 from Cell Signaling (Beverly, MA). The PAR-2-activating peptides SLIGRL-NH2 and 2-furoyl-LIGRLO-NH2 were synthesized in house by Dr. Denis McMaster in the University of Calgary Core Peptide Synthesis Service.
Values are means ± SE. Statistical analyses were conducted using GraphPad Instat version 3.0 (GraphPad Software, San Diego, CA). Comparisons of two groups were made using Student's t-test for unpaired data. Comparisons of more than two groups were made using a two-way analysis of variance with post hoc Tukey's test.
PAR-2 expression and effects of activation on ion transport.
SCBN and T84 cell monolayers constitutively expressed PAR-2 mRNA (Fig. 1A) and protein (Fig. 1B). PAR-2 was expressed primarily in the plasma membrane, with some punctate immunoreactivity in the cytoplasm (Fig. 1B). Furthermore, on the basis of z stacks constructed from optical sections through the monolayers, it was determined that SCBN and T84 cells express PAR-2 on the apical and basolateral membranes, although the majority of immunoreactivity appeared to be basolateral (Fig. 1D). Monolayers incubated without secondary antibody showed no immunofluorescence (Fig. 1C).
Activation of PAR-2 by SLIGRL-NH2 or 2-furoyl-LIGRLO-NH2 resulted in a concentration-dependent increase in Isc (Fig. 2A). The increase in Isc began within 1 min of addition of activating peptide, peaked at ∼3 min, and returned to baseline by 5 min (Fig. 2B). 2-Furoyl-LIGRLO-NH2 was more potent than SLIGRL-NH2, consistent with the structure-activity data obtained for these PAR-2-selective receptor agonists in previous studies (14, 15). Similarly, T84 cells exhibited a concentration-dependent response to basolateral SLIGRL-NH2, although the responses to each SLIGRL-NH2 concentration were smaller than those observed in SCBN cells (Fig. 2). In SCBN cells, the signal from PAR-2 appeared to originate principally from the basolateral aspect of the cell monolayer, since activation of PAR-2 by 50 μM SLIGRL-NH2 on the basolateral membrane resulted in a change in Isc that was significantly greater than that mediated by apical activation of PAR-2 (4.8 ± 0.6 vs. 26.4 ± 1.7 μA/cm2, P < 0.001). There was no response to apically applied 50 μM SLIGRL-NH2 in T84 cells, indicating a functional basolateral distribution of the receptor in this cell line. Treatment of the cells with HEPES, the vehicle for SLIGRL-NH2 and 2-furoyl-LIGRLO-NH2, or the reverse peptide LRGILS-NH2, had no effect on Isc (data not shown).
Activation of PAR-2 by 50 μM SLIGRL-NH2 in cell monolayers incubated in Cl−-free buffer resulted in an Isc response that was substantially smaller than that of cells incubated in regular Krebs buffer (Fig. 3). Similar results were obtained in cell monolayers exposed to 10 μM FSK (Fig. 3). When Cl−-free buffer was replaced with regular buffer, increases in Isc mediated by 10 μM FSK were similar to those in controls, indicating the viability of the cell preparation (Fig. 3).
PAR-2-induced Cl− secretion is Ca2+, PKA, PKC, and COX dependent.
Pretreatment of the cells with 250 nM thapsigargin to deplete intracellular Ca2+ stores or 30 μM BAPTA-AM to chelate released Ca2+, but not the Ca2+ channel blocker verapamil (30 μM), significantly reduced the Isc response to PAR-2 activation by SLIGRL-NH2 (Fig. 4). Thapsigargin resulted in a sustained increase in Isc (9.6 ± 1.5 μA/cm2). This increase in thapsigargin-induced Isc was not, however, affected by pretreatment of the cells with the PKA inhibitor H-89 (5 μM; data not shown), nor did thapsigargin affect FSK-induced Isc responses (data not shown). Pretreatment of the cells with 5 μM H-89 significantly reduced the Isc response to PAR-2 activation by 50 μM SLIGRL-NH2 (Fig. 4). Furthermore, the combination of H-89 (5 μM) and thapsigargin (250 nM) significantly reduced the Isc response to PAR-2 activation, suggesting parallel pathways for PKA and Ca2+ signaling in PAR-2-mediated Cl− secretion (Fig. 4). In addition, PAR-2 stimulation by 50 μM SLIGRL-NH2 significantly increased intracellular Ca2+ and cAMP in SCBN cells (Fig. 5A and B). Finally, a role for PKC was demonstrated, since the pan-PKC inhibitor GFX (1 μM) significantly reduced PAR-2-mediated ion transport (Fig. 4).
Pretreatment of SCBN cells with an inhibitor of CFTR (glibenclamide, 100 μM) significantly reduced PAR-2-mediated Cl− secretion (Fig. 6). Furthermore, apical pretreatment with the Ca2+-dependent Cl− channel blocker DIDS (300 μM) significantly reduced PAR-2-mediated changes in Isc (Fig. 6). These results suggest an important role for CFTR and the Ca2+-dependent Cl− transport in PAR2-induced changes in Isc.
Exposure of the cells to the nonselective COX inhibitor indomethacin (10 μM) also significantly reduced the Cl− secretory response to PAR-2 activation by 50 μM SLIGRL-NH2 (18.5 ± 1.3 and 9.3 ± 2.2 μA/cm2 for vehicle control and indomethacin, respectively, P < 0.005).
PAR-2-induced stimulus-secretion coupling.
Treatment of SCBN cells with the broad-spectrum tyrosine kinase inhibitor genistein (100 μM) significantly reduced the Isc response to SLIGRL-NH2, suggesting a role for tyrosine kinase signaling in the response (Fig. 7). Genistein resulted in a small increase in Isc (3.2 ± 0.2 μA/cm2). Inhibition of EGFR tyrosine kinase activity with PD-153035 (1 μM) significantly reduced the Isc response, suggesting that PAR-2-induced Cl− secretion is due, at least in part, to transactivation of the kinase activity of the EGFR. Previous studies have implicated Src kinase and MMPs in the EGFR transactivation event (33, 43). Thus we examined the actions of PP1 (1 μM), a selective inhibitor of Src kinase, and GM-6001 (20 μM), an inhibitor of MMPs, by adding these inhibitors to Ussing chambers before PAR-2 activation. Pretreatment with PP1, but not GM-6001, significantly reduced the Isc response (Fig. 7).
To further explore whether Src kinase is involved in the EGFR transactivation mechanism, immunoblots were performed to monitor the level of phosphorylated EGFR in the presence or absence of specific inhibitors during stimulation with SLIGRL-NH2. After PAR-2 activation, there was an increase in EGFR phosphorylation compared with control (Fig. 8A). This increase was comparable to the phosphorylation observed after treatment with EGF. EGFR phosphorylation after PAR-2 activation was diminished in the presence of the Src kinase inhibitor PP1, but not in the presence of the MMP inhibitor GM-6001 (Fig. 8, A and C). Immunoblot analysis also demonstrated that PAR-2 activation increased Src phosphorylation, which reflects the activation of this enzyme (Fig. 8B).
To assess a potential role for the proline-rich tyrosine kinase Pyk2 in the PAR-2-mediated signaling events, immunoblots for phosphorylated Pyk2 were conducted in the presence of specific inhibitors. PAR-2 activation by SLIGRL-NH2 significantly increased Pyk2 phosphorylation, and this event was inhibited by PP1 and BAPTA-AM, but not by PD-153035 or GFX (Fig. 9). These results suggest an important role for Ca2+- and Src-dependent Pyk2 signaling after PAR-2 activation.
To further elucidate the PAR-2-induced interaction between Src/Pyk2 and the EGFR, SCBN cells were exposed to SLIGRL-NH2 and EGFR was immunoprecipitated. Immunoblots for Src or Pyk2 were conducted, but neither Src nor Pyk2 was detected (data not shown).
MAP kinase activation.
To determine a role for MAP kinase signaling in PAR-2-mediated ion transport, SCBN cells grown on Snapwell supports were mounted in Ussing chambers and pretreated with the specific MEK inhibitors PD-98059 (50 μM) and U-0126 (25 μM) before PAR-2 activation with SLIGRL-NH2. Pretreatment with PD-98059 or U-0126 resulted in a significant reduction in the PAR-2-induced change in Isc (Fig. 7). To further substantiate these findings, immunoblots for phosphorylated ERK1/2 and the upstream kinase cRaf were conducted. Activation of PAR-2 by SLIGRL-NH2 significantly increased cRaf phosphorylation compared with control. This response was not inhibited by PP1 or PD-153035 (Fig. 10). However, the response to PP1 was not significantly different from the vehicle control. Interestingly, we found that inhibition of PKC with GFX significantly reduced cRaf phosphorylation. Since activation of EGFR can result in downstream ERK1/2 activation and the PAR-2-induced change in Isc was blocked by inhibitors of ERK1/2 activation, immunoblots for phosphorylated ERK1/2 were conducted in the presence of specific inhibitors for Src, EGFR, MEK, or PKC. Activation of PAR-2 by SLIGRL-NH2 significantly increased ERK1/2 phosphorylation, and pretreatment with inhibitors of Src, EGFR, MEK, or PKC significantly reduced this phosphorylation (Fig. 11). These findings suggest that, after EGFR transactivation, there is downstream signaling, which results in ERK1/2 phosphorylation and a subsequent Cl− secretory response.
The present study has demonstrated that basolateral PAR-2 activation stimulates Cl− secretion in intestinal epithelial cell monolayers via a Src-mediated mechanism dependent on intracellular Ca2+, PKA, PKC, EGFR transactivation, Raf, Pyk2, and MAP kinase, along with the participation of COX activity. These signaling pathways contribute to Cl− secretion via cAMP- and Ca2+-dependent pathways. PARs have a wide tissue distribution and are expressed by many cell types within the GI tract (12, 24, 34, 46), suggesting diverse functional roles for these receptors in the gut. PAR-2 is expressed on the apical and basolateral membranes of enterocytes within the GI tract (16), and it has been demonstrated that activation of PAR-2 can result in Cl− secretion in isolated intestinal tissue segments (15, 24, 39). Although PAR-2 activation can result in epithelial Cl− secretion, the signaling pathways regulating this process in intestinal epithelial cells are not clear. The present study provides insight into the roles of a number of signaling pathway elements in the secretion of Cl− stimulated by the activation of PAR-2.
We chose the SCBN cell line for most of our studies, inasmuch as these cells form electrically tight monolayers amenable to the study of ion transport under voltage-clamp conditions, as we previously demonstrated (8, 9). We showed that SCBN cells express PAR-2, as shown by RT-PCR and immunofluorescence confocal microscopy. The latter revealed that PAR-2 was localized to the apical and basolateral plasma membrane. In most cells, it appeared that the distribution of immunoreactivity was primarily basolateral. This was reflected in the ion transport response to PAR-2 stimulation, since the Isc response to apically applied PAR-2-activating peptide was substantially smaller than that elicited by basolateral application. This observation was not peculiar to SCBN cells; similar observations were made in T84 cells. PAR-2 expression on these cell lines may differ from that observed on native epithelia. Nevertheless, in in vitro preparations of pig ileum (24), rat ileum (55), and human colon (39), a change in Isc was stimulated more effectively by contraluminal than by luminal exposure to PAR-2 activators. These results suggest that, physiologically, a subepithelial PAR-2 activator, such as mast cell tryptase, is more likely to stimulate a secretory response than would a luminal activator, such as trypsin. Thus PAR-2 activation could contribute to the Cl− secretion that characterizes intestinal anaphylaxis (51).
Using the selective PAR-2-activating peptides SLIGRL-NH2 and 2-furoyl-LIGRLO-NH2, we constructed concentration-response curves for the change in Isc. From our studies and consistent with previous structure-activity relationship experiments using SLIGRL-NH2 and 2-furoyl-LIGRLO-NH2 (44), we found that 2-furoyl-LIGRLO-NH2 is a more potent activator of PAR-2-induced changes in Isc. Nevertheless, we used SLIGRL-NH2 for the bulk of our experiments, since its specificity has been clearly demonstrated (28) and it was available in greater quantities from our peptide synthesis facility. It has also been previously demonstrated that physiological activators of PAR-2, such as trypsin and mast cell tryptase, elicit downstream signaling pathways similar to those of activating peptides (26). Experiments using Cl−-free buffer showed that the PAR-2-induced changes in Isc are primarily the result of Cl− secretion. Partial inhibition with apical glibenclamide suggests that CFTR is, in part, responsible for the change in Isc. Our data also suggest that Ca2+-dependent Cl− transport could play a role in PAR-2-induced changes in Isc. Experiments using apically applied DIDS to block Ca2+ significantly reduced the secretory response to PAR-2 stimulation. Because DIDS was added apically, we interpreted the reduced Isc as being due to an effect on apical Ca2+-dependent Cl− channels. Nevertheless, we have not ruled out a nonselective effect of DIDS at other ion transporters in these experiments. Indeed, the fact that the Isc response was not completely abolished in Cl−-free buffer implies a residual flux of other ions, such as bicarbonate ion, which also contributes to the net apically directed anion flux in intestinal epithelium (56). DIDS can also block basolateral anion exchange; when we added it basolaterally, we observed a decreased response to PAR-2 activation (data not shown).
We have conducted experiments implicating Ca2+- and cAMP-dependent pathways in the Cl− secretory response to PAR-2 activation. In most cells, PAR-2 signals through Gqα to stimulate an increase in intracellular Ca2+ (49). Using fluo 3-loaded SCBN cells, we have demonstrated that PAR-2 activation in these cells increases intracellular Ca2+. Consistent with these observations, we showed that the Isc response to PAR-2 activation is dependent on the release of Ca2+ from intracellular stores, since the response was inhibited by pretreatment of cells with BAPTA-AM or thapsigargin, whereas the L-type Ca2+ channel blocker verapamil had no effect.
We also provided evidence for the involvement of the PKA-dependent Cl− channel CFTR. Previous studies in HEK-293 cells showed that PAR-2 can elicit responses that result in the generation of cAMP and, thus, in the activation of PKA (3). To determine whether PAR-2 can increase cAMP in SCBN cells, we performed an ELISA and found that stimulation of PAR-2 with SLIGRL-NH2 did in fact result in elevated cAMP levels. We found that inhibition of PKA with H-89 reduced PAR-2-mediated Cl− secretion. This observation is consistent with the ability of glibenclamide to inhibit the PAR-2-induced change in Isc, since CFTR channel opening is a cAMP-PKA-dependent event (23). The Ca2+- and PKA-dependent pathways appear to work in parallel, since the combination of thapsigargin and H-89 had an additive inhibitory effect on PAR-2-induced changes in Isc.
PAR-2-induced increases in intracellular Ca2+ and cAMP could activate multiple ion channels. First, these signaling pathways could stimulate cAMP-dependent (CFTR) and Ca2+-dependent Cl− channels directly. Second, these pathways could also stimulate the activation of cAMP- or Ca2+-dependent K+ channels on the basolateral membrane. K+ channel activation is necessary to maintain the activity of the Na+-K+-ATPase that drives epithelial ion transport (5). Indeed, previous studies demonstrated that PAR-2-induced Cl− secretion in Calu-3 airway epithelial cells was mediated in part by K+ efflux via Ca2+-dependent activation of the basolateral human IK1 channel (52).
Having provided evidence for the ion channels participating in PAR-2-induced changes in Isc in SCBN cells, we sought to determine the signaling pathways involved. We investigated three possible pathways: 1) signaling mediated by Ca2+ and PKC, 2) the role of EGFR transactivation, and 3) the role of COXs. Ca2+ is an important regulator of PKC, which can mediate cRaf phosphorylation and, hence, ERK1/2 MAP kinase activation (41). PKC is also capable of modulating CFTR activation and, thereby, can regulate Cl− secretion through this channel (30) and is an important regulator of Ca2+-dependent signaling events in SCBN cells (10). We showed that inhibition of PKC with the pan-PKC inhibitor GFX reduced PAR-2-induced ion transport. In an earlier study, it was demonstrated that, in the KNRK cell line, PAR-2 coupled to Gαq/11 and phospholipase Cβ, resulting in the formation of inositol trisphosphate, the mobilization of Ca2+, and the activation of PKC and ERK1/2 (17). Further studies are needed to determine exactly where PKC activation fits into PAR-2 stimulus-secretion coupling and the isoforms of this kinase that are involved; however, a complete examination of these factors is beyond the scope of the present study.
The transactivation of EGFR has been shown to play a role in secretory events linked to G protein-coupled receptor activation in intestinal epithelial cells. For example, EGFR transactivation occurs after muscarinic receptor activation by carbachol in the T84 colonic epithelial cell line (43). This effect was found to depend on the MMP-dependent release of transforming growth factor-α, which then resulted in EGFR activation of subsequent signaling events (43). We showed that activation of PAR-2, also a G protein-coupled receptor, resulted in EGFR phosphorylation. Furthermore, a selective inhibitor of EGFR tyrosine kinase activity blocked PAR-2-induced secretion. However, in contrast to studies that demonstrated a role for MMPs in EGFR transactivation (11, 22, 58), we showed that the transactivation event was independent of metalloproteinase activity, since it was not blocked by the pan-MMP inhibitor GM-6001.
To further investigate EGFR-mediated signaling following PAR2 activation, we examined the roles of Src and Pyk2 in the response. EGFR transactivation following muscarinic receptor activation in T84 cells also requires the formation of a Src-Pyk2 complex (33, 43). Consistent with these observations, we have demonstrated that, on activation of PAR-2 by SLIGRL-NH2, EGFR transactivation and Pyk2 phosphorylation occurred by a Src-dependent mechanism, since EGFR and Pyk2 phosphorylation events were blocked by the Src inhibitor PP1. Others have demonstrated that Src and Pyk2 form a complex that associates with EGFR, leading to its activation via phosphorylation events (33, 37). Furthermore, Pyk2 has also been suggested as a potential mediator of EGFR transactivation acting downstream of Src and Ca2+ signaling events (21). Since we found that Pyk2 phosphorylation was Src and Ca2+ dependent, we speculated that Src and Ca2+ may be acting via Pyk2, leading to EGFR transactivation. Thus, to address whether Src and/or Pyk2 physically interacted with EGFR after PAR-2 activation, we conducted experiments in which EGFR was immunoprecipitated from SCBN whole cell lysates following PAR-2 activation, with subsequent immunoblotting for Src and Pyk2. However, on the basis of our coimmunoprecipitation experiments, neither Src nor Pyk2 appears to form a complex with EGFR. This would suggest that either this complex is not detectable by conventional coimmunoprecipitation protocols or the response involves another Src-dependent signaling event that leads to the transactivation of EGFR.
On EGFR activation, a number of signaling pathways are stimulated, including the activation of phosphatidylinositol 3-kinase, Grb2, SOS, Ras, cRaf, and downstream MAP kinase, specifically ERK1/2 (29, 40, 42, 57). Since EGFR transactivation plays an important role in PAR-2-induced Cl− secretion in SCBN cells, other kinases downstream of EGFR activation are likely to be involved. We showed that, on PAR-2 activation, there was an upregulation in cRaf phosphorylation, an important signaling molecule upstream of ERK1/2. Interestingly, we found that neither Src nor EGFR inhibition significantly reduced cRaf phosphorylation. However, both significantly reduced ERK1/2 phosphorylation, suggesting other potential mechanisms of cRaf activation following PAR-2 activation. PKC is known to stimulate Ras/cRaf signaling events (36) and, on the basis of our finding that PKC inhibition significantly reduces PAR-2-induced Cl− secretion, PKC may be important in the activation of the Ras/cRaf signaling pathway. Indeed, we found this to be the case; not only did PKC inhibition significantly reduce cRaf phosphorylation, it also completely inhibited ERK1/2 phosphorylation. These findings suggest an important role for PKC- and Src-mediated EGFR transactivation in the stimulation of ERK1/2. A similar dual signaling pathway that converges at the level of Ras/cRaf has been reported for bradykinin signaling, where EGFR transactivation and PKC stimulated ERK1/2 phosphorylation in COS-7 cells (1). In addition, we showed that ERK1/2 was phosphorylated following PAR-2 activation and that this effect was blocked by inhibitors of Src, EGFR tyrosine kinase, and MEK. Importantly, these inhibitors also blocked PAR-2-dependent changes in Isc, suggesting that ERK1/2 activation is a key event in the response.
One of the possible downstream effectors of ERK1/2 includes the activation of COX. We showed that COX activity was required for PAR-2-induced Cl− secretion. Previous studies have also demonstrated that PAR-2-induced Cl− secretion may be dependent on COX activity and arachidonic acid metabolites such as prostaglandins (24, 50). Recent studies showed that PAR-2 activation can lead to the induction of COX-2 through an ERK1/2-dependent mechanism in epithelial (32) and endothelial (54) cells. However, given the time course of the PAR-2-induced change in Isc (∼5 min), the COX dependence of the response must be due to constitutively expressed COX, rather than induction of new enzyme. Because we used a nonselective COX inhibitor, indomethacin, we cannot speculate on which COX isoform is responsible for the COX dependency of the PAR-2-evoked change in Isc. Although we have not provided evidence for a link between ERK1/2 and COX in the PAR-2-mediated response, previous studies with PAR-1, a GPCR that also couples to Gqα, indicate an ERK1/2-dependent activation of phospholipase A2, which liberates arachidonic acid, which could then be metabolized to secretory prostanoids by constitutively expressed COXs (8).
Taken together, these results suggest the activation of a complex set of signaling events leading to the rapid upregulation of Cl− secretion following PAR-2 activation. Given the prominent role of Cl− and water secretion in intestinal epithelial barrier function and the central role of barrier disruption in chronic inflammatory bowel diseases such as Crohn's disease, it will be critical to identify the proteinases that are released and could activate epithelial PAR-2 in the setting of inflammatory bowel disease.
This work was supported by grants to M. D. Hollenberg and W. K. MacNaughton from the Canadian Institutes of Health Research. W. K. MacNaughton is a Senior Scholar of the Alberta Heritage Foundation for Medical Research.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2008 the American Physiological Society