Abstract

The thrombin receptor, protease-activated receptor-1 (PAR-1), has wide tissue distribution and is involved in many physiological functions. Because thrombin is in the intestinal lumen and mucosa during inflammation, we sought to determine PAR-1 expression and function in human intestinal epithelial cells. RT-PCR showed PAR-1 mRNA expression in SCBN cells, a nontransformed duodenal epithelial cell line. Confluent SCBN monolayers mounted in Ussing chambers responded to PAR-1 activation with a Cl-dependent increase in short-circuit current. The secretory effect was blocked by BaCl2 and the Ca2+-ATPase inhibitor thapsigargin, but not by the L-type Ca2+ channel blocker verapamil or DIDS, the nonselective inhibitor of Ca2+-dependent Cl transport. Responses to thrombin and PAR-1-activating peptides exhibited auto- and crossdesensitization. Fura 2-loaded SCBN cells had increased fluorescence after PAR-1 activation, indicating increased intracellular Ca2+. RT-PCR showed that SCBN cells expressed mRNA for the cystic fibrosis transmembrane conductance regulator (CFTR) and hypotonicity-activated Cl channel-2 but not for the Ca2+-dependent Cl channel-1. PAR-1 activation failed to increase intracellular cAMP, suggesting that the CFTR channel is not involved in the Cl secretory response. Our data demonstrate that PAR-1 is expressed on human intestinal epithelial cells and regulates a novel Ca2+-dependent Clsecretory pathway. This may be of clinical significance in inflammatory intestinal diseases with elevated thrombin levels.

  • thrombin
  • epithelium
  • ion transport
  • serine endopeptidases
  • chloride channels

protease-activated receptors (PARs) are a unique class of G protein-coupled receptors activated by serine proteases that cleave specific regions of the extracellular NH2 terminus of the molecule to reveal a new NH2 terminus that acts as a “tethered ligand.” The tethered ligand, by binding to other extracellular domains on the PAR molecule, stimulates G protein-dependent signaling (34,35). Four PARs have been cloned to date (17, 18, 24, 34,38), and others have been predicted based on pharmacological structure-activity relationships (26, 28, 33). PAR-1, -3, and -4 are activated by the coagulation factor thrombin, whereas PAR-2 is activated by trypsin and possibly by mast cell tryptase (5,21, 23). PAR-1 was the first PAR cloned and is the prototypical thrombin receptor. It is found in a wide variety of cell types, including platelets, endothelial cells, fibroblasts, monocytes, T cell lines, osteoblast-like cells, smooth muscle cells, neurons, and glial cells, and in certain tumor cell lines (7).

Thrombin has long been known to be involved in inflammation and has been implicated in the pathogenesis of inflammatory bowel disease. Thrombin and PAR-1 have critical proinflammatory effects such as platelet aggregation, vasodilatation and vasoconstriction, increased vascular permeability, and granulocyte chemotaxis (8). Patients with Crohn's disease show various coagulation abnormalities, and intestinal vascular injury has been proposed as a major pathogenic factor (30, 36). Chronic inflammatory bowel disease, especially ulcerative colitis, is associated with a thrombotic tendency (12). Thus thrombin would be in a position to affect epithelial function in the inflamed gut.

The epithelium plays a central role in host defense. The ability of epithelial cells in the crypt region to secrete chloride ions and water prevents or slows the translocation of bacteria, bacterial products, and antigens from the intestinal lumen to the mucosa (4). A loss of basal secretion or of the ability to respond to secretagogues renders the organism susceptible to infection. Conversely, hyperresponsiveness to secretagogues tips the absorption and secretion balance in the other direction, resulting in diarrhea that could lead to excessive electrolyte and water loss. Electrolyte-transporting epithelial cells from nonintestinal tissues, including renal proximal tubule cells (13), have been shown to express PAR-1. However, intestinal epithelial expression of PAR-1 has not been demonstrated nor has thrombin been linked to intestinal epithelial cell function or dysfunction.

In this study, we tested the hypothesis that PAR-1 is expressed on intestinal epithelial cells and can modulate epithelial secretion of chloride and, hence, water. To do this, we assessed the expression and function of PAR-1 in a nontransformed, chloride-secreting human crypt cell line. Furthermore, because chloride secretion by intestinal epithelial cells can occur through calcium- or adenylate cyclase-dependent pathways, we conducted further experiments to determine which of these pathways is involved in PAR-1-induced chloride transport.

MATERIALS AND METHODS

SCBN cells.

SCBN is a nontransformed, chloride-secreting, duodenal epithelial crypt cell line (25). SCBN cells were grown according to previously published methods (25, 29). Briefly, cells ofpassages 23–33 were grown to confluence (∼5 days) in either 75-cm2 flasks or on Snapwell semipermeable supports (Corning, Corning, NY). Cells in Snapwell supports were fed every day with DMEM supplemented with 5% fetal bovine serum,l-glutamine, streptomycin, and tylosin. Cells in flasks were fed with fresh medium every 2–3 days.

RT-PCR.

RT-PCR was conducted to determine the expression of PAR-1 and the chloride channels, cystic fibrosis transmembrane conductance regulator (CFTR), calcium-dependent chloride channel-1 (CLCA-1), and chloride channel-2 (ClC-2) in SCBN cells. The procedure for RT-PCR was conducted using the “primer-dropping” method (37) as we (20) have previously described. Briefly, monolayers of SCBN cells were grown in T25 flasks to near confluence as determined by light microscopy. Cells were scraped from the flasks and RNA extracted using 1.5 ml TRIzol reagent per flask. Samples were frozen on dry ice and stored in TRIzol at −80°C until processed for RNA extraction. Samples were thawed, and 0.3 ml of chloroform was added to the tube. Samples were centrifuged at 11,700 g for 30 min at 4°C, and 500 μl of the aqueous phase were removed and mixed with 750 μl of isopropanol to precipitate RNA. The samples were centrifuged at 12,000 g for 20 min. The RNA pellet was then washed in 75% ethanol and centrifuged for 5 min at 7,500 g and redissolved in 50 μl of ultrapure autoclaved water. The purity and concentration of the RNA were measured using a Gene Quant II nucleic acid analyzer (Pharmacia Biotech, Uppsala, Sweden). RNA (2 μg) was added to a reaction mixture containing 2 μl of 10× PCR buffer, 2 μl of 10 mM dNTPs, 2 μl N6, and 0.5 μl of RNAguard. Superscript enzyme (300 U, GIBCO BRL, Burlington, ON, Canada) was added for reverse transcription. RNA samples were first incubated for 10 min, and the reaction mixture was heated to 42°C for 50 min and then to 95°C (to destroy the Superscript enzyme) in a DNA Engine thermal cycler (MJ Research, Waltham, MA). PCR was performed on either the cDNA from the RT reaction or RT negative samples to control for contamination with genomic DNA, using the primer sequences shown in Table1. PCR for PAR-1 was stopped after 44 cycles (denaturation at 94°C for 17 s, annealing at 53°C for 1 min, and elongation at 72°C for 1 min), for CFTR after 44 cycles (denaturation at 94°C for 30 s, annealing at 54°C for 30 s, and elongation at 72°C for 1 min), for CLCA-1 after 44 cycles (denaturation at 94°C for 30 s, annealing at 54°C for 30 s, and elongation at 72°C for 1 min), and for ClC-2 after 46 cycles (denaturation at 94°C for 17 s, annealing at 62°C for 30 s, and elongation at 72°C for 1 min). PCR for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was stopped after 22 cycles (denaturation at 94°C for 30 s, annealing at 55°C for 30 s, and elongation at 72°C for 1 min). PCR products were purified using the QIAquick PCR purification kit (Qiagen, Mississauga, ON, Canada) and sequenced by the University of Calgary Core DNA Service. Sequences were compared with those published on the National Institutes of Health Genbank database.

View this table:
Table 1.

Primer sequences used in RT-PCR

Assessment of chloride secretion.

SCBN monolayers were grown to confluence on Snapwell semipermeable supports. Confluence was determined by the increase in resistance across the monolayers as measured by an electrovoltohmmeter (EVOM, World Precision Instruments, Sarasota, FL). Only monolayers with a resistance >1,000 Ω/cm2 were used. Snapwell supports were mounted in modified Ussing chambers and bathed on the apical side with Krebs buffer containing 10 mM mannitol and on the basolateral side with Krebs buffer containing 10 mM glucose. Krebs buffer contained in mM: 115 NaCl, 2 KH2PO4, 2.4 MgCl2, 25 NaHCO3, 8 KCl, and 1.3 CaCl2. The transepithelial potential difference was clamped to 0 V by applying a short-circuit current (I sc) with a voltage clamp apparatus (EVC-4000, World Precision Instruments). The change inI sc was an indicator of the change in the net electrogenic electrolyte flux across the monolayer.I sc was recorded with a digital data acquisition system (MP100, BioPac, San Diego, CA) and analyzed with AcqKnowledge software (version 3.1.3, BioPac).

PAR-1 activation was accomplished with addition of either thrombin or the PAR-1-activating peptides TFLLR-NH2 (16) or Ala-parafluoroPhe-Arg-cyclohexyl-Ala-Cit-Tyr-NH2(Cit-NH2) (19). When used, inhibitors were added to both the basolateral and apical sides of the monolayers, with the exception of DIDS, which was added to the apical surface only.

Determination of cAMP.

CFTR intracellular trafficking (31) and channel function (10) are stimulated by increases in the intracellular concentration of cAMP. To determine the effects of PAR-1 activation on cAMP levels, experiments were conducted in which five 75-cm2 flasks of confluent SCBN cells were rinsed with sterile PBS. Fetal bovine serum-free DMEM (4 ml) was then added to each flask, and cells were scraped using a cell scraper. Cells and medium were pooled into a 50-ml centrifuge tube, and 1-ml aliquots of cell suspension were transferred to microfuge tubes. The phosphodiesterase inhibitor B-8279 (Sigma Chemical, Mississauga, ON, Canada) was added to each tube for a final concentration of 100 μM. Immediately after, forskolin (10 μM), Cit-NH2 (5 μM), thrombin (5 U/ml), or the vehicle control HEPES were added and the suspensions allowed to incubate for 5 min at room temperature. The tubes were then immersed in liquid nitrogen, cycled three times between liquid nitrogen and a 40°C water bath to lyse the cells, and centrifuged at 15,300g for 5 min at 4°C. Concentrations of cAMP were determined in supernatants using a commercial ELISA kit (R & D Systems, Minneapolis, MN) according to the manufacturer's instructions.

Materials.

Thrombin (human; sp act, 2,800 NIH U/mg protein) was purchased from Calbiochem (San Diego, California). The PAR-1-activating peptides were synthesized as carboxyamides in-house by the University of Calgary Peptide Synthesis Facility (directed by Dr. Dennis McMaster). Thapsigargin was obtained from Alomone (Jerusalem, Israel). Routine buffer reagents were purchased from BDH (Toronto, ON, Canada). Unless otherwise specified, all other drugs and reagents were purchased from Sigma Chemical.

Statistics.

Data are expressed as means ± SE. Comparison of more than two groups was made by ANOVA with the post hoc Tukey's test using Instat version 3.00 software (GraphPad Software, San Diego, CA). Comparison of two groups was made using Student's t-test for unpaired data. P < 0.05 was considered significant.

RESULTS

PAR-1 expression and response to activation.

SCBN cells constitutively express PAR-1 mRNA as shown in Fig.1. The PCR product was sequenced and compared with the Genbank database, which confirmed human PAR-1 (Genbank accession no. M62424). The level of expression of PAR-1 was similar to that of the housekeeping gene GAPDH.

Fig. 1.

Reverse image of RT-PCR gel for glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and protease-activated receptor-1 (PAR-1). Lane 1, base pair marker lane;lane 2, PCR conducted with primers and cDNA obtained from RT (PAR-1 and GAPDH were present); lane 3, PCR conducted without cDNA; lane 4, PCR conducted without primers.

SCBN cells grown to confluence on Snapwell supports exhibited a baseline I sc of −1 ± 0.3 μA/cm2. Basolateral exposure of monolayers to thrombin or to the PAR-1-activating peptides Cit-NH2 or TFLLR-NH2 caused a rapid increase inI sc that reached a peak within 2 min and subsequently returned to baseline over the next 5 min (Fig.2). The maximal change inI sc observed after basolateral application of PAR-1 activators was concentration dependent (Fig.3). Apical application of thrombin caused no change in I sc, whereas apical application of the PAR-1-activating peptides caused only slight increases inI sc, which were much lower than those observed after basolateral application. The control peptide FSLLR-NH2 (100 μM) did not cause a change inI sc when added to the basolateral side of SCBN monolayers (Fig. 2). The monolayers exposed to FSLLR-NH2responded to subsequent addition of the cAMP-dependent secretagogue forskolin (10 μM), with an increase in I sc, indicating that these preparations were still viable. Incubation of monolayers in chloride-free Krebs buffer significantly reduced the increases in I sc that occurred after basolateral application of thrombin or Cit-NH2 (Figs. 2 and4), suggesting that the change inI sc observed after PAR-1 activation was due primarily to stimulation of chloride secretion. A residual increase inI sc was observed in the presence of chloride-free buffer, however, and was likely due to bicarbonate ion transport (6). In addition, the secretory response to thrombin or Cit-NH2 was reduced by pretreatment of the monolayers with the nonselective potassium channel blocker barium chloride (2 mM; Fig. 4).

Fig. 2.

Representative traces of short-circuit current (I sc) responses to thrombin and PAR-1-activating peptides. The y-axis showing absoluteI sc for each trace is included as baselines varied slightly from one preparation to another. A: exposure to thrombin, Cit-NH2, and TFLLR-NH2 caused a rapid, transient increase in I sc. B: the inactive control peptide FSLLR-NH2 (FS, 100 μM) did not change I sc. This preparation was still viable because it responded positively to basolateral application of forskolin (FSK). C: the increases inI sc in response to basolateral application of thrombin or Cit-NH2 were reduced or abolished in Cl-free Krebs buffer. D: theI sc responses to thrombin and Cit-NH2 were mediated through the same receptor, because treatment with one agonist caused desensitization to subsequent application of the other agonist. Cell monolayers were still viable as they responded normally to subsequent basolateral application of forskolin.

Fig. 3.

Concentration-response relationships for increases inI sc induced by activation of PAR-1. Thrombin (A), Cit-NH2 (B), and TFLLR-NH2 (C) all caused concentration-dependent increases in I sc when applied basolaterally. The response to TFLLR-NH2 was minimal at the highest concentration tested (100 μM). Responses to apically applied PAR-1 agonists were substantially smaller than those observed after basolateral activation.

Fig. 4.

I sc responses of monolayers of SCBN cells mounted in Ussing chambers to 5 U/ml thrombin and 10 μM Cit-NH2. A: cells were incubated in normal Krebs buffer or Cl-free modified Krebs buffer. B: monolayers were incubated in normal Krebs buffer or Krebs buffer containing 2 mM BaCl2. * P < 0.05; ** P < 0.01.

To ensure that the I sc responses observed were due to activation of PAR-1, studies were conducted in which thrombin or a PAR-1-AP was added basolaterally to the monolayers to elicit an increase in I sc. Once baselineI sc had been reestablished (5–8 min), thrombin or the PAR-1-AP was added basolaterally again to assess desensitization. Activation of PAR-1 with either thrombin or an activating peptide causes rapid desensitization of the receptor and, hence, unresponsiveness to subsequent attempts to elicit a PAR-1-dependent response (32). Exposure to a PAR-1 activator caused almost complete desensitization to a subsequent application of the same PAR-1 activator (Fig.5). Furthermore, activation of PAR-1 caused crossdesensitization to subsequent application of a different PAR-1 activator (Figs. 2 and 5).

Fig. 5.

I sc responses of monolayers of SCBN cells mounted in Ussing chambers to Cit-NH2 (Cit; 5 μM), TFLLR-NH2 (TF; 25 μM), or thrombin (Thr; 5 U/ml). Cells were not pretreated (control; C) or were pretreated with the same PAR-1 activator to demonstrate autodesensitization or with a different PAR-1 activator to demonstrate crossdesensitization. * P< 0.05; ** P < 0.01; *** P < 0.001.

Calcium dependency of PAR-1-induced chloride secretion.

PAR-1 activation leads to a rapid and transient increase in intracellular calcium (16). To determine if the chloride secretion induced by PAR-1 activation was due to increased intracellular calcium concentration, we first determined if PAR-1 activation caused increased cytosolic calcium in SCBN monolayers grown on coverslips and loaded with the calcium indicator fluorophore fura 2. Figure 6 shows representative micrographs with a fluorescence ratio of 1 × 105 cells/monolayer exposed to thrombin (50 U/ml) or Cit-NH2 (100 μM). Both thrombin and Cit-NH2 caused an increase in fluorescence ratio. However, not all cells responded with an increase in fluorescence. Both thrombin and Cit-NH2 caused crossdesensitization to the other PAR-1 agonist (Fig. 6). The higher concentration of agonist required to activate PAR-1 in this system, compared with the chloride secretion experiments, may be due to the monolayers being grown on glass coverslips. This would limit access of apically added compounds to the basolateral surface of the cells, where the receptor is likely located.

Fig. 6.

Ca2+ signaling responses in SCBN cells grown on coverslips and loaded with fura 2. A: representative traces of the change in fluorescence ratio after PAR-1 activation. Both thrombin and Cit-NH2 increased the intracellular concentration of Ca2+ and caused crossdesensitization to the other PAR-1 agonist. B: representative micrographs of SCBN cells grown on glass coverslips and loaded with fura 2. Increasing brightness represents computer-generated images of increasing fluorescence ratio. Brighter areas indicate increased Ca2+in response to thrombin (50 U/ml) or Cit-NH2 (100 μM). The 0-s time point indicates the fluorescence ratio immediately before addition of the PAR-1 activator.

To determine if the increases in calcium elicited by PAR-1 activation were coupled to chloride secretion, we conducted experiments with SCBN monolayers grown on Snapwell supports and mounted in Ussing chambers. The responses to thrombin and Cit-NH2 were assessed in the presence of the intracellular calcium reuptake inhibitor thapsigargin (250 nM), the L-type calcium channel blocker verapamil (10 μM), or DIDS (300 μM), which blocks CLCA, but not CFTR. Only thapsigargin reduced the responsiveness of the SCBN cells to thrombin or Cit-NH2 compared with vehicle controls (Fig.7). Neither DIDS nor verapamil affected the CFTR-dependent response to forskolin. Thapsigargin caused a prolonged increase in I sc, and subsequent exposure to thrombin resulted in a decrease inI sc. Thapsigargin pretreatment did not affect the response to forskolin (Fig. 7).

Fig. 7.

Effect of Ca2+ blockers on theI sc response to thrombin (5 U/ml; A) or Cit-NH2 (10 μM; B) by SCNB cell monolayers in Ussing chambers. Monolayers were treated with drug 20–30 min before addition of either thrombin or Cit-NH2.C: representative trace of I scresponses in the presence of thapsigargin. Thapsigargin caused an increase in I sc and reversed the response to subsequent exposure to thrombin but did not affect the response to forskolin. Con, control; Thaps, thapsigargin; Verap, verapamil. ** P < 0.01; *** P < 0.001.

Chloride transport in intestinal epithelial cells can occur through several channels. Apically directed chloride transport occurs through either the CFTR, which is modulated through cAMP, or CLCA. Furthermore, ClC-2 has recently been implicated in vectorial chloride secretion by intestinal epithelial cells (22). Because the SCBN cell monolayers responded to the cAMP-dependent secretagogue forskolin and because the I sc response to PAR-1 activation was calcium dependent, we assumed that both chloride conductance pathways were present in SCBN cells. However, because the channel responsible for chloride secretion in SCBN cells has not been characterized, we conducted RT-PCR to look for expression of CFTR, ClC-2, and CLCA-1. PCR products were purified using the Qiagen PCR purification kit and sequenced at the University of Calgary Core DNA Service. Sequences of amplified DNA were compared with those available in the Genbank database. The results of RT-PCR indicate that SCBN cells express mRNA for both CFTR and ClC-2. Furthermore, when PCR was performed on RT-negative samples to control for possible contamination with genomic DNA, no product was found. We were unable to show CLCA-1 mRNA expression via this method (Fig. 8).

Fig. 8.

Reverse image of RT-PCR for cystic fibrosis transmembrane conductance regulator (CFTR), Ca2+-dependent Cl channel-1 (CLCA-1), Cl channel-2 (ClC-2), and GAPDH in SCBN cells. A: CFTR was expressed in these cells, whereas CLCA-1, expected at 509 bp based on our primer sequence, was not; −, the negative control for CFTR, ruling out contamination with genomic DNA. B: SCBN cells express ClC-2 mRNA; −, negative control.

To rule out a role for cAMP-dependent transport in theI sc response to PAR-1 activation, we conducted further experiments in which SCBN cells were exposed to vehicle, thrombin, Cit-NH2 or forskolin and subsequently assayed for cAMP using ELISA. PAR-1 activation caused no increase in cAMP compared with vehicle control. In contrast, exposure of the cells to forskolin caused a large increase in cAMP levels in these cells (Fig.9).

Fig. 9.

Adenylate cyclase activity assessed in SCBN cells after exposure to HEPES buffer (control), thrombin (Throm, 5 U/ml), Cit-NH2 (10 μM), or forskolin (10 μM). Activity is represented as the generation of cAMP in the presence of a phosphodiesterase inhibitor. PAR-1 activation failed to activate adenylate cyclase. * P < 0.05.

DISCUSSION

In the present study, we have shown that PAR-1 is expressed in a nontransformed, human small intestinal epithelial crypt cell line (SCBN) and that activation of PAR-1 on the basolateral surface of SCBN cells results in apically directed chloride secretion. Through the use of RT-PCR and selective PAR-1-activating peptides, we have shown that it is indeed PAR-1, and not a different PAR, being activated by thrombin. Furthermore, we have shown that PAR-1 activation leads to chloride secretion through a mechanism that is dependent on release of calcium from intracellular stores and independent of adenylate cyclase activity.

Four PARs have been cloned to date. Of these, PAR-1 and PAR-4 in conjunction with PAR-3 are activated by thrombin whereas PAR-2 is activated by trypsin and perhaps tryptase. An important role for PARs is suggested by their wide tissue distribution. For example, PAR-1 is expressed in human platelets, endothelial cells, fibroblasts, monocytes, T cell lines, osteoblast-like cells, smooth muscle cells, neurons, and glial cells in the brain and periphery and certain tumor cell lines (7). Our study is the first to demonstrate PAR-1 expression in the intestinal epithelium and to link its activation to ion transport.

We chose the SCBN duodenal epithelial cell line because it is nontransformed, derived from humans, and capable of apically directed chloride transport and it forms electrically tight monolayers, thus making it amenable to study in the Ussing chamber (25). In our hands, SCBN monolayers consistently demonstrated electrical resistances of >2,000 Ω/cm2 (data not shown). The SCBN cell line also expresses PAR-1 mRNA as shown by RT-PCR. Sequencing of the PCR product showed identity with the sequence in the Genbank database (accession no. M62424), confirming that our product was indeed human PAR-1. Furthermore, in Ussing chamber studies in which the SCBN monolayers were studied under voltage clamp conditions, application of thrombin or the PAR-1 activating peptides Cit-NH2 and TFLLR-NH2 caused a rapid increase inI sc when applied to the basolateral side of the monolayers and almost no response when applied apically, suggesting that PAR-1 expression is polarized and restricted to the basolateral aspect of the cells. The fact that there was a slightI sc response to apical application of the PAR-1-activating peptides suggests that either the peptides were able to translocate across the epithelium to activate basolateral PAR-1 or there is a sparse population of apical PAR-1. PAR-1 localization in renal proximal tubular epithelium is also polarized and basolateral (13). The I sc changes observed after PAR-1 activation were due to the movement of chloride ion, because responses were significantly reduced in chloride-free buffer. The residual increase in I sc seen in the absence of chloride was likely due to bicarbonate secretion (6). Reduction of the secretory response to PAR-1 activation by inhibition of potassium conductance using barium chloride also confirmed the involvement of the basolateral potassium channel, which is involved in maintaining the energetics favorable to chloride transport (2).

Having demonstrated that PAR-1 activation could stimulate vectorial chloride transport, we next sought to determine which chloride channel was involved. Intestinal epithelial cells express a number of chloride channels. The channels involved in apical, outwardly directed chloride currents are CFTR and CLCA (2, 9). Furthermore, ClC-2, which in intestinal cells has previously been linked to volume regulation, has recently been shown (22) to have a polarized distribution in the cells and to contribute to chloride currents. CFTR activation is controlled by several intracellular signaling pathways that control channel phosphorylation and dephosphorylation (11). Conspicuous among these is a cAMP-dependent pathway that stimulates CFTR trafficking to the apical membrane (31) and channel opening (11). SCBN cells respond to forskolin (25), which activates adenylate cyclase, suggesting that CFTR is present in these cells. In this study, we have confirmed the responsiveness of this cell line to forskolin and shown that CFTR mRNA is expressed. The latter was determined by RT-PCR and confirmed by sequencing of the PCR product (Genbank accession no. NM000492). However, despite the presence of CFTR in this cell line, the response to PAR-1 activation is not due to a CFTR-mediated chloride current. PAR-1 activation failed to increase intracellular cAMP concentrations, whereas forskolin caused a 70-fold increase. Furthermore, inhibition of calcium-dependent signaling pathways blocked the response to PAR-1 activation but did not affect the response to forskolin.

Activation of calcium-dependent pathways is a hallmark of PAR-1-initiated signaling (16, 19). Using fura 2-labeled SCBN cells, we showed that PAR-1 activation caused a rapid, transient increase in intracellular calcium that followed a time course similar to that observed for the increase in I sc in ion transport studies. The transient nature of the calcium signaling response is characteristic of PAR-1 activation (16, 19). Not all cells responded to PAR-1 activators. This could be due to a heterogeneous population of SCBN cells, in which only a subset of cells express PAR-1, or to lack of accessibility of the activator to the basolaterally located receptor. The fact that larger concentrations of thrombin and Cit-NH2 were required to elicit a calcium signal compared with those required to stimulate changes inI sc can be explained by the latter point. Furthermore, we showed that PAR-1-mediated chloride secretion in SCBN cells was calcium dependent because it was inhibited by thapsigargin. The fact that the chloride secretory response to PAR-1 activation is reduced by depletion of intracellular calcium stores with thapsigargin, and not by blocking the entry of extracellular calcium with verapamil, suggests that intracellular reserves are being used in the signaling process. As expected, thapsigargin on its own stimulated an increase inI sc, likely due to the initial liberation of calcium from intracellular stores. Interestingly, pretreatment with thapsigargin resulted in a reversal of the current response to subsequent PAR-1 activation. We speculate that removal of stored calcium uncovers an additional signal transduction pathway that inhibits chloride secretion, but which is masked by the larger stimulatory signal present when intracellular calcium is available. However, a thorough investigation of this inhibitory pathway was beyond the scope of this study.

The calcium dependency of PAR-1-induced chloride secretion suggested that one of the members of the CLCA family was involved. CLCA-1 is the isoform that is expressed in human small intestine (14). However, we were unable to detect the presence of human CLCA-1 by RT-PCR using previously published (14) primers and conditions. Nevertheless, this does not dismiss the possibility that other chloride channels of this type, distinct from human CLCA-1, are present in SCBN cells. New members of the calcium-activated chloride channel family have recently been characterized (1, 9) and more will undoubtedly be cloned. CLCA can be demonstrated pharmacologically by inhibition with DIDS (14). In our study, PAR-1-mediated chloride secretion is not inhibited by apical DIDS. Taken together, the weight of evidence points to the selective activation of a chloride channel distinct from CFTR and CLCA by PAR-1 activation.

It is possible that ClC-2 is involved in PAR-1-induced chloride secretion in this cell line. Mohammad-Panah et al. (22) have shown that ClC-2 is expressed in human intestinal epithelial cells and that this distribution may be polarized. Using RT-PCR, we showed that SCBN cells express ClC-2. Further study is necessary to show unequivocally that ClC-2 mediates the chloride secretory response to PAR-1 activation.

The presence of PAR-1 on intestinal epithelial cells has important implications for barrier function during tissue injury. The intestinal epithelium provides an important barrier to the translocation of luminal pathogens or antigens to the mucosal lamina propria (4). This barrier consists of tight junctions preventing paracellular permeation of luminal constituents, plus the ability of crypt cells to secrete chloride and water (4). A compromised barrier allows access of luminal contents to the lamina propria, which in turn may initiate, exacerbate, or perpetuate inflammation. Thrombin has been shown to play a role in the inflammatory response, including that which characterizes inflammatory bowel disease (27). The epithelium, particularly the basolateral surface, would be exposed to thrombin during inflammation-associated hemorrhage. Two scenarios can be envisioned for the interaction of thrombin with epithelial PAR-1 in vivo. First, PAR-1 activation could lead to a secretory response that could contribute to the diarrhea symptomatic of inflammatory conditions of the gut. Conversely, PAR-1 activation during inflammation may represent an activation of the secretory component of the epithelial host defense function.

Acknowledgments

This work was supported in part by grants from the Canadian Institutes of Health Research (formerly the Medical Research Council of Canada; M. D. Hollenberg, J. L. Wallace, and W. K. MacNaughton) and the Natural Sciences and Engineering Research Council (A. Buret). M. C. Buresi is funded by an Alberta Heritage Foundation for Medical Research studentship and N. Vergnolle by the University of Calgary NicOx Chair in Inflammation Research. W. K. MacNaughton is an Alberta Heritage Foundation for Medical Research Scholar. J. L. Wallace is an Alberta Heritage Foundation for Medical Research Senior Scientist and a Canadian Institutes of Health Research Senior Scientist.

Footnotes

  • Address for reprint requests and other correspondence: W. MacNaughton, Mucosal Inflammation Research Group, Univ. of Calgary, 3330 Hospital Dr NW, Calgary, Alberta, T2N 4N1 Canada (E-mail:wmacnaug{at}ucalgary.ca).

  • 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.

REFERENCES

View Abstract