INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Keywords: tryptase; inflammation; asthma

In addition to degrading proteins, certain serine proteases can directly signal to cells by cleaving and triggering proteinase-activated receptors (PARs), members of a new family of heptahelical receptors (1). Proteases hydrolyze at specific sites in the extracellular tails of their receptors to expose tethered ligand domains that bind to and activate the cleaved receptors. In this manner thrombin, which is generated in the circulation during coagulation, activates PAR1, PAR3, and PAR4, and trypsin and tryptase, a major protein of human mast cells, activate PAR2. PARs couple to multiple signaling events, most of which mediate tissues responses to injury, such as inflammation and repair. Thus, proteases and their receptors are important therapeutic targets that have been implicated in human diseases.

Proteases that are known to activate PAR2 have major effects on airway function that may be relevant to human disease. Tryptase, which is released from mast cells when they degranulate, acts on airway epithelial cells to induce expression of intercellular adhesion molecule-1 (ICAM-1) and stimulate secretion of interleukin-6 (IL-6), IL-8, and granulocyte macrophage colony-stimulating factor (GM-CSF) (4, 5). Tryptase also increases vascular permeability (6, 7), contributes to airway hyperresponsiveness (8), and is mitogenic for airway smooth muscle cells and fibroblasts (11). Tryptase inhibitors are also effective in experimental models of asthma (15) and in the treatment of asthma in humans (16). Trypsinogens are also expressed by airway epithelial cells (17), but their regulation and function remain to be elucidated. The observations that trypsin and tryptase activate PAR2 raise the possibility that the widespread effects of these proteases are due to activation of this receptor.

PAR2 is expressed by epithelial cells, endothelial cells, and myocytes in the guinea-pig and mouse airways (17, 18). Agonists of PAR2, including proteases and peptides corresponding to the tethered ligand domain, affect airway tone in experimental animals, although the effects depend on the species and the model. For example, in mice, PAR2 agonists cause bronchodilation by an epithelium-dependent and eicosanoid-dependent mechanism, leading to the suggestion that PAR2 agonists are protective (17). In contrast, in guinea pigs, in vivo PAR2 agonists cause bronchoconstriction (18), supporting the hypothesis that tryptase and PAR2 are implicated in asthma. However, little is known about the role of PAR2 in the human airways. In view of the proposed use of tryptase inhibitors to treat asthma in humans (16), it is important to determine the location and function of PAR2 in the human airways.

The purpose of the present study was to determine the localization and the possible motor function of PAR2 in human airways. Our aims were to (1) localize PAR2 messenger RNA (mRNA) and protein in primary cultures of airway smooth muscle cells; (2) determine the ability of trypsin, tryptase, and a peptide corresponding to the tethered ligand of PAR2 to signal directly to cultured smooth muscle cells; (3) localize immunoreactive PAR2 in normal human airway; and (4) evaluate the effect and mechanism of PAR2 agonists on motility of isolated human bronchi.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

PAR Agonists

Sources of trypsin, tryptase, and thrombin have been described (19- 22). Peptides corresponding to the tethered ligands of human PAR2 (SLIGKV-NH₂, SL-NH₂), PAR1 (SFLLR-NH₂), PAR4 (GYPGQV-NH₂), and their reverse sequences were synthesized by solid-phase methods.

Human Tissues

Lobar or segmental bronchial rings from grossly normal tissue were taken from 24 patients undergoing lung resection for a solitary peripheral carcinoma. The study conformed to the Declaration of Helsinki and was approved by the ethic committee of the University of Ferrara.

Human Airway Smooth Muscle Cells (HASMC)

HASMC were isolated from human tissue and were maintained as described (23). HASMC were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum (FCS), insulin (1 µM), penicillin (100 U/ml), streptomycin (100 µg/ml), and glutamine (2 mM) at 37° C in a humidified incubator. Cells were passaged using enzyme-free cell dissociation buffer and were used for experiments between the fifth and the sixth passages. HASMC retain native contractile protein expression, as demonstrated by staining for smooth muscle-specific actin.

PAR2 Reverse Transcriptase/Polymerase Chain Reaction (RT-PCR)

Total RNA from HASMC (0.5 µg) was reverse transcribed (19, 20). Primers to human PAR2 (forward: 5'-CCC TTT GTA TGT CGT GAA GC-3', reverse: 5'-TTC CTG GAG TGT TTC TTT GAG G-3') were chosen to amplify a 525 bp fragment. After denaturation of 7 min, 35 cycles at 94° C for 1 min, 62° C for 1 min, and 72° C for 1 min were followed by elongation at 72° C for 7 min. Reverse transcriptase was omitted in controls. KNRK (normal rat kidney cells transformed by Kirsten sarcoma virus) overexpressing human PAR2 were used as a positive control. Products were analyzed on a 1% agarose ethidium bromide gel.

PAR2 Localization

HASMC were fixed in 4% paraformaldehyde in 100 mM phosphate-buffered saline (PBS), pH 7.4 for 20 min at 4° C (19, 20). PAR2 was localized by indirect immunofluorescence using rabbit antiserum B5 to rat PAR2: ³⁰GPNSKGR SLIGRLDT⁴⁶P-YGGC ( cleavage site, YGGC for conjugation) (24). To verify retention of their phenotype, muscle cells were similarly stained for smooth muscle actin. Bronchial rings were fixed in 1% paraformaldehyde in 100 mM PBS, pH 7.4 for 6 h, washed with PBS/15% sucrose, and frozen sections were prepared. PAR2 was detected using the B5 antiserum by the streptavidin-biotin complex peroxidase method with nickel enhancement (25, 26). Controls included preabsorption of antiserum with 1 mM to 10 µM of receptor fragment, or omission of the primary antibody.

PAR2-induced Ca²⁺ Mobilization

HASMC were incubated with 2.5 mM Fura-2/acetoxymethyl (AM) for 40 min at 37° C (19, 20). Fluorescence was measured in individual cells at 340 and 380 nm excitation and 510 nm emission. For concentration- response analyses, fresh cells were studied with each agonist. For desensitization, cells were repetitively challenged with agonists without an intervening wash. A minimum of 10 cells were analyzed in three to four different experiments.

PAR2-induced Contraction of Isolated Human Bronchi

Bronchial rings (2 to 5 mm diameter) were mounted in organ baths (37° C, 95% O₂, 5% CO₂) to record isometric contractions (27). A tension of 2.5 g was applied and after 90 min tissue was challenged with acetylcholine (ACh) (1 mM), washed, and equilibrated for another 90 min. Cumulative concentration-response curves were generated to trypsin (10 nM to 0.3 µM), tryptase (1 to 30 nM), thrombin (10 nM to 1 µM), activating and reverse peptides (0.1 µM to 100 µM). The effects of indomethacin (5 µM, 45 min before the stimulus), N^G-nitro-L-monomethylarginine (L-NMMA) or inactive D-NMMA (100 µM, 15 min before the stimulus) were examined on a second cumulative concentration-response curve (90 min after the first). In some experiments the epithelium was removed with a swab, and effective removal was verified by histology.

Statistics

Results are expressed as mean ± SE. Differences between groups were examined by Student's t test.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

HASMC Express PAR2 mRNA

We examined the expression of PAR2 mRNA in primary cultures of HASMC by RT-PCR using primers to human PAR2. We used a KNRK cell line overexpressing human PAR2 as positive control. An agarose gel analysis of the PCR reactions showed amplification of a predicted 525 bp fragment of PAR2 mRNA in HASMC (Figure 1). This product migrated similarly to product amplified from KNRK-PAR2 cells, which served as a positive control. PAR2 mRNA was not detected in controls where reverse transcriptase was omitted. Thus, HASMC express PAR2 mRNA.

View larger version (62K): [in this window] [in a new window]   Figure 1.   Detection of PAR2 mRNA in HASMC. A PAR2 product of the predicted size of 525 bp was amplified from RNA from smooth muscle cells and a KNRK cell line expressing human PAR2 (positive control). No bands for PAR2 mRNA were detected in negative controls (without reverse transcriptase, not shown).

HASMC Express Immunoreactive PAR2

To confirm expression of PAR2 protein in HASMC, we used immunofluorescence and confocal microscopy. Immunoreactive PAR2 was detected in all cells (Figure 2A) where it was localized to the plasma membrane and in intracellular vesicles (Figure 2C). HASMC stained for smooth muscle actin, and thus retained their smooth muscle characteristics (Figure 2B). Staining was abolished when the PAR2 antibody was preabsorbed with the peptide that was used for immunization (Figure 2D), confirming specificity. In KNRK-PAR2 cells (positive control), immunoreactive PAR2 was detected at the plasma membrane and in intracellular vesicles (data not shown) (28). Thus, HASMC express immunoreactive PAR2.

View larger version (65K): [in this window] [in a new window]   Figure 2.   Localization of immunoreactive PAR2 in HASMC. Confocal imaging revealed the presence of PAR2 in smooth muscle cells (A, C ) where it was detected at the plasma membrane (arrowheads) and in intracellular locations (arrows). Preabsorption of the antibody with the peptide used for immunization abolished the signal (D), confirming specificity. Cells were stained for smooth muscle actin (B), confirming that they maintained their phenotype in culture. Scale bar = 10 µm.

Proteases Directly Signal to HASMC

To verify that HASMC express functional PAR2 and to identify potential agonists of PAR2 in this system, we measured Ca²⁺ mobilization in individual cells. SL-NH₂ (50 µM), which corresponds to the tethered ligand of human PAR2 that is exposed after proteolytic cleavage and that selectively activates PAR2, induced a prompt increase in [Ca²⁺]_i in all cells (Figure 3). Responses were maximal within approximately 20 s and declined to basal levels within 1 min. Trypsin, tryptase, and SL-NH₂ induced a dose-dependent increase in [Ca²⁺]_i. The half maximal effective concentration (EC₅₀) values were not determined because maximal responses were not measured (owing to inadequate amounts of certain agonists, notably tryptase). However, based on concentrations of agonists required to increase to 340/380 nm ratio by 0.1 unit, the rank order was trypsin (~ 1 nM to increase ratio by 0.1), which was approximately 10-fold more potent than tryptase (10 nM) and approximately 10,000-fold more potent than SL-NH₂ (~ 10,000 nM) (Figure 4).

View larger version (45K): [in this window] [in a new window]   Figure 3.   Effect of SL-NH2 on Ca2+ mobilization in HASMC. (A) Pseudocolor images of the cells. The inset shows the [Ca2+]i in nM. (B) Traces from individual cells shown in A. Each line represents a single cell.

View larger version (20K): [in this window] [in a new window]   Figure 4.   Effects of graded concentrations of trypsin, tryptase, and SL-NH2 on [Ca2+]i in HASMC. Results are expressed as change from basal and are mean ± SE of observations of > 10 cells.

HASMC also express PAR1, which can be activated by thrombin (29). We examined desensitization of Ca²⁺ mobilization to repeated stimulation with agonists to obtain evidence that PAR2 agonists were selective. Exposure of HASMC to 50 µM SL-NH₂ for 2 min strongly desensitized responses to a second challenge with 50 µM SL-NH₂ applied 2 min later (Figures 5A and 5B). Similarly, 10 nM trypsin desensitized responses to a second challenge 2 min later (Figure 5C). However, 50 µM SL-NH₂ or 10 nM trypsin did not desensitize responses to 10 nM thrombin (Figures 5A-5C). These results indicate that trypsin and SL-NH₂ stimulate Ca²⁺ mobilization by triggering PAR2 rather than a receptor for thrombin (PAR1, PAR3, or PAR4). Furthermore, desensitization to SL-NH₂ or trypsin is due to a receptor-specific event and not to depletion of intracellular Ca²⁺ pools, because robust responses to thrombin persisted.

View larger version (16K): [in this window] [in a new window]   Figure 5.   Desensitization of PAR2-induced mobilization in HASMC. Cells were exposed to the indicated agonists at intervals of 2 min, without an intervening wash. Panel A shows the trace from a single cell. Panels B and C show the mean ± SE of maximal responses, expressed as change from basal, of observations from > 10 cells. Panels A and B show that exposure to SL-NH2 desensitized subsequent responses to SL-NH2 but not to thrombin. Panel C shows that exposure to trypsin desensitized subsequent responses to trypsin but not to thrombin.

We examined the cross-desensitization of different PAR2 agonists to assess whether they activated the same receptor on HASMC. Exposure of HASMC to 30 µM SL-NH₂ for 2 min strongly desensitized responses to a second challenge with 1 nM trypsin applied 2 min later (75% desensitization compared with untreated cells). Similarly, 1 nM trypsin desensitized responses to 10 µM SL-NH₂ by 80% and abolished responses to 30 nM tryptase. Thus, SL-NH₂, trypsin, and tryptase probably activate the same receptor (PAR2) on HASMC to mobilize Ca²⁺.

Human Bronchiolar Smooth Muscle and Epithelium Express Immunoreactive PAR2

PAR2 expression in human central airways was confirmed by using immunohistochemistry. Immunoreactive PAR2 was detected in bronchial smooth muscle, which confirms the results obtained from isolated cells (Figure 6A). In addition, PAR2 was detected in the epithelium, bronchial glands, and in the endothelium and smooth muscle of bronchial vessels (Figures 6A and 6B). Preabsorption of PAR2 antibody with the nonconjugated immunogenic peptide completely abolished the staining, confirming specificity (not shown).

View larger version (146K): [in this window] [in a new window]   Figure 6.   Localization of immunoreactive PAR2 in human bronchus. (A) Immunoreactive PAR2 was detected in bronchial smooth muscle (sm) and epithelium (ep). (B) PAR2 was also detected in bronchial glands (bg) and blood vessels (bv). Original magnification: ×312.

PAR2 Agonists Induce Contraction of Human Bronchi

Our results indicate that HASMC express PAR2 mRNA and protein, and that PAR2 agonists such as trypsin and tryptase mobilize intracellular Ca²⁺ in cultured cells. To evaluate the functional consequences of PAR2 activation, we measured contractile responses of rings of human bronchi in an organ bath. ACh (1 mM) increased the tone (1.42 ± 0.23 mg, n = 18) in 92% of tissues, and only these preparations were included in the study. Trypsin (10 nM to 0.3 µM) and SL-NH₂ (0.1 to 100 µM) induced a small, but concentration-dependent contraction in 52% of the preparations tested. The maximal contraction to trypsin and SL-NH₂ was 12.7 ± 1.3% (n = 9) and 14.5 ± 1.8% of the response to ACh (1 mM), respectively (Figure 7A). Trypsin was more potent than SL-NH₂, although the EC₅₀ could not be determined, as responses did not become maximal to high concentrations of agonists. In nonprecontracted tissues or in tissues precontracted with ACh (1 mM), neither trypsin nor SL-NH₂ caused detectable relaxation (data not shown). In most of the precontracted bronchi, the highest concentrations of PAR2 agonists produced a moderate contraction superimposed on the ACh-induced contraction (data not shown). The reverse sequence of the activating peptide (100 µM), which does not activate the PAR2 receptor, was without effect.

View larger version (29K): [in this window] [in a new window]   Figure 7.   Contractile effect of isolated human bronchi to PAR agonists. (A) Trypsin and SL-NH2. (B) SL-NH2 in epithelium-denuded preparations. (C ) SL-NH2 in the presence of 5 µM indomethacin. (D) Thrombin and SFLLRN-NH2. Mean ± SE of at least five experiments. *p < 0.05.

Tryptase (1 nM) did not produce any contractile response. Higher concentrations of tryptase (3 nM and 10 nM) caused a 7 ± 2% (n = 4) and 14 ± 5% (n = 4) increase in bronchial tone, respectively. Tryptase (30 nM, n = 2) increased the tone by 11% and 21% of the response to ACh.

Contractile responses to maximally effective concentrations of SL-NH₂ (100 µM) were significantly increased after removal of the epithelium (Figure 7B). Pretreatment with indomethacin (5 µM) induced a moderate, although significant reduction of the contractile effect induced by SL-NH₂ (Figure 7C). In contrast, pretreatment with L-NMMA (100 µM) (data not shown) did not affect the motor response to SL-NH₂.

Thrombin (10 nM to 0.3 µM) and the human PAR1 activating peptide (0.1 to 100 µM) caused a concentration-dependent contraction in 58% and 63% of the isolated human bronchi tested, respectively (Figure 7D). The maximal effects were 20.0 ± 2.2% (n = 9) and 17.4 ± 1.5% (n = 5) of the response to ACh (1 mM), respectively. PAR4 activating peptide did not cause any measurable contractile or relaxant response in isolated human bronchial rings (data not shown).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Our results show that HASMC express PAR2. Agonists of PAR2, including trypsin, tryptase, and peptides corresponding to the tethered ligand domain that is unmasked by proteolysis can directly signal to these cells in culture by triggering PAR2. PAR2 agonists cause contraction of intact human bronchi. The contractile effects of PAR2 agonists are dampened by release of a bronchodilator factor from the epithelium, which also expresses PAR2, and are enhanced by the release of contractile products of cyclooxygenase. Together these results support the hypothesis that tryptase and PAR2 induce constriction of the human airway.

PAR2 Is Expressed by Human Primary Airway Smooth Muscle Cells

Several observations indicate that PAR2 is expressed by human airway smooth muscle. First, immunoreactive PAR2 was specifically localized to primary cultures of HASMC and to bronchiolar muscle cells in tissue sections using an antiserum to PAR2. Second, PAR2 was detected in smooth muscle in primary culture using RT-PCR. Third, trypsin, tryptase, and SL-NH₂ stimulated Ca²⁺ mobilization in HASMC, confirming expression of functional PAR2. PAR2 couples to generation of inositol(1,4,5)tris-phosphate and Ca²⁺ mobilization in transfected cell lines and in cells that naturally express PAR2 (19, 20, 24, 30), which supports our results. To our knowledge, this is the first observation that PAR2 is present and is functional in human bronchial smooth muscle cells in culture.

An important question in our experiments is whether the responses of cultured airway smooth muscle cells to tryptase, trypsin, and SL-NH₂ are mediated by PAR2 or are caused by the activation of a different receptor. The observation that the potency of tryptase to activate Ca²⁺ mobilization in HASMC was high and similar to that observed in other cell lines suggests that PAR2 could be involved in the response (19, 20, 24). The effect of trypsin on [Ca²⁺]_i in HASMC was reproduced by a peptide corresponding to the tethered ligand of PAR2. Moreover, after desensitization of PAR2 with trypsin or SL-NH₂, responses to thrombin were maintained. These findings exclude involvement of thrombin receptors (PAR1, PAR3, and PAR4) in responses to PAR2 agonists. Finally, the findings that SL-NH₂ desensitized responses to trypsin, and that trypsin also desensitized responses to SL-NH₂ and tryptase suggest that these agonists activate the same receptor, PAR2. Studies with PAR2 knockout animals or with PAR2 agonists, which are not currently available, are required to confirm this suggestion.

PAR2 Agonists Induce Contraction of Human Bronchiolar Smooth Muscle

PAR2 agonists caused a moderate contractile response in the majority of the human isolated bronchi studied. As reported previously (29), we observed similar contractile responses with PAR1 agonists. Several observations support the view that the contraction produced by SL-NH₂, trypsin, and tryptase is mediated by PAR2. First, the peptide with the reverse sequence of SL-NH₂ did not affect the motor function of human bronchi. Second, the PAR4 activating peptide was also without effect. Thus, although trypsin may activate PAR4 (31), a role of this receptor on the motor effect evoked by PAR2 agonists seems unlikely. In addition, SL-NH₂, in contrast with the PAR1 activating peptide, has been reported to be rather selective for PAR2 (32). Because of the limitation in protease availability, tryptase was studied in a limited number of experiments. However, the threshold concentration of tryptase to cause a measurable contraction was lower than that of trypsin, suggesting a potent effect of tryptase in the airway.

The findings that PAR2 agonists induced cross desensitization in HASMC suggest that they also trigger the same receptor in intact tissues. However, experiments with PAR2 antagonists are required to unequivocally confirm that the agonists act through a common mechanism in human airway tissues. Whereas trypsin was approximately 10,000-fold more potent than SL-NH₂ at mobilizing Ca²⁺ in HASMC, it was only approximately 100-fold more potent than SL-NH₂ in stimulating motility of tissues. Although this different pharmacology could suggest activation of different receptors, there are many other potential explanations. These include differences in susceptibility of SL-NH₂ to degradation or trypsin to inhibition in different preparations, and differential accessibility of peptide and protease agonists to the PAR2-expressing cells.

Several mechanisms contribute to the motor effects of PAR2 agonists in human bronchi. Most of the contractile response is probably mediated by a direct activation of a smooth muscle receptor. The immunohistochemical localization of PAR2 in smooth muscle cells either isolated in culture or in tissue sections supports this hypothesis, which is also favored by the ability of PAR2 agonists to stimulate Ca²⁺ mobilization in HASMC. However, indirect mechanisms may also contribute to PAR2-mediated contraction, because removal of the epithelium and indomethacin increased and reduced the response, respectively. Thus, it is possible that activation of an epithelial PAR2 releases a factor that limits the contraction produced by direct stimulation of PAR2 in the smooth muscle. Experiments in the presence of the nitric oxide (NO) synthase inhibitor, L-NMMA, indicate that this factor is not NO. In contrast, experiments with indomethacin suggest that prostanoid release, after PAR2 activation, contributes to the bronchial contraction. It remains to be determined whether PAR2 agonists induce release of prostanoids from these tissues. Alternative explanations for the increased potency of PAR2 agonists after removal of the epithelium include enhanced penetration of the tissues without the mucosa, and removal of epithelial peptidases that may degrade the agonist peptide.

Conclusions

A previous study showed that in the mouse trachea in vitro and in rats in vivo PAR2 agonists cause remarkable relaxation (17), and that PAR2 agonists can also relax isolated human bronchi. In contrast, in guinea-pig bronchi in vitro, activation of PAR2 results in either a contraction or a relaxation, depending on the tissue, whereas in vivo PAR2 activation causes a robust bronchoconstriction (18). There is also recent evidence that trypsin causes a contraction of isolated guinea-pig bronchi (33). However, this effect of trypsin, which was mediated by tachykinin release from bronchial sensory nerve endings, was not mimicked by the PAR2 activating peptide (33). We and others have shown that in isolated human bronchi activation of both PAR1 (29) and PAR2 causes a moderate contractile response that, at least in part, is mediated by a direct activation of smooth muscle cell receptors. Thus, motor responses to PAR2 agonists, as those described for agonists of tachykinin or bradykinin receptors, are species-dependent and show marked variations according to the experimental condition (in vitro or in vivo) (34).

Tryptase is the major protease released from human mast cells during inflammation and anaphylaxis (37). Some of the inflammatory effects of tryptase are mediated by degradation of extracellular proteins and peptides (38). In addition, tryptase is now recognized as a signaling molecule that activates PAR2 (19, 20, 39). Tryptase inhibitors have been successfully tested in experimental models of asthma and in asthmatic patients (15, 16). The present findings suggest that part of the beneficial effect of tryptase inhibitors may be due to the ability of these inhibitors to limit the signaling properties of the protease. However, the physiologic and pathophysiologic contribution of the interaction between tryptase and protease receptors in the airways, and, in particular, the role of the moderate contraction resulting from PAR2 activation, remains to be determined.