It is demonstrated that human serine proteinases including thrombin, tryptase, elastase and trypsin are potent stimuli of IL-8 secretion from A549 cells, which suggests that they are likely to play a role in IL-8 related airway inflammatory disorders such as asthma, chronic obstructive pulmonary disease and cystic fibrosis.
As little as 5.6 nM of thrombin is able to stimulate approximately 2 fold increase in IL-8 secretion, and 56 nM of thrombin induces 8 fold increase in IL-8 release, indicating that this proteinase is a potent secretagogue of IL-8 release from A549 cells. Human mast cell tryptase, an established mediator of inflammation  at a concentration as low as 3.7 nM induces twice more IL-8 secretion over baseline release, and trypsin, a potential mediator of airway inflammation  at a concentration of 0.042 nM provokes approximately 3 fold increase in IL-8 secretion from A549 cells, suggesting that tryptic enzyme in airways may play a role in stimulation of IL-8 hypersecretion from airway epithelium. Similarly, elastase, a well-established mediator of airway inflammation at a concentration of 10.2 nM elicits 4.4-fold increase in IL-8 release, indicating that it is a potent secretagogue of IL-8 release from A549 cells as well. At a concentration of 345 nM, elastase was also found to be able to induce IL-8 and MCP-1 secretion from human gingival fibroblasts . However, at the concentrations higher than 62.5 nM elastase could disarm PAR-2 within 10 min following incubation with human lung epithelial cells . These findings suggest that elastase at lower concentrations induce cytokine release from A549 cells, but at higher concentrations may inactivate PAR-2 on human lung epithelial cells including A549 cells. It was impossible for us to examine the effect of elastase at the concentrations higher than 62.5 nM with our experimental system as at the concentration of 0.6 μg/ml (20.4 nM), elastase was able to dissociate A549 cells from plate after 8 h incubation and the suspended cells died soon after (assessed by trypan blue staining). This phenomenon may explain the reason for which elastase at the concentration of 0.6 μg/ml fails to enhance IL-8 release. The similar phenomenon is also observed with trypsin at the concentrations higher than 1 μg/ml. These findings implicate that the detachment of bronchial epithelium observed in chronic airway inflammation may result from the hydrolytic activities of elastase and trypsin. Time course study shows that IL-8 release induced by thrombin, tryptase and trypsin initiates within 2 h following incubation, whereas the response to elastase occurs only after 8 h incubation period. This indicates that elastase and the other proteinases tested may adopt different mechanisms in induction of IL-8 release from A549 cells. The concentrations of tryptase and elastase used in the present study should be achievable under pathological conditions as the level of tryptase in asthmatic bronchial alveolar lavage fluid was 13.2 ng/ml  and the levels of elastase in asthmatic and cystic fibrosis sputum were 27 and 466 ng/ml, respectively . While information on the levels of thrombin and trypsin in respiratory fluids are not available, a report described that trypsin-like activity was 46.9 mU/ml in mucoid sputum from patients with asthma  might implicate that the concentrations of thrombin and trypsin used in the current study ought to be achievable in the inflammatory airways.
A549 cells have been reported to secrete some 10 pg/ml of elafin and 3 ng/ml of secretory leukocyte protease inhibitor (SLPI) following 24 h incubation . This concentration of elafin, an inhibitor of elastase should not affect the action of elastase on A549 cells, but the concentrations of SLPI, an inhibitor of trypsin and elastase may reduce the stimulatory action of the lower concentrations of trypsin or elastase on A549 cells.
Hirudin inhibites approximately 87% thrombin induced IL-8 secretion; SBTI and α1-antitrypsin completely abolish trypsin induced IL-8 secretion and MSACK eliminats 97% elastase induced IL-8 secretion, suggesting strongly that the actions of these proteinases on A549 cells are dependent upon their intact catalytic sites. Since the known substrates of these proteinases on cells are PARs, the expression of PARs on A549 cells was investigated in the present study. To our surprise, benzamidine and leupeptine at a concentration of 30 μg/ml (a quite high concentration for the study on cells based on our previous work  are only able to inhibit tryptase induced IL-8 secretion by 47.5% and 6.5%, which suggests that IL-8 secretion induced by tryptase may not depend on its enzymatic activity, and there may be a receptor other than PARs being involved in the process. The similar findings on tryptase were observed previously in other studies . However, to our knowledge, this is the first work examining the effects of thrombin, trypsin, tryptase and elastase on IL-8 release from airway epithelial cells under the same conditions.
It has been reported that a number of human cell types express more than one member of PAR family. Thus, platelets express PAR-1 and PAR-4 genes [14, 26], endothelial cells express PAR-1, PAR-2, and possibly PAR-3 [13, 27], fibroblast express PAR-1, PAR-2, PAR-3 and PAR-4 genes , smooth muscle cell express PAR-1, PAR-2, and PAR-3 genes  and respiratory epithelial cells express PAR-1, PAR-2, PAR-3 and PAR-4 genes and possibly proteins . In the present study, we find that A549 cells express all four PARs at both protein and mRNA levels. Since expression of the PARs was observed under both permeabilized and non-permeabilized conditions, it is most likely that all these four PARs are located in both the cytoplasma and the plasma membrane surface of the cells.
The PAR family consists of a group of four G protein-coupled receptors including PAR-1, PAR-2, PAR-3 and PAR-4. They share a unique mechanism of activation involving the proteolytic cleavage of the receptor by serine proteinases to unmask a new N-terminal sequence, the so-called "tethered ligand" that autoactivates the receptor . Thus, thrombin and trypsin cleavage PAR-1 within the amino-tail (LDPR41↓S42FLLRN, where↓denotes cleavage), trypsin, tryptase and elastase cleavages PAR-2 at the site of SKGR34↓S35LIGKV, thrombin also cleavages PAR-3 at the site LPIK38↓T39FRGAP and PAR-4 at PAPR47↓G48YPGQV . Synthetic receptor-activating peptides corresponding to the new amino termini of the cleaved receptors can also activate PARs . Thus, SFLLR-NH2 , SLIGKV-NH2  and tc-LIGRLO-NH2 , TFRGAP-NH2 , GYPGQV-NH2  represented agonists of PAR-1, PAR-2, PAR-3 and PAR-4, respectively in the current study. For comparison, their reverse peptides were used as controls.
SFLLR-NH2, tc-LIGRLO-NH2, SLIGKV-NH2 and GYPGQV-NH2 stimulates approximately 15.6, 79, 6.6, and 3.5 fold increase in release of IL-8, implicating that there are appropriate mechanisms to carry out IL-8 release process in response to PAR-1, PAR-2 and PAR-4 activation in A549 cells. However, A549 cells do not show any response (in terms of IL-8 release) to PAR-3 activation. Activation of A549 cells to release IL-8 by agonists of PARs indicates that the actions of thrombin, tryptase, elastase and trypsin on A549 cells are most likely carried out through hydrolytic cleavage of N-termini of PARs. The time course shows that the influence of agonists of PAR-1 and PAR-2 on A549 cells initiates within 2 h following incubation, but the action of agonist of PAR-4 on cells appears only after 8 h incubation. These observations suggest that the actions of thrombin on A549 cells are mainly (if not all) carried out through PAR-1, but not PAR-4, whereas the influence of trypsin on cells is most likely through both PAR-1 and PAR-2. It is hard to understand the slower response of cells to elastase and at least partially enzymatic activity independent actions of tryptase on A549 cells. These obviously require further investigation. Using various concentrations of agonist peptides of PARs to stimulate A549 cells may better reflect the actions of these peptides on the cells, which reinforces the previous finding .
Up-regulation of expression of IL-8 gene in A549 cells by thrombin, trypsin, tryptase, elastase, PAR-1 and PAR-2 agonist peptides indicates that IL-8 released from A549 cells induced by these stimuli is most likely being newly generated, rather than pre-stored in the cells. The observation that relatively small quantity of IL-8 was released during the first 2 h of incubation in response to the above stimuli also supports our view. While the influence of tryptase and trypsin on IL-8 gene expression does not appear to have been studied previously, the report which found elastase [36, 37] and thrombin  up-regulated IL-8 gene expression in human epithelial cells may support our current findings. To our knowledge, this is the first work examining IL-8 gene expression in response to several serine proteinases in epithelial cells under the same conditions. The parallel investigation of the actions of serine proteinases on A549 cells may contribute to easier understanding of the role of these proteinases in regulation of IL-8 gene expression. It is difficult to understand the reason why GYPGQV-NH2 does not significantly up-regulate IL-8 gene expression, but stimulates IL-8 release from A549 cells at 16 h following incubation. It could be that the significantly increased IL-8 gene expression occurs between 8 and 16 h incubation period, but we did not examine it.
Induction of inflammatory mediator release from airway epithelial cells by agonists of PARs has been demonstrated previously. Thus, agonist of PAR-1 stimulated platelet-derived growth factor secretion from lung epithelial cells ; agonists of PAR-2 stimulated IL-8 secretion from 16 HBE cells , GM-CSF and eotaxin release from human pulmonary epithelial cells  and matrix metalloproteinase-9 release from A549 and primary cultured small airway epithelial cells , and agonist of PAR-4 stimulated IL-8 secretion from human respiratory epithelial cells .
Our findings further strengthen the view that through activation of PARs, serine proteinases are actively involved in the pathogenesis of airway inflammation.
However, since A549 cells is not a normal airway epithelial cells, it may not fully represent the events happening in normal airway epithelial cells in response to the stimuli above in real life.