Role of Mast Cell and Neutrophil Proteases in Airway Secretion 1 - 3 JAY A. NADEL
Introduction Cystic fibrosis and chronic bronchitis are chronic diseases associated with mucus hypersecretion and inflammatory cell infiltration. Neutrophils and mast cells are believedto play important roles in the pathogenesis of airway inflammation. Wehypothesized that proteases released from these cells during exocytosis may stimulate secretion in submucosal glands and could explain, at least in part, the hypersecretion that exists in these pathologic conditions. Using a line of cultured airway gland cells, we studied the secretion of 35S-labeled macromolecules. The results of these studies indicate that mast cell chymase, neutrophil elastase, and cathepsin G are the most potent secretagogues ever described. Our studies with chymase suggest that the secretion occurs by a novel mechanism. Methods The methods are presented in detail in the original publications and will be described here only briefly.
Culture of Bovine Tracheal Gland Serous Cells The bovine tracheal gland serous cells were cultured as reported previously (1).These cells maintained the characteristics of differentiated serous cells, and they incorporated radiolabeled precursors into macromolecules and secreted them in response to various mediators. Cells were seeded onto tissue culture plastic coated with human placental collagen (15 ug/cm-) at a density of 2 x 104 cells/ em- in medium containing 40070 Dulbecco's modified Eagle's H21 medium, 40070 Ham's F12 medium, 20070 fetal calf serum, and 50 ug/ml gentamicin. Flasks were maintained at 37° C in 5070 CO 2/95070 air. The medium was changed every 3 days. Release of "Sslabeled Macromolecules On Day 9 of culture, confluent monolayers of serous cells cultured in human placental collagen-coated tissue culture flasks (surface area, 75 em") were incubated with 10 ml of medium containing 20 u.Ci/ml Na/ 5S0 4 • After 24 h, the medium containing the radiolabel was removed, and the cells were washed three times with Dulbecco's phosphate-buffered saline. Antibiotic- and serum-free medium was then added (10 ml/flask) and was changed every 30 min. At 210 min, the medium was collected and replaced with either fresh medium alone (baseline control) or medium containing the secretagogues; antagonists were added concurrently. At 240 min, the medium was collected again. The harvested medium from the 180-to 210-min and the 210-to 240min incubation periods was dialyzed (Spec548
SUMMARY To investigate the hypothesis that mast cell and neutrophil proteases stimulate airway gland secretion, we studied the effects of two mast cell proteases (tryptase and chymase) and two neutrophil enzymes (human neutrophil elastase and cathepsin G) on secretion of 35S-labeled macromolecules from cultured bovine airway gland serous cells. Tryptase had no effect, but the other three enzymes stimulated secretion. Threshold concentrations of the enzymes (~ 10-1 0 M) were lower by two orders of magnitude than other agonists (e.g., histamine, prostaglandins, ~-adrenergic agonists). Only proteases induced maximal secretory response (~ 800/0depletion of 35S-labeled macromolecules), and these responses were> 10-fold larger than those of other agonists. The active catalytic sites of the enzymes are required for their secretory activities. These findings suggest a role for these enzymes in the pathogenesis of inflammatory airway diseases associated with hypersecretion, and they suggest that the use of selective site-directed inhibitors of these enzymes may provide a novel strategy for intervention in inflammatory diseases of the airways associated with AM REV RE8PIR DI8 1991; 144:848-851 hypersecretion (e.g., cystic fibrosis, chronic bronchitis).
trapor tubing molecular mass cutoff, 12,000 to 14,000 D) against distilled water containing 10 mg/L sodium azide to remove unincorporated 35S04 -2. Nondialyzable 35S-labeled macroITlolecules were counted after addition of scintillation fluid (Hydrofluor; National Diagnostics, Inc., Somerville, NJ) by scintillation spectroscopy to an accuracy of 2070 (beta counter model LS7500; Beckman Instruments, Palo Alto, CA). Secretion is expressed as a percentage increase of release of 35S_ labeled macromolecules during incubation with the agonists over the release during the immediately preceding time period (secretory index) and is corrected for the declining baseline (determined in controls incubated with medium alone).
Purification of Mastocytoma Cell Proteases Tryptase and chymase were purified from tumors of "BR" mastocytoma cells (2, 3). Results Mast Cell Proteases Under basal conditions, cultured serous cells released nondialyzable 35S-la,beled macromolecules spontaneously and continuously. Purified tryptase had no significant effect on the release of 35S-labeled macromolecules, but purified chyrnase caused a profound response (p < 0.001, n = 5) (figure 1), with a threshold of ~ 10-9 M. At 10-8 M, the secretory index averaged 1,530 ± 80070 (mean ± SE); a maximal response was not reached in the concentrations tested (10-11 to 10-8 M). In comparison, another component of mast cells, histamine, stimulated tlie secretion of 35S-labeled macromolecules, with a threshold of ~ 10-8 M (p < 0.001, n = 6) (figure 2). With increasing concentrations of histamine, a plateau response was reached of 177 ± 90070 at 10-6 M. Thus, the maximal response to histamine was more than an order of magnitude less than the response to chymase. The response to
chymase was noncytotoxic: trypan blue dye exclusion and lactic dehydrogenase release were unaffected. Inhibitors of the active site of chymase, soybean trypsin inhibitor (100ug/ ml), and chymostatin (10 ug/ml) reduced or abolished the secretory response to 10-8 M chymase. The secretory response induced by chymase was completely prevented by the metabolic inhibitors azide (100 mM), dicoumarol (1mM), and 2,4-dinitrophenol (1 mM) (n = 5 each, p < 0.01), indicating that the secretory response required an unimpaired cellular energy metabolism. To compare the effect of mediator-rich supernatant from mast cells with the effect of highly purified chymase, we challenged cultured serous cells with degranulation supernatant from dog mastocytoma cells. Mast cell supernatant from 3 x 104 mastocytoma cells stimulated the release of 35S-labeled macromolecules (secretory index, 145 ± 10070; p < 0.05). This concentration of supernatant produced a final concentration of histamine of 2 x 10-8 M in the incubation medium, an amount of histamine that had no significant effect on secretion. When supernatant from 107 mastocytoma cells was added per milliliter incubation medium (final concentration of histamine, 6 x 10-6 M), the secretory response reached 1,570 ± 80070 (n = 6). The secretory response to mastocytoma supernatant was unaffected by the combined cyclooxygenaseand lipoxygenase inhibitor BW755C (10-4 ) , but soybean trypsin inhibitor (an inhibitor of
1 From the Cardiovascular Research Institute and the Departments of Medicine and Physiology, University of California, San Francisco, San Francisco, California. 2 Supported in part by a Research Development Program Grant from the Cystic Fibrosis Foundation. 3 Correspondence and requests for reprints should be addressed to Jay A. Nadel, M.D., Cardiovascular Research Institute, Box 0130, University of California, San Francisco, CA 94143-0130.
549
ROLE OF MAST CELL AND NEUTROPHIL PROTEASES IN AIRWAY SECRETION
2000 Fig. 1. Effect of chymase (opensquares) and tryptase (closed squares) on secretion of 35S-labeled macromolecules from cultured tracheal gland serous cells. Secretion is greatly stimulated by chymase (p < 0.0001, n = 5), whereas tryptase has no significant effect. Values are mean ± SEM; n = 5. (Reproduced with permission from reference 7.)
Secretion
[0/0 above baseline] 1000
o
QI
o
10-11
10-10
10-9
Agonist [M]
Fig. 2. Effect of degranulation supernatant from mastocytoma cells (open squares) and of exogenous histamine (closed squares) on secretion of 35S-labeled macromolecules from cultured tracheal gland serous cells. The position of the concentration-response curves is adjusted according to the final histamine concentration in the incubation medium. Mastocytoma cell supernatant stimulates secretion much more than equivalent concentrations of histamine (p < 0.0001),suggesting the presence of other secretagogues in mastocytoma supernatant. Values are mean ± SEM; n = 6. (Reproduced with permission from reference 7.)
2000 Secretion
[0/0 above baseline] 1000
0
~
10-8
0
10-7
10-5
10-6
Histamine [M] i
i
10 4
105
i
i.
10
6
10
7
Mast Cells [per ml]
2000 Fig. 3. Effects of human neutrophil cathepsin G (open squares) and elastase (closed squares) and of histamine (closed triangles) on secretion of 35S_ labeled macromolecules from cultured tracheal gland serous cells. Secretion is stimulated markedly by cathepsin G and elastase (p < 0.0001,n = 5), whereas histamine has a comparatively small effect (p < 0.0001, n = 6). Values are mean ± SEM. (Reproduced with permission from reference 31.)
1500
1000
500
o , o
1 0 -11 10 -10 10 -9 10 -8
10 -7
10 -6
10 -5
Agonist [M]
Cathepsin G Cathepsin G + Chymostatin 1Ouq/rn' Cathepsin G + Chymostatin 50~g/ml Fig. 4. Effect of active site inhibitors on secretion induced by 10-8 M human neutrophil cathepsin G (top panel) and elastase (bottom panel). Soybean trypsin inhibitor, 100 Ilg/ml; PMSF, 10-4 M; N-methoxysuccinyl-Ala-Ala-Pro-Val-chloromethylketone (AAPV-CK), 10-5 M; u 1-proteinase inhibitor, 100 Ilg/ml. Values are mean ± SEM; n = 5. (Reproduced with permission from reference 31.)
Cathepsin G + Soybean Trypsin Inhibitor I----..,....----r---~----,
250
500
750
1000
Secretion [% above Baseline]
Elastase Elastase + PMSF Elastase + AAPV-CK Elastase + u1 -Proteinase Inhibitor 0
500
1000
1500
Secretion [0/0 above Baseline]
2000
chymase, 100 ug/ml) reduced the secretory response to 33 ± 180/0 (n = 5). Addition of the Hj-receptor antagonist metiamide (10-4 M) further reduced the response, so it was no longer significant (42 ± 50/0; n = 6). A second chymase inhibitor, chymostatin (10 ug/ml), also significantly reduced the secretory response (390 ± 140/0; n = 5), whereas the tryptase inhibitor aprotinin (30 J.1g/ ml) did not affect the respone (1,300 ± 500/0; n = 6). These findings show that chymase is the main secretagogue produced by mast cells and that the effect of released histamine is an order of magnitude smaller. 'To study the mechanism of stimulationsecretion coupling, we investigated the role of cyclic AMP (cAMP), inositol phosphates (IP), protein kinase C (PKC), and Ca 2 + in stimulation-secretion coupling in the cultured gland cells (4). Activation of the cells with histamine (10-5 M) increased cAMP 17-fold, whereas chymase (10-9 M) and bradykinin (10-6 M) had no effect; the phosphodiesterase inhibitor 3-isobutyl,1-methylxanthine (10-5 M) augmented the secretory response to histamine but not to chymase. Incorporation of 3H-myoinositol into IP increased 5-fold in response to bradykinin but not to histamine or chymase. Depletion of PKC from cells by preincubation with phorbol12-myristate 13acetate (10-6 M) reduced the secretoryresponses to histamine and to bradykinin but not to chymase. Bradykinin and histamine caused a transient rise in cytoplasmic Ca 2 + (Fura-2 fluorescence), but chymase had no effect. Furthermore, the intracellular Ca 2 + antagonist TMB-8 (2 x 10-4 M) decreased the secretory response to histamine but not to chymase.
Neutrophil Proteases Both human neutrophil elastase and cathepsin G stimulated secretion of 35S-labeled macromolecules in a concentration-dependent manner (figure 3). The threshold concentration of both elastase and cathepsin G was ~ 10-10 M, and the maximal secretoryresponse of both proteases was similar. However, elastase was more potent than cathepsin G (p < 0.001),with responses at 10-8 M of 1,810 ± 600/0 and 970 ± 200/0 above baseline, respectively (n = 5). By comparison, histamine caused secretion, with a threshold of 10-6 M and a maximal response of only 185 ± 80/0 at 10-5 M (n = 6). The responses to neutrophil enzymes were not associated with cytotoxicity, as studied morphologically, by vital dye exclusion or by lactic dehydrogenase release. Inhibitors of cathepsin G and neutrophil elastase, respectively, inhibited the secretagogue responses to the two enzymes (figure 4). The inhibitory effects of the protease inhibitors were similar to their effects on the amidolytic activity of the proteases. To study whether the effects of purified neutrophil proteases could be reproduced by products released directly from neutrophils, we purified human neutrophils from peripheral blood of healthy subjects (~ 950/0 neutrophils), suspended them in complete Hanks'
550
balanced salt solution (HBSS) (0.6 x 108 cells/ml), activated them with 5 ug/rnl cytochalasin Band 10-6 M FMLP for 10min, and prepared cell-free supernatant by centrifugation. When degranulation supernatant from 105 neutrophils was added per milliliter of incubation medium of serous cells, the secretory response averaged 2,095 ± 54070 above baseline (p < 0.001, n = 6). This response was markedly diminished by preincubation with the human neutrophil elastase inhibitor N-methoxysuccinyl-Ala-Ala-Pro-Val-chloromethylketone (10-5 M). The remaining response (363 ± 22070) was further decreased by the cathepsin G inhibitor chymostatin (50 ug/ml) to 102 ± 19070. Concentrations of cytochalasin Band FMLP equal to those present in supernatant from activated neutrophils had no effect on baseline secretion (0 ± 8070; n = 4). These results indicate that neutrophil elastase and, to a lesser extent, cathepsin G released directly from neutrophils cause gland secretion.
Discussion The present findings indicate that mast cell chymase and neutrophil proteases are remarkably potent secretagogues 0 f cultured airway submucosal glands. In contrast to other agonists such as histamine, prostaglandins, and ~-adrenergic agonists, in which concentrations in excess of 10-8 M are required to induce secretory responses (1,5,6), the threshold concentration for proteases (~ 10-10 M) is lower by two orders of magnitude. In addition, only proteases induce maximal secretory responses (~ 80070 depletion of 35S-labeled macromolecules), and these responses to proteases are> 10-fold larger than those of other agonists (7). These are the most potent secretagogues ever described, and they are derived from two types of inflammatory cells, neutrophils and mast cells. Neutrophils are believed to play an important role in the pathogenesis of airway inflammation. Exposure to air pollutants (e.g., ozone and sulfur dioxide) causes an influx of neutrophils into the airway (8-11); this exposure may be associated with bronchitis and hypersecretion (10, 12).Airway neutrophils also occur in higher numbers in the airways and in bronchoalveolar lavage of cigarette smokers than in those of nonsmokers (13, 14),and free neutrophil elastase activity may be present in the airways of cigarette smokers and in patients with bronchitis (14-18). Increased numbers of neutrophils are found in patients with chronic bronchitis (19). Among conditions associated with increased neutrophils in airways, patients with cystic fibrosis have an immense neutrophil "load" in the airways, and this is associated with high concentrations of neutrophil elastase in the sputum (20, 21). Cystic fibrosis is associated with mucus hypersecretion and progressive airway obstruction, and the discovery of the potent secretagogue effect of neutrophil proteases suggests the possibility that selectiveprotease antagonists may playa role in the therapy of this disease. Neu-
JAY A. NADEL
trophils contain approximately 1 to 2 ug of elastase (22, 23) and 2 to 4 ug of cathepsin G (24, 25) per 106 cells. Thus, the release of 10070 of the proteases from 105 neutrophils/ em" of submucosal tissues may be sufficient to cause a significant secretory response. The number of neutrophils found in the submucosa of mongrel dogs is higher (0.5 x 10' /cm") and is considerably increased after ozone exposure (9). Thus, a secretory response is likely even if> 99070 of the proteases are inhibited by endogenous proteases. In diseases that are characterized by excessiveand abnormal airway secretions (e.g., cystic fibrosis, chronic bronchitis), sputum concentrations of uninhibited neutrophil proteases are increased markedly and greatly exceed the concentrations used in this study (20). These findings suggest a potential role of inhibitors of neutrophil proteases in the treatment of these diseases. Mast cells are also recognized to be important inflammatory cells and are implicated in various inflammatory lung diseases. They are abundant in the airway submucosa (26, 27), where they are in close contact with submucosal glands. Although mast cells containing chymase are rare in the peripheral tissue of normal lungs (28), the proportion of mast cells that contain chymase is much higher in the bronchial epithelium, where the glands are located (29). Furthermore, in certain pathologic conditions, the proportion of mast cells containing chymase may increase. For example, the proportion of chymase-containing mast cells is increased dramatically in bronchoalveolar lavagein patients with cystic fibrosis (30). Degranulation of these mast cells could lead to high local concentrations of chymase, causing activation and secretion from gland cells. The mechanism by which proteases induce gland cell degranulation remains to be determined. The active catalytic site of the proteases is required for degranulation. The high molecular weight of the proteases (e.g., M, of chymase = 30,000 D) makes it likely that their site of action is on the cell surface, although pinocytosis of the protease molecules is not ruled out. Whether the active site interacts with a cell-surface protein or an intracellular protein, studies of chymase action suggest a novel mode of action in relation to stimulation-secretion. Thus, in cultured bovine airway gland serous cells, in contrast to histamine and bradykinin, chymase did not increase cAMP, and the secretory response was unaffected by the phosphodiesterase inhibitor 3-isobutyl-1-methylxanthine (10-5 M); incorporation of 3H-myoinositol into inositol phosphate was unaffected, and depletion of protein kinase C by phorbol 12-myristate did not affect the secretory response; and secretion was not associated with a rise in cytoplasmic Ca 2 +. Thus none of the classic second messengers appears to be involved in chymase-induced secretion (4). In summary, mast cells and neutrophils are the most potent secretagogues of airway sub-
mucosal gland secretion thus far described. Chymase causes secretion by a novel mechanism involving the active site of the enzyme, and the other enzymes may act by a similar mechanism. It is proposed that the release of these enzymes near glands in inflammatory diseases associated with hypersecretion may play an important role in the pathogenesis of these diseases. Therefore, the use of selective site-directed inhibitors of these enzymes may provide an important strategy for therapeutic intervention. Acknowledgment The writer thanks Patty Snell for assistance in manuscript preparation.
References 1. Finkbeiner WE, Nadel JA, Basbaum CB. Establishment and characterization of a cell line derived from bovine tracheal glands. In Vitro Cell Dev BioI 1986; 22:561-7. 2. CaugheyGH, Viro NF, Ramachandran J, Lazarus SC, Borson DB, Nadel JA. Dog mastocytoma tryptase: affinity purification, characterization and amino-terminal sequence. Arch Biochem Biophys 1987; 258:555-63. 3. Caughey GH, Viro NF, Lazarus SC, Nadel JA. Purification and characterization of dog mastocytoma chymase. Identification of an octapeptide conserved in chymotryptic leukocyte proteases. Biochim Biophys Acta 1988; 952:142-9. 4. Sommerhoff CP, Caughey GH, Nadel JA. classical second messengers are not involved in the stimulation-secretion coupling of chymase-induced degranulation of airway gland serous cells.FASEB J 1990; 4:AI940. 5. Sommerhoff CP, Finkbeiner WE, Nadel JA, Basbaum CB. Mediators of anaphylaxis stimulate 35S-labeledmacromolecule secretion from cultured bovine tracheal gland serous cells. Bull Eur Physiopathol Respir 1987; 23:363s. 6. Sommerhoff CP, Finkbeiner WE, Nadel JA, Basbaum CB. Prostaglandin D 2 and prostaglandin E 2 stimulate 35S-labeled macromolecule secretion from cultured bovine tracheal gland serous cells (abstract). Am Rev Respir Dis 1987; 135:A363. 7. Sommerhoff CP, Caughey CH, Finkbeiner WE, Lazarus SC, Basbaum CB, Nadel JA. Mast cell chymase: a potent secretagogue for airway gland serous cells. J Immunol 1989; 142:2450-6. 8. Seltzer J, Bigby BG, Stulbarg M, et al. 0 3 induced change in bronchial reactivity to methacholine and airway inflammation in humans. J Appl Physiol 1986; 60:1321-6. 9. Holtzman MJ, Fabbri LM, O'Byrne PM, et al. Importance of airway inflammation for hyperresponsiveness induced by ozone in dogs. Am Rev Respir Dis 1983; 127:686-90. 10.. Seltzer J, Scanlon PD, Drazen JM, Ingram RH J r, Reid L. Morphologic correlation of physiologic changes caused by S02-induced bronchitis in dogs: the role of inflammation. Am Rev Respir Dis 1984; 129:790-7. 11. Shore SA, Kariya ST, Anderson K, et al. Sulfurdioxide-induced bronchitis in dogs: effects on airway responsiveness to inhaled and intravenously administered methacholine. Am Rev Respir Dis 1987; 135:840-7. 12. Phipps RJ, Denas SM, Sielczak MW, Wanner A. Effects of 0.5 ppm ozone on glycoprotein secretion, ion and water fluxes in sheep trachea. J Appl Physiol 1986; 60:918-27. 13. Reynolds HY, Merrill WW. Airway changes
ROLE OF MAST CELL AND NEUTROPHIL PROTEASES IN AIRWAY SECRETION
in young smokers that may antedate chronic obstructive lung disease. Med Clin North Am 1981; 65:667-89. 14. Stockley RA, Burnett D. Alpha l-antitrypsin and leukocyte elastase in infected and non-infected sputum. Am Rev Respir Dis 1979; 120:1081-6. 15. Ohlsson K, TegnerH. Granulocyte collagenase, elastase and plasma protease inhibitors in purulent sputum. Eur 1 Clin Invest 1975; 5:221-7. 16. Weitz 11,Crowley KA, Landman SL, Lipman BI, Yu 1. Increased neutrophil elastase activity in cigarette smokers. Ann Intern Med 1987;107:680-2. 17. lanoff A, Raju L, Dearing R. Levelsof elastase activity in bronchoalveolar lavage fluids of healthy smokers and nonsmokers. Am Rev Respir Dis 1983; 127:540-4. 18. Tetley TD, Smith SF, Burton GH, Winning Al, Cooke NT, Guz A. Effects of cigarette smoking and drugs on respiratory tract proteases and antiproteases. Eur Respir 1 [Suppl]1987; 153:93-102. 19. Gibson PG, Girgis-Gabarado A, Morris MM, et ale Cellular characteristics of sputum from patients with asthma and chronic bronchitis. Thorax 1989; 44:693-9. 20. Goldstein W, Doring G. Lysosomal enzymes
from polymorphonuclear leukocytesand proteinase inhibitors in patients with cystic fibrosis. Am Rev Respir Dis 1986; 134:49-56. 21. Suter S, Schaad UB, TegnerH, Ohlsson K, Desgrandchamps D, WaldvogelFA. Levelsof free granulocyte elastase in bronchial secretions from patients with cystic fibrosis: effect of antimicrobial treatment against Pseudomonas aeruginosa. 1 Infect Dis 1986; 153:902-9. 22. Baugh Rl, Travis 1. Human leukocyte granule elastase: rapid isolation and characterization. Biochemistry 1976; 15:836-41. 23. Heck LW, Darby WL, Hunter FA, Bhown A, Miller El, Bennett lC. Isolation, characterization, and amino-terminal amino acid sequence analysis of human neutrophil elastase from normal donors. Anal Biochem 1985; 149:153-62. 24. Heck LW, Rostand KS, Hunter FA, Bhown A. Isolation, characterization, and amino-terminal amino acid sequence analysis of human neutrophil cathepsin G from normal donors. Anal Biochem 1986; 158:217-27. 25. Senior RM, Campbell El. Cathepsin G in human mononuclear phagocytes:comparison between monocytes and U927 monocyte-like cells. 1 Immu-
551 nol 1984; 132:2547-51. 26. Guerzon GM, Pare PD, Michoud M-C, Hogg lC. The number and distribution of mast cells in monkey lungs. Am Rev Respir Dis 1979;119:59-66. 27. Shanahan F, MacNiven I, Dyck N, Denburg lA, Bienenstock 1, Befus AD. Human lung mast cells: distribution and abundance of histochemically distinct populations. Int Arch Allergy Appl Immunol 1987; 83:329-31. 28. Schwartz LB, Irani A-MA, Roller K, Castells MC, Schechter NM. Quantitation of histamine, tryptase and chymase in dispersed human T and TC mast cells. 1 Immunol 1987; 138:2611-5. 29. Irani AA, Schechter NM, Craig SS, DeBlois G, Schwartz LB. Two types of human mast cells that have distinct neutral protease compositions. Proc Natl Acad Sci USA 1986; 83:4464-8. 30. Schwartz LB, Irani AA, Mroueh S, Spock A. Reversed ratio of T/TC types of human mast cells in cysticfibrosis (abstract). Clin Res 1987; 35:255A. 31. Sommerhoff CP, Nadel lA, Basbaum CB, Caughey GH. Neutrophil elastase and cathepsin G stimulate secretion from cultured bovine airway gland serous cells. 1 Clin Invest 1990; 85:682-9.
Introduction Cystic fibrosis and chronic bronchitis are chronic diseases associated with mucus hypersecretion and inflammatory cell infiltration. Neutrophils and mast cells are believedto play important roles in the pathogenesis of airway inflammation. Wehypothesized that proteases released from these cells during exocytosis may stimulate secretion in submucosal glands and could explain, at least in part, the hypersecretion that exists in these pathologic conditions. Using a line of cultured airway gland cells, we studied the secretion of 35S-labeled macromolecules. The results of these studies indicate that mast cell chymase, neutrophil elastase, and cathepsin G are the most potent secretagogues ever described. Our studies with chymase suggest that the secretion occurs by a novel mechanism. Methods The methods are presented in detail in the original publications and will be described here only briefly.
Culture of Bovine Tracheal Gland Serous Cells The bovine tracheal gland serous cells were cultured as reported previously (1).These cells maintained the characteristics of differentiated serous cells, and they incorporated radiolabeled precursors into macromolecules and secreted them in response to various mediators. Cells were seeded onto tissue culture plastic coated with human placental collagen (15 ug/cm-) at a density of 2 x 104 cells/ em- in medium containing 40070 Dulbecco's modified Eagle's H21 medium, 40070 Ham's F12 medium, 20070 fetal calf serum, and 50 ug/ml gentamicin. Flasks were maintained at 37° C in 5070 CO 2/95070 air. The medium was changed every 3 days. Release of "Sslabeled Macromolecules On Day 9 of culture, confluent monolayers of serous cells cultured in human placental collagen-coated tissue culture flasks (surface area, 75 em") were incubated with 10 ml of medium containing 20 u.Ci/ml Na/ 5S0 4 • After 24 h, the medium containing the radiolabel was removed, and the cells were washed three times with Dulbecco's phosphate-buffered saline. Antibiotic- and serum-free medium was then added (10 ml/flask) and was changed every 30 min. At 210 min, the medium was collected and replaced with either fresh medium alone (baseline control) or medium containing the secretagogues; antagonists were added concurrently. At 240 min, the medium was collected again. The harvested medium from the 180-to 210-min and the 210-to 240min incubation periods was dialyzed (Spec548
SUMMARY To investigate the hypothesis that mast cell and neutrophil proteases stimulate airway gland secretion, we studied the effects of two mast cell proteases (tryptase and chymase) and two neutrophil enzymes (human neutrophil elastase and cathepsin G) on secretion of 35S-labeled macromolecules from cultured bovine airway gland serous cells. Tryptase had no effect, but the other three enzymes stimulated secretion. Threshold concentrations of the enzymes (~ 10-1 0 M) were lower by two orders of magnitude than other agonists (e.g., histamine, prostaglandins, ~-adrenergic agonists). Only proteases induced maximal secretory response (~ 800/0depletion of 35S-labeled macromolecules), and these responses were> 10-fold larger than those of other agonists. The active catalytic sites of the enzymes are required for their secretory activities. These findings suggest a role for these enzymes in the pathogenesis of inflammatory airway diseases associated with hypersecretion, and they suggest that the use of selective site-directed inhibitors of these enzymes may provide a novel strategy for intervention in inflammatory diseases of the airways associated with AM REV RE8PIR DI8 1991; 144:848-851 hypersecretion (e.g., cystic fibrosis, chronic bronchitis).
trapor tubing molecular mass cutoff, 12,000 to 14,000 D) against distilled water containing 10 mg/L sodium azide to remove unincorporated 35S04 -2. Nondialyzable 35S-labeled macroITlolecules were counted after addition of scintillation fluid (Hydrofluor; National Diagnostics, Inc., Somerville, NJ) by scintillation spectroscopy to an accuracy of 2070 (beta counter model LS7500; Beckman Instruments, Palo Alto, CA). Secretion is expressed as a percentage increase of release of 35S_ labeled macromolecules during incubation with the agonists over the release during the immediately preceding time period (secretory index) and is corrected for the declining baseline (determined in controls incubated with medium alone).
Purification of Mastocytoma Cell Proteases Tryptase and chymase were purified from tumors of "BR" mastocytoma cells (2, 3). Results Mast Cell Proteases Under basal conditions, cultured serous cells released nondialyzable 35S-la,beled macromolecules spontaneously and continuously. Purified tryptase had no significant effect on the release of 35S-labeled macromolecules, but purified chyrnase caused a profound response (p < 0.001, n = 5) (figure 1), with a threshold of ~ 10-9 M. At 10-8 M, the secretory index averaged 1,530 ± 80070 (mean ± SE); a maximal response was not reached in the concentrations tested (10-11 to 10-8 M). In comparison, another component of mast cells, histamine, stimulated tlie secretion of 35S-labeled macromolecules, with a threshold of ~ 10-8 M (p < 0.001, n = 6) (figure 2). With increasing concentrations of histamine, a plateau response was reached of 177 ± 90070 at 10-6 M. Thus, the maximal response to histamine was more than an order of magnitude less than the response to chymase. The response to
chymase was noncytotoxic: trypan blue dye exclusion and lactic dehydrogenase release were unaffected. Inhibitors of the active site of chymase, soybean trypsin inhibitor (100ug/ ml), and chymostatin (10 ug/ml) reduced or abolished the secretory response to 10-8 M chymase. The secretory response induced by chymase was completely prevented by the metabolic inhibitors azide (100 mM), dicoumarol (1mM), and 2,4-dinitrophenol (1 mM) (n = 5 each, p < 0.01), indicating that the secretory response required an unimpaired cellular energy metabolism. To compare the effect of mediator-rich supernatant from mast cells with the effect of highly purified chymase, we challenged cultured serous cells with degranulation supernatant from dog mastocytoma cells. Mast cell supernatant from 3 x 104 mastocytoma cells stimulated the release of 35S-labeled macromolecules (secretory index, 145 ± 10070; p < 0.05). This concentration of supernatant produced a final concentration of histamine of 2 x 10-8 M in the incubation medium, an amount of histamine that had no significant effect on secretion. When supernatant from 107 mastocytoma cells was added per milliliter incubation medium (final concentration of histamine, 6 x 10-6 M), the secretory response reached 1,570 ± 80070 (n = 6). The secretory response to mastocytoma supernatant was unaffected by the combined cyclooxygenaseand lipoxygenase inhibitor BW755C (10-4 ) , but soybean trypsin inhibitor (an inhibitor of
1 From the Cardiovascular Research Institute and the Departments of Medicine and Physiology, University of California, San Francisco, San Francisco, California. 2 Supported in part by a Research Development Program Grant from the Cystic Fibrosis Foundation. 3 Correspondence and requests for reprints should be addressed to Jay A. Nadel, M.D., Cardiovascular Research Institute, Box 0130, University of California, San Francisco, CA 94143-0130.
549
ROLE OF MAST CELL AND NEUTROPHIL PROTEASES IN AIRWAY SECRETION
2000 Fig. 1. Effect of chymase (opensquares) and tryptase (closed squares) on secretion of 35S-labeled macromolecules from cultured tracheal gland serous cells. Secretion is greatly stimulated by chymase (p < 0.0001, n = 5), whereas tryptase has no significant effect. Values are mean ± SEM; n = 5. (Reproduced with permission from reference 7.)
Secretion
[0/0 above baseline] 1000
o
QI
o
10-11
10-10
10-9
Agonist [M]
Fig. 2. Effect of degranulation supernatant from mastocytoma cells (open squares) and of exogenous histamine (closed squares) on secretion of 35S-labeled macromolecules from cultured tracheal gland serous cells. The position of the concentration-response curves is adjusted according to the final histamine concentration in the incubation medium. Mastocytoma cell supernatant stimulates secretion much more than equivalent concentrations of histamine (p < 0.0001),suggesting the presence of other secretagogues in mastocytoma supernatant. Values are mean ± SEM; n = 6. (Reproduced with permission from reference 7.)
2000 Secretion
[0/0 above baseline] 1000
0
~
10-8
0
10-7
10-5
10-6
Histamine [M] i
i
10 4
105
i
i.
10
6
10
7
Mast Cells [per ml]
2000 Fig. 3. Effects of human neutrophil cathepsin G (open squares) and elastase (closed squares) and of histamine (closed triangles) on secretion of 35S_ labeled macromolecules from cultured tracheal gland serous cells. Secretion is stimulated markedly by cathepsin G and elastase (p < 0.0001,n = 5), whereas histamine has a comparatively small effect (p < 0.0001, n = 6). Values are mean ± SEM. (Reproduced with permission from reference 31.)
1500
1000
500
o , o
1 0 -11 10 -10 10 -9 10 -8
10 -7
10 -6
10 -5
Agonist [M]
Cathepsin G Cathepsin G + Chymostatin 1Ouq/rn' Cathepsin G + Chymostatin 50~g/ml Fig. 4. Effect of active site inhibitors on secretion induced by 10-8 M human neutrophil cathepsin G (top panel) and elastase (bottom panel). Soybean trypsin inhibitor, 100 Ilg/ml; PMSF, 10-4 M; N-methoxysuccinyl-Ala-Ala-Pro-Val-chloromethylketone (AAPV-CK), 10-5 M; u 1-proteinase inhibitor, 100 Ilg/ml. Values are mean ± SEM; n = 5. (Reproduced with permission from reference 31.)
Cathepsin G + Soybean Trypsin Inhibitor I----..,....----r---~----,
250
500
750
1000
Secretion [% above Baseline]
Elastase Elastase + PMSF Elastase + AAPV-CK Elastase + u1 -Proteinase Inhibitor 0
500
1000
1500
Secretion [0/0 above Baseline]
2000
chymase, 100 ug/ml) reduced the secretory response to 33 ± 180/0 (n = 5). Addition of the Hj-receptor antagonist metiamide (10-4 M) further reduced the response, so it was no longer significant (42 ± 50/0; n = 6). A second chymase inhibitor, chymostatin (10 ug/ml), also significantly reduced the secretory response (390 ± 140/0; n = 5), whereas the tryptase inhibitor aprotinin (30 J.1g/ ml) did not affect the respone (1,300 ± 500/0; n = 6). These findings show that chymase is the main secretagogue produced by mast cells and that the effect of released histamine is an order of magnitude smaller. 'To study the mechanism of stimulationsecretion coupling, we investigated the role of cyclic AMP (cAMP), inositol phosphates (IP), protein kinase C (PKC), and Ca 2 + in stimulation-secretion coupling in the cultured gland cells (4). Activation of the cells with histamine (10-5 M) increased cAMP 17-fold, whereas chymase (10-9 M) and bradykinin (10-6 M) had no effect; the phosphodiesterase inhibitor 3-isobutyl,1-methylxanthine (10-5 M) augmented the secretory response to histamine but not to chymase. Incorporation of 3H-myoinositol into IP increased 5-fold in response to bradykinin but not to histamine or chymase. Depletion of PKC from cells by preincubation with phorbol12-myristate 13acetate (10-6 M) reduced the secretoryresponses to histamine and to bradykinin but not to chymase. Bradykinin and histamine caused a transient rise in cytoplasmic Ca 2 + (Fura-2 fluorescence), but chymase had no effect. Furthermore, the intracellular Ca 2 + antagonist TMB-8 (2 x 10-4 M) decreased the secretory response to histamine but not to chymase.
Neutrophil Proteases Both human neutrophil elastase and cathepsin G stimulated secretion of 35S-labeled macromolecules in a concentration-dependent manner (figure 3). The threshold concentration of both elastase and cathepsin G was ~ 10-10 M, and the maximal secretoryresponse of both proteases was similar. However, elastase was more potent than cathepsin G (p < 0.001),with responses at 10-8 M of 1,810 ± 600/0 and 970 ± 200/0 above baseline, respectively (n = 5). By comparison, histamine caused secretion, with a threshold of 10-6 M and a maximal response of only 185 ± 80/0 at 10-5 M (n = 6). The responses to neutrophil enzymes were not associated with cytotoxicity, as studied morphologically, by vital dye exclusion or by lactic dehydrogenase release. Inhibitors of cathepsin G and neutrophil elastase, respectively, inhibited the secretagogue responses to the two enzymes (figure 4). The inhibitory effects of the protease inhibitors were similar to their effects on the amidolytic activity of the proteases. To study whether the effects of purified neutrophil proteases could be reproduced by products released directly from neutrophils, we purified human neutrophils from peripheral blood of healthy subjects (~ 950/0 neutrophils), suspended them in complete Hanks'
550
balanced salt solution (HBSS) (0.6 x 108 cells/ml), activated them with 5 ug/rnl cytochalasin Band 10-6 M FMLP for 10min, and prepared cell-free supernatant by centrifugation. When degranulation supernatant from 105 neutrophils was added per milliliter of incubation medium of serous cells, the secretory response averaged 2,095 ± 54070 above baseline (p < 0.001, n = 6). This response was markedly diminished by preincubation with the human neutrophil elastase inhibitor N-methoxysuccinyl-Ala-Ala-Pro-Val-chloromethylketone (10-5 M). The remaining response (363 ± 22070) was further decreased by the cathepsin G inhibitor chymostatin (50 ug/ml) to 102 ± 19070. Concentrations of cytochalasin Band FMLP equal to those present in supernatant from activated neutrophils had no effect on baseline secretion (0 ± 8070; n = 4). These results indicate that neutrophil elastase and, to a lesser extent, cathepsin G released directly from neutrophils cause gland secretion.
Discussion The present findings indicate that mast cell chymase and neutrophil proteases are remarkably potent secretagogues 0 f cultured airway submucosal glands. In contrast to other agonists such as histamine, prostaglandins, and ~-adrenergic agonists, in which concentrations in excess of 10-8 M are required to induce secretory responses (1,5,6), the threshold concentration for proteases (~ 10-10 M) is lower by two orders of magnitude. In addition, only proteases induce maximal secretory responses (~ 80070 depletion of 35S-labeled macromolecules), and these responses to proteases are> 10-fold larger than those of other agonists (7). These are the most potent secretagogues ever described, and they are derived from two types of inflammatory cells, neutrophils and mast cells. Neutrophils are believed to play an important role in the pathogenesis of airway inflammation. Exposure to air pollutants (e.g., ozone and sulfur dioxide) causes an influx of neutrophils into the airway (8-11); this exposure may be associated with bronchitis and hypersecretion (10, 12).Airway neutrophils also occur in higher numbers in the airways and in bronchoalveolar lavage of cigarette smokers than in those of nonsmokers (13, 14),and free neutrophil elastase activity may be present in the airways of cigarette smokers and in patients with bronchitis (14-18). Increased numbers of neutrophils are found in patients with chronic bronchitis (19). Among conditions associated with increased neutrophils in airways, patients with cystic fibrosis have an immense neutrophil "load" in the airways, and this is associated with high concentrations of neutrophil elastase in the sputum (20, 21). Cystic fibrosis is associated with mucus hypersecretion and progressive airway obstruction, and the discovery of the potent secretagogue effect of neutrophil proteases suggests the possibility that selectiveprotease antagonists may playa role in the therapy of this disease. Neu-
JAY A. NADEL
trophils contain approximately 1 to 2 ug of elastase (22, 23) and 2 to 4 ug of cathepsin G (24, 25) per 106 cells. Thus, the release of 10070 of the proteases from 105 neutrophils/ em" of submucosal tissues may be sufficient to cause a significant secretory response. The number of neutrophils found in the submucosa of mongrel dogs is higher (0.5 x 10' /cm") and is considerably increased after ozone exposure (9). Thus, a secretory response is likely even if> 99070 of the proteases are inhibited by endogenous proteases. In diseases that are characterized by excessiveand abnormal airway secretions (e.g., cystic fibrosis, chronic bronchitis), sputum concentrations of uninhibited neutrophil proteases are increased markedly and greatly exceed the concentrations used in this study (20). These findings suggest a potential role of inhibitors of neutrophil proteases in the treatment of these diseases. Mast cells are also recognized to be important inflammatory cells and are implicated in various inflammatory lung diseases. They are abundant in the airway submucosa (26, 27), where they are in close contact with submucosal glands. Although mast cells containing chymase are rare in the peripheral tissue of normal lungs (28), the proportion of mast cells that contain chymase is much higher in the bronchial epithelium, where the glands are located (29). Furthermore, in certain pathologic conditions, the proportion of mast cells containing chymase may increase. For example, the proportion of chymase-containing mast cells is increased dramatically in bronchoalveolar lavagein patients with cystic fibrosis (30). Degranulation of these mast cells could lead to high local concentrations of chymase, causing activation and secretion from gland cells. The mechanism by which proteases induce gland cell degranulation remains to be determined. The active catalytic site of the proteases is required for degranulation. The high molecular weight of the proteases (e.g., M, of chymase = 30,000 D) makes it likely that their site of action is on the cell surface, although pinocytosis of the protease molecules is not ruled out. Whether the active site interacts with a cell-surface protein or an intracellular protein, studies of chymase action suggest a novel mode of action in relation to stimulation-secretion. Thus, in cultured bovine airway gland serous cells, in contrast to histamine and bradykinin, chymase did not increase cAMP, and the secretory response was unaffected by the phosphodiesterase inhibitor 3-isobutyl-1-methylxanthine (10-5 M); incorporation of 3H-myoinositol into inositol phosphate was unaffected, and depletion of protein kinase C by phorbol 12-myristate did not affect the secretory response; and secretion was not associated with a rise in cytoplasmic Ca 2 +. Thus none of the classic second messengers appears to be involved in chymase-induced secretion (4). In summary, mast cells and neutrophils are the most potent secretagogues of airway sub-
mucosal gland secretion thus far described. Chymase causes secretion by a novel mechanism involving the active site of the enzyme, and the other enzymes may act by a similar mechanism. It is proposed that the release of these enzymes near glands in inflammatory diseases associated with hypersecretion may play an important role in the pathogenesis of these diseases. Therefore, the use of selective site-directed inhibitors of these enzymes may provide an important strategy for therapeutic intervention. Acknowledgment The writer thanks Patty Snell for assistance in manuscript preparation.
References 1. Finkbeiner WE, Nadel JA, Basbaum CB. Establishment and characterization of a cell line derived from bovine tracheal glands. In Vitro Cell Dev BioI 1986; 22:561-7. 2. CaugheyGH, Viro NF, Ramachandran J, Lazarus SC, Borson DB, Nadel JA. Dog mastocytoma tryptase: affinity purification, characterization and amino-terminal sequence. Arch Biochem Biophys 1987; 258:555-63. 3. Caughey GH, Viro NF, Lazarus SC, Nadel JA. Purification and characterization of dog mastocytoma chymase. Identification of an octapeptide conserved in chymotryptic leukocyte proteases. Biochim Biophys Acta 1988; 952:142-9. 4. Sommerhoff CP, Caughey GH, Nadel JA. classical second messengers are not involved in the stimulation-secretion coupling of chymase-induced degranulation of airway gland serous cells.FASEB J 1990; 4:AI940. 5. Sommerhoff CP, Finkbeiner WE, Nadel JA, Basbaum CB. Mediators of anaphylaxis stimulate 35S-labeledmacromolecule secretion from cultured bovine tracheal gland serous cells. Bull Eur Physiopathol Respir 1987; 23:363s. 6. Sommerhoff CP, Finkbeiner WE, Nadel JA, Basbaum CB. Prostaglandin D 2 and prostaglandin E 2 stimulate 35S-labeled macromolecule secretion from cultured bovine tracheal gland serous cells (abstract). Am Rev Respir Dis 1987; 135:A363. 7. Sommerhoff CP, Caughey CH, Finkbeiner WE, Lazarus SC, Basbaum CB, Nadel JA. Mast cell chymase: a potent secretagogue for airway gland serous cells. J Immunol 1989; 142:2450-6. 8. Seltzer J, Bigby BG, Stulbarg M, et al. 0 3 induced change in bronchial reactivity to methacholine and airway inflammation in humans. J Appl Physiol 1986; 60:1321-6. 9. Holtzman MJ, Fabbri LM, O'Byrne PM, et al. Importance of airway inflammation for hyperresponsiveness induced by ozone in dogs. Am Rev Respir Dis 1983; 127:686-90. 10.. Seltzer J, Scanlon PD, Drazen JM, Ingram RH J r, Reid L. Morphologic correlation of physiologic changes caused by S02-induced bronchitis in dogs: the role of inflammation. Am Rev Respir Dis 1984; 129:790-7. 11. Shore SA, Kariya ST, Anderson K, et al. Sulfurdioxide-induced bronchitis in dogs: effects on airway responsiveness to inhaled and intravenously administered methacholine. Am Rev Respir Dis 1987; 135:840-7. 12. Phipps RJ, Denas SM, Sielczak MW, Wanner A. Effects of 0.5 ppm ozone on glycoprotein secretion, ion and water fluxes in sheep trachea. J Appl Physiol 1986; 60:918-27. 13. Reynolds HY, Merrill WW. Airway changes
ROLE OF MAST CELL AND NEUTROPHIL PROTEASES IN AIRWAY SECRETION
in young smokers that may antedate chronic obstructive lung disease. Med Clin North Am 1981; 65:667-89. 14. Stockley RA, Burnett D. Alpha l-antitrypsin and leukocyte elastase in infected and non-infected sputum. Am Rev Respir Dis 1979; 120:1081-6. 15. Ohlsson K, TegnerH. Granulocyte collagenase, elastase and plasma protease inhibitors in purulent sputum. Eur 1 Clin Invest 1975; 5:221-7. 16. Weitz 11,Crowley KA, Landman SL, Lipman BI, Yu 1. Increased neutrophil elastase activity in cigarette smokers. Ann Intern Med 1987;107:680-2. 17. lanoff A, Raju L, Dearing R. Levelsof elastase activity in bronchoalveolar lavage fluids of healthy smokers and nonsmokers. Am Rev Respir Dis 1983; 127:540-4. 18. Tetley TD, Smith SF, Burton GH, Winning Al, Cooke NT, Guz A. Effects of cigarette smoking and drugs on respiratory tract proteases and antiproteases. Eur Respir 1 [Suppl]1987; 153:93-102. 19. Gibson PG, Girgis-Gabarado A, Morris MM, et ale Cellular characteristics of sputum from patients with asthma and chronic bronchitis. Thorax 1989; 44:693-9. 20. Goldstein W, Doring G. Lysosomal enzymes
from polymorphonuclear leukocytesand proteinase inhibitors in patients with cystic fibrosis. Am Rev Respir Dis 1986; 134:49-56. 21. Suter S, Schaad UB, TegnerH, Ohlsson K, Desgrandchamps D, WaldvogelFA. Levelsof free granulocyte elastase in bronchial secretions from patients with cystic fibrosis: effect of antimicrobial treatment against Pseudomonas aeruginosa. 1 Infect Dis 1986; 153:902-9. 22. Baugh Rl, Travis 1. Human leukocyte granule elastase: rapid isolation and characterization. Biochemistry 1976; 15:836-41. 23. Heck LW, Darby WL, Hunter FA, Bhown A, Miller El, Bennett lC. Isolation, characterization, and amino-terminal amino acid sequence analysis of human neutrophil elastase from normal donors. Anal Biochem 1985; 149:153-62. 24. Heck LW, Rostand KS, Hunter FA, Bhown A. Isolation, characterization, and amino-terminal amino acid sequence analysis of human neutrophil cathepsin G from normal donors. Anal Biochem 1986; 158:217-27. 25. Senior RM, Campbell El. Cathepsin G in human mononuclear phagocytes:comparison between monocytes and U927 monocyte-like cells. 1 Immu-
551 nol 1984; 132:2547-51. 26. Guerzon GM, Pare PD, Michoud M-C, Hogg lC. The number and distribution of mast cells in monkey lungs. Am Rev Respir Dis 1979;119:59-66. 27. Shanahan F, MacNiven I, Dyck N, Denburg lA, Bienenstock 1, Befus AD. Human lung mast cells: distribution and abundance of histochemically distinct populations. Int Arch Allergy Appl Immunol 1987; 83:329-31. 28. Schwartz LB, Irani A-MA, Roller K, Castells MC, Schechter NM. Quantitation of histamine, tryptase and chymase in dispersed human T and TC mast cells. 1 Immunol 1987; 138:2611-5. 29. Irani AA, Schechter NM, Craig SS, DeBlois G, Schwartz LB. Two types of human mast cells that have distinct neutral protease compositions. Proc Natl Acad Sci USA 1986; 83:4464-8. 30. Schwartz LB, Irani AA, Mroueh S, Spock A. Reversed ratio of T/TC types of human mast cells in cysticfibrosis (abstract). Clin Res 1987; 35:255A. 31. Sommerhoff CP, Nadel lA, Basbaum CB, Caughey GH. Neutrophil elastase and cathepsin G stimulate secretion from cultured bovine airway gland serous cells. 1 Clin Invest 1990; 85:682-9.
