Eosinophils Cause Acute Edematous Injury in Isolated Perfused Rat Lungs 1 , 2
JUDITH L. ROWEN, DALLAS M. HYDE, and RUTH J. MCDONALD3
Introduction
Eosinophils, like neutrophils, undergo a respiratory burst that leads to the production of toxic oxygen metabolites such as superoxide anion and hydrogen peroxide. In fact, the eosinophil respiratory burst may be more potent and more sustained than that of the neutrophil (1-4). Both cell types contain a peroxidase that converts hydrogen peroxide produced during the respiratory burst to toxic hypohalous acids. Eosinophil peroxidase (EPO) differs from the more familiar myeloperoxidase (MPO) in its affinity for specific halides; whereas MPO utilizes chloride almost exclusively, EPO has a higher affinity for bromide (5) and iodide (6), thus leading to production of the acids HOBr and HOI. EPO has also been shown to participate in the formation of singlet oxygen, another highly toxic oxygen species (7). In addition to oxygen metabolites, the eosinophil produces other potentially injurious substances such as eosinophil cationic protein (6, 8), major basic protein (MBP) (6,8), and leukotriene C4 (LTC4 ) (9-12). Eosinophils predominantly release LTC4 in contrast to neutrophils, which produce primarily LTB4 (9, 10). Despite this formidable armamentarium, the potential for eosinophils to contribute to acute lung injury has not been investigated. Several lines of evidence suggest this is a fruitful avenue to pursue.Eosinophils havebeen shown to cause damage to tracheal explants (13, 14) and pulmonary cell lines in tissue culture (15-18). Eosinophil MBP is localized in the crystalloid core of the eosinophil granule. At a concentration of 100ug/ml, MBP caused exfoliation of guinea pig tracheal explants (14). At lower concentrations, MBP caused ciliostasis in circumscribed areas of the tracheal epithelium. MBP inhibits active ciliated cells by direct impairment of axone ural function, probably by the inhibition of ATPase activity (13). Davis and coworkers (17) demonstrated eosinophil cytotoxicity
SUMMARY Eosinophils produce oxidants and other toxic substances and thus have the potential to cause acute lung injury. We found that addition of normal human eosinophils and the respiratory burst stimulant phorbol myristate acetate to isolated perfused rat lungs caused acute edematous injury as reflected in weight gain and morphologic changes. Lung to body weight ratio (x 103 ) was 16.7 ± 3.3 in the experimental group with stimulated eosinophils added compared with 4.7 ± 0.38 for the control group. Morphologic examination showed both epithelial and endothelial damage. This injury was ameliorated by the addition of catalase, which neutralizes hydrogen peroxide produced during the respiratory burst. Lung/body weight ratio in the group with stimulated eosinophils plus catalase was 7.8 ± 1.1, and the specimens were indistinguishable from control specimens by histopathologic examination. Our results indicate that eosinophils are capable of causing acute lung injury. This injury is mediated, at least in part, by toxic oxygen products. AM REV RESPIR DIS 1990; 142:215-220
against parenchymal and epithelial cell lines, apparently mediated by oxygenradicals. According to work by Agosti and coworkers (18), eosinophil peroxidase activity caused cytolysis of type II pneumocytes. All three components of the EPO system, eosinophil peroxidase, halide (especially iodide), and hydrogen peroxide were required for this injury to occur. Clinical evidence also suggests that eosinophils playa role in acute lung injury. Eosinophil products, specifically eosinophil cationic protein and LTC4 , were found in increased amounts in the plasma and bronchoalveolar lavage fluid of patients with the adult respiratory distress syndrome (ARDS) (19-21). These patients did not have peripheral eosinophilia; presumably, the eosinophils had migrated to the site of injury. The elevation in eosinophil cationic protein correlated with the severity of illness (19, 20), and the rise in LTC4 preceded respiratory failure by 3 days (21), further supporting a possible role in the acute injury. In this report, we directly test the ability of stimulated eosinophils to cause injury in the isolated perfused rat lung, a model of acute edematous lung injury. Methods Materials Phorbol myristate acetate (PMA), catalase, and f-met-Ieu-phe (FMLP) were purchased from Sigma Chemical Co., (St. Louis, MO).
PMA was diluted in dimethylsulfoxide and stored in 50-~g aliquots at - 70° C. The PMA was diluted immediately before use with 4 ml of Hanks' balanced salt solution (HBSS) (Grand Island Biological Company, Grand Island, NY). Catalase was diluted with sterile, filtered phosphate-buffered saline (PBS) (Sigma) to a concentration of 23,900 Vlml. The dilute solution was stored at 4° C. The FMLP was diluted in PBS to 10-3 M and stored in 0.2-ml aliquots at -70° C. An aliquot was thawed shortly before use and diluted further with PBS to 10-5 M.
Eosinophil Isolation The procedure was adapted from the method of Roberts and Gallin (22), which uses FMLP to stimulate neutrophils to change shape and thus their density, allowing separation of eosinophils from neutrophils on Percoll gradients. Blood donors had 1to 10% eosinophils on peripheral smears as determined from a loo-cell count differential. Many donors suffered from allergic rhinitis, but none was receiving medication. After appropriate consent was obtained, 240 ml of venous blood (Received in original form September 25, 1989 and in revised form January 12, 1990) I From the Department of Pediatrics, School of Medicine, and the California Primate Research Center at the University of California, Davis, California. 2 Correspondence and requests for reprints should be addressed to Ruth J. McDonald, M. D., Department of Pediatrics, University of California, Davis Medical Center, 4301 X Street, Sacramento, CA 95817. J Recipient of Clinical Investigator Award K08 HL-01980 from the National Institutes of Health.
215
216
was drawn into 60-ml syringes containing 4 to 5 ml of 3.8070 sodium citrate. The blood was transferred to 50-ml centrifuge tubes and centrifuged at 275 x g for 20 min at room temperature. The overlying plasma was removed and centrifuged at 800 x g for 20 min. The platelet-poor plasma was decanted and reserved. The cell pellet was resuspended with 8 ml of 6% dextran (480,000 average MW; Sigma) and sterile, filtered isotonic saline. The red blood cells were sedimented for 35 min, after which the leukocyte-rich supernatant was removed and washed once at 275 x g for 5 min with HBSS without calcium or magnesium (GIBCO). The cell pellet was then resuspended in 2.25 ml of platelet-poor plasma. The FMLP was added to each pellet for a final concentration of 10-6 to 10-7 M. The pellets were then incubated for 20 min in a shaking water bath at 37° C. During incubation, Percoll (Sigma) gradients were prepared. One gradient was prepared for each 50 ml of whole blood. The materials used were kept on ice. Each 100ml of Percoll had 800 mg NaCI (1.37 M), 115 mg NH2 P 04 (0.0081 M), 20 mg KCI (0.0027 M), and 20 mg KHP04 (0.0015 M) added. For each gradient, 3 ml of 60% (by volume, diluted with PBS) was underlaid with 3 ml of 750/0 Percoll. The cell pellets wereresuspended and overlaid on the gradients. The gradients were centrifuged at 400 x g for 25 min. The eosinophils, found at the interface of the two Percolliayers, wereaspirated and washed twice with HBSS without Ca 2 + or Mg2+. The cells were counted in a hemacytometer and diluted to 101 cells/ml with HBSS without Ca 2 + or Mg 2 +. A cytocentrifuge specimen was prepared and stained with a rapid, modified Wright's stain (Leukostat; Fisher Scientific, Springfield, NJ) for differential cell counting.
Preparation of Isolated, Perfused Lungs The perfusate used was a modified Krebs-Henseleit buffer with 4% Ficoll (Ficoll 70; Sigma). The Ficoll, NaHC0 3 (19.45mM), csci, (1.6 mM), and either KBr or KI (50 to 100 J,1M) were added the day of the experiment. The isolated perfused lungs were prepared using Sprague-Dawley rats weighing 250 to 500 g (Bantin & Kingman, Inc., Fremont, CA, or inborn). The method used is as previously described (23).The pulmonary artery was cannulated, and a large-bore cannula was placed in the left atrium. The pulmonary circulation was washed free of blood in a nonrecirculating fashion as the heart and lungs were dissected en bloc from the chest. Once removed, the lungs were suspended in a warmed, humidified chamber and perfused in a recirculating fashion at an isogravimetric flow rate. Pulmonary artery pressure was continuously monitored, and the lungs were suspended from a force displacement transducer to allow continuous monitoring of changes in lung weight. A 20-min stabilization period elapsed prior to each experimental protocol. In experiments requiring addition of eosinophils, the cells were injected over 5 min directly into the pulmonary circulation through a three-
ROWEN, HYDE, AND MCDONALD
way stopcock. When catalase was added, 1 ml (23,900 U) was added directly to the perfusate at the end of the stabilization period. PMA was also added directly to the perfusate (for a final concentration of 0.0125 J,1g/ ml), but at 10 min after the start of the experimental protocol. Each preparation was monitored for 120min. At the end of this period the heart, trachea, and esophagus weredissected away, and the lungs wereweighed. The weight was divided by the total body weight to correct for the difference in animal size; this lung/total body weight ratio is reported throughout as value x 103 •
trimmed down, thin-sectioned, placed on grids, and stained with uranyl acetate and lead citrate before viewing on a Hitachi 600 transmission electron microscope.
Statistics The data were tested by factorial analysis of variance using Statview 512+ software (Brainpower, Inc, Calabasas, CA) on a Macintosh SE computer. Results
Measurement of Hydrogen Peroxide Production Production of H 202 was assessed by stimulating 1 x 106 eosinophils with PMA followed by reaction with potassium thiocyanate. After incubation, samples were extracted with trichloroacetic acid (TCA) and centrifuged, and supernatants weremixed with 10mM ferrous ammonium sulfate and incubated with 2.5 M potassium thiocyanate for 10min. Absorbance at 480 nm was determined spectrophotometrically and compared with standards of known H 20 2 concentration (26, 27).
Eosinophil Isolation and Activity In Vitro Using the eosinophil isolation method described, the purity of the isolated eosinophils (n = 34) was 87.7 ± 1.5070. Contaminants were almost solely neutrophils, rarely lymphocytes. The isolated cells were capable of mounting a respiratory burst in response to PMA as demonstrated by two in vitro assays.They produced superoxide anion as reflected in cytochrome c reduction (80.0 ± 4.6 nmol reduced/2 x 106 eosinophils, n = 13)and hydrogen peroxide (119.1 ± 10.8 J.1M, n = 5). The yield ranged from 0.26 to 9.3 x 107 eosinophils. As the yield varied, the number of cells injected into experimental lungs varied. However, the mean number of cells (table 1) was not significantly different between the experimental groups.
Morphologic Examination Lungs were preserved for morphologic examination by tracheal instillation of 440 mosm Karnovsky's solution at a pressure of 30 em H 2 O. After a minimum of 12 h of fixation, lungs were cut transversely into 5-mm slabs. Samples weretaken from parenchymal regions on two slabs from the right cranial, right caudal, and left lung lobes. All samples werepostfixed in 1% osmium tetroxide, dehydrated in a series of ethanol baths, and embedded in Araldite. One-micron sections were cut and then stained with methylene blue and basic fuchsin. After the lesions were identified by light microscopy, representative regions were
Effect of Stimulated Eosinophils in Isolated Perfused Rat Lungs Lungs perfused with eosinophils alone or PMA alone had lung/total body weight ratios that were not significantly different from those perfused with buffer alone (table 1). These lungs had normalappearing bronchiolar epithelium and vascular endothelium (figure 1). By light microscopy, the buffer and PMA control groups were indistinguishable. The eosinophil control group was also similar to the buffer perfused group except for occasional eosinophils seen in the
Measurement of Superoxide Anion Production Superoxide anion production was assayed by measuring SOD-inhibitable reduction of cytochrome c (1.49mg/ml) spectrophotometrically at 550 nm after a 20-min incubation period with PMA (1.25 ug/ml) at 37° C (24,25).
TABLE 1 LUNGITOTAL BODY WEIGHT RATIO IN ISOLATED LUNGS PERFUSED WITH EOSINOPHILS Experimental Group
n
Buffer PMA Eosinophils Eos+PMA Catalase Eos/PMAICatalase
5 8 5 9 6 6
Definition • Values t Values :j: Values
LunglTotal Body Weight* (x 10')
4.7 4.9 5.8 16.7 5.2 7.8
± ± ± ± ± ±
0.38 0.24t 0.46t 3.3:1: 0.21t 1.1t
Eosinophils* (x 107 )
1.51 ± 0.24 1.42 ± 0.11 1.54 ± 0.17
of abbreviations: PMA = phorbol myristate acetate. are mean ± 1 SE. not significantly different (p < 0.01) from values obtained with buffer alone. significantly different (p < 0.01) from values obtained with buffer alone.
217
EOSINOPHILS CAUSE EDEMATOUS INJURY IN PERFUSED RAT WNGS
B
A
I
L
small vessels. When eosinophils and PMA were both added, however, the lungs became grossly edematous, with a significantly higher lung/total body weight ratio, 16.7 ± 3.3 versus 4.7 ± 0.38 for buffer control lungs. Epithelial and endothelial injury were apparent in the eosinophil plus PMA treated lungs (figure 2). Epithelial injury in some locations was so severe that 50 to 100 urn lengths of bronchiolar epithelium were missing from the surface. Additionally, electron microscopy revealed necrotic epithelial cells in various stages of lysis. Eosinophils were observed contiguous to the endothelium and in aggregations in the vascular lumina. The halide added to the buffer, iodide, or bromide did not result in a significantly different degree of injury (data not shown). Lungs perfused with buffer alone or eosinophils alone had steady pulmonary artery pressures throughout the 2-h protocol. The PMA control lungs showed a mild, transient increase in pulmonary artery pressure. The eosinophils plus PMA group had a more sustained but again mild increase in pulmonary artery pressure before notable weight gain (data not shown).
Addition of Catalase to Isolated Perfused Rat Lungs Control lungs with catalase added to the buffer did not weigh significantly more than control lungs with buffer alone (table 1), nor were their morphologic features different from those of the control lungs. Addition of catalase alone caused a transient, slight increase in pulmonary artery pressure; the increase seen when catalase, eosinophils, and PMA were added was more sustained but returned to baseline by the end of the 2-h period (data not shown). Addition of catalase to the lungs perfused with eosinophils and PMA markedly reduced the injury, resulting in a mean lung/total body weight ratio of 7.8 ± 1.1, significantly less than the 16.7 ± 3.3 seen with eosinophils plus PMA. The lung/total body weight ratio for the eosinophil, PMA, and catalase group was not significantly different from that of any of the control Fig . 1. Micrographs of tissue from an isolated rat lung perfused with unstimulated eosinophils. A. Light micrograph of a bronchiole (B) and an arteriole (A) from t-urn araldite section stained with methylene blue and basic fuchsin . Magnification: x350. B. Transmission electron micrograph of a bronchiolar wall. Magnification: x4,800 (C = ciliated cell; NCB = nonciliated bronchiolar cell; L = bronchiolar lumen). C. Transmission electron micrograph of an arteriole wall . Magnification: x4,300 (E = endothelial cell; S = smooth muscle cell; L = arteriolar lumen).
218
ROWEN. HYDE, AND MCDONALD
groups. These lungs were indistinguishable from the control groups by morphologic examination as well (figure 3).
,
L
Discussion
Our work is the first to demonstrate acute injury induced by eosinophils in an intact organ system model, thus bridging the gap between work in tissue cultures and the indirect clinical evidence that eosinophils participate in acute injury (1321). Using the isolated perfused rat lung, we have shown that PMA- stimulated human eosinophils cause weight gain and morphologic changes representative of acute lung injury. Our data also support a role for oxygen metabolites in this injury. Lungs with resting human eosinophils or PMA alone had lung/total body weight values similar to those for buffer-perfused lungs (table I). However, when the lungs were perfused with eosinophils that were subsequently stimulated with PMA, the lung /total body weight ratio was markedly increased. Morphologically, there was evidence of both epithelial and endothelial injury (figure 2). Addition of catalase, which scavenges hydrogen peroxide produced during the respiratory burst, led to a decrease in the injury seen in the eosinophil plus PMA group, and the lungs resembled control lungs morphologically (figure 3). Much attention has been focused on the neutrophil as an effector cell in ARDS (28). Our results indicate that the contribution of other cell types such as the eosinophil must not be overlooked. Eosinophil activation appears to be part of the inflammatory process in ARDS as evidenced by a significant correlation between BALF eosinophil cationic protein and neutrophil MPO (20). Our previous work with neutrophils in the isolated perfused rat lung used 4.00 x 107 cells to produce a similar degree of injury (29) obtained in this study with 1.42 X 107 eosinophils. The injury we describe was thus unlikely to be due to the small number of neutrophils contaminating the isolated eosinophils. Additionally, there was no correlation between the small degree of neutrophil contamination and the lung! body weight ratio (data not shown). We isolated eosinophils using a modified version of the Roberts and Gallin method (22). This method ha s been criticized for its use of FMLP, as some investigators maintain that eosinophils are primed by the concentrations of FMLP used (30). Although this is a point worth debating when using this isolation meth-
, A L
,
I
c
219
EOSINOPHILS CAUSE EDEMATOUS INJURY IN PERFUSED RAT WNGS
i
Fig. 2. Micrographs of tissue from an isolated rat lung perfused with eosinophils and PMA. A and C. Araldite sections 1 urn th ick, stained with methylene blue and basic fuchsin. A. Light micrograph of a bronch iole with epithel ial damage (arrowheads) . L = bronchiolar lumen. Magnification: x350. B. Transmission electron micrograph of damaged bronchiolar epithelium (arrowheads). C = ciliated cell; NCB = nonciliated bronchiolar cell (L = bronch iolar lumen). Magnif ication: x 4.300. C. Light micrograph of eosinophil aggregat ion in a venule (arrowheads) . Magnification: x350. O. Transmission electron micrograph of a damaged eosinophil (El in a capillary in an interalveolar septum . The endothelial cell (En) has vesicles and membrane damage (arrowheads) . Magnification: x 4,500.
od to study some aspects of eosinophil biology in vitro, it is not relevant to our work. Our data lead us to conclude that eosinophils, when properly stimulated, can cause acute lung injury. If FMLP contributes to the stimulation, our conclusion is still valid. One subpopulation of activated eosinophils, the "hypo dense"
eosinophils, exhibit greater cytotoxicity (6). Future investigations into the differential capability of eosinophil subpopulations to induce acute lung injury may be warranted. The finding that catalase prevented weight gain indicates that oxygen metabolites, specifically hydrogen peroxide, are
B
required for the injury to occur, either directly or through peroxidase-generated hypohalous acids. The injury may be a result of synergy between eosinophil granule proteins and these toxic oxygen products. As seen in figure 2C and D, eosinophils interact with the endothelium. Presumably, many eosinophil products are released during adherence and activation and thus would be available to cause injury directly or in concert with oxygen metabolites. Again, work with neutrophils is a paradigm, as the damage to isolated lungs is intensified by the presence of both neutrophil elastase and oxygen metabolites (29). Eosinophils do not produce elastase (6), but cytotoxicity to a variety of human cell lines is increased synergistically with a combination of MBP, EPa, and hydrogen peroxide (31). Synergy has also been described between eosinophil cationic protein and oxygen metabolites in killing of schistosomula (32). Eosinophil products could potentially interact with the products of other cell types such as mast cells and basophils. Major basic protein activates basophil and mast cell histamine release (33). Mast cells do produce neutral proteases (34), which could conceivably interact with eosinophil-derived oxygen metabolites much in the same way neutrophil products interact. In conclusion, our work demonstrates that eosinophils cause acute edematous injury in the isolated perfused rat lung model. The mechanism of this injury clearly involves toxic oxygen metabolites, but the exact mechanism needs further delineation. Our data provide further evidence to support the hypothesis that eosinophils contribute to the pathogenesis of ARDS. Acknowledgment The writers thank Linda V. Bruckner and Jana Levin for expert technical assistance. They also wish to thank their blood donors.
References
I ./
•
Fig. 3. Micrograph of tissue from an isolated rat lung perfused with eosinophils, PMA, and catalase. Light micrograph of a bronchiole (B) and an adjacent arteriole (A). Araldite section 1urn thick stained with methylene blue and basic fuchsin. Magnification : x350.
I. Petreccia DC, Nauseef WM, Clark RA. Respiratory burst of normal human eosinophils. J leukocyte Bioi 1987; 41:283-8. 2. Shull PA, Graziano FM, Wallow IH, Busse WW. Comparison of superoxide generation and luminol-dependent chemiluminescence with eosinophils and neutrophils from normal individuals. J Lab Clin Med 1985; 106:638-45. 3. DeChatelet LR, Shirley PS, McPhail LC, Hunt ley CC, Muss HB, Bass DA. Oxidative metabolism of the human eosinophil. Blood 1977; 50:525-35 . 4. Sedgwick 18, VrtisRF, Gourley MF, Busse WW. Stimulus-dependent differences in superoxide anion generation by normal human eosinophils and neutrophils. J Allergy Clin Immunol 1988; 81: 876-83.
220 5. WeissSJ, TestST,Eckmann CM, Roos D, Regiani S. Brominating oxidants generated by human eosinophils. Science 1986; 234:200-3. 6. Gleich GJ, Adolphson CR. The eosinophilic leukocyte: structure and function. Adv Immunol 1986; 39:177-253. 7. Kanofsky JR, Hoogland H, Wever R, WeissSJ. Singlet oxygen production by human eosinophils. J BioI Chern 1988; 263:9692-6. 8. Gleich GJ, Loegering DA. Immunobiology of eosinophils. Annu Rev Immunol1984; 2:429-59. 9. Weller PF, Lee CW, Foster DW, Corey EJ, Austen KF, LewisRA. Generation and metabolism of 5-lipoxygenase pathway leukotrienes by human eosinophils: predominant production of leukotriene C 4 • Proc Natl Acad Sci USA 1983; 80:7626-30. 10. Owen WF, Soberman RJ, YoshimotoT, Sheffer AL, Lewis RA, Austen KF. Synthesis and release of leukotriene C4 by human eosinophils. J Immunol 1987; 138:532-8. 11. Shaw RJ, Cromwell 0, Kay AB. Preferential generation of leukotriene C4 by human eosinophils. Clin Exp Immunol 1984; 56:716-22. 12. Jorg A, Henderson WR, Murphy RC, Klebanoff SJ. Leukotriene generation byeosinophils. J Exp Med 1982; 155:390-402. 13. Hastie AT, Loegering DA, Gleich GJ, Kueppers F. The effect of purified human eosinophil major basic protein on mammalian ciliary activity. Am Rev Respir Dis 1987; 135:848-53. 14. Frigas E, Loegering DA, Gleich GJ. Cytotoxic effects of the guinea pig eosinophil major basic protein on tracheal epithelium. Lab Invest 1980; 42:35-43. 15. Cook RM, Ashworth RF, Musgrove NRJ. Eosinophil- and neutrophil-mediated injury of human lung fibroblast cells. Int Arch Allergy Appl Immunol 1987; 83:428-31. 16. Ayars GH, Altman Le, Gleich GJ, Loegering DA, Baker CB. Eosinophil- and eosinophil granule-
ROWEN, HYDE, AND MCDONALD
mediated pneumocyte injury. J Allergy Clin Immunol 1985; 76:595-604. 17. Davis WB, Fells GA, Sun XH, Gadek JE, Venet A, Crystal RG. Eosinophil-mediated injury to lung parenchymal cells and interstitial matrix. J Clin Invest 1984; 74:269-78. 18. Agosti JM, Altman Le, Ayars GH, Loegering DA, Gleich GJ, Klebanoff SJ. The injurious effect of eosinophil peroxidase, hydrogen peroxide, and halides on pneumocytes in vitro. J Allergy Clin Immunol 1987; 79:496-504. 19. Modig J, Samuelsson T, Hallgren R. The predictive and discriminative value of biologically active products of eosinophils, neutrophils and complement in bronchoalveolar lavage and blood of patients with adult respiratory distress syndrome. Resuscitation 1986; 14:121-34. 20. Hallgren R, Samuelsson T, Venge P, Modig J. Eosinophil activation in the lung is related to lung damage in adult respiratory distress syndrome. Am Rev Respir Dis 1987; 135:639-42. 21. Knoller J, Schonfeld W, Jolia T, Sturm J, Konig W. Generation of leukotrienes in polytraumatic patients with adult respiratory distress syndrome (ARDS). Prog Clin BioI Res 1987; 236A:311-6. 22. Roberts RL, Gallin JI. Rapid method for isolation of normal human peripheral blood eosinophils on discontinuous Percoll gradients and comparison with neutrophils. Blood 1985; 65: 433-40. 23. McDonald RJ, Berger EM, Repine JE. Neutrophil-derived oxygenmetabolites stimulate thromboxane release,pulmonary artery pressure increases, and weight gains in isolated perfused rat lungs. Am Rev Respir Dis 1987; 135:957-9. 24. Curnutte JT, Whitten DM, Babior BM. Defective superoxide production by granulocytes from patients with chronic granulomatous disease. N Engl J Med 1974; 290:593-7. 25. Hoidal JR, Beall GD, Rasp FL, Holmes B,
White JG, Repine JE. Comparison of the metabolism of alveolar macrophages from humans, rats and rabbits. Response to heat-killed bacteria or phorbol myristate acetate. J Lab Clin Med 1978; 93:787-95. 26. Weiss SJ, Young J, LoBuglio AF, Slivka A, Nimeh NF. Role of hydrogen peroxide in neutrophil mediated destruction of cultured endothelial cells. J Clin Invest 1981; 68:714-21. 27. Thurman RG, Ley HG, Scholz R. Hepatic microsomal ethanol oxidation: hydrogen peroxide formation and role of catalase. Eur J Biochem 1972; 25:420-30. 28. TateRM, Repine JE. Neutrophils and the adult respiratory distress syndrome. Am Rev Respir Dis 1983; 128:552-9. 29. McDonald RJ, Bruckner LV, Repine JE. Neutrophil elastase augments acute edematous injury in isolated rat lungs perfused with neutrophil cytoplasts. Am Rev Respir Dis 1989; 140:1825-7. 30. Yazdanbakhsh M, Eckmann CM, Koenderman L, Verhoeven AJ, Roos D. Eosinophils do respond to fMLP. Blood 1987; 70:379-83. 31. Samoszuk MK, Petersen A, Gidanian F, Rietveld C. Cytophilic and cytotoxic properties of human eosinophil peroxidase plus major basic protein. Am J Pathol 1988; 132:455-60. 32. Yazdanbakhsh M, Tai P, Spry CJF, Gleich GJ, Roos D. Synergismbetween eosinophil cationic protein and oxygenmetabolites in killingof schistosomula of Schistosoma mansoni. J Immunol 1987; 138:3443-7. 33. O'Donnell MC, Ackerman SJ, Gleich GJ, Thomas LL. Activation of basophil and mast cell histamine release by eosinophil granule major basic protein. J Exp Med 1983; 157:1981-91. 34. Schwartz LB, Austen KF. Structure and function of the chemical mediators of mast cells. Prog Allergy 1984; 34:271-321.
JUDITH L. ROWEN, DALLAS M. HYDE, and RUTH J. MCDONALD3
Introduction
Eosinophils, like neutrophils, undergo a respiratory burst that leads to the production of toxic oxygen metabolites such as superoxide anion and hydrogen peroxide. In fact, the eosinophil respiratory burst may be more potent and more sustained than that of the neutrophil (1-4). Both cell types contain a peroxidase that converts hydrogen peroxide produced during the respiratory burst to toxic hypohalous acids. Eosinophil peroxidase (EPO) differs from the more familiar myeloperoxidase (MPO) in its affinity for specific halides; whereas MPO utilizes chloride almost exclusively, EPO has a higher affinity for bromide (5) and iodide (6), thus leading to production of the acids HOBr and HOI. EPO has also been shown to participate in the formation of singlet oxygen, another highly toxic oxygen species (7). In addition to oxygen metabolites, the eosinophil produces other potentially injurious substances such as eosinophil cationic protein (6, 8), major basic protein (MBP) (6,8), and leukotriene C4 (LTC4 ) (9-12). Eosinophils predominantly release LTC4 in contrast to neutrophils, which produce primarily LTB4 (9, 10). Despite this formidable armamentarium, the potential for eosinophils to contribute to acute lung injury has not been investigated. Several lines of evidence suggest this is a fruitful avenue to pursue.Eosinophils havebeen shown to cause damage to tracheal explants (13, 14) and pulmonary cell lines in tissue culture (15-18). Eosinophil MBP is localized in the crystalloid core of the eosinophil granule. At a concentration of 100ug/ml, MBP caused exfoliation of guinea pig tracheal explants (14). At lower concentrations, MBP caused ciliostasis in circumscribed areas of the tracheal epithelium. MBP inhibits active ciliated cells by direct impairment of axone ural function, probably by the inhibition of ATPase activity (13). Davis and coworkers (17) demonstrated eosinophil cytotoxicity
SUMMARY Eosinophils produce oxidants and other toxic substances and thus have the potential to cause acute lung injury. We found that addition of normal human eosinophils and the respiratory burst stimulant phorbol myristate acetate to isolated perfused rat lungs caused acute edematous injury as reflected in weight gain and morphologic changes. Lung to body weight ratio (x 103 ) was 16.7 ± 3.3 in the experimental group with stimulated eosinophils added compared with 4.7 ± 0.38 for the control group. Morphologic examination showed both epithelial and endothelial damage. This injury was ameliorated by the addition of catalase, which neutralizes hydrogen peroxide produced during the respiratory burst. Lung/body weight ratio in the group with stimulated eosinophils plus catalase was 7.8 ± 1.1, and the specimens were indistinguishable from control specimens by histopathologic examination. Our results indicate that eosinophils are capable of causing acute lung injury. This injury is mediated, at least in part, by toxic oxygen products. AM REV RESPIR DIS 1990; 142:215-220
against parenchymal and epithelial cell lines, apparently mediated by oxygenradicals. According to work by Agosti and coworkers (18), eosinophil peroxidase activity caused cytolysis of type II pneumocytes. All three components of the EPO system, eosinophil peroxidase, halide (especially iodide), and hydrogen peroxide were required for this injury to occur. Clinical evidence also suggests that eosinophils playa role in acute lung injury. Eosinophil products, specifically eosinophil cationic protein and LTC4 , were found in increased amounts in the plasma and bronchoalveolar lavage fluid of patients with the adult respiratory distress syndrome (ARDS) (19-21). These patients did not have peripheral eosinophilia; presumably, the eosinophils had migrated to the site of injury. The elevation in eosinophil cationic protein correlated with the severity of illness (19, 20), and the rise in LTC4 preceded respiratory failure by 3 days (21), further supporting a possible role in the acute injury. In this report, we directly test the ability of stimulated eosinophils to cause injury in the isolated perfused rat lung, a model of acute edematous lung injury. Methods Materials Phorbol myristate acetate (PMA), catalase, and f-met-Ieu-phe (FMLP) were purchased from Sigma Chemical Co., (St. Louis, MO).
PMA was diluted in dimethylsulfoxide and stored in 50-~g aliquots at - 70° C. The PMA was diluted immediately before use with 4 ml of Hanks' balanced salt solution (HBSS) (Grand Island Biological Company, Grand Island, NY). Catalase was diluted with sterile, filtered phosphate-buffered saline (PBS) (Sigma) to a concentration of 23,900 Vlml. The dilute solution was stored at 4° C. The FMLP was diluted in PBS to 10-3 M and stored in 0.2-ml aliquots at -70° C. An aliquot was thawed shortly before use and diluted further with PBS to 10-5 M.
Eosinophil Isolation The procedure was adapted from the method of Roberts and Gallin (22), which uses FMLP to stimulate neutrophils to change shape and thus their density, allowing separation of eosinophils from neutrophils on Percoll gradients. Blood donors had 1to 10% eosinophils on peripheral smears as determined from a loo-cell count differential. Many donors suffered from allergic rhinitis, but none was receiving medication. After appropriate consent was obtained, 240 ml of venous blood (Received in original form September 25, 1989 and in revised form January 12, 1990) I From the Department of Pediatrics, School of Medicine, and the California Primate Research Center at the University of California, Davis, California. 2 Correspondence and requests for reprints should be addressed to Ruth J. McDonald, M. D., Department of Pediatrics, University of California, Davis Medical Center, 4301 X Street, Sacramento, CA 95817. J Recipient of Clinical Investigator Award K08 HL-01980 from the National Institutes of Health.
215
216
was drawn into 60-ml syringes containing 4 to 5 ml of 3.8070 sodium citrate. The blood was transferred to 50-ml centrifuge tubes and centrifuged at 275 x g for 20 min at room temperature. The overlying plasma was removed and centrifuged at 800 x g for 20 min. The platelet-poor plasma was decanted and reserved. The cell pellet was resuspended with 8 ml of 6% dextran (480,000 average MW; Sigma) and sterile, filtered isotonic saline. The red blood cells were sedimented for 35 min, after which the leukocyte-rich supernatant was removed and washed once at 275 x g for 5 min with HBSS without calcium or magnesium (GIBCO). The cell pellet was then resuspended in 2.25 ml of platelet-poor plasma. The FMLP was added to each pellet for a final concentration of 10-6 to 10-7 M. The pellets were then incubated for 20 min in a shaking water bath at 37° C. During incubation, Percoll (Sigma) gradients were prepared. One gradient was prepared for each 50 ml of whole blood. The materials used were kept on ice. Each 100ml of Percoll had 800 mg NaCI (1.37 M), 115 mg NH2 P 04 (0.0081 M), 20 mg KCI (0.0027 M), and 20 mg KHP04 (0.0015 M) added. For each gradient, 3 ml of 60% (by volume, diluted with PBS) was underlaid with 3 ml of 750/0 Percoll. The cell pellets wereresuspended and overlaid on the gradients. The gradients were centrifuged at 400 x g for 25 min. The eosinophils, found at the interface of the two Percolliayers, wereaspirated and washed twice with HBSS without Ca 2 + or Mg2+. The cells were counted in a hemacytometer and diluted to 101 cells/ml with HBSS without Ca 2 + or Mg 2 +. A cytocentrifuge specimen was prepared and stained with a rapid, modified Wright's stain (Leukostat; Fisher Scientific, Springfield, NJ) for differential cell counting.
Preparation of Isolated, Perfused Lungs The perfusate used was a modified Krebs-Henseleit buffer with 4% Ficoll (Ficoll 70; Sigma). The Ficoll, NaHC0 3 (19.45mM), csci, (1.6 mM), and either KBr or KI (50 to 100 J,1M) were added the day of the experiment. The isolated perfused lungs were prepared using Sprague-Dawley rats weighing 250 to 500 g (Bantin & Kingman, Inc., Fremont, CA, or inborn). The method used is as previously described (23).The pulmonary artery was cannulated, and a large-bore cannula was placed in the left atrium. The pulmonary circulation was washed free of blood in a nonrecirculating fashion as the heart and lungs were dissected en bloc from the chest. Once removed, the lungs were suspended in a warmed, humidified chamber and perfused in a recirculating fashion at an isogravimetric flow rate. Pulmonary artery pressure was continuously monitored, and the lungs were suspended from a force displacement transducer to allow continuous monitoring of changes in lung weight. A 20-min stabilization period elapsed prior to each experimental protocol. In experiments requiring addition of eosinophils, the cells were injected over 5 min directly into the pulmonary circulation through a three-
ROWEN, HYDE, AND MCDONALD
way stopcock. When catalase was added, 1 ml (23,900 U) was added directly to the perfusate at the end of the stabilization period. PMA was also added directly to the perfusate (for a final concentration of 0.0125 J,1g/ ml), but at 10 min after the start of the experimental protocol. Each preparation was monitored for 120min. At the end of this period the heart, trachea, and esophagus weredissected away, and the lungs wereweighed. The weight was divided by the total body weight to correct for the difference in animal size; this lung/total body weight ratio is reported throughout as value x 103 •
trimmed down, thin-sectioned, placed on grids, and stained with uranyl acetate and lead citrate before viewing on a Hitachi 600 transmission electron microscope.
Statistics The data were tested by factorial analysis of variance using Statview 512+ software (Brainpower, Inc, Calabasas, CA) on a Macintosh SE computer. Results
Measurement of Hydrogen Peroxide Production Production of H 202 was assessed by stimulating 1 x 106 eosinophils with PMA followed by reaction with potassium thiocyanate. After incubation, samples were extracted with trichloroacetic acid (TCA) and centrifuged, and supernatants weremixed with 10mM ferrous ammonium sulfate and incubated with 2.5 M potassium thiocyanate for 10min. Absorbance at 480 nm was determined spectrophotometrically and compared with standards of known H 20 2 concentration (26, 27).
Eosinophil Isolation and Activity In Vitro Using the eosinophil isolation method described, the purity of the isolated eosinophils (n = 34) was 87.7 ± 1.5070. Contaminants were almost solely neutrophils, rarely lymphocytes. The isolated cells were capable of mounting a respiratory burst in response to PMA as demonstrated by two in vitro assays.They produced superoxide anion as reflected in cytochrome c reduction (80.0 ± 4.6 nmol reduced/2 x 106 eosinophils, n = 13)and hydrogen peroxide (119.1 ± 10.8 J.1M, n = 5). The yield ranged from 0.26 to 9.3 x 107 eosinophils. As the yield varied, the number of cells injected into experimental lungs varied. However, the mean number of cells (table 1) was not significantly different between the experimental groups.
Morphologic Examination Lungs were preserved for morphologic examination by tracheal instillation of 440 mosm Karnovsky's solution at a pressure of 30 em H 2 O. After a minimum of 12 h of fixation, lungs were cut transversely into 5-mm slabs. Samples weretaken from parenchymal regions on two slabs from the right cranial, right caudal, and left lung lobes. All samples werepostfixed in 1% osmium tetroxide, dehydrated in a series of ethanol baths, and embedded in Araldite. One-micron sections were cut and then stained with methylene blue and basic fuchsin. After the lesions were identified by light microscopy, representative regions were
Effect of Stimulated Eosinophils in Isolated Perfused Rat Lungs Lungs perfused with eosinophils alone or PMA alone had lung/total body weight ratios that were not significantly different from those perfused with buffer alone (table 1). These lungs had normalappearing bronchiolar epithelium and vascular endothelium (figure 1). By light microscopy, the buffer and PMA control groups were indistinguishable. The eosinophil control group was also similar to the buffer perfused group except for occasional eosinophils seen in the
Measurement of Superoxide Anion Production Superoxide anion production was assayed by measuring SOD-inhibitable reduction of cytochrome c (1.49mg/ml) spectrophotometrically at 550 nm after a 20-min incubation period with PMA (1.25 ug/ml) at 37° C (24,25).
TABLE 1 LUNGITOTAL BODY WEIGHT RATIO IN ISOLATED LUNGS PERFUSED WITH EOSINOPHILS Experimental Group
n
Buffer PMA Eosinophils Eos+PMA Catalase Eos/PMAICatalase
5 8 5 9 6 6
Definition • Values t Values :j: Values
LunglTotal Body Weight* (x 10')
4.7 4.9 5.8 16.7 5.2 7.8
± ± ± ± ± ±
0.38 0.24t 0.46t 3.3:1: 0.21t 1.1t
Eosinophils* (x 107 )
1.51 ± 0.24 1.42 ± 0.11 1.54 ± 0.17
of abbreviations: PMA = phorbol myristate acetate. are mean ± 1 SE. not significantly different (p < 0.01) from values obtained with buffer alone. significantly different (p < 0.01) from values obtained with buffer alone.
217
EOSINOPHILS CAUSE EDEMATOUS INJURY IN PERFUSED RAT WNGS
B
A
I
L
small vessels. When eosinophils and PMA were both added, however, the lungs became grossly edematous, with a significantly higher lung/total body weight ratio, 16.7 ± 3.3 versus 4.7 ± 0.38 for buffer control lungs. Epithelial and endothelial injury were apparent in the eosinophil plus PMA treated lungs (figure 2). Epithelial injury in some locations was so severe that 50 to 100 urn lengths of bronchiolar epithelium were missing from the surface. Additionally, electron microscopy revealed necrotic epithelial cells in various stages of lysis. Eosinophils were observed contiguous to the endothelium and in aggregations in the vascular lumina. The halide added to the buffer, iodide, or bromide did not result in a significantly different degree of injury (data not shown). Lungs perfused with buffer alone or eosinophils alone had steady pulmonary artery pressures throughout the 2-h protocol. The PMA control lungs showed a mild, transient increase in pulmonary artery pressure. The eosinophils plus PMA group had a more sustained but again mild increase in pulmonary artery pressure before notable weight gain (data not shown).
Addition of Catalase to Isolated Perfused Rat Lungs Control lungs with catalase added to the buffer did not weigh significantly more than control lungs with buffer alone (table 1), nor were their morphologic features different from those of the control lungs. Addition of catalase alone caused a transient, slight increase in pulmonary artery pressure; the increase seen when catalase, eosinophils, and PMA were added was more sustained but returned to baseline by the end of the 2-h period (data not shown). Addition of catalase to the lungs perfused with eosinophils and PMA markedly reduced the injury, resulting in a mean lung/total body weight ratio of 7.8 ± 1.1, significantly less than the 16.7 ± 3.3 seen with eosinophils plus PMA. The lung/total body weight ratio for the eosinophil, PMA, and catalase group was not significantly different from that of any of the control Fig . 1. Micrographs of tissue from an isolated rat lung perfused with unstimulated eosinophils. A. Light micrograph of a bronchiole (B) and an arteriole (A) from t-urn araldite section stained with methylene blue and basic fuchsin . Magnification: x350. B. Transmission electron micrograph of a bronchiolar wall. Magnification: x4,800 (C = ciliated cell; NCB = nonciliated bronchiolar cell; L = bronchiolar lumen). C. Transmission electron micrograph of an arteriole wall . Magnification: x4,300 (E = endothelial cell; S = smooth muscle cell; L = arteriolar lumen).
218
ROWEN. HYDE, AND MCDONALD
groups. These lungs were indistinguishable from the control groups by morphologic examination as well (figure 3).
,
L
Discussion
Our work is the first to demonstrate acute injury induced by eosinophils in an intact organ system model, thus bridging the gap between work in tissue cultures and the indirect clinical evidence that eosinophils participate in acute injury (1321). Using the isolated perfused rat lung, we have shown that PMA- stimulated human eosinophils cause weight gain and morphologic changes representative of acute lung injury. Our data also support a role for oxygen metabolites in this injury. Lungs with resting human eosinophils or PMA alone had lung/total body weight values similar to those for buffer-perfused lungs (table I). However, when the lungs were perfused with eosinophils that were subsequently stimulated with PMA, the lung /total body weight ratio was markedly increased. Morphologically, there was evidence of both epithelial and endothelial injury (figure 2). Addition of catalase, which scavenges hydrogen peroxide produced during the respiratory burst, led to a decrease in the injury seen in the eosinophil plus PMA group, and the lungs resembled control lungs morphologically (figure 3). Much attention has been focused on the neutrophil as an effector cell in ARDS (28). Our results indicate that the contribution of other cell types such as the eosinophil must not be overlooked. Eosinophil activation appears to be part of the inflammatory process in ARDS as evidenced by a significant correlation between BALF eosinophil cationic protein and neutrophil MPO (20). Our previous work with neutrophils in the isolated perfused rat lung used 4.00 x 107 cells to produce a similar degree of injury (29) obtained in this study with 1.42 X 107 eosinophils. The injury we describe was thus unlikely to be due to the small number of neutrophils contaminating the isolated eosinophils. Additionally, there was no correlation between the small degree of neutrophil contamination and the lung! body weight ratio (data not shown). We isolated eosinophils using a modified version of the Roberts and Gallin method (22). This method ha s been criticized for its use of FMLP, as some investigators maintain that eosinophils are primed by the concentrations of FMLP used (30). Although this is a point worth debating when using this isolation meth-
, A L
,
I
c
219
EOSINOPHILS CAUSE EDEMATOUS INJURY IN PERFUSED RAT WNGS
i
Fig. 2. Micrographs of tissue from an isolated rat lung perfused with eosinophils and PMA. A and C. Araldite sections 1 urn th ick, stained with methylene blue and basic fuchsin. A. Light micrograph of a bronch iole with epithel ial damage (arrowheads) . L = bronchiolar lumen. Magnification: x350. B. Transmission electron micrograph of damaged bronchiolar epithelium (arrowheads). C = ciliated cell; NCB = nonciliated bronchiolar cell (L = bronch iolar lumen). Magnif ication: x 4.300. C. Light micrograph of eosinophil aggregat ion in a venule (arrowheads) . Magnification: x350. O. Transmission electron micrograph of a damaged eosinophil (El in a capillary in an interalveolar septum . The endothelial cell (En) has vesicles and membrane damage (arrowheads) . Magnification: x 4,500.
od to study some aspects of eosinophil biology in vitro, it is not relevant to our work. Our data lead us to conclude that eosinophils, when properly stimulated, can cause acute lung injury. If FMLP contributes to the stimulation, our conclusion is still valid. One subpopulation of activated eosinophils, the "hypo dense"
eosinophils, exhibit greater cytotoxicity (6). Future investigations into the differential capability of eosinophil subpopulations to induce acute lung injury may be warranted. The finding that catalase prevented weight gain indicates that oxygen metabolites, specifically hydrogen peroxide, are
B
required for the injury to occur, either directly or through peroxidase-generated hypohalous acids. The injury may be a result of synergy between eosinophil granule proteins and these toxic oxygen products. As seen in figure 2C and D, eosinophils interact with the endothelium. Presumably, many eosinophil products are released during adherence and activation and thus would be available to cause injury directly or in concert with oxygen metabolites. Again, work with neutrophils is a paradigm, as the damage to isolated lungs is intensified by the presence of both neutrophil elastase and oxygen metabolites (29). Eosinophils do not produce elastase (6), but cytotoxicity to a variety of human cell lines is increased synergistically with a combination of MBP, EPa, and hydrogen peroxide (31). Synergy has also been described between eosinophil cationic protein and oxygen metabolites in killing of schistosomula (32). Eosinophil products could potentially interact with the products of other cell types such as mast cells and basophils. Major basic protein activates basophil and mast cell histamine release (33). Mast cells do produce neutral proteases (34), which could conceivably interact with eosinophil-derived oxygen metabolites much in the same way neutrophil products interact. In conclusion, our work demonstrates that eosinophils cause acute edematous injury in the isolated perfused rat lung model. The mechanism of this injury clearly involves toxic oxygen metabolites, but the exact mechanism needs further delineation. Our data provide further evidence to support the hypothesis that eosinophils contribute to the pathogenesis of ARDS. Acknowledgment The writers thank Linda V. Bruckner and Jana Levin for expert technical assistance. They also wish to thank their blood donors.
References
I ./
•
Fig. 3. Micrograph of tissue from an isolated rat lung perfused with eosinophils, PMA, and catalase. Light micrograph of a bronchiole (B) and an adjacent arteriole (A). Araldite section 1urn thick stained with methylene blue and basic fuchsin. Magnification : x350.
I. Petreccia DC, Nauseef WM, Clark RA. Respiratory burst of normal human eosinophils. J leukocyte Bioi 1987; 41:283-8. 2. Shull PA, Graziano FM, Wallow IH, Busse WW. Comparison of superoxide generation and luminol-dependent chemiluminescence with eosinophils and neutrophils from normal individuals. J Lab Clin Med 1985; 106:638-45. 3. DeChatelet LR, Shirley PS, McPhail LC, Hunt ley CC, Muss HB, Bass DA. Oxidative metabolism of the human eosinophil. Blood 1977; 50:525-35 . 4. Sedgwick 18, VrtisRF, Gourley MF, Busse WW. Stimulus-dependent differences in superoxide anion generation by normal human eosinophils and neutrophils. J Allergy Clin Immunol 1988; 81: 876-83.
220 5. WeissSJ, TestST,Eckmann CM, Roos D, Regiani S. Brominating oxidants generated by human eosinophils. Science 1986; 234:200-3. 6. Gleich GJ, Adolphson CR. The eosinophilic leukocyte: structure and function. Adv Immunol 1986; 39:177-253. 7. Kanofsky JR, Hoogland H, Wever R, WeissSJ. Singlet oxygen production by human eosinophils. J BioI Chern 1988; 263:9692-6. 8. Gleich GJ, Loegering DA. Immunobiology of eosinophils. Annu Rev Immunol1984; 2:429-59. 9. Weller PF, Lee CW, Foster DW, Corey EJ, Austen KF, LewisRA. Generation and metabolism of 5-lipoxygenase pathway leukotrienes by human eosinophils: predominant production of leukotriene C 4 • Proc Natl Acad Sci USA 1983; 80:7626-30. 10. Owen WF, Soberman RJ, YoshimotoT, Sheffer AL, Lewis RA, Austen KF. Synthesis and release of leukotriene C4 by human eosinophils. J Immunol 1987; 138:532-8. 11. Shaw RJ, Cromwell 0, Kay AB. Preferential generation of leukotriene C4 by human eosinophils. Clin Exp Immunol 1984; 56:716-22. 12. Jorg A, Henderson WR, Murphy RC, Klebanoff SJ. Leukotriene generation byeosinophils. J Exp Med 1982; 155:390-402. 13. Hastie AT, Loegering DA, Gleich GJ, Kueppers F. The effect of purified human eosinophil major basic protein on mammalian ciliary activity. Am Rev Respir Dis 1987; 135:848-53. 14. Frigas E, Loegering DA, Gleich GJ. Cytotoxic effects of the guinea pig eosinophil major basic protein on tracheal epithelium. Lab Invest 1980; 42:35-43. 15. Cook RM, Ashworth RF, Musgrove NRJ. Eosinophil- and neutrophil-mediated injury of human lung fibroblast cells. Int Arch Allergy Appl Immunol 1987; 83:428-31. 16. Ayars GH, Altman Le, Gleich GJ, Loegering DA, Baker CB. Eosinophil- and eosinophil granule-
ROWEN, HYDE, AND MCDONALD
mediated pneumocyte injury. J Allergy Clin Immunol 1985; 76:595-604. 17. Davis WB, Fells GA, Sun XH, Gadek JE, Venet A, Crystal RG. Eosinophil-mediated injury to lung parenchymal cells and interstitial matrix. J Clin Invest 1984; 74:269-78. 18. Agosti JM, Altman Le, Ayars GH, Loegering DA, Gleich GJ, Klebanoff SJ. The injurious effect of eosinophil peroxidase, hydrogen peroxide, and halides on pneumocytes in vitro. J Allergy Clin Immunol 1987; 79:496-504. 19. Modig J, Samuelsson T, Hallgren R. The predictive and discriminative value of biologically active products of eosinophils, neutrophils and complement in bronchoalveolar lavage and blood of patients with adult respiratory distress syndrome. Resuscitation 1986; 14:121-34. 20. Hallgren R, Samuelsson T, Venge P, Modig J. Eosinophil activation in the lung is related to lung damage in adult respiratory distress syndrome. Am Rev Respir Dis 1987; 135:639-42. 21. Knoller J, Schonfeld W, Jolia T, Sturm J, Konig W. Generation of leukotrienes in polytraumatic patients with adult respiratory distress syndrome (ARDS). Prog Clin BioI Res 1987; 236A:311-6. 22. Roberts RL, Gallin JI. Rapid method for isolation of normal human peripheral blood eosinophils on discontinuous Percoll gradients and comparison with neutrophils. Blood 1985; 65: 433-40. 23. McDonald RJ, Berger EM, Repine JE. Neutrophil-derived oxygenmetabolites stimulate thromboxane release,pulmonary artery pressure increases, and weight gains in isolated perfused rat lungs. Am Rev Respir Dis 1987; 135:957-9. 24. Curnutte JT, Whitten DM, Babior BM. Defective superoxide production by granulocytes from patients with chronic granulomatous disease. N Engl J Med 1974; 290:593-7. 25. Hoidal JR, Beall GD, Rasp FL, Holmes B,
White JG, Repine JE. Comparison of the metabolism of alveolar macrophages from humans, rats and rabbits. Response to heat-killed bacteria or phorbol myristate acetate. J Lab Clin Med 1978; 93:787-95. 26. Weiss SJ, Young J, LoBuglio AF, Slivka A, Nimeh NF. Role of hydrogen peroxide in neutrophil mediated destruction of cultured endothelial cells. J Clin Invest 1981; 68:714-21. 27. Thurman RG, Ley HG, Scholz R. Hepatic microsomal ethanol oxidation: hydrogen peroxide formation and role of catalase. Eur J Biochem 1972; 25:420-30. 28. TateRM, Repine JE. Neutrophils and the adult respiratory distress syndrome. Am Rev Respir Dis 1983; 128:552-9. 29. McDonald RJ, Bruckner LV, Repine JE. Neutrophil elastase augments acute edematous injury in isolated rat lungs perfused with neutrophil cytoplasts. Am Rev Respir Dis 1989; 140:1825-7. 30. Yazdanbakhsh M, Eckmann CM, Koenderman L, Verhoeven AJ, Roos D. Eosinophils do respond to fMLP. Blood 1987; 70:379-83. 31. Samoszuk MK, Petersen A, Gidanian F, Rietveld C. Cytophilic and cytotoxic properties of human eosinophil peroxidase plus major basic protein. Am J Pathol 1988; 132:455-60. 32. Yazdanbakhsh M, Tai P, Spry CJF, Gleich GJ, Roos D. Synergismbetween eosinophil cationic protein and oxygenmetabolites in killingof schistosomula of Schistosoma mansoni. J Immunol 1987; 138:3443-7. 33. O'Donnell MC, Ackerman SJ, Gleich GJ, Thomas LL. Activation of basophil and mast cell histamine release by eosinophil granule major basic protein. J Exp Med 1983; 157:1981-91. 34. Schwartz LB, Austen KF. Structure and function of the chemical mediators of mast cells. Prog Allergy 1984; 34:271-321.
