- Email: [email protected]
Increased soluble E-Selectin is associated with lung inflammation, and lung injury in hyperoxia-exposed rats
Toxicology Letters Toxicology Letters 87 (1996) 157- 165
Increased soluble E-Selectin is associated with lung inflammation, and lung injury in hyperoxia-exposed rats P.L. Ramsaya, R.S. Geskeb, C.A. Montgomeryb,
S.E. Welty*”
“Department of Pediatrics, Division of’ Neonatology, Baylor College of Medicine, Houston, TX 77030. USA bCenter jbr Comparative Medicine, Baylor College of’ Medicine, Houston. TX 77030. USA
Received I May 1996; revised 24 May 1996; accepted 28 May 1996
Abstract
The pulmonary damage caused by prolonged exposure to high oxygen concentrations is accompanied by lung inflammation, which may contribute to the expression of hyperoxic lung injury. In turn, adhesion molecules are crucial for initiating inflammatory responses. The goal of the present study was to investigate the association of contents of soluble adhesion molecules in plasma or alveolar fluids of hyperoxic rats with lung expression of adhesion molecules, lung inflammation and lung injury. We exposed adult Sprague-Dawley rats to > 95% oxygen for up to 60 h and measured the contents of intercellular adhesion molecule-l (ICAM-1) and E-Selectin in plasma and lung tissue expression of the same molecules, and we assessed lung myeloperoxidase (MPO) activities and lung water contents as indices of lung inflammation and injury, respectively. We also assessed ICAM- content in lavage samples, because ICAM-I may be shed from the alveolar epithelium. Lung water was elevated at 60 h of hyperoxia-exposure, and this effect was preceded by increases in lung MPO activities. Lung ICAM- expression was more than doubled at 48 h, although soluble ICAM- contents were not elevated in plasma or lavage. Soluble E-Selectin was increased by more than 50% at 24 h of hyperoxia-exposure, while lung expressions of E-Selectin were not increased until 48 h. The sequence of the events observed in the present studies suggests that E-Selectin contributes to lung inflammation in hyperoxia and the acceleration of lung injury immediately following the inflammatory response suggests a pivotal role for inflammation in this injury. Keywords:
Adhesion molecules; Lung diseases; Inflammation;
Oxygen toxicity
1. Introduction Lung injury secondary to breathing supplemental oxygen continues to be a major problem in the
*Correspondingauthor, One Baylor Plaza, Room 337D, Houston, TX 77030, USA. Tel: +713 7986190. Fax: +713 7985691; e-mail: [email protected]
treatment of patients experimental animals, peroxic lung injury is injury to the capillary
037%4274/96/$15.00 Q 1996 Elsevier Science Ireland Ltd. All rights reserved PII SO378-4274(96)03773-3
with respiratory failure. In the pathophysiology of hycharacterized by progressive endothelium [1,2] leading to
158
P.L. Ramsay et al. / Toxicology
pulmonary edema [3,4], along with injury to the alveolar epithelium. The onset of capillary injury is closely associated with neutrophil accumulation [2], and recent studies suggest that this inflammation contributes to the course of hyperoxic lung injury [5,6]. Neutrophil accumulation leading to alveolar capillary cell injury is probably mediated through upregulation of endothelial cell adhesion molecules that are ligands for the complementary adhesion molecules expressed on neutrophils. A recent study by Keeny and coworkers suggests that lung E-Selectin is upregulated in the course of hyperoxia-exposure in rats [7], although, to our knowledge, studies to investigate the effect of anti-E-Selectin interventions have not been carried out. Another adhesion molecule that has been implicated in hyperoxia-induced lung inflammation is intercellular adhesion molecule-l (ICAM-1). However, there is no evidence that ICAM- content is increased in the circulatory compartment of the lung [S]. A study by Kang and coworkers found that although ICAM- expressions were increased in the lungs of hyperoxia-exposed mice, the change was due to the increased expression of ICAM- on alveolar epithelial cells. While general mechanisms of tissue inflammation and mechanisms specific to hyperoxic lung injury are being investigated intensively, a series of studies of inflammatory disease has found that soluble adhesion molecules are increased in the deleterious inflammatory processes [9, lo]. Soluble E-Selectin and ICAM- concentrations have been found to correlate with the extent of inflammation in a number of inflammatory diseases [lo], although the precise functions of soluble adhesion molecules in the disease processes are not understood completely. Despite the uncertainties reincreased the specific functions, garding concentrations of specific soluble adhesion molecules appear to correlate well with ongoing tissue inflammation. A particularly prevalent lung disease in which inflammation and hyperoxia contribute significantly is the chronic lung disease observed in premature infants that is termed bronchopulmonary dysplasia (BPD). Substantial efforts have
Letters
87 (1996) 157-165
gone into determining markers that correlate with lung inflammation and the likelihood of developing BPD, although studies to date have concentrated on measuring inflammatory mediators in bronchoalveolar lavage samples or in tracheal aspirates [l 1,121. A recent study found that increased concentrations of soluble ICAMin tracheal aspirates from premature infants was a relatively good predictor for the development BPD [ 131. This finding in human neonates suggests that 1CAM-l is increased in the airspace compartment, and the observation is consistent with studies in which ICAMexpression was found to be increased on alveolar type I and II cells in hyperoxia-exposed animals [8]. The findings in human infants and hyperoxiaexposed animals suggest that increased concentrations of soluble ICAM- in tracheal aspirates may reflect ongoing upregulation of lung tissue ICAM1 expression on the respiratory epithelium. However, this compartment-specific upregulation does not address inflammatory events in the circulatory compartment in the lung and injury to the vascular endothelium. In the study in premature infants, there were no differences in concentrations of soluble ICAM- in the sera, suggesting that ICAMis not upregulated on the pulmonary endothelium. Therefore, it is important to elucidate the pattern of compartment-specific (e.g. vascular, air-space, interstitium) adhesion molecule expression in order to understand the mechanisms by which neutrophils adhere to the pulmonary endothelium, extravasate, and contribute to lung injury. While the data in premature infants suggests airspace involvement of ICAM- 1, measuring soluble E-Selectin in plasma may be an indicator of critical adhesive events in the circulatory compartment, because E-Selectin is expressed only on endothelial cells. The goal of the present study was to investigate the effects of hyperoxia on E-Selectin in the vascular compartment and ICAM- in the vascular and air-space compartments, and correlate the findings with lung tissue expressions. Although relative time course responses are not sufficient to distinguish cause from effect or from irrelevant or nonspecific responses, a cause cannot follow its effects and temporal resolutions can provide use-
ful clues regarding mechanisms of injury and response. In addition, serial assessments of the specific soluble adhesion molecule(s) might provide a useful parameter for monitoring ongoing processes early in the course of hyperoxic lung inflammation.
2. Materials
and methods
Male Sprague-Dawley rats weighing 180&200 g were studied (Harlan Industries, Houston Texas). The rats were placed in a Plexiglas covered chamber that allowed them free access to food and water, while exposing them to > 95% oxygen continuously by administration of pure oxygen at 5 l/min. Excess carbon dioxide was removed with soda lime (Sodasorb; Dewey & Almy Chemical Division. Grace and Co., Lexington, MA). The rats were monitored for signs of illness and the oxygen concentration in the chamber was measured twice daily during hyperoxia-exposure. To obtain samples of lung tissue and blood, rats were anesthetized with 200 mg/kg of pentobarbital ip. 16 gauge catheters were placed in the tracheas, and whole blood samples were obtained with cardiac punctures into syringes containing powdered heparin (Fisher Scientific, Pittsburgh, PA). After collecting blood, the left lung of each animal was lavaged with 5 ml of iced normal saline via the tracheostomy tube and the fluid recovered (between 75%) and 85% of instilled volume) and stored at 4” until processing. The blood and lavage samples were centrifuged in a microfuge for 5 min at 4°C to obtain plasma and BAL supernatant samples. respectively. The cell pellets were resuspended in 1 ml of phosphate-buffered saline, pH 7.4, at 4”C, and divided into 2 equal aliquots for subsequent immunohistochemical and protein analyses. The right lungs, which were not lavaged, were obtained for assessments of lung tissue [CAM-l and E-Selectin expressions so that lung tissue expressions of ICAMand E-Selectin represents expression in the lung as well as in cells accumulating in the lung. Air-breathing rats served as controls in all experiments and samples
from them were obtained and analyzed in parallel with those obtained from hyperoxia-exposed rats.
Lung tissue myeloperoxidase activities were measured as reflections of pulmonary neutrophil accumulation using techniques described by Goldblum and coworkers [14]. Briefly. lungs were homogenized in 0.5% hexadecyltrimethylammonium bromide (HTAB) in 50 mM phosphate buffer (pH = 6.0, approximately 5.0 ml HTAB,‘g tissue). followed by sonication, 3 freeze-thaw cycles and centrifugation (3000 g, 30 min) at 4°C. The myeloperoxidase activities were then determined in the sample supernatants by measuring the of change in A,,,, resulting from decomposition HZO, in the presence of o-dianisidine. Myeloperoxidase activities were expressed as units/lung.
In order to analyze the magnitude of lung injury, extravascular lung water contents were measured by a modification [15] of the wet-to-dry determination described by Pearce et al. [16].
Protein concentrations were determined for the BAL supernatant and the plasma samples using a technique described by Bradford [17]. The cells in the BAL were lysed with serial freeze thaw cycles and the protein concentration determined. Lung tissue was placed in phosphate buffer homogenized and protein concentrations determined. For ICAMdetection. equivalent amounts of protein were electrophoresed on a 10% acrylamide denaturing gel [ 181. For E-Selectin determinations. gels and samples were prepared identically to those described for ICAMexcept for exclusion of SDS and 2-mercaptoethanol.
Details of Western blotting for ICAMbeen described previously [18] and involve
have elec-
160
P.L. Ramsay ef al. / Toxicology Letters 87 (1996) 157-165
trophoretic transfer of protein to a filter, incubation with a primary and secondary antibody, followed by incubation with 10 ~1 of 1 pCci/pl lZ51 protein A (ICN Radiochemicals, Irvine, CA). ICAM- detection employed as a primary antibody lA29 (monoclonal antibody made by immunizing mice against rat ICAM-1). E-Selectin was assayed with a primary antibody CL-37 (monoclonal antibody made by immunizing mice against human E-Selectin cross reactive with rat E-Selectin). Quantitation was by densitometry (ICAMl), or Phosphor Screen and a Phosphoimage Analyzer (Molecular Dynamics, Sunnyvale, CA) (E-Selectin). 2.6. Lavage cell count and immunostaining for ICAM- 1
removed and anti-rat ICAM- was applied at a concentration of 10 pug/ml. The preparations were incubated overnight in a moist chamber at room temperature. After overnight incubation, slides were rinsed in PBS/Tween/BSA then incubated for 45 min at room temperature with a secondary antibody conjugated with biotin. Endogenous peroxidase activity was blocked with methanol and hydrogen peroxide, and the preparations incubated with an avidin-biotin complex using a peroxidase reporter laboratories, (Vector Burlingame, CA), and diaminobenzidine (Sigma Chemical Company, St. Louis, MO) was used to visualize ICAM presence and distribution. The slides were rinsed in PBS/Tween/BSA and counterstained with hematoxylin. 2.7. Statistics
One aliquot of lavage fluid was assayed on a Miles Technicon H-l hematology analyzer (Tarrytown, NY) to establish total leukocyte counts and relative cell populations; another was diluted 1:4 with PBS. Aliquots (100 ~1) of the diluted samples were used for cytological preparations using a Miles Cyto-Tek (Miles Scientific, Elkhart, IN) and CODE-ON plus slides (Fisher Scientific, Pittsburgh, PA). Specimens were centrifuged at 265 g for 20 min, then fixed in ice cold zinc formalin (ANATECH, Battle Creek, MI) for 10 min. The cell preparations were rinsed in PBS containing 0.1% Tween 20 (Sigma Chemical, St. Louis, MO). One slide from each sample was stained using a Diff-Quick staining kit (Scientific Products). Immunocytochemistry was performed using a computer driven CODE-ON system (Instrumentation Laboratories, Lexington, KY). The preparations were treated with Redusol (Research Genetics, Huntsville, AL) for 4 min at room temperature to enhance capillary action then rinsed in PBS/O.l% Tween. All reagents were prepared the PBS/O. 1% Tween to which 0.5% Bovine Serum Albumin (BSA) was added (Gibco BRL, Grand Island, NY). To minimize background staining, sections were blocked with normal goat serum (Vector Laboratories, Burlingame, CA) for 20 min at room temperature prior to application of the primary antibody. Blocking serum was
The continuous variables are expressed as means _+ standard errors. To compare quantitative data, we used unpaired r-tests or one-way ANOVA, with Student-Newman Keuls, depending on the number of groups. All statistics were analyzed using SPSS for windows version 6.0 (Chicago, IL), and statistical differences were considered to be significant at P < 0.05 [19].
3. Results No rats died during hyperoxia-exposure, and no rats had to be sacrificed early because of severe clinical deterioration. Lung inflammation, as assessed by MPO activities (Fig. lA), preceded evidence of pulmonary edema, as assessed by lung water contents (Fig. 1B). Lung tissue ICAMexpressions were more than doubled by 48 h of hyperoxia (Fig. 2A), whereas simultaneous plasma and lavage soluble ICAM- concentrations were not increased (Fig. 2B). In contrast, the cell pellet fractions of BAL fluids contained more than 6 times as much ICAM- than was found in the cell fractions obtained from air-breathing animals (Fig. 2C). E-Selectin expression in the lungs were more than doubled at 48 h of hyperoxia (Fig. 3A), whereas soluble E-Selectin contents were increased by 24 h (Fig. 3B). Interestingly,
P.L. Ramsay
et al.
I Toxicology
Letters
87 (i996)
157-165
161 l
lung E-Selectin decreases back to control levels at 60 h of hyperoxia-exposure, which most likely represents a dilutional effect of the severe protein leak in the lung, rather than an actual decrease in lung E-Selectin expression. Although the number of cells recovered in BAL samples from rats exposed to hyperoxia increased relative to air-breathing rats, there were no changes in the proportion of the cells that were alveolar macrophages and neutrophils (Table l), indicating that the increased content of ICAMin the cellular component of the BAL samples was not a result of a change in relative proportions of the cells recovered. These data implicate increases in ICAM- expression in one or more specific cell types, such as alveolar macrophages. Immunohistochemical analyses of the cell pellet specimens
l
i
125,
-
75
! q50 e 25
Fig. 2. Quantification of ICAM-I expression in hyperoxia-exposed rat lung tissue and in the plasma and BAL. Western blots for ICAM-I revealed that lung expression of ICAM-I was more than doubled at 48 h (2A), while soluble ICAM-I in the plasma or BAL was not different than at the same time of hyperoxia exposure (28). Lavage cell ICAM-I was × greater in cells exposed to hyperoxia for 48 h (2C). Data are expressed as means k S.E.M., n = 4 or 5 in each group. *P < 0.05 vs. 0 h or air-breathing animals. Equivalent amounts of protein were loaded for all Western blots. Fig. I. Lung injury and inflammation in the course of hyperoxia-exposure. Sprague-Dawley rats were exposed to hyperoxia-exposure for up to 60 h, and lung MPO activities (A) and extravascular lung water contents (B) were determined as measures of lung inflammation and injury, respectively. Extravascular lung water content was increased at 60 h and lung MPO activities increased at 48 h. Data are expressed as means F S.E.M., n = 6 in each group. *P -C 0.05 vs. 0 h or air-breathing animals.
revealed that ICAM- expression is almost entirely limited to alveolar macrophages (Fig. 4A and B from cells exposed to 48 h and 0 h of hyperoxia, respectively). In this figure, ICAM- is observed as brown staining in the cells (see arrow in 4A). Alveolar macrophages are identifiable by
162
P. L. Rroustr~~ PI ol.
To.\icdo,q~
their large size and the mononuclear appearance. The uniform brown staining in lymphocytes (Fig. 4C, see arrow), a cell type that expresses ICAM-I constitutively, supports the specificity of the analysis for ICAM-I. Alveolar macrophages from rats exposed to hyperoxia were 50 f 3% positive for ICAM-I, whereas I5 f 2% of the alveolar macrophages from control rats were positive (mean + S.E.M., II = 4), with 10 high power fields per sample examined (P < 0.001). The six-fold increase in ICAMcontent in lavage cell pellets and the three-fold increase in the percentage of macrophages showing protein expression further suggests a greater extent of ICAM-I expression per cell in the hyperoxic rats.
Lctrc~~\ 87 (I 9%) Table
IS7
IhJ
I
Effect of hyperoxia
on bronchoalveolar
lavage cell count
Air breathing
Hyperoxic
__~
_
Cell count (cells 181)
460 2
‘I%Neutrophils ‘X Alveolar
*P
<
I I9
905 *
7*4
macrophages
~- ..-
I IO*
7*5 86 + 4
91 i-5
0.001.
II = 4 in hyperoxic
group and 6 in air breathing
Rats were placed in either
group.
> 95 ‘% oxygen or room air for 48 h.
After anesthesia. 5 ml of iced saline was instilled into one lung. The recovered Invage was spun at IO 000 g for IO min at 4°C. and the cell pellet was resuspended in white
cell count
and
differential
I
ml of sterile salrnc. and
were obtained.
Reco\crcd
lavagc in every cahe was between 75 and 85’%#instilled volume.
4. Discussion
* 3x7
3x
ax 71
8
s
ml
0
0
24
I!6
48
60
ttWSG$XEEdtO>950/0~ Fig. 3. Lung and soluble E-Selectin Sprague of
Dawley
> 95% oxygen.
plasma
After
hyperoxia-exposure.
samples were analyzed
content,
in hyperoxia-exposed
respectively
by
lung tissue and
for E-Selectin
Western
blotting.
expression Signal
E-Selectin
was found to be more than doubled
Similarly,
plasma
E-Selectin
was increased
Data are expressed as means
*P
<
0.05
f
S.E.M.,
vs. 0 h or air-breathing
of protein
for
and lung
after 48 h (A).
by approximately
50% at 24 h of exposure and more than doubled
amounts
rats.
rats were exposed to room air or up to 60 h
at 48 h (B).
n = 6 in catch group. animals.
were loaded for all Wcstcrn
Equivalent hlol\.
The most significant finding in the present study is that the sequential increases in E-Selectin. ICAM-I, lung MPO and lung injury is consistent with the hypothesis that E-Selectin plays a significant role in initiation of hyperoxic lung inflammation and that lung inflammation contributes to lung injury. While the sequential interactions suggest a role for inflammation in hyperoxic lung injury. more definitive studies are necessary to determine the extent of the contribution. Furthermore. subtle evidence of lung injury may be cvident prior to neutrophil accumulation. indicating that there is ;I component of neutrophil independent lung injury [2]. In the animal model employed in the present studies, significant lung injury is manifested only after 60 h of exposure, and this injury is preceded by increases in lung MPO activities. The increases in lung MPO activities coincides with increases in lung ICAM-I and E-Selectin expression; however, soluble E-Selectin contents are increased as early as 24 h after exposure to hyperoxia. The increases in contents of soluble E-Selectin observed at 24 h suggest that the expression of E-Selectin on the pulmonary vascular endothelium is a significant early event in hyperoxia-exposure that leads to adherence of neutrophils. The interaction of the neutrophils with rhe endothelium leads to cleavage of E-ScIcctin, which increases measurably in the plasma. The increase in plasma E-Selectin prior to the incrcnse in lung infkunmation (MPO activities)
163
164
P.L. Ramsay
et al. 1 Toxicology
suggests that neutrophils are adhering to lung endothelium, but that neutrophil clearance equals neutrophil adherence. Another explanation is that the number of neutrophils is actually increasing but that the neutrophils have reduced cellular contents of myeloperoxidase. The data available at present suggest that the observed increases in soluble E-Selectin reflect increased cellular expression, in combination with cleavage of E-Selectin from the cell surface as a result neutrophil interaction with E-Selectin [9]. Hence, increased soluble E-Selectin may be a marker of E-Selectin-mediated lung/neutrophil interactions, which is supported by the parallel increases in lung and soluble E-Selectin preceding increases in lung MPO activities. In addition to the sequence of events suggesting a primary role for E-Selectin in hyperoxic lung injury, our findings suggest that lung inflammatory events may be detectable in plasma samples, although plasma samples may not be specific for events occurring in the lung. Despite the concern of specificity, if the same correlations with soluble adhesion molecules in the present study are expressed in humans, measuring soluble adhesion molecules in the plasma or serum may prove to be useful for detection of processes leading to lung inflammation prior to significant injury, thereby directing the rational and timely interventions with anti-inflammatory therapeutics. It is important to note that studies from our lab suggest that E-Selectin is not upregulated in hyperoxia-exposed mice, although P-selectin is upregulated 1201, suggesting that there are species differences in the adhesive mechanisms for the development of lung inflammation. However, the general sequence of events from Selectin-mediated transient adhesion to neutrophil activation and firm adhesion appears to be retained.
Letters
87 (1996) 157-165
The lack of increase in soluble ICAM- in the plasma of hyperoxia-exposed rats is consistent with an earlier report by Kojima and coworkers, in which no increases in soluble ICAM- were observed in the sera of premature infants [13]. Our findings in the BAL differ somewhat from the findings in premature neonates in that Kojima et al. found an increase in soluble ICAM- in tracheal aspirates, whereas we did not observe an increase in soluble ICAM- contents in BAL samples. The simplest interpretation of the difference between studies is that the models of lung injury examined in the two studies are fundamentally different. The study by Kojima et al. involved hyperoxic exposure and barotrauma on immature lungs. In contrast, the model we studied was one of pure oxygen toxicity in spontaneously breathing adult animals. An additional finding from the present study is that increased ICAM- expression is associated with alveolar macrophages in the cellular component of the BAL. It is not clear whether the increase ICAM- represents increased expressions of ICAM- on these cells, or whether the ICAMhas been phagocytosed by these cells. In either event, further studies are needed to determine the causes and implications of the increases in ICAM1 expressions associated with alveolar macrophages during the evolution of hyperoxic lung injury. In summary, the fact that contents of soluble E-Selectin increased in the plasma compartment prior to increases in expressions of E-Selectin in lung tissue, and that both preceded lung inflammation suggests that E-Selectin is a significant mediator of lung inflammation in hyperoxic lung injury, and further studies to determine the effect of interventions affecting E-Selectin expression or function as a means of attenuating lung inflammation and injury are warranted. In addition,
Fig. 4. Immunohistochemical analysis of lavage cells For ICAM-I. Cell pellets from lavages of rats exposed to either room air or 48 h of z 95% oxygen were placed on a slide, incubated with anti-rat ICAM-I, followed by incubation with a biotin-conjugated secondary antibody. ICAM-I signal was visualized after incubations with avidin-biotin complex, peroxidase reporter and diaminobenzidine. In section (A), alveolar macrophages from hyperoxia-exposed rats are visualized and have substantial staining for ICAM-I (see arrow), while in section (B) there are alveolar macrophages from room air-exposed rats with no staining for ICAM-I. In section (C), there is a lymphocyte, a cell type known to express ICAM-1, and in the section it is intensely stained for ICAM-I (see arrow), supporting the specificity of the method.
increases in soluble E-Selectin may prove to be valuable as an indicator of early inflammatory events, and could be used to guide early anti-inflammatory interventions in patients at high risk for development of deleterious responses to lung inflammation.
Acknowledgements This work was supported by NIH grant AI19031, NIH grant 5 P30 HD27823-04, and an AHA grant, Texas Affiliate.
References
lung injury: Biochemical. [II Smith, L.J. (1985) Hyperoxic cellular, and morphologic characterization in the mouse. J. Lab. Clin. Med. 106, 2699278. changes in pulmonary PI Crapo, J.D. (1986) Morphologic oxygen toxicity. Ann. Rev. Physiol. 48, 721-731. [31 Bressack, M.A., McMillan, D.D. and Bland, R.D. (1979) Pulmonary oxygen toxicity: increased microvascular permeability to protein in anesthetized lambs. Lymphology I?, 133 139. T.A. and Bland, R.D. ( 1982) [41 Hansen, T.N., Hazinski, Vitamin E does not prevent oxygen-induced lung injury in newborn lambs. Pediatr. Res. 16, 5833587. C.D.. Wolyniec. W.W., LaPlante, A.M., [51 Wegner, Marschman. K., Lubbe. K., Haynes, N., Rothlein, R. and Letts. G.L. (1992) Intercellular adhesion molecule-l contributes to pulmonary oxygen toxicity in mice: Role of leukocytes revisited. Lung 170. 267 -279. [61 Ramsay, P.L., Welty. SE., Bree, M.P. and Hansen, T.N. (1994) Lung injury and inflammation are reduced in hyperoxic mice by inhibition of neutrophil /12 integrin CDI la:CDI8. Pediatr. Res. 35, 396A. S.E.. Nowooryta, K. and Schmalstieg. F.C. (71 Keeney. (1995) The effect of hyperoxia on pulmonary E-Selectin and ICAM-I in neonatal rats. Pediatr. Res. 37. 337A. PI Kang. B.H.. Crapo. J.D., Wegner, CD.. Letts, L.G. and Chang. L.Y. (I 993) Intercellular adhesion molecule-l expression on the alveolar epithelium and its modification by hyperoxia. Am. J. Resp. Cell. Mol. Biol. 9, 350-355.
[91 Gearing. adhesion
A.J.H. molecules
and
Newman.
in disease.
W. (1993) Immunol.
Circulating
Today
14. 506~
512. 1.. Pigott, R.. Hughes, J., [lOI Gearing, A.J.H.. Hemingway. Rees, A.J. and Cashman, S.J. (1992) Soluble forms of vascular adhesion molecules. E-Selectin. ICAM-I. and VCAM-I: Pathological significance. Ann New York Acad. Sci. 667. 324 33 I. Groneck. P.. Reuss, D.. Gotze-Speer, B. and Speer. C.P. [I II (1993) Effects of dexamethasone on chemotactic activity and inflammatory mediators in tracheobronchial aspirates of preterm infants at risk for chronic lung disease. J. Pediatr. 122, 938 -944. B.. Oppermann. M.. Eiffert. H. LIZI Groneck. P.. Gotze-Speer, and Speer, C.P. (1995) Association of pulmonary inflammation and increased microvascular permeability during the development of bronchopulmonary dysplasia: a sequential analysis of inflammatory mediators in respiratory fluids of high-risk preterm neonates. Pediatrics 93. 712 718. T.. Sasai, M. and Kobayaski. Y. (1993) In[I31 Kojima. creased soluble ICAM-I in tracheal aspirates of infants with bronchopulmonary dysplasia. Lancet 342. IO23 I O-74. S.E.. Kuei-Meng, W. and Jay. M. (1985) Lung [I41 Goldblum, myeloperoxidase as a measure of pulmonary leukostasis in rabbits. J. Appl. Physiol. 59, 1978 1985. A.J.. Vaughan III, T.R., Brigham. K.L.. [I51 Erdmann, Woolverton. W.C. and Staub, N.C. (1975) Effect of increased vascular pressure on lung fluid balance in unanesthetized sheep. Circ. Res. 37, 271 284. J. and Beazell, J. (1965) Mea[I61 Pearce, M.L., Yamashita. surement of pulmonary edema. Circ. Rcs. 16, 482 -488. iI71 Bradford, M.M. (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72. 248. [I81 Welty. S.E., Rivera, J.L., Elliston, J.. Smith. C.V.. Zeb. T., Ballantyne, C.M.. Montgomery. C.A. and Hansen. T.N. (1993) increases in lung tissue expression of intercellular adhesion molecule-l are associated with hyperoxic lung injury and inflammation in mice. Am J. Resp. Cell. Mol. Biol. 9. 3933400. Analysis, 2nd ed. Prentice [I91 Zar. J.H. (1984) Biostatistical Hall. Englewood Cliffs, New Jersey. pp. 43 44. P.L.. Piedboeuf. B.. ct al. ( 1995) PO1 Welty. S.E., Ramsay. P-Selectin is upregulated in hyperoxic mouse lungs. Am. J. Respir. Crit. Care Med. 151. 4542.
Increased soluble E-Selectin is associated with lung inflammation, and lung injury in hyperoxia-exposed rats P.L. Ramsaya, R.S. Geskeb, C.A. Montgomeryb,
S.E. Welty*”
“Department of Pediatrics, Division of’ Neonatology, Baylor College of Medicine, Houston, TX 77030. USA bCenter jbr Comparative Medicine, Baylor College of’ Medicine, Houston. TX 77030. USA
Received I May 1996; revised 24 May 1996; accepted 28 May 1996
Abstract
The pulmonary damage caused by prolonged exposure to high oxygen concentrations is accompanied by lung inflammation, which may contribute to the expression of hyperoxic lung injury. In turn, adhesion molecules are crucial for initiating inflammatory responses. The goal of the present study was to investigate the association of contents of soluble adhesion molecules in plasma or alveolar fluids of hyperoxic rats with lung expression of adhesion molecules, lung inflammation and lung injury. We exposed adult Sprague-Dawley rats to > 95% oxygen for up to 60 h and measured the contents of intercellular adhesion molecule-l (ICAM-1) and E-Selectin in plasma and lung tissue expression of the same molecules, and we assessed lung myeloperoxidase (MPO) activities and lung water contents as indices of lung inflammation and injury, respectively. We also assessed ICAM- content in lavage samples, because ICAM-I may be shed from the alveolar epithelium. Lung water was elevated at 60 h of hyperoxia-exposure, and this effect was preceded by increases in lung MPO activities. Lung ICAM- expression was more than doubled at 48 h, although soluble ICAM- contents were not elevated in plasma or lavage. Soluble E-Selectin was increased by more than 50% at 24 h of hyperoxia-exposure, while lung expressions of E-Selectin were not increased until 48 h. The sequence of the events observed in the present studies suggests that E-Selectin contributes to lung inflammation in hyperoxia and the acceleration of lung injury immediately following the inflammatory response suggests a pivotal role for inflammation in this injury. Keywords:
Adhesion molecules; Lung diseases; Inflammation;
Oxygen toxicity
1. Introduction Lung injury secondary to breathing supplemental oxygen continues to be a major problem in the
*Correspondingauthor, One Baylor Plaza, Room 337D, Houston, TX 77030, USA. Tel: +713 7986190. Fax: +713 7985691; e-mail: [email protected]
treatment of patients experimental animals, peroxic lung injury is injury to the capillary
037%4274/96/$15.00 Q 1996 Elsevier Science Ireland Ltd. All rights reserved PII SO378-4274(96)03773-3
with respiratory failure. In the pathophysiology of hycharacterized by progressive endothelium [1,2] leading to
158
P.L. Ramsay et al. / Toxicology
pulmonary edema [3,4], along with injury to the alveolar epithelium. The onset of capillary injury is closely associated with neutrophil accumulation [2], and recent studies suggest that this inflammation contributes to the course of hyperoxic lung injury [5,6]. Neutrophil accumulation leading to alveolar capillary cell injury is probably mediated through upregulation of endothelial cell adhesion molecules that are ligands for the complementary adhesion molecules expressed on neutrophils. A recent study by Keeny and coworkers suggests that lung E-Selectin is upregulated in the course of hyperoxia-exposure in rats [7], although, to our knowledge, studies to investigate the effect of anti-E-Selectin interventions have not been carried out. Another adhesion molecule that has been implicated in hyperoxia-induced lung inflammation is intercellular adhesion molecule-l (ICAM-1). However, there is no evidence that ICAM- content is increased in the circulatory compartment of the lung [S]. A study by Kang and coworkers found that although ICAM- expressions were increased in the lungs of hyperoxia-exposed mice, the change was due to the increased expression of ICAM- on alveolar epithelial cells. While general mechanisms of tissue inflammation and mechanisms specific to hyperoxic lung injury are being investigated intensively, a series of studies of inflammatory disease has found that soluble adhesion molecules are increased in the deleterious inflammatory processes [9, lo]. Soluble E-Selectin and ICAM- concentrations have been found to correlate with the extent of inflammation in a number of inflammatory diseases [lo], although the precise functions of soluble adhesion molecules in the disease processes are not understood completely. Despite the uncertainties reincreased the specific functions, garding concentrations of specific soluble adhesion molecules appear to correlate well with ongoing tissue inflammation. A particularly prevalent lung disease in which inflammation and hyperoxia contribute significantly is the chronic lung disease observed in premature infants that is termed bronchopulmonary dysplasia (BPD). Substantial efforts have
Letters
87 (1996) 157-165
gone into determining markers that correlate with lung inflammation and the likelihood of developing BPD, although studies to date have concentrated on measuring inflammatory mediators in bronchoalveolar lavage samples or in tracheal aspirates [l 1,121. A recent study found that increased concentrations of soluble ICAMin tracheal aspirates from premature infants was a relatively good predictor for the development BPD [ 131. This finding in human neonates suggests that 1CAM-l is increased in the airspace compartment, and the observation is consistent with studies in which ICAMexpression was found to be increased on alveolar type I and II cells in hyperoxia-exposed animals [8]. The findings in human infants and hyperoxiaexposed animals suggest that increased concentrations of soluble ICAM- in tracheal aspirates may reflect ongoing upregulation of lung tissue ICAM1 expression on the respiratory epithelium. However, this compartment-specific upregulation does not address inflammatory events in the circulatory compartment in the lung and injury to the vascular endothelium. In the study in premature infants, there were no differences in concentrations of soluble ICAM- in the sera, suggesting that ICAMis not upregulated on the pulmonary endothelium. Therefore, it is important to elucidate the pattern of compartment-specific (e.g. vascular, air-space, interstitium) adhesion molecule expression in order to understand the mechanisms by which neutrophils adhere to the pulmonary endothelium, extravasate, and contribute to lung injury. While the data in premature infants suggests airspace involvement of ICAM- 1, measuring soluble E-Selectin in plasma may be an indicator of critical adhesive events in the circulatory compartment, because E-Selectin is expressed only on endothelial cells. The goal of the present study was to investigate the effects of hyperoxia on E-Selectin in the vascular compartment and ICAM- in the vascular and air-space compartments, and correlate the findings with lung tissue expressions. Although relative time course responses are not sufficient to distinguish cause from effect or from irrelevant or nonspecific responses, a cause cannot follow its effects and temporal resolutions can provide use-
ful clues regarding mechanisms of injury and response. In addition, serial assessments of the specific soluble adhesion molecule(s) might provide a useful parameter for monitoring ongoing processes early in the course of hyperoxic lung inflammation.
2. Materials
and methods
Male Sprague-Dawley rats weighing 180&200 g were studied (Harlan Industries, Houston Texas). The rats were placed in a Plexiglas covered chamber that allowed them free access to food and water, while exposing them to > 95% oxygen continuously by administration of pure oxygen at 5 l/min. Excess carbon dioxide was removed with soda lime (Sodasorb; Dewey & Almy Chemical Division. Grace and Co., Lexington, MA). The rats were monitored for signs of illness and the oxygen concentration in the chamber was measured twice daily during hyperoxia-exposure. To obtain samples of lung tissue and blood, rats were anesthetized with 200 mg/kg of pentobarbital ip. 16 gauge catheters were placed in the tracheas, and whole blood samples were obtained with cardiac punctures into syringes containing powdered heparin (Fisher Scientific, Pittsburgh, PA). After collecting blood, the left lung of each animal was lavaged with 5 ml of iced normal saline via the tracheostomy tube and the fluid recovered (between 75%) and 85% of instilled volume) and stored at 4” until processing. The blood and lavage samples were centrifuged in a microfuge for 5 min at 4°C to obtain plasma and BAL supernatant samples. respectively. The cell pellets were resuspended in 1 ml of phosphate-buffered saline, pH 7.4, at 4”C, and divided into 2 equal aliquots for subsequent immunohistochemical and protein analyses. The right lungs, which were not lavaged, were obtained for assessments of lung tissue [CAM-l and E-Selectin expressions so that lung tissue expressions of ICAMand E-Selectin represents expression in the lung as well as in cells accumulating in the lung. Air-breathing rats served as controls in all experiments and samples
from them were obtained and analyzed in parallel with those obtained from hyperoxia-exposed rats.
Lung tissue myeloperoxidase activities were measured as reflections of pulmonary neutrophil accumulation using techniques described by Goldblum and coworkers [14]. Briefly. lungs were homogenized in 0.5% hexadecyltrimethylammonium bromide (HTAB) in 50 mM phosphate buffer (pH = 6.0, approximately 5.0 ml HTAB,‘g tissue). followed by sonication, 3 freeze-thaw cycles and centrifugation (3000 g, 30 min) at 4°C. The myeloperoxidase activities were then determined in the sample supernatants by measuring the of change in A,,,, resulting from decomposition HZO, in the presence of o-dianisidine. Myeloperoxidase activities were expressed as units/lung.
In order to analyze the magnitude of lung injury, extravascular lung water contents were measured by a modification [15] of the wet-to-dry determination described by Pearce et al. [16].
Protein concentrations were determined for the BAL supernatant and the plasma samples using a technique described by Bradford [17]. The cells in the BAL were lysed with serial freeze thaw cycles and the protein concentration determined. Lung tissue was placed in phosphate buffer homogenized and protein concentrations determined. For ICAMdetection. equivalent amounts of protein were electrophoresed on a 10% acrylamide denaturing gel [ 181. For E-Selectin determinations. gels and samples were prepared identically to those described for ICAMexcept for exclusion of SDS and 2-mercaptoethanol.
Details of Western blotting for ICAMbeen described previously [18] and involve
have elec-
160
P.L. Ramsay ef al. / Toxicology Letters 87 (1996) 157-165
trophoretic transfer of protein to a filter, incubation with a primary and secondary antibody, followed by incubation with 10 ~1 of 1 pCci/pl lZ51 protein A (ICN Radiochemicals, Irvine, CA). ICAM- detection employed as a primary antibody lA29 (monoclonal antibody made by immunizing mice against rat ICAM-1). E-Selectin was assayed with a primary antibody CL-37 (monoclonal antibody made by immunizing mice against human E-Selectin cross reactive with rat E-Selectin). Quantitation was by densitometry (ICAMl), or Phosphor Screen and a Phosphoimage Analyzer (Molecular Dynamics, Sunnyvale, CA) (E-Selectin). 2.6. Lavage cell count and immunostaining for ICAM- 1
removed and anti-rat ICAM- was applied at a concentration of 10 pug/ml. The preparations were incubated overnight in a moist chamber at room temperature. After overnight incubation, slides were rinsed in PBS/Tween/BSA then incubated for 45 min at room temperature with a secondary antibody conjugated with biotin. Endogenous peroxidase activity was blocked with methanol and hydrogen peroxide, and the preparations incubated with an avidin-biotin complex using a peroxidase reporter laboratories, (Vector Burlingame, CA), and diaminobenzidine (Sigma Chemical Company, St. Louis, MO) was used to visualize ICAM presence and distribution. The slides were rinsed in PBS/Tween/BSA and counterstained with hematoxylin. 2.7. Statistics
One aliquot of lavage fluid was assayed on a Miles Technicon H-l hematology analyzer (Tarrytown, NY) to establish total leukocyte counts and relative cell populations; another was diluted 1:4 with PBS. Aliquots (100 ~1) of the diluted samples were used for cytological preparations using a Miles Cyto-Tek (Miles Scientific, Elkhart, IN) and CODE-ON plus slides (Fisher Scientific, Pittsburgh, PA). Specimens were centrifuged at 265 g for 20 min, then fixed in ice cold zinc formalin (ANATECH, Battle Creek, MI) for 10 min. The cell preparations were rinsed in PBS containing 0.1% Tween 20 (Sigma Chemical, St. Louis, MO). One slide from each sample was stained using a Diff-Quick staining kit (Scientific Products). Immunocytochemistry was performed using a computer driven CODE-ON system (Instrumentation Laboratories, Lexington, KY). The preparations were treated with Redusol (Research Genetics, Huntsville, AL) for 4 min at room temperature to enhance capillary action then rinsed in PBS/O.l% Tween. All reagents were prepared the PBS/O. 1% Tween to which 0.5% Bovine Serum Albumin (BSA) was added (Gibco BRL, Grand Island, NY). To minimize background staining, sections were blocked with normal goat serum (Vector Laboratories, Burlingame, CA) for 20 min at room temperature prior to application of the primary antibody. Blocking serum was
The continuous variables are expressed as means _+ standard errors. To compare quantitative data, we used unpaired r-tests or one-way ANOVA, with Student-Newman Keuls, depending on the number of groups. All statistics were analyzed using SPSS for windows version 6.0 (Chicago, IL), and statistical differences were considered to be significant at P < 0.05 [19].
3. Results No rats died during hyperoxia-exposure, and no rats had to be sacrificed early because of severe clinical deterioration. Lung inflammation, as assessed by MPO activities (Fig. lA), preceded evidence of pulmonary edema, as assessed by lung water contents (Fig. 1B). Lung tissue ICAMexpressions were more than doubled by 48 h of hyperoxia (Fig. 2A), whereas simultaneous plasma and lavage soluble ICAM- concentrations were not increased (Fig. 2B). In contrast, the cell pellet fractions of BAL fluids contained more than 6 times as much ICAM- than was found in the cell fractions obtained from air-breathing animals (Fig. 2C). E-Selectin expression in the lungs were more than doubled at 48 h of hyperoxia (Fig. 3A), whereas soluble E-Selectin contents were increased by 24 h (Fig. 3B). Interestingly,
P.L. Ramsay
et al.
I Toxicology
Letters
87 (i996)
157-165
161 l
lung E-Selectin decreases back to control levels at 60 h of hyperoxia-exposure, which most likely represents a dilutional effect of the severe protein leak in the lung, rather than an actual decrease in lung E-Selectin expression. Although the number of cells recovered in BAL samples from rats exposed to hyperoxia increased relative to air-breathing rats, there were no changes in the proportion of the cells that were alveolar macrophages and neutrophils (Table l), indicating that the increased content of ICAMin the cellular component of the BAL samples was not a result of a change in relative proportions of the cells recovered. These data implicate increases in ICAM- expression in one or more specific cell types, such as alveolar macrophages. Immunohistochemical analyses of the cell pellet specimens
l
i
125,
-
75
! q50 e 25
Fig. 2. Quantification of ICAM-I expression in hyperoxia-exposed rat lung tissue and in the plasma and BAL. Western blots for ICAM-I revealed that lung expression of ICAM-I was more than doubled at 48 h (2A), while soluble ICAM-I in the plasma or BAL was not different than at the same time of hyperoxia exposure (28). Lavage cell ICAM-I was × greater in cells exposed to hyperoxia for 48 h (2C). Data are expressed as means k S.E.M., n = 4 or 5 in each group. *P < 0.05 vs. 0 h or air-breathing animals. Equivalent amounts of protein were loaded for all Western blots. Fig. I. Lung injury and inflammation in the course of hyperoxia-exposure. Sprague-Dawley rats were exposed to hyperoxia-exposure for up to 60 h, and lung MPO activities (A) and extravascular lung water contents (B) were determined as measures of lung inflammation and injury, respectively. Extravascular lung water content was increased at 60 h and lung MPO activities increased at 48 h. Data are expressed as means F S.E.M., n = 6 in each group. *P -C 0.05 vs. 0 h or air-breathing animals.
revealed that ICAM- expression is almost entirely limited to alveolar macrophages (Fig. 4A and B from cells exposed to 48 h and 0 h of hyperoxia, respectively). In this figure, ICAM- is observed as brown staining in the cells (see arrow in 4A). Alveolar macrophages are identifiable by
162
P. L. Rroustr~~ PI ol.
To.\icdo,q~
their large size and the mononuclear appearance. The uniform brown staining in lymphocytes (Fig. 4C, see arrow), a cell type that expresses ICAM-I constitutively, supports the specificity of the analysis for ICAM-I. Alveolar macrophages from rats exposed to hyperoxia were 50 f 3% positive for ICAM-I, whereas I5 f 2% of the alveolar macrophages from control rats were positive (mean + S.E.M., II = 4), with 10 high power fields per sample examined (P < 0.001). The six-fold increase in ICAMcontent in lavage cell pellets and the three-fold increase in the percentage of macrophages showing protein expression further suggests a greater extent of ICAM-I expression per cell in the hyperoxic rats.
Lctrc~~\ 87 (I 9%) Table
IS7
IhJ
I
Effect of hyperoxia
on bronchoalveolar
lavage cell count
Air breathing
Hyperoxic
__~
_
Cell count (cells 181)
460 2
‘I%Neutrophils ‘X Alveolar
*P
<
I I9
905 *
7*4
macrophages
~- ..-
I IO*
7*5 86 + 4
91 i-5
0.001.
II = 4 in hyperoxic
group and 6 in air breathing
Rats were placed in either
group.
> 95 ‘% oxygen or room air for 48 h.
After anesthesia. 5 ml of iced saline was instilled into one lung. The recovered Invage was spun at IO 000 g for IO min at 4°C. and the cell pellet was resuspended in white
cell count
and
differential
I
ml of sterile salrnc. and
were obtained.
Reco\crcd
lavagc in every cahe was between 75 and 85’%#instilled volume.
4. Discussion
* 3x7
3x
ax 71
8
s
ml
0
0
24
I!6
48
60
ttWSG$XEEdtO>950/0~ Fig. 3. Lung and soluble E-Selectin Sprague of
Dawley
> 95% oxygen.
plasma
After
hyperoxia-exposure.
samples were analyzed
content,
in hyperoxia-exposed
respectively
by
lung tissue and
for E-Selectin
Western
blotting.
expression Signal
E-Selectin
was found to be more than doubled
Similarly,
plasma
E-Selectin
was increased
Data are expressed as means
*P
<
0.05
f
S.E.M.,
vs. 0 h or air-breathing
of protein
for
and lung
after 48 h (A).
by approximately
50% at 24 h of exposure and more than doubled
amounts
rats.
rats were exposed to room air or up to 60 h
at 48 h (B).
n = 6 in catch group. animals.
were loaded for all Wcstcrn
Equivalent hlol\.
The most significant finding in the present study is that the sequential increases in E-Selectin. ICAM-I, lung MPO and lung injury is consistent with the hypothesis that E-Selectin plays a significant role in initiation of hyperoxic lung inflammation and that lung inflammation contributes to lung injury. While the sequential interactions suggest a role for inflammation in hyperoxic lung injury. more definitive studies are necessary to determine the extent of the contribution. Furthermore. subtle evidence of lung injury may be cvident prior to neutrophil accumulation. indicating that there is ;I component of neutrophil independent lung injury [2]. In the animal model employed in the present studies, significant lung injury is manifested only after 60 h of exposure, and this injury is preceded by increases in lung MPO activities. The increases in lung MPO activities coincides with increases in lung ICAM-I and E-Selectin expression; however, soluble E-Selectin contents are increased as early as 24 h after exposure to hyperoxia. The increases in contents of soluble E-Selectin observed at 24 h suggest that the expression of E-Selectin on the pulmonary vascular endothelium is a significant early event in hyperoxia-exposure that leads to adherence of neutrophils. The interaction of the neutrophils with rhe endothelium leads to cleavage of E-ScIcctin, which increases measurably in the plasma. The increase in plasma E-Selectin prior to the incrcnse in lung infkunmation (MPO activities)
163
164
P.L. Ramsay
et al. 1 Toxicology
suggests that neutrophils are adhering to lung endothelium, but that neutrophil clearance equals neutrophil adherence. Another explanation is that the number of neutrophils is actually increasing but that the neutrophils have reduced cellular contents of myeloperoxidase. The data available at present suggest that the observed increases in soluble E-Selectin reflect increased cellular expression, in combination with cleavage of E-Selectin from the cell surface as a result neutrophil interaction with E-Selectin [9]. Hence, increased soluble E-Selectin may be a marker of E-Selectin-mediated lung/neutrophil interactions, which is supported by the parallel increases in lung and soluble E-Selectin preceding increases in lung MPO activities. In addition to the sequence of events suggesting a primary role for E-Selectin in hyperoxic lung injury, our findings suggest that lung inflammatory events may be detectable in plasma samples, although plasma samples may not be specific for events occurring in the lung. Despite the concern of specificity, if the same correlations with soluble adhesion molecules in the present study are expressed in humans, measuring soluble adhesion molecules in the plasma or serum may prove to be useful for detection of processes leading to lung inflammation prior to significant injury, thereby directing the rational and timely interventions with anti-inflammatory therapeutics. It is important to note that studies from our lab suggest that E-Selectin is not upregulated in hyperoxia-exposed mice, although P-selectin is upregulated 1201, suggesting that there are species differences in the adhesive mechanisms for the development of lung inflammation. However, the general sequence of events from Selectin-mediated transient adhesion to neutrophil activation and firm adhesion appears to be retained.
Letters
87 (1996) 157-165
The lack of increase in soluble ICAM- in the plasma of hyperoxia-exposed rats is consistent with an earlier report by Kojima and coworkers, in which no increases in soluble ICAM- were observed in the sera of premature infants [13]. Our findings in the BAL differ somewhat from the findings in premature neonates in that Kojima et al. found an increase in soluble ICAM- in tracheal aspirates, whereas we did not observe an increase in soluble ICAM- contents in BAL samples. The simplest interpretation of the difference between studies is that the models of lung injury examined in the two studies are fundamentally different. The study by Kojima et al. involved hyperoxic exposure and barotrauma on immature lungs. In contrast, the model we studied was one of pure oxygen toxicity in spontaneously breathing adult animals. An additional finding from the present study is that increased ICAM- expression is associated with alveolar macrophages in the cellular component of the BAL. It is not clear whether the increase ICAM- represents increased expressions of ICAM- on these cells, or whether the ICAMhas been phagocytosed by these cells. In either event, further studies are needed to determine the causes and implications of the increases in ICAM1 expressions associated with alveolar macrophages during the evolution of hyperoxic lung injury. In summary, the fact that contents of soluble E-Selectin increased in the plasma compartment prior to increases in expressions of E-Selectin in lung tissue, and that both preceded lung inflammation suggests that E-Selectin is a significant mediator of lung inflammation in hyperoxic lung injury, and further studies to determine the effect of interventions affecting E-Selectin expression or function as a means of attenuating lung inflammation and injury are warranted. In addition,
Fig. 4. Immunohistochemical analysis of lavage cells For ICAM-I. Cell pellets from lavages of rats exposed to either room air or 48 h of z 95% oxygen were placed on a slide, incubated with anti-rat ICAM-I, followed by incubation with a biotin-conjugated secondary antibody. ICAM-I signal was visualized after incubations with avidin-biotin complex, peroxidase reporter and diaminobenzidine. In section (A), alveolar macrophages from hyperoxia-exposed rats are visualized and have substantial staining for ICAM-I (see arrow), while in section (B) there are alveolar macrophages from room air-exposed rats with no staining for ICAM-I. In section (C), there is a lymphocyte, a cell type known to express ICAM-1, and in the section it is intensely stained for ICAM-I (see arrow), supporting the specificity of the method.
increases in soluble E-Selectin may prove to be valuable as an indicator of early inflammatory events, and could be used to guide early anti-inflammatory interventions in patients at high risk for development of deleterious responses to lung inflammation.
Acknowledgements This work was supported by NIH grant AI19031, NIH grant 5 P30 HD27823-04, and an AHA grant, Texas Affiliate.
References
lung injury: Biochemical. [II Smith, L.J. (1985) Hyperoxic cellular, and morphologic characterization in the mouse. J. Lab. Clin. Med. 106, 2699278. changes in pulmonary PI Crapo, J.D. (1986) Morphologic oxygen toxicity. Ann. Rev. Physiol. 48, 721-731. [31 Bressack, M.A., McMillan, D.D. and Bland, R.D. (1979) Pulmonary oxygen toxicity: increased microvascular permeability to protein in anesthetized lambs. Lymphology I?, 133 139. T.A. and Bland, R.D. ( 1982) [41 Hansen, T.N., Hazinski, Vitamin E does not prevent oxygen-induced lung injury in newborn lambs. Pediatr. Res. 16, 5833587. C.D.. Wolyniec. W.W., LaPlante, A.M., [51 Wegner, Marschman. K., Lubbe. K., Haynes, N., Rothlein, R. and Letts. G.L. (1992) Intercellular adhesion molecule-l contributes to pulmonary oxygen toxicity in mice: Role of leukocytes revisited. Lung 170. 267 -279. [61 Ramsay, P.L., Welty. SE., Bree, M.P. and Hansen, T.N. (1994) Lung injury and inflammation are reduced in hyperoxic mice by inhibition of neutrophil /12 integrin CDI la:CDI8. Pediatr. Res. 35, 396A. S.E.. Nowooryta, K. and Schmalstieg. F.C. (71 Keeney. (1995) The effect of hyperoxia on pulmonary E-Selectin and ICAM-I in neonatal rats. Pediatr. Res. 37. 337A. PI Kang. B.H.. Crapo. J.D., Wegner, CD.. Letts, L.G. and Chang. L.Y. (I 993) Intercellular adhesion molecule-l expression on the alveolar epithelium and its modification by hyperoxia. Am. J. Resp. Cell. Mol. Biol. 9, 350-355.
[91 Gearing. adhesion
A.J.H. molecules
and
Newman.
in disease.
W. (1993) Immunol.
Circulating
Today
14. 506~
512. 1.. Pigott, R.. Hughes, J., [lOI Gearing, A.J.H.. Hemingway. Rees, A.J. and Cashman, S.J. (1992) Soluble forms of vascular adhesion molecules. E-Selectin. ICAM-I. and VCAM-I: Pathological significance. Ann New York Acad. Sci. 667. 324 33 I. Groneck. P.. Reuss, D.. Gotze-Speer, B. and Speer. C.P. [I II (1993) Effects of dexamethasone on chemotactic activity and inflammatory mediators in tracheobronchial aspirates of preterm infants at risk for chronic lung disease. J. Pediatr. 122, 938 -944. B.. Oppermann. M.. Eiffert. H. LIZI Groneck. P.. Gotze-Speer, and Speer, C.P. (1995) Association of pulmonary inflammation and increased microvascular permeability during the development of bronchopulmonary dysplasia: a sequential analysis of inflammatory mediators in respiratory fluids of high-risk preterm neonates. Pediatrics 93. 712 718. T.. Sasai, M. and Kobayaski. Y. (1993) In[I31 Kojima. creased soluble ICAM-I in tracheal aspirates of infants with bronchopulmonary dysplasia. Lancet 342. IO23 I O-74. S.E.. Kuei-Meng, W. and Jay. M. (1985) Lung [I41 Goldblum, myeloperoxidase as a measure of pulmonary leukostasis in rabbits. J. Appl. Physiol. 59, 1978 1985. A.J.. Vaughan III, T.R., Brigham. K.L.. [I51 Erdmann, Woolverton. W.C. and Staub, N.C. (1975) Effect of increased vascular pressure on lung fluid balance in unanesthetized sheep. Circ. Res. 37, 271 284. J. and Beazell, J. (1965) Mea[I61 Pearce, M.L., Yamashita. surement of pulmonary edema. Circ. Rcs. 16, 482 -488. iI71 Bradford, M.M. (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72. 248. [I81 Welty. S.E., Rivera, J.L., Elliston, J.. Smith. C.V.. Zeb. T., Ballantyne, C.M.. Montgomery. C.A. and Hansen. T.N. (1993) increases in lung tissue expression of intercellular adhesion molecule-l are associated with hyperoxic lung injury and inflammation in mice. Am J. Resp. Cell. Mol. Biol. 9. 3933400. Analysis, 2nd ed. Prentice [I91 Zar. J.H. (1984) Biostatistical Hall. Englewood Cliffs, New Jersey. pp. 43 44. P.L.. Piedboeuf. B.. ct al. ( 1995) PO1 Welty. S.E., Ramsay. P-Selectin is upregulated in hyperoxic mouse lungs. Am. J. Respir. Crit. Care Med. 151. 4542.