Vitronectin in Bronchoalveolar Lavage Fluid Is Increased in Patients with Interstitial Lung Disease 1 - 4
W. R. POHl, MAUREEN G. CONLAN, AUSTIN B. THOMPSON, RONALD F. ERTl, DEBRA J. ROMBERGER, DEANE F. MOSHER, and S. I. RENNARD
Introduction
Vitronectin is a glycoprotein found in human serum and extracellular matrix (1-5). There are two forms: a one-chain form of molecular weight 75,000 and a two-chain form with peptide chains of molecular weight 65,000 and 10,000 (6, 7). Vitronectin has the Arg-Gly-Asp amino acid sequence found in other adhesive proteins and possesses several binding sites that interact with cell surfaces and components of the extracellular matrix (6-8). By virtue of these properties, vitronectin promotes cell substrate and cell-cell interactions (2, 3, 9). In addition, vitronectin modulates both interaction of antithrombin III with heparin and the function of the terminal attack complex of the complement system (1014). These functions suggest that vitronectin may be a modulator of the processes of inflammation and repair. Interstitial lung diseases are frequently associated with the development of pulmonary fibrosis, a process in which the normal pulmonary parenchyma is replaced by a collagenous matrix (15-17). An accumulation of inflammatory cells, cell mediators, and connective tissue elements can be observed in the epithelial lining fluid recovered from the lower respiratory tract by bronchoalveolar lavage (BAL). A number of these factors have been suggested to modulate cell recruitment, attachment, and proliferation (16, 17). We hypothesized that vitronectin may also playa role in interstitial lung disease. The present study was designed, therefore, to determine if vitronectin is present in the epithelial lining fluid of the respiratory tract of normal volunteers and if levelsare increased in patients with interstitial lung diseases. Methods Subjects In order to investigate a possible role for vitronectin in the pathogenesis of interstitial lung diseases, BAL fluids were measured for vitronectin from a group of subjects with sar-
SUMMARY Vitronectin, also known as S-proteln, is a 75,OOO-dalton serum glycoprotein that has a variety of functions, including the capacity to Interact with the terminal components of the complement cascade, the coagulation system, and cell surfaces. By virtue of its ability to interact with cells, vitronectin Is capable of mediating ceil-spreading and adhesion and may also Influence cell differentiation and cell growth. To investigate the possibility that vitronectln might contribute to the pathogenesis of interstitial lung disease, vltronectin was measured in bronchoalveolar lavage fluid from patients with sareoldosis, Idiopathic pUlmonary fibrosis, and, for comparison, normal volunteers. Vltronectin was detected in lavage fluid and serum of all study SUbjects. Increased lavage fluid concentrations were found in patients with interstitial lung disease when compared with normal subjects (p < 0.005), and glucocorticoid-treated patients with Interstitial lung disease had lower vitronectin levels than did untreated patientS. Furthermore, on 50S-PAGEand Western blot analysis lavage fluid vitronectin comigrated with serum vltronectin, suggesting similar molecular size. ThUS, vltronectin is a normal constituent of the epithelial lining fluid, and lavage fluid vitronectin is similar to serum vltronectln. The increase of vitronectln concentrations in the epithelial lining fluid of patients with Interstitial lung disease suggests that vltronectin may contribute to the pathogenesis of interstitial lung disease. AM REV RESPIR DIS 1991; 143:1369-1375
coidosis (n = 22), subjects with idiopathic pulmonary fibrosis (n = 18), and, for comparison, 10 subjects with chronic bronchitis without evidence of interstitial lung disease and nine normal nonsmoking volunteers. A detailed description of these subjects follows; data are expressed as mean ± SEM. The subjects with sarcoidosis included nine men and 13 women with biopsy-proved sarcoidosis and a mean age of 47 ± 3.1 yr, of whom six were smokers. Eleven of the subjects with sarcoidosis were receiving prednisone therapy (mean daily dose, 25 ± 7 mg) at the time of bronchoscopy and, of these, three had repeat bronchoscopies while still receivingcorticosteroids. The mean FVC was 77.8 ± 1.4070 predicted, FEV, was 78.4 ± 4.1% predicted, TLC was 83.0 ± 4.5% predicted, and the single-breath diffusion capacity for carbon monoxide (DLco) was 78.3 ± 3.3% of predicted. In 18 study subjects (12 men, 6 women) idiopathic pulmonary fibrosis (IPF) was diagnosed by established criteria (15), including open lung biopsy.The mean age was 55 ± 2.9 yr, Seven of these subjects werereceiving prednisone therapy (mean daily dose, 33 ± 5.4 mg) at the time of their BAL, and 11 patients- were untreated. This group of subjects had the following values: FVC, 82.0 ± 2.3 predicted; FEV,/FVC%, 75.4 ± 3.6 observed; TLC, 68.4 ± 3.6% predicted; DLCO, 68.3 ± 4.7% predicted. The 10subjects with chronic bronchitis included seven men and three women with a
mean age of 40 ± 3.2 yr. The nine normal nonsmoking volunteers included seven men and two women, with a mean age of 35 ± 2.8 yr. The normal subjects were free of any history of lung disease and had normal chest roentgenograms and pulmonary function tests. The subjects with chronic bronchitis had a mean FVC of 88.1 ± 4.7% predicted, FEV,/FVC 70.2 ± 3.9% observed, TLC of 96.2 ± 3.7% predicted, and DLCO 92 ± 5.8% predicted. /
(Received in original form August 27, 1990and in revised form January 3, 1991) 1 From the Department of Internal Medicine, Pulmonary and Critical Care Medicine Section, University of Nebraska Medical Center, Omaha, Nebraska, the Health Science Center Medical School, University of Texas at Houston, Texas, and the Medical Science Center, University of'Wisconsin, Madison, Wisconsin. 2 Supported by Grant HL-29586 from the National Institutes of Health and by Grant MST 17 175 86201 from the Max Kade Foundation. 3 Presented in part at the 55th Annual Meeting of the American College of Chest Physicians, Boston, October 1989. 4 Correspondence and requests for reprints should be addressed to Dr. S. I. Rennard, Department of Internal Medicine, Pulmonary and Critical Care Medicine Section, University of Nebraska Medical Center, 600 South 42nd Street, Omaha, NE 68198-2465.
1369
1370
The subjects with interstitial lung disease underwent bronchoscopy after giving informed consent as part of their clinical evaluation. The subjects with chronic bronchitis and the normal subjects gave additional informed consent as part of a research study. Both consent forms had been approved by the University of Nebraska Institutional Review Board for the Protection of Human Research Subjects.
Bronchoalveolar Lavage BAL was performed as previously described (18-20). Briefly, bronchoscopy was performed under local anesthesia. The bronchoscope was advanced transorally and gently "wedged" in a segmental or subsegmental bronchus. Three different sites were lavaged by consecutively instilling and immediately retrieving with gentle suction five 20-ml aliquots of sterile 0.90/0 saline, for a total of 15aliquots. The first aliquot from each site (the bronchial sample) was processed separately from the final four aliquots (the alveolar sample). For the purposes of this report, only the results for the alveolar samples are reported. The mean recovery for the alveolar samples was 165 ± 8.4 ml (55.4%). Mucus was removed from the fluid by filtration with a single layer of nylon mesh. The number of cells in the lavage fluid wasthen determined by a hemocytometer. Cell differentials were determined (by two separate technicians whose results were averaged) for 200 cells prepared by cytocentrifugation (Cytospin; Shandon Instruments, Sewickley, PA). The lavage fluid samples were centrifuged (400 x g for 8 min) to separate the cells, and the cell-free fluids were stored at - 80° C until further analysis. Standard solutions of vitronectin were stored at - 80° C. Subsequent measurements of vitronectin demonstrated no decrease in concentration. However, repeated freezing and thawing lead to loss of vitronectin and were thus avoided.
POHL. CONLAN, THOMPSON, ERTL, ROMBERGER, MOSHER, AND RENNARD
used were peroxidase-conjugated rabbit antimouse IgG and goat antirabbit IgG (lCN, Immuno Biologicals, Lisle, IL). Flat-bottom microtiter plates (Dynatech Laboratories, Chantilly, VA) were coated with the human vitronectin at a concentration of 2.5 ug/ml, The first antibodies were used at dilutions of 1:2,000 and 1:500 for the monoclonal and polyclonal preparations, respectively. The second antibodies were used at a dilution of 1:1,000. A solution of O-phenylenediamine (10mg/loo ml PBS) and H 2 0 2 (0.0030/0) was used to develop the plates, and the absorbance of the resulting chromaphore was measured at 492 nm with a spectrophotometer (Model 2550EIA Reader; Bio-Rad, Richmond, CA). Each plate included two standard curves. Results were calculated by using the method of Rodbard (24). The ELISA tests were both sensitive to 10 ng/ml and did not detect fibronectin. In order to ensure that BAL fluids could be accurately quantified, BAL fluids were spiked with standard amounts of purified human vitronectin. In all cases, the total amount of vitronectin determined by ELISA corresponded to that present in the neat BAL fluid plus the amount spiked (data not shown). There was a good correlation (r = 0.89) between the two tests for 53 BAL samples over a wide range of concentrations. The reported data represent measurements with the monoclonal antibody.
Fibronectin Measurements Fibronectin was purified by gelatin sepharose chromatography according to the method of Engvall and Ruoslahti (25). Measurements were done by an indirect ELISA as described (21). The first antibody was goat antihuman fibronectin (1:1,000) prepared by our laboratory, and the second antibody wasperoxidaseconjugated rabbit antigoat IgG (ICN, Immuno Biologicals) (1:1,000).
Albumin and Urea Measurements Quantification of Vitronectin Levels in To allow for correction of the dilution that Lavage Fluid and Serum occurs during BAL, albumin and urea were An indirect or competitive enzyme-linked im- measured in the lavage fluids and serum munosorbent assay (ELISA) (21)was used to samples. Albumin was determined using an indirect quantify vitronectin in both serum and lavage fluid. The ELISA wasperformed as previously ELISA test, as described previously (21). Hudescribed (21, 22) with some modifications. man albumin was obtained from Sigma Human vitronectin was purified from hu- Chemical (St. Louis, MO), goat antihuman man plasma by heparin affinity chromatog- albumin from Atlantic Antibodies (Scarraphy using a modification of the method of borough, ME), and peroxidase-conjugated Yathogo and coworkers (23). Two antibody antigoat IgG from ICN Immuno Biologicals. preparations to human vitronectin wereused Vitronectin/albumin data were expressed as to develop ELiSAs. Vitronectin in 53 sam- micrograms per milligram. Urea measurements and estimation of conples for which sufficient fluid was available was quantified with both assays. One anti- centrations in epithelial lining fluid (ELF) body preparation was mouse ascitic fluid con- were done as previously described (26). ELF taining monoclonal antibody to human S- vitronectin data wereexpressedas micrograms protein (Boehringer Mannheim Biochemicals, per milliliter. Indianapolis, IN), and the other was a rabbit SDS-PAGE, Western Blotting polyclonal antiserum to human vitronectin (Calbiochem Corp., La Jolla, CA). These two In order to evaluate if the vitronectin detected antibodies wereboth used as the monoclonal in epithelial lining fluid was similar to serum guaranteed specificity, but the polyclonal vitronectin, polyacrylamide gel electropho(which we also believeto be specific) had im- resis in sodium dodecyl sulfate (SDS-PAGE) proved sensitivity. The "second" antibodies was performed according to the method of
Laemmli (27)under reducing conditions using a 7.5% separating gel and a 4% stacking gel. Lavage fluid and serum samples wereprecipitated with 90% trichloroacetic acid and washed with 99% ethanol at - 20° C. After centrifugation (14,000 X g for 10 min), the precipitated proteins were dissolved in gel buffer, and electrophoresis was performed. Proteins separated by SDS-PAGE were electroblotted onto a nitrocellulose sheet (Schleicher and Schnell, Inc., Keene, NH) using the methods of Towbin and coworkers (28). Vitronectin blotted onto the sheet was allowed to react with the mouse monoclonal antibody to human S-protein (1:500 dilution) or the polyclonal antibody, rabbit antihuman vitronectin antiserum (1:500 dilution) followed by the second antibody, peroxidase-conjugated rabbit antimouse IgG (1:1,000 dilution) or goat antirabbit IgG (1:1,000 dilution). Antibody binding was detected by incubation with diaminobenzidine and H 2 0 2 •
Statistical Analysis Data are reported as mean ± SEM. All comparisons between patient groups were made by using the non parametric Mann-Whitney test. Correlations wereperformed using Spearman's rank correlation test.
Results
BAL Fluid Results Subjects with sarcoidosis and idiopathic pulmonary fibrosis (lPF) as well as chronic bronchitis had a significantly higher total cell count than did normal subjects (P < 0.05) (table 1).The differential cellcounts demonstrated an increased proportion of lymphocytes in patients with sarcoidosis (p < 0.001 compared with that in patients with IPF or with bronchitis, p < 0.01 compared with normal subjects). The mean percentages of lymphocytesin the normal subjects in the current study are somewhat higher than in previous reports. This is due to a single outlier with 50010 lymphocytes. The normal volunteers were paid, and there may have been some tendency to deny recent viral infection. However, as the normal subjects were "normal" by all criteria available, rigor requiresinclusion of all. Importantly, exclusion of this outlier in no way changes any conclusions. The percentage of neutrophils was significantly elevated in patients with IPF but not in those with sarcoidosis when compared with normal subjects (p. < 0.05). Increased lavage fluid albumin concentrations were found in subjects with sarcoidosis (p < 0.001) and in subjects with IPF (p < 0.01)but not in those with chronic bronchitis (p > 0.3 compared with normal subjects).
VITRONECTIN IN SAL FWIO IS INCREASED IN INTERSTITIAL WNG OISEASE
1371
TABLE 1 BRONCHOALVEOLAR LAVAGE PARAMETERS IN STUDY SUBJECTS' Differential Cell Count t
Patients Sarcoid, n = 22 IPF, n = 18 Bronchitis, n = 9 Normal, n = 9
Total Cell Count (1 x 108 )
41.8 49.8 41.7 22.3
± ± ± ±
Macrophages (%)
65 70 80 82
6.3* 9.9* 12.01* 5
± ± ± ±
4,2§ 4,2 5,B 4.4
Neutroph ils (%)
Lymphocytes (%)
24.0 9.6 7.5 16.3
± ± ± ±
3.411 3.1 3.6 4
8.3 13.8 7.4 2.3
Albumin • (JJ.glm/)
± 2.5 ± 3.7. ± 5,3 ± 0,8
105 55 24 27
± ± ± ±
16" 15.2 4.5 6.2
• Ail dala are expressed as mean ± SEM. Numbers represent the different cell count as a percentage of total cell count. p < 0.05 compared with normal subjects. § p < 0.02 compared with normal subjects. II p < 0.001 compared with patients with IPF or with bronchitis. , p < 0.05 compared with normal subjects. •• p < 0.01 compared with patients with IPF and p < 0.001 compared with normal subjects and patients with bronchitis.
t
*
A. 1.1
----E
1,0
N
0.9
c
m v
"-"
W. R. POHl, MAUREEN G. CONLAN, AUSTIN B. THOMPSON, RONALD F. ERTl, DEBRA J. ROMBERGER, DEANE F. MOSHER, and S. I. RENNARD
Introduction
Vitronectin is a glycoprotein found in human serum and extracellular matrix (1-5). There are two forms: a one-chain form of molecular weight 75,000 and a two-chain form with peptide chains of molecular weight 65,000 and 10,000 (6, 7). Vitronectin has the Arg-Gly-Asp amino acid sequence found in other adhesive proteins and possesses several binding sites that interact with cell surfaces and components of the extracellular matrix (6-8). By virtue of these properties, vitronectin promotes cell substrate and cell-cell interactions (2, 3, 9). In addition, vitronectin modulates both interaction of antithrombin III with heparin and the function of the terminal attack complex of the complement system (1014). These functions suggest that vitronectin may be a modulator of the processes of inflammation and repair. Interstitial lung diseases are frequently associated with the development of pulmonary fibrosis, a process in which the normal pulmonary parenchyma is replaced by a collagenous matrix (15-17). An accumulation of inflammatory cells, cell mediators, and connective tissue elements can be observed in the epithelial lining fluid recovered from the lower respiratory tract by bronchoalveolar lavage (BAL). A number of these factors have been suggested to modulate cell recruitment, attachment, and proliferation (16, 17). We hypothesized that vitronectin may also playa role in interstitial lung disease. The present study was designed, therefore, to determine if vitronectin is present in the epithelial lining fluid of the respiratory tract of normal volunteers and if levelsare increased in patients with interstitial lung diseases. Methods Subjects In order to investigate a possible role for vitronectin in the pathogenesis of interstitial lung diseases, BAL fluids were measured for vitronectin from a group of subjects with sar-
SUMMARY Vitronectin, also known as S-proteln, is a 75,OOO-dalton serum glycoprotein that has a variety of functions, including the capacity to Interact with the terminal components of the complement cascade, the coagulation system, and cell surfaces. By virtue of its ability to interact with cells, vitronectin Is capable of mediating ceil-spreading and adhesion and may also Influence cell differentiation and cell growth. To investigate the possibility that vitronectln might contribute to the pathogenesis of interstitial lung disease, vltronectin was measured in bronchoalveolar lavage fluid from patients with sareoldosis, Idiopathic pUlmonary fibrosis, and, for comparison, normal volunteers. Vltronectin was detected in lavage fluid and serum of all study SUbjects. Increased lavage fluid concentrations were found in patients with interstitial lung disease when compared with normal subjects (p < 0.005), and glucocorticoid-treated patients with Interstitial lung disease had lower vitronectin levels than did untreated patientS. Furthermore, on 50S-PAGEand Western blot analysis lavage fluid vitronectin comigrated with serum vltronectin, suggesting similar molecular size. ThUS, vltronectin is a normal constituent of the epithelial lining fluid, and lavage fluid vitronectin is similar to serum vltronectln. The increase of vitronectln concentrations in the epithelial lining fluid of patients with Interstitial lung disease suggests that vltronectin may contribute to the pathogenesis of interstitial lung disease. AM REV RESPIR DIS 1991; 143:1369-1375
coidosis (n = 22), subjects with idiopathic pulmonary fibrosis (n = 18), and, for comparison, 10 subjects with chronic bronchitis without evidence of interstitial lung disease and nine normal nonsmoking volunteers. A detailed description of these subjects follows; data are expressed as mean ± SEM. The subjects with sarcoidosis included nine men and 13 women with biopsy-proved sarcoidosis and a mean age of 47 ± 3.1 yr, of whom six were smokers. Eleven of the subjects with sarcoidosis were receiving prednisone therapy (mean daily dose, 25 ± 7 mg) at the time of bronchoscopy and, of these, three had repeat bronchoscopies while still receivingcorticosteroids. The mean FVC was 77.8 ± 1.4070 predicted, FEV, was 78.4 ± 4.1% predicted, TLC was 83.0 ± 4.5% predicted, and the single-breath diffusion capacity for carbon monoxide (DLco) was 78.3 ± 3.3% of predicted. In 18 study subjects (12 men, 6 women) idiopathic pulmonary fibrosis (IPF) was diagnosed by established criteria (15), including open lung biopsy.The mean age was 55 ± 2.9 yr, Seven of these subjects werereceiving prednisone therapy (mean daily dose, 33 ± 5.4 mg) at the time of their BAL, and 11 patients- were untreated. This group of subjects had the following values: FVC, 82.0 ± 2.3 predicted; FEV,/FVC%, 75.4 ± 3.6 observed; TLC, 68.4 ± 3.6% predicted; DLCO, 68.3 ± 4.7% predicted. The 10subjects with chronic bronchitis included seven men and three women with a
mean age of 40 ± 3.2 yr. The nine normal nonsmoking volunteers included seven men and two women, with a mean age of 35 ± 2.8 yr. The normal subjects were free of any history of lung disease and had normal chest roentgenograms and pulmonary function tests. The subjects with chronic bronchitis had a mean FVC of 88.1 ± 4.7% predicted, FEV,/FVC 70.2 ± 3.9% observed, TLC of 96.2 ± 3.7% predicted, and DLCO 92 ± 5.8% predicted. /
(Received in original form August 27, 1990and in revised form January 3, 1991) 1 From the Department of Internal Medicine, Pulmonary and Critical Care Medicine Section, University of Nebraska Medical Center, Omaha, Nebraska, the Health Science Center Medical School, University of Texas at Houston, Texas, and the Medical Science Center, University of'Wisconsin, Madison, Wisconsin. 2 Supported by Grant HL-29586 from the National Institutes of Health and by Grant MST 17 175 86201 from the Max Kade Foundation. 3 Presented in part at the 55th Annual Meeting of the American College of Chest Physicians, Boston, October 1989. 4 Correspondence and requests for reprints should be addressed to Dr. S. I. Rennard, Department of Internal Medicine, Pulmonary and Critical Care Medicine Section, University of Nebraska Medical Center, 600 South 42nd Street, Omaha, NE 68198-2465.
1369
1370
The subjects with interstitial lung disease underwent bronchoscopy after giving informed consent as part of their clinical evaluation. The subjects with chronic bronchitis and the normal subjects gave additional informed consent as part of a research study. Both consent forms had been approved by the University of Nebraska Institutional Review Board for the Protection of Human Research Subjects.
Bronchoalveolar Lavage BAL was performed as previously described (18-20). Briefly, bronchoscopy was performed under local anesthesia. The bronchoscope was advanced transorally and gently "wedged" in a segmental or subsegmental bronchus. Three different sites were lavaged by consecutively instilling and immediately retrieving with gentle suction five 20-ml aliquots of sterile 0.90/0 saline, for a total of 15aliquots. The first aliquot from each site (the bronchial sample) was processed separately from the final four aliquots (the alveolar sample). For the purposes of this report, only the results for the alveolar samples are reported. The mean recovery for the alveolar samples was 165 ± 8.4 ml (55.4%). Mucus was removed from the fluid by filtration with a single layer of nylon mesh. The number of cells in the lavage fluid wasthen determined by a hemocytometer. Cell differentials were determined (by two separate technicians whose results were averaged) for 200 cells prepared by cytocentrifugation (Cytospin; Shandon Instruments, Sewickley, PA). The lavage fluid samples were centrifuged (400 x g for 8 min) to separate the cells, and the cell-free fluids were stored at - 80° C until further analysis. Standard solutions of vitronectin were stored at - 80° C. Subsequent measurements of vitronectin demonstrated no decrease in concentration. However, repeated freezing and thawing lead to loss of vitronectin and were thus avoided.
POHL. CONLAN, THOMPSON, ERTL, ROMBERGER, MOSHER, AND RENNARD
used were peroxidase-conjugated rabbit antimouse IgG and goat antirabbit IgG (lCN, Immuno Biologicals, Lisle, IL). Flat-bottom microtiter plates (Dynatech Laboratories, Chantilly, VA) were coated with the human vitronectin at a concentration of 2.5 ug/ml, The first antibodies were used at dilutions of 1:2,000 and 1:500 for the monoclonal and polyclonal preparations, respectively. The second antibodies were used at a dilution of 1:1,000. A solution of O-phenylenediamine (10mg/loo ml PBS) and H 2 0 2 (0.0030/0) was used to develop the plates, and the absorbance of the resulting chromaphore was measured at 492 nm with a spectrophotometer (Model 2550EIA Reader; Bio-Rad, Richmond, CA). Each plate included two standard curves. Results were calculated by using the method of Rodbard (24). The ELISA tests were both sensitive to 10 ng/ml and did not detect fibronectin. In order to ensure that BAL fluids could be accurately quantified, BAL fluids were spiked with standard amounts of purified human vitronectin. In all cases, the total amount of vitronectin determined by ELISA corresponded to that present in the neat BAL fluid plus the amount spiked (data not shown). There was a good correlation (r = 0.89) between the two tests for 53 BAL samples over a wide range of concentrations. The reported data represent measurements with the monoclonal antibody.
Fibronectin Measurements Fibronectin was purified by gelatin sepharose chromatography according to the method of Engvall and Ruoslahti (25). Measurements were done by an indirect ELISA as described (21). The first antibody was goat antihuman fibronectin (1:1,000) prepared by our laboratory, and the second antibody wasperoxidaseconjugated rabbit antigoat IgG (ICN, Immuno Biologicals) (1:1,000).
Albumin and Urea Measurements Quantification of Vitronectin Levels in To allow for correction of the dilution that Lavage Fluid and Serum occurs during BAL, albumin and urea were An indirect or competitive enzyme-linked im- measured in the lavage fluids and serum munosorbent assay (ELISA) (21)was used to samples. Albumin was determined using an indirect quantify vitronectin in both serum and lavage fluid. The ELISA wasperformed as previously ELISA test, as described previously (21). Hudescribed (21, 22) with some modifications. man albumin was obtained from Sigma Human vitronectin was purified from hu- Chemical (St. Louis, MO), goat antihuman man plasma by heparin affinity chromatog- albumin from Atlantic Antibodies (Scarraphy using a modification of the method of borough, ME), and peroxidase-conjugated Yathogo and coworkers (23). Two antibody antigoat IgG from ICN Immuno Biologicals. preparations to human vitronectin wereused Vitronectin/albumin data were expressed as to develop ELiSAs. Vitronectin in 53 sam- micrograms per milligram. Urea measurements and estimation of conples for which sufficient fluid was available was quantified with both assays. One anti- centrations in epithelial lining fluid (ELF) body preparation was mouse ascitic fluid con- were done as previously described (26). ELF taining monoclonal antibody to human S- vitronectin data wereexpressedas micrograms protein (Boehringer Mannheim Biochemicals, per milliliter. Indianapolis, IN), and the other was a rabbit SDS-PAGE, Western Blotting polyclonal antiserum to human vitronectin (Calbiochem Corp., La Jolla, CA). These two In order to evaluate if the vitronectin detected antibodies wereboth used as the monoclonal in epithelial lining fluid was similar to serum guaranteed specificity, but the polyclonal vitronectin, polyacrylamide gel electropho(which we also believeto be specific) had im- resis in sodium dodecyl sulfate (SDS-PAGE) proved sensitivity. The "second" antibodies was performed according to the method of
Laemmli (27)under reducing conditions using a 7.5% separating gel and a 4% stacking gel. Lavage fluid and serum samples wereprecipitated with 90% trichloroacetic acid and washed with 99% ethanol at - 20° C. After centrifugation (14,000 X g for 10 min), the precipitated proteins were dissolved in gel buffer, and electrophoresis was performed. Proteins separated by SDS-PAGE were electroblotted onto a nitrocellulose sheet (Schleicher and Schnell, Inc., Keene, NH) using the methods of Towbin and coworkers (28). Vitronectin blotted onto the sheet was allowed to react with the mouse monoclonal antibody to human S-protein (1:500 dilution) or the polyclonal antibody, rabbit antihuman vitronectin antiserum (1:500 dilution) followed by the second antibody, peroxidase-conjugated rabbit antimouse IgG (1:1,000 dilution) or goat antirabbit IgG (1:1,000 dilution). Antibody binding was detected by incubation with diaminobenzidine and H 2 0 2 •
Statistical Analysis Data are reported as mean ± SEM. All comparisons between patient groups were made by using the non parametric Mann-Whitney test. Correlations wereperformed using Spearman's rank correlation test.
Results
BAL Fluid Results Subjects with sarcoidosis and idiopathic pulmonary fibrosis (lPF) as well as chronic bronchitis had a significantly higher total cell count than did normal subjects (P < 0.05) (table 1).The differential cellcounts demonstrated an increased proportion of lymphocytes in patients with sarcoidosis (p < 0.001 compared with that in patients with IPF or with bronchitis, p < 0.01 compared with normal subjects). The mean percentages of lymphocytesin the normal subjects in the current study are somewhat higher than in previous reports. This is due to a single outlier with 50010 lymphocytes. The normal volunteers were paid, and there may have been some tendency to deny recent viral infection. However, as the normal subjects were "normal" by all criteria available, rigor requiresinclusion of all. Importantly, exclusion of this outlier in no way changes any conclusions. The percentage of neutrophils was significantly elevated in patients with IPF but not in those with sarcoidosis when compared with normal subjects (p. < 0.05). Increased lavage fluid albumin concentrations were found in subjects with sarcoidosis (p < 0.001) and in subjects with IPF (p < 0.01)but not in those with chronic bronchitis (p > 0.3 compared with normal subjects).
VITRONECTIN IN SAL FWIO IS INCREASED IN INTERSTITIAL WNG OISEASE
1371
TABLE 1 BRONCHOALVEOLAR LAVAGE PARAMETERS IN STUDY SUBJECTS' Differential Cell Count t
Patients Sarcoid, n = 22 IPF, n = 18 Bronchitis, n = 9 Normal, n = 9
Total Cell Count (1 x 108 )
41.8 49.8 41.7 22.3
± ± ± ±
Macrophages (%)
65 70 80 82
6.3* 9.9* 12.01* 5
± ± ± ±
4,2§ 4,2 5,B 4.4
Neutroph ils (%)
Lymphocytes (%)
24.0 9.6 7.5 16.3
± ± ± ±
3.411 3.1 3.6 4
8.3 13.8 7.4 2.3
Albumin • (JJ.glm/)
± 2.5 ± 3.7. ± 5,3 ± 0,8
105 55 24 27
± ± ± ±
16" 15.2 4.5 6.2
• Ail dala are expressed as mean ± SEM. Numbers represent the different cell count as a percentage of total cell count. p < 0.05 compared with normal subjects. § p < 0.02 compared with normal subjects. II p < 0.001 compared with patients with IPF or with bronchitis. , p < 0.05 compared with normal subjects. •• p < 0.01 compared with patients with IPF and p < 0.001 compared with normal subjects and patients with bronchitis.
t
*
A. 1.1
----E
1,0
N
0.9
c
m v
"-"
