INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Fibronectin is a large, multifunctional glycoprotein found in plasma and connective tissue, and is present in lung matrix and in bronchoalveolar lavage (BAL) fluid from asthma patients (1, 2). Fibronectin could contribute to the airway inflammatory process in asthma by increasing chemotaxis and altering the phenotype of leukocytes (3). In addition, fibronectin can recruit and cause the proliferation of fibroblasts, which can be involved in lung repair and fibrosis (1). For more than a decade, it has been recognized that subepithelial fibrosis is a consistent component of airway remodeling, which is presumed to be a mechanism for asthma severity and lack of reversibility seen in some patients (4). Deposits of fibronectin have been demonstrated in association with subepithelial fibrosis in bronchial biopsies of asthmatic subjects, but not in normal control subjects (4, 5). Increased concentrations of fibronectin in the airway could be caused by either the enhanced leakage from the plasma that accompanies inflammation or stimulation of production by local cells. The production of fibronectin by airway epithelial cells is increased in asthma compared with normal subjects (6), and BAL levels of fibronectin are greater in asthma than normal subjects (2). Histamine, which is typically increased in the airway of asthmatic subjects, has been shown to induce fibronectin secretion by airway epithelial cells and macrophages (7, 8).

One of the key questions in the pathogenesis of persistent asthma is how acute allergic airway inflammation is related to the process of airway remodeling and fibrosis. Because of the multiple potential functions of fibronectin in remodeling, we evaluated the effect of antigen-induced airway inflammation on airway levels of fibronectin as measured by ELISA. To identify the potential source of fibronectin, we used Western blot analysis to differentiate cellular (locally produced) from plasma (liver produced) forms. To accomplish these goals, bronchoscopy and segmental bronchoprovocation (SBP) with antigen and saline were performed in 17 atopic subjects. SBP was followed by BAL at 5 min and 48 h after the challenge. The acute response to antigen was evaluated by measuring BAL concentrations of histamine and the late effect was assessed by changes in cell numbers and their phenotype. BAL concentrations of fibronectin, and its potential source, were determined and comparisons made among the different sampling conditions (Saline Day 0, Antigen Day 0, Saline Day 2, and Antigen Day 2).

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Materials

All chemical reagents were purchased from Sigma Chemical Co. (St. Louis, MO) unless otherwise noted. Hanks' balanced salt solution (HBSS) was purchased from Life Technologies (Grand Island, NY). Plasma fibronectin (9), polyclonal anti-fibronectin antisera, and Lab monoclonal antibody (mAb) (10) specific for the type III domains 4 through 8 of fibronectin were isolated as described previously. IST-9 mAb (11) specific for the extra domain-A (ED-A) domain of cellular fibronectin was purchased from Accurate Chemical & Scientific Corp. (Westbury, NY). Cellular fibronectin was purified from the conditioned medium of human foreskin fibroblasts by affinity chromatography on gelatin-agarose (12).

Subject Selection

Seventeen atopic subjects were recruited for participation in these studies. Each subject had a positive skin prick test to one or more aeroallergens with a correlative history. Each subject underwent a medical history and physical examination. Histamine challenge was performed to determine nonspecific bronchial responsiveness as previously described (13). All subjects had a history of allergic rhinitis, with normal lung functions. Informed consent was obtained from each subject prior to participation. The study was approved by the University of Wisconsin-Madison Center for Health Sciences Human Subjects Committee.

Bronchoscopy, SBP, and BAL

All studies were done out of allergen season. At least 1 mo prior to bronchoscopy, a graded nebulized challenge of antigen was performed in each subject to determine the antigen dose that provoked a 20% fall in FEV₁ (Ag PD₂₀). The Ag PD₂₀ was calculated from a cumulative dose-response curve as described previously (13). Bronchoscopy and SBP were done as described previously (14, 15). Briefly, one bronchopulmonary segment was identified, and the fiberoptic bronchoscope was wedged into that segment. A sham SBP was performed by injecting 10 ml of 0.9% NaCl followed by 5 ml of air to clear the bronchoscope channel. The bronchoscope was held in a wedge position for 5 min, and BAL was performed. For BAL, three or six 40-ml aliquots of warm (37° C) 0.9% NaCl was used in each segment. The fluid was sequentially recovered by gentle hand suction. Then a second, separate segment was identified and the bronchoscope was wedged into that segment; the antigen dose (10% of the calculated Ag PD₂₀) was diluted in 10 ml of 0.9% NaCl and injected into that segment followed by 5 ml of air to clear the bronchoscope channel. The bronchoscope was again held in a wedge position for 5 min, and then BAL was performed. A second bronchoscopy was done 48 h later; at that time, the two previously challenged segments were identified and BAL was performed from each segment in a similar fashion to that performed on Day 0. In each subject the same total volume of 0.9% NaCl was used in each segment on both study days. In one individual, BAL was also performed at 7 d.

BAL Fluid Analysis

BAL fluid return from each segment was pooled and centrifuged (400 × g × 10 min) to sediment BAL cells. The supernatant was removed and kept at 70° C for later analysis. BAL cells were washed twice with HBSS containing 2% newborn calf serum. A cell count was performed using a hemocytometer and cells were adjusted to a final concentration of 2 × 10⁶ cells per milliliter. Cytocentrifuge slides were prepared, air dried, fixed in methanol, and stained (Diff-Quik Scientific Products, Chicago, IL). For differential cell counts, 300 leukocytes were enumerated and identified as lymphocytes, neutrophils, eosinophils, macrophages, or epithelial cells on the basis of staining and morphologic characteristics. Cell numbers were calculated by multiplying the total cell count by the percentage as determined by the differential cell count. BAL total protein levels were measured by a Lowry assay. Albumin concentrations were determined using a commercially available ELISA (Albuwell; Exocell, Inc., Philadelphia, PA). Quantification of histamine was performed by a radioenzymatic assay which utilizes a histamine specific n-methyl transferase to transfer a tritiated methyl group from s-adenosyl-L-methionine to histamine forming N-tele-methyl histamine (16). A series of solvent extractions was performed to isolate a radiolabeled product. The standard curve is run with each assay to allow data comparison of unknown standard. The assay has a linear detection range between 30 and 50,000 pg/ml.

Fibronectin Assay

The concentrations of fibronectin present in BAL fluids were determined using a competitive ELISA method (17). In brief, 96-well plates (Fisher Scientific, Pittsburgh, PA) were coated overnight with plasma fibronectin at 4 µg/ml in Tris-buffered saline (TBS), pH 7.4 containing 0.1% bovine serum albumin (BSA). In a separate 96-well plate, BAL fluids and several concentrations of a standard fibronectin solution were incubated with polyclonal anti-fibronectin rabbit antisera diluted 1:10,000 in TBS + 0.1% BSA overnight at 4° C. Plates coated with fibronectin were blocked with 1% BSA in TBS at room temperature for 30 min. After rinsing three times with TBS + 0.05% Tween-20, 100 µl of antigen-antibody incubations were added to plates and incubated for 2 h at room temperature. The wells were washed, and free antibody bound to the plate was detected using alkaline phosphatase-conjugated goat anti-rabbit antibody (Sigma) and alkaline phosphatase substrate 104 (Sigma). Signal was detected using an EL 340 microplate reader (Bio-Tek instruments, Winooski, VT). Concentrations of fibronectin were calculated from the standards using Delta Soft ELISA analysis software (BioMetalics, Princeton, NJ). The minimal amount of fibronectin that could be detected was 5 ng/ml.

Western Blot for Detecting Fibronectin in BAL Fluids

BAL fluid samples were concentrated by lyophilization as follows. A volume of 10 ml of freshly obtained BAL fluid was brought to 1.8 M urea, 0.6% sodium dodecyl sulfate (SDS), and 2% -mercaptoethanol (sample buffer), and then heated to 100° C for 10 min, and dialyzed against 0.1% SDS. After lyophilization, pellets were resuspended in 250 µl sample buffer (without SDS) containing 0.2 mg/ml bromophenol blue to concentrate the sample 40×. Assuming 500 ng/ml fibronectin in unconcentrated BAL fluid on average for antigen-challenged segments after 48 h (see RESULTS), 100 µl of 40× concentrated BAL fluid would contain 2 µg fibronectin. Plasma samples were diluted 1:10 in TBS-containing sample buffer with 0.2 mg/ml bromophenol blue. Because the mean concentration of fibronectin in plasma is 300 µg/ml plasma fibronectin (18), 100 µl of plasma diluted 1:10 would contain 3 µg of fibronectin. Aliquots of 100 µl of each sample were run on a 7.5% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto Protran nitrocellulose membranes (Intermountain Scientific, Bountiful, UT). The difference in the amount of BAL fluid versus plasma loaded onto the gel, expressed in terms of original volume, was 400-fold, but the two samples should contain the same approximate amount of fibronectin. Transferred proteins were visualized by ponceau S (Sigma) to ensure complete transfer, and locations of protein standards were marked with a pencil. Membranes were blocked with 5% nonfat dry milk in TBS + 0.05% Tween-20 for 1 h at room temperature, rinsed, and incubated with either IST-9 mAb (11) diluted 1:200 in TBS + 0.05% Tween-20 + 0.1% BSA, or Lab mAb (10) diluted 1:100 in TBS + 0.05% Tween-20 + 0.1% BSA overnight at 4° C. Unbound primary antibody was removed by rinsing, and membranes were incubated with peroxidase-conjugated rabbit anti-mouse antibody (Sigma) diluted 1:1,000 in TBS + 0.05% Tween-20 + 2.5% nonfat dry milk for 1 h at room temperature. Membranes were rinsed with five changes of TBS + 0.05% Tween-20, incubated with ECL chemiluminescence substrate (DuPont-NEN Research Products, Boston, MA) for 1 min at room temperature and exposed to X-OMAT film (Eastman Kodak Co., Rochester, NY) for 10 min at room temperature. Controls included purified plasma fibronectin and cellular fibronectin (9, 19). Images of the films were scanned and analyzed using NIH Image software with gel plotting macros (National Institutes of Health, Bethesda, MD). Aliquots of 30 µl of each sample were also run on a 7.5% SDS-PAGE and stained with Coomassie brilliant blue to visualize total protein.

Data Analysis

Most of the data were not normally distributed. Therefore, data are presented as medians with 25%, 75% interquartiles unless otherwise noted. Comparison between the results from the four separate BAL samples (Saline Day 0, Saline Day 2, Antigen Day 0, and Antigen Day 2) was done using analysis of variance (ANOVA) on ranks (Kruskal-Wallis) with post hoc pairwise testing (Student-Newman-Keuls method or Dunn's method when uneven sample numbers were compared). Correlations between the fibronectin concentrations and concentrations of histamine or proportions of various BAL cells were done using Spearman Rank correlation test.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Subject Characteristics

Subject characteristics are shown in Table 1. Spirometry was normal on both study days. FEV₁ percentage of predicted was 102 ± 3.5% on Day 0 (mean ± SEM) and 101 ± 3.5% on Day 2. Bronchoscopy, SBP, and BAL were tolerated well.

Cellular Profiles of BAL Fluid

The volumes of lavage fluid recovered from the two lavage segments on both days were similar and the fluid recovery averaged 75% of the injected fluid. The cell numbers are outlined in Table 2. In general, there was an increase in total cells on Day 2 which was significantly more prominent in the segment that received antigen. There was a 100-fold increase in BAL eosinophils 2 d after antigen challenge, compared with only a twofold increase in the number of macrophages, less than a threefold increase in the number of lymphocytes, and a 10-fold increase in the numbers of neutrophils. The marked increase in eosinophils was specific to the antigen-challenged segment, whereas lesser increases of the other cell types were found in saline-treated and antigen-challenged segments.

Determination of Histamine, Fibronectin, Albumin, and Total Protein in BAL Fluid

The histamine content of BAL fluid was increased 3-fold 5 min after antigen SBP when compared with the saline-treated segment (Table 3). The histamine levels in the antigen-challenged segment dropped on Day 2, but remained significantly elevated compared with the saline segments. On Day 2, there were significant increases in BAL albumin and total protein that were more marked in the antigen-challenged segment (Table 3). A modest increase in albumin and total protein was seen in the saline-challenged segment on Day 2 (Table 3). This has been noted in several previous studies (14). A statistically significant increase in fibronectin concentration was seen on Day 2 in both segments. This was significantly higher in the antigen-challenged segment compared with the saline-challenged segment (Table 3). Interestingly, the concentratons of fibronectin in the 48-h antigen-challenged segments correlated with the levels of histamine in these segments 5 min after the challenge (r = 0.53, p < 0.05).

The increase in fibronectin on Day 2 was greater than the increase in total protein, suggesting local synthesis of fibronectin as opposed to vascular leakage. The normal concentration of fibronectin in plasma is 300 µg/ml (18), or about 0.4% of the total proteins. Fibronectin represented 0.05 to 0.06% of the protein in BAL fluid collected immediately after antigen challenge. The percentage increased to 1.1% in the saline segments and 1.7% in the antigen-challenged segment. Protein staining of SDS-PAGE profiles of concentrated BAL revealed an increase in many plasma proteins, including the 68 kD albumin band (Figure 1A). Transudates (or exudates) would result in similar weight-to-weight ratios in BAL fluid as in blood plasma. Indeed, when amounts of albumin in samples obtained 48 h postchallenge were compared with total protein, the values were similar to what would be expected in blood (Figure 2A), assuming 7 mg/ml total protein, and 4.5 mg/ml albumin (20). Fibronectin concentrations, in contrast, were greater than would be expected from a transudate alone (Figure 2B).

View larger version (43K): [in this window] [in a new window]   Figure 1.   BAL fluid contains cellular fibronectin. SDS treated and reduced plasma and BAL fluids from one patient taken at Day 0, Day 2, or Day 7 were separated by SDS-PAGE and analyzed by protein staining and Western blot analysis. (A) Coomassie-stained gel of concentrated BAL fluid or diluted plasma. (B) ED-A+ fibronectin detected by IST-9-probed Western blot of BAL fluid or plasma. Film was exposed for 10 min. (C ) Total fibronectin detected using Lab mAb-probed Western blot of BAL fluid or plasma. Film was exposed for 10 min. M, high-molecular-weight markers, including plasma fibronectin, with weights indicated; C, 1 µg cellular fibronectin; D0, 2, 7, samples obtained on Day 0, Day 2, or Day 7; P, plasma; S, BAL from saline-treated segment; and A, BAL from antigen-treated segment. Arrows indicate the mobility of plasma fibronectin in the size marker mix.

View larger version (15K): [in this window] [in a new window]   Figure 2.   Ratios of total protein and albumin or fibronectin in BAL fluid compared with the ratio found in blood. The concentrations of (A) total protein and albumin or (B) total protein and fibronectin are shown for each condition tested. The line represents the expected weight-to-weight ratio found in blood. Points on or above the line indicate leakage from the plasma, and points below the line indicate local production.

Identifying the Splice Variant of Fibronectin by Western Blot Analysis

Fibronectin exists as a number of different splice variants. During embryogenesis (21) and wounding (22), variants containing the ED-A and ED-B type III modules are expressed more than in resting tissue. Blood plasma and extracts of normal lung contain largely fibronectins which lack ED-A and ED-B (23, 24). We used Western blot analysis with an ED-A-specific mAb on BAL fluids of six of the 17 patients (Figures 1 and 3). The cellular ED-A+ form of fibronectin was detected in concentrated BAL fluid samples; an example from one patient is shown in Figure 1. The fibronectin was intact, inasmuch as smaller bands were not detected either on Western blot of ED-A+ fibronectin or total fibronectin (Figures 1B and 1C) or when a polyclonal antibody was used to detect total fibronectin (data not shown). When companion plasma samples were analyzed, no ED-A+ fibronectin was detected (Figure 1B). The amounts of fibronectin present in BAL fluid from the antigen-challenged segment and saline-challenged segments were compared for six patients. Scanned blots were analyzed using NIH Image software to determine the optical density of the resulting bands. When expressed as a ratio of signal in antigen samples over saline samples, a ratio greater than 1 indicates more fibronectin in the antigen-challenged segment. When compared in this way, the ratio of antigen:saline on Day 0 was not greater than 1 on either blot (Figure 3). In contrast, the ratio of antigen:saline on Day 2 was greater than 1 in five of the six patients tested (Figure 3).

View larger version (15K): [in this window] [in a new window]   Figure 3.   Comparison of fibronectin concentrations in antigen-challenged segments and saline-challenged segments by Western blot analysis. SDS treated and reduced BAL fluids from six patients taken at Day 0 or Day 2 were separated by SDS-PAGE and analyzed by Western blot analysis. Total fibronectin and cellular fibronectin were detected using IST-9 and Lab mAb as described in Figure 1. Optical density units determined using NIH Image were compared by dividing units from antigen-challenged segments by units from saline-challenged segments. Data are expressed as the mean ± SEM.

Correlation between Concentrations of Fibronectin and Cell Numbers 2 d after Antigen Challenge

The concentrations of fibronectin correlated with the total numbers of cells present in BAL fluid 2 d after antigen challenge (Table 4). The numbers of macrophages and neutrophils correlated with levels of fibronectin, although numbers of lymphocytes did not (Table 4). The strongest correlation was observed between fibronectin concentrations and numbers of eosinophils present in the BAL fluid obtained 2 d after the antigen challenge (Table 4, scattergram shown in Figure 4).

View larger version (14K): [in this window] [in a new window]   Figure 4.   The correlation between fibronectin concentrations and eosinophil numbers in BAL fluid 48 h after antigen SBP. A significant correlation was noted as described in Table 4 (Spearman rank correlation test, n = 17).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In this study, we demonstrated increased amounts of fibronectin in BAL fluid 48 h after SBP. Fibronectin concentrations were higher in BAL in the allergen-challenged segments and correlated with cellular markers of airway inflammation as reflected by the total cells and numbers of eosinophils, neutrophils, and alveolar macrophages. Western blot analysis showed that ED-A+ cellular fibronectin was present in BAL samples, with the highest amount found in antigen-challenged segments after 2 d, whereas no ED-A+ fibronectin was detected in plasma. Therefore, the source of fibronectin is likely related to local production by airspace cells with an additional contribution from microvascular leakage associated with local inflammation. This is the first study to demonstrate antigen- induced, in vivo production of fibronectin in association with cellular inflammation in the airways of allergic human subjects.

It has long been appreciated that fibronectin is increased in the airway of patients with interstitial lung disease (25) and asthma (2). Patients with hypersensitivity pneumonitis have increased concentratons of fibronectin and vitronectin in BAL fluid compared with normals (26). Interestingly, hypersensitivity pneumonitis patients with recent exposure to the offending antigen ( 4 d) had significantly higher fibronectin levels compared with those with more remote exposure (5 to 9 d) (26). Recently, Chakir and colleagues performed immunohistochemical staining of bronchial biopsies obtained from normal, asthmatic, and rhinitic subjects (5); they found no fibronectin staining in normals, whereas every asthma subject had positive fibronectin staining. Five of the eight rhinitis subjects had positive staining. Increased subepithelial expression of fibronectin in asthma subjects with thickened subepithelial bands has also been described by Roche and colleagues (4).

Concentrations of fibronectin in BAL fluid from our subjects were greater than would be expected if the fibronectin diffused from the circulation as a result of increased vascular permeability (Figure 2B). When concentrations of fibronectin were compared with concentrations of total protein in the lavage fluid, the levels of fibronectin were greater than the weight-to-weight ratio of fibronectin to total protein found in blood. This was not the case for the concentrations of albumin, which could only arise from increased vascular permeability (Figure 2A). The form of fibronectin located in normal adult lung is the type secreted by hepatic cells into the circulation (plasma fibronectin) (1). Splicing in most cultured cells results in forms containing an extra domain between the type III-11 and type III-12 repeats, known as ED-A (27). The ED-A region is lacking in liver fibronectin message but is present in the message expressed in fibroblasts, thus it is used as a marker for cellular expression of fibronectin as opposed to hepatic, plasma fibronectin (24). Because the concentrations of fibronectin compared with total protein suggested local production, we used a mAb specific for the ED-A region of fibronectin, IST-9 (11), to determine the splice variant found in BAL after antigen challenge (Figure 1B). IST-9 specifically identified purified cellular fibronectin and failed to recognize purified plasma fibronectin, thus confirming the specificity of the mAb for cellular fibronectin. BAL fluid contained bands recognized by IST-9. Samples from segments lavaged 2 d after antigen challenge contained more cellular and total fibronectin than companion saline segments in the majority of the patients tested (Figure 3). This indicates that although the increase of fibronectin in the BAL fluid arises in part from plasma fluid leakage, a considerable component of the increase is due to local production.

We observed a statistically significant correlation between eosinophil, neutrophils, and macrophages and concentrations of fibronectin in this study (Table 4). Although it is not possible from these studies to determine the source of fibronectin in the BAL fluid, it is most likely derived by local synthesis by cells within the lung tissue or within the airway space. Alveolar macrophages isolated from BAL fluids from patients with interstitial lung diseases or asthma produce greater amounts of fibronectin than cells isolated from normal patients (28, 29). Incubation of normal alveolar macrophages with histamine for 24 h results in a significant increase in production of fibronectin (7). The increase in production is due to de novo protein synthesis. Epithelial cells obtained by bronchial brushing from subjects with asthma produce greater amounts of fibronectin than those obtained from normal control subjects (6). Fibronectin released from injured bronchoepithelial cells has been presumed to have a role in epithelial repair following injury. Histamine increases fibronectin production by normal bronchoepithelial cells in a time- and dose-dependent fashion (8). Whereas we have not established the specific cell type or types responsible for the increased fibronectin levels, it is possible that the increase in BAL histamine levels 5 min after antigen challenge (Table 3) contributed to the stimulation of cells within the airway leading to increased concentrations of fibronectin in the BAL fluid (7, 8).

Within the airway, fibronectin could contribute to disease severity by altering the phenotype of resident leukocytes, leading to increases in cell survival and recruitment. Fibronectin can enhance T-cell survival and can serve a costimulatory role in causing cytokine secretion ([30] and Jarjour and coworkers, unpublished observations). Fibronectin is chemotactic for eosinophils and can alter their phenotypic state by increasing oxidative metabolism, leukotriene C₄ production, and release of granule proteins (31). The cellular form of fibronectin increases survival of eosinophils, whereas plasma fibronectin is not as effective (32). We have recently shown that eosinophil survival is enhanced by cellular fibronectin present in the solution phase, such as the fibronectin present in BAL fluid (Meerschaert and coworkers, unpublished observations). It cannot be determined if the correlative relationship between numbers of leukocytes and concentrations of fibronectin is a causal relationship of increased fibronectin leading to increased leukocytes or increased leukocytes leading to increased fibronectin production, but it is likely that both scenarios are involved.

Extending our observations to the pathogenesis of airway remodeling in asthma must be done with caution. The model used in this study employed local airway challenge with a relatively large dose of relevant antigen in a group of atopic individuals. Our studies to date have not included detailed assessment of the time course for initiation and resolution of this enhanced fibronectin generation, nor can we state what is the cell source for fibronectin in the airway. Determining the factors responsible for initiating and modulating fibronectin release and the differences, if any, between allergic subjects with asthma and those without asthma should contribute significantly to our understanding of this disease and factors associated with its persistence and severity.

In summary, the concentrations of fibronectin increased after antigen challenge, and these values correlated with increased inflammatory cell presence in the airway. At least part of the fibronectin in the airway is the cellular form, containing the ED-A region. This finding suggests that antigen challenge resulted in local production of fibronectin by airway cells. Increased generation of fibronectin in the airway could be a key factor in promoting airway remodeling and disease severity in allergic subjects after antigen exposure.