The airway surface is covered by a fluid, the airway surface liquid, interposed between the mucous layer and the epithelium. The airway surface liquid contains proteins, secreted by different cell types, that may have pro-/anti-inflammatory or bactericidal functions or have a role in the mucociliary clearance. We have used a proteomics approach to identify the proteins secreted by an isolated in vitro model of human airway epithelium, at resting and under proinflammatory conditions, as a strategy to define the factors involved in epithelial barrier function. To this aim, we have analyzed the airway surface liquid from human bronchial epithelial cells grown as polarized monolayers in the presence and absence of inflammatory stimuli such as IL-4, IL-1β, TNF-α, and IFN-γ. Two-dimensional electrophoresis followed by mass spectrometry analysis has allowed the identification of ∼175 secreted protein spots, among which are immune-related proteins, structural proteins, an actin severer, some protease inhibitors, and a metalloproteinase. Comparisons between treated and untreated conditions have shown that the expression of several proteins was significantly modified by the different cytokines. Our results indicate that the surface epithelium is an active player in the epithelial barrier function and that inflammatory conditions may modulate protein secretion.
- human bronchial epithelial cells
- tumor necrosis factor-α
the airway surface is covered by a thin fluid layer, the airway surface liquid (ASL), which has an essential role as a barrier to protect the underlying epithelium. In fact, particles and microorganisms contained in breathed air are trapped into the mucous layer that lies over the ASL and moved toward the oropharinx by the continuous beating of cilia. For an optimal mucociliary clearance to occur, it is essential that the ASL maintains the proper depth and fluidity (40). Recently, other authors and we (12, 16, 18) have shown that interleukin (IL)-4, IL-13, interferon (IFN)-γ, and tumor necrosis factor (TNF)-α modify the ion transport properties of the cultured cells by changing the expression and activity of ion channels like the cystic fibrosis transmembrane conductance regulator, the epithelial sodium channel, and the Ca2+-dependent Cl− channel. The consequent changes in the volume and composition of ASL secretions might have considerable influence on the mucociliary clearance and fluidification of mucous secretion.
The ASL has other very important functions besides mucociliary clearance. Indeed, airway epithelial cells respond to whole bacteria, bacterial components, and proinflammatory cytokines by secreting, into the ASL, factors that control the inflammatory response, proteins that directly attack bacteria or proteins that recruit cells with bactericidal activity (21). This reaction to airway colonization by bacteria may involve macromolecules such as lactoferrin, lysozyme, immunoglobulins, β-defensins (39), and probably several still unknown proteins. The identification of factors involved in this response could be extremely important in understanding the mechanisms working in the control of infection and in inflammatory lung pathology. As stated above, the expression and activity of some epithelial ion channels change in the presence of some proinflammatory stimuli (12, 16, 18). Similarly, it is possible that some of the proteins secreted into the ASL to improve mucociliary clearance or involved in infection/inflammation control might be modulated by cytokines.
Recently, a database of proteins recovered by bronchoalveolar lavage (BAL) has been created using two-dimensional (2D) electrophoresis (31). In addition, Magi et al. (27) have used a proteomics approach to study the BAL fluid (BALF) from patients with sarcoidosis and idiopathic pulmonary fibrosis. In this way, these authors have identified an important number of proteins with statistically significant disease-related variations. Analysis of the protein content of BALF is of outstanding interest since it enables accurate diagnosis and follow up of a number of lung diseases, such as acute respiratory distress syndrome (22) and allergic lung diseases (30) among others. Nevertheless, the proteins present in BALF may originate from a broad range of sources. Indeed, beside those secreted by epithelial cells, the BALF may contain proteins leaked from serum across the surface epithelium and proteins secreted by T lymphocytes, alveolar macrophages, etc.
We have undertaken a proteomic study on an in vitro model of airway epithelium in the absence of other tissues and cells. Our purpose was to identify those proteins produced and secreted by the epithelium under physiological and pathological conditions as a strategy to define the factors involved in the barrier function of airway epithelium. To this aim, we have analyzed the ASL from human bronchial epithelial cells grown as polarized monolayers in the presence and absence of inflammatory stimuli such as IL-4, IL-1β, TNF-α, and IFN-γ. 2D gel electrophoresis followed by mass spectrometry analysis has allowed the identification of ∼175 protein spots. In addition, comparisons between treated and untreated conditions have shown that several proteins were modulated by the different cytokines.
MATERIALS AND METHODS
dl-dithiothreitol, gelatin from pig skin, Triton X-100, aprotinin, leupeptin, PMSF, pepstatin, and IL-4, were from Sigma Aldrich (St. Louis, MO). Acrylamide, Coomassie brilliant blue G-250, 2-mercaptoethanol, and N,N′-methylenebisacrylamide (Bis) were from Bio-Rad (Hercules, CA). IL-1β, TNF-α, and IFN-γ were from Boehringer. All other chemicals of analytical and electrophoretic grade were purchased from Fluka (Buchs, Switzerland).
Human bronchi were obtained from lung lobectomies from patients with diagnosis of lung cancer, pulmonary hypertension, or lung emphysema. The study was approved by the local Research Ethical Committee. Bronchial epithelial cells were detached by overnight digestion with protease XIV and cultured on flasks in a serum-free medium as previously described (17, 41). To reduce the influence of different genetic backgrounds on the abundance of proteins, cells in this study were from bronchi of four different individuals. We found no important donor-dependent differences in the proteins obtained in basal or stimulated conditions. To obtain polarized epithelia, human bronchial epithelial cells were plated at high density, in serum- and hormone-supplemented medium, on Transwell Clear permeable supports with a diameter of 24 mm (Corning Costar, Sigma-Aldrich, Milan, Italy). Experiments started 6–7 days after plating, when the transepithelial resistance was ∼1,000 Ω·cm2. The apical culture medium was removed, the apical side was washed with saline solution to remove the remaining proteins, and epithelia were maintained in an air-liquid interface, i.e., with culture medium only at the basolateral side. Under these conditions, the epithelia maintain a thin layer of liquid similar to the ASL in vivo (23). The treatment by the cytokines started 2 days later, thus on days 9 or 10 after plating, and lasted 5 days. At the end of the experiment, the apical medium was added, and the transepithelial resistance was measured to confirm epithelium integrity. Only the fluid obtained from those filters that maintained a high transepithelial resistance (>500 Ω·cm2) was considered appropriate for subsequent analysis.
Cytokine stimulation and ASL collection.
We used 10 ng/ml IL-4, 0.5 ng/ml IL-1β, 40 ng/ml TNF-α, or 100 U/ml IFN-γ to stimulate our cells, since these are the lower concentrations able to produce in 24 h a full change in ion transport characteristics of human bronchial epithelial cells (16, 18). Cytokines were added to the basolateral medium. The medium, with or without cytokines, was replaced every 24 h. Collection of ASL proteins was achieved by washing the apical side of polarized epithelium with 500 μl of a Ringer solution with the following composition: 126 mM NaCl, 0.38 mM KH2PO4, 2.13 mM K2HPO4, 1 mM MgSO4, 1 mM CaCl2, 10 mM glucose, and 24 mM NaHCO3. This was done every 24 h, leaving the cells in air-liquid interface after each collection. The first 2 days, the apical solution was discarded to eliminate any residual culture medium protein, whereas the samples of the subsequent 5 days were immediately frozen. Fluid recovery from two Transwell inserts yielded enough protein (∼60–100 μg/ml) to prepare a single analytical gel, whereas fluid from six to eight Transwells was necessary to load a preparative gel.
2D gel electrophoresis.
ASL proteins were cleaned, delipidated, and concentrated as previously reported (7, 19, 28). Aliquots of 25 μg were dissolved in the focusing solution, i.e., 7 M urea, 2 M thiourea, 4% wt/vol 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS), 0.5% ampholyte, pH 3.5–10, 5 mM tributylphosphine, 20 mM iodoacetamide, 40 mM Tris, and 0.1 mM EDTA, pH 8.8. Before isoelectric focusing, the sample was incubated in this solution for 1 h to allow proper reduction and alkylation. To prevent overalkylation during the focusing step, excess iodoacetamide was destroyed by adding an equimolar amount of dithiothreitol. Samples were separated in the first dimension on homemade soft immobilized pH gradient (IPGs) in a nonlinear pH 3.5–10 interval (4, 9). In the second dimension, proteins were separated based on their molecular weight in polyacrylamide gels having the following size: 180 × 160 × 1.5 mm. At the end of separation, proteins were visualized by a double staining procedure: first methyltrichloroacetate negative staining (10) followed by silver staining (8) for the analytical image or colloidal Coomassie (6) for preparative gels and mass spectrometry analysis. Analytical gels were loaded with 80 μg of protein, whereas preparative gels were loaded with 400 μg of protein.
Monodimension gel zymography.
The samples were cleaned, delipidated, concentrated, and protein quantified (9). Aliquots of 10 μg were used for each zymographic assay. Monodimensional gelatin zymography was performed under nonreducing conditions on 8–16% acrylamide gradient gel copolymerized with 0.1% gelatin from pig skin. SDS was then removed from gels by Triton X-100 wash. The zymograms were developed for 48 h at 37°C in TBS buffer and 5 mM CaCl2 (14). Gels were stained with Coomassie blue, and unstained areas, corresponding to zones of digestion, were visualized after destaining with 30% methanol and 7% acetic acid.
To identify the proteases, after SDS removal the zymograms were developed for 48 h at 37°C in TBS buffer plus 1) 5 mM EDTA for inhibition of metalloproteases; 2) 20 μg/ml aprotinin, 20 μg/ml leupeptin, and 5 mM PMSF for inhibition of serine and cysteine proteases; 3) 20 μg/ml pepstatin for inhibition of aspartyl proteases; or 4) 20 μg/ml bestatin for inhibition of aminopeptidases. Gels were stained with Coomassie blue and destained with methanol and acetic acid. The condition leading to disappearance of unstained areas indicated the class of proteases contained in the sample.
Second-dimension gel zymography.
Zymography on second-dimension gels was performed as follows: proteins were cleaned, delipidated, and concentrated as described above and preliminarily separated by isoelectric focusing with homemade soft IPGs in a nonlinear pH 3.5–10 (4). Aliquots of 25 μg of proteins were dissolved in 4 M urea, 4% CHAPS, and 40 mM Tris. The strips were reswollen in a solution consisting of 4 M urea, 4% CHAPS, and 1% wt/vol carrier ampholyte cocktail containing 60% of pH 4–8 and 40% of pH 3.5–10 solutions. After the run, the gel strips were equilibrated for 30 min in 4 M urea, 30% glycerol, 1% SDS, and 50 mM Tris·HCl, pH 6.8, and applied to a second-dimension gel. Second-dimension electrophoresis and successive steps were carried out as described for monodimension gels.
Samples separated by monodimension electrophoresis under nonreducing conditions, on 8–16% acrylamide gradient gel, were electrotransferred onto nitrocellulose membrane (Protean BA 83; Schleicher & Schuell BioScience, Dassel, Germany). The membranes were blocked with bovine serum albumin and then reacted with 5 μg/ml monoclonal anti-human matrix metalloproteinase MMP-2 (MAb-10752; Immunological Sciences, Rome, Italy), MMP-7 (MAb-10756; Immunological Sciences), or MMP-9 (MAb-11072; Immunological Sciences). Anti-mouse peroxidase labeled antibodies (Vector Laboratories, Burlingame, CA) were utilized as second antibodies.
Image acquisition and data analysis.
Coomassie blue, colloidal Coomassie blue, and ammoniacal silver-stained gels were digitized with a GS-800 scanner (Bio-Rad) and immunoblots with a Versa Doc 4000 (Bio-Rad). The 2D-PAGE and 1D-PAGE images were subsequently analyzed with PDQuest 7.3.1 and Quantity One software (Bio-Rad), respectively. An initial image analysis showed that the presence of cytokines produced marked differences in the density of several protein spots compared with untreated conditions (see ⇓⇓Fig. 3). The largest differences in protein levels were clearly apparent after 48 h of treatment and were maintained for at least two more days. Therefore, we always compared the effect of cytokines after 48 h. We used silver staining for the differential expression analysis (control vs. treated conditions) because the high sensitivity of ammoniacal silver allowed us to load significantly less protein per gel. This is an important point given that we are using primary human cells, for which availability is limited. However, silver staining displays a poor linear range. To overcome, in part, the nonlinearity problem, and to permit a proper quantitative image analysis, the development time of silver staining was the same for control and treated conditions and short enough to avoid saturation. Thus most of the protein level changes were in the linear range. Nevertheless, we cannot exclude that saturation in some cases had taken place and, therefore, some differences could be underrated. To allow comparison between different gels, the volume (density multiplied by the area) of each spot was calculated, the background was subtracted, and then the volume was normalized. Briefly, the relative volume of each spot was obtained by dividing the volume of every single spot by the sum of the volumes of all spots present in the gel. The resulting relative spot volume was used for comparison. An internal control was that β-γ-actin volume (spots 119–122) remained constant in all gels. Subsequently, PDQuest software compared data from two conditions and established the significance of the differential protein expression observed by means of a Student's t-test. In addition, for each condition analyzed, a synthetic average image (from 4 to 6 gels) was created by the software, representing all the spots constantly present in that condition. Quantitative differences between the experimental conditions were considered real variations when the relative volume of a spot fulfilled two conditions: 1) it changed at least twofold when comparing synthetic average images, and 2) the Student's t-test indicated that the difference was statistically significant (P < 0.05).
Mass spectrometry analysis and protein identification.
Electrophoretic spots were excised from colloidal Coomassie blue-stained gels (6), and proteins were digested as described (2). These gels were run under narrower pH ranges to permit a better separation of spots (compare Figs. 1 and 2). The different conditions used in preparative and analytical gels could, in a few cases, have limited the confidence of spot matching. Peptide mass fingerprinting (PMF) spectra were obtained using the Ettan MALDI-TOF Pro mass spectrometer (Amersham Biosciences), and mass fingerprinting database searching in the NCBInr, SWISS-PROT, and TrEMBL was carried out using ProFound (www.proteometrics.com), PeptIdent (http://www.expasy.org),and MASCOT (www.matrixscience.com) online available software. The searches were performed with a peptide mass tolerance of 50 ppm carbamylation modification of cysteine residues and allowing a single missed tryptic cleavage. Peptide sequencing by tandem mass spectrometry (MS/MS) analysis was performed to confirm the sequence of some peptides when the PMF was not clear, using a nanospary/LCQ DECA IT mass spectrometer (Thermo, West Palm Beach, FL). Database searching was carried out using TurboSEQUEST (Thermo) and MASCOT MS/MS ion search software (www.matrixscience.com).
RESULTS AND DISCUSSION
Identification of ASL proteins.
2D-gels prepared with ASL proteins recovered from the apical surface of airway epithelia showed hundreds of silver-stained spots, some of them grouped into clusters or charge trains (Fig. 1). The presence of charge trains of the same protein indicates various degrees of in vivo modifications, such as glycosylation, phosphorylation, or acetylation, which affect electrophoretic mobility (32). To identify the proteins secreted by the airway epithelium, we generated four preparative gels (see example in Fig. 2), and protein spots were excised, digested, and identified by mass spectrometry. We analyzed 175 spots (see Table 1) that resulted in being part of 48 known and 6 unknown protein groups. With the aim of classifying these proteins, we have ordered them in five functional groups, namely 1) immune-related proteins, 2) enzymes and related proteins, 3) structural proteins, 4) keratins and 5) unknown proteins (Fig. 3). Enzymes and immune-related proteins were the most represented categories, together accounting for ∼65% of all protein spots. In the group of immune-related proteins, we have included immunoglobulin receptors or binding proteins, complement components, and, interestingly, four proteins with putative antimicrobial role, lipocalin-2, intelectin, von Ebner minor salivary gland protein, and palate lung nasal epithelial clone (PLUNC). The last two secreted proteins are predicted to share significant similarity to bactericidal/permeability-increasing protein and lipopolysaccharide binding protein, two of the proteins critical to the spread of signals from gram-negative bacteria (3, 25). In addition, lipocalin-2 and intelectin may be involved in the innate immune response to bacterial infection. Usually, bacteria acquire much of the iron from the host by synthesizing siderophores that scavenge iron and transport it into the pathogen. Lipocalin-2 limits bacterial growth by sequestrating the iron-siderophore complex (15). Intelectin is also called endothelial lectin HL-1 or lactoferrin receptor. It may specifically recognize carbohydrate chains of pathogens and bacterial components containing galactofuranosyl residues, in a calcium-dependent manner, and may be involved in iron metabolism (38). Another worthy-of-note protein identified in the ASL is the actin severer gelsolin, which breaks down actin filaments. We have recently described the upregulation of gelsolin by IL-4 and in asthma and suggested a role for this protein in the airways (7). In fact, gelsolin might improve the fluidity of airway surface liquid by breaking down filamentous actin that may be released by dying cells during inflammation. Several enzyme-related proteins were found in the ASL. Of particular interest, we have identified three protease inhibitors, i.e., leukocyte elastase inhibitor (LEI), squamous cell carcinoma antigen 1 (SCCA-1), and α1-antichymotrypsin. LEI is a serine protease inhibitor that regulates the activity of neutrophil elastase, cathepsin G, and proteinase-3. In addition, it has been suggested that, after cleavage of its reactive group, it is transformed into a DNase that may digest nuclear DNA during apoptosis (37). Interestingly, a potential role for inhibitors of neutrophil elastase has been proposed for the treatment of inflammatory diseases in the airways (see Ref. 11 for review). SCCA-1 inhibits cysteine proteinases such as cathepsin S, K, L, and papain and thus seems to have a role in autoimmune diseases, bone remodeling, and epidermal homeostasis (29). α1-Antichymotrypsin is an inhibitor of two macrophage ectoenzymes that cleave promacrophage-stimulating enzyme and in this way modulates the production of active macrophage-stimulating protein (MSP). In turn, MSP induces adhesion, motility, and replication of epithelial cells and on macrophages stimulates motility, induces phagocytosis, and inhibits the upregulation of nitric oxide synthase by inflammatory stimuli (36). At the moment, it is hard to depict a clear role for these protease inhibitors. However, we know that the balance between proteases and antiproteases is essential in the regulation of epithelia remodeling and that the loss of balance might be involved in the development of inflammatory diseases in the airways (21). We have also found protease activity in the ASL and identified the enzyme responsible for this activity (see below). Some other proteins, with less clear roles in the airways, were also identified in the ASL.
Modulation of proteins secreted to the ASL by cytokines.
We have used IL-1β, TNF-α, IL-4, and IFN-γ as proinflammatory stimuli with the aim of understanding how the airways respond to infection/inflammation by modulating certain proteins. Whereas the first two cytokines are involved in the inflammatory response to bacteria, IFN-γ is implicated in the response to viral infections and IL-4 in asthma. The addition of cytokines to the basolateral side of the epithelia produced marked differences in the intensity of several protein spots. The largest differences in protein levels were clearly apparent after 48 h of treatment and were maintained for at least two more days. Side by side comparison of gels prepared with samples from cultures treated with and without cytokines revealed changes in intensity and/or mobility of some of the spots (Fig. 4). With a twofold change threshold, we found that 109 spots out of 168 (63%) had altered abundance in the presence of at least one cytokine. The number of protein spots that were upregulated in cytokine-treated epithelia was 39 (Table 2), whereas 61 spots were downregulated (Table 3). Interestingly, only a single protein, the gelsolin precursor, was upregulated by IL-4 and downregulated by the other cytokines. All other proteins, if modulated, were either up- or downregulated. This might suggest that cytokines modify the production/secretion of each protein in a single direction, either up or down.
We have noticed a pattern where many proteins (∼30%) were modulated in the same direction, increased or decreased, or resulted unchanged, in the presence of IL-1β, TNF-α, and IFN-γ. Conversely, ASL proteins from IL-4-treated cells seem to follow a different pattern (only ∼15% of proteins modulated in the same direction of any of the other cytokines). For example, the complement C3 α-chain (spot 143) resulted in downregulation by IL-4 and was unaffected by the other cytokines, or the gelsolin precursor (spots 18 to 23), which was upregulated by IL-4 and downregulated by IL-1β, TNF-α, and IFN-γ. This result is consistent with the different role played by IL-4 and T helper type 2 (Th2) cell stimulation in immune pathways. In fact, IL-1β and TNF-α act in a paracrine fashion on Th1 cells, augmenting the expression of IFN-γ, and all of them act synergistically with one another to potentiate the Th1 cell activation and to stimulate other cells (34). To allow a better comparison of the effect of cytokines on either protein type, different tables have been constructed for immune-related proteins (Table 4), enzyme-related proteins (Table 5, see last column for the function of each enzyme), and structural proteins (Table 6). It is worth noting that 6 out of 54 protein families have not been identified, and 3 of them were strongly modulated by some of the cytokines. We will continue to search the databases, and, given the increasing rate of protein identification, we are confident we will be able to reduce this category in a short time.
Analysis of the proteolytic activity in the ASL.
Among the different classes of proteins that could be found in the ASL, proteases are particularly interesting. Extracellular proteolytic activity is important in the regulation of airway remodeling under physiological and pathological conditions. MMPs are major proteolytic enzymes that are involved in extracellular matrix turnover, and one of them, MMP-9, is known to be produced by bronchial epithelial cells and has been associated with inflammatory lung diseases (33, 35). Instead, MMP-7 has an important role in the maintenance of innate immunity in organs such as the lungs and intestines where it proteolytically activates antibacterial peptides such as prodefensins (5). We hypothesized that these or other proteases might be secreted by airway epithelium in the ASL, where they could be modulated by inflammatory cytokines and could be involved in lung damage either directly or by the activation of inflammatory cells like neutrophils.
To evaluate the proteolytic activity of human bronchial ASL, we carried out monodimension gel zymography. To this aim, we used SDS-PAGE gels in the presence of casein as protease substrate. Gel staining with Coomassie blue revealed sites of proteolysis as clear bands on a dark blue background where the substrate had been degraded by the enzyme (Fig. 5). We did not find any protease activity in control ASL, whereas it was present in the ASL from epithelia treated with IL-4, IL-1β, TNF-α, and IFN-γ (Fig. 5A). To identify the involved enzyme, before staining the gels with Coomassie, we incubated them in the presence of specific inhibitors: EDTA (for MMPs); aprotinin, leupeptin, and PMSF (for serine and cysteine proteases); pepstatin (for aspartyl proteases); or bestatin (for aminopeptidases). By doing so, we have identified bands of proteolytic activity between 70 and 90 kDa with a specific and reproducible inhibition by EDTA but not by other inhibitors, thus suggesting the presence of MMPs only. This finding has been further investigated by using specific antibodies against different MMPs with the aim of identifying the enzyme responsible for the proteolytic activity. We found that only the antibody against MMP-2, and not the ones against MMP-7 and MMP-9, was able to recognize the band with proteolytic activity in the ASL from epithelia treated with cytokines (Fig. 5C). MMP-2 is synthesized as a 72-kDa inactive form and is thought to be activated proteolytically into 68-, 62-, or 45-kDa forms (13). To identify whether specific isoforms of MMP-2 were induced by the different proinflammatory cytokines, we carried out a 2D gel zymography. Interestingly, we found two spots of different molecular weight with MMP activity (Fig. 6). The presence of two spots with the same isoelectric point but different molecular weight probably indicates that the upper MMP is a precursor with intermediate activity, whereas the lower one is the full active form (26). Alternatively, it is possible that the upper MMP is bound to a second protein, perhaps to a tissue inhibitor of MMP. To be noted, all cytokines induced the same MMP proteolytic isoforms (1).
The balance between proteinase inhibitors and proteinases determines local proteolytic activity. As mentioned before, the presence of proteases in the ASL is not unexpected. Instead, their modulation by IL-4, IL-1β, TNF-α, and IFN-γ is of relevance, suggesting a role in different types of inflammatory conditions. For example, various studies have demonstrated the importance of serine proteinase inhibitors in regulating the activity of serine proteinases released by leukocytes during inflammation. However, these inhibitors may also contribute, directly or indirectly, in the defense against microorganisms, in extracellular matrix production, and in cell growth (20). The identification of MMP-2 together with the identification of protease inhibitors gives an outline of some of the elements involved in airway remodeling and in inflammation. It has been shown on a murine model of asthma that the use of tissue inhibitors of MMPs prevents allergen-induced airway inflammation (24). Our results suggest that inhibitors of MMP-2 could be important in inflammatory diseases involving release of IL-4, IL-1β, TNF-α, and IFN-γ.
In conclusion, by using a proteomic approach to study the protein content of the bronchial epithelium ASL, we have attained very valuable results. Many proteins secreted into the ASL have been identified, and it has been established that an important fraction of them is modulated by various proinflammatory stimuli. Interestingly, we have observed that many proteins are modulated in the same way by IL-1β, TNF-α, and IFN-γ, but not by IL-4. This result is consistent with the different pathways controlled by IL-4 (Th2 immune response) and by the other three cytokines (Th1 immune response). Among the various proteins identified in the ASL, there are very interesting candidates that deserve future studies. In particular, we have recently concentrated our interest in gelsolin, which probably improves the fluidity of the ASL by breaking down filamentous actin released by dying cells during inflammation (7). In addition, we have identified four proteins with a putative antibacterial role: von Ebner minor salivary gland protein, PLUNC, intelectin, and lipocalin. We anticipate that our data will be a stimulus to other researchers to study the role of such proteins in the innate immune defense of the airways. The identification of MMP-2, induced by IL-4, IL-1β, TNF-α, and IFN-γ, together with the identification of a few protease inhibitors in the ASL, gives an important indication of some of the elements involved in airway remodeling and in inflammation. In fact, MMP-2 could be, in part, responsible for airway destruction even in the absence of infection. Actually, MMP-2 might be a good target for novel therapy in inflammatory diseases involving release of IL-4, IL-1β, TNF-α, and IFN-γ.
This study was supported by a grant from the Italian Cystic Fibrosis Research Foundation, Fondo per gli investimenti nella nicera di base (FIRB) funds to Dipartimento di Pediatria, Università degli Studi di Genova (O. Zegarra-Moran), and by Renal Child Foundation, Genoa, Italy (M. Bruschi and L. Musante).
We thank Oliviero Promontorio for the supply of antibodies.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society