Differential Proteomic Analysis of Bronchoalveolar Lavage Fluid in Asthmatics following Segmental Antigen Challenge

EXPERIMENTAL PROCEDURES

Chemicals and Reagents—

A multiple affinity protein removal column (4.6 × 50 mm) and corresponding buffer solutions were purchased from Agilent (Wilmington, DE). The affinity column was used to remove six abundant proteins (serum albumin, IgG, IgA, haptoglobin, transferrin, and antitrypsin) using affinity-purified polyclonal antibodies packed into an LC column. Molecular weight filters (Microcon; molecular weight cutoff, 3,000) were purchased from Millipore (Bedford, MA). A 10% Tricine SDS-PAGE gel was obtained from Invitrogen. Sequencing-grade modified trypsin was from Promega (Madison, WI). ELISA assay kits for MMP-9 were purchased from R&D Systems (Minneapolis, MN). All other chemicals were purchased from Sigma and used without further purification.

Study Design—

Four patients with mild asthma and three healthy control subjects were recruited for the study at The Cleveland Clinic (Table I summarizes the characteristics of the study population). Asthma was defined by the National Institutes of Health guidelines including episodic respiratory symptoms, reversible airflow obstruction (documentation of variability of FEV₁ by 12% and 200 ml, either spontaneously or after inhaled albuterol), or a positive methacholine challenge test (34). All subjects underwent allergy prick skin testing with a standard panel of 10 aeroallergens (Bayer, Spokane, WA), normal saline solution (negative control), and histamine (positive control). A positive reaction was defined as a wheal ≥3 mm in diameter. The prick needle was a two-pronged lancet from Allergy Laboratories of Ohio, Inc. Skin tests were read after 15 min. Allergic asthmatic patients had at least two positive skin tests with one of them as ragweed or timothy grass. Healthy controls were non-smokers and skin test-negative, and they had a negative methacholine provocation test (according to the method of Chai et al. (35), decrease in FEV₁, <20% with maximum concentration of methacholine of 25 mg/ml over five stages). All subjects had a base-line FEV₁ >50% predicted, no asthma exacerbation, or respiratory tract infection within 1 month.

At the initial screening visit, subjects were classified as allergic asthmatics or non-allergic healthy controls. Each subject underwent a medical history, physical examination, and pulmonary function testing. Allergic asthmatic patients underwent whole-lung aerosol allergen challenge with ragweed or grass to demonstrate antigen responsiveness and to determine the provocative dose that caused a 20% reduction in FEV₁ (antigen PD₂₀). A stock solution of allergen (10,000 protein nitrogen units (PNU)/ml, Greer Laboratories, Lenoir, NC) was diluted with saline solution to produce six concentrations (1, 3.16, 31.6, 100, 1,000, and 3,162 PNU/ml). The allergen preparations at the concentrations used were all below the USP acceptable limit of endotoxin for injection (0.25 endotoxin units/ml) as determined by the limulus amebocyte lysate assay. The patient took 5 breaths at each concentration per the method of Chai et al. (35) through a Rosenthal-French dosimeter. The antigen PD₂₀ in PNU was determined. The dose used for segmental allergen challenge was 10% antigen PD₂₀ diluted into 10 ml of saline solution into each segment. The healthy non-allergic controls were challenged with 100 PNU allergen. At least 4 weeks (see Table I, column “Days after aerosol challenge”) after the aerosol challenge, a bronchoscopy was performed to obtain a base-line BAL. Three 50-ml aliquots of normal saline solution warmed to 37 °C were instilled in the lingula and withdrawn by gentle hand aspiration. Processing of BAL has been described previously (36). Allergen (in 10 ml of saline solution) was instilled into two subsegments of the right middle lobe. Subjects underwent a second bronchoscopy 24 h later to obtain a BAL of 150 ml from each of the two allergen-challenged subsegments. The BALFs thus obtained were filtered, and cells were separated by centrifugation at 400 × g for 10 min at 4 °C. The BALF supernatants were further concentrated with Centricon (molecular weight cutoff, ∼3,000) and were stored at −80 °C. Because the amount of BALF proteins is highly variable among individuals and can be dramatically different before and after SAC due to the enhanced airway leakage (8), for quantitative proteomic analysis, we chose to normalize the samples based on the apparent epithelial lining fluid (per milliliter of epithelial lining fluid) with urea as a dilution marker during BAL collection (4, 10). This study was approved by the institutional review board of the Cleveland Clinic Foundation, and informed consent was obtained from all subjects.

Depletion of Abundant Proteins by Immunoaffinity Chromatography—

All chromatographic fractionations were performed according to the manufacturer’s instructions at ambient temperature on an Agilent 1100 HPLC system with automated sample injector and fraction collector both set at 4 °C. Briefly the BALF samples from asthmatic patients and healthy controls were normalized to the equivalent volume of initial BALF volume and then diluted accordingly to a final volume of 250 μl with sample loading buffer (buffer A) of the Agilent depletion system. The samples were injected automatically onto the affinity column in buffer A at a flow rate of 0.25 ml/min for 9 min. The proteins bound to the column were eluted with buffer B at a flow rate of 1.0 ml/min for 3.5 min, and the column was regenerated by equilibrating with buffer A for 7.5 min with a total run cycle of 20 min. The depleted BALF (flow-through) and proteins bound to the column (column retentate) were collected at 1.75–5 min and 10.25–12 min, respectively.

SDS-PAGE Separation and In-gel Digestion—

The collected fractions were concentrated in Microcon centrifugal concentrators (molecular weight cutoff, 3,000) and were buffer-exchanged into ∼30 μl of solution containing 50 mm Tris-HCl and 0.1% SDS at pH 8.5. Proteins were reduced with 20 mm DTT at 60 °C for 1 h and alkylated with 50 mm iodoacetamide at ambient temperature in the dark for 30 min. The samples were mixed with 10× SDS sample loading buffer prior to SDS-PAGE separation in a 10% Tricine minigel. The gel was stained using colloidal Coomassie Blue. The whole gel lane was cut evenly into 20 slices. Each slice was minced into 1 × 1-mm size pieces and subsequently subjected to in-gel digestion with modified trypsin (0.53 μg/gel slice) in a digestion robot (DigestPro, ABIMED, Analysen-Technik GmbH, Langenfeld, Germany). Finally the tryptic digests were concentrated in a SpeedVac to a final volume of ∼40 μl before MS analysis.

Nano-LC-MS/MS Analysis—

Tryptic digests were analyzed by an automated nano-LC-/MS/MS system using a Famos autosampler (LC Packings, San Francisco, CA) and an Agilent 1100 HPLC binary pump coupled to an LTQ ion trap mass spectrometer equipped with a nanospray ionization source (ThermoFinnigan, San Jose, CA). 10 μl of the digest solution were injected onto a reversed-phase PicoFrit column packed with Magic C₁₈ (5-μm particle, 200-Å pore size, 75 μm × 12 cm; New Objective, Woburn, MA). Peptides were eluted from the column with a 90-min linear gradient from 2 to 55% B at a flow rate of 0.2 μl/min (mobile phase A: 0.1% formic acid aqueous solution; mobile phase B: 90% acetonitrile and 0.1% formic acid). The spray voltage was 1.8 kV, the heated capillary temperature was maintained at 180 °C, and the collision energy for MS/MS was set at 35 units. Automated data-dependent MS analysis was carried out using the dynamic exclusion feature built into the MS acquisition software (Xcalibur 1.3, ThermoFinnigan). Each MS full scan (m/z 350–2,000) was followed by MS/MS scans of the first three most intense peaks detected in the prior MS scan to obtain as many CID spectra as possible.

Protein Identification and Semiquantitation—

Proteins from each BALF sample were identified using SpectrumMill (version 3.1, Agilent) by searching against the mammalian subset of the National Center for Biotechnology Information (NCBI) non-redundant protein database (updated as of March 31, 2004). Search parameters included a static modification on cysteine residues (carbamidomethylation), a 50% minimum matched peak intensity, ±2.5-Da tolerance on precursor ions and ±0.7-Da tolerance on product ions, one missed tryptic cleavage, and ESI-ion trap scoring parameters as defined by the searching algorithm. To address the database redundancy issue, the proteins that share common peptides were grouped together and displayed as a single protein group in “Protein Centric Columns” mode as described in the SpectrumMill search engine. Within any given protein group, the human protein with the highest score was selected as the most likely correct search result. If the scores were the same, the proteins were listed in an alphabetic order by the search software, and the human proteins listed on top were selected. In a few cases where human proteins are not found in a protein group, then no protein was selected from the group. All of the database search results were further validated by applying the designated protein and peptide scores as well as the following user-defined criteria: 1) protein validation mode: protein score, >18; peptide scored percent intensity, >70% for all charge states; peptide score, >7 for peptide charge +1, >8 for peptide charge +2, and >9 for peptide charges ≥+3; 2) peptide validation mode: peptide scored percent intensity, >70%; peptide score, >13 for all charge states. In short, a protein can be considered positively identified if either multiple spectra of moderate quality or better or at least one spectrum of high quality was obtained. The searching criteria used here would result in a false positive rate of <4% as justified by Liao et al. (16). To estimate the protein relative abundance, the peak areas of the extracted ion chromatograms for each peptide precursor ion in the full scan (total peptide ion current) was calculated in the region ±1.4 m/z and ±75 scans using SpectrumMill. As data acquisition was so fast (0.4 s/spectrum average), there are always enough data points to ensure accurate measurement of chromatographic peak areas of the peptides. The variation for protein quantitation using such a method has been shown to be about 20% by separate groups when using protein standards (19–21). The abundance of an individual protein was then calculated as the sum of the total ion current measured for all peptide precursor ions derived from that protein. Thus the relative concentration of each protein was determined by comparing total MS intensities of all identified peptides from that protein in one sample and those from the other samples. ELISA assay of MMP-9 in each sample was also performed according to the protocol recommended by the vendor.

Functional Categorization of BALF Proteins—

The annotation of protein cellular localization and biological function was performed using its Gene Ontology (GO) biological process terms (17, 37). Among the 1,592 proteins identified from NCBI non-redundant database, about two-thirds have Swiss-Prot records of which 63% had GO process terms associated with them. However, more than 95% of the differentially expressed proteins had corresponding GO process terms.

RESULTS

Affinity Depletion of Abundant Proteins—

BALF contains an abundance of serum proteins permeated through the air-blood barrier. This is especially true in allergen-challenged samples due to SAC-induced plasma protein leakage (8). The existence of highly abundant serum proteins masks the detection of physiologically relevant proteins that are usually present at very low levels. An affinity column with mixed polyclonal antibodies was used to remove serum albumin, IgG, IgA, transferrin, haptoglobin, and antitrypsin, which represent the six most abundant proteins in human plasma. Fig. 1 shows the SDS-PAGE images of proteins in an asthmatic BALF (patient MM) before (Lane A) and after protein depletion (Lane B) as well as proteins bound to the affinity column (column retentate) (Lane C). The undepleted BALF (Lane A) is dominated by HSA and different IgGs. After depletion, about 90% of the total protein mass was removed, and the mixture was greatly simplified (Lane B). Accordingly the dynamic range of the protein mixture was reduced roughly by 2 orders of magnitude. In-gel digestion followed by LC-MS/MS analysis of undepleted BALF proteins (entire Lane A) resulted in the identification of 441 distinct proteins of which 281 were identified from two or more peptides and 56 of them were immunoglobulins. Whereas a total of 889 distinct proteins were identified from the depleted BALF (entire Lane B) with 686 proteins identified by two or more peptides and 10 of them as immunoglobulins. The efficiency of the depletion is also reflected in the significant reduction of the total peptide ion intensity and the abundance rank of the six target proteins in the depleted BALF sample compared with undepleted sample and column retentate (Table II). From the comparison of total peptide ion intensities of column flow-through and retentate, we estimated that ∼90% of the IgA and haptoglobin were depleted, and the removal efficiency for the other four proteins was greater than 95%. The specificity of the depletion column was evaluated by the identification of in-gel digested proteins bound to the column. Among a total of 118 nonspecifically bound proteins identified from the column, most of them were also found in flow-through with much higher abundance. There were many protein bands below the major HSA spot at 66 kDa, but they were mostly identified as HSA degradation products.

Protein Profiling in BALFs of Asthmatics and Healthy Control Subjects—

Analysis of BALF samples by LC-MS/MS in combination with database searching using SpectrumMill against NCBI non-redundant database resulted in long lists of proteins identified from each sample. In addition, the information about the number of peptides identified as well as the total peptide MS ion intensities for each protein in each sample was also generated by SpectrumMill. The most noticeable similarity between the patterns of protein expression across different samples is that, as expected, plasma proteins are always present as predominant components. The majority of the proteins identified were represented across all samples with little intensity or abundance change with a subset of them changing in either asthmatic samples or challenged asthmatic samples. The basis for the selection of interesting candidates that could be potential pathogenic mediators or biomarkers for asthma (see below for more detailed selection criteria) are the differences in total peptide ion intensity of a particular protein as well as the frequency of its appearance between asthmatics and healthy subjects or between asthmatics before and after challenge. Altogether a total of 1,592 distinct proteins representing a wide spectrum of functional classes was identified with high confidence in 14 discrete BALFs from four asthmatic patients and three healthy subjects. The majority of these proteins on the list have not been previously reported in BALF (Supplemental Table I).

Differential Expression of BALF Proteins—

To evaluate the relative changes in the protein expression responding to SAC, the protein levels before and 24 h after SAC were compared. Protein expression levels from the asthma group were also compared with those from the control group to find proteins that are specific to asthmatics. Only proteins identified from three or more unique peptides were qualified for the selection. Proteins were considered to be differentially expressed if either of the following three criteria was met: 1) the protein displays more than 5-fold change in three or more asthmatics and in no more than one healthy subject before and after SAC; 2) the protein has more than 5-fold change in two or more asthmatics and in no healthy subject before and after SAC; 3) expression in asthmatics is at least 100-fold higher than that in healthy controls. For example, calgranulin B was detected in all four asthmatics and three control samples. Its expression in asthmatics was significantly higher after SAC (27-, 52-, 1-, and 6-fold increase, respectively) but to a much lesser extent in healthy controls (0.6-, 5-, and 4-fold, respectively). In another example, eosinophil granule major basic protein, eosinophil-derived neurotoxin, and eosinophil cationic protein were selected because they were all detected in two challenged asthmatics (MM and HB) that had good response to SAC but in none of the healthy controls. By comparing the protein profiles of four asthmatics and three healthy controls with the above criteria, about 160 proteins were selected that were differentially expressed in challenged asthmatic patients. These proteins are summarized in Table III based on their cellular functions and localizations. Most of the proteins in Table III exhibited elevated expression after SAC except a small group of proteins that were down-regulated after challenge. The differentially expressed proteins listed here represented a wide range of biological categories including serum proteins such as apolipoprotein, metabolic enzymes such as protein prostaglandin-H₂ D-isomerase, calcium-binding proteins such as calgranulin C, cytokine/chemokines such as high mobility group protein 1 (HMG-1) and CCL17, matrix metalloproteinase such as MMP-9 and its inhibitor tissue inhibitors of metalloproteinase 1 (TIMP-1), cellular signaling proteins such as neutrophil gelatinase-associated lipocalin, and proteins involved in lung remodeling such as hepatoma-derived growth factor and its inhibitor as well as a number of proteins whose properties and functions have not been well characterized.

To evaluate the accuracy of relative quantitation of proteins directly based on their total MS signals, the concentration of MMP-9 in each sample was also measured by ELISA and was plotted against the total MS ion intensities of MMP-9 from the same samples. As shown in Fig. 2, a relatively good linear correlation (R² = 0.9827) was observed between the concentration of MMP-9 measured by ELISA (x axis) and the total MS ion intensity of MMP-9 measure by MS (y axis), reflecting the capability of the MS-based method for gross relative protein quantitation in the global analysis of complex protein mixtures such as the BALF proteome.

Gene Ontology Classification—

Proteins identified in BALF were categorized by GO to define their molecular functions and cellular localization as well as biological processes. For this purpose, all 1,592 proteins from the entire 14 experimental data sets as well as the differentially expressed proteins were summarized separately as an overview (Fig. 3, A and B) to show the distribution of BALF proteins using the GO indexing system. The total protein complement identified in the BALF (Fig. 3A) covered a broad range of protein functional classes associated with diverse biological processes such as proteolysis, immune response, inflammatory responses, cell adhesion, cell mobility, cell proliferation, metabolism, and signal transduction. In addition to extracellular proteins, a number of membrane and intracellular proteins were also identified. However, the differentially expressed proteins (Fig. 3B) are primarily extracellular/secreted proteins that are involved in various aspects of immune responses and airway damage and protection such as inflammation, proteolysis, response to oxidative stress, apoptosis, or defense response to bacteria.

DISCUSSION

Controlled provocation of airway inflammation using bronchoscopy and segmental allergen challenge is a powerful tool for the in vivo study of airway inflammation. The unique features and conclusions from our study are as follows. 1) This is the first comprehensive differential proteomic analysis of bronchoalveolar lavage fluid from healthy subjects as well as mild asthmatic patients at base line and 24 h after segmental allergen challenge. 2) Over 1,500 distinct proteins were identified of which about 10% displayed significant up-regulation specific to the asthmatic patients after SAC (Supplemental Table I). 3) A large proportion of these proteins and their expression changes were identified for the first time from BALF; many of these have biological relevance to the pathophysiology of asthma. In this study, we utilized a high throughput proteomic approach to study the differential expression of proteins in BALFs. In contrast to traditional biochemical approaches that can only study one or a few particular protein species, proteomics is able to analyze thousands of proteins and thus provides comprehensive BALF proteome information that is essential in understanding the molecular mechanisms underlying asthma.

The profile of abundant proteins in BALF mirrors that of plasma proteins largely due to the permeability of air-blood barrier. In agreement with the literature, the most abundant plasma proteins such as serum albumin, IgG, and transferrin were also found to be most abundant in BALF from both asthmatic patients and healthy controls. As shown in Fig. 1, HSA accounted for as much as 90% of the total proteins in this particular BALF sample. After SAC, the albumin level can further increase in BALF due to plasma leakage from pulmonary blood vessels into lung tissues as a result of inflammatory damage to the alveolar capillary barrier (8). The large dynamic range in the protein expression level presented a great challenge to the successful detection and relative quantitation of interesting BALF proteins that are usually present at very low abundance. As a critical step, we used an affinity column to remove the six most abundant proteins. The ratio of the mixed affinity resins in the column was originally optimized for the removal of such proteins from plasma, but it worked equally well in the case of BALF proteins. As shown in Table II, virtually quantitative depletion of all six proteins was achieved.

Quantitation of proteins directly based on MS signal intensities without internal standards has historically drawn limited attention primarily due to limited linearity of correlation between protein concentration and MS signal strength and variable ionization efficiency, especially for proteins present in a complex mixture. However, reports from several groups have shown that relative changes in total signal intensities of peptides correlated well with their concentration changes in one sample versus another (16, 19–21). In general, we also found our approach worked well for proteins identified from three or more peptides and proteins showing noticeable expression changes. For example, the relative quantity of MMP-9 in each sample measured by MS correlated very well with that obtained by ELISA (Fig. 2). However, it still should be noted that the method is semiquantitative and should be used with caution for comparison of proteins with low abundance and proteins with relatively small changes.

As shown by the gene ontology chart (Fig. 3, A and B), a large number of extracellular proteins and protein families were identified in BALF. These proteins are involved in numerous functions such as proteolysis, inflammatory responses, cell adhesion, cell mobility, cell proliferation, metabolism, and signal transduction. One interesting observation was that a variety of intracellular proteins were also found in the BALF from both asthmatics and controls with some of them clearly elevated after SAC (Table III). For example, seven proteins in annexin family (annexins A1, A2, A3, A5, A6, A8, and A11) were identified with high confidence of which A3 and A11 were elevated up to 116-fold following SAC (data not shown). It is not yet clear why these intracellular proteins were present in high abundance in the BALF. The possible causes may include: 1) cellular proteins shed into lung airway as a result of increased cell apoptosis and lysis, (2) increased leakage of those proteins circulating in the blood, and (3) protein secretions by pathways that remain largely unknown. Because only a few ribosomal proteins, one of the most abundant intracellular protein families, were identified and with very low level, it is unlikely that cell lysis is a major reason for the identification of these intracellular proteins. On the other hand, it is more likely that these putative intracellular proteins play certain extracellular roles that are currently unknown because many proteins have pleiotropic cellular functions. One such example is annexin A1, which can bind to extracellular protein S100A11. The expression and secretion of annexin A1 is induced by glucocorticoids (38). Although the mechanism for the secretion of annexin A1 is unknown, the extracellular annexin A1 can interfere with granulocyte recruitment, migration, and/or activation at sites of inflammation (39). Another example is HMG-1. HMG-1 is a nuclear protein but can be secreted by activated monocytes and macrophages (40, 41). Outside the cell, it binds to receptor for advanced glycation end products (RAGE) and is a potent proinflammatory mediator that activates downstream cytokine release. It is well known to participate in the development of lethal sepsis (42, 43). The same scenario may apply to its other two family members, HMG-2 and high mobility group protein 1-like 2 (HMG-1L2) (see Supplemental Table I).

The pathophysiology of allergic asthma is complex and involves many cell types (1–3, 5, 44, 45). The initial allergic response is dominated by activation of resident mast cells. Twenty-four hours following allergen challenge, a variety of cell types including Th2 cells, eosinophils, macrophages, and neutrophils are recruited from the circulation and invade the airway epithelial lamina. Those immune effector cells together with the local fibroblasts, smooth muscle cells, and epithelial cells contribute to the persistent airway inflammation and remodeling through release of inflammatory mediators such as chemokines, cytokines, and proteases. Overall there is an increase in the cell numbers and expression levels of proteins within allergen-challenged subjects compared with those in unchallenged controls. As shown in Table III, most of the differentially expressed proteins found in our study can be correlated to the different aspects of the complex molecular events happening during an asthmatic attack. These molecules include chemokines and cytokines, proteases, complement factors, acute phase proteins, and monocyte-specific granule proteins that are proinflammatory; local matrix proteins that are associated with the increased mucus production and maintenance of airway function; and abundant plasma proteins that indicate the increased plasma infiltration. In addition, the overall level of protein expression changes also correlated well with the extent of the response of an individual asthmatic to SAC based on clinical observations with good responders (patients MM and HB) having much more protein abundance compared with those for poor responders (patients KG and JH) (Table I). Our observations in BALF protein expression changes are consistent with what is known about the pathophysiology of asthma.

Over a dozen chemokines and cytokines were identified in BALFs (Table IV). Some of the proteins (CCL17, megakaryocyte-stimulating factor, macrophage-stimulatory protein, and high mobility group protein 1) were shown to be distinctly elevated after SAC according to our criteria. Others were present in much lower abundance with only three or fewer peptides identified, and thus the change in expression is not discernable. It is also interesting to note that some of the well known Th2-specific cytokines such as IL4, IL5, and IL13, and chemokines such as eotaxin and monocyte chemotactic protein were not detected. Possible reasons include the timing of the BALF harvest and tissue or cell type distribution profiles, and if present they were below the sensitivity limit of mass spectrometric detection.

Up-regulated proteins associated with potentiation of inflammation represented the largest group in Table III. Those include proteases, complement factors, acute phase proteins, and eosinophil and neutrophil granule proteins. It has been well documented that activation of eosinophils, once being recruited to the airway, leads to secretion of many eosinophil-specific granule proteins such as eosinophil cationic protein, major basic protein, eosinophil peroxidase, eosinophil-derived neurotoxin, and Charcot-Leyden crystal protein, etc. (1, 3, 44). All but eosinophil peroxidase was identified in BALFs with highly elevated expression in asthmatics after SAC (Fig. 4). Asthmatics MM and HB who had stronger overall responses to SAC showed significant increases in the expression of these proteins. This observation was also supported by a previous report that the elevated level of eosinophil cell counts and their cationic protein release in BAL was related to the magnitude of the late phase response (5). Such observation on the response of the asthmatic patients to SAC is consistent with the phenomenon of eosinophilia during a typical allergic asthma exacerbation and thus further verifies the SAC as a clinical model for asthma exacerbation. In addition to the eosinophil-specific granule proteins, a number of neutrophil granule proteins, such as myeloperoxidase, C5a (neutrophil chemotactic factor), lipocalin 2, and cysteine-rich secretory protein-3 precursor (CRISP-3) were also significantly elevated in the airway of asthmatic patients with strong SAC responses, whereas the expression levels of those proteins in healthy controls are either unchanged or are elevated to a much lesser extent following SAC. From the molecular level, such changes in these proteins demonstrate the involvement of neutrophils in the late phase reaction after SAC.

MMPs are a large family of proteinases that play important roles in asthma by their influence on the function and migration of inflammatory cells and matrix deposition and degradation (46–50). MMP-7, MMP-8, MMP-9, and MMP-20 were identified with high confidence in BALFs among which MMP-9 was predominant with 21 unique peptides identified. MMP-9 plays a critical role in tissue repair and remodeling because of its function to degrade type IV collagen, which is an important constituent of basement membrane where airway remodeling occurs (46, 47, 50). A dramatic increase in MMP-9 expression was consistently observed in asthmatics after SAC. Interestingly TIMP-1, which regulates the activity of MMP-9, was also elevated in challenged asthmatics, although the absolute concentration of TIMP-1 seemed to be lower than that of MMP-9.

A number of proteins secreted by local epithelial cells, smooth muscle cells, and fibroblasts in the airway were also highly up-regulated in BALF after SAC. Some of these proteins such as megakaryocyte-stimulating factor (also called lubricin) are possibly involved in the maintenance of normal airway function. Others including pulmonary surfactants, LPLUNC1, protein PLUNC, chitinase 3-like 2, chitinase 3-like 1, lipopolysaccharide-binding protein, mannose-binding lectin (MBL), ficolins 2 and 3, and intelectin are well known host defense proteins that are part of the innate immunity against pathogens (51–54). As one would expect, MBL-associated serine protease that is required for MBL and ficolins to trigger the innate immunity through lectin-complement pathway upon recognition of infectious agents is also clearly up-regulated after SAC. Among these host defense proteins, pulmonary surfactant family contains four unique proteins, surfactant protein (SP)-A, SP-B, SP-C, and SP-D, which are all synthesized by lung epithelial cells. Both SP-A and SP-D are hydrophilic surfactant proteins and share similar roles by regulating immune cell functions and participate in first line defense within the lung against invading pathogens, whereas the hydrophobic surfactant proteins, SP-B and SP-C, are responsible for surface activities.

An interesting observation is the increased expression of two growth factors, insulin-like growth factor-I (IGF-I) and hepatocyte growth factor-like protein (HGFL, also named macrophage-stimulating protein), and their related proteins (Table III). The increase of IGF-I and its binding proteins plays an important role in airway remodeling because there is a strong correlation between IGF-I expression and collagen thickening after inhaled corticosteroid treatment in asthmatics (55). On the other hand, HGFL is a multifunctional factor regulating cell adhesion and motility, growth, and survival. The increase of HGFL seems to play at least dual roles during the late phase inflammation reaction after SAC. It may contribute to the recruitment and activation of macrophages while also preventing the airway epithelial cells from apoptosis in an inflammatory environment (56, 57). More interestingly, the hepatocyte growth factor activator, which is a serine protease required for the activation of hepatocyte growth factor, was also up-regulated, whereas the level of hepatocyte growth factor activator inhibitor was decreased in the BALF after SAC. This observation clearly indicates a coordinated response at the molecular level to SAC in asthmatics.

CONCLUSIONS

Although the value of BALF analysis in determining the pathogenesis of lung disorders is well established, to our knowledge this study provides the most comprehensive database of the proteins present in BALFs of both healthy subjects and asthmatic patients. This is the first report of differential proteomic analysis of BALFs from both asthmatics and healthy subjects before and after SAC. The analysis of four asthmatics and three healthy individuals resulted in the identification of 1,592 proteins among which about 10% displayed altered expression for asthmatic patients at the peak of the late phase reaction to allergen challenge. Although the present findings are based on a limited number of asthmatic individuals, the results, from the protein level, clearly reflect the complex pathophysiology of asthma, such as inflammation, eosinophilia, airway remodeling, tissue damage and repair, mucus production, and plasma infiltration. This study provides new insights for uncovering novel inflammatory mediators that can be potential therapeutic targets or biomarkers for asthma.