INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

₁-Acid glycoprotein (AGP) is a plasma protein that belongs to the group of acute-phase proteins. It is produced mainly in the liver, but messenger RNA (mRNA) expression for AGP has been described in hyperplastic alveolar type II cells, although neither AGP protein nor mRNA could be detected in normal lung tissue (1, 2). The physiologic role of AGP is poorly understood, but it has a modulatory influence on the immune response. It has been shown to protect against inflammation-induced tissue injury (3) and tumor necrosis factor-- induced cell killing (4), and it acts on cells that are involved in the inflammatory process, such as polymorphonuclear cells (5, 6), macrophages (7, 8), lymphocytes (9, 10), and platelets (11).

AGP has five N-linked glycans that can be of the di, tri, and tetraantennary type (12, 13). Variations in the degree of branching of these glycans occur during inflammation and pregnancy. Decreased branching is induced by acute inflammatory conditions, such as occur during severe trauma and intercurrent infections in rheumatoid arthritis and systemic lupus erythemathosus (14). In contrast, a slightly increased branching of the glycans of AGP has been described during some chronic inflammatory conditions (e.g., rheumatic diseases) (18), and a strong increase has been reported during pregnancy (19). The degree of branching of the glycans of AGP might be important, because it has been shown to affect various immunomodulatory functions of AGP (1, 8, 11, 20).

To evaluate the degree of branching of the glycans of AGP in asthma, we compared the concentration and glycosylation pattern of AGP in plasma and epithelial lining fluid of patients with atopic asthma (AA), atopic nonasthmatic (ANA) subjects, and patients with interstitial lung disease (ILD), and correlated the differences with lung-function and inflammatory parameters.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patient Selection

Patients were selected on the basis of their history of asthma and allergy. Spirometry was performed and bronchial hyperreactivity (BHR) was evaluated by means of the histamine threshold (a provocative concentration of histamine causing a 20% decrease in FEV₁ [PD₂₀] < 16 mg/ml was defined as representing hyperreactivity). Allergy was confirmed by skin-prick tests with a set of 16 common aeroallergens (ALK-Abelló bv., Nieuwegein, The Netherlands) and with radioallergosorbent tests (RASTs) (Dermatophagoides pteronyssinus, cat, ragweed, and fungi). The AA patients had a history of mild/moderate asthma and had not been admitted to a hospital in the previous year.

The ANA subjects did not have any pulmonary complaints, and had normal lung function and a normal histamine PD₂₀. The third group consisted of patients with various noninfectious ILDs (one with eosinophilic pneumonitis, five with sarcoidosis, one with idiopathic pulmonary fibrosis, one with ILD associated with rheumatoid arthritis) in which bronchoalveolar lavage (BAL) was performed for diagnostic purposes. The ILD patients had no previous history of asthma and/or allergy. Glucocorticoid therapy, smoking, and pregnancy were used as exclusion criteria in all patient groups. The ethics committees of the authors' institutions approved the study, and informed consent was obtained from each subject before the study.

BAL and Blood Sampling

BAL was performed according to a standardized procedure (21), and NaCl (0.9%) was used as lavage fluid. The proximal airway compartment was investigated through the fluid recovered after instillation of the first aliquot of 50 ml (22); the subsequent 100 ml and 50 ml aliquots were pooled after recovery and were used to reflect the alveolar compartment. The bronchoalveolar lavage fluid (BALF) was kept on ice to ensure the viability of the recovered cells. The cells were counted and spun down, and cytospin slides were made. Cell staining was done with May-Grünwald-Giemsa stain and cell differentiation was done by counting 200 cells per slide. The BALF samples were stored in portions at 20° C. In order to analyze the glycosylation of AGP in epithelial lining fluid (ELF), the BALF had first to be concentrated. Portions of 25 ml of BALF were lyophilized, recovered in water, dialyzed over a membrane (molecular weight cutoff: 12 to 14 kD; CelluSep; Roth, Karlsruhe, Germany) against distilled water, lyophilized for a second time, and resuspended in 25 µl of distilled water. In this way BALF proteins were concentrated by a factor of 1,000.

AGP Concentration in ELF and Blood Plasma

The plasma AGP concentration was measured nephelometrically (34). Concentrations of AGP in BALF concentrates were measured through rocket electrophoresis according to the method of Laurell (23), using monospecific rabbit antiserum against human AGP (obtained from Dr. Andrzej Mackiewicz, Great Poland Cancer Centre, Poznan, Poland) for precipitation. The human serum protein calibrator (HSPC: Lot No. 127; Dakopatts, Glostrup, Denmark), a pool of 30 sera from healthy donors, was used as standard (24, 25).

Crossed Affinoimmunoelectrophoresis

Crossed affinoimmunoelectrophoresis (CAIE) was performed according to a modification (24) of the Bøg-Hansen method (26), using 8% polyacrylamide in a 24.3 mM diethylphenobarbituric acid/Tris buffer (pH 8.6) containing 0.4 mM calcium lactate and 0.02 NaN₃. The lectin concanavalin A (Con A, 1 mg/ml; type V; Sigma Chemical Co., St. Louis, MO) was included in the first-dimension gel as the diantennary-specific affinocomponent. The separation of the different glycoforms of AGP in blood plasma and BALF concentrate was done via electrophoresis of these fluids through a Con A-containing polyacrylamide slab gel, using a Mini-Protean II dual slab gel apparatus (BioRad Inc., Hercules, CA). AGP lacking glycans of the diantennary type is not retarded by Con A, whereas AGP containing one or more diantennary glycans binds to Con A and as a result is electrophoretically retarded in the gel. Detection of the separated glycoforms was achieved through electrophoresis in the second dimension, using the monospecific, precipitating antiserum (rabbit antihuman-AGP-IgG) in a 1% agarose gel, with HSPC as a standard (27). The resulting precipitation lines were stained with Coomassie brilliant blue, and the relative occurrence of lectin-retarded and -nonretarded glycoforms was calculated from the areas under the curves, as determined by analysis with Summagraph ACECAD D-9000 software (Monterey, CA).

Analysis of Glycan Structures of AGP

AGP was isolated from 0.5 ml plasma from each of two representative individuals in each study group by the method of Chan and Yu (28). The purity of the preparations was controlled with sodium dodecylsulfate-polyacrylamide gel electrophoresis as well as chromatography on Superose 12 HR (Pharmacia, Uppsala, Sweden). The glycans of 100 µg of AGP were released by Peptide:N-glycosidase F (PNGase-F) treatment. Incubations were done for 24 h at 37° C under reducing conditions in 50 mM sodium phosphate buffer containing 1% Nonidet P-40 and 1,000 U PNGase-F (total volume = 50 µl). After the volume of each sample preparation was increased to 1 ml with demineralized and filtered water (milliQ water; Millipore, Etten-Leur, The Netherlands), the released glycans were bound to Carbograph SPE columns (Alltech Corporation, Deerfield, IL) and washed with 5 ml milliQ water, and the neutral and acidic glycans were eluted with 3 ml of 25% (vol/vol) CH₃CN and 25% CH₃CN + 0.05% trifluoroacetic acid, respectively (29). The glycan fraction was lyophilized, dissolved in 100 µl milliQ water, and analyzed with high-performance anion-exchange chromatography with pulsed amperometric detection, using a Carbopack PA-100 column (0.4 × 25 cm; Dionex Corp., Sunnyvale, CA) with 0.1 M NaOH at a rate of 1 ml/min as the mobile phase. Prior to each sample injection, the column was washed for 5 min with 0.1 M NaOH/0.5 M sodium acetate and then for 15 min with 0.1 M NaOH. Elution was isocratic with 0.1 M NaOH for 10 min, after which sodium acetate was applied in a gradient by increasing the concentration from 0 to 0.25 M over a period of 100 min. Di-, tri-, and tetrasialylated complex-type oligosaccharide structures were used to characterize the Carbopack PA-100 column; top fractions eluted at 55, 68, and 80 min, respectively.

Statistical Analysis

All differences were tested for statistical significance with Student's t test. Multivariate regression analysis was performed to study the relationship between different variables (30).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Population

The selection of the study groups was based on allergic and asthmatic history. The AA and ANA groups were of comparable age and sex distribution, but the ILD group was older and showed a higher male-to-female ratio (Table 1). The most frequent allergens eliciting positive skin prick tests were grass pollen, Dermatophagoides pteronyssinus, cat and dog hair, and bird feather. The numbers and percentages of subjects in the AA and ANA groups sensitized to these allergens, as measured with skin tests, were 64 (83%), 73 (33%), 64 (42%), 64 (50%), and 27 (50%), respectively. The selected patient groups showed marked differences in lung function and BHR. As expected, the selection of the study groups resulted in clear differences in airway hyperreactivity to histamine (PD₂₀) (p < 0.001) (Table 1). Lung function, as measured with FEV₁ %predicted was decreased in the AA group (p < 0.05) and was unrelated to age or sex.

The percentages of eosinophils in the lung (airway and alveolar fraction) and in the blood were significantly increased in asthmatic subjects (p < 0.05 and p < 0.01, respectively) (Table 2), and the IgE concentration showed a tendency to be higher in the AA as compared with the ANA group (p = 0.05) (Table 1). No significantly greater BALF recovery or red blood cell (RBC) count was found for any of the study groups versus any other, and the percentages and numbers of epithelial cells and lymphocytes were equal in all three groups (data not shown).

AGP Concentrations

The concentrations of plasma AGP in both the AA and the ANA groups were slightly lower than the control value, whereas the concentration in the ILD group was about twofold higher than the control (p < 0.0005) (Table 3). AGP was also detected in BALF, albeit at a concentration about 1,000-fold lower than in plasma. Again no differences were found between the AA and the ANA group, whereas not enough material was available to determine the AGP concentration in the BALF of ILD patients.

Degree of Branching of the Glycans of AGP

CAIE with the lectin Con A as the affinocomponent in the first-dimension gel was used to determine the relative occurrence of glycoforms of plasma and BALF AGP, indicating whether none (AGP-C0), one (AGP-Cw), or two or more (AGP-Cs) of the five N-linked complex type glycans of AGP were of the diantennary type (Figure 1). The AA group had a significantly higher proportion of plasma and BALF AGP glycoforms lacking diantennary glycans (C0) than did the HSPC standard (Table 4). The AA group also had a higher C0 glycan fraction than did either the ILD or the ANA group (Figure 2, Table 4). This was confirmed by HPAEC-PAD analyses of the glycans released by PNGase-F from isolated plasma AGP in the various study groups, showing that the AA group appeared to have a much lower fraction of disialylated glycans of AGP than did the other three groups (Figure 3, arrow). Comparison of the elution patterns at positions A and B in Figure 3 shows that AGP in the ILD group was more highly fucosylated than was AGP in the AA group, since position A also represents elution of difucosylated, trisialylated glycans and position B also represents monofucosylated, trisialylated glycans. The amount of AGP in BALF was too small to permit HPAEC-PAD analysis. Statistical analysis showed a relation between the degree of AGP glycan branching and gender (p = 0.017; r = 0.51) that was independent of asthma status. The relative presence of plasma AGP-C0 was highest in female subjects, at 58 ± 14%, versus 45 ± 7% in male subjects (p < 0.01).

View larger version (44K): [in this window] [in a new window]   Figure 1.   Fractionation of AGP in BALF according to diantennary glycan content. Typical examples of AGP glycoforms in BALF of an asthmatic patient (A) and a healthy subject (B). CAIE of concentrated BALF, with Con A in the first-dimension gel, was performed as described in METHODS. Electrophoresis was run from right to the left for the first dimension and from bottom to top for the second dimension (arrows); only the second dimension gels are shown. CO = glycoforms of AGP not retarded by Con A, containing five tri- or tetraantennary glycans; Cw and Cs = AGP glycoforms in which one (Cw) and two or more (Cs) of the five glycans of AGP are of the diantennary type (13). The areas under the curves were used to calculate the relative occurrence of the various glycoforms. Note the high C0 content of AGP in atopic asthma patient's BALF relative to the C0 content of AGP in healthy subject's BALF.

View larger version (13K): [in this window] [in a new window]   Figure 2.   Relative occurrence of diantennary glycoforms of AGP in plasma (A) and BALF (B). Percent distribution of the C0 glycoforms is given for plasma and BALF of the various study groups. The box plots show the median value (dashed line), interquartile range (boxes), and outliers (solid line). See Figure 1 for definition of AGP glycoforms.

View larger version (21K): [in this window] [in a new window]   Figure 3.   Typical HPAEC-PAD profiles of glycans in plasma AGP of a healthy subject (Ctrl) and of individuals in the ANA, AA, and ILD groups. The preparation of the glycans released by PNGase-F from isolated AGP, and the subsequent analysis with HPAEC-PAD, using a PA-100 column is described in METHODS. The column was calibrated with complex-type standard glycans as follows: A = disialylated glycans; C = trisialylated glycans; D = tetrasialylated glycans. Mono- and difucosylated, trisialylated, complex-type glycans were recovered in B and A, respectively (D. C. W. Poland, C. A M. Koeleman, and W. Van Dijk, unpublished observation). The multiple peaks in each region are due to heterogeneity in type of glycan substitution (2  6- or 2  3-linkage) with sialic acid (32). Note that only in AA patients is the peak comprising the disialylated glycans of AGP (arrow) sharply reduced relative to the peaks of the tri- and tetrasialylated glycans.

Relation to Asthma Parameters

The relationship between specific AGP glycoforms and a number of asthma-related parameters was tested statistically, after correction for gender, through multiple linear regression analysis. The degree of branching of AGP glycans was closely related to the eosinophil percentage in both BALF and blood, and was also related to the eosinophil number in BALF (no data were available on the number of eosinophils in blood) (Table 5). The BALF was separated into a bronchial and an alveolar fraction, and the eosinophil percentages in both fractions were found to be significantly related to the proportion of AGP-C0 in BALF (p = 0.008, r = 0.69, and p = 0.019, r = 0.65, respectively). This correlation was also present, although less strongly so, with plasma AGP. Thus, an increased number and percentage of eosinophils in BALF was accompanied by an increase in the degree of glycan branching in AGP. Additionally, the eosinophil percentage in peripheral blood was related to the C0 fraction of BALF and plasma AGP (p = 0.003, r = 0.72; and p = 0.003, r = 0.61, respectively).

Both of the functional parameters of asthma, spirometrically evaluated lung function and PD₂₀ were correlated with the degree of branching of the AGP glycans. FEV₁ showed a negative association with the relative C0 fraction of AGP in BALF (p = 0.004, r = 0.76) (Table 5). Thus, a decreased FEV₁ was accompanied by an increase in AGP-C0. A similar but less strong association was found for PD₂₀, with increased BHR (and thus a lower PD₂₀) accompanied by an increased AGP-C0 (C0: p = 0.041, r = 0.57). The relation with both PD₂₀ and FEV₁ existed only at the local level; the differential glycosylation of plasma AGP was not related to lung function.

In contrast to its relation to eosinophils, the glycosylation pattern of AGP was not found to relate either to the type of allergen to which patients were sensitized or to the plasma IgE concentration.

Chronic inflammation of the airways renders the bronchial epithelium more vulnerable to injury, which facilitates minor bleeding during bronchoscopy. We used the RBC count and the percentage of epithelial cells in the leukocyte differentiation in BALF as measures of epithelial vulnerability (31). However, these parameters, like allergen type and plasma IgE concentration, were unrelated to the AGP glycoform pattern, indicating that epithelial damage was also unrelated to the degree of branching of the glycans of AGP.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

In this study we investigated the relative occurrence of diantennary N-linked glycans in the acute-phase protein AGP in AA and ANA patients and patients with ILD. Our data indicate that AGP in BALF and plasma from asthma patients shows increased branching of its glycans, and that this phenomenon is correlated with both inflammatory and functional parameters.

Despite the modest size of the patient groups in the study, significant differences were detected in the degree of AGP glycan branching. The differences were confirmed by comparison with an HSPC standard consisting of pooled serum from 30 healthy individuals. In addition, the lung and the blood plasma data in our study can be seen as two sets of data that are in perfect agreement with one another. However, although plasma AGP can be transported to the extravascular space (33), the two sets of data might be independent, since AGP can be produced locally in the lung (1). Asthmatic subjects in our study did not show an increased concentration of AGP, suggesting that the effect of asthma on AGP is restricted to its glycosylation.

In ILD, which is often associated with chronic inflammation, no significant differences were found in AGP glycan branching, but increases were observed in the total plasma AGP concentration and in AGP fucosylation. This is comparable to the state of glycosylation of AGP found in chronic inflammatory conditions, in which a normal (diabetes mellitus) or slightly increased branching of AGP glycans (rheumatoid arthritis) was accompanied by strongly increased fucosylation of AGP (19, 22, 34). In our study HPAEC-PAD analysis indicated no increase in fucosylation of the glycans of asthma patients' AGP (Figure 3, fraction B). Therefore, the observed strong increase in branching of the glycans of AGP in asthma is more comparable to changes in the glycosylation of AGP induced by pregnancy and high-dose oral estrogen treatment (19, 21). Nevertheless, the state of glycosylation of AGP is clearly related to asthma-induced inflammation, since it was found to correlate with factors related to chronic airway inflammation in the form of an increased number of eosinophils, decreased lung function, and increased BHR.

Several confounding factors can influence the glycosylation of AGP. These are: (1) inflammatory state (17); (2) pregnancy (19); (3) drug/glucocorticoid use (35); and (4) oral contraceptive use (21). The selection of the AA and ANA groups in our study was based on the presence of asthma symptoms and an atopic constitution, and excluded patients with other, confounding inflammatory diseases as well as those using steroids and pregnant individuals. Atopy might have influenced our findings, but the degree of branching of the glycans of AGP was unrelated to the allergens examined in the study or to the plasma IgE concentration; additionally, the ANA group showed a normal pattern of AGP glycan branching. Moreover, the data found for the ILD group suggest that not all inflammatory diseases of the lung influence the extent of glycan branching of AGP. However, use of oral contraceptives, which affect the relative proportions of AGP glycoforms, was not excluded in our study (21), and might explain the relationship found between AGP glycosylation and sex. Nevertheless, the results of a multivariate regression analysis showed that the correlation between the degree of AGP glycan branching and asthma was independent of sex, and the correlation of such branching with the inflammatory and lung function parameters examined in the study also cannot be explained by oral contraceptive use. The study data instead indicate that chronic inflammation of the airways is the most plausible explanation for the increased branching of AGP glycans found in asthma patients. The factors responsible for the altered glycosylation of AGP observed in asthma remain unknown. It does not seem likely that this alteration is a nonspecific effect induced by a hypermetabolic state. Such a state also occurs, for example, during acute inflammation, which is characterized by decreased AGP glycan branching and increased fucosylation (see the previous discussion). Furthermore, in vitro studies with rat and human hepatocytes have shown that specific combinations of inflammatory cytokines and glucocorticosteroids are responsible for the changes in glycosylation of AGP during acute inflammation and in the development of rheumatoid arthritis, most probably by inducing increased activity of specific glycosyltransferases (17). Therefore, it is likely that asthma-related factors are involved in inducing the increased branching of AGP glycans seen in our study.

AGP, as a secondary phase acute-phase protein, is supposed to dampen the harmful systemic effects induced in the first phase of an acute inflammatory process. The large inflammation-dependent changes in plasma concentrations of specific AGP glycoforms are important in this respect. For example, increased branching increases the capacity of AGP to inhibit the proliferation of lymphocytes (20) and to induce the secretion by macrophages of an inhibitor of interleukin-1 activity (8). Furthermore, inhibition of the complement cascade by AGP has been shown to depend on the composition of its glycans (2), and increased fucosylation of AGP has been suggested to have an inhibitory effect on selectin-mediated interactions between lymphocytes and the endothelium (16, 17). Because of this, it is important to identify specific asthma- related immunomodulatory properties of AGP glycoforms, as well as to identify the factors that regulate their expression.

In asthma, clinical and functional parameters are used to assess treatment and disease activity. Inflammatory markers that can be detected in blood or in other body fluids might be useful in monitoring treatment of asthma in order to prevent exacerbations. The increased branching of AGP glycans observed in our study was related to the two inflammatory parameters of histamine PD₂₀ and eosinophil number, and might therefore constitute a noninvasive marker of inflammation in asthma. It will therefore be of interest to confirm these data in larger patient groups, and to study changes in glycosylation of AGP glycans during exacerbations and after the start of therapy for asthma.

In conclusion, the present study shows an increased degree of branching of the N-linked glycans of AGP in asthmatic individuals, at both the level of the airways and the systemic level, that correlates with eosinophil numbers as well as lung-function parameters.