Published ahead of print on November 3, 2003, doi:10.1164/rccm.200305-619OC
© 2004 American Thoracic Society Free Secretory Component from Cystic Fibrosis Sputa Displays the Cystic Fibrosis Glycosylation PhenotypeSchool of Pharmacy and Biomedical Sciences, University of Portsmouth, Portsmouth, and Department of Respiratory Paediatrics, Royal Brompton Hospital, London, United Kingdom Correspondence and requests for reprints should be addressed to Lindsay J. Marshall, Ph.D., School of Pharmacy and Biomedical Sciences, St. Michael's Building, White Swan Road, University of Portsmouth, Portsmouth PO1 2DT, UK. E-mail: lindsay.marshall{at}port.ac.uk
Secretory IgA contributes to humoral defense mechanisms against pathogens targeting mucosal surfaces, and secretory component (SC) fulfills multiple roles in this defense. The aims of this study were to quantify total SC and to analyze the form of free SC in sputa from normal subjects, subjects with asthma, and subjects with cystic fibrosis (CF). Significantly higher levels of SC were detected in CF compared with both other groups. Gel filtration chromatography revealed that SC in CF was relatively degraded. Free SC normally binds interleukin (IL)-8 and inhibits its function. However, in CF sputa, IL-8 binding to intact SC was reduced. Analysis of the total carbohydrate content of free SC signified overglycosylation in CF compared with normal subjects and subjects with asthma. Monosaccharide composition analysis of free SC from CF subjects revealed overfucosylation and undersialylation, in agreement with the reported CF glycosylation phenotype. SC binding to IL-8 did not interfere with the binding of IL-8 to heparin, indicating distinct binding sites on IL-8 for negative regulation of function by SC and heparin. We suggest that defective structure and function of SC contribute to the characteristic sustained inflammatory response in the CF airways.
Key Words: polymeric immunoglobulin receptor • glycoproteins • inflammation • respiratory mucosa The cystic fibrosis (CF) "glycosylation phenotype" (1) was first described in 1959 as an altered ratio of fucose:sialic acid in CF duodenal mucins (2). The gene mutated in CF encodes for the CF transmembrane conductance regulator (CFTR), a cAMP-dependent chloride channel in the apical membrane of epithelial cells. There is mounting evidence for a direct role of defective CFTR in the altered terminal glycosylation observed in CF. For example, overexpression of mutant CFTR resulted in decreased sialylation of glycoconjugates compared with cells expressing wild-type CFTR (3). Conversely, transfection of wild-type CFTR into airway epithelial cells cultured from CF subjects normalized both sialic acid and fucose content of membrane glycoproteins (3). Although the exact mechanism remains to be elucidated, the alterations in glycosylation in CF have been attributed to the function of CFTR as an intracellular chloride channel (4). Sialylation and fucosylation are important posttranslational modifications of proteins, as terminal sialic acid and fucose residues are implicated in several types of interaction, including, importantly, host–pathogen interactions. For example, the invasion of Chinese hamster ovary cells by Trypanosoma cruzi (5) and the adherence of Pseudomonas aeruginosa to mouse tracheal cells (6) both appear to depend on sialyl residues on host cell surfaces. Several bacteria, including P. aeruginosa, express fucose-binding lectins (7), which facilitate adherence to the host. Secretory component (SC) forms the extracellular domain of the polymeric immunoglobulin receptor (pIgR), which is expressed uniquely by epithelial cells and is responsible for transcytosis of dimeric IgA across the epithelium (8). SC is composed of approximately 15% by weight N-linked carbohydrate (9), although estimates of carbohydrate content vary from 9.5–23.4% (10–15). Proteolytic cleavage of pIgR at the apical surface releases secretory IgA (SIgA), comprising SC bound to dimeric IgA. The closely packed quaternary structure of SIgA confers protease resistance to both SC and IgA (16), enhancing mucosal immunity. Basolateral to apical movement of pIgR in the absence of ligand results in the apical release of free SC (8). Of the total SC in mucosal secretions, up to 60% is present as free SC (17), which has been shown to possess both antiinflammatory (18) and anti-infective properties (19). We recently demonstrated a role for the carbohydrate component of free SC in binding and inhibiting the chemotactic activity of interleukin (IL)-8 (20), a potent neutrophil chemoattractant that is implicated in the persistence of the chronic inflammatory response in the CF airways (21). We hypothesized that aberrantly glycosylated SC in CF could not bind or inhibit IL-8 bioactivity and would thereby contribute to the sustained inflammation characteristic of CF airways. Therefore, the aim of this study was to examine, ex vivo, the form of SC isolated from sputa from normal subjects, subjects with asthma, and subjects with CF. Some of the results of these studies have been previously reported in the form of an abstract (22).
Sputa from 20 normal subjects (mean age of 23 ± 4.6 years), 22 subjects with asthma (mean age of 35 ± 6.4 years), and 41 subjects with cystic fibrosis (mean age of 12.5 ± 0.6 years) were analyzed. This study was approved by the ethical committee of each institution, and written consents were obtained from the patients or guardians where appropriate. Further patient details are available in the online supplement. Sputum samples were induced in normal subjects and subjects with asthma as described previously (23) or were spontaneously expectorated by CF patients. Sputa were either processed immediately with an equal weight of phosphate-buffered saline (PBS) (23) or snap frozen and stored at -80°C before processing. Total IL-8 and total SC were measured by ELISA as described previously (23). Immunoprecipitation, sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and Western blot analysis of total SC, carbohydrate content, and SC-associated IL-8 were performed as described previously (20). Sputa (250 µl) were fractionated on a Superose 6 column (Pharmacia, Uppsala, Sweden) at a flow rate of 0.5 ml/minute in PBS using the Biologic low-pressure liquid chromatography system (BioRad, Herts, UK). Fractions (250 µl) were collected and 20 µL transferred to nitrocellulose using a dot-blot hybridization manifold (Jencons PLS, Bedfordshire, UK). Blots were stained as described previously (24). The chemotactic activity of the fractions was assessed using neutrophil chemotaxis in micro Boyden chambers, as described previously (20). For monosaccharide composition analysis, SC was immunoprecipitated from sputa using the Seize X kit (Perbio Science UK Ltd., Cheshire, UK). After recovery of free SC, monosaccharides were released by acid hydrolysis and analyzed by fluorophore-assisted carbohydrate electrophoresis using a method adapted from Jackson (25). Briefly, half the volume of immunoprecipitated SC was used for sialic acid analysis and was treated with 0.2 N trifluoroacetic acid for 1 hour at 80°C. The remaining half was treated with 4 N trifluoroacetic acid at 100°C for 5 hours (for analysis of fucose). Hydrolyzed samples were cooled at -20°C for 30 minutes and dried in a vacuum evaporator. The dried pellets were fluorescently labeled with 5 µL of 0.1 M 2-aminoacridone (dissolved in glacial acetic acid: dimethyl sulfoxide at 3: 17, vol/vol) and 5 µL of 0.1 M sodium cyanoborohydride at 37°C overnight before drying again. In parallel with the samples, standards of N-acetylneuraminic acid (25 µmol/ml) and fucose (1 µmol/ml) were fluorescently labeled. Samples and standards, loaded at 4 µl per lane in dimethyl sulfoxide:glycerol:water (2:1:7), were electrophoresed at 4°C on 20% polyacrylamide gels in 0.1 M Tris-borate buffer, pH 8.3. Gels were imaged using the UVP Mini darkroom Gel Documentation System and the intensity of labeling of samples expressed relative to the intensity of the standards. Analysis of monosaccharide standards showed that the fluorescence intensity was linear in the range of 0 to 8 nmol. To demonstrate IL-8 binding, 5 ng/ml of human recombinant IL-8 (the gift of Dr. Ivan Lindley, Novartis, Vienna) was incubated with SC (500 ng/ml) in Hanks' balanced salt solution/20 mM N-2-hydroxyethylpiperazine-N'-ethane sulfonic acid, pH 7.4, for 2 hours at 20°C. The mixture (250 µl) was applied to a gel filtration column (Superdex 200; Pharmacia), separated, and analyzed as described previously here. To demonstrate the effect of SC on IL-8 binding to heparin, IL-8 (0.36 µg/ml), SC (3.6 µg/ml), or a mixture incubated as described previously here was applied in a final volume of 1 ml to a 5-ml heparin affinity column (Econocolumn; BioRad) connected to the Biologic system. The column was washed at 0.5 ml/minute with 0.05-M Tris-HCl, pH 7.4, and samples eluted with 10 ml of a linear concentration gradient (0–100%) of 1 M NaCl. Sequential 0.5-ml fractions were collected and analyzed for SC and IL-8 by immunoblot, as described previously here. To assess the effects of IL-8 on SC synthesis, primary cultures of bronchial epithelial cells (Cambrex Bio Science Wokingham, Berkshire, UK), cultured using established methods (20), were incubated with human recombinant IL-8 for 48 hours. Supernatants were harvested and SC analyzed by ELISA as described (23).
Data Analysis
Quantification of SC and IL-8 in Sputum Samples The concentrations of SC and IL-8 in sputum samples were quantified using ELISA and the molar ratio of IL-8:SC calculated, based on molecular weights of 8,300 and 80,000 D, respectively (Table 1) .
The concentrations of SC and IL-8 in sputum from normal subjects were 7.84 ± 2.1 µg/ml and 0.35 ± 0.2 ng/ml, respectively, and calculation of the molar ratio of IL-8:SC revealed a 100-fold molar excess of SC. In comparison, significantly higher levels of both SC (p < 0.01) and IL-8 (p < 0.05 and p < 0.01 for asthma and CF, respectively) were detected in sputum from subject with asthma and subjects with CF. In asthma, the molar ratio of IL-8:SC was not significantly different compared with normal. However, in CF, levels of IL-8 were 500-fold higher than those in normal sputum. This, in combination with a fivefold increase in SC, resulted in a significantly higher molar ratio of IL-8:SC than that found in sputum from either normal subjects or subjects with asthma, reflecting the massive increase in IL-8 characteristic of the CF airways. Nevertheless, the IL-8:SC molar ratio remained less than 1.0.
The Molecular Size of SC in Sputum from Normal Subjects, Subjects with Asthma, and Subjects with CF
After gel filtration, chromatography of CF sputum on Superose 6, which separates in the molecular weight range 5–5,000 kD, SC was detected in high molecular weight complexes (more than 5,000 kD) in the void volume of the column, in fractions corresponding to SIgA (approximately 450 kD), and consistent with Western blot data, at 90 kD (Figure 2A) . Degraded SC was also detected in CF, with molecular weights ranging from 62 kD to 500 D.
Quantification of the staining intensity as a percentage of the total revealed that free SC and SIgA accounted for 7.5 ± 1.2% and 12.0 ± 1.1%, respectively, whereas 15.9 ± 2.1% was degraded SC of 60 kD or less (n = 3). Analysis of IL-8 in the same fractions (Figure 2B) revealed a peak of IL-8 immunoreactivity in the 80- to 90-kD range, indicating an association of IL-8 with SC in these samples. The 90-kD SC detected on Western blots may therefore represent SC (80 kD) in complex with IL-8 (8.3 kD), for which further evidence is given later here. Of the total IL-8, 51.9 ± 6.3% was detected at molecular weights of more than 100,000 kD; 13.2 ± 1.6% was detected in the same molecular weight range as free SC, and 4.0 ± 1.2% of the total was free, 8-kD IL-8.
The Chemotactic Activity of Sputa
IL-8 in Immunoprecipitates of SC We previously demonstrated that IL-8 was released from primary bronchial epithelial cells in culture as a complex with SC (20). To investigate whether IL-8 was similarly complexed with SC in the airway, SC was immunoprecipitated from sputum samples, and the immunoprecipitate was stained with an antibody to IL-8 (Figure 4A) . For each subject group, IL-8 was coimmunoprecipitated with SC at an apparent molecular weight of 90 kD. Densitometric analysis of the amount of IL-8, which coimmunoprecipitated with SC, after correction for the amount of SC in the sample, revealed no differences in sputum samples from normal subjects compared with subjects with asthma (Figure 4B). However, analysis of the amount of IL-8 coimmunoprecipitated with SC from CF sputum revealed 20-fold less IL-8 compared with sputum from either normal subjects or subjects with asthma (p < 0.005). This was an unexpected finding given that the ELISA data indicated a significantly higher ratio of IL-8 to SC in sputum from CF subjects (Table 1).
As we had previously demonstrated that the carbohydrate moiety of SC was important for IL-8 binding (20), we hypothesized that the reduced binding of IL-8 to SC in CF reflected an abnormal glycosylation pattern of SC in CF. We therefore examined the glycosylation pattern of SC in the airways.
The Carbohydrate Content of Total SC
The Fucose and Sialic Acid Content of Free SC We investigated whether SC fitted the recognized CF glycosylation phenotype (1) by examining the monosaccharide composition of free SC from sputa of normal and CF subjects. Western blotting for SC and ELISA for SIgA revealed that the first fraction to elute from the beads contained a high molecular weight form of SC and that SIgA was present in the first fraction only (data not shown). This first fraction was excluded from further analysis. Analysis of fractions 2–5 indicated that more than 95% of SC was in the form of free SC; thus, these fractions were pooled and used for monosaccharide composition analysis. Figures 6A and 6B illustrate the analysis of neutral sugars and sialic acid residues, respectively, hydrolyzed from free SC isolated from normal (N) and CF sputa. The lower band in the standard (lane S) on the fucose gel (Figure 6A) is glucose contamination. Fucose was the most abundant neutral sugar detected in all samples (Figure 6A). The intensity of labeling was expressed relative to the known concentration of standard fucose or sialic acid and adjusted for the concentration of SC in each sample (Figure 6C). Free SC from CF sputa contained almost twice the amount of fucose (7.1 ± 1.7 nmol/µg SC) compared with normal subjects (4.3 ± 1.5 nmol/µg SC), although this difference did not achieve statistical significance (Figure 6C). In contrast, analysis of the sialic acid content of free SC (Figures 6B and 6C) revealed a significant (p < 0.001) fivefold decrease from 2.7 ± 0.2 nmol/µg SC in normal sputa to 0.5 ± 0.2 nmol/µg SC in CF sputa. The ratio of fucose:sialic acid was significantly (p = 0.02) almost 13-fold higher in SC from CF sputa (Figure 6D). These data agree with the established CF glycosylation phenotype of increased fucosylation and decreased sialylation.
IL-8 Binds SC and Heparin at Different Sites We confirmed the association of SC with IL-8 in vitro. A mixture of IL-8 and SC separated by gel filtration chromatography resulted in one major peak at the molecular weight of SC detected in sputum samples, approximately 90 kD (Figure 7A) . Immunoblot analysis confirmed the presence of both IL-8 and SC in this peak, and not in other fractions. All chemokines bind to heparin; thus, IL-8 applied to a heparin affinity column required 0.6 M NaCl for elution (Figure 7B). When SC was applied to the column alone, the majority (77.8 ± 1.5%) was eluted in the wash buffer with only 22.3 ± 1.5% retained on the column. However, preincubation of SC with IL-8 resulted in retention of both IL-8 and SC on the column (closed symbols and solid line, Figure 7B). In this case, 89.8 ± 1.4% of SC was retained on the column. Bound SC was eluted with 0.6 M NaCl, indicating that the retention of SC on the column was mediated via IL-8 binding both heparin and SC, and indicating distinct binding sites on IL-8 for heparin and SC.
IL-8 Increases SC Production We investigated whether IL-8 could influence SC release from airway epithelial cells, as previously demonstrated in gut-associated lymphoid tissues (26). Incubation of primary cultures of bronchial epithelial cells with human recombinant IL-8 resulted in a dose-dependent increase in SC production (Figure 8) . Concentrations of IL-8 greater than 10-9 M significantly increased SC release (p < 0.05), with a maximum increase of 174% of control levels observed with 10-7 M IL-8.
In summary, we have shown that SC is more abundant in sputa from CF subjects than patients with asthma or normal subjects and that SC is relatively degraded and overglycosylated in CF samples. Less IL-8 is bound to free SC in samples from the airways of CF subjects compared with samples from normal subjects and subjects with asthma, despite significantly higher concentrations of IL-8 in CF subjects. The increase in the ratio of fucose:sialic acid residues on free SC in CF is in agreement with the known glycosylation phenotype of CF. We suggest that changes in the carbohydrate composition of SC have significant functional consequences in CF.
SC is a multifunctional component of the mucosal defense system. The established role of SC as the extracellular ligand-binding domain of pIgR is binding and epithelial transcytosis of dimeric IgA (8). The continued association of SC with dimeric IgA after release of the complex into mucosal secretions protects IgA against proteolysis, increasing its efficacy (16). However, SC synthesis is generally in excess of that required for conjugation with IgA and up to 60% of SC in stimulated parotid saliva exists as free SC (17). Both antiinflammatory and anti-infective properties of free SC have been proposed. For example, it has been suggested that SC produced by cultures of primary human keratinocytes inhibited IFN- The initial finding that SC isolated from the sputum of CF patients was degraded relative to that isolated from either normal subjects or subjects with asthma was not unexpected. Uninhibited protease activity in the CF airways is well documented (27), with levels of elastolytic activity in CF airway lining fluid measured at almost 89-fold greater than normal (28). Furthermore, it was recently shown that neutrophil-derived serine proteases cleave SC (29). However, it was of interest to note that sputa from normal subjects and subjects with asthma apparently contained no degraded SC (Figure 1). Elastase levels, although higher in CF than healthy control subjects, are not significantly different in spontaneously expectorated and induced sputum samples from patients with CF (30), and the different profile in CF is therefore not likely to be an artifact of different sampling methods. Other studies have shown that fragmentation of normal human colostral SC occurred on storage at 4°C, resulting in molecular weights of 42–35 kD (31). As sputum samples from normal subjects, subjects with asthma, and subjects with CF were processed and stored (at -80°C) the same way, it would appear that the degraded SC that we observed in sputa from CF subjects was specific for this subject group and not an artifact of processing or storage conditions.
Increased levels of SC were detected in both CF and asthmatic sputa compared with normal. Proinflammatory cytokines such as IL-4, IFN- The measurements of high IL-8 concentrations (Table 1) agree with previously published data regarding the high concentrations of IL-8 in sputum and bronchoalveolar lavage fluid from patients with asthma and CF (40, 41). Given the high levels of both SC and IL-8 in CF sputa, it was surprising to find that SC isolated from CF sputa bound significantly less IL-8 than SC from normal subjects or subjects with asthma. The free IL-8 detected in CF samples on gel filtration chromatography was active in neutrophil chemotaxis assays, although other IL-8–containing fractions were not (Figure 3). Free IL-8 was undetectable in samples from normal subjects or subjects with mild asthma (20, 40). IL-8 has been shown to increase SC production in gut-associated lymphoid tissues (26), and we have demonstrated that IL-8 increased SC production in primary human bronchial epithelial cells (Figure 8). This suggests that the high levels of active IL-8 present in the CF airways are likely to increase SC synthesis further. Additionally, because normal SC binds IL-8 and inhibits neutrophil chemotactic activity in vitro (20), an important self-regulatory loop is indicated, which may be defective in CF. The ELISA used to quantify SC measures both free SC and SC bound to IgA, that is, SIgA. The synthesis of SIgA requires cooperation between differentiated B cells in the lamina propria (IgA and J-chain production) and epithelial cells (pIgR). However, the presence of free SC in secretions from hypogammaglobulinemic and IgA-deficient subjects (17) demonstrates that free SC production can occur independently of SIgA transcytosis, and the release of free SC into saliva has been shown to be independent of its dissociation from SIgA (42). Thus, the high levels of SC observed in the CF airways may reflect upregulated production of pIgR, independent of SIgA transcytosis. The finding that the carbohydrate moiety of colostral SC was important for IL-8 binding (20) led us to investigate the glycosylation pattern of SC isolated from sputa. SC isolated from CF sputa was relatively overglycosylated compared with normal, which suggested that the monosaccharide composition, rather than the extent of glycosylation, was the factor determining the IL-8 binding capacity of SC. Fluorophore-assisted carbohydrate electrophoresis analysis revealed that free SC isolated from CF sputa was relatively overfucosylated and undersialylated in comparison to free SC from normal subjects. Glycosylation abnormalities have been suggested to result from hyperacidification of the Golgi and suboptimal sialyltransferase activity (43), expression of proinflammatory cytokines (44, 45), or alterations in the Golgi compartmentalization of glycosyltransferases (46–48), issues that remain to be resolved. Transfection of CFTR into CF airway epithelial cells has been shown to rescue glycosylation patterns of membrane glycoproteins and glycolipids (3). The study by Rhim and colleagues (1) found that as the expression of transfected wild-type CFTR decreased, the ratio of sialic acid:fucose reduced 8.6-fold, returning levels to those observed in the parent CF cells. We observed a similar 7.2-fold decrease in the ratio of sialic acid:fucose on free SC from CF compared with normal subjects. This is the first report of both increased fucosylation and decreased sialylation in a nonmucin molecule in CF and the first analysis of free SC isolated from the airways. The ratio of fucose:sialic acid in free SC isolated from normal sputa was 1.6, which is similar to the values of 1.0, 2.2, and 2.3 calculated from previously published data for SC from human milk (9, 15, 49). Our data also indicate 706.6 ng fucose/µg SC and 835.1 ng sialic acid/µg SC compared with previous reports of 12.4 ng fucose/µg SC and 15.9 ng sialic acid/µg SC for SC isolated from colostrum (12). Previous data on the monosaccharide composition of human SC (9–15, 49) analyzed SC purified from human milk or colostrum. The values for fucose and sialic acid content vary considerably between these studies, with estimates of fucose content from 5.1 mol/mol protein (12) to 19 mol/mol protein (14). Glycosylation in the mammary gland is under developmental and hormonal regulation (50), and the carbohydrate content and composition of specific proteins has been shown to alter significantly during the course of lactation (51). Therefore, differences in the monosaccharide composition of SC isolated from sputum and that from colostrum or milk are not unexpected. Alterations in the monosaccharide composition of glycoconjugates have been implicated in several disease states other than CF, including diabetes, rheumatoid arthritis, and ulcerative colitis (47, 52–54), although the physiologic implications of these changes have yet to be fully defined. In addition, terminal fucosylation of cell-surface glycoproteins is required for co-ordination of epithelial cell migration, as fucosidase treatment of primary airway epithelial cells completely blocks wound repair (55). Further understanding of the nature of the interaction between SC and IL-8 and the functional consequences of increased fucosylation and decreased sialylation of SC in CF may help in the development of novel antiinflammatory strategies.
The authors thank Dr. I. J. Lindley for his generous gifts of human recombinant IL-8 and anti–IL-8 antibody.
Supported by the Cystic Fibrosis Trust of Great Britain. This article has an online supplement, which is accessible from this issue's table of contents online at www.atsjournals.org Conflict of Interest Statement: L.J.M. has no declared conflict of interest; B.P. has no declared conflict of interest; K.B. has no declared conflict of interest; R.S. has no declared conflict of interest; A.B. has no declared conflict of interest; J.K.S. has no declared conflict of interest. Received in original form May 7, 2003; accepted in final form October 27, 2003
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
METHODS
0.05. 
function (
are known to upregulate the expression of pIgR (