Lectin ELISA for Analysis of α1-Acid Glycoprotein Fucosylation in the Acute Phase Response

In recent years, increasing attention has been directed to the specific changes in glycosylation of glycoproteins that occurs in response to certain physiologic and pathologic conditions (1)(2)(3)(4)(5)(6). α1-Acid glycoprotein (AGP; orosomucoid) is a heavily glycosylated acute phase protein with five glycosylation sites carrying N-linked, complex-type oligosaccharides (N-glycans). The function of AGP is not well understood, but immunomodulating properties have been suggested that may be dependent on the expression of the sialyl LewisX (SLeX) epitope, in which fucose is a necessary component (7). However, methods for analysis of the glycosylation of acute phase proteins have been time-consuming and not suitable for routine analysis in a clinical laboratory (3)(6)(8). A lectin ELISA has been used previously for analysis of a tumor marker (9). The Aleuria aurantia lectin (AAL) was used in an ELISA for analysis of fucosylation of serum cholinesterase in liver disease (10). We developed a lectin ELISA that uses biotinylated AAL and a capture antibody specific for AGP to study the fucosylation on AGP in the acute phase response. As a model for acute inflammation, we monitored the daily changes in AGP fucosylation in patients with severe burns. The plasma concentration of AGP and C-reactive protein (CRP) was analyzed on a Cobas Integra 700 (Roche). AAL was purified from locally picked mushrooms as described previously (7). AAL fractions were pooled, lyophilized, and stored at −20 °C until biotinylation was performed according to the manufacturer’s instructions (ImmunoprobeTM Biotinylation Kit; Sigma). Biotinylated AAL was stored at 4 °C. Microtiter plates (Nunc-ImmunoTM Maxisorp) were coated with polyclonal antibodies directed against human AGP (anti-human orosomucoid, cat. no. A0011; Dako), diluted 1:100 in coating buffer (15 mmol/L Na2CO3, 35 mmol/L NaHCO3, and 0.2 g/L NaN3, pH 9.6), for 12 h at 4 °C. The following procedures were then performed at room temperature. Two hundred microliters of blocking agent [phosphate-buffered saline (PBS), pH 7.4 containing 50 g/L bovine serum albumin (BSA)] was added to the wells, and the plates were incubated on a shaker for 60 min. The wells were then washed four times with a washing solution (9 g/L NaCl, 7.5 g/L Tween). Patient samples (serum or plasma) were diluted 1:200 in PBS containing 10 g/L BSA. The diluted samples (100 μL) were individually added to wells and incubated on a shaker for 60 min. Samples were added in duplicates for each patient. The plates were washed six times with washing solution and incubated with 100 μL of biotinylated AAL diluted 1:100 in PBS containing 10 g/L BSA, incubated for 60 min on a shaker, washed six times, and incubated with ExtrAvidin (Sigma) diluted 1:1000 in PBS containing 10 g/L BSA. After incubation for 60 min on a shaker, the plates were washed six times and incubated with 100 μL of the substrate (1 tablet of ABTS dissolved in 5 mL ABTS buffer; Boehringer Mannheim) for 30 min on a shaker. Absorbances were read at 405 nm. The Statistica 5.1 software package (StatSoft) was used for statistical analysis. There was no difference between the lectin ELISA results for serum, EDTA plasma, or heparin plasma. The specificity of lectin binding was tested by addition of free l-fucose to the AAL solution before addition to the microtiter plate. Fucose inhibited lectin binding in a dose-dependent manner, with complete inhibition at a concentration of 0.1 g/L. Samples from 30 patients with increased AGP (range, 1.2–3.5 g/L) were analyzed both before and after adjustment of AGP concentration to 0.7 ± 0.2 g/L. A very minor effect of the AGP concentration on lectin binding was noted. However, to avoid any interference in a clinical study, the AGP concentration was adjusted to 0.7 ± 0.2 g/L in samples having an AGP concentration >1.1 g/L. To test for linearity, a pool of sera with highly fucosylated AGP (AGP-FR = 5.5) was diluted with the control pool (AGP-FR = 1.0). Both pools had the same concentration of AGP. A linear correlation (r2 = 0.98) was found in this interval, as shown in Fig. 1 A. The within-run coefficient of variation (CV) was 5% for the control pool and 3% for a sample with highly fucosylated AGP. The total CV for the control pool was 10%. Evaluation of linearity by dilution of a pool with highly fucosylated AGP with a control pool (A) and comparison of lectin ELISA with HPAEC-PAD (B). (A), the x-axis denotes the content of the highly fucosylated pool (percentage of total). (B), comparison of lectin ELISA AGP-FR with fucosylated AGP N-glycans according to HPAEC-PAD chromatograms (percentage of total). AGP fucosylation (AGP-FR) = (mean A405 patient sample − blank)/(mean A405 control pool − blank).<./> We compared the lectin ELISA results to high-pH anion-exchange chromatography with pulsed amperometric detection (HPAEC-PAD), which has been used previously to separate N-glycans released from patient AGP (3). The AGP-FR was compared with the ratio of fucosylated N-glycans calculated from the HPAEC-PAD chromatograms for eight different samples (AGP-FR range, 0.2–7.4), as shown in Fig. 1B . The ELISA response was low for AGP with <50% fucosylated oligosaccharides but increased several-fold in samples with 55–80% fucosylated N-glycans. This nonlinear relationship may be explained by the enhanced binding of AAL to multiply fucosylated AGP because AGP has five glycosylation sites and AAL is a dimeric molecule with two binding sites for fucose (11). In the experiment in which highly fucosylated AGP was diluted (Fig. 1A), multiply fucosylated oligosaccharides were present even at lower concentrations, providing linearity. The AGP-FR was determined in samples from 80 donors (40 men, 40 women; age range, 18–65 years), showing a log gaussian distribution; absorbances ranged from 0 to 1.1, with a median of 0.26 for men and 0.12 for women. The total median was 0.18, whereas the mean absorbance from repeated measurements of the pooled donor plasma used as the control was 0.19. The 2.5 and 97.5 percentiles expressed as AGP-FR were 0.2–3.5 and 0.1–2.8 for men and women, respectively. The difference in AGP fucosylation between men and women was significant at P <0.05 (t-test). This difference could be explained in part by the intake of oral estrogens by some female donors because estrogen therapy has been shown to decrease fucosylation and expression of SLeX on AGP (4). Patients with severe burns were monitored for AGP fucosylation and plasma concentrations of AGP and CRP every 24 h for 7 days after admission. The results are shown in Table 1. The median values showed a general increase for AGP and CRP during the first days after admittance, whereas there was no significant change in AGP fucosylation the first 3 days (Friedman ANOVA). In four patients, fucosylation initially decreased from day 1 to day 2, when it started to increase. There was no significant correlation between AGP fucosylation and AGP concentration when calculating the values for the first 3 days (Spearman r <0.1). This indicates, as suggested in a previous study of patients with burns, that changes in glycosylation of AGP do not correlate directly to the increase in the acute phase protein synthesis (12). Individual values for AGP fucosylation,1 AGP concentration, and CRP concentration in patients with severe burns. AGP-FR = (mean A405 patient sample − blank)/(mean A405 control pool − blank). Individual values for AGP fucosylation,1 AGP concentration, and CRP concentration in patients with severe burns. AGP-FR = (mean A405 patient sample − blank)/(mean A405 control pool − blank). AAL recognizes α-linked fucose irrespective of linkage position (11). However, AGP has been shown to exclusively carry fucose linked α1–3 to N-acetylglucosamine, and AAL does not bind to AGP isolated from individuals lacking fucosyl transferase VI activity, which is responsible for α1–3 fucosylation of AGP in the liver (7)(13). The carbohydrate epitopes LewisX (LeX, Galβ1–4[Fucα1–3]GlcNAcβ1-) and sialyl LewisX (SLeX, Neu5Acα2–3Galβ1–4[Fucα1–3]GlcNAcβ1-) both contain fucose in this linkage. However, in studies using monoclonal antibodies directed against these epitopes in analysis of AGP isolated both from healthy individuals and patients with different inflammatory diseases, only SLeX was detected (6). This suggests that AAL binding to AGP could be used as an indirect measurement of SLeX expression. Although monoclonal antibodies have been used in the study of SLeX expression on AGP, these antibodies are generally weak and not well suited for routine analysis. Expression of SLeX on circulating leukocytes is involved in mediating binding to activated vascular endothelium in the inflammatory response (14). It has been proposed that the enhanced, multivalent expression of this epitope on AGP during inflammatory conditions may have an immunomodulatory function by inhibiting SLeX-mediated cell-cell interactions (7). Therefore, the method presented, in which AAL is used to determine the fucosylation of AGP as an indirect measurement of SLeX expression, may be useful in future clinical studies and may potentially increase our knowledge of how AGP may affect the inflammatory response. Further studies are in progress in our laboratory to correlate fucosylation changes in AGP to clinical and other laboratory findings in patients with different inflammatory conditions. This study was supported by grants from the Health Research Council in the Southeast of Sweden, and the Swedish Medical Research Council (Grant 13X-2). The study was approved by the local ethics committee. Serum samples from patients with burns were kindly provided by Dr. Olof Ljunghusen. We thank Inga-Lill Björkholm, Inger Gustafsson, and Ingela Nilsson for skillful technical assistance.