Quantitation of the oligosaccharides of human serum IgG from patients with rheumatoid arthritis: a critical evaluation of different methods

Abstract

Several different chromatographic methods and a lectin-based assay have been compared for the quantitation of oligosaccharides released from immunoglobulin G (IgG). The analysis of a series of IgG samples purified from the serum of rheumatoid arthritis patients was carried out by these methods to evaluate the percentage of the glycoforms having 0, 1 or 2 galactose residues (G0, G1 and G2) in order to (a) identify the method that can be most widely used for quantitation, (b) accurately define the range of G0 values found in patients with rheumatoid arthritis, and (c) make available a series of characterised standards for distribution to clinical chemistry laboratories. The chromatographic methods involved: release of oligosaccharides by glycoamidase A after protease digestion followed by HPLC analysis of aminopyridine derivatives on reverse phase and normal phase columns; hydrazinolysis treatment with exoglycosidases (G0 mix) and Biogel P4 chromatography of 2-aminobenzamide (2-AB) derivatives; hydrazinolysis and weak anion exchange or normal phase HPLC of 2-AB derivatives; release of oligosaccharides by PNGase F and either Biogel P4 chromatography of 2-AB derivatives or HPAEC-PAD analysis of native oligosaccharides. The G0 values given by these methods compared favourably with each other and a dot blot assay of denatured IgG interaction with Ricinus communis agglutinin and Bandeiraea simplicifolia lectin II. The HPLC and HPAEC methods give additional information that may be important in less routine assays.

Introduction

Although oligosaccharides only account for 2–3% of the total mass of the IgG molecule, it has been shown that glycosylation of the Fc region is essential for optimal activation of effector mechanisms leading to the clearance and destruction of pathogens. This suggests that occupancy of the glycosylation sites of recombinant IgG antibody molecules by naturally occurring oligosaccharides will be an essential requirement for production of antibodies with biological effector activities. In addition, numerous studies have shown that while the defining biological activity of a glycoprotein molecule may not be dependent on its glycosylation, other essential characteristics may be compromised in unnaturally glycosylated forms, e.g., stability, pharmacokinetics and antigenicity (Gahmberg and Tolvanen, 1996; O'Connor and Imperiali, 1996; Bagriaçik et al., 1996; Sapan, 1997).

Human antibody molecules of the IgG class have N-linked oligosaccharide attached at the amide side chain of Asn-297 in both of the C_H2 domains of the Fc region (Deisenhofer, 1981). The oligosaccharide moiety is of the complex biantennary type having a heptasaccharide `core' structure (GlcNAc<sub>2</sub>Man<sub>3</sub>GlcNAc<sub>2</sub>) and variable outer arm `non-core' sugar residues, such as fucose, bisecting GlcNAc, Gal and NeuAc (Fig. 1). Thus, at least 36 structurally unique oligosaccharide chains may be attached at each Asn-297 residue. Analysis of monoclonal and polyclonal IgG demonstrates the presence of all the oligosaccharide species predicted in Fig. 1 (Parekh et al., 1985; Takahashi et al., 1995; Farooq et al., 1997), although more than 60% of normal human IgG oligosaccharides are accounted for by six main glycans (Wormald et al., 1997). In addition, it has been reported that 15–30% of polyclonal IgG molecules bear a complex N-linked oligosaccharide in the Fab region (Abel et al., 1968; Rudd et al., 1991; Youings et al., 1996; Wormald et al., 1997). It is apparent, therefore, that glycosylation is a post-translational modification that can introduce a significant structural and, possibly, functional heterogeneity into the IgG molecule (Dwek et al., 1995).

Changes in the profile of oligosaccharides attached to the Fc region of IgG isolated from the sera of patients with rheumatoid arthritis (RA) and some other inflammatory diseases has been reported (Parekh et al., 1985; Youings et al., 1996). The main finding is a reduction in galactosylation and consequently an increase in Fc oligosaccharides devoid of Gal, referred to as G0. This is mirrored in the IgG produced in lymphocytes of patients with RA (Bodman et al., 1992) and the IgG of lpr/lpr mice (Mizuochi et al., 1990) that spontaneously develop an RA-like disease. Aggregated agalactosyl IgG has been shown to be able to activate complement in vitro via an oligosaccharide-mediated interaction with mannose-binding protein (Malhotra et al., 1995a, Malhotra et al., 1995b). On the other hand, chain-terminating galactose has been shown to be important in binding of IgG by C1q (Tsuchiya et al., 1989) and Fc receptors (Lund et al., 1995). Furtherance of our understanding of the role and/or significance of differing Fc glycosylation patterns relies on the availability of accurate and sensitive techniques for the analysis of glycosylation. There remains much discussion about which system provides the best compromise between accuracy, sensitivity, ease of use and cost. This debate has been joined by the authors of the present study as a result of a collaborative EU-funded project (EUROCARB²), within which a panel of IgG preparations were analysed using different techniques. The purpose of this report is to present a critical evaluation of the protocols employed including isolation and characterisation of the IgG proteins and release of the oligosaccharide and their analysis.