With the rapid development of monoclonal antibody (mAb) and antibody-drug conjugate (ADC) pharmaceuticals, mammalian cell culture processes are continuously updated and optimized.
The biological activity of mAb drugs is determined not only by amino acid sequences but also significantly impacted by post-translational modifications (PTMs). Common PTMs include deamidation, disulfide bond formation, C-terminal lysine clipping, and glycosylation. Glycosylation, a complex PTM, mainly occurs in the endoplasmic reticulum (ER) and Golgi apparatus. Based on the mechanism of action of monoclonal antibodies, the glycosylation sites, glycan types and their abundance may affect the efficacy, safety and quality stability of products. Therefore, glycosylation is widely recognized as one of the critical quality attributes (CQAs) of mAb drugs.
Classification of Glycosylation
Protein glycosylation modifications are mainly divided into N-glycosylation and O-glycosylation. N-glycosylation takes place at the characteristic sequence NXT/S in the primary protein structure (where X is any amino acid except proline). O-glycans can attach to any amino acid containing a hydroxyl (-OH) group, with serine (S) and threonine (T) being the most common modification sites. For N-glycans, when nascent proteins enter the endoplasmic reticulum, the oligosaccharide unit Glc3Man9GlcNAc2 is transferred to the side-chain amino group of asparagine residues. Subsequent glycan processing in the ER and Golgi apparatus generates three major N-glycan types: high-mannose, complex, and hybrid types.
The biosynthesis of O-glycans occurs after N-glycosylation, folding and polymerization of proteins. O-glycans consist of a series of monosaccharide structures, most commonly N-acetylgalactosamine (GalNAc), N-acetylglucosamine (GlcNAc), xylose, mannose and fucose. Mucin-like O-glycan structures extend from the core galactosamine.
Not all N-glycosylation consensus sequences or serine/threonine residues are necessarily modified during glycosylation, resulting in site heterogeneity of glycosylation. Meanwhile, due to the competitive involvement of different enzymes in biosynthesis, each glycosylation site of both N-glycans and O-glycans usually presents distinct glycan structures, defined as micro-heterogeneity.
Given the profound impact of N-glycosylation on the stability, biological function and structural integrity of biotherapeutics, this chapter focuses primarily on N-glycosylation, particularly the N-glycosylation modification of monoclonal antibodies.
Biosynthesis of N-Glycans
The initiation of N-linked glycan synthesis relies on the donor structure Glc3Man9GlcNAc2, which binds to dolichol via a pyrophosphate bond. Dolichol inserts into the lipid bilayer in a helical or folded conformation. Glycan assembly on the head group of dolichol proceeds in two phases: the first occurs on the cytosolic side of the ER membrane, and the second within the ER lumen.
Enzymes catalyzing the binding of 2 GlcNAc residues and 5 mannose residues directly utilize nucleotide donors UDP-GlcNAc and GDP-Man. The lipid-linked glycan then undergoes transmembrane translocation, exposing the elongated glycan chain to the luminal side of the ER membrane for further monosaccharide addition. Dolichol-linked glycans serve as intermediate donors for the subsequent addition of 4 mannose residues and 3 glucose residues. These two dolichol-bound glycans are synthesized on the cytosolic surface of the ER membrane through the reaction of dolichol phosphate with UDP-Glc or GDP-Man. Under the catalysis of oligosaccharyl transferase (OST), the oligosaccharide is transferred from dolichol to the N-glycosylation site (asparagine residue) of nascent proteins, accompanied by correct protein folding.
Subsequently, during the transport of proteins from the ER to the Golgi apparatus, a series of specific enzymes cleave partial terminal monosaccharides of the glycan chain, including α-glucosidase I and α-mannosidase I/II. Meanwhile, other specific enzymes incorporate UDP-activated or CMP-activated monosaccharides into the glycan chain, such as N-acetylglucosaminyltransferase I/II/III (GnT I/II/III), fucosyltransferase, galactosyltransferase, and sialyltransferase. Proteins may exit the Golgi apparatus at different stages of glycan synthesis, forming diverse N-glycan patterns and leading to high heterogeneity of protein N-glycans.
Understanding the biosynthetic pathway of N-glycans is essential for recognizing why N-glycans serve as critical quality attributes of therapeutic protein products, and facilitates cell line engineering to modify the biological characteristics of therapeutic proteins. Different growth states of host cells and production processes result in distinct combinations of glycan types (i.e., glycan profiles) of therapeutic proteins. Thus, glycan profiles act as sensitive indicators of production process stability, and manufacturers are required to conduct routine N-glycan detection.
N-glycans can only be transferred from dolichol to asparagine residues of proteins under the action of oligosaccharyl transferase. To eliminate potential Fc effector functions of Atezolizumab, Roche adopted a strategy of mutating the asparagine residue at the Fc glycosylation site to another amino acid, as Fc N-glycans are indispensable for mediating Fc effector functions (detailed in Section 2 of this chapter).
In addition, to enhance the antibody-dependent cellular cytotoxicity (ADCC) effect of mAbs, fucosyltransferase genes in host cells can be knocked out or GnT III can be overexpressed to eliminate or reduce core fucose content in glycans. Supplementing culture media with appropriate concentrations of magnesium ions or butyrate can increase terminal galactose levels on mAbs, thereby enhancing complement-dependent cytotoxicity (CDC) activity.
N-Glycan Nomenclature
The structural complexity of N-glycans poses challenges to standardized description. Accurate characterization of N-glycans covers three key aspects: monosaccharide composition, linkage position, and anomeric configuration.
The most authoritative standardized nomenclature was proposed by the International Union of Pure and Applied Chemistry (IUPAC), which elaborates monosaccharide types, anomeric configurations and linkage positions in textual form. It features high accuracy but poor readability and convenience for application.
For the common mAb glycan G2FS1, IUPAC provides one linear description and three two-dimensional (2D) descriptions (complete, modified and simplified versions). Later, the Consortium for Functional Glycomics (CFG) and Oxford Glycobiology Institute successively proposed simplified nomenclature systems for N-glycans, which are now widely adopted in academic communication among glycobiologists.
The CFG nomenclature uses different colored geometric shapes to represent distinct monosaccharides, with textual annotations for anomeric configurations and linkage positions, offering greater intuitiveness than the IUPAC system. Since the linkage modes and positions of N-glycans in human endogenous proteins are relatively conserved, such details are often omitted in routine description, which is a common practice for characterizing N-glycan profiles of therapeutic protein products.
The Oxford nomenclature adopts unique geometric shapes for different monosaccharides; solid lines represent β-linkages, dashed lines represent α-linkages, and linkage positions are indicated by the spatial orientation of geometric shapes. This method is concise, intuitive and structurally accurate.
For most monoclonal antibodies, N-glycosylation is predominantly located at the asparagine residue around position 297 of the heavy chain within the CH2 domain, with relatively simple glycan patterns. Most mAbs are expressed in mammalian cell lines such as CHO, SP2/0 and NS0, with highly conserved anomeric configurations and linkage positions. Therefore, a simplified textual nomenclature is applicable for routine characterization.
Significance of N-Glycan Quality Control
In 2006, Genzyme obtained FDA approval for alglucosidase alfa (brand name: Myozyme), a replacement therapy for the rare genetic disorder Pompe’s disease.
To meet growing patient demand, Genzyme scaled up the manufacturing process from 160 L to 2000 L the same year. Although clinical trial data confirmed equivalent efficacy between the small-scale and large-scale products, the FDA deemed the N-glycan profiles of the scaled-up product significantly different from the original version and required a full biologics license application re-submission. The scaled-up process was finally approved by the FDA in 2008 under a new brand name Lumizyme. This case fully demonstrates the critical importance of N-glycan profile control for biotherapeutic products.
N-glycans profoundly affect the pharmacological efficacy of recombinant proteins. Erythropoietin (EPO) contains three N-glycosylation sites and one O-glycosylation site, and sialic acid at the non-reducing terminus of glycans is closely correlated with its in vivo half-life. Darbepoetin alfa, engineered by mutating 5 amino acid sites to introduce 2 additional N-glycosylation sites, exhibits a markedly prolonged in vivo half-life. Tissue plasminogen activator (t-PA) contains high-mannose glycans on its kringle and EGF domains, which accelerate in vivo clearance; targeted mutation of these glycosylation sites extends its half-life, enabling reduced dosage or less frequent administration.
The core fucose content of mAb Fc N-glycans is highly correlated with ADCC activity. Every 1% increase in afucosylated antibody content can elevate ADCC activity by tens of percentage points. N-glycan profiles also influence the immunogenicity of recombinant proteins. Cetuximab expressed in murine SP2/0 cells carries α-galactose residues at the glycan termini. Upon administration to humans, these residues bind to pre-existing anti-α-galactose antibodies and trigger severe hypersensitivity reactions, highlighting the necessity of strict control over non-human monosaccharide moieties on recombinant proteins.
Manufacturing processes exert a decisive impact on recombinant protein glycan profiles. Elevated levels of magnesium ions and butyrate in culture media can increase the abundance of G1F and G2F glycan isoforms on mAbs, further modulating CDC activity. Stable manufacturing processes ensure quality comparability before and after process changes, as exemplified by the alglucosidase alfa case mentioned above.
Moreover, biosimilar development requires detailed quality characterization of multiple batches of reference listed drugs (RLDs), with particular emphasis on glycan profile analysis. Manufacturing processes are then rationally developed to produce biosimilars with highly comparable glycan profiles to the reference product.
