The role of glycan characterisation in ATMP development and evaluation

Advanced therapeutic medicinal products (ATMP) aim to transform treatment strategies for unmet clinical needs. ATMP development requires product standardisation, high‐quality consistent starting materials, and safety and efficacy, while reducing adverse reactions. Glycosylation is a common post‐translational modification of glycans to specific peptide sites. Changes in glycosylation in starting materials or culture can potentially affect the bioactivity, safety and efficacy of glycosylated products. Glycan analysis is applied to few biological products, most commonly monoclonal antibodies and glycoconjugate vaccines. Regulatory authorities are starting to scrutinise the glycan profiles of biomedical products more carefully since glycosylation‐mediated activity has been recognised to lead to adverse effects in this class of therapies, including allergic/hypersensitivity reactions, anti‐drug antibodies and product heterogeneity in biologic and advanced treatments. This review highlights the impact of glycosylation in cell and tissue substrates for use in advanced therapies, identifying the need for glycan analysis to design safer, consistent products.


| INTRODUCTION
Advanced therapeutic medicinal products (ATMP) are increasingly developing to transform treatment strategies for unmet clinical needs. Currently, ATMPs face challenges around standardising products, quality and consistency of starting materials, the impact of safety parameters on efficacy and allergic reaction and rejection issues. Glycosylation-mediated activity has been shown to play a role in these challenges, yet no standard currently exists for glycan analysis in ATMPs. Glycosylation is a common post-translational modification involving enzymatic conjugation of carbohydrates (or glycans) to specific sites on proteins. Changes in protein glycosylation in the starting materials or culture can potentially affect the bioactivity, safety and efficacy of glycosylated biological products. 1 The diversity of these protein glycoforms must be characterised to reduce risks associated with glycosylated materials and products' biological activity and patient safety in the clinical setting. Many factors influence the glycosylation of biotherapeutics from species-specific materials and bioreactor processes to final purification. This applies to purified glycoproteins, where much regulatory interest has been focused on complex biological products such as advanced therapy medicinal products. Several reviews have considered the importance of glycan analysis in biopharmaceuticals in the context of purified glycoproteins. [2][3][4][5] However, an integrated overview of glycosylation considerations in complex biological systems such as ATMPs is lacking. An integrated approach to glycan analysis of ATMPs may offer a means of standardisation and quality control of these complex biological products. Glycosylation will play an essential part in light of next-generation drugs, especially in the context of ATMPs, for product characterisation, improved efficacy and sustained and targeted delivery. 3

CHARACTERISATION IN REGULATORY PROCESSES
Current standards for glycan analysis are applied to a few biological products, most commonly monoclonal antibodies and glycoconjugate vaccines. Other drugs may require glycan analysis if glycosylated motifs are shown to be instrumental in drug activity. Monoclonal antibodies (mAb) are a driving force of growth in the biopharmaceutical industry, enhanced by increasing biosimilars. 3 mAb glycosylation affects binding and effector functions, and thus regulations now require stringent glycan characterisation to ensure product safety and efficacy. With the expiry of patents on legacy mAb products, biosimilars have risen in market share. A biosimilar is a biological medicine, usually a mAb, that is highly similar to another already approved mAb (the 'reference medicine' or legacy product). 6 Guidelines regulating biosimilars emphasise glycan analysis and comparability. 6 Biosimilars must be demonstrated to be like the original biologic agent, with comparative characterisation studies to assess similarity and demonstrate proof for regulatory agencies. The biosimilar's post-translational modifications (PTMs) can determine immunoreactivity, potency, pharmacokinetic profile, clearance and immunogenicity. Enhanced immunogenicity can vary from sub-clinical to serious, lifethreatening complications. Therefore, characterisation of both the biosimilar and innovator products is required to evaluate immunogenicity. Biosimilar analytical comparability should identify and assess all differentiating features of a biosimilar compared to the legacy product, as outlined by the International Council for Harmonisation (ICH) of Technical Requirements for Pharmaceuticals for Human Use that sets international specifications for biotechnological and biological products supporting new marketing applications (ICH Q6B). 7 The EMA has issued a guideline on the development, characterisation, production and specification for mAbs and other related products as follows ( Figure 1 The carbohydrate content (including neutral sugars, amino sugars and sialic acids) should be determined. In addition, the structure of the carbohydrate chains, the oligosaccharide pattern (antennary profile), the glycosylation site(s) and occupancy should be analysed. Typically, monoclonal antibodies have one Nglycosylation site on each heavy chain located in the Fc region. The light chain is usually not glycosylated. However, additional glycosylation site(s) in the heavy chains may occur, and thus their presence or absence should be confirmed." Aberrant expression of glycan structures or variants that do not present in the legacy product must be characterised, specifying the degree of decoration and terminal monosaccharides, for example, mannose, galactose, fucose and sialic acid. Glycan species that vary from legacy products require justification, especially if they contain non-human epitopes.
Characterisation of final products alone is insufficient to capture the biological properties of variable PTMs in a product. To identify potential immunogenicity and adverse effects, further evaluation of product-related impurities that may arise in degradation products should be isolated and characterised. 7 This is especially relevant to glycan characterisation where conformational changes and motif presentation can alter immunogenicity. In addition, degradation products may occur during manufacture and storage, and such changes should be monitored and investigated appropriately within established acceptance criteria. The EMA proposes using high-performance liquid chromatography, capillary electrophoresis, mass spectroscopy, circular dichroism, and other analytical methods relevant to detect and characterise conjugated PTMs where indicated. 7 The integrity of biological products should include the evaluation of glycan stability during manufacture and storage to ensure retention of potency and established properties.
Another class of drugs that has required specific glycan analysis during market authorisation are glycoconjugate vaccines. 9 Glycoconjugate vaccines incorporate bacterial polysaccharides to protein carriers to induce a T-cell-mediated immune response. These products undergo rigorous testing, such as characterising bulk polysaccharides, including identity, % O-acetylation, purity, contaminants, molecular weight, endotoxin content and bioburden. Activated polysaccharides are investigated for the degree of activation, localisation of activation sites, batch consistency and stability. 9 Monovalent conjugates undergo routine tests, including polysaccharide identity, carrier identity, polysaccharide: protein ratio, "free" saccharide/protein contents and immunogenicity. Final vaccine products must undergo all aforementioned tests with additional sterility testing, excipients and adjuvant tests. 9 Enzyme replacement therapies have exploited glycandependent endocytosis through the mannose-6-phophate (M6P) dependent pathway. Glyco-engineering techniques have been developed to increase the M6P content of therapeutic enzymes. M6P can be used to decorate peptides, proteins and nanoparticles. Recombinant acid αglucosidase (GAA), an enzyme therapy, was approved under the brand 'Myozyme', which was then changed to 'Lumizyme' during scale up due to differing glycan profiles and by definition a different product. M6P content is measured by glycoprotein hydrolysis, derivatization and high-performance liquid chromatography (HPLC) analysis of isolated glycans. 10 Robust analytical methods such as this make it possible to differentiate variable glycan expression on biological products.
In recent years, interest in the glycan composition of biomedical products has grown among regulators, reflected by the issuance of technical guidelines that include specific glycan characterisation requirements (e.g., ICH Q5E and ICH Q6B). An emphasis has been placed on the products that express complex glycan profiles as these can drift out of previously defined specifications during batch processing and scale up production. A turning point was reached in 2008 when the Food and Drug Administration (FDA) rejected an authorisation application from Gen-zyme® (Myozyme®, (alglucosidase alfa as mentioned above) for the treatment of Pompe disease because glycosylation-specific differences in the drug from large batch production versus small batch was significant and was no longer considered a biological equivalent of the original material. 11 Considering these regulatory demands, manufacturers must carefully characterise product glycosylation and define its biological and clinical role in medication. 11 The FDA's decision on Myozyme® emphasises the evolving interest in glycans and the need to understand the role of glycosylation in material properties, stability, bioactive potential, immunogenicity and ultimately, the clinical safety and effectiveness of a drug. 12 The focus is shifting from purified glycoproteins alone to other classes of biological products.
The FDA has emphasised the importance of glycan analysis in viral vector-based gene therapies, such as adeno-associated vectors (AAVs). Specifically, the FDA recommends the inclusion of biochemical characteristics including, but not limited to, glycosylation sites. 13 Mass spectrometry-based characterisation of recombinant AAVs has demonstrated that differential glycosylation affects in vivo performance, product potency and yield in bioprocessing. 14 The role of glycosylation on viral vectors is not entirely clear and is highly dependent on the characterised glycosylation sites and the functionality of the sites.

CHARACTERISATION IN ATMPS
ATMPs include gene therapies, somatic cell therapies (sCTMPs) and tissue-engineered products (TEPs). These advanced therapies arise from many autologous, allogeneic, xenogeneic and synthetic sources and are usually introduced into the body to treat, prevent or diagnose diseases. Biomaterial products used in these therapies are often highly modified with PTMs, including phosphorylation, sulphation and glycosylation, which modulate properties and function. 15 This necessitates characterisation to establish a PTM function and tune the desired PTM profile to the target tissue. 1 Specific alterations in protein glycosylation have been recognised to lead to adverse effects in this class of therapies. Allergic/hypersensitivity reactions, 16 anti-drug antibodies that reduced efficacy, 17 product and potency heterogeneity have been documented in biologic and advanced therapies using allogeneic and xenogeneic materials (Figure 2). 18 Additionally, graft versus host disease (GvHD) is a significant complication that has been observed in cell therapies. 19 The human immune system generates responses to specific glycan motifs, including N-glycolylneuraminic acid (Neu5Gc) 20 and terminal α-1,3-linked galactose (αgal). 21 These glycosylated motifs occur naturally in cells and tissues originating from non-human primates. Their inclusion as PTMs on cell and tissue-based products thus presents antigenic potential and an immunogenic risk. Elimination of immunogenic glycosylated motifs from non-human materials is of significant interest for ATMPs. Human cells in culture can incorporate Neu5Gc through a salvage pathway, an effect that is propagated in vitro with inadvertent media supplementation of exogenous precursors to Neu5Gc. 20 Nonhuman cells used in industry integrate both Neu5Gc and α-gal into N-linked glycans, potentiating a glycosylation-mediated immunogenic response. 22,23 Hypersensitivity events are recorded in up to 3% of patients treated with biological products, where adverse events range from minor irritation to cardiovascular collapse. 24 This hypersensitivity to glycan immunogens is IgE antibody-mediated. 21,25 Interestingly, tick bites increase anti-α-gal IgE antibody titres due to α-gal motifs on salivary proteins, 26 explaining why diet-associated and cetuximab α-gal hypersensitivity incidence is higher in regions with higher tick populations. 27 Cetuximab hypersensitivity is more remarkable than other biologics that also contain α-gal, as cetuximab α-gal occurs in the Fab region, which is accessible to IgE binding. In contrast, other biologics like infliximab and palivizumab are glycosylated in the Fc region, which is less accessible to IgE binding. 28 While these specific modifications are analysed in detail in glycoprotein preparations such as mAbs, they are rarely considered in the setting of ATMPs.
Another specific adverse reaction, Acute GvHD (aGvHD), is observed after allogeneic hematopoietic stem cell transplantation (alloHSCT), a somatic cell therapy and the incidence of aGvHD is around 30%-50%. 19,29 Recent evidence has suggested that this complication arises partly due to aberrant glycan expression. 30 Glycan characterisation could significantly improve alloHSCT procedures by mitigating against complications, such as GVHD. 30 Current guidelines for cell characterisation do not capture the full heterogeneity of autologous and allogeneic cells. This translates to biological variability and inconsistent therapeutic responses. Bioreactor conditions influence the glycosylation of cell-expressed proteins, including dissolved oxygen tension, pH, temperature and nutrient supplementation. 3 To maintain glycan expression in cell therapy products, tight regulation of bioreactor conditions is necessary.
The EMA has released guidance on xenogeneic cellbased medicinal products 31 defined as: "medicinal products that contain viable animal cells or tissues as an active substance". While these guidelines set clear goals for xenogeneic materials, no specific glycan analysis is recommended despite the recognised immunogenic epitopes in these materials. Xenogeneic cells have demonstrated pro-immunogenic effects when implanted in an immunecompetent host exposed to the xenogeneic cells or their products. In contrast, medical devices derived from xenogeneic material undergo significant chemical crosslinking with glutaraldehyde to eliminate biologically active epitopes, as in bioprosthetic heart valves. 32 It is recommended that studies investigate the immune response to xenogeneic cells and their products when used in ATMPs. It is suggested that xenogeneic antigen deletion/modification should be pursued to reduce immune responses and induce tolerance.
TEPs use starting materials of variable origin and nature, determined by therapeutic indication, risk profile and intended effects, and material availability, cost and suitable grade. 18 The heterogeneous nature of these starting materials would benefit from scrutinised and innovative development approaches, including appropriate analytical techniques such as glycan profiling to characterise the material, determine material consistency and study identified unique risks and/or effects when incorporated into a TEP. A reflection paper on TEPs recommends further product characterisation. 10 Finally, drug delivery is an important aspect of ATMP development that is affected by glycosylation regulation. The role of glycosylation in the delivery, activity and breakdown of mAbs has been emphasised, glycosylation in ATMP delivery is an emerging challenge in ATMP development. As ATMPs become more complex integrated systems, the various cell and tissue types within these systems must be modified to interact with each other and in product-host interfaces to optimise therapeutic delivery of active molecules and cells. Many drug modifications have been described to create glycanconjugated systems for cell-specific delivery. 33 The truncation of N-glycosylation pathways in mammalian cells has been described as a method to control the glycosylation of secreted factors and proteins, 34 while glycan modifications of nanoparticles have been used across multiple disciplines to target aberrant lectin expression in disease for tissue localisation. 35 These approaches are likely to combine in the near future to create cell therapies to both target specific cells and tissues to deliver a discreet and highly regulated secretome using glycan engineering. In gene therapies, adeno-associated viruses F I G U R E 2 Glycan-mediated adverse effects characterised in ATMPs. Substrates, cells and tissues from non-human sources contain pro-immunogenic glycosylated motifs that induce pro-inflammatory responses. Glycan heterogeneity arises from the complex regulation of glycan synthesis and processing sensitive to species, cell type and culture/microenvironment conditions. This leads to variable glycosylation of materials, cells and final products, leading to varying product potency and clinical responses. Several glycan-mediated adverse reactions have been characterised, including graft versus host disease and tumorigenic transformation.
have been re-engineered to modulate capsid-glycan interactions for improved gene delivery. 36

RECOMMENDATIONS FOR ADVANCED CHARACTERISATION IN GLYCOSYLATED PRODUCTS
Glycosylation plays a role in product pharmacokinetics and can modulate the immunogenic properties of a material. Therefore, glycan characterisation protocols should be outlined in the designing and manufacturing of advanced therapies incorporating glycosylated components. Additionally, such an output would offer additional information regarding product consistency to support marketing authorisation applications. Manufacturers and ATMP developers should consider the need for glycan analysis early in systems that incorporate biosynthesis of glycoproteins, either in isolation or as complex products that have characterised glycosylation sites. Special attention should be paid to past examples where variable glycosylation of specific products has led to variation in potency and safety. Initial risk assessment for scale up should consider glycosylation stability of any product to avoid product variability that has occurred in the past. Careful consideration should be given to the appropriateness of glycan characterisation for specific substrates dependent on the intended and potential impact of altered glycosylation on bioactivity in each therapeutic context, for example, The functional dependency of localisation of alpha-galactosylated motifs on mAbs demonstrates the need for integrated glycoproteomics for glycan structure and localisation. On the other hand, lectin microarray profiles may be sufficient to screen glycosylation profiles of cell therapies to confirm consistent glycophenotype across passages and batches. This review calls for greater awareness of glycosylation in cell and tissue substrates for intended use in advanced therapies, identifying the need for glycan analysis in product manufacturing for better-informed risk assessment and quality control. Glycan characterisation of a subset of ATMPs throughout development may aid in the design of safer products and more efficient testing and manufacturing.