BioPharm

16 BioPharm International eBook October 2018 www.biopharminternational.com Biopharma Laboratory Best Practices Glycan Analysis The most common prep methods include fractionation, enrichment and release of analytes, chemical or enzymatic fragmentation, derivatiza- tion to enhance detection, or sep- aration and cleanup. These can be performed in various permutations and combinations. The methods must be reliable and validated and should be stan- dardized as far as possible. When developi ng t hem you need to balance quality, speed, through- put, ease, and cost to suit your needs. For exa mple, at Ludger we prioritise quality above other considerations because we sup - port regulatory submissions and also develop methods for clients for their drug batch release. To achieve high quality, we aim to design robust, standardized sample prep workflows that have high effi- ciency, minimum steps, and that preser ve the stoichiometr y and structural integrity of analyte spe- cies. We also seek to reduce inter- fering contaminants such as salts, detergents, proteins, peptides, and en z y mes and w ill va lidate t he methods with a wide range of ana- lyte types. We then optimize the other factors as far as we can, pro- vided that in doing so we main- tain quality. We ensure that our methods are straightforward for manual operation as well as when we automate them and add paral- lelism to increase reliability and sample throughput. For costs, we consider material costs of kits and reagents as well as overheaded labor for hands-on time and instrument use. Finally, speed is important, but we don't go for rapid methods or cut out important steps such as sample cleanup if it means compromising quality. For example, with PNGase F treatment to release N-glycans we use 10-minute reac- tions for [quality-by-design (QbD)]- type work but some glycan species have slow release kinetics—so for lot release and regulatory work we may increase the reaction times up to 18 hours. "As with any analytical method, appropriate validation and controls should be included." — Aled Jones, ProZyme Jones (ProZyme): The complexity of the sample preparation will depend on the level of glycan analysis taken. As with any analytical method, appropriate validation and controls should be included. BioPharm: How does one set speci- fication limits on glycans? W i d d o w s o n ( T h e r m o F i s h e r Scientific): For all types of biophar- maceutical glycoprotein, safety and efficacy are critical, and the spec- ification limits should be set with this in mind. All quality attributes should be risk assessed to determine whether they should be deemed as critical; and thus, should be rou- tinely monitored. Examples of criti- cal glycan attributes include levels of high mannose glycans, which have an effect on clearance (4), and levels of immunogenic moieties such as the α-Gal (Galα1-3Gal) moi- ety, which is commonly expressed when murine cell lines are used (5). Knowledge of the clinical effects and safety profile of the molecule is required for appropriate specification limits to be set in the development of innovator products. In biosimi- lar development, specification limits are set through analysis of multiple batches of innovator material. Jones (ProZyme): [The specifica- tion limit] can depend on the mol- ecule and the biological role of the attached glycans. For biosimilars, it's been seen in some examples that glycan profiles may vary slightly compared to an innovator, but cell- based assay and [pharmacokinetics] data ultimately show the difference to have limited clinical relevance. However, in those cases, the glycan differences were documented and explained, which is critical. Fernandes (Ludger): This is an area where biopharma companies can get into serious trouble with regulators. We address this in our GlyShape pro- gram, which provides free-of-charge training for biopharma companies on how to implement streamlined QbD for drug glycosylation. Our approach builds on the premise that to set meaningful glycan speci- fications you must understand the impact that your drug's glycosylation patterns and likely variations have on its clinical performance. This requires knowledge gained from GlycoSARS (Glycoform Structure- Activity Relationship Studies). Our focus is on those glycoforms bearing either of two types of glycosylation critical quality attributes (GCQAs): non-safe glycoforms (i.e., those with glycosylation features such as non- human glycan motifs that could trigger adverse reactions) and active glycoforms (which have glycans that confer high or medium ther- apeutic activity). We then develop mathematical models describing the clinical performance of two entities: the glycoform population and corre- sponding glycan patterns of the ide- alized drug batch (this relates to your quality target product profile, QTPP, and serves as a reference spec) and the range of glycoform distributions and corresponding glycosylation patterns resulting from any man- ufacturing process in your design space (DS, which relates to accept- able variations around the reference). The models are like a relief map asso- ciating clinical performance with

• Cover