Improved manufacturability of mAb variants
In the pharmaceutical and biotechnology industry, development programs of therapeutic mAbs are often abandoned due to poor manufacturability. Poor manufacturability of an antibody includes low expression cultures or formation of aggregates. Developability studies are now being carried out early during the process to identify any problems associated with protein stability, solubility that is insufficient to meet dosing need or sensitivity to stress. These experiments include short-term stability studies at different temperatures, freeze-thaw cycles, forced degradation studies and viscosity determination at high concentration of samples (X. Yang et al., 2013). Considering that drug substance and drug product quality attributes are affected by minimal changes in the manufacturing process, all steps may be critical to material quality and patient safety. Thus, it is mandatory to minimize and control the level of antibody variants in the drug product.
As early as translation, variants can be formed in production cell lines through the misincorporation of amino acids, such as the incorporation of serine instead of asparagine in CHO cells. While both are neutral, polar amino acids with similar physical properties, this results in the generation of a different amino acid sequence, a phenomenon found to be caused by starvation of a particular amino acid in cell medium (Khetan et al., 2010; Parker, Pollard, Friesen, & Stanners, 1978; Wen et al., 2009; X. C. Yu et al., 2009). Despite this, most variants originate from post-translational modifications (PTMs) or are generated during the manufacturing process. Modifications of a drug are considered critical quality attributes (CQA), if they are linked to stability, activity, or efficacy (Jacobi et al., 2014). Major quality modifications are extensively described in literature (Jacobi et al., 2014; Sissolak, Lingg, Sommeregger, Striedner, & Vorauer-Uhl, 2019), and glycosylation variants are commonly known to influence pharmacokinetics, antigen binding and immunogenicity (Eon-Duval, Broly, & Gleixner, 2012; Mimura et al., 2018). A priori knowledge and glycoengineering advances of therapeutic IgG antibodies provide opportunities to optimize safety, functionality and efficacy of the drug (Sissolak et al., 2019). Examples of applying protein engineering techniques include amino acid exchanges to prevent aggregation or avoiding methionine residues in complementary-determining regions (CDRs) to prevent impactful oxidation. mAb oxidation occurs primarily on methionine residues, leading to more polar side chains by the formation of methionine sulfoxide. Additionally, replacement of N-terminal residues by glutamine can be designed to reduce the number of charge variants arising upon pyroglutamate formation (Beck, Wurch, Bailly, & Corvaia, 2010; Y. D. Liu, Goetze, Bass, & Flynn, 2011). Improvements of protein sequence modifications by in silico engineering and structure-based rational design simulations have provided advancements in predicting aggregation-prone regions, stability calculations and solubility of a monoclonal antibody (Arslan, Karadag, & Kalyoncu, 2019; Kuroda & Tsumoto, 2020; Sormanni, Amery, Ekizoglou, Vendruscolo, & Popovic, 2017). Furthermore, improvement of the thermodynamic stability of a monoclonal antibody can be monitored by experimental stress studies, including temperature, pH and protease incubations, to select molecules more resistant to aggregation (Arslan et al., 2019; Enever, Pupecka-Swider, & Sepp, 2015; Tesar et al., 2017). Monoclonal antibody structures may be altered when exposed to different temperatures, pH and stress conditions, causing unwanted products which may exhibit increased immunogenicity as well as reduced efficacy and activity (Cui, Cui, Chen, Li, & Guan, 2017). Further protein engineering approaches are beyond the scope of this review and have been described elsewhere (Chiu & Gilliland, 2016; Popplewell, 2015; C. Yang, Gao, & Gong, 2018).
Along with protein engineering efforts, the central goal of process development of biologics is the establishment of manufacturing technologies and processes that will generate consistency in different batches (Carson, 2005). Quality-by-design (QbD) was established and first developed by Dr. Joseph M. Juran in the early 1990s, who believed that quality should be designed into a product, and that most quality crises and complications arose from the way a product was designed (Juran, 1992). Along with GMP, regulatory agencies encourage risk-based approaches and have adopted the QbD principles since early 2000s (L. X. Yu et al., 2014). By providing guidance on pharmaceutical development, QbD aims at facilitating the design of products and processes, ultimately maximizing the product’s efficacy and safety profile while enhancing product manufacturability (Alt et al., 2016). The principles are described by the International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH) guideline documents, ICH Q8-11. The guideline documents cover: pharmaceutical development in ICH Q8, quality risk management in ICH Q9, pharmaceutical quality system in ICH Q10, and development and manufacture of drug substance in ICH Q11 (ICH ; Yu et al., 2014). The objectives of pharmaceutical QbD include: achieving meaningful product quality specifications based on clinical performance; increasing process capability and reducing product variability and defects by enhancing product and process design; facilitating the ability to identify root causes of failures; and improving product development and manufacturing efficiencies (L. X. Yu et al., 2014). Of the QbD milestones,critical quality attributes (CQA) identifies the crucial characteristics of a product required to ensure quality from a patient’s perspective. By identifying specifications or numerical ranges for CQAs, candidates may be ranked and discarded accordingly (Somma, 2007). Several QbD milestones are summarized in Figure 2.
For improved manufacturability, the QbD paradigm has been incorporated into mAb-specific aspects. Karlbert et al. (2018) reported the adoption of quantitative structure-activity relationship (QSAR)-type modelling by exploiting the structural characteristics of mAbs for directed QbD implementation, increasing both product and process understanding. For upstream processes, Nagashima et al. (2013) applied the QbD approach to mAb production in CHO cells, focusing on quality risk management using failure mode and effects analysis. By doing so, the authors identified potential critical process parameters (CPPs) and key performance indicators (KPI) that may impact quality attributes and gained further knowledge to mitigate any future cell culture-related issues (Nagashima et al., 2013). On a scale beyond single processes, Genentech participated in a pilot program launched in 2008 by the FDA in an effort to expand the implementation of QbD, leading to imperative improvement of QbD tools and concepts for subsequent products (Alt et al., 2016; Finkler & Krummen, 2016). Today, QbD activities are fully integrated into product development phases and have led to more robust, productive manufacturing processes with increased clinical efficiency (Gronemeyer et al., 2014). A recent report has also demonstrated the efficient implementation and feasibility of using QbD-based similarity assessment of a biosimilar mAb, using an adalimumab biosimilar to Humira® (Zhang et al., 2020).