Recovery and Purification Processing for Bispecific Antibody Production

Introduction

The biologics landscape is evolving into more innovative and enhanced functional configurations of which bispecific antibodies (BsAbs) make up the next generation of key immuno-oncology therapeutics. BsAbs target multiple antigens within a single molecule to increase tumor selectivity and thus promote immune cell recruitment to tumor cells. This unique molecular structure has led to a wide range of applications including blocking distinct signaling pathways, dual targeting of different disease mediators, and delivering payloads to targeted sites (Kontermann, 2015). Since BsAbs were initially described (Nisonoff, 1960), more than a hundred BsAbs are either currently on the market or under development (Labrijn, 2019; Sedykh, 2018).

The main formats of BsAbs can be categorized into IgG-like and non- IgG-like. IgG-like BsAbs retain the full function of the Fc region, thus engaging immune effector functions and have the advantage of increased serum half-life and stability (Kontermann 2015; Brinkmann, 2017). Emerging engineering approaches that build upon the bispecific structure include trispecific formats. Trispecifics can bind and engage Natural Killer (NK) cells in conjunction with T- and B-cell activation or strengthen T-cell activation for enhanced potency (Labrijin, 2019, Yun, 2018, Garfall, 2019, Gantke, 2017).

Initially, bispecific antibodies were synthesized using hybridoma technology (Milstein, 1984) or reduction and re-oxidation to couple antigen-binding fragments (Fabs) of two different antibodies expressed in different cell types (Brennan, 1985). However, these types of synthesized BsAbs were characterized by low production yield due to mispairing and thus technical difficulties in downstream purification. There have been many advancements in protein engineering to overcome these limitations and drive heterodimer formation. These include Heavy Chain (HC) steric engineering, such as knobs-in-holes (KiH) technology (Ridgway, 1996) (Merchant, 1998) and electrostatic interactions (Gunasekaran, 2010), common Light Chain (LC) fabrication (Merchant, 1998) and CrossMab (Klein, 2016) strategies to prevent mis-association. With these improvements in antibody engineering technology, the majority of therapeutic BsAbs are currently produced in mammalian cell lines. This allows similar monoclonal antibody/protein purification technologies, including depth filtration, chromatography, and membrane filtration to be applied in the production of drug substance (DS). Therefore, downstream process and manufacturing controls play a critical role to ensure the safety, quality, purity, and efficacy of BsAb products that consequently meet regulatory requirements for clinical trials and commercialization.

Overview of Bispecific Antibody Purification

Although potentially more therapeutically robust with enhanced functionality and improved safety profiles in contrast to monospecific antibodies, unique challenges arise in downstream purification of these molecules. Due to the complexity of BsAb architecture, many combinations of mispaired species can result, hindering correct heterodimer association (Wang, 2019; Krah, 2018). However, even with improved engineering approaches as mentioned above, unique panels of product-related variants can still be assembled during the cell culture process. These product-related impurities in expressed product pools include free HCs, LCs, homodimers, half molecules, and other mispaired HC/LC species that must be cleared and controlled when developing a downstream process for manufacturing BsAbs. Thus, the increasing complexity of bispecific formats give rise to more potential product-related species that can form and the need for a robust downstream process to clear these additional entities. Chromatography operations are favored technologies to selectively purify BsAbs by removing process- and product-related impurities. Affinity purification is one of the most valuable chromatography operations used in biotechnology industries. Protein A/G/L, KappaFabSelect, and LamdaFabselect, which bind to the Fc, or Kappa and Lamda Fab, respectively, have been used for IgG-like BsAb affinity purification (Chen, 2020; Zwolak, 2017; Ollier, 2019; Tustian, 2016, Fischer, 2015).

For most BsAbs, one affinity capture column step (usually Protein A) is insufficient to reach the level of purity required for clinical studies, and product-related variants (free HCs, LCs, homodimers, half molecules, and other mispaired HC/LC species) cannot be separated from the BsAb. Moreover, depending on the Protein A ligand properties, variable region-containing species can impact binding and strongly influence elution parameters (Ghose, 2005). Other chromatography strategies, such as Ion Exchange (IEX) (Gramer, 2013, Sharkey, 2017), Hydrophobic Interaction Chromatography (HIC) (Manzke O., et al 1997; Spiess 2013), Size-Exclusion Chromatography (SEC) (Schaefer, 2011, Huang, 2016, Zhao, 2014), and Mixed-Mode or Multimodal (MM) Chromatography (Zito 2016, Lee, 2017, Bertl 2019), have been applied to further provide separation from product-related variants for small scale material generation.

Although these various chromatographic tools are available to achieve high BsAb purity under linear gradient elution conditions, it remains a significant challenge to develop an economical downstream process platform which meets the strict regulatory requirements (Li, 2019). Following the affinity capture step, the polishing steps require significantly more development than a typical mAb due to the myriad of process-related variants (mentioned above) that need to be removed. Here we describe a robust downstream platform that was effective in generating high quality DS for three different IgGlike bispecific molecules despite harboring different formats. These include IgG (1 + 1), (2 + 1), and single chain variable fragment (scFv) formats, denoted by the number of binding sites, also described in (Labrijn, 2019).

Platform Purification Process for Bispecific Antibodies

In the biotechnology industry, the use of a platform approach for mAb purification has several advantages including rapid early-stage process development as well as resource- and cost-savings, particularly in raw material sourcing, technology transfer, analytical development, and regulatory fi ling. However, due to the aforementioned reasons, a downstream BsAb platform process that can be applied across various BsAb formats and structures has not been reported. We have developed a downstream purification platform for IgG-like BsAbs and applied to three distinct BsAb molecules. This permitted significant reduction of development time and cost-savings for early-stage development of BsAb entities in our pipeline.

The BsAb downstream platform process consists of harvest/clarification by depth filtration, affinity capture chromatography followed by low pH viral inactivation, two subsequent polishing chromatography steps, viral reduction filtration, ultrafiltration/diafiltration (UF/DF), and final DS filtration and fill steps, depicted in Figure 1. This downstream platform process provides effective separation of process- and product-related impurities, including residual host cell DNA and Host Cell Protein (HCP) and High Molecular Weight (HMW) and Low Molecular Weight (LMW) species, respectively. It also demonstrates good viral clearance capability, flexibility, and manufacturability. More importantly, this platform maintains the flow of a typical mAb platform thereby enabling easy facility fit and multi-product harmonization at manufacturing scale.

Harvest Clarification

The first step in the downstream BsAb platform process is to recover the BsAb from the cell culture bulk by removing cells, cell debris, and other process- and product-related impurities, such as DNA, lipids, LMW species, etc. Similar to the mAb harvest clarification process, depth filtration with in-line sterile filtration yields a clarified cell culture supernatant prior to chromatography purification steps. Three stages of filtration include primary (a large pore size range depth filter), secondary (a tighter and smaller pore size range depth filter), and tertiary sterile microfiltration (0.45/0.2 um) stages to separate the soluble products from the larger, insoluble cells and cell debris. For larger-scale production volumes of 2000 L or greater, diskstack continuous centrifugation coupled with depth filtration can be employed. Depth filter matrix, pore size, charge properties, and capacity need to be evaluated for individual BsAb. Table 1 compares harvest clarification yields for three BsAb molecules with different formats across lab- and pilot-scales. Similar depth filter trains were employed for all three BsAbs. The titers of depth-filtered clarified cell culture supernatant were reduced compared to bioreactor titers of cell culture bulk with centrifugation, which may be due to removal of some BsAb product-related variant species as well as dilution from the buff er chase. The step yields for each BsAb varied, potentially due to differing depth filter configurations and matrix, but were overall satisfactory and consistent across scales.

Affinity and Polishing Chromatography

Following clarification, a Protein A affinity column is employed as the capture step in this platform process to separate the target bispecific molecule from process- and product-related impurities. Under the typical production bioreactor cell culture conditions at neutral pH, the BsAb and product-related species containing the Fc-portion (HC, homodimer, half molecule and ¾ BsAb) bind to the Protein A column and are co-eluted with low pH elution buff er, while the process-related impurities (host DNA and HCPs) and non-Fc-containing product-related impurities flow through the column. Some non-specific weak binding species, such as LC and Fabs, may form non-covalent associations and co-elute with the BsAb as well (data not shown). Development work is required however to optimize the loading capacity for maximizing the step yield and the product purity for each BsAb antibody since non-ideal resin-binding capacities can negatively impact Protein A eluate yield and purity. Moreover, scFv, such as found in BsAb2, and three chain BsAbs tend to form aggregates (Cao, 2018), and elution buff er pH screening can optimize yield and minimize aggregate levels, which present challenges for polishing column purification if in excess. As presented in Table 2, under similar Protein A column conditions, the step yields were comparable, ranging from 81 – 94%. The SEC main peak purity of the Protein A eluate pools varied from 49 to 98% and CE-SDS Non-Reduced (NR) main peak purity from 56 – 94%, attributed to varying formats, structures, and stabilities of the BsAbs. In particular, BsAb1 and BsAb2 had high LMW levels due to a unique panel of product-related variants associated with these molecules. Residual DNA and HCP impurity levels were not significant challenges following the Protein A column step.

Subscribe to our e-Newsletters
Stay up to date with the latest news, articles, and events. Plus, get special offers
from American Pharmaceutical Review – all delivered right to your inbox! Sign up now!

Most BsAb purification processes will need at least two or more polishing chromatography steps to further remove product-related species including free HCs, LCs, homodimers, half molecules, and other mispaired HC/LC species (¾ BsAb, HC heterodimer), which are not normally observed in typical mAb Protein A elution pools. These unique BsAb product-related variants remained problematic for these molecules, and one affinity capture column step was insufficient to achieve the product quality required for clinical studies. In combination with IEX and MM, the two polishing chromatography steps utilized in this BsAb platform process, product-related species including homodimer and half molecule were effectively removed (Table 2). For BsAb1, the CE-SDS (NR) purity was improved from 71 to 90% post-IEX due to significant removal of homodimer, half molecules, ¾ BsAb, free LC and HC. These residual product variants were further removed by the second polishing MM column. The final CE-SDS (NR) purity for BsAb1 is > 97%. For BsAb3, IEX promoted further removal of HC, half molecules and HC heterodimer species, and subsequent MM chromatography achieved auxiliary separation from LC, half molecules, HC heterodimer and ¾ BsAb variants, resulting in 98% CE-SDS (NR) purity. Similarly, the CE-SDS (NR) purity for BsAb2 was remarkably improved from approximately 56% to 97% post-MM and -IEX chromatography polishing steps. Notably, after two polishing chromatography steps, all three BsAbs attained high purity and product quality required to meet the regulatory requirements for clinical trials. However, to achieve this level of purity for BsAb1 and BsAb2, lower step yields were a trade-off and just below 60% for one of two polishing steps (BsAb2) or both polishing steps (BsAb1). Moreover, endotoxin and bioburden levels were within the established specification limits (data not shown).

Viral Clearance

The capability of the BsAb downstream process for removal/ inactivation of endogenous and adventitious viruses from mammalian cell-derived BsAb products is an integral component of product safety assurance strategies. For mAb products, typical clearance values are approximately > 12 log10 for endogenous retroviruses (Xenotropic Murine Leukemia Virus (X-MuLV)) and > 6 log10 for adventitious viruses (Murine Minute Virus (MMV)) (Miesegaes, 2010). Each step of the platform purification process including Protein A chromatography, low pH Viral Inactivation, two polishing chromatography steps, and viral reduction filtration for BsAbs1-3 was assessed for its capacity to clear or inactivate viral challenge.

Log reduction values (LRVs) for each step are presented in Table 3. All purification process steps for each BsAb efficiently cleared both MMV and X-MuLV, with LRVs greater than 6 logs and 18 logs, respectively.

Conclusion

With the increasing number of bispecific antibodies entering various stages of development as the next generation of novel antibody therapeutics, biopharmaceutical companies increasingly focus on speed to clinic, efficiency, and cost-savings. The use of a downstream platform process enables rapid early-stage development without compromising quality and production. The downstream platform process, consisting of three column steps, that we developed for BsAb production has demonstrated robust clearance of process and product-related impurities, consistent process performance, reduction of development time, and major resource- and cost-savings. IEX and MM polishing chromatography, also supported in (Gramer, 2013; Lee, 2017; Wan, 2020), are very effective in removal of novel product-related variants associated with BsAbs. In particular, MM resins offer unique separation capabilities better tailored for clearing BsAb-related variants.

A platform process and engineering technologies including KiH, CrossMab, charge variant manipulations to promote HC-HC heterodimerization and HC-LC pairing (Gunasekaran, 2010), and modification of Protein A binding avidity (Tustian, 2016), aid in generation of high quality BsAbs. In addition, BsAb product quality-based clone selection approaches during cell line development are complementary solutions to overcome the challenges of therapeutic BsAb production.

References

1. Bertl S. et al (2019). Separation of bispecific antibodies and bispecific antibody production side products using hydroxyapatite chromatography. US 10316059 B2.
2. Brennan, M. D. et al (1985). Preparation of bispecific antibodies by chemical recombination of monoclonal immunoglobin G1 fragments. Science, 229 (4708), 81-83.
3. Kontermann, K. E. et al (2015). Bispecific Antibodies. Drug Discovery Today, 20(7), 838-847.
4. Brinkmann, U. et al (2017). The making of bispecific antibodies. MAbs, 9, 182-212.
5. Cao, M. et al (2018). Characterization and analysis of scFv-IgG bispecific antibody size variants. MAbs, 10(8), 1236-1247.
6. Chen, X. et al (2020). Protein L chromatograpy: A useful tool for monitoring/separating homodimers during the purification of IgG-like asymmetric bispecific antibodies. Protein Expr Purif, 175, 105711
7. Fahrner R. L. et al (2003). Development and operation of a cation-exchange chromatography process for large-scale purification of a recombinant antibody fragment. In: Rathore A. S. et al (Eds): Scale-up and optimization in preparative chromatography: principles and biopharmaceutical applications. Marcel Dekker, New York, USA 2003, pp-231-250.
8. Fischer, N. et al (2015). Exploiting light chains for the scalable generation and platform purification of native human bispecific IgG. Nature Communications, 6, 6113.
9. Gantke, T. W. et al (2017). Trispecific antibodies for CD16-directed NK cell engagement and dual-targeting of tumor cells. Protein Eng Des Sel, 30 (9), 673-684.
10. Garfall, A.L. et al (2019). Trispecific antibodies off er a third way forward for anticancer immunotherapy. Nature, 575, 450-451.
11. Ghose, S. et al (2005). Antibody variable region interactions with Protein A: Implications for the development of generic purification processes. Biotechnol Bioeng, 92(6), 665-673.
12. Gramer, M. et al (2013). Production of stable bispecific IgG1 by controlled Fab-arm exchange: scalability from bench to large-scale manufacturing by application of standard approaches. MAbs, 5(6), 962-973.
13. Gunasekaran, K. et al (2010). Enhancing antibody Fc heterodimer formation through electrostatic steering effects: applications to bispecific molecules and monovalent IgG. J Biol Chem, 285(25), 19637-19646.
14. Huang, Y. et al (2016). Engineered Bispecific Antibodies with Exquisite HIV-1-Neutralizing Activity. Cell, 165, 1621–1631.
15. Klein, C. S. et al (2016). The use of CrossMAb technology for the generation of bi- and multispecific antibodies. MAbs, 8(6), 1010-1020.
16. Kontermann, RE, et al (2015). Bispecific antibodies. Drug Discovery Today, 20, 838-847.
17. Krah, S. et al (2018). Engineering IgG-like Bispecific Antibodies-An Overview. Antibodies, 7, 1-10.
18. Labrijn, A. F., et al (2019). Bispecific antibodies: a mechanistic review of the pipeline. Nature Reviews, 18, 585-608.
19. Lee, Y. F., et al (2017). Modeling of bispecific antibody elution in mixed-mode cation exchange chromatography. J Sep Sci, 40(18), 3632-3645.
20. Li, Y. (2019). A brief introduction of IgG-like bispecific antibody purification: Methods for removing product-related impurities. Protein Expression and Purification, 155, 112-119.
21. Manzke O., et al (1997). Single-step purification of bispecific monoclonal antibodies for immonotherapeutic use by hydrophobic interaction chromatography, J. Immunol. Methods, 208, 65-73.
22. Merchant, A. et al (1998). An efficient route to human bispecific IgG. Nat Biotechnol, 16, 677-681.
23. Miesegaes, G. et al (2010). Analysis of viral clearance unit operations for monoclonal antibodies. Biotechnol and Bioeng, 106(2) 238-246.
24. Milstein, C. et al (1984). Hybrid hybridomas and the production of bi-specific monoclonal antibodies. Immunol. Today, 10, 299-304.
25. Nisonoff, A., et al (1960). Properties of the major component of a peptic digest of rabbit antibody. Science, 132, 1770-1771.
26. Ollier, R. et al (2019). Single-step Protein A and Protein G avidity purification methods to support bispecific antibody discovery and development. MAbs, 11(8), 1464-1478.
27. Ridgway, J. P. et al (1996). “Knobs-into-holes engineering of antibody CH3 domains for heavy chain heterodimerization. Protein Engineering, 9(7), 617-621.
28. Schaefer, W. et al (2011). Immunoglobulin domain crossover as a generic approach for the production of bispecific IgG antibodies. Proc. Natl. Acad. Sci. USA, 108, 11187.
29. Sedykh, S. E., et al (2018). Bispecific antibodies: Design, therapy, perspectives. Drug Des. Dev. Ther., 12, 195-208.
30. Sharkey B., et al (2017). Purification of Common Light Chain IgG-like bispecific antibodies using highly linear pH gradients. MAbs, 9(2), 257-268.
31. Spiess C., et al (2013). Bispecific antibodies with natural architecture produced by coculture of bacteria expressing two distict half-antibodies. Nature Biotechnol, 31 (8), 753-759.
32. Tustian, A. D. et al (2016). Development of purification processes for fully human bispecific antibodies based upon modification of protein A binding avidity. mAbs, 8(4), 828-838.
33. Wan, Y. et al (2020). Removing light chain-missing byproducts and aggregates by Capto MMC ImpRes mixed-mode chromatography during the purification of two WuXi Body based bispecific antibodies. Protein Expr Purif, 175, 105712.
34. Wang, Q. et al (2019). Design and production of bispecific antibodies. Antibodies, 8, 43.
35. Yun H. D. et al (2018). Trispecific killer engager CD16xIL15xCD33 potently induces NK cell activation and cytotoxicity against neoplastic mast cells. Blood Adv, 2, 1580-1584.
36. Zhao, L. et al (2014). Combating non-Hodgkin lymphoma by targeting both CD20 and HLADR through CD20-243 CrossMab. MAbs, 6, 740–748.
37. Zito, A. et al (2016). A Murine bispecific monoclonal antibody simultaneously recognizing β-Glucan and MP65 Determinant in Candida specifies. PLoS One, e014874.
38. Zwolak, A. et al (2017). Modulation of protein A binding allows single-step purification of mouse bispecific antibodies that retain FcRn binding. MAbs, 9(8), 1306-1316.