Advanced Techniques for Improving mAb Purity and Efficacy

Sponsored JUNE 2024 22(6) BioProcess International 41 T he new generation of engineered antibodies has an increased tendency for aggregation, creating new challenges for downstream processing. Aggregates are large, tangled clusters of antibody molecules that can form irreversibly during upstream production, downstream processing, and storage (1). Extreme levels of pH, ionic strength, temperature, concentration, shear forces, and other processing conditions exacerbate aggregation. The resulting particles can expose different epitopes, which decreases overall product efficacy, increases immunogenicity, and (depending on particle size) even introduces the potential to block blood vessels in recipients (2). Given the importance of aggregate removal, shortcomings among standard techniques create the need for economical and effective purification solutions. A new chromatography resin offers unique selectivity with favorable process economics for robust removal of aggregates, host-cell proteins (HCPs), and leached protein A from antibody manufacturing processes. An Opportunity To Expand Operating Ranges Several methods are used for aggregate removal based on different separation principles including size, charge, and hydrophobicity (3). Although such approaches are effective, they also present a number of challenges. Ion-exchange chromatography (IEC) has a relatively low binding capacity, hydrophobicinteraction chromatography (HIC) resins are difficult to clean, and both size-exclusion and hydroxyapatite media present obstacles for packing columns and scaling up processes. On the next page, the left panel of Figure 1 summarizes typical operating conditions for chromatographic aggregate removal, with shaded areas representing conditions that provide for high yield and antibody purity with each approach. HIC in flow-through mode can process feed with a 5–15% aggregate levels at a mass loading of 65–200 g/L resin. Cation-exchange chromatography (CEX) in bind–elute mode can process feed with ≤5% aggregate at a maximum loading of 100 g/L resin; overloaded CEX resins can process similar aggregate levels at much higher mass loading. However, CEX gives low monomer recovery at low mass loading, so processdevelopment operations require substantial amounts of feed material. The right panel of Figure 1 shows corresponding operating conditions. Blue boxes define the opportunity to expand typical operating ranges using the POROS Caprylate mixed-mode CEX resin, which offers unique selectivity. Resin Features POROS Caprylate resin balances the removal of aggregates and other impurities, such as HCPs and leached protein A, with robust monoclonal antibody (mAb) recovery — a combination that is challenging for other mixed-mode resins operated in flow-through mode. The ability to operate in flow-through mode increases mass loading capability, requires lower volumes of buffer and resin, uses a smaller equipment footprint, and can be accomplished in shorter processing times than bind– elute mode takes. Additional process benefits are enabled by these product attributes: • Caprylic acid functionality enables high aggregate selectivity, reducing the number of chromatography steps needed while increasing process yield. SUPPLIER SIDE High Aggregate Levels with Engineered Monoclonal Antibodies An Innovative Approach to Addressing the Challenge Ying Chen and Al de Leon POROS Caprylate bead (50 μm) and ligand structure 42 BioProcess International 22(6) JUNE 2024 Sponsored • The 50-µm bead size and large through-pores provide high aggregate capacity for increased productivity and reduced cost. • The poly(styrene-codivinylbenzene) backbone offers predictable scalability with a linear pressure-flow curve. Optimizing Aggregate Removal and mAb Recovery To demonstrate the ability of POROS Caprylate resin to remove aggregates with high monomer recovery, we exposed a purified immunoglobulin G (IgG1) mAb to high and low pH multiple times to generate aggregates at levels up to 10%. Batch and column experiments started with highthroughput screening (HTS) to define and provide directionality on the effects of pH, salt concentration, and loading density on both aggregate removal and monomer recovery. HTS experiments included 200-g/L and 100-g/L loading densities with initial aggregate levels of 5% and 10%, respectively, at different pH levels and with salt concentrations. Results showed that aggregate removal was favored at low pH, and high monomer recovery was achievable with higher salt concentrations. Use of those conditions could improve efficiency by requiring fewer adjustments to the process stream following protein A affinity capture. We used a column-based design of experiments (DoE) study centered on conditions of pH 5.25 and 28.6-mS/cm conductivity to confirm aggregate removal and monomer recovery at 100-g/L loading and 10% aggregates. As Figure 2 shows, the operating condition for POROS Caprylate resin can be broadened depending on aggregate-level acceptability. Note that ≥95% reduction in HCP and leached protein A also were demonstrated. A loading-density study showed that POROS Caprylate resin facilitated effective polishing with high monomer yield and purity. The feed had 89.4% monomer purity and 10.6% aggregates, and residence time was three minutes. Figure 3 shows cumulative monomer recovery and aggregate percentage in the flow-through as a function of loading density. A 1% aggregate level corresponds to an 85.6 g/L loading density and 80.4% monomer recovery. Both the loading density and monomer Figure 1: Typical operating ranges for intermediate polishing steps reveal the opportunity (blue boxes) to expand these ranges using a chromatographic technique with unique selectivity; B–E = bind–elute, CEX = cation-exchange chromatography, FT = flow-through, HIC = hydrophobic-interaction chromatography. Load Aggregate (%) pH Mass Loading (g/L) Load Conductivity (mS/cm) 20 15 10 5 0 8 7 6 5 4 10 100 1,000 0 10 20 30 FT HIC FT POROS Caprylate Overloaded CEX B–E CEX FT HIC B–E CEX Overloaded CEX FT POROS Caprylate Figure 2: White area of contour plot represents operating conditions that resulted in an aggregate level of ≤1% and monomer recovery ≥75%; HCP = host-cell protein, LPrA = lipoprotein A pH Salt Concentration (mM) 500 400 300 200 100 0 4.60 4.85 5.10 5.35 5.60 5.85 100-g/L loading density with 10% aggregate ≤1% aggregate ≥75% recovery ≥95% HCP and LPrA clearance Figure 3: Monomer recovery and aggregate percentage in flowthrough fraction as a function of loading density Loading Density (g/L resin) Monomer Recovery (%) 100 80 60 40 20 0 0 50 100 150 200 250 300 350 Aggregate (%) 10 9 8 7 6 5 4 3 2 1 0 Aggregate Loading Density Monomer Recovery 1% 85.6 g/L resin 80.4% 2% 181.9 g/L resin 96.3% 3% 256.8 g/L resin 99.2% Sponsored JUNE 2024 22(6) BioProcess International 43 recovery increased with higher aggregate levels. A range of different buffer conditions also can be used, including lower pH and salt concentrations. The ability to use a lower salt concentration would eliminate the need for sample manipulation before the next chromatographic step. Comparing POROS Caprylate resin with a commercially available alternative that also is designed to remove aggregates in flow-through mode revealed that the POROS resin had 75% higher monomer recovery and 230% higher HCP reduction at a mass loading of 100 g/L using optimized buffer conditions. POROS Caprylate resin also enabled removal of many high-risk and challenging HCPs. Table 1 classifies those proteins according to their risk and difficulty to remove. Most identified HCPs were undetectable; others were reduced by at least an order magnitude. A Flow-Through Polish Step Aggregates will remain a challenge for the next generation of antibody products. POROS Caprylate resin offers a new tool to polish mAb products with aggregate levels of up to 20% while operating at a broad range of pH and conductivity and reducing both HCPs and leached protein A. Because the resin was designed to operate in flow-through mode, using it reduces processing times and buffer use — and thus overall costs — compared with IEC in bind–elute mode. References 1 Joshi V, Yadav N, Rathore AS. Aggregation of Monoclonal Antibody Products: Formation and Removal. BioPharm Int. 26(3) 2013; https://www. biopharminternational.com/view/ aggregation-monoclonal-antibodyproducts-formation-and-removal. 2 Lundahl MLE, et al. Aggregation of Protein Therapeutics Enhances Their Immunogenicity: Causes and Mitigation Strategies. RSC Chem. Biol. 2(4) 2021: 1004–1020; https://doi.org/10.1039/ D1CB00067E. 3 Vázquez-Rey M, Lang DA. Aggregates in Monoclonal Antibody Manufacturing Processes. Biotechnol. Bioeng. 108(7) 2011: 1494–1508; https:// doi.org/10.1002/bit.23155.  Al de Leon is a research and development manager (al.deleon@thermofisher.com), and Ying Chen is a research and development staff scientist (ying.chen@thermofisher.com), both at Thermo Fisher Scientific in Bedford, MA. POROS resins are pharmaceutical-grade reagents for manufacturing and laboratory use only. A range of different buffer conditions also can be used, including lower pH and salt concentrations. The ability to use a lower salt concentration would ELIMINATE the need for sample manipulation before the next step. Table 1: Host-cell protein (HCP) classification and relative signal intensity (total ion count) after the protein A step and following POROS Caprylate purification HCP Classification Identified HCP Detection in IgG1 Material After Purification with . . . Protein A Affinity Resin POROS Caprylate Resin High risk 8-kDa glucose-regulated protein (GRP78, BiP) 6.64 × 105 Not detected Alpha-enolase (2-phospho-D-glycerate hydrolysase 2.43 × 104 Not detected Cathepsin B (CatB) 1.00 × 106 Not detected Cathepsin L (CatL) 4.16 × 104 Not detected Cathepsin Z (CatZ) 7.52 × 104 Not detected Glutathione S-transferase P1 (GSTP1) 4.06 × 105 Not detected Lysosomal acid lipase (LAL) 2.67 × 105 Not detected Matrix metalloproteinase 19 (MMP-19) 2.08 × 105 Not detected Phospholipase B-like 2 (PLBL2) 1.67 × 105 Not detected Monocyte chemoattractant protein 1 (MCP-1) 1.72 × 106 1.02 × 105 Peroxiredoxin 1 (PRDX1) 4.20 × 105 1.12 × 105 Difficult to remove Cathepsin D (CatD) 8.43 × 104 Not detected Insulin-like growth-factor binding protein 4 7.46 × 104 Not detected Metalloproteinase inhibitor 2.08 × 105 Not detected Galectin-3 binding protein 2.15 × 105 3.31 × 104 Lipoprotein lipase 2.72 × 106 7.42 × 105 High risk and difficult to remove Clusterin (Clu) 2.24 × 107 1.56 × 106