Comparing Performance of New Protein A Resins for Monoclonal Antibody Purification

Friday, February 16, 2018

Abstract

Protein A chromatography has been used as the mAb capture step in the majority of FDA submissions. A recent review article on Protein A resin highlighted the advances in Protein A performance over the last 40 years, focusing on improvements in their capacity, productivity, and operational velocity. In this study, we compared the performance of recently launched Protein A resins with one of the most widely used Protein A resins for their dynamic and static binding capacities, and impurity clearance capabilities. Three of the tested resins demonstrated higher capacities and productivities using flow rates that are typical of Protein A operations. The clearance of host cell proteins and high molecular species were found to depend on the Protein A ligand and bead base chemistry. Implementing commonly used high salt intermediate wash buffers resulted in comparable impurity clearances and moderately increased yield losses. Despite the increased yield losses, the productivities of the newly available Protein A resins were 30-60% higher than conventional Protein A resins. The increased productivities of high capacity Protein A resins can better accommodate increasing bioreactor titers while providing comparable product quality attributes.

Keywords

Protein A, SpA, mAb, monoclonal antibody, affinity chromatography, purification

Introduction

The standard process for monoclonal antibody (mAb) purification from harvested cell culture fluid typically involves at least two chromatography steps: Protein A affinity and ion exchange chromatography.1,2 Protein A chromatography separates proteins from impurities based on a reversible interaction between the Fc portion of a mAb and a Protein A ligand immobilized on a chromatography matrix. Impurities flow through the matrix while the mAb stays bound to the column. The column is subsequently washed to achieve additional impurity removal. Product is eluted from the matrix by a reduction in pH. A typical Protein A chromatography step yields >99% pure mAb, (i.e. >99% free of process impurities such as growth media components, host cell proteins (HCP) and Deoxyribonucleic acid (DNA)3 ).

The first therapeutic antibody, Orthoclone, was approved in 1986 and utilized Protein A chromatography as a capture step in the purification process.4 Protein A was subsequently used as the mAb capture step in 86% of FDA submissions.5

Advances in media, strategies to produce and identify highly productive clones, and nutrient feeding strategies have made it possible to get expression titers of 1–5 g/L.6-8 In a recent review article, Bolton et al. reported continued improvements in the performance of protein A chromatography over its full history, as indicated by capacity, operational flow rate, and productivity (rate of mAb production per liter of resin) to provide insights into the reasons for its consistent use.9 As shown for 62 technologies, improvements in any technology is typically driven by product sales.10 The improvements in Protein A technology were attributed to the increased Protein A sales, which closely paralleled the sales of mAb therapeutics. Both increased rapidly in 2000 after the first major mAb therapeutics were approved and the markets were developed.9

Since 1978, both the performance and availability of Protein A resins have steadily increased.9 Newer beads are more rigid, yet highly porous, allowing fast flow rates, low pressure drops, and high mAb binding capacities.11 Changes to the number of Protein A domains and the amino acids in those domains has led to improved Protein A capacity, lifetime, caustic stability, and mAb Fc specificity along with milder elution conditions.12,13

While improved capacities and productivities of Protein A resin compliment improvements in cell culture productivity, there is limited literature describing the impact of these protein A improvements on impurity clearance and product quality. In this study, the capacity, yield, and impurity clearance of a range of available Protein A resins were evaluated using two therapeutic IgG1 molecules and three therapeutic IgG2 molecules. Outputs from the different resin evaluations included yield, pool volume, purity (based on levels of HCP and HMW species), and chromatographic profiles. In addition, static and dynamic binding capacities and pressure drops were measured.

Materials and Methods

Protein A Resins, Columns and Instruments

Five resins utilizing three base bead chemistries were tested. Three resins utilized cross linked agarose beads, one resin utilized hydroxylated methacrylate polymer beads, and one resin utilized silica-based beads. The properties of the resins are summarized in Table 1.

Table 1. Characteristics of Protein A resins evaluated.

All chromatography experiments were conducted using an AKTA Avant 150 purification system. Chromatography resins were packed in 0.66 cm ID Omnifit columns to a bed height of 10.0 ± 1.0 cm following vendor’s recommended protocol. The target compression factor range, based on vendor recommendations was 1.10 – 1.25. The packing solution and the packing flow velocity used were as recommended by the vendor for each resin. The packed columns were tested for their HETP and asymmetry using recommended parameters from the vendor including linear velocity, injection solution, and injection volume. Provance™ Protein A was a prepacked column received from the vendor. The acceptance criteria for HETP and asymmetry were ≤ 0.10 cm and 0.7 to 1.4, respectively. The columns were held at room temperature in a bacteriostatic storage solution between their uses. Protein concentrations were determined by absorbance at 280 nm (A280).

Antibody

Five Chinese Hamster Ovarian (CHO) cell derived antibodies were used in this study. Three were IgG2 isotypes (mAb1, mAb2, mAb4) and two were IgG1 isotype (mAb5 and mAb7). The molecular weight of deglycosylated antibody and their isoelectric point (pI) data are summarized in Table 2.

Table 2. Antibody type, Molecular weight, and pI

Dynamic Binding Capacity (DBC) Studies

To determine the dynamic binding capacity (DBC), the neutralized Protein A pool for each mAb was diluted to 10 g/L (unless otherwise specified), and the conductivity and pH were adjusted to typical Protein A load conditions for each mAb (as indicated in Table 3). The mAbs were loaded onto columns at residence times ranging from one to 30 minutes. The A280 of the load material was determined with the column offline and the amount of load required to achieve 10% of breakthrough (i.e. 10% of the A280 of load material) was recorded as the 10% breakthrough DBC (Qd).

Table 3. Protein A operating parameters for DBC and performance runs for all mAbs.

The 10% breakthrough Qd (mg of protein bound per mL of resin) was plotted as a function of residence time and then analyzed by the nonlinear relationship according to Eq. 1. In this relationship, Qinf is the binding capacity at infinite residence time, τ is the residence time, and K is fitting parameter with no physical significance. Qinf and K were determined using the Solver function in Microsoft Excel.19

Qd=(Qinf * τ)/(K+τ) (1)

Static Binding Capacity Studies

The static binding capacity (SBC) was determined by incubating the protein solution with the resin until equilibrium. The static binding capacity was determined using different load concentrations and conductivities. Dried resin was weighed into 2 mL microcentrifuge tubes and protein solutions were added to the tubes at fixed protein load (mg) to resin mass (mg) ratios. The mixture of protein solutions and resins were incubated at either room temperature or 37°C on an orbital shaker at 120 RPM for 48 h. The resin-protein mixtures were then transferred completely in to Ultrafree®-CL Centrifugal Filter Units. The filter units were spun using an Allegra® 6 Series Centrifuge operated at 800 g for 5 min. The resins were retained by the filters, while the protein flowed through. The protein concentration in the filtrate was quantified by A280. The concentration of protein in the load and the supernatant were used to determine the amount of bound protein, and thus the static binding capacity (mg of bound protein per mg of resin). A correlation between the dried resin mass and the gravity settled volume was determined, and taking in to account for the compression factor the resin mass was converted into the resin packed column volume, which was subsequently used to determine the binding capacity per volume of resin (mg of bound protein per mL of resin).

Performance at Target Loading

The process specific operating conditions (load, wash, and elution steps) are summarized in Table 3. Thawed and sterile filtered harvested cell culture fluid (HCCF) was used as the load material (mAb 1 and mAb2) to evaluate and compare the quality of eluate pool from the Protein A resins. The runs were performed at loading specific for each molecule and resin as summarized in Table 4, but not exceeding 90% of the Qdyn10% determined for each resin.

Table 4. Loading on three resins for performance run of mAb1 and mAb2

Load titer was determined using an Affinity Protein A UPLC Method employing a POROS® A 20 µm, 2.1 x 30 mm, 0.1 mL Column, and Waters H-Class UPLC with ACQUITY TUV detector. The yield was determined from the eluate pool protein concentrations measured using A280. Elution pools were neutralized with 2 M Tris-base and analyzed for high molecular weight (HMW) and Host cell protein (HCP) levels. SEC-UPLC was used to monitor HMW levels using an analytical size exclusion column (BEH200, 1.7 µm particle size, 4.6 x 150. HCPs were measured by ELISA.

Wash studies

mAb1 was selected for intermediate wash buffer screening studies. These experiments were performed as described for the performance runs in Table 3, except that two column volumes (CVs) of a unique intermediate wash buffer was employed for each experiment. A number of different intermediate wash buffers were tested for their ability to lower HCP levels on Toyopearl® AF-rProtein A HC-650F (Tosoh Toyopearl) loaded at 50 mg/mL. As a control, an extended equilibrium wash buffer was used on MabSelect SuRe™ loaded at 35 and 42 mg/ mL. Eluate pools were collected and analyzed for HCP content.

Results and Discussion

Dynamic Binding Capacity Studies

The DBC is used to determine column sizing, which in turn defines the cycling strategy and processing time for the Protein A capture step at a commercial scale. The DBC was measured at 10% breakthrough (the Qdyn10%) for the five resins and the five mAbs. The DBC at 10% breakthrough is plotted as a function of residence time for the five resins and the five mAbs in Figure 1.

Figure 1. Dynamic binding capacity at 10% breakthrough as function of residence time for fi ve resins and fi ve mAbs.

MabSelect SuRe LX™ and Tosoh Toyopearl resin had higher capacities than MabSelect SuRe™, which in turn was either slightly better than (for mAb2) or comparable (mAb7) to Provance™ Protein A resin. At shorter residence times, MabSelect SuRe PCC™ had higher capacities than all other resins for the molecules (mAb1, mAb2, mAb5, mAb7) that were tested. The higher capacity resins were the focus of further evaluation. Binding capacities are known to be strongly dependent on resin morphology and ligand density. As a rule-of-thumb, large pores enable fast mass transfer but have lower surface area, which may lead to lower equilibrium capacities. High ligand densities result in high capacities but reduced mass transfer, since the immobilized Protein A reduces the pore diameter.20 As a result, higher dynamic binding capacity (DBC) is achieved by balancing particle size, pore size and ligand density. The pore sizes for the 5 resins (Table 1) are comparable. Therefore significant differences between the resins due to pore size would be unexpected. MabSelect SuRe LX™ and MabSelect SuRe™ have the same particle size and pore size, while MabSelect SuRe LX™ has higher ligand densities (as claimed by the manufacturer). As a result, MabSelect SuRe LX™ exhibited higher dynamic binding capacities than MabSelect SuRe™ at longer residence times, while at shorter residence times they exhibited similar dynamic binding capacities due to pore diffusion limitations.

At shorter residence times, MabSelect SuRe PCC™ exhibited higher dynamic binding capacities than the other resins. MabSelect SuRe PCC™ has a similar pore size as MabSelect SuRe LX™ and MabSelect SuRe™, but its smaller particle size provides higher surface area. As a result, at shorter residence times when pore diffusion is a limitation, MabSelect SuRe PCC™ had higher capacities. This was also the case for Toyopearl® AF-rProtein A HC-650F, which has a smaller particle size than MabSelect SuRe LX™ and MabSelect SuRe™. However, at longer residence times, the Toyopearl® AFrProtein A HC-650F, MabSelect SuRe PCC™, and MabSelect SuRe LX™ exhibited comparable capacities that were significantly higher than MabSelect SuRe™. This could be due to the effects of surface area and ligand density.

Using the binding capacities (Qd) measured at various residence times, the capacity of each resin at infinite residence time (Qinf ) was calculated using equation 1. The Qinf of Toyopearl® AF-rProtein A HC-650F, MabSelect SuRe PCC™, and MabSelect SuRe LX™ were predicted to be comparable and 30% greater than for MabSelect SuRe™ (Figure 2).

Figure 2. Static binding capacity, Qinf and recovered capacity after batch binding for mAb5.

Among the high capacity resins listed above, we observed that the Qinf were lower for the IgG2 than the IgG1 isotypes. However, there was not a consistent difference in Qdyn10% between IgG isotypes at shorter residence times (up to 30 min), as demonstrated in Figure 1. For instance, at shorter residence times, binding capacities of the resins with the different mAbs were comparable, excluding mAb7, an IgG1, which had lower capacities on all resins. However this mAb exhibited similar or better capacities at longer residence times. It is possible that this mAb was more prone to self-interaction and therefore had slower mass transfer, however, the aggregate levels of this mAb were comparable to the aggregate levels of the other mAbs.

It has been reported that higher concentrations either improve21,22 or do not affect Protein A capacity.23-25 In this study, load concentrations of mAb5 from 1 g/L to 17 g/L were tested and the load concentration did not impact the capacity of Tosoh Toyopearl resin.

Static Binding Capacity Studies

The static binding capacities for a subset of resins were tested as described in the materials and methods section. Static binding capacities were 15-30% higher than the Qinf (Figure 2). The volume of resin suspension was corrected for the compression factor, and the static binding capacities are reported as normalized to packed resin volume. Since, the SBC studies were performed using resin suspensions, more surface area may have been available for mAb binding compared to packed resin. The packed resin loses about 20% of the mAb binding sites on the external surface due to steric hindrance from the adjacent beads. The excess protein during the SBC study also appeared to be loosely bound and was fully removed by the first wash step. The static capacity determined from mAb recovered during elution was consequently comparable to the Qinf.

The effect of various load conditions (conductivity, concentration, and temperature) on the static binding capacity of MabSelect SuRe LX™ was investigated using mAb5. The load conductivity and temperature had no significant effect (Table 5). This indicates that the capacity is likely limited by the number of available binding sites and not limited by binding strength, which may be improved by changing load conditions.26

Table 5. Static binding capacity of mAb5 on MabSelect SuRe LX™ as a function of load conditions.

Pressure-Flow Characteristics

The average pressure drop for mAb1 and mAb2 as a function of velocity for the different resins is provided in Figure 3. Pressure drop is affected by bead size, bed porosity, and by resin compressibility. MabSelect SuRe LX™ and MabSelect SuRe™ have similar beads and demonstrated similar pressure-flow relationships. MabSelect SuRe PCC™ has similar characteristics to other agarose resins but is composed of smaller beads and therefore had higher pressures at the various flow rates (Figure 3). The Toyopearl® AF-r Protein bead diameter is similar to that of the MabSelect SuRe PCC™ (PCC). However, the Tosoh Toyopearl resin had lower pressures at the different flow rates. It is possible that this was due to the higher rigidity of the Tosoh Toyopearl methacrylate beads.19 The PCC and Tosoh Toyopearl were packed to compression factors of 1.10 and 1.15, respectively for this experiment. The pressure drop with the silica bead based Provance™ Protein A was higher than that of the other resins (Figure 3).

Figure 3. Average of mAb1 and mAb2 pressure drop versus fl ow rate for the diff erent resins packed in 10 cm columns.

Performance at Target Loading

Three resins were evaluated for yield and pool volume with mAb1 and mAb2 (IgG2 isotypes) at 35 mg/mL and 39 mg/mL loading, respectively. Two high capacity resins, MabSelect SuRe LX™ and Toyopearl® AF-rProtein A HC-650F, and one standard resin, MabSelect SuRe™ were tested at loading as shown in Table 4. The higher loading corresponds to 90% of the Qdyn10% (DBC at 10% breakthrough) obtained from Figure 1. The levels of host cell proteins (HCP) and high molecular weight (HMW) species were measured in the elution pools. For mAb2 at 39 mg/mL loading, the yield was about 8% higher for the two high capacity resins. While there was a slight decline in the yield at higher loading for the two high capacity resins, overall, the higher loading resulted in an improvement in the productivity by ~22 – 27% for mAb2.

For mAb1 at 35 mg/mL loading, the yield was comparable or slightly better for the two high capacity resins. At higher loading, the yield for all resins exceeded 90% and were similar to their corresponding values at lower loading. The capacity for mAb1 was higher than mAb2, and allowed higher loading. This led to an increase in productivity of 20, 45, and 60% for MabSelect SuRe™, MabSelect SuRe LX™ and Toyopearl® AF-rProtein A HC-650F, respectively.

The pool volumes of mAb2 from the Toyopearl® AF-rProtein A HC-650F resin were higher (4 versus 2 column volumes, as shown in Figure 4a) and the pool HMW levels were lower (2 versus 1.5 percent, as shown in Figure 4a). The pool volumes for mAb1 were similar (1.7 column volumes, as shown in Figure 4b) for all three resins. The mAb1 pool HMW levels from the Tosoh resin were lower (3 versus 2 percent, as shown in Figure 4b). It is possible that the lower HMW levels observed using the Tosoh Toyopearl resin were due to higher ligand affinity or adsorption of HMW to the bead surfaces. The Tosoh Toyopearl resin is composed of hydroxylated methacrylic and the other resins are composed of agarose. The higher pool volume for mAb2 with the Tosoh Toyopearl resin could be due to tighter binding of mAb2 to the Tosoh ligand, and the higher elution pH used with mAb2 compared to mAb1. The HMW and HCP levels in the mAb1 and mAb2 elution pools from the different resin were not affected by loading.

Figure 4. Elution pool volumes and HMW content for (a) mAb2 and (b) mAb1 from the performance runs. The resins were loaded to 90% of DBC at 10% breakthrough.

For mAb1, levels of HCP were similar for two of the resins, MabSelect SuRe™ and MabSelect SuRe LX™ but higher for Toyopearl® AF-rProtein A HC-650F (Figure 5). It is possible that the difference was due to the different polymers that make up the beads. In addition, wash optimization may improve eluate pool HCP levels.27 Although, the HCP level in the mAb 2 eluate pool using Toyopearl® AF-rProtein A HC-650F was slightly higher than those of the agarose resins, it was low enough for removal by the subsequent polishing column step.

Figure 5. HCP levels in the eluate pools from the mAb1 and mAb2 performance runs. The resins were loaded to 90% of DBC at 10% breakthrough.

Wash Studies

The HCP levels in mAb1 eluate pool were above 3000 ppm using Toyopearl® AF-rProtein A HC-650F, and higher than those obtained using agarose resins (1200-1400 ppm for MabSelect SuRe™ at 35 – 42 mg/mL loading). While these elevated HCP levels have been shown to be effectively removed during the polishing step, they may cause resin fouling and affect the robustness of the overall downstream process. It is possible that using the equilibration buffer during the wash step limited the HCP removal.

A number of wash buffers were investigated for improvement of HCP removal on Tosoh Toyopearl resin. Buffers with pH values between 5 and 9 were tested. Buffers were tested with different additives including urea, sodium chloride, L-Arginine, and Isopropyl alcohol (IPA). The HCP removal was better for wash buffers at higher pH values for mAb1. The presence of additives improved the HCP clearance, but also resulted in significant yield loss (Table 6). For comparison across different resins, an intermediate wash buffer was selected that had a higher yield (preferably ≥85%) while maintaining HCP levels comparable to those obtained using MabSelect SuRe™. While the use of urea or a combination of urea and IPA significantly reduced the HCP level, the yield dropped to ~80% and 65%, respectively. The use of arginine as an additive worked the best at both pH 9 and 7.5, and we selected the use of arginine at pH 7.5 to avoid potential deamidation at higher pH values.28,29 Using this wash buffer condition, the Protein A operation was conducted for the four resins (including MabSelect SuRe PCC™). The resins were loaded to 35 mg/mL, and 90% of the Qdyn10% (Figure 1). Higher loading resulted in improved productivity (Supplementary Figure 1) despite the higher yield loss (4 – 8%) due to the implementation of intermediate wash step (Figure 6).

Table 6. Intermediate wash buff er screening for HCP reduction on Toyopearl® AF-rProtein A HC-650F loaded at 50 mg/mL
Figure 6. Eff ect of wash condition and loading on mAb1 Protein A yield.

 

Figure 7. Eff ect of wash condition and loading on mAb1 eluate pool HCP levels for four resins.

The HCP levels in the eluate pools were below 1000 ppm for the methacrylate and agarose based resins using the intermediate wash compared to 1600-2500 ppm and 3500-4000 ppm without an intermediate wash for agarose and methacrylate resins, respectively. The moderate decrease in yield using the intermediate wash may be acceptable given the improvement in the product quality combined with the higher productivity achieved with the high capacity resins.

Conclusions

While the sequence of steps used for mAb purification has not changed since the first therapeutic antibody was licensed, the performance of Protein A chromatography, as indicated by capacity, operational flow rate and productivity has steadily improved. The results presented in this work represent a comprehensive comparison of Protein A resins currently on the market with respect to static and dynamic binding capacities, productivities, yield, pool volumes, and HMW and HCP reduction capabilities. The differences in observed impurity removal capabilities were attributed to differences in the Protein A ligands and the base bead chemistries. It was also demonstrated that the use of an appropriately selected intermediate wash buffer could deliver comparable product quality attributes across different resins irrespective of the ligand and bead backbone material.

Acknowledgements

The authors acknowledge the help and assistance of Bingyi Yao, Mahsa Rohani, Caleb Chase, Benj Guzman, Nana Osei-Owusu, Anjanesh Venkatesh, Katherine Chaloupka, Catherine Grim, Justin Ladwig, Mihaela Trumbo, Ganesh Vedantham and Nick Keener of Amgen for their contributions to this work.

References

  1. D Kelley, B., et al., Weak partitioning chromatography for anion exchange purification of monoclonal antibodies. Biotechnology and bioengineering, 2008. 101(3): p. 553-566.
  2. Rathore, A.S., et al., Evolution of the monoclonal antibody purification platform. BioPharm Int http:// www. biopharminternational. com/biopharm/ Downstream+ Processing/Evolution-of-theMonoclonal-Antibody-Purification-/ArticleStandard/ Article/detail/827131, 2013.
  3. Ueda, S.K.Y., Comparing Protein A Resins for Monoclonal Antibody Purification. BioPharm International, 2013. 26(12).
  4. Lain, B., Protein A: The Life of a Disruptive Technology. BioProcess International, 2013. 11(8): p. 29-38.
  5. Miesegaes, G.R., et al., Monoclonal antibody capture and viral clearance by cation exchange chromatography. Biotechnology and Bioengineering, 2012. 109(8): p. 2048-2058.
  6. Chadd, H.E. and S.M. Chamow, Therapeutic antibody expression technology. Current opinion in biotechnology, 2001. 12(2): p. 188-194.
  7. Hacker, D.L., M. De Jesus, and F.M. Wurm, 25 years of recombinant proteins from reactorgrown cells—where do we go from here? Biotechnology advances, 2009. 27(6): p. 1023- 1027.
  8. Agrawal, V. and M. Bal, Strategies for rapid production of therapeutic proteins in mammalian cells. Bioprocess Int, 2012. 10(4): p. 32-46.
  9. Bolton, G.R. and K.K. Mehta, The role of more than 40 years of improvement in protein A chromatography in the growth of the therapeutic antibody industry. Biotechnology Progress, 2016.
  10. Nagy, B., et al., Statistical basis for predicting technological progress. PloS one, 2013. 8(2): p. e52669.
  11. Fahrner, R.L., et al., Performance comparison of protein A affinity-chromatography sorbents for purifying recombinant monoclonal antibodies. Biotechnol Appl Biochem, 1999. 30 ( Pt 2): p. 121-8.
  12. GE Healthcare, MabSelect SuRe™. Data file 11-0011-65 AC.
  13. Ghose, S., et al., Antibody variable region interactions with Protein A: Implications for the development of generic purification processes. Biotechnology and Bioengineering, 2005. 92(6): p. 665-673.
  14. GE Healthcare, MabSelect SuRe. Data file 11-0011-65 AC.
  15. GE Healthcare, MabSelect SuRe™ LX. Data file 28-9870-62 AA.
  16. GE Healthcare, MabSelect SuRe™ PCC. Data file 29177558 AA.
  17. Grace, Grace-ProVance-eBook. https://grace.com/pharma-and-biotech/en-us/ Documents/Grace-ProVance-eBook.pdf.
  18. LLC, T.B., Catalog - Process Media, Protein A Chromatography. http://www.separations. us.tosohbioscience.com/Products/ProcessMedia/ByMode/ProteinA/ToyopearlAFrProteinAHC-650.htm.
  19. McCaw, T.R., E.K. Koepf, and L. Conley, Evaluation of a novel methacrylate-based protein a resin for the purification of immunoglobulins and Fc-fusion proteins. Biotechnology progress, 2014. 30(5): p. 1125-1136.
  20. Hahn, R., et al., Comparison of protein A affinity sorbents: II. Mass transfer properties. Journal of Chromatography A, 2005. 1093(1): p. 98-110.
  21. Genscript, Monofinity A Resin for Antibody Purification 2010: PBM140721.
  22. Müller, E. and J. Vajda, Routes to improve binding capacities of affinity resins demonstrated for Protein A chromatography. Journal of Chromatography B, 2016.
  23. GE Healthcare, Dynamic binding capacity study on MabSelect SuRe LX for capturing hightiter monoclonal antibodies. Application Note 28-9875.
  24. Preisig, C., Innovations In Purification: Lowering Cost And Accelerating Drug Development With Pre-Packed Columns For Single Use. Bioprocess Online, 2014.
  25. Chakrabarti, A., et al., Characterization of a Newly Developed High Capacity, Alkaline Resistant, Recombinant Protein A Resin. Tosoh Biosciences TP198, 2014.
  26. Bolton, G.R., et al., Inactivation of viruses using novel protein A wash buffers. Biotechnology progress, 2015. 31(2): p. 406-413.
  27. Shukla, A.A. and P. Hinckley, Host cell protein clearance during protein A chromatography: development of an improved column wash step. Biotechnology progress, 2008. 24(5): p. 1115-1121.
  28. Wakankar, A.A. and R.T. Borchardt, Formulation considerations for proteins susceptible to asparagine deamidation and aspartate isomerization. Journal of pharmaceutical sciences, 2006. 95(11): p. 2321-2336.
  29. Pace, A.L., et al., Asparagine deamidation dependence on buffer type, pH, and temperature. Journal of pharmaceutical sciences, 2013. 102(6): p. 1712-1723.
  • <<
  • >>

Join the Discussion