Comparing Protein A Resins for Monoclonal Antibody Purification

With greater economic pressure on monoclonal antibody (mAb) production for therapeutic and research uses, antibody titers in mammalian cell culture have increased dramatically over the past 20 years. As a consequence, downstream processing must accept and handle higher titers of mAbs in harvested cell-culture fluid (HCCF), and vendors of mAb purification technologies must develop chromatography resins with high binding capacity to meet the demand. In addition, more cost-efficient cleaning procedures are necessary to extend the lifetime of chromatography resins and to reduce the cost for cleaning and validation. A team from Chugai Pharmaceutical (Japan) investigated the mAb purification performance of a new alkali-tolerant, prototype Protein A resin (Resin 3, MabSelect SuRe LX prototype, GE Healthcare), which has the potential to address the demand for a more advanced, cost-effective mAb purification technology.

The methodology for the production of monoclonal antibodies from a cell line by hybridization of mouse myeloma and mouse spleen cells from an immunized donor was first published in 1975 (1). As a technology that permits the generation of monoclonal antibodies against almost any target molecule, it immediately gained great interest as a source for potential drug candidates. Although it took some time for the first therapeutic mAb to become commercially available in 1986 (2), the market for therapeutic mAbs has since grown rapidly. MAbs have proved to be successful as targeted therapeutics for a variety of diseases, including several forms of cancer, multiple sclerosis, and immunological disorders such as rheumatoid arthritis and psoriasis. In 2007, mAbs accounted for almost half of the top-20 best-selling biotechnology drugs in the US alone, establishing them as an important group of molecules (3). Today, mAbs constitute the single largest class of biological drugs and accounts for about 36% of the total biologics market with an annual sales growth rate of approximately 10% (4).

Commercial-scale production challenges
The rapid growth in mAb demand has triggered industry efforts to increase manufacturing capacity, with the consequence that the antibody titers in mammalian cell culture have increased dramatically. Today, a typical process accumulates titers of 1-5 g/L, but expression levels as high as 10-13 g/L have been reported (5). The increase in upstream productivity creates a subsequent demand on downstream processing to address high-titer HCCF.

Commercial-scale purification of mAbs usually contains two or three chromatographic steps. Protein A is the affinity chromatography ligand of choice for the first antibody capture step, because its high selectivity gives excellent purity (typically > 99%) and high yields. Furthermore, Protein A-based resins form the basis of almost all mAb-purification platforms as they are easy to use at both small and large scale with generic experimental protocols.

Increased antibody titers create a potential purification challenge because of the limited capacity of current Protein A resins. To handle the high titers, new resins with significantly greater capacity are needed. In addition, Protein A resins with the ability to withstand repeated cleaning-in-place (CIP) with low-cost sodium hydroxide (NaOH) considerably improves process economics.

Protein A Resins
Protein A is a bacterial protein from Staphylococcus aureus, with the capacity to bind mammalian antibodies of class immunoglobulin G (IgG) with high affinity. The gene for Protein A has been cloned and expressed in Escherichia coli (6, 7) allowing for the production of large quantities of recombinant Protein A.

Although recombinant Protein A is widely used as an affinity ligand for the capture and purification of antibodies, its sensitivity to alkaline conditions prevents the use of rigorous and cost-effective CIP and sanitization protocols based on NaOH.

Compared to conventional Protein A resins, one of the affinity chromatography resins investigated, Resin 2 (MabSelect SuRe, GE Healthcare) is based on a modified alkali-tolerant Protein A ligand. Through protein engineering, the amino acids in one of the IgG-binding domains particularly sensitive to alkali were identified and substituted with more stable ones.

A novel prototype Resin 3 offers an increased dynamic binding capacity (DBC) at a slightly longer residence time.  

Evaluating capacity and reusability
Table I shows the Protein A resins that were compared for performance and cost-efficiency. According to the vendor, Resin 3 exhibits higher DBC than Resin 2 at longer residence times (8). A residence time of 6 min is expected to give a DBC (at 10% breakthrough) of approximately 60 g antibody per liter resin. In this study, the authors were able to confirm this behavior, and for further studies a residence time of 6 min and a loading of 50 g antibody per liter resin, corresponding to approximately 80% of the DBC (at 10% breakthrough), were selected.

Table I. Properties of Protein A resins; degree of alkali resistance is indicated by +/-.

Ligand

Average particle size (µm)

Alkali resistance

Matrix

Binding capacity*

Resin 1 (rProteinA Sepharose 4 Fast Flow)

rProtein A

90

+/-

Agarose

~27 g/L resin

Resin 2 (Mab Select SuRe)

Alkali-tolerant Protein A

85

+++

Agarose

~35 g/L resin

Resin 3 (Mab Select SuRe LX prototype)

Alkali-tolerant Protein A

85

+++(+)

Agarose

~60 g/L resin

* Typical dynamic binding capacities according to resin manufacturer data

Table II presents the outline of the lifetime study. An amount of cell-culture supernatant corresponding to 50-g antibody per liter resin was applied to the column at 6 min. residence time. This was followed by a two-step washing procedure to remove unbound particles. Bound antibody was eluted with 50 mM acetic acid and the column resin was cleaned in place with five column volumes (CV) of 0.1 M NaOH.

Table II. Study outline. Column height = 100 mm and inner diameter = 10 mm, run time = 5 h, CV is column volume.

Solution

Volume

Flow rate

Equilibrium

20 mM citrate-phosphate buffer
pH 7.5, 1 mol/L sodium chloride (NaCl)

5 CV

300 cm/h

Load

Harvested cell-culture fluid

50 g/L resin
(residence time 6 min)

100 cm/h

Wash 1

20 mM citrate-phosphate buffer
pH 7.5, 1 mol/L NaCl

5 CV

300 cm/h

Wash 2

10 mM citrate-phosphate buffer
pH 7.7

5 CV

300 cm/h

Elution

50 mM acetic acid

6 CV

300 cm/h

Regeneration

0.1 mol sodium hydroxide

3 CV

120 cm/h

Storage (per 4 cycles)

2% benzyl alcohol, 50 mM sodium acetate, pH 5.0

5 CV

120 cm/h

The results show that the step yield was consistently over 95%, and high log-reduction factors of host-cell proteins (HCP) and DNA were achieved (see Figure 1). In this study, carryover was evaluated each 28th cycle and was found to be less than 0.1% (i.e., after cycle 28, 56, 84, and 112). The lifetime study with mAb-containing feedstock demonstrates that the product quality, DBC, and yield with Resin 3 were stable for more than 100 purification cycles. No increase in pressure was observed during the study.

Figure 2: Cost-performance of Resin 3 prototype compared to conventional resins (Resin 1 and Resin 2). Product amount is 500 kg, fermenter size is 10,000 L (for 1 g/L) or 5000 L (for 3.5 g/L), column size is 20 cm bed height, column lifetime is 120 cycles for Resin 1 and 200 cycles for Resin 2 and 3, process time is 10-15 h.

With Resin 3, in the case of a titer level of 3 g/L, the authors found that the overall purification cost can be even further reduced by 26%. With a lower titer level of 1 g/L, however, the decrease was only 4%.

Under the selected conditions, Resin 1 was not suitable for purification of antibody from a 5 g/L titer. The purification cost per kilogram of produced antibody from a 5 g/L titer using Resin
2 (7000 USD) can be reduced by 32% by using Resin 3 (see Figure 2).

These results show that the use of Resin 3 in purification of antibodies from high-titer feeds significantly improves process economy (see Figure 2).

Summary
These data demonstrate that the Resin 3 prototype has high capacity and reusability with stable step yield and impurity clearance (e.g., DNA, HCP) for more than 100 cycles. The engineered Protein A ligand allows for the use of rigorous and cost-effective CIP and sanitization protocols based on NaOH. Furthermore, the ligand is protease stable, which leads to lower ligand leakage, and the highly cross-linked agarose matrix allows for high flow velocities at production scale.

In conclusion, process economy can be significantly improved by the use of Resin 3 in purification of monoclonal antibodies from high-titer cell culture supernatants.

Acknowledgements
The authors wish to thank GE Healthcare Life Sciences (Uppsala, Sweden) for providing Resin 3 prototype resin.

References
1. G. Köhler and C. Milstein, Nature 256 (5517) 495-497 (1975).
2. S. Kozlowski and P. Swann, Adv. Drug Deliv. Rev. 58 (5-6) 707- 722 (2006).
3. P.A. Scolnik, mAbs 1 (2), 179-184 (2009).
4. S. Aggarwal, Nat. Biotechnol. 29 (12) 1083-1089 (2011).
5. B. Kelley, mAbs 1 (5), 443-452 (2009).
6. S. Lofdahl, et al.,  Proc. Natl. Acad. Sci. USA 80 (3) 697-701 (1983).
7. D. Colbert, et al.,  J. Biol. Response Mod. 3 (3) 255-259 (1984).
8. GE  Healthcare, “Dynamic binding capacity study on MabSelect SuRe LX for capturing high-titer monoclonal antibodies,” Application Note, 28-9875-25, Edition AA.