February 15, 2007 (Vol. 27, No. 4)

Sustaining Cost-efficient, Antibody-based Biopharmaceutical Production

Affinity resins based on recombinant protein A ligand are the industry standard for the primary captures in large-scale antibody purification processes. Inherent sensitivity of protein-based affinity ligands, such as protein A, to the most commonly employed cleaning in place (CIP) reagent 1 M sodium hydroxide (NaOH) has led to the development of alternative CIP regiments, such as 6 M guanidine hydrochloride or urea. These methods, however, add significantly to the overall cost of goods (COG) and are often not as efficient as 1 M NaOH.

Deamidation of the side–chains of asparagine residues at ambient alkalinity has been recognized as the major contributor to the alkaline sensitivity of protein A. Replacing asparagine residues with other amino acids can stabilize the protein and enable it to withstand harsh alkaline solutions without loss of activity. This further demonstrates the potential of recombinant methods to enhance industrial use of affinity chromatography based on protein ligands.

MabSelect SuRe™ from GE Healthcare (www.gehealthcare.com) is a base-tolerant affinity resin designed with a genetically engineered recombinant protein A derived ligand as its functional component.

Figure 1

Dynamic Binding Capacity

The dynamic binding capacity (DBC) at 10% breakthrough for polyclonal human immunoglobulin G (hIgG) has been determined in duplicate for both MabSelect SuRe and a conventional protein A resin. A series of tests simulating pH alterations encountered during Mab purification were undertaken to demonstrate the resilient nature of MabSelect SuRe resin toward an alkaline environment. CIP protocols varying in the strength of the CIP reagent and its contact time with the MabSelect SuRe resin were incorporated into the test protocols. Results showed that MabSelect SuRe resin is tolerant toward moderate concentrations of sodium hydroxide solutions for prolonged periods of time (Figure 1).

A more industrially relevant study was then undertaken. Humanized IgG1 and IgG4 were purified hundreds of times with each purification cycle followed up by CIP, comprising 15-minutes contact time with 0.1 M NaOH solution. The residence times for the Mabs on the MabSelect SuRe resin and the column dimensions were varied to generate data over a range of conditions.

The volumes of the mammalian cell culture-clarified supernatants applied to the MabSelect SuRe resin were monitored so that 80% of the binding capacity was utilized. The DBC at 10% breakthrough of MabSelect SuRe resin for humanized IgG1 and IgG4 were determined in duplicate at regular intervals throughout the purification campaigns (Figure 2).

A visual examination of the MabSelect SuRe resin from the top end and the bottom end of the packed column used for 150 purification cycles of humanized IgG1 in conjunction with 150 CIP cyles showed no signs of bead aggregates or deposits. A closer examination of these samples at 500X magnification under a scanning electron microscope revealed no apparent fouling or damage to the individual beads. These results support previous chromatographic data as to the inherent ability of MabSelect SuRe to retain performance despite prolonged exposure to NaOH and provide a consistent and low COG and method to purify antibody-based pharmaceutical products.

Figure 2

CIP Protocol

Traditionally, during a Mab-purification campaign a CIP cycle is recommended following every purification cycle. The CIP protocols provide a thorough removal of host cell protein (HCP) contaminants adsorbed to the resin. CIP also disengages residual aggregates so as to leave a thoroughly regenerated resin capable of binding the next batch of antibody in a purification campaign. Frequent CIP following every purification cycle, however, is not only time consuming but also cost inefficient.

MabSelect SuRe resin enables reduction of the total numbers of CIP cycles by spacing their intervals up to as far as every tenth cycle while still maintaining low production costs due to the potential to use higher NaOH solution concentrations. The DBC of the MabSelect SuRe resin for humanized IgG1 is maintained at more than 99% of its initial value following 21 purification cycles in conjunction with the optimized CIP protocol.

This optimized CIP protocol, performed following every tenth purification cycle on MabSelect SuRe resin, will reduce the overall COG due to the realization of increased number of purification cycles (Figure 3).

In this 21-cycle study all the recovered fractions of pure humanized IgG1 were 99% monomer as analyzed by size exclusion chromatography. This is in keeping with a yield of greater than 90% pure IgG1 during all 150 cycles of the earlier study. Additionally, the yield of IgG4 was greater than 93% over 100 purification cycles.

Measurement of HCP using ELISA or other methods is often used to indicate the consistency of bioprocesses. Regulatory authorities do not set global limits for the levels of residual HCPs present in final product. Dose and route of administration of the final product are among several factors that determine the acceptable levels of HCPs. However, the HCP clearance profile for MabSelect SuRe resin shows that during the 21 purification cycles, the levels of HCP per mg soluble monomer humanized IgG1 were were consistently at 1,600 ppm.

HCP in the pure IgG4 eluates from 100 purification cycles on MabSelect SuRe were screened for presence of HCP contaminants on a Western blot by probing the membrane with polyclonal antibodies raised against the HCP and against the immunoglobulin (Figure 4).

The absence of protein bands in lanes 18 and 19 indicate neither carryover nor cross-contamination in low pH-mediated eluates from cycles lacking feedstock applications performed following the 50th and 100th purification cycles respectively. The presence of bovine serum albumin and hIgG in lane 2, as the negative and positive controls, respectively, marks the specificity of the two sets of antibodies. Lane 1 shows the molecular marker proteins.

The trace amounts of ligand leakage from a protein-based affinity resin need to be removed by subsequent purification steps like multimodal chromatography, ion-exchange, or hydrophobic interaction chromatography. Protein A levels from MabSelect SuRe resin are low because of high protease stability of the genetically engineered Protein A ligand. Nonetheless, it is essential to monitor the levels of ligand in the desorbed fractions in order to stay within the acceptable limits.

Figure 3

Figure 4

Aman Mottaqui–Tabar is a research engineer for biomolecular characterization at GE Healthcare Biosciences.
Web: www.gehealthcare.com.
E-mail: [email protected].

Previous articleInvitrogen Sells Contract Services Business for $210M
Next articleRBM Awarded $750,000 from DoD