Improving Protein Purification
“Should we just make things bigger or go for alternatives?” asked Duncan Low, Ph.D., scientific director, process development, at Amgen (www.amgen.com). The question is whether existing technology can actually deal with more than 100 kg lots, noting that there has been a shift from concern over manufacturing capacity to concern over costs. Purification is the most significant factor in cost of goods, with protein A and viral inactivation being the most expensive steps, according to data from Brian Kelley, Ph.D., director of purification process development at Genetics Institute-Wyeth.
Dr. Low described how he sees the future of mAb purification. Cell culture yields are likely to be more than 10 g per liter in the near future, thanks to greater cell density, which also leads to more host cell protein and more debris. He thinks that protein A may be approaching its performance ceiling of 50 g per liter and a flow rate of 750 cm per hour flow rate.
Alternatives to protein A include bacterial IgG binding protein, lectins, immunoaffinity columns, single domain antibodies (highly selective and the closest to protein A), and synthetic ligands. Tests, however, show that selectivity decreases as complexity of the ligand decreases, but in economic terms, the less complex ligands are cheaper. Engineering solutions to the cost problem might include simulated moving beds instead of columns, which leads to significant savings in resin costs.
Alternatives to chromatography include precipitation, crystallization, magnetic separation, and aqueous two-phase systems. Amgen carried out a comparison of Protein A against precipitation that suggested that the latter gave equivalent or better performance with a clean product. Membranes too can play a role in viral inactivation and in many other processes.
“In the next five years, there will be incremental improvements such as high-throughput resins and increases in selectivity, and we will be asking questions about chromatography and whether there is room for nonchromatography modalities,” Dr. Low concluded
The need for higher capacity and greater selectivity during protein purification could be addressed by rational affinity ligand design, which is being developed by Chris Lowe, Ph.D., director of the Institute of Biotechnology at the University of Cambridge. The institute uses molecular modeling to find ligands that may bind to at least part of the target protein binding site; these are then synthesized and immobilized on a column. Those that do bind target protein can then be optimized, with those combining high-binding capacity with high-elution capacity being the most promising.
mAbs are clearly major targets for this work, and experiments on a synthetic ligand version of both protein A and protein L have been carried out.
Meanwhile, Dr. Lowe has also used the synthetic ligand approach, in application to NovoNordisk’s (www.novonordisk.com) Factor VIIa, to replace the antibody in an immunochromatography purification step with a small molecule alternative. “This antibody saturates its capacity for target protein rapidly,” Dr Lowe said. “The synthetic ligands proved to have a higher capacity, with the final choice molecule having 35 times more binding capacity than the antibody.”
More recently, Dr. Lowe’s team has been targeting glycoforms, looking for synthetic ligands that will bind to proteins with specific glycosylation patterns. This would enable a company to pull out a particular glycoform from a product, something which is likely to gain the approval of the regulatory authorities. Finally, they are also developing ligands for prion proteins that could help in removal of this form of contamination and are interested in 3-D ligands, which can bind in an even more specific way to their target.