Bioseparation: The Power of Affinity Purification

By John Curling

Separation and purification sciences are intrinsically linked with the development of high-purity therapeutic proteins. Following the invention of partition chromatography by Archer J.P. Martin, PhD and Richard L.M. Synge, PhD — which led to the Nobel Prize in chemistry in 1952 — chromatography has been the dominant separation technology over at least four decades and now accounts for the ubiquitous critical steps in the purification train to the final bulk active protein.

In a 2004 review, Lightfoot and Moscariello¹ succinctly summarize: “Bioseparations are typically difficult. All products of interest are labile and thus require mild processing conditions. They typically arrive at downstream processing contaminated with closely related species, and the required product purity is usually very high.” For example, in monoclonal antibody products, host cell and media proteins, as well as impurities from processing such as leachates from solid supports, need to be at the parts-per-million (ppm) level. Nucleic acids derived from the cellular hosts need to be at 10 to 100 pg/dose, and endotoxins (fever-causing pyrogens) must be less than five units.

Separations are based on the interaction of the target protein with its environment, falling into three categories: differential solubility, in which the protein is generally caused to precipitate; differential interaction, with physical fields encompassing centrifugation, thermal denaturation, etc.; and differential interaction, with solid media. Chromatography and membrane separations belong to this last category and are the techniques that truly enable purification.

Membrane-based (and/or centrifugal) separations are typically used at the beginning of processes to remove particulate matter and to clarify feed streams for the purification process. In the process train, ultrafiltration/diafiltration is the workhorse of the industry to effect product concentration and to change the ionic environment between purification steps. Nanofiltration is frequently used to reduce the potential load of adventitious or contaminating viruses in purified products and final, sterile filtration is a mandatory step in lieu of harsh physical or chemical sterilizing techniques. However, these separation technologies do not contribute to the biochemical purity of the product but enable chromatographic processing. Anionic membrane filters are used to reduce the level of DNA derived from host mammalian cell lines and may also be effective in reducing the lipopolysaccharide pyrogen contaminants from E. coli or other bacterial hosts. Derivatized membranes may prove to be the matrices of choice for the removal of biocontaminants, such as viruses and bacteria, and for the capture of proteins present in starting materials at very low concentration. Nonetheless, these techniques all leave the main purification issues to be resolved by chromatography.

Downstream processing of clarified biomass is generally divided into three distinct operations: capture of the target biopharmaceutical protein, intermediate purification/separation or fractionation and “polishing” to final purity. The capture step is adsorptive and serves to concentrate the protein from the culture supernatant or fermentation broth, primarily removing water and frequently providing some degree of purification. The fractionation steps that follow are the main purification stages in the process. The final polishing steps remove remaining impurities and bring the product to the target purity, in addition to assuring a consistent impurity profile.

Chromatographic techniques are either group-specific in nature — exploiting general properties of ionic charge, size or hydrophobicity — or are specific to the biochemical and biological properties of the target. Of all the chromatographic techniques in the bioseparations tool kit, affinity chromatography is the most powerful since it is directed to the uniqueness of the protein target. At its introduction in the late ’60s, affinity chromatography was used to purify classes of proteins dependent on their properties or function — such as antibody binding, hormone binding, enzyme inhibition, etc. Over the last decade, development of affinity chromatography has enabled monoclonal antibody purification through the unique affinity of Protein A (a protein originally from Staphylococcus aureus), with the constant region of immunoglobulins. However, it is seldom that nature is so serendipitous. In more recent years, computational chemistry, molecular modelling and combinatorial chemistry have enabled an unprecedented renaissance of affinity chromatography and provided opportunity for chromatographic adsorbent development that enables purification by design. In this mode, a specific adsorbent is constructed to the target biopharmaceutical moiety, often in a customized program between the biopharmaceutical company and the adsorbent vendor.

The most successful approach to the design of affinity chromatography resins employs high throughput screening techniques of libraries constructed in 96-well plates, containing “microcolumns” of defined synthetic affinity adsorbents. The column libraries are constructed using high throughput, parallel synthesis robotics. Using sensitive protein assays, candidate adsorbents that selectively recognize the target protein can be identified for further optimization and development. ProMetic BioSciences Ltd. (Cambridge, U.K.), a subsidiary of ProMetic BioSciences Inc. (Montreal, QC), uses robust chemical scaffolds based on triazine derivatized with known, non-toxic amines. These “ligands” are immobilized onto a neutral, beaded agarose support that can be used in a standard packed bed chromatographic mode. This state-of-the-art chemistry has been used to develop high-performance affinity adsorbents for a number of biopharmaceutical targets, including antibodies, fusion proteins and proteins in blood plasma.

In a unique co-operation between the American Red Cross (Washington, DC) and ProMetic BioSciences, the Cascade process has been designed to capture the most therapeutically relevant proteins from plasma recovered from blood donations, or source plasma collected by apheresis. These proteins include factor VIII/von Willebrand factor used in the treatment of hemophilia; immunoglobulin G indicated in immune deficiency; and alpha-1 proteinase inhibitor, used in North America for the treatment of hereditary emphysema. The Cascade also provides other potentially valuable proteins, such as fibrinogen, for tissue sealants and wound-healing applications, and plasminogen, which has potential as a thrombolytic agent. These proteins are efficiently recovered at exceptionally high yields in a trunk process that facilitates downstream processing.

Yield and purity are critical to the process designer. However, the cost of downstream processing is increasingly at the centre of biopharmaceutical development debate. In part, this is due to the escalating costs of Protein A adsorbents — now significantly more than $10,000 US per litre — and the industry development to larger volume manufacturing, causing the unit cost of 100-litre antibody purification columns to exceed $1 million US. Downstream purification costs are cited as being between one-third and 80 per cent of the total production cost of biopharmaceuticals, depending on how calculations are made, with chromatography being notoriously expensive.

With bioreactor titres rising beyond the gram-product-per-litre level, downstream processing, and thus chromatography, are becoming rate-limiting because of cost and throughput. It is not surprising that yield and cost become focal points. Figure 1 shows a typical recovery sequence after clarification of the feed stream.

In this simplistic model, there is a 15 per cent yield loss in the first capture step and then 10 per cent losses in each of the intermediate purification steps, followed by five per cent losses at each of the following stages. If intermediate purification is reduced by one step, the resulting yield gain is 10 per cent units of final product, and an additional 10 per cent revenue from the same fermentation, cell culture or blood plasma pool.

It is here that affinity chromatography shows its true power. The challenge of bioseparations lies in distinguishing and extracting the target protein from its environment. For example, about 300 proteins in blood plasma have been identified amid the many thousands present at low concentration. In antibody purification from cell culture, the target must be separated from other proteins produced by the cell machinery. Finding a single protein will rely on anomalous protein behaviour unless designed affinity “fishing” is used. Affinity chromatography combines selectivity with high yield. Synthetic adsorbents using Mimetic Ligand™ technology select the target protein but are generally not as exquisitely specific as antibodies to the target used in immunoaffinity techniques. However, protein ligand adsorbents introduce an unwanted biological derivative into the manufacturing process and are to be avoided. Additionally, there are significant cost differences in favour of synthetic ligand chromatography compared to all types of affinity separations that use biological ligands — either existing in nature or discovered by phage display technology and engineered into cellular expression systems. Re-engineered Protein A is more expensive than its predecessors and is not a long-term solution for monoclonal antibody purification.

Chromatography will remain the technology of choice for downstream processing, despite the many efforts to find new solutions that avoid the use of packed beds. Affinity chromatography using custom-designed synthetic ligands is likely to play an ever-increasing role in the bioseparations arsenal, while efforts continue through European and other partnerships to find alternatives to the Protein A ligand. New formats for chromatography may emerge in the next decade and manufacturers will be challenged with balancing the introduction of cost saving, higher yielding technology with the expense of demonstration of comparability required by the regulatory agencies.

Reference
(1) Lightfoot, E.N. and J.S. Moscariello. Bioseparations. Biotechnol. Bioeng. 87 (3): 259-273.

John Curling is a senior scientist and consultant with ProMetic BioSciences Ltd. (Cambridge, U.K.)