Liverpool is best known as the scruffy British industrial city that gave the Beatles their start back in the 60s, but it is recognized today for reinventing its image as a home to an expanding biotech presence. SAFC and Sartorius Stedim Biotech’s recent “Downstream Processing Day” seminar was held in the city and brought together investigators from a number of companies, many with facilities at the Liverpool Science Park.
Purification issues are critical to the development of high-quality vaccine proteins. François Champluvier, Ph.D., a research scientist in downstream industrialization at GlaxoSmithKline, presented his investigations on removal of host-cell proteins during vaccine purification.
The standard purification protocol is based on anion-exchange chromatography followed by cation-exchange chromatography and a final step using hydroxyapatite. “Our challenge was to achieve a lower final host-cell protein content, but with minimum modification to the existing process and without impacting the process overall yield,” Dr. Champluvier explained.
Since the purification process uses Sartorius Stedim Biotech membranes, it was necessary to work closely with the company in order to arrive at an optimal solution. The Sartobind membranes are composed of chemically modified cellulose—Q being a strong anion, S a strong cation, and HIC a phenyl ligand.
The approach that Dr. Champluvier and his colleagues adopted was to perform prospective evaluations at a small scale with different filters to identify the optimal conditions, then include these conditions in the existing process. The small-scale trials revealed that the Sartobind S filters provided the best combination of recovery of protein while maximizing host-cell protein elimination.
“We found that in the case of the higher host-cell protein input, the Sartobind S shows a very good efficacy with very good antigen yields, whereas at the lower end of the host-cell protein amounts the benefits were not visible. So with minimal changes, we were able to achieve substantially improved purity with minimal impact on the subsequent steps.”
Estimates of viral contamination during bioreactor runs at production scale range from 1 in 360 to 1 in 500, making such events quite unlikely. But because of their high impact, their consequences are extremely grave, as discussed by Clare Medlow, technical manager at SAFC. Moreover, few after-the-fact troubleshooting investigations are successful, given that it is difficult to identify the root cause and source of the virus. Thus, one strategy is to install broad preventive measures, removing the components that present the highest risk.
There are a number of mechanisms by which a production train can become contaminated including the cell line itself, the compressed gases that buffer the medium, the water, and the components added to the medium. Of these possibilities, the feeds and additions to the medium may be the most likely, given the many sources involved and the fact that many are of animal origin.
In many instances, multiple layers of security must be applied including irradiation of solutions, pasteurization, removal of animal products from the formulations, and screening for viruses. All of these can mitigate against contamination at various point in the process.
Viral inactivation of solutions through a pasteurization process can be used on material that does not contain serum. Precise control of temperature and residence times in the heating vessels are required. Medlow described two types of high-temperature, short-time liquid treatment options.
The first is the shell-and-tube heat-exchange system from Cotter Brothers/GEA Processing Engineering, consisting of several tubes encased in a larger tube. The other device is the Actini Electric Actijoule®, in which application of a low-voltage current to a steel tube heats the liquid inside the tube. According to Medlow, 4–6 logs of virus reduction can be achieved under specific pilot conditions and then evaluated at expanded scale.
“It is essential that such protocols be critically validated, otherwise the process is meaningless.”
Steven J. Burton, Ph.D., CEO of ProMetic Biosciences, considered the use of mimetic ligands in the purification of biotherapeutic proteins. These synthetic ligands, obtained through the use of computational chemistry, are tough and can be reused for many purification cycles. Because they are made through conventional chemistry rather than by biological systems, there is no risk of contamination, and they are economical to manufacture compared to ligands from traditional biological sources such as protein A.
Mimetic ligands can be integrated into a purification platform with an affinity capture step that will achieve a high-capacity recovery of the product with excellent removal of host-cell proteins, according to Dr. Burton. “Nonantibody proteins represent 65% of the biotherapeutics market. This segment, including growth factors, hormones, cytokines, therapeutic enzymes, and protein vaccines, is large and diverse, so effective purification technologies are very much in demand.”
Active ligands can be generated through modeling and optimizing existing ligands, referred to as analogue synthesis. In this case the active site is known, simplifying the modeling of the complementary ligand structures—a process referred to as rational design. However, in some cases, neither the active site nor ligand is known, necessitating a systematic screening using ligand arrays.
ProMetic scientists have developed an approach to improve their synthetic ligands. This method is referred to as molecular docking, in which they prioritize ligands by a virtual screening strategy. By developing algorithms that place a ligand into a binding site and then score the resulting pose, the candidate molecules can be ranked. Those that would be disallowed for an immobilized ligand can be removed from further consideration. Such an approach saves time and resources since it allows the scientists designing the ligand to avoid fruitless and unproductive molecules.
As a successful application of the rational design strategy, Dr. Burton described his company’s development of synthetic affinity ligands in which x-ray crystallographic data for transferrin was analyzed using the ConSurf Server program, which identifies conserved and variable residues.
The higher the level of conservation, the greater the likelihood of functional importance and applicability as a target site for affinity ligand design. “We used a program of blind docking to computationally roam our virtual ligands over the entire protein surface and identify minimum energy binding sites.”