Efficient influenza virus capture from cell cultivation broth by a novel affinity chromatography

Influenza is one of the most worldwide spread diseases. Every year several million people are getting infected. Besides antiviral medical treatments, prophylactic vaccinations are crucial to control seasonal influenza epidemics. Hence, every year large amounts of vaccine doses have to be produced. In the case of pandemic outbreaks these production processes have to be shifted to the appropriate pandemic influenza strain, reducing the production capacity of seasonal vaccines. Conventionally, embryonated chicken eggs are used for human influenza vaccine production. This production process has only a limited scalability. In addition, these vaccines contain egg derived proteins which may cause allergic reactions. Hence, cell culture based vaccine production processes have been developed, which require an adapted downstream processing strategy. Here, the capture step is the most important unit operation concerning overall process economics. Product selectivity and concentration factor of this step define the efforts required for all further purification steps. Introduction of an affinity chromatography step at the beginning of the downstream process allows a high product concentration factor. Furthermore, it separates the target biomolecule from the major bulk of contaminants with a single step. Studies have shown that Euonymus europaeus lectin (EEL) is a suitable ligand for an affinity capture step to purify human influenza A/Puerto Rico/8/34 virus from Madin Darby canine kidney (MDCK) cell culture supernatant. The lectin EEL was immobilized on polymer beads and binds to terminal alpha-galactose of glycosylated viral envelope proteins such as the hemagglutinin (HA). The dissociation of the virus-ligand complex was done by competitive elution with lactose. Thereby more than 90 % of the influenza virus HA-activity was recovered as product fraction while the majority of host cell proteins and nucleic acids were separated [1]. For a chromatographic step, matrix selection plays an important role regarding purification efficiency. Therefore, different matrices for EEL as ligand were screened in a second part of the study. These supports included stabilized reinforced cellulose membranes, polymer and porous glass particles, cellulose and agarose beads. During experimental investigation of virus binding and product recovery studies some of the tested materials indicated advantages for virus purification when compared to cellufine sulfate and heparin. These are two alternative affinity matrices currently used for influenza virus purification. Strong virus binding was achieved by the cellulose membranes and the polymer based adsorbent. Furthermore, reinforced cellulose membranes had a far better binding capacity than other tested adsorbents. To determine the general applicability of EEL-affinity chromatography, studies were extended for other virus strains. Therefore, two of the topical viruses from the flu seasons 2006/07 and 2007/08 (A/Wisconsin/67/2005 (subtype H3N2) and B/Malaysia/2506/2004; both produced in 33016 MDCK PF cell culture, Novartis Behring, Marburg, Germany) were screened for lectin binding. Here, EEL showed no virus strain dependency and bound efficiently to the investigated influenza viruses. Further purification of the virus samples demonstrated that EEL-affinity chromatography is a valuable alternative to capture MDCK cell derived influenza viruses. Furthermore, the impact of host cells on lectin-affinity chromatography has been evaluated. Therefore, influenza A/Puerto Rico/8/34 virus was propagated in MDCK and Vero cells and screened for ligand binding. In contrast to previous findings with MDCK cells, Vero cell derived influenza virus A/Puerto Rico/8/34 bound to beta-galactose specific Erythrina christagalli lectin (ECL) and only very limited to the alpha-galactose specific lectin EEL. Lectin binding studies are planned to be extended to a recombinant subtype of H5N1. In addition, kinetic measurements are being performed to characterize the virus ligand binding in more detail. [1] Opitz, L. et al. 2007. Vaccine 25: 939-947