Integrated process optimization: lessons from retrovirus and virus-like particle production.

The optimization of production and purification processes is usually approached by engineers from a strictly biotechnological point of view. The present paper envisages the definition and application of an optimization model that takes into account the impact of both biological and technological issues upon the optimization protocols and strategies. For this purpose, the optimization of three analogous but different systems comprising animal cell growth and bioparticle production is presented. These systems were: human immunodeficiency 1 (HIV-1) and porcine parvovirus (PPV) virus-like particles (VLPs) produced in insect cells and retrovirus produced in mammalian cells. For the systematization of the optimization process four levels of optimization were defined-product, technology, design and integration. In this paper, the limits of each of the optimization levels defined are discussed by applying the concept to the systems described. This analysis leads to decisions regarding the production of VLPs and retrovirus as well as on the points relevant for further process development. Finally, the definition of the objective function or performance index, the possible strategies and tools for bioprocess optimization are described. Although developed from the three described processes, this approach can, based on the recent literature evidence reviewed here, be applied more universally for the process development of complex biopharmaceuticals.

Large scale production and downstream processing of a recombinant porcine parvovirus vaccine.

Porcine parvovirus (PPV) virus-like particles (VLPs) constitute a potential vaccine for prevention of parvovirus-induced reproductive failure in gilts. Here we report the development of a large scale (25 l) production process for PPV-VLPs with baculovirus-infected insect cells. A low multiplicity of infection (MOI) strategy was efficiently applied avoiding the use of an extra baculovirus expansion step. The optimal harvest time was defined at 120 h post-infection at the MOI used, with the cell concentration at infection being 1.5x10(6) cells/ml. An efficient purification scheme using centrifugation, precipitation and ultrafiltration/diafiltration as stepwise unit operations was developed. The global yield of the downstream process was 68%. Baculovirus inactivation with Triton X-100 was successfully integrated into the purification scheme without an increase in the number of process stages. Immunogenicity of the PPV-VLPs tested in guinea pigs was similar to highly purified reference material produced from cells cultured in the presence of serum-containing medium. These results indicate the feasibility of industrial scale production of PPV-VLPs in the baculovirus system, safety of the product, and the potency of the product for vaccine application.

Expression and characterization of recombinant human alpha-3/4-fucosyltransferase III from Spodoptera frugiperda (Sf9) and Trichoplusia ni (Tn) cells using the baculovirus expression system.

The human alpha-3/4-fucosyltransferase III (Fuc-TIII) participates in the synthesis of Lewis determinants. The enzyme from human sources is scarce and heterogeneous. In this paper we describe the expression of a secreted form of Fuc-TIII (SFT3) in two insect cell lines, Spodoptera frugiperda (Sf9) and Trichoplusia ni (Tn), using the baculovirus expression system. The Sf9 cells secreted approx. 0.4 unit/l (1 mg/l) of the enzyme. The Tn cells secreted approx. 3-fold this amount. A large proportion of active protein was accumulated in the two cell lines (50 and 75% respectively for Sf9 and Tn cells, on the fourth day after infection) indicating a possible limitation not only of the folding machinery, but also a saturation of the secretory pathway. SFT3 was purified by cation-exchange chromatography followed by affinity chromatography. The enzyme from the Tn cell line had a lower global charge, possibly due to post-translational modifications, such as phosphorylation or sulphation. The two glycosylation sites from SFT3 were occupied. SFT3 secreted by Sf9 cells was completely deglycosylated by peptide-N-glycanase F, whereas 50% of SFT3 secreted by Tn cells was resistant to deglycosylation by this enzyme. The apparent kinetic parameters determined with the type I acceptor were k(cat)=0.4 s(-1) and K(m)=0.87 mM for the SFT3 secreted by Tn cells, and k(cat)=0.09 s(-1) and K(m)=0.76 mM for the SFT3 secreted by Sf9 cells, indicating that the enzymes had substrate affinities within the same order of magnitude as their mammalian counterpart. Furthermore, SFT3 secreted by either cell type showed a clear preference for type 1 carbohydrate acceptors, similarly to human Fuc-TIII.