With an increasing demand for efficacious, safe, and affordable vaccines for human and animal use, process intensification in cell culture-based viral vaccine production demands advanced process strategies to overcome the limitations of conventional batch cultivations. However, the use of fed-batch, perfusion, or continuous modes to drive processes at high cell density (HCD) and overextended operating times has so far been little explored in large-scale viral vaccine manufacturing. Also, possible reductions in cell-specific virus yields for HCD cultivations have been reported frequently. Taking into account that vaccine production is one of the most heavily regulated industries in the pharmaceutical sector with tough margins to meet, it is understandable that process intensification is being considered by both academia and industry as a next step toward more efficient viral vaccine production processes only recently. Compared to conventional batch processes, fed-batch and perfusion strategies could result in ten to a hundred times higher product yields. Both cultivation strategies can be implemented to achieve cell concentrations exceeding 107 cells/mL or even 108 cells/mL, while keeping low levels of metabolites that potentially inhibit cell growth and virus replication. The trend towards HCD processes is supported by development of GMP-compliant cultivation platforms, i.e., acoustic settlers, hollow fiber bioreactors, and hollow fiber-based perfusion systems including tangential flow filtration (TFF) or alternating tangential flow (ATF) technologies. In this review, these process modes are discussed in detail and compared with conventional batch processes based on productivity indicators such as space-time yield, cell concentration, and product titers. In addition, options for the production of viral vaccines in continuous multi-stage bioreactors such as two- and three-stage systems are addressed. While such systems have shown similar virus titers compared to batch cultivations, keeping high yields for extended production times is still a challenge. Overall, we demonstrate that process intensification of cell culture-based viral vaccine production can be realized by the consequent application of fed-batch, perfusion, and continuous systems with a significant increase in productivity. The potential for even further improvements is high, considering recent developments in establishment of new (designer) cell lines, better characterization of host cell metabolism, advances in media design, and the use of mathematical models as a tool for process optimization and control.

Increasing the yield and the productivity in cell culture-based vaccine manufacturing using high-cell-density (HCD) cultivations faces a number of challenges. For example, medium consumption should be low to obtain a very high concentration of viable host cells in an economical way but must be balanced against the requirement that accumulation of toxic metabolites and limitation of nutrients have to be avoided. HCD cultivations should also be optimized to avoid unwanted induction of apoptosis or autophagy during the early phase of virus infection. To realize the full potential of HCD cultivations, a rational analysis of the cultivation conditions of the appropriate host cell line together with the optimal infection conditions for the chosen viral vaccine strain needs to be performed for each particular manufacturing process. We here illustrate our strategy for production of the modified vaccinia Ankara (MVA) virus isolate MVA-CR19 in the avian suspension cell line AGE1.CR.pIX at HCD. As a first step we demonstrate that the adjustment of the perfusion rate strictly based on the measured cell concentration and the glucose consumption rate of cells enables optimal growth in a 0.8 L bioreactor equipped with an ATF2 system. Concentrations up to 57 × 10 cells/mL (before infection) were obtained with a viability exceeding 95%, and a maximum specific cell growth rate of 0.019 h (doubling time = 36.5 h). However, not only the cell-specific MVA-CR19 virus yield but also the volumetric productivity was reduced compared to infections at conventional-cell-density (CCD). To facilitate optimization of the virus propagation phase at HCD, a larger set of feeding strategies was analyzed in small-scale cultivations using shake flasks. Densities up to 63 × 10 cells/mL were obtained at the end of the cell growth phase applying a discontinuous perfusion mode (semi-perfusion) with the same cell-specific perfusion rate as in the bioreactor (0.060 nL/(cell d)). At this cell concentration, a medium exchange at time of infection was required to obtain expected virus yields during the first 24 h after infection. Applying an additional fed-batch feeding strategy during the whole virus replication phase resulted in a faster virus titer increase during the first 36 h after infection. In contrast, a semi-continuous virus harvest scheme improved virus accumulation and recovery at a rather later stage of infection. Overall, a combination of both fed-batch and medium exchange strategies resulted in similar cell-specific virus yields as those obtained for CCD processes but 10-fold higher MVA-CR19 titers, and four times higher volumetric productivity.

A cultivation strategy to increase the productivity of Modified Vaccinia Ankara (MVA) virus in high-cell density processes is presented. Based on an approach developed in shake flask cultures, this strategy was established in benchtop bioreactors, comprising the growth of suspension AGE1.CR.pIX cells to high cell densities in a chemically defined medium before infection with the MVA-CR19 virus strain. First, a perfusion regime was established to optimize the cell growth phase. Second, a fed-batch regime was chosen for the initial infection phase to facilitate virus uptake and cell-to-cell spreading. Afterwards, a switch to perfusion enabled the continuous supply of nutrients for the late stages of virus propagation. With maximum infectious titers of 1.0 × 10 10 IU/mL, this hybrid fed-batch/perfusion strategy increased product titers by almost one order of magnitude compared to conventional batch cultivations. Finally, this strategy was also applied to the production of influenza A/PR/8/34 (H1N1) virus considered for manufacturing of inactivated vaccines. Using the same culture system, a total number of 3.8 × 10 10 virions/mL was achieved. Overall, comparable or even higher cell-specific virus yields and volumetric productivities were obtained using the same cultivation systems as for the conventional batch cultivations. In addition, most viral particles were found in the culture supernatant, which can simplify further downstream operations, in particular for MVA viruses. Considering the current availability of well-described perfusion/cell retention technologies, the present strategy may contribute to the development of new approaches for viral vaccine production. Electronic supplementary material The online version of this article (10.1007/s00253-019-09694-2) contains supplementary material, which is available to authorized users.

No abstract

El método de co-simulación estocástica espacial, basado en cópulas, es un método general que permite simular variables con cualquier tipo de dependencia y funciones de distribución de probabilidad. Esta flexibilidad proviene del uso de un modelo de cópula para la representación de la función de distribución de probabilidad conjunta. El método se ha implementado principalmente a través de un enfoque no paramétrico utilizando cópulas de Bernstein y se ha aplicado con éxito para la simulación de propiedades petrofísicas usando atributos sísmicos elásticos como variables secundarias. En el presente trabajo este método se implementa mediante otros dos enfoques: paramétrico y semi-paramétrico. Específicamente, para el enfoque paramétrico se usa la familia de cópulas Arquimedianas. Primero, el enfoque paramétrico se valida con un caso publicado y luego se realiza una comparación de los tres enfoques en términos de precisión y rendimiento. Los resultados mostraron que el enfoque paramétrico es el que peor reproduce las estadísticas de los datos y presenta mayor incertidumbre con un menor costo computacional, mientras que el enfoque no-paramétrico resultó el que mejor reproduce la dependencia de los datos a un alto costo computacional. El enfoque semi-paramétrico reduce un 10% el costo computacional respecto al no-paramétrico, pero se degrada significativamente su precisión.