(563g) Advanced Process Control of a Continuous Viral Bioreactor

Viral particles which can include whole virus particles and virus-like particles (VLPs) are an indispensable part of modern biotechnology, finding use in many applications such as the production of vaccines (e.g., inactivated or attenuated whole viral vaccines), recombinant protein production, and gene therapies. Increasingly, large-scale manufacturing of viral particles is shifting to continuous cell culture-based processes due to benefits such as improved product quality control and manufacturability. However, viral infection kinetics and the propensity of some viruses to produce defective interfering particles (DIPs) can introduce complex and undesirable process dynamics into the bioreactor which can affect steady-state operation and productivity [1], [2]. DIPs, which interfere with standard virus particle (STV) replication, are seen either as a nuisance/ impurity that adversely impacts manufacturing and product quality or as a desirable product by themselves with potential use as antiviral and antitumor therapeutics [3]. Correspondingly, it would be significant if the same bioreactor configuration can be employed to produce either a high-purity STV or DIP product stream. Currently, strategies used to minimize DIP production such as optimizing cell lines and virus strains to reduce DIP generation or using novel bioreactor configurations are costly and complex. In comparison, advanced process control, which has been unexplored in this setting, can be used to achieve successful operation. This work explores how both classical and modern control strategies such as state feedback and nonlinear economic model predictive control can be used to achieve process objectives such as generating high-purity product streams and maximizing bioreactor productivity. A model predictive control strategy is proposed for viral production in continuous bioreactors and demonstrated to be robust even under plant-model mismatch and insensitive to realistic levels of sensor noise and disturbances. Comparisons are made to more traditional control strategies, in terms of the control actions taken, robustness to model uncertainties, and closed-loop performance.

References

[1] T. Frensing, “Defective interfering viruses and their impact on vaccines and viral vectors,” Biotechnol. J., vol. 10, no. 5, pp. 681–689, May 2015

[2] T. Frensing et al., “Continuous Influenza virus production in cell culture shows a periodic accumulation of defective interfering particles,” PLoS One, vol. 8, no. 9, p. e72288, Sep. 2013.

[3] M. D. Hein et al., “Cell culture-based production and in vivo characterization of purely clonal defective interfering influenza virus particles,” BMC Biol., vol. 19, no. 1, p. 91, Dec. 2021.