(373b) Mathematical Model for Measles Virus Production in Batch Bioreactors

Measles virus (MeV) propagation has been studied extensively for production of recombinant MeV vectors for viral vaccines and gene therapy applications as well as for oncolytic virus therapy. Recombinant MeV vectors have demonstrated high safety and efficacy as well as long-lasting immunity toward antigens [1]. In addition, MeV is a suitable candidate for an oncolytic virus because tumor cells overexpress the MeV receptor CD46 [2].

MeV is commonly produced in bioreactors with microcarrier culture using Vero cells as a production cell line [3]. This process is yet to provide sufficient concentrations of MeV required for therapeutic applications primarily due to the stability of the virus. MeV, an enveloped virus, is sensitive to shear stress, low pH, and UV light [3], [4]. However, the main challenge of MeV production is its high sensitivity to temperature. At the growth temperature of Vero cells (37°C), MeV has a half-life of one hour [3]. This temperature-dependent virus inactivation reduces the maximum virus yield in two ways: initial infection by inoculation and the overall production yield [3]. Therefore, it is imperative to know the optimal timing of the infection (TOI) process and the time of virus release to predict the optimal time of harvest (TOH). Our objective is to develop a modeling approach to predict the kinetics of MeV infection in Vero cells for the purpose of virus production.

Our model is adapted from an unstructured dynamic mass balance model by Möhler et al. describing the populations of uninfected cells, infected cells, and free virus over the course of infection in a vaccine production bioreactor producing influenza virus [5]. Our model constitutes seven dynamic mass balance equations which, in addition to uninfected and infected cells, includes balances for dead cells due to infection and infected cells that detach from the microcarriers into the supernatant. Our model also captures both infectious virus particles and non-infectious virus particles. Our model includes syncytia formation, a common cytopathic effect in MeV-induced infections, and reduction in infectious virus particles due to defective interfering particles [3], [6].

To determine an optimal set of model parameters from a limited data set, we used a global optimization algorithm. Data available for use in parameter estimation included concentrations of total cells, dead cells, total virus, and infectious virus from five similar batch bioreactor runs. A bootstrapping technique was utilized to determine confidence intervals for our parameter estimates. Model performance and validation were evaluated using datasets not included in the parameter estimation. We have further performed a global sensitivity analysis to determine which subset of parameters have the most impact on process performance and to suggest how culture operational parameters, TOI and TOH, may be manipulated to optimize productivity.

References:

[1] M. Lu et al., “A safe and highly efficacious measles virus-based vaccine expressing SARS-CoV-2 stabilized prefusion spike”, PNAS, vol. 118, no. 12, Feb. 2021, doi: 10.1073/pnas.2026153118.

[2] S. Aref, K. Bailey, and A. Fielding, “Measles to the Rescue: A Review of Oncolytic Measles Virus,” Viruses, vol. 8, no. 10, Oct. 2016, doi: 10.3390/V8100294.

[3] T. A. Grein, D. Loewe, H. Dieken, D. Salzig, T. Weidner, and P. Czermak, “High titer oncolytic measles virus production process by integration of dielectric spectroscopy as online monitoring system,” Biotechnol Bioeng, vol. 115, no. 5, pp. 1186–1194, May 2018, doi: 10.1002/BIT.26538.

[4] S. Kiesslich and A. A. Kamen, “Vero cell upstream bioprocess development for the production of viral vectors and vaccines,” Biotechnol Adv, vol. 44, Nov. 2020, doi: 10.1016/J.BIOTECHADV.2020.107608.

[5] L. Möhler, D. Flockerzi, H. Sann, and U. Reichl, “Mathematical model of influenza A virus production in large-scale microcarrier culture,” Biotechnol Bioeng, vol. 90, no. 1, pp. 46–58, Apr. 2005, doi: 10.1002/bit.20363.

[6] T. Whistler, W. J. Bellini, and P. A. Rota, “Generation of Defective Interfering Particles by Two Vaccine Strains of Measles Virus,” Virology, vol. 220, no. 2, pp. 480–484, Jun. 1996, doi: 10.1006/VIRO.1996.0335.