Production of high-titer human influenza A virus with adherent and suspension MDCK cells cultured in a single-use hollow fiber bioreactor
Introduction
Cell culture-based manufacturing of influenza viruses offers several advantages compared to the conventional production using fertilized chicken eggs. The easier and faster scalability in case of a pandemic [1], the possibility to run a fully characterized process [2], comparable immunogenicity with egg-derived vaccines [3], [4], [5], and high virus yields [6] are some advantages of cell cultures. These reasons have led to the approval of cell culture-derived vaccines as Optiflu® and Flucelvax® (both manufactured by Novartis Vaccines and Diagnostics, Germany) in Europe and the USA, respectively.
Continuous cell lines as MDCK, Vero, HEK.293, and PER.C6® [1], [7], [8] have been characterized for production of influenza virus in industry and research. These cells can be cultivated to large scales with the use of bioreactors, thus improving virus yields and achieving commercially relevant production capacities [9], [10]. Fed-batch culture, and the use of microcarriers in stirred tank bioreactors (STR) are strategies that can lead to high cell concentration, with values in the order of 1 × 107 cells/mL [11], [12]. Nevertheless, limitations of medium compounds and the accumulation of toxic metabolites at these high densities are known to induce shifts in the cell-specific virus yield [13], [14]. This so called “cell density effect” could be avoided with the use of perfusion systems as filtration, cross flow filters, spin-filters, acoustic filter [12], and alternating tangential flow systems (ATF®) [15], [16], [17], [18]. In this scenario, the use of hollow fiber bioreactors (HFBRs), which is a type of perfusion system, has been only recently proposed for production of influenza virus [19].
HFBRs [20], [21] have been widely used in tissue engineering [22] as well as for production of monoclonal antibodies [23] and recombinant proteins [24]. These platforms are able to support the growth of in vivo-like cell densities for weeks and even months, and cell densities can be even 100-fold higher than common suspension cell cultures [25]. HFBRs consist of a cartridge with up to thousands of capillaries, which separate the cartridge volume in two compartments, known as the intra-capillary space (ICS) and the extra-capillary space (ECS). Typically, cells are grown protected from shear stresses in the ECS, while medium flows through the ICS and can exchange substrates and inhibitors with the ECS through the hollow fiber membranes. These bioreactors have been used to study pharmacodynamics of antiviral and antimicrobial compounds [26] and, in a first try, Hirschel et al. [19] have proposed the use of HFBRs as production platform for influenza A virus at high concentrations. The work of Hirschel et al. showed that, by infecting adherent MDCK (ATCC) cells with the pandemic strain A/Mexico/4108/2009, infectious virus concentrations up to 1 × 109 virions/mL and total virus concentrations between 1 × 1010 and 1 × 1011 virions/mL can be obtained. These results motivated to continue with a more thorough study.
In this work, we characterize a process for production of influenza A virus using single-use HFBRs. For this purpose, an adherent MDCK (ECACC) and a suspension MDCK.SUS2 cell line derived herefrom [27] were cultured in the ECS of the HFBRs. Both cell lines were infected with the A/PuertoRico/8/34 (H1N1), and the A/Mexico/4108/2009 (H1N1) pandemic strains. Total and infectious virus productions were characterized for both cells, and finally, cell-specific virus yield and space-time virus yield of the HFBRs were estimated and compared to the data reported for other production platforms as classical STR, wave bioreactors (Wave®), and ATF® system.
Section snippets
Cell lines and cell cultivation
Adherent MDCK cells (ECACC, #84121903) (hereafter referred as MDCK) were scaled up in vented tissue culture flasks (Cellstar, Greiner bio-one) at 37 °C, 5% CO2 with serum-free medium Episerf (Gibco Invitrogen) supplemented with 2 mM l-glutamine (Sigma), 2 mM pyruvate (Sigma), and 20 mM d-glucose (Roth). Cultivations in flasks were started with approximately 4.0 × 104 cells/cm2.
Suspension MDCK.SUS2 cells [27] (hereafter referred as MDCK.SUS2) were maintained and scaled-up in 125 or 250 mL vented shaker
Number of MDCK and MDCK.SUS2 cells grown in ECS at time of infection
A total of 1.4 × 109 MDCK cells were counted from the run A1 (see Table 1), and therefore the number of MDCK cells grown in HFBRs was assumed to be between 1 × 109 and 2 × 109 (TMCM, see Section 2.6) cells. In contrast, a slightly lower number of 8 × 108 MDCK.SUS2 cells were counted from S1. However, subsequent HFBRs runs showed that the number of MDCK.SUS2 cells growing inside the ECS was larger (up to 1.8 × 109 cells), therefore the cell number used to calculate the MOI was different for each run with
Cell growth and glucose uptake rate
Rodrigues et al. proposed the use of GUR as an indirect method for monitoring cell growth [30]. Nevertheless, a more recent work indicated that strong changes in pH could induce variations in the GUR [31]. This can explain the higher GUR (0.91 mmol/h) observed in ARM2, as the fresh medium added to the system just before time of infection increased the pH from 7.2 to 7.6 (pH data of ARM2 not shown). Nevertheless, the use of GUR is recommended as long as process conditions are carefully controlled.
Conclusions
It was demonstrated that single-use HFBRs appear to be a promising alternative for production of influenza A virus with very high HA (up to 3.87 log10 (HA units/100 μL)) and TCID50 (up to 5.6 × 1010 virions/mL) titers, and similar CSVY when compared to other established production platforms. Further research with HFBRs for production of other seasonal and pandemic influenza A and B strains could explore, for example, infections at different MOIs and their impact on virus yields, as well as the scale
Acknowledgment
The authors would like to thank the Naval Health Research Center, USA, for sending the A/Mexico/4108/2009 influenza strain.
References (43)
- et al.
Influenza A (H1N1) vaccine efficacy in animal models is influenced by two amino acid substitutions in the hemagglutinin molecule
Virology
(1989) - et al.
Comparison of large-scale mammalian cell culture systems with egg culture for the production of influenza virus A vaccine strains
Vaccine
(2001) - et al.
Scalable production of influenza virus in HEK-293 cells for efficient vaccine manufacturing
Vaccine
(2010) - et al.
Industrial choices for protein production by large-scale cell culture
Curr Opin Biotechnol
(2001) - et al.
Large-scale mammalian cell culture
Curr Opin Biotechnol
(1997) - et al.
A new MDCK suspension line cultivated in a fully defined medium in stirred-tank and wave bioreactor
Vaccine
(2010) - et al.
Establishment of a mink enteritis vaccine production process in stirred-tank reactor and Wave® Bioreactor microcarrier culture in 1–10 L scale
Vaccine
(2007) - et al.
Statistical optimization of influenza H1N1 production from batch cultures of suspension Vero cells (sVero)
Vaccine
(2011) - et al.
Wave microcarrier cultivation of MDCK cells for influenza virus production in serum containing and serum-free media
Vaccine
(2006) - et al.
Productivity, apoptosis, and infection dynamics of influenza A/PR/8 strains and A/PR/8-based reassortants
Vaccine
(2012)