OR WAIT null SECS
The authors conclude that miniature bioreactors can adequately predict the cell culture kinetics in scaled-up reactors using equal mixing times.
There is a growing need to accelerate bioprocess development for mammalian cell culture. Major pharmaceutical and biotech firms are facing challenges to reduce process development costs and cultivation times (1). The conventional method for mammalian cell-line development usually involves a series of shake flasks for screening the cell-line prior to large-scale cultivation. The shortcomings of this method include long development times, laborious operation, and limited experimental throughput, which result in slow bioprocess development of mammalian cell cultures.
Various scale-down miniature bioreactors have been designed to speed up the bioprocess development of mammalian cell cultures. Generally, it is accepted practice to perform the small-scale experiment in a high throughput and highly parallel manner. Current technology endeavours to enable high-throughput process development include the use of microtiter plates, miniature stirred-tank bioreactors, and microbioreactors (2, 3, 4).
Miniature stirred-tank bioreactors (MBRs)-based on the conventional stirred-tank reactor (STR)-enable a rapid and scalable experimental process development. Experiments are usually carried out in 4-16 parallel reactors running simultaneously at scales of 10 mL to 500 mL. The main advantages of these reactors are reduced cultivation times and costs, and the ability to conduct continuous monitoring and real-time visualization of key process parameters in each single bioreactor (3, 4). Moreover, the capacity of miniature stirred bioreactors for inline monitoring and control of pH, dissolved oxygen (DO), and temperature could make these reactors an excellent alternative to shake-flask systems for early-stage mammalian cell-culture bioprocess development.
Scale translation of miniature bioreactors to benchtop reactors remains a crucial issue for mammalian cell-culture processes. Mixing directly affects the heat transfer, gas dispersion, and blending of different media components in the reactor. Furthermore, poor mixing in bioreactors can result in pH, nutrient, and temperature gradients, as well as poor control of operating parameters (5).
Mixing in mammalian cell-culture reactors is achieved using either marine or pitched-blade impellers to minimize shear damage. These impeller designs will create low shear stress and gently mix the culture (6). Detailed engineering characterization, such as mixing time, is imperative to understand the performance of a bioreactor. By characterizing the system during the typical operating range, suitable parameters for scale translation can be identified. In this study, mixing time has been determined in a prototype version of a commercial MBR system and used as a criterion for translation to bench-scale stirred bioreactors.
Materials and methodsCell lines and media
All experiments in this work were carried out using a Chinese hamster ovary (CHO) cell line expressing an IgG antibody. The medium used was of an animal-free origin and was chemically defined.
All MBR work was carried out with a 0.5-L miniature bioreactor system (HEL Ltd, UK) run in parallel with a working volume of 0.35 L. The operating parameters of pH, DO, and temperature were controlled at 7.1, 30%, and 37°C, respectively. Agitation was provided either by a single three-blade marine impeller (direct driven) or a four-bladed marine impeller (magnetic-bottom-driven) with an agitation rate set to match the mixing time in the 5L-STR. Aeration in the bioreactor vessel was achieved either by a horseshoe sparger or singular hole sparger.
The experiments were carried out in 5-L stirred-tank bioreactors (Biostat B-DCU, Sartorius) run in parallel with a working volume of 3.5 L using the same operating parameters as given for the MBRs. Agitation was provided by a single three-blade segment marine impeller. Aeration in the bioreactor vessel was achieved via a horseshoe-type sparger.
In all experiments at both scales, the mode of operation was batch phase from day zero until day six, followed by fed batch from day seven onwards. The cultures were bolus-fed once a day to maintain the glucose concentration at 2 gL-1.
Mixing time was measured in the HEL-BioXplore miniature bioreactor using the pH tracer method (6). The experiment was conducted using two types of impellers; direct driven and magnetic driven with the horseshoe and singular hole sparger. The mixing time in the 5L-STR as a function of agitation rate was previously determined by investigator Silk (6).
Cell number and viability
The cell cultures were sampled daily and cell concentration and viability were analyzed using an automated cell-counting device, VI-Cell XR (Beckman Coulter), which automates the trypan blue dye exclusion method.
IgG HPLC assay
The quantification of IgG produced was determined through the use of a high-performance liquid chromatography (HPLC) system (Agilent Technologies) with a Protein-G affinity column.
Results and discussion
Mixing time is a useful parameter to measure the mixing efficiency in a reactor and homogeneity of a fluid when agitated by an impeller. These mixing time values vary depending on the impeller and/or sparger designs and geometry of the reactor. Because the MBR exhibits a similar geometry as a conventional STR, an empirical correlation proposed by Nienow (7) was applied to compare experimental and predicted values for the different MBR configurations (Equation 1).
Figure 1 shows that mixing time is inversely proportional to agitation rate. The mixing time achieved ranges from 5-14 seconds and 4-11 seconds for experimental and predicted values, respectively. Although both experimental data sets show the same trend as the calculated values, the measured values are either slightly higher (magnetic driven) or lower (direct driven) compared with theoretical predictions. These deviations may be due to differences in average energy dissipation rate that were not measured. Nevertheless, the results suggest that the Nienow (7) correlation can be used to predict the mixing times in the MBR with good accuracy considering the experimental variations.
Matched mixing time between miniature bioreactors and bench-stirred tank reactors
In this case study, a typical fed-batch CHO cell culture process was chosen and mixing time was used as a scale-translation criterion. The operating conditions for each reactor system with regard to agitation speed, aeration rate, and feed addition were adjusted accordingly.
Figure 2: Chinese hamster ovary (CHO) growth kinetics in fed-batch cultures for two reactors: a 5-L benchtop (solid line) and a miniature bioreactor (MBR) (dashed line). Graph (A) represents viable cell concentration; graph (B) shows cell viability for each type of bioreactor. The arrows (↓) indicate the points in the process in which feed is added.
The CHO cell-growth profile and percentage viability for the two types of reactors, miniature bioreactors, and bench 5-L stirred-tank bioreactors are depicted in Figure 2. Both cultures had a prolonged exponential phase but reached the peak viable cell concentration (VCC) at different days (Figure 2A). The parallel MBR cultures exhibited a slower exponential growth between 70-168 hours of cultivation compared with the benchtop cultures. The peak viable cell concentrations for the two reactors, however, are almost identical (Table I).
The benchtop culture showed a considerably longer stationary phase compared with the MBR and achieved higher cell viability on day 14. Final percentage viability of the MBR was at 60%, while the viability in the 5-L stirred-tank reactor was at 80% after >300 hours (Figure 2B). The higher viability observed in the stirred-tank reactor on day 14 might be due to the better monitoring and control of gas sparging in the larger reactor. Besides that, the feed addition, which was performed daily from day seven, enabled prolonged cell viability and antibody production until the harvest day.
Figure 3 shows the final antibody production for both bioreactor cultures. Based on the HPLC analyses, the MBR culture reached a final IgG titer of 0.69 gL-1, which is 17% lower than that in the 5 L STR with 0.83 gL-1. The derived growth parameters for both systems, however, show excellent agreement in terms of specific production rate (qP) and generally good comparability with regard to the cumulative integral viable cell concentration (CiVC) (Table I).
Figure 3: Antibody production in Chinese hamster ovaries (CHO) growth kinetics in fed-batch culture for two reactors: a 5-L stirred-tank reactor (solid line) and a miniature bioreactor (dashed line).
This study has demonstrated the potential of miniature bioreactors for scale translation using equal mixing time as a criterion. The mixing times obtained for experimental and predicted values in the miniature bioreactors were comparable, which suggested that the correlation can be applied to predict mixing times at this scale. Scale comparison cultivations for fed-batch CHO cell cultures were carried out using miniature bioreactors and compared with a standard 5-L stirred-tank reactor at matched mixing time. The results show that the MBR gave comparable CHO cell growth and product kinetics compared to that at the 5-L scale. The results of this investigation suggest that MBRs can be used as scale-down models of bench-scale bioreactors.
Mohd Helmi Sani thankfully acknowledges the funding by the Ministry of Higher Education, Malaysia and Universiti Teknologi Malaysia.
1. H. Zhang et al., Cur. Pharm. Biotech. 11 (1), pp. 103-112 (2010).
2. M.A. Hanson and G.Rao, “Miniaturization of bioreactors,” in Encyclopedia of Industrial Biotechnology: Bioprocess, Bioseparation and Cell Technology, M.C. Flickinger, Ed. (John Wiley and Sons, 2010).
3. J. Betts and F. Baganz, Microbial Cell Factories, 5 (1), pp. 21-35 (2006).
4. R. Bareither and D. Pollard, Biotech. Prog., 27 (1), pp. 2-14 (2011).
5. Z. Xing et al., Biotech. Bioeng. 103 (4), pp. 733-746 (2009).
6. N.J. Silk, “High Throughput Approaches to Mammalian Cell Culture Process Development,” (EngD thesis), University College London (2014).
7. A. Nienow, App. Mech. Rev. 51 (1), pp. 3-32 (1998).
All figures are courtesy of the authors.
About the Authors
Mohd Helmi Sani is research engineer in the Department of Biotechnology and Medical Engineering at the Universiti Teknologi Malaysia; and Frank Baganz is senior lecturer in the Department of Biochemical Engineering at the University College London.
Article DetailsBioPharm International
Vol. 29, No. 1
Pages: 47–49, 55
When referring to this article, please cite it as M.H. Sani and F. Baganz, BioPharm International 29 (1) 2016.