Evaluation of a Single-Use Bioreactor for the Fed-Batch Production of a Monoclonal Antibody

Abstract

In this study, we tested the combination of a disposable bioreactor and a disposable dissolved oxygen sensor as a replacement for our standard bioreactors. The possibility to run a fed-batch cell culture process developed for the production of a monoclonal antibody in a 50-L single-use bioreactor was investigated. The single-use bioreactor was assessed both as a seed train and as a production bioreactor. Therefore, three configurations corresponding to different combinations of the 50-L disposable bioreactor and the reference 5-L glass bioreactor (fully scalable up to 300 L) were compared.

In the past decade, biopharmaceutical manufacturing processes have undergone multiple changes resulting in significant improvements in efficiency. In parallel with the development of high producing cell lines and robust chemically defined media for cell culture, the constant evolution of disposables has led to simpler operations. The use of disposables eliminates cleaning and sterilization steps, as well as cleaning validation, thus reducing costs and the time of operation per batch. Traditional disposable devices such as filters, tubing, bags, bottles, and syringes have commonly been used in biopharmaceutical manufacturing since the 1990s. The breakthrough in disposable bioreactor technology development in terms of larger capacities was the use of bag systems for cell culture. The 20-L rocker system was commercially available in 1998 and the technology became a success, especially in cell expansion operations. It allowed working with complete sterility, thereby securing the industrial cell culture processes. Further development of this system led to higher volumes and to the equipment available today on the market.

(MERCK SERONO SA)

The next step was the development of stirred-tank bioreactors with a configuration similar to conventional stainless steel bioreactors. Compared to traditional wave bioreactors, stirred-tank bioreactors offer the benefits of sparging, stirring of the suspension, and a higher utilization rate of the bag size as cultivation space. Such bioreactors are considered for use in the industry in cell amplification processes, to reach higher cell densities, or as production bioreactors. The single-use disposable bioreactor (SUB) from HyClone (Thermo Fisher Scientific) entered the market in 2006. Currently, disposable bioreactors up to 2,000 L (SUB, Hyclone and XDR, Xcellerex) culture volume are commercially available and plans are to develop 3,000 L bioreactor systems over the next few years.¹

In addition to disposable bioreactors, innovative single-use sensors are currently being developed, which will allow a cell culture process to run long-term in a fully disposable system. Single-use sensors also will solve technical problems sometimes raised by using plastic bags in single-use bioreactors, which can interfere with the functioning of stainless steel sensors because of static electricity problems, and therefore create drifts especially for pH measurement.²

In this study, we tested the combination of a disposable bioreactor and a disposable dissolved oxygen (DO) sensor as a replacement of a standard bioreactors to run a fed-batch cell culture process developed for the production of a monoclonal antibody (MAb). The disposable equipment selected was the 50-L stirred-tank SUB by HyClone coupled to the TruLogic RDPD controller based on DeltaV technology and the disposable TruFluor DO probe by Finesse Solutions.

The SUB was assessed both as a seed train and as a production bioreactor. Therefore, three configurations corresponding to different combinations of the 50-L disposable bioreactor and the reference 5-L glass bioreactor (fully scalable up to 300 L) were compared (Figure 1): seed train and production in a 5-L glass bioreactor (named 5L/5L); seed train in the 50-L SUB; production in 5-L glass bioreactor (named SUB/5L); and seed train and production in the 50-L SUB (named SUB/SUB).

Material and Methods

Cell Culture and Bioreactor Operation

The cell line used in this study was a Chinese hamster ovary (CHO) cell line developed for the production of a MAb. The cell culture process was performed using chemically defined cell culture media and feeds. A typical seed train was used for the production runs (Figure 1). After vial thawing, cells were grown in T-flasks and shake flasks for six days at 37°C, 5% CO₂, and then transferred into a 2-L Cultibag RM Optical (Sartorius Stedim Biotech GbmH), followed by a 50-L wave bag. The two last steps of the seed train were performed in bioreactors (N-2 and N-1 steps). To properly assess the performance of the SUB as a seed train bioreactor and a production bioreactor, the cell suspension from the 50-L wave bag was split in a 50-L SUB (Hyclone, Thermo Fisher Scientific Inc.) and a 5-L benchtop-scale glass bioreactor (BIOSTAT B-DCU Quad, Sartorius-Stedim Biotech GbmH) for the N-2 step. The N-1 step consisted of the passage of the N-2 bioreactor, removing part of the suspension, and adding fresh media. Then, the seed train in the 5-L glass bioreactor was used to inoculate two 5-L glass bioreactors for the production phase. The seed train in the SUB was used to inoculate, in parallel, one SUB and two 5-L glass bioreactors for the production phase. The experimental plan is described in Figure 1 and was performed twice.

Figure 1. Description of the process and experimental scheme. The cell amplification was performed in T-flasks and shake-flasks, followed by a passage in a 2-L bag rocker and a 50-L bag rocker. The inoculum was splitted to inoculate the N-2 bioreactors (SUB and 5-L glass vessel bioreactor). The N-1 SUB seed train bioreactor was used to inoculate two 5-L glass vessel bioreactors and itself as production bioreactor with the remaining inoculum. The 5-L glass vessel seed train bioreactor was used to inoculate one other 5-L bioreactor and itself as production bioreactor.

The process developed in the 5-L glass bioreactors was adapted for the SUB as follows: pH, temperature, DO setpoints, and feeding strategy were unchanged compared to the 5-L bioreactor. Values for gas flow rates and agitation setpoint were scaled up by keeping the headspace renewal rate and the power per volume constant, respectively. The pH regulation was performed using CO₂ in overlay and sodium hydroxide for the 5-L glass bioreactors.

Partial pressure of carbon dioxide (pCO₂), pH, and oxygen (pO₂) were controlled offline on ABL5 (Radiometer Medical). Viable cell density (VCD) and viability were measured using a Vi-Cell automatic cell counter (Beckman Coulter). Metabolites and electrolytes were controlled offline using a Nova Bioprofile 100+ (Nova Biomedical). Protein production was determined using a Protein A–based assay on the Gyrolab platform (Gyros).

Using a Disposable DO Probe

The 50-L SUB was operated using a TruLogic RDPD (R&D and process development bioprocess) controller (Finesse Solutions). The disposable TruFluor DO probe (Finesse Solutions) was compared to the InPro6800 polarographic probe (Mettler Toledo) at a setpoint of 40%. The disposable DO probe was inserted in a sleeve manufactured with the SUB 50-L bag. The sleeve contained the disposable sensor, and the probe contained a non-invasive reader connected to the transmitter. The standard DO probe was inserted with a Kleenpack connector (Pall Corporation) after sterilization. After insertion, the standard DO probe was calibrated as usual. The DO was controlled using the InPro6800 probe, but signals from both probes were recorded on the controller.

Results

Process Scale-Up in the SUB

The seed trains performed in the 50-L SUB and in the 5-L glass bioreactors resulted in comparable cell density and viability (Figure 2). For all configurations, viable cell densities at the end of the growth phase ranged from 2.7 to 3.2 × 10⁶ cells/mL, and viability ranged from 95.5 to 98.7%.

Figure 2. Profiles of viable cell density (SUB VCD and 5-L Bio VCD) and viability (SUB Viab and 5-L Bio Viab, in %) of the seed train bioreactors. The passage of the wave bag into the N-2 bioreactors was performed at working day 20, and the passage from N-2 to N-1 bioreactor was performed at working day 23. VCD and Viab were determined using the Vi-cell cell counter (Beckman Coulter). The graphs show mean values with errors bars of two different runs. The variability of the viability is so low that the error bars are not visible.

The production phase performed in the 50-L SUB and in the 5-L glass bioreactors also gave similar results for the maximum VCD obtained at production day 7 (working day 33) and integral viable cells (IVC) (Figure 3 and 4). The IVC curves (Figure 4) show comparable cell growth throughout the process between different scales and configurations, whether the seed train material was coming from the 50-L SUB or the 5-L bioreactor. The metabolites profiles such as glucose, lactate, glutamine, and glutamate were all comparable between different scales and configurations (data not shown).

Figure 3. Viable cell density bar graph at production day 7 (working day 33) of the production bioreactors. The viable cell density reached a maximum value at production day 7 (working day 33) and started decreasing on the next day. The values shown here are means of two or three values with error bars.

The MAb titers obtained at production day 7 (corresponding to the highest point of viable cell concentration) were comparable for all configurations (Figure 5), again demonstrating the similar performance of the SUB and the 5-L bioreactor in seed train and production. The titer was plotted against the IVC (slopes: 5L/5L 22.62, SUB/SUB 21.81, SUB/5L 19.23) and the relationship was found to be linear, as expected.³ The slopes were very close for the three configurations, demonstrating that the same specific productivity is reached in different configurations.

Figure 4. Integral viable cell density of the three configurations (SUB/SUB, SUB/5L, and 5L/5L). The graphs are means of two or three runs with error bars.

The pH in the bioreactor was regulated using CO₂ in overlay in the 5-L glass bioreactors and in the 50-L SUB. The pCO₂ levels were compared between the SUB and the 5-L runs to assess the CO₂ stripping performance of the SUB (Figure 6). As shown in Figure 6, the 50-L SUB cultures or the 5-L bioreactor coming from the SUB seed train have appropriate pCO₂ levels below 60 mmHg. In addition, to assess the impact of pH regulation using CO₂ instead of acid, pCO₂ data were compared to historical data from a 5-L acid-regulated run and showed that the 50-L SUB production run with the 50-L SUB seed train have even lower pCO₂ levels than the acid-regulated runs (40–54 mmHg).

Figure 5. Product titer bar graph at production day 7 (working day 33). The viable cell density reached a maximum value at production day 7 (working day 33) and started decreasing on the next day. The values showed here are means with error bars of two or three values.

DO Probe Comparison

A drawback of single-use bioreactors is the possible impairment of sensors based on electric potential differences because of the static electricity phenomena.⁴ However, disposable sensors now are available based on fluorescence, which may not cause the same problems. One of these probes was tested in the single-use bioreactor to monitor the density of oxygen, in comparison to a reference stainless steel polarographic probe. The trends of the two probes are identical, even if a small constant difference (<2%) is noted between the two curves, probably because of calibration (data not shown). Thus, in this specific configuration, no impairment of the stainless steel DO probe was observed in the environment of the SUB. The disposable fluorescence DO probe showed a comparable performance, and could probably be used to replace the reference stainless steel probe.

Figure 6. The pCO2 profiles for the different configurations (SUB/SUB, SUB/5 L, and 5 L/5 L), measured with ABL-5 gas analyzer (Radiometer). The values showed here are means with error bars of two or three values.

Discussion and Conclusion

The aim of this study was to asses a disposable bioreactor in combination with a disposable probe for a fed-batch MAb production process. The equipment chosen was the HyClone SUB coupled to the TruLogic RDPD controller and the disposable TruFluor DO probe by Finesse Solutions. The single-use DO sensor showed comparable results to conventional ones. The equipment was readily implemented because the set-up time was only one day.

The SUB gave results comparable to the 5-L glass vessel bioreactor (small-scale reference for the process) for the seed train and the production steps. This shows that despite different material, agitation, and aeration, the disposable bioreactor had a performance similar to standard bioreactors, at least for the fed-batch process tested here. The scale-up to 50-L also was straightforward. Given that all HyClone disposable bioreactors have the same overall reactor geometry ratio up to 2,000 L, it can be expected that the scale-up to larger volumes such as 300 L could be performed using the same principles. The scale-up to a higher volume such as 1,000 L could be more complex because process scale-up is rarely linear between such different scales.

The single-use bioreactor showed the capacity to be used either as a seed train bioreactor or a production bioreactor, or both. If this double use is to be implemented, the bag aeration configuration should be carefully defined to be able to cope with different oxygen demand in cell expansion and production. In this case, the same bag was used for both phases, a limitation in oxygen flow rate appeared toward the end of the culture.

The bioprocess container bag used for these experiments was equipped with a 20-mm sparger membrane. The bubbles released by this system were small enough to have sufficient oxygen transfer to the culture, but big enough to strip CO₂. The bag is now available with a dual sparge system, consisting in a 20-mm porous frit for the oxygen transfer and an open pipe for CO₂ stripping, enlarging the range for pCO₂ stripping. Many different disposable bioreactors systems coexist on the market and new versions are frequently released, showing the high dynamism of single-use technology. Each system presents its own features and advantages. Some other disposable bioreactors currently are being assessed in our company.⁵

This study enabled us to demonstrate the applicability of using a single-use bioreactor for producing a MAb at 50-L scale, and we can expect that further scale-up to at least 300-L can be achieved. In the future, it can be expected that disposable bioreactors will become far more common in biopharmaceutical manufacturing. Their use is of specific interest when producing material for early clinical trials to avoid a capital investment early, when the final production bioreactor volume, as well as the future of the molecule, are unknown. Some people claim that the use of disposable bioreactors also has big advantages when building a new facility, because the need for utilities might be reduced in a fully disposable environment, therefore reducing start up time, installation costs, and campaign turnaround.⁶ On the other hand, some concerns exist about the environmental impact of disposables, although assessing the latter is far from simple. The reduced use of purified water, clean and pure steam, and cleaning chemicals compared to stainless equipment has to be balanced with the increased plastic waste. One way to reduce the impact of such waste could be to convert back part of the 32.6 GJ/ton of energy stored in plastic in waste-to-energy incineration facilities, not necessarily solving the issue of carbon footprint.⁷ The ultimate solution might reside in recycling these disposable products, requiring further development on innovative transformation methods. Some other interesting future directions with respect to single-use bioreactors could be the development of systems for perfusion process applications, as well as more insights on leachables and extractables.

Emmanuelle Cameau is a biotech process sciences upstream specialist, Georges De Abreu is a biotech central services manager, Alain Desgeorges, PhD, is a biotech process sciences upstream coordinator, Elodie Charbaut Taland, PhD, is a biotech process sciences manager, and Henri Kornmann, PhD, is a biotechnology production director, all at Merck Serono SA, Aubonne, Switzerland, +41(0)218217111, emmanuelle.cameau@merckserono.net

References

1. Brecht R. Disposable Bioreactors: Maturation into pharmaceutical glycoprotein manufacturing. In: Eibl R, Eibl D, editors. Disposable Bioreactors. Springer: Advances in Biochemical Engineering/Biotechnology; 2009. p. 1–31.

2. Selker M, Paldus B. Single-use sensors for Upstream applications. Next Gen Pharm. 2009; 16. Available from: http://www.ngpharma.com/article/Single-use-Sensors-for-Upstream-Applications.

3. Smolke C, editor. The metabolic pathway engineering handbook. Boca Raton, FL: CRC Press; 2009.

4. Parmeggiani L. Encyclopaedia of occupational health and safety: A-K. Switzerland: International Labour Office; 1983.

5. Poles A, et al. Comparison of fed batch cell culture performances between stainless steel and disposable bioreactors, submitted to Biopharm Int.

6. Ravisé A, Cameau E, De Abreu G, Pralong A. Hybrid and disposable facilities for manufacturing of biopharmaceuticals: Pros and cons. In: Eibl R, Eibl D, editors. Disposable bioreactors. Springer: Advances in Biochemical Engineering/Biotechnology; 2009. p. 185-219.

7. Porter R, Roberts T, editors. Energy savings by wastes recycling, Commissioned by European Economic Communities. Elsevier, London; 1985.