Single-Use Bioreactors for the Rapid Production of Preclinical and Clinical Biopharmaceuticals
Fed-batch processes were scaled up from traditional bench-scale bioreactors to large-scale single-use systems.
In the current competitive biopharmaceutical production landscape, the substantially higher process development costs mean that new candidate molecules are increasingly under pressure to advance to clinical proof-of-concept stage in the shortest possible time with a minimum commitment of capital resources. Single-use bioreactors have the potential to alleviate both of these constraints because they confer specific advantages over conventional stainless-steel bioreactors. Acquisition and implementation costs of disposable bioreactors tend to be much lower compared to traditional stainless-steel bioreactor systems. Some of these advantages can be attributed to their modular nature, which enables the use of existing manufacturing facilities with minimal modifications to the current infrastructure, allowing execution of a straightforward “roll in-roll out” concept. Furthermore, because the product contact surface is changed with each experiment/campaign, carry over between fermentations is non-existent, thus removing the need to perform cleaning verification/validation and enabling a more rapid turn-around between products. As a result, single-use systems from various manufacturers are now available on the market, and pharmaceutical development groups have increased the frequency with which these systems are used to produce biopharmaceuticals (1, 2).
Although the benefits concerning this technology are fairly clear, one of the primary unresolved concerns focuses on performance comparability between single-use and conventional stainless-steel systems. In this study, the authors present data from two different fed-batch processes producing monoclonal antibodies (mAbs) using both system types. To achieve good correlations between bench-scale development reactors and the single-use bioreactors, a continuous stirred tank reactors (CSTR) design was used throughout the study as described by De Wilde et al. (3). One major outcome of this work was the development of an entirely single-use upstream platform for the manufacture of biopharmaceuticals from vial thaw to harvest clarification.
Materials and methods
Recombinant Chinese hamster ovary (CHO) cell lines were used to produce two different fully human mAbs in fed-batch processes. Process A was used to produce an IgG2 and Process B to produce an IgG1. Both processes used a combination of commercially available “off the shelf” and proprietary basal media and feed solutions.
Initial process development was performed in traditional bioreactor designs at the 2-L, 5-L, and/or 10-L scale (Applikon Biotechnology, Netherlands). Preclinical and clinical material was generated in the single-use Biostat STR (Sartorius Biotech, Germany) at the 200-L and 1000-L scales. A standard bag design consisting of two marine-type impellers, a macro-sparger and optical sensors (PreSens, Germany) for pH and dissolved oxygen was used. Stirrer geometry, tip speed, and overall power input were matched as closely as possible among all bioreactors, similar to the approach taken by De Wilde and Adams (4) and Noack et al. (5).Figure 1 illustrates tip speed and power input values among the different bioreactors. Although the power input of the 10-L system is slightly higher than the rest, the overall tip speed remains comparable to that of the disposable units.
Standard shaker flasks were used during the seed train expansion process. Cells were cultivated in a CO2 incubator maintained at 37.0 °C and 5-7% CO2. Scale-up passaging occurred every two to three days until sufficient cell mass had accumulated to directly inoculate the bioreactors regardless of scale, with the exception of the 1000 L. The 200-L bioreactor was inoculated at an initial volume of 50-80 L operation range due to variation in the accumulated cell mass of the seed train. Basal medium was added to the reactor to “passage” the cells by dilution and considered in “scale-up” mode until a predefined total cell number was reached; this time point was defined as fed-batch day zero. If required, the 200-L unit was used as part of the seed train expansion process and culture was directly transferred to the 1000-L reactor to achieve a target starting cell density on inoculation day.
Fermenter process parameters (e.g., pH, temperature, pO2) were maintained by biocontrollers manufactured by either Sartorius (DCU 2/3/Biostat STR, Sartorius Biotech, Germany) or Applikon (iControl, Applikon Biotechnology, Netherlands). All bioreactors were configured with a macro-sparger to deliver mixed gas consisting of air, N2, CO2, and O2. Dissolved oxygen was controlled at 40% air saturation, pH at 6.7-7.3, and the initial process temperature was set to 36.5 °C. For Process B, a temperature shift to 33.0 °C was performed once the cell density reached 8 x 106 cells/mL.
The bioreactors were sampled daily to determine viable cell density and viability (Cedex cell counter, Roche Diagnostics, Germany) as well as glucose and lactate values (YSI, Yellow Springs Instruments, OH, USA). Dissolved oxygen, pH, and pCO2 levels were measured offline using a blood gas analyzer (Siemens Diagnostics, NJ, USA) and used to verify the pH and dissolved oxygen probes of the bioreactors. mAb titers were measured using a protein A high-performance liquid chromatography (HPLC) method. The fed-batch process was terminated after reaching predefined harvest criteria (viability and/or time based), and the crude harvest was clarified by dead-end filtration.
The 200-L Biostat, in addition to being the seed source for the 1000-L system, was also used to verify the bench-scale fed-batch process and to produce preclinical material. The 1000-L reactor was used to produce GMP material for clinical trials. The entire clinical upstream process consisting of the 200-L and 1000-L bioreactors as depicted in Figure 2, uses disposable materials throughout, including bags for basal media, feeds, and clarified harvest.
Results and discussion
To compare the performance of the single-use bioreactors to traditional fermenters, critical process parameters for each vessel type are plotted together. Data from multi-use reactors are plotted as grey dots to define a point cloud of expected values and single-use disposable runs are shown as continuous lines. Figure 3 and Figure 4 summarize the data obtained in Process A and B, respectively. As shown for Process A, cell growth, viability, glucose, and lactate concentrations obtained with the single-use systems follow the general trends defined by the bench-scale reusable bioreactors. Lactate values were slightly higher in the bench-scale reactor cultures. These cultures were used for the initial process development, specifically to optimize the feeding process. The slightly higher viability can be attributed to slight differences in the Cedex cell counter used during these cultures. Offline pH and pO2 measurements are well aligned, demonstrating that single-use bioreactor control is comparable and was not affected by vessel type or the measurement technology (i.e., conventional glass electrodes and Clark oxygen probes vs. single-use fluorescence patches). As a result, most Biostat STR data were highly comparable to the data obtained at small scale, and certain parameters shifted towards more desirable profiles (i.e., reduced lactate concentrations and increased titer per unit time). Process B in Figure 4 shows similar behavior, with tighter clustering of data among the reactor types and scales. For some of the bench-scale reactors as well as the 1000-L disposable unit the temperature shift was programed into the control unit at a rate of 0.5 °C per 30 minutes. The temperature decline was linear over the 3.5 h time span with less than 0.5 °C under shoot. The 200-L disposable unit used a heating blanket, therefore, active cooling like in the 1000-L unit was not possible. The temperature shift here took approximately 11 h with a slightly larger under shoot (data not shown). Nevertheless, this second process also demonstrates the feasibility of implementing temperature shift control schemes with single-use bioreactors. Overall, the temperature shift for Process B was necessary to obtain the correct product quality attributes. In addition, it increased the space-time productivity of the antibody.
In addition to similar in-process culture performance, critical product quality attributes of the harvest like protein aggregation rates, charge heterogeneity, and glycosylation profiles remained consistent across all bioreactor systems (data not shown). Overall, the data show good comparability between the bench-scale cultures and the large-scale fermentations, demonstrating both the scalability of each processes and the successful integration of single-use bioreactor technology into the authors’ clinical manufacturing platform for mAb fed-batch processes.
One of the primary concerns surrounding the adoption of single-use reactor technology for full commercial manufacturing is bag robustness. Due to the nature of the materials employed, defects can potentially be introduced at any point during the life of the bag; at manufacture, shipping, unpacking/setup, or operation. During the preliminary implementation at Bayer HealthCare, a 200-L culture was terminated one day after inoculation because a 24 h medium sterility-hold at 50 L did not reveal a pinhole defect. That incident triggered an in-house integrity test of all single-use reactors prior to use which consisted of pressurizing a mounted bag to 0.25-0.35 psi (17-24 mbar) and monitoring the pressure for 10 min (200 L) or 20 min (1000 L), shown in Figure 5. An intact bag will only lose ~0.02 psi (1.4 mbar) during the test period whereas a defective bag will rapidly depressurize during the test procedure. Unfortunately, this practice does not guarantee success, as the metal wall of the bag holder can mask potential defects during the pressure hold and despite a passing test. Bag failure can result in contamination and lost cultures.
To mitigate this risk, Sartorius Stedim Biotech developed and qualified a pressure test comparable to the one described by the authors but utilizing a patented fleece that inserts a mesh gap between the bag and holder surfaces to prevent these potential masking effects. The approach allows post bag installation and pre-use testing of the entire single-use bioreactor assembly (6). The fleece is designed to be easily removed prior to the start of a run, to ensure normal bioreactor temperature control. This method has been qualified for reliable defect detection for various bag sizes (i.e., a correlation between reliably detectable defect size and pressure decay has been established) (7). This method is, therefore, a more reliable detection method and fully encompasses any typical defects that might be incurred during transportation, storage, unpacking, and installation that cannot be identified by simple visual inspection. In essence, this approach is similar to conventional stainless-steel bioreactor practices, where the sterilization method is qualified during installation qualification/operation qualification (IQ/OQ) and a pressure hold test is often performed for risk mitigation purposes to detect any miss-assemblies during regular operation. The sterilization IQ/OQ can be compared to the vendor assembly and sterilization qualification of single-use bags and the pressure decay testing serves in a comparable way as a risk mitigation tool.
Conclusion
As more and more single-use bioreactors and bags are used in late-phase and commercial production, especially with the availability of 2000-L single-use stirred tank bioreactors, the need for consistent bag quality, robustness, improved assurance of supply, change management, and business continuity planning become crucial. To satisfy these requirements, Sartorius Stedim Biotech developed a new polyethylene film in close collaboration with resin and film suppliers to meet future industry needs and to further improve bag consistency, robustness, and performance for single-use bioprocessing applications. During development, attention was paid to working with vendors to ensure a stable supply chain, clearly defining and controlling the resin and additive packages, establishing acceptable film extrusion ranges with design space studies, and optimizing of the bag welding process. The result was a system with significantly improved strength and flexibility characteristics, alleviating many of the potential avenues to introduce defects during the manufacture and handling of these single-use bioreactors (8). Furthermore, cell culture and leachable studies were performed during all stages of the new film development to ensure the new formulation is not toxic and does not impede cell growth (9).
All these aspects help to pave the way towards wide implementation of commercial scale single-use biomanufacturing, benefitting from the initially mentioned advantages of reduced upfront investment, flexibility, quick change-over, and minimal validation effort.
References
1. C. Heath and R. Kiss, Biotechnol. Prog. 23, 46-51 (2007).
2. M. George et al., Biotechnol. Bioeng. 106, 609-917 (2010).
3. D. De Wilde et al., BioProcess Int. 7 (Suppl 4) 36-41 (2009).
4. D. De Wilde and T. Adams, Eur. J. Parent. Pharm. Sci. 15, 41-46 (2010).
5. U. Noack et al., “Single-Use Stirred Tank Reactor BIOSTAT CultiBag STR: Characterization and Applications,” in Single-Use Technology in Biopharmaceutical Manufacture (Eds. R. Eibl and D. Eibl), John Wiley & Sons, Inc. (Hoboken, NJ, USA, 2010).
6. M. Stering et al., Genetic Engineering & Biotechnology News, 33 (11) (2013).
7. M. Stering et al., BioProcess Int. 12 (Suppl 5) 58-61 (2014).
8. E. Vachette et al., BioProcess Int. 12 (Suppl 5) 38-42 (2014).
9. E. Jurkiewicz et al., Verification of a new biocompatible single-use film formulation with optimized additive content for multiple bioprocess applications. Biotechn. Prog., Epub ahead of print, accepted: 21 May 2014, DOI: 10.1002/btpr.1934.
About the Authors
Rüdiger Heidemann, PhD, senior staff development scientist, Cell Culture Development, Global Biological Development, Bayer HealthCare LLC
Christopher R. Cruz, senior associate development scientist, Cell Culture Development, Global Biological Development, Bayer HealthCare LLC
Paul Wu, PhD, director, Upstream Development, Cell Culture Development, Global Biological Development, Bayer HealthCare LLC
Mikal Sherman, application specialist, Fermentation Technologies, Sartorius Stedim North America
Jessica Martin, field marketing manager, Single-Use Bioreactors, Sartorius Stedim North America
Christel Fenge, PhD, vice-president of marketing for fermentation technologies at Sartorius Stedim Biotech, Göttingen, Germany.
