The purpose of this study was to show Xcellerex™ X-platform bioreactor range scalability performance through a well characterized CHO fed-batch cell culture process, expressing an immunotherapeutic IgG monoclonal antibody.
We successfully scaled the IgG antibody producing process using the Xcellerex™ X-platform bioreactor systems from 50 to 2000 L (Fig 1) confirming that they are consistent pilot to production bioreactors. Similar peak viable cell density (VCD), growth profile, cell viability and titer were achieved across all scales. Metabolite, nutrient, partial pressure of CO2 (pCO2), pH and dissolved oxygen (DO) trends all tracked similarly across each bioreactor highlighting the scalability of the platform.
Our study showed that X-platform bioreactor systems offer easy scalable performance with the ability to reproduce small-scale results without multiple production runs.
Fig 1. Xcellerex™ X-platform cell culture bioreactor scale-up providing a working volume from 10 to 2000 L.
Introduction
We conducted two consecutive CHO fed-batch process campaign runs expressing the immunotherapeutic IgG monoclonal antibody, trastuzumab (Herceptin), firstly using the Xcellerex™ X-500 as the terminal production bioreactor and secondly using the Xcellerex™ X-2000 as the terminal production bioreactor. Additionally, the X-200 was used as the N-1 bioreactor in the seed train for the X-500 production batch. During the X-2000 production batch, we used the X-200 as the N-2 vessel and the X-500 as the N-1 vessel. The X-50 was used as a satellite production bioreactor during both campaigns and run alongside the larger production bioreactors to further demonstrate scalable performance from the same starting material.
The cell culture scale up strategy used was to maintain the same power per volume (P/V) of 63 W/m3 across all scales, with the agitation speed adjusted after feeding to keep P/V constant throughout the run duration. The cumulative initial air gas flow through sparge 1 and sparge 2 on a vessel volume per minute (vvm) basis was similar across scales. In order to accomplish this, we configured the DO controller to supply an air and oxygen mix in response to increasing demand from the cell culture; the sparge 2 air flow rate was manually adjusted daily, as needed, to keep total vvm approximately constant at each scale. This methodology was effective at maintaining a similar pCO2 trend amongst the bioreactors.
Materials and methods
Table 1. Consumables and reagents used during this study
| Type | Name |
| Cell line | Chinese hamster ovary (CHO) cell line expressing the immunotherapeutic mAb, trastuzumab |
| Seed train flask | 125 mL shake flask |
| Seed train flask | 500 mL shake flask |
| Seed train flask | 2000 mL shake flask |
| Bioreactor | Cellbag™ 50 L, BC11, pHOPT, DOOPT II |
| Bioreactor | X-50 bag, Fed-batch II |
| Bioreactor | X-200 bag, Fed-batch II |
| Bioreactor | *X-500 bag, Perfusion II |
| Bioreactor | *X-2000 bag, Perfusion II |
| Manifold | Click-in probe sheath |
| Reagent | ActiPro™ medium |
| Reagent | HyClone™ Cell Boost™ 7a supplement |
| Reagent | HyClone™ Cell Boost™ 7b supplement |
| Reagent | HyClone™ ADCF antifoam, irradiated |
| Reagent | 50% (w/w) glucose |
| Reagent | Sodium bicarbonate |
| Reagent | L-Methionine sulfoximine (MSX) |
| Reagent | Sodium hydroxide |
*Perfusion II bags can be used in fed-batch although they are not essential.
Reagent preparation for the fed-batch cell culture process
We prepared the following reagents:
- HyClone™ ActiPro™ cell culture medium supplemented with 37.5 µM methionine sulfoximine (MSX) prepared by sterile addition of 375 µL of 100 mM MSX to 1 L of the liquid medium.
- ActiPro™ powder medium was hydrated with a 3000 L Xcellerex™ magnetic mixer according to the recommended procedure and sterile filtered with 0.2 µm Supor™ EKV Kleenpak™ Nova Capsules.
- HyClone™ Cell Boost™ 7a supplement was hydrated according to the recommended procedure and sterile filtered with 0.2 µm Supor™ EKV Kleenpak™ Nova Capsules. The batch for the X-500 production run was mixed with a stand mixer. The batch for the X-2000 production run was mixed with a 3000 L Xcellerex™ magnetic mixer.
- HyClone™ Cell Boost™ 7b supplement was obtained in liquid form. The batch for the X-500 production run and both satellite X-50 bioreactors was sterile-transferred from bottles to carboys. The batch for the X-2000 production run was pooled and sterile filtered with a 0.2 µm Supor™ EKV Kleenpak™ Nova Capsule.
- Glucose 50% (w/w) was purchased in liquid form. The batch for the X-500 production run and both satellite X-50 bioreactors was sterile transferred from bottles to carboys. The batch for the X-2000 production run was pooled and sterile filtered with a 0.2µm Supor™ EKV Kleenpak™ Nova Capsule.
- Sodium bicarbonate was prepared from powder to formulate a 1 M solution and sterile filtered into bags using a 0.2µm Ulta™ Pure HC PES filter.
Seed train for the fed-batch cell culture process
We started the seed train with a vial thaw and culture initiation in a 125 mL flask after washing and reconstituting the pellet. During shake flask passages, we used ActiPro™ medium supplemented with 37.5 µM MSX to selectively maintain a population of cells expressing the gene of interest. The cells were kept in exponential growth phase and passaged to maintain a viable cell density less than 6 ×106 cells/mL. The flowchart of the seed train for each run is shown in Figures 2 and 3.
We installed our ReadyToProcess™ WAVE™ 25 bioreactor, then filled and conditioned the medium for at least 24 h using the settings detailed in Table 2. The pH and DO sensor calibration information was set in the UNICORN™ control software. Prior to inoculation a sample was taken and a pH 1-pt standardization performed. We used the WAVE™ 25 bioreactor as the N-2 bioreactor during the X-500 run and the N-3 bioreactor during the X-2000 run.
We installed the Xcellerex™ X-200 bioreactor bag and conducted a pressure decay test. DO and pH sensors were calibrated and installed into the Click-in probe sheath and autoclaved. After we installed the probes into the bag, we filled and conditioned the medium for at least 24 h using the settings in Table 3. The bioreactor was brought up to the 37°C temperature setpoint and a 1-pt pH standardization performed. The X-200 bioreactor was inoculated from the WAVE™ 25 bioreactor and settings adjusted as per the values in Table 4.
We then installed the X-500 bioreactor bag and performed a pressure decay test. DO and pH sensors were calibrated and installed into the Click-In probe sheath and autoclaved. After we installed the probes into the bag, we filled and conditioned the medium for at least 24 h using the settings in Table 5. The bioreactor was brought up to the 37°C temperature setpoint and a 1-pt pH standardization performed. The X-500 was inoculated from the X-200 bioreactor and settings adjusted per the values in Table 6.
Fig 2. Seed train schematic for Run 1.
Fig 3. Seed train schematic for Run 2.
Table 2. ReadyToProcess WAVE™ 25 rocker set points
| Parameter | Set point |
| Temperature | 37°C |
| pH control | N/A |
| DO set point | 40% |
| DO control | Air/oxygen |
| Rocking speed | 25 rpm |
| Rocking angle | 6° |
| Rocking motion | 30% |
| Air flow | 1.0 L/min |
| CO2 concentration | 7.5% |
| Starting volume (after inoculation) | 22 L |
| Target inoculation concentration | 0.4 ± 0.1 × 106 cells/mL |
| Target final concentration | < 6 × 106 cells/mL |
Table 3. Xcellerex™ X-200 bioreactor medium conditioning set points
| Parameter | Set point |
| Temperature | 37°C |
| pH ± deadband set point | 7.2 ± 0.1 |
| DO control | N/A |
| Agitation speed | 52 rpm |
| Sparge 1 air flow | 0.4 slpm |
Table 4. Xcellerex™ X-200 bioreactor culture set points
| Parameter | Set point |
| Temperature | 37°C |
| pH ± deadband set point | 6.95 ± 0.15 |
| pH control | CO2 (sparger 1)/NaHCO3 (1 M) |
| DO set point | 40% |
| DO control | Air and O2 (sparger 1) |
| Sparge 1 air flow (initial) | 0.4 slpm |
| Agitation speed | 103 rpm |
| Agitation direction | Upflow |
| Power input per volume after inoculation | 63 W/m3 |
| Starting volume (after inoculation) | 75 L |
| Target inoculation concentration | 0.9 ± 0.1 × 106 cells/mL |
| Target final concentration | < 6 × 106 cells/mL |
Table 5. Xcellerex™ X-500 bioreactor medium conditioning set points
| Parameter | Set point |
| Temperature | 37°C |
| pH ± deadband set point | 7.2 ± 0.1 |
| DO control | N/A |
| Agitation speed | 57 rpm |
| Sparge 1 air flow | 2.2 slpm |
Table 6. Xcellerex™ X-500 bioreactor culture set points as the N-1 vessel
| Parameter | Set point |
| Temperature | 37°C |
| pH ± deadband set point | 6.95 ± 0.15 |
| pH control | CO2 (sparger 1)/NaHCO3 (1 M) |
| DO set point | 40% |
| DO control | Air and O2 (sparger 1) |
| Sparge 1 air flow (initial) | 1 slpm |
| Agitation speed | 103 rpm |
| Agitation direction | Upflow |
| Power input per volume after inoculation | 63 W/m3 |
| Starting volume (after inoculation) | 350 L |
| Target inoculation concentration | 0.9 ± 0.1 × 106 cells/mL |
| Target final concentration | < 6 × 106 cells/mL |
Xcellerex™ X-platform bioreactor cell culture (50, 500 and 2000 L)
We installed the bioreactor bag into each vessel and performed a pressure decay test. A foam camera was mounted to the bag (X-500 and X-2000 only) while the bag was under pressure to ensure proper connection. DO, pH, VCD and pCO2 sensors were calibrated and installed into the Click-in probe sheath and then autoclaved. After the probes were installed into the bag, we filled with medium and the exhaust control phase was then initiated to heat the filters to 55°C and start the condensate pump. We brought the bioreactor up to a temperature of 37°C and performed a 1-pt pH standardization. The medium was conditioned for a minimum of 24 h in the X-500 and X-2000 bioreactors using the settings detailed in Table 5 and 7. The X-50 bioreactors did not require medium addition and conditioning because the entire inoculated culture was transferred from either the X-500 or X-2000 parent bioreactor. At least 2 h prior to the transfer of inoculum into the X-50, we started temperature control of the jacket (fixed at 37.5°C) so that the vessel could maintain the temperature of the inoculum during transfer. We also started the X-50 exhaust control phase to bring the filter heaters up to the 60°C setpoint.
We inoculated the X-500 bioreactor with 66 L of the cell culture from the X-200 to a starting cell density of 1.08 × 106 cells/mL. The pH setpoint was updated, DO control initiated, sparge 2 gassing started, and agitation speed updated according to the values detailed in Table 9. We then connected a sterile transfer line to one of the spare X-500 sensor ports to the top of the X-50 bioreactor bag and transferred 36 L of inoculum immediately after inoculation of the X-500 bioreactor. Agitation at 80 rpm was started once the volume in the X-50 bioreactor submerged the impeller, and the cell culture was gently mixed during inoculation transfer. After the probes were adequately submerged, pH and DO control were initiated and sparge 2 gassing started. The agitation speed was increased to the settings detailed in Table 8.
We inoculated the X-2000 bioreactor with 281 L of the culture from the X-500 bioreactor to a starting cell density of 1.09 × 106 cells/mL. The pH setpoint was updated, DO control initiated, sparge 2 gassing started and agitation speed updated according to the values detailed in Table 10. We connected a sterile transfer line to one of the spare X-2000 bioreactor sensor ports to the top of the X-50 bioreactor bag and 36 L of inoculum was transferred to the X-50 bioreactor immediately after inoculation of the X-2000 bioreactor. We followed the same transfer process setup during this run and then increased the agitation speed to the setting detailed in Table 8.
We sampled all the bioreactors post-inoculation and each day thereafter for cell counts, viability, pH, pCO2, osmolality, metabolites, and nutrients. Titer was analyzed from day 2 and continued until just prior to harvest on day 14. On day 3 feeding was initiated post sampling by adding 2.5% HyClone™ Cell Boost™ 7a and 0.25% HyClone™ Cell Boost™ 7b of the starting process volume for each bioreactor. The glucose concentration in the bioreactor after the Cell Boost™ feeds was calculated based on the pre-feed glucose measurement and quantity of Cell Boost™ 7a added, which has glucose in the supplement. Glucose was then added to reach the target glucose concentration listed in Table 11, if needed, using a 50% (w/w) solution. Post-feed samples were taken to ensure the target glucose concentration was achieved.
We used Hyclone™ ADCF™ antifoam (~ 3% active simethicone) to control the foam level in the bioreactors via scheduled additions through the pump transfer phase on the system software. We delivered a dose of 0.05 mL/L antifoam to the process starting volume at a starting frequency of once per day after feeding. The start of antifoam additions was based on foam development in each bioreactor and began between days 3 and 5. The frequency was increased to a maximum of three times per day as gas flow and cell density peaked. This frequency was then reduced again as gas flows and cell densities decreased later in the culture. For all bioreactors the total active simethicone concentration was kept below 30 ppm with foam levels well managed.
The final feeds were performed on day 13 and the bioreactors were harvested on day 14 after completing the final samples.
Table 7. Xcellerex™ X-2000 bioreactor medium conditioning set points
| Parameter | Bioreactor set point |
| Temperature | 37°C |
| pH ± deadband set point | 7.2 ± 0.1 |
| DO control | N/A |
| Agitation speed | 37 rpm |
| Sparge 1 air flow | 9 slpm |
Table 8. Xcellerex™ X-50 bioreactor set points
| Parameter | Bioreactor set point |
| Temperature | 37°C |
| pH ± deadband set point | 6.95 ± 0.15 |
| pH control | CO2 (sparger 1)/NaHCO3 (1 M) |
| DO set point | 40% |
| DO control | Air and O2 (sparger 1) |
| Sparge 1 air flow (initial) | 0.1 slpm |
| Sparge 2 air flow (initial) | 0.04 slpm |
| Agitation speed | 167 rpm |
| Agitation direction | Upflow |
| Power input per volume after inoculation | 63 W/m3 |
| Starting volume (after inoculation) | 36 L |
| Target inoculation concentration | 1.0 ± 0.1 × 106 cells/mL |
| Target viability at harvest | 80% |
| Culture duration | 14 d |
Table 9. Xcellerex™ X-500 bioreactor set points
| Parameter | Bioreactor set point |
| Temperature | 37°C |
| pH ± deadband set point | 6.95 ± 0.15 |
| pH control | CO2 (sparger 1)/NaHCO3 (1 M) |
| DO set point | 40% |
| DO control | Air and O2 (sparger 1) |
| Sparge 1 air flow (initial) | 1 slpm |
| Sparge 2 air flow (initial) | 0.4 slpm |
| Agitation speed | 103 rpm |
| Agitation direction | Upflow |
| Power input per volume after inoculation | 63 W/m3 |
| Starting volume (after inoculation) | 360 L |
| Target inoculation concentration | 1.0 ± 0.1 × 106 cells/mL |
| Target viability at harvest | 80% |
| Culture duration | 14 d |
Table 10. Xcellerex™ X-2000 bioreactor set points
| Parameter | Bioreactor set point |
| Temperature | 37°C |
| pH ± deadband set point | 6.95 ± 0.15 |
| pH control | CO2 (sparger 1)/NaHCO3 (1 M) |
| DO set point | 40% |
| DO control | Air and O2 (sparger 1) |
| Sparge 1 air flow (initial) | 4 slpm |
| Sparge 2 air flow (initial) | 1.6 slpm |
| Agitation speed | 70 rpm |
| Agitation direction | Upflow |
| Power input per volume after inoculation | 63 W/m3 |
| Starting volume (after inoculation) | 1440 L |
| Target inoculation concentration | 1.0 ± 0.1 × 106 cells/mL |
| Target viability at harvest | 80% |
| Culture duration | 14 d |
Table 11. Glucose feed criteria
| Process stage | Lactate condition | Target glucose concentration |
| Net lactate production | N/A | 4 g/L |
| Net lactate consumption | > 2.0 g/L | 4 g/L |
| Net lactate consumption | 1 g/L ≤ [Lac] ≤ 2.0 g/L | 5 g/L |
| Net lactate consumption | < 1 g/L | 5–6 g/L |
Results and discussion
All bioreactors achieved similar peak viable cell densities in the range of 43 to 44 × 106 cells/mL with viability exceeding 92% for the duration of the 14 d run. Cell culture for the X-500 bioreactor and its satellite, the X-50 bioreactor, both peaked on day 9, whereas the X-2000 bioreactor and its X-50 bioreactor satellite had cell cultures that peaked on day 10. Both peak days are within the historical range for this process. The complete trend is shown in Figure 4.
Fig 4. VCD and viability for the Xcellerex™ X-50, X-500, and X-2000 bioreactors.
The glucose and lactate trend shown in Figure 5 was similar for all runs with a clear lactate shift occurring by day 7 and net lactate consumption maintained for the culture duration. Glucose and lactate whole broth samples were analyzed using a Nova Biomedical BioProfile FLEX2.
 |
 |
Fig 5. Glucose and lactate profiles for the Xcellerex™ X-50, X-500, and X-2000 bioreactors.
Ammonium, glutamate, and glutamine trend similarly across all runs displaying expected glutamine synthetase activity for this cell line. These graphs are shown in Figure 6. By day 8, glutamate was depleted between each feeding event while ammonium was kept within acceptable levels and glutamine production sufficient for cell metabolism. Ammonium was analyzed using whole broth samples on a Nova Biomedical BioProfile FLEX2, whereas glutamate and glutamine were analyzed using the Roche Cedex Bio Analyzer from a cell-free supernatant sample that was filtered with a 0.22 µm syringe filter.
 |
 |
Fig 6. Ammonium, glutamate, and glutamine for the Xcellerex™ X-50, X-500 and X-2000 bioreactors.
Titer production tracked similarly for all bioreactors with a final concentration in the range of 4.7 to 4.8 g/L. IgG titer was measured on a Roche Cedex Bio Analyzer from a cell-free supernatant sample that was filtered with a 0.22 µm syringe filter and the complete trend is shown in Figure 7. Titer measurements were also confirmed by affinity HPLC (data not shown) using a Waters Alliance HPLC with PDA detector.
Fig 7. mAb titer profiles for the Xcellerex™ X-50, X-500, and X-2000 bioreactors.
Offline dissolved gas measurement was performed with the Nova Biomedical BioProfile FLEX2 on whole broth samples within one min of the sample being taken. The pCO2 trend is shown in Figure 8 and trended similarly for all runs. The profile shown is typical for this process where pCO2 levels are lower earlier in the culture as the pH is controlled at the lower deadband of 6.8 and lactate is accumulating. As the culture progresses and lactate shifts to net consumption, pH trends upwards and is controlled to the upper deadband value of 7.1. At this stage of the culture, CO2 is actively added to maintain pH and the corresponding pCO2 levels equilibrate at a higher level.
Fig 8. pCO2 profile for the Xcellerex™ X-50, X-500, and X-2000 bioreactors.
Osmolality in all runs followed a similar pattern as shown in Figure 9. This indicates that the feeding regime and base utilization were consistent amongst all systems.
Fig 9. Osmolality profile for the Xcellerex™ X-50, X-500, and X-2000 bioreactors.
The pH was controlled well in each bioreactor and followed a similar trend where pH declined on day 1 to the lower deadband value of 6.8 and then increased to the upper deadband value of 7.1 by day 7. The high pH spikes correspond to feeding events, specifically the Cell Boost™ 7b that is highly alkaline.
Fig 10. Online pH trend for the Xcellerex™ X-50, X-500, and X-2000 bioreactors.
DO was similarly controlled during all of the runs. The X-50 DO showed increased noise towards the end of the run as the CO2 mass-flow controller (MFC) was activated intermittently to control pH to the upper deadband. Control loop tuning was not optimized during the run for this phase of the culture but can be adjusted to decrease oscillations. The periodic drops coincide with feeding events where cell metabolism increases and subsequently increases oxygen demand momentarily causing a brief oscillation in the DO reading.
 |
 |
Fig 11. Online DO trend for the Xcellerex™ X-50, X-500, and X-2000 bioreactors.
The X-500 and X-2000 are equipped with a thermal camera that exploits the temperature difference of the foam surface to the process fluid to generate a signal that correlates with foam accumulation. In this study, active control of the foam using the thermal camera was not performed in order to match the same functionality as the X-50 system. However, in Figure 12 the foam PV response is shown over the course of the run. Each cycle corresponds to foam generation within the bioreactor and subsequent elimination after each dose of antifoam.
 |
 |
Fig 12. Foam PV trend for the Xcellerex™ X-500 and X-2000 bioreactors.
Conclusion
Our bioreactor scale-up study showed that a fed-batch CHO process producing an IgG antibody was successfully scaled using Xcellerex™ X-platform bioreactors (X-50, X-500, and X-2000) from 50 to 2000 L. The seed train included shake flasks, a ReadyToProcess™ WAVE 25 bioreactor, and Xcellerex™ X-platform bioreactors and was the same between the two runs with the addition of the X-500 as the N-1 stage for the X-2000 cell culture run. The bioreactor scaling strategy detailed here (constant P/V) combined with the consistent geometry across all scales means that scaling using the X-platform bioreactors is a straightforward process. The results show that X-platform are consistent pilot to production bioreactors. Similar peak VCD, growth profile, cell viability, and titer were achieved across all scales. Metabolite, nutrient, pCO2, pH and DO trends all tracked similarly across each bioreactor confirming the scalability of the platform.
Foam was well controlled in all systems through the use of an automated antifoam addition schedule, which can significantly reduce the amount of antifoam needed. In the X-500 and X-2000 where exhaust filters are located at the ground level, the condensate return pump prevented any build up of condensate in the exhaust lines and returned the fluid to the bioreactor.
CY56867