We describe a broadly applicable and efficient workflow for finding the optimal feed combination for high-performing fed-batch bioprocessing.
- We cultured a monoclonal antibody (mAb)-producing CHO cell line in HyClone™ CDM4NS0 cell culture medium and screened individual Cell Boost™ feed supplements in subsequent spiked batch and fed-batch experiments.
- Process development was supported by a design of experiment (DoE) approach to reduce the number of cultures required.
The developed fed-batch culture conditions supported mAb titers of approximately 4 g/L in bioreactor cultures.
Introduction
Fed‑batch processes are commonly applied in cell culture production to supply critical nutrients at physiological conditions and drive cell performance towards excellence. The optimal high‑performing basal medium and feed solution and the optimal feed regimen are often combined for a given cell line to boost antibody titers to g/L in straightforward fed‑batch processes. The workflow outlined in Figure 1 is exemplified for a mAb-producing CHO cell line after medium screening and adaptation to HyClone™ CDM4NS0 basal medium (Step 0). After the initial basal medium screening, we used a DoE approach to identify the optimal combination of different supplements that enhanced cell culture performance in spiked batch experiments (Step 1). In this DoE study, levels ranged from no feed addition to a maximum addition to reach a final osmolality of 400 mOsm/kg in the spiked basal medium. We then applied the optimal combination of supplements in a fed‑batch process by daily bolus addition of the feed solution (Step 2).
At this stage, we applied a DoE approach to fine‑tune the feed ratio added to the cell culture. The selected supplements were added to the basal medium to reach 400 to 600 mOsm/kg after 10 simulated feed additions. In the two consecutive DoE experiments at small (30 mL) scale, we identified the combination of supplements and their optimal amounts given as daily bolus feed additions for the investigated fed‑batch process. Optionally, the feeding strategy can be optimized by exploring different constant or dynamic feeding strategies based on cell culture demands (Step 3). This step, however, was excluded from this study.
Finally, we used the optimal supplement combination in a controlled bioreactor run at 500 mL scale, which showed similar or even slightly higher final product titers (Step 4).
Fig 1. Proposed workflow towards a high‑performing fed‑batch process.
Materials and methods
Cell line and media
We used the CHO‑DG44 cell line, expressing an IgG1 antibody directed against TNF‑α, in all experiments. The CHO cells were adapted to HyClone™ CDM4NS0 basal medium (Step 0). Eight different Cell Boost™ supplements (1, 2, 3, 4, 5, 6, 7a, 7b) were selected for supplementation of the chemically defined basal medium to drive culture performance to its peak maximum (Table 1).
The supplements differ in their nutrient composition and concentration. Defining a molar ratio based on total amino acid content let us establish a normalization factor for the individual supplements in the supplement mixtures.
The Cell Boost™ 1 supplement was arbitrarily set as 100%.
Table 1. Molar ratio of amino acids in stock solutions of Cell Boost™ supplements
|
Cell Boost™ supplement |
Stock solution (%) |
Molarity of total amino acids (mM) |
Molar amino acid ratio |
| 1 | 10 | 208.0 | 1.00 |
| 2 | 10 | 351.6 | 0.59 |
| 3 | 5 | 115.6 | 1.80 |
| 4 | 10 | 365.5 | 0.57 |
| 5 | 5 | 131.3 | 1.58 |
| 6 | 5 | 135.6 | 1.53 |
| 7A | 18.1 | 603.0 | 0.34 |
| 7B | 9.5 | 313.3 | 0.03 |
Note! For Cell Boost 7b, 10% of the amount of Cell Boost™ 7a supplement is used.
Step 1. Screening of feed supplements
We determined the optimal combination of Cell Boost™ supplements using a DoE approach (Fig 2). In the MODDE 12 software package (Umetrics AB), all eight supplements were entered as quantitative factors, using low, middle, and high DoE levels specified as ‑1, 0, and +1, respectively.
Using the software’s Design wizard, a balanced subset of the full factorial design at two levels was recommended as first choice, using three center points and a linear model. With this approach, we could define an experimental plan using various combinations of different supplements spiked at different concentrations (i.e., DoE levels) to the basal medium.
The experimental plan included the following controls:
- Basal medium without Cell Boost™ supplements (i.e., DoE level ‑1 in experiment no. 1)
- All supplements spiked at maximum concentrations (i.e., DoE level 1 in experiment no. 16)
- All supplements spiked at half‑maximum concentrations (i.e., DoE level 0 in experiment no. 17 to 19) as triplicate cultures.
Next, we defined the different DoE levels (‑1, 0, and +1). We used the total amino acid concentration of each Cell Boost™ feed supplement as a normalization factor. By mixing all supplements according to their molar ratio of total amino acid content (Table 1), each supplement contributes with the same total amino acid concentration (i.e., 10 mM) when this mix is spiked into the basal medium.
According to Table 1, 1 mL of supplement 1 was mixed with 0.59, 1.8, 0.57, 1.58, 1.53, 0.34, and 0.03 mL of supplements 2, 3, 4, 5, 6, 7a, and 7b, respectively, to prepare an “all supplement feed mix”. We spiked this mix of balanced feeds into the basal medium to reach a maximum osmolality of 400 mOsm/kg to define the maximum DoE level +1.
Fig 2. Experimental design matrix and definition of DoE levels for Cell Boost™ supplement-spiked batch cultures of the mAb‑producing CHO cell line grown in HyClone™ CDM4NS0 medium (Step 1). Each supplement contributes with 10 mM of the total amino acids..
A high osmolality above 400 mOsm/kg is often considered a critical upper limit in spiked media, above which cell performance is detrimentally impacted. The lowest DoE level of ‑1 is defined as no supplement addition, and DoE level 0 is defined as the half‑maximum supplement addition.
Using this approach, we added the “all supplement feed mix” to the basal medium at 0%, 26.89%, and 53.78% for DoE levels ‑1, 0, and +1, respectively, resulting in final osmolalities of 295, 364, and 410 mOsm/kg, respectively. Each feed additionally provided 0, 7.5, or 15 mM of total amino acids to the basal medium at level ‑1, 0, and +1, respectively.
We then used the different supplement-spiked basal media for simple batch cultivations. We used the mAb‑producing cell line at a starting cell concentration of 0.3 × 106 c/mL in 30 mL cultures in the basal medium. This was supplemented with 6 mM L‑glutamine in 50 mL shake tubes. Cultures were maintained at 220 rpm, 37°C, 7% CO2, and 90% relative humidity.
We performed sampling of cell concentration, viability, antibody concentration, and metabolites (i.e., glucose, lactate, glutamine, glutamate, and ammonium) directly after inoculation on day 3, and then every day onwards. We terminated cultures when the viability dropped below 60%.
Step 2. Defining ratio of feed supplements
The optimal combination of Cell Boost™ supplements identified in Step 1 was used in fed‑batch culture using another DoE matrix (Fig 3). In this experimental plan, Cell Boost™ supplement 3 was kept at a constant level of 6.5% of the working volume, as this feed showed good performance when added to the basal medium. The established experimental plan consisted of a total of 16 tubes. Additionally, DoE levels were extended to 1.5 for selected conditions.
As fed‑batch culture was applied in Step 2, we defined the different DoE levels (‑1, 0, +1, and +1.5) using a different approach compared with Step 1. Only the selected Cell Boost™ 1, 2, 3, 7a, and 7b supplements were now mixed together according to their total amino acid ratio.
We observed that the 7b supplement might form precipitates when mixed with other Cell Boost supplements. Hence, the 7b supplement was always added separately to the culture.
Selected supplements were spiked into the basal medium to reach 400, 500, and 600 mOsm/kg for DoE level ‑1, 0 and +1, respectively. We assumed that this would be the final osmolality after 10 theoretical feed days. Therefore, the total amount of supplements we used to reach 400, 500, or 600 mOsm/ kg was divided by 10 to obtain the final daily feed addition for each supplement at a certain DoE level.
Cultures were seeded at 0.3 × 106 c/mL in basal medium supplemented with 6 mM L‑glutamine in 50 mL shaking tubes at a working volume of 25 mL at 220 rpm, 37°C, 7% CO2, and 90% relative humidity.
Starting on day 3, feeds were added once daily according to Figure 3.
Step 4. Process verification in bioreactor runs
The mAb5 cell line in the basal medium was cultured in a 0.5 L DASGIP (Eppendorf) fed‑batch fermentation run to verify the developed Cell Boost™ mix and feeding strategy (Table 2).
Table 2. Different formulations of the optimized feed solution and daily concentration based on the current working volume (WV)
|
Feed solution |
Included supplements |
Bolus feed addition |
Comment
|
| Reference feed | Cell Boost™ 1, 2, 3, 7a, 7b | 8.77% of WV |
Cell Boost™ supplements prepared individually and mixed as liquids |
| Feed mix | Cell Boost™ 1, 2, 3, 7a | 8.70% of WV |
Cell Boost™ 7b added separately at 0.07% of WV |
Fig 3. Experimental design matrix and definition of DoE levels for fed‑batch cultures of the mAb‑producing CHO cell line grown in HyClone™ CDM4NS0 basal medium (Step 2).
The bioreactor was filled with 0.5 L of the basal medium. After equilibration to pH 7.0 and 37°C, viable cells were seeded at 0.3 × 106 c/mL. Dissolved oxygen (DO) was maintained above 30% by adjusting air flow and oxygen concentration and the suspension was agitated constantly at 80 rpm. Starting on day 3, the culture was fed once daily with a constant 8.77% feed amount using different feed solutions prepared from powder or liquid Cell Boost™ formulations.
For Feed mix 2, we added the 7b supplement separately into the bioreactor, using an additional feed tubing line. We measured critical metabolites (glucose, lactate, L‑glutamine, glutamate, and ammonia) daily. Cultures were harvested once the viability reached 70%.
Results and discussion
Screening of Cell Boost™ supplements in spiked batches (Step 1)
Using a DoE approach, different combinations of supplements were spiked in the basal medium in batch cultures. For the optimal‑performing combinations, peak cell concentrations could be increased from 8 × 106 c/mL of the unspiked control up to 15 × 106 c/mL (Fig 4). The integral of viable cells increased from 23 × 106 c × d/mL to more than 37 × 106 c × d/mL, resulting in a 2.6‑fold increase in final mAb titer from 1.2 g/L up to 2.5 g/L on day 9.
In‑depth analysis of the DoE experiment by regression analysis and model generation indicated a combination of Cell Boost™ 1, 2, 3, 7a, and 7b supplements to be optimal to enhance batch performance of the investigated cell line grown in the HyClone™ basal medium. We observed that Cell Boost™ 3 has a beneficial effect on cell growth and productivity, whereas Cell Boost™ 5 and 6 supplements showed low performance and we therefore excluded them from further studies.
Optimization of Cell Boost™ supplement ratios in fed-batch cultures (Step 2)
The selected combination of Cell Boost™ 1, 2, 3, 7a, and 7b supplements from Step 1 was further applied in a fed‑batch process (Fig 5). Using a DoE approach, the supplement combination was fed to the cultures to find the optimal daily feed ratio of each Cell Boost™ solution. Peak cell concentrations ranged from 8 × 106 to 24 × 106 c/mL, reaching final titers of up to 3.5 g/L.
The broad spectrum of experimental conditions lets us empirically determine the optimal ratios of supplement combinations. However, as the experiment was designed according to a DoE approach, models can also be established using the MODDE software, allowing calculation of optimized conditions that cannot be identified empirically. As a final output, the DoE approach suggested a daily optimal feed ratio of 1, 2, 3, 7a, and 7b supplements at 1.59%, 2.37%, 3.44%, 1.30%, and 0.07%, respectively, of the working volume of the fed-batch process. To lower osmolality of the final feed solution, the 7b supplement can be excluded from the feed combination.
Fig 4. Cell growth and productivity for Step 1. (A) Total cell concentrations (CC), (B) cell viability, (C) viable cumulative cell days (VCCD), and (D) mAb titer. CB = Cell Boost™ media.
Fig 5. Cell growth and productivity for Step 2. (A) Total cell concentrations (CC), (B) cell viability, and (C) mAb titer.
Process verification in bioreactor cultures (Step 4)
Figure 6 summarizes cell culture data for the two DoE screening steps (Steps 1 and 2) and the final controlled bioreactor run (Step 4), in which the optimized feed composition and daily constant 8.77% feed regimen were verified. We prepared the same supplement combination as different formulations and fed to the cultures according to Figure 7A. Feed solution was prepared from premixed powder formulations (Feed mix).
We recommended that the 7b supplement is added separately to the culture, as this supplement can form extensive precipitates when mixed with other Cell Boost™ supplements. As a reference, we prepared the selected Cell Boost™ supplements individually and added them separately (Reference feed).
Comparable peak cell concentrations of 23 × 106 c/mL were reached in both the reference culture and in the Feed mix culture to which we added the 7b supplement separately (Fig 7B). Similarly, maximum peak titers of 4 g/L were reached under both tested conditions after 10 d (Fig 7C). The nutrient composition was the same for both tested feed solutions.
For both the Reference feed and Feed mix, the osmolality remained relatively constant at 300 mOsm/kg over the entire fed‑batch process (Fig 7D).
The glycan distribution was comparable between the two runs (Fig 7E).
Fig 6. Summary of cell culture data for two DoE screening steps (Steps 1 and 2) and final bioreactor run (Step 4).
Fig 7. (A) Used feed solutions, (B) cell densities and viabilities, (C) mAb titers, (D) osmolalities of the bioreactor cultures, and (E) relative abundance of critical metabolites including glycans. Feed supplements were added at constant 8.77% daily feed regimen.
Conclusion
This work demonstrates a suitable workflow to establish a high‑performing fed‑batch process that is also applicable to other cell culture media and feed systems.
- Based on two consecutive DoE studies in batch and fed‑batch mode, we identified a combination of Cell Boost™ feed supplements for maximum culture performance.
- For the investigated mAb‑producing CHO cell line grown in HyClone™ CDM4NS0 medium, a constant 8.77% feed regimen, we found that using a feed mix containing Cell Boost™ 1, 2, 3, 7a, and 7b supplements (at 1.59%, 2.37%, 3.44%, 1.30%, and 0.07%, respectively) drove antibody production beyond 4 g/L in a fed‑batch process.
- Using the described workflow, we rapidly established a fed‑batch process to boost mAb production several‑fold from unspiked batch culture, thereby simplifying optimization and increasing overall yield.
Key takeaways from the article
- What was the goal of the study?
To develop a high-performing fed-batch workflow that identifies optimal feed combinations and feeding strategies for a mAb-producing CHO cell line.
- What cell line and medium were used?
A CHO-DG44 cell line expressing monoclonal antibody was cultured in HyClone™ CDM4NS0 basal medium.
- How were feed supplements selected?
Eight Cell Boost™ supplements (1, 2, 3, 4, 5, 6, 7a, 7b) were screened using a design-of-experiments (DoE) approach in spiked fed-batch cultures.
- Which supplements were found to be optimal?
Cell Boost™ 1, 2, 3, 7a, and 7b delivered the strongest improvements in cell performance and antibody titers.
- What feeding strategy achieved the best results?
A daily bolus addition using an optimized ratio of the five selected supplements, totaling a constant 8.77% of feed per day.
- What final titers were achieved?
The optimized process produced up to 4 g/L of monoclonal antibody in bioreactor runs.
- Why was osmolality important?
Feed additions were calibrated to maintain osmolality between 400 to 600 mOsm/kg in the model system, as excessive osmolality can negatively impact CHO cell performance.
- Can the workflow apply to other systems?
This DoE-based method, with modification, could be broadly applicable to other CHO cell lines, media systems, and feed formulations.
CY15106