Automated Bioreactor Sampling with the Bioprocess Autosampler from Eppendorf

In cell culture processes, optimal growth and high productivity are crucial. Monitoring process parameters and analyzing samples frequently are key to achieving this, yet manual sampling during working hours can miss critical events occurring outside these times. The Bioprocess Autosampler from Eppendorf automates bioreactor sampling, allowing scientists to focus on other tasks and allowing extended analysis beyond business hours without additional staffing costs. It also decreases the risks of user error and variations in sampling techniques from person to person, providing extended and consistent data throughout the process.

In this study we used the Bioprocess Autosampler for automated sampling at various timepoints from four cell culture bioreactors that were operated in parallel.

Key findings:

• Automated bioreactor sampling enhanced productivity by allowing scientists to focus on other tasks while ensuring consistent data collection.
• Automated bioreactor sampling facilitated extended analysis beyond business hours.
• Cell culture metabolites and monoclonal antibodies produced remained stable while being stored in the Bioprocess Autosampler Peltier-cooled sample storage at 4 °C over the weekend.

The Bioprocess Autosampler takes samples from multiple bioreactors and stores them tempered for later analysis. It is compatible with glass and single-use bioreactors operated with a DASbox^® Mini Bioreactor System , DASGIP^® Parallel Bioreactor System , SciVario^® twin or BioFlo^® 320 bioreactor control system .

• Enables regular and process-triggered automated sampling and bolus addition 24/7
• Saves space, because no sterile hood is necessary for operation
• Keeps you flexible, because it is compatible with differently sized single-use and glass bioreactors
• Suitable for cell culture and microbial applications

The Bioprocess Autosampler is fully integrated in and controlled by the DASware^® control 6 software , providing simplicity and ease of use.
For this study, the Bioprocess Autosampler was equipped with 4 tools. Each tool holds a 5 mL syringe containing a 23 gauge conical needle. The tools are equipped with a needle guide containing a sterile filter and tubing. This provides a sterile air flow around the needle tip, creating additional protection from contamination. This sterile air flow is automatically activated and controlled by the robot head movements. A multiport module installed on the Bioprocess Autosampler allows for automated sampling and bolus addition to take place (Figure 2). Each multiport module reserves one sample port and one bolus addition port per vessel. These ports contain a septum that provides a sterile boundary to the tubing connected to the bioreactor.
The Bioprocess Autosampler was used in combination with DASGIP Spinner Vessels with pitched-blade impeller (working volume 0.2 to 1 L) that were controlled by the DASGIP Parallel Bioreactor System .

Bioreactor sample storage
A temperature-controlled Peltier-cooled sample storage enables storage of the samples taken, either in 1.5 mL or 10 mL sterilized glass vials. For this study, we used 1.5 mL glass vials to store our samples. The Peltier-cooled sample storage has 3 drawers that holds 2 racks each (6 racks total). A total of 54 samples can be stored in each rack when using the 1.5 mL sized vials. The temperature range for the Peltier-cooled sample storage is 4 °C to 40 °C. For this study, we chose to store our samples at 4 °C.

Bioreactor sampling strategy and scheduling
The Bioprocess Autosampler takes samples in an aseptic operation using a combination of thorough washing cycles to clean the syringe tools as well as dispensing a 70 % ethanol overlay on top of the septum in the sampling port prior to sampling. This layer of ethanol on top of the septum port provides additional protection against contamination.
The Bioprocess Autosampler offers two methods to take a sample: serial and clustered sampling. The serial sampling method uses the same syringe tool to sample each bioreactor (for microbial applications). The clustered sampling method dedicates one syringe tool per bioreactor, significantly lowering the risks of cross contamination between vessels (for cell culture applications and contamination-sensitive microbial applications). The clustered sampling method is the standard method for cell culture use.

For this study, the Bioprocess Autosampler was used to take samples from four CHO cell culture bioreactors run in batch mode using the clustered sampling method. All vessels were run in parallel for 8 days. The clustered sampling method cleans all tools ahead of the samples being taken which allows for close sampling times from vessel to vessel.
We scheduled the Bioprocess Autosampler to take samples every 12 hours (twice daily) and end sampling after a total of 16 samples per vessel were taken. Our sampling schedule is shown in Figure 4.

An additional washing step (initial wash) is scheduled ahead of the sampling workflow when using the clustered sampling method. This ensures that each tool is thoroughly cleaned prior to executing the sampling workflow. Washing steps are indicated in blue on the sampling scheduler and sampling steps are shown in green (Figure 4). An additional washing step can be added at the end of the sampling schedule, which ensures that the syringe is fully cleaned at the end of the experiment and ready for the next process to start.
Once the workflow has been started and each vessel has been inoculated, the initial wash and sampling is started. To collect the highest quality sample from the growing culture, the Bioprocess Autosampler takes a dead volume from the sampling line and discards it before taking the actual sample. This dead volume can be adjusted to up to 5 mL. In this study, we used a dead volume of 5 mL. The volume of our taken and stored sample was 1.5 mL.

Thawing cells and inoculum preparation

For this study, we used a proprietary suspension CHO cell line from TPG Biologics Inc. that produces the human monoclonal antibody (hMAb) Immunglobulin G (IgG).We cultivated the cells in Dynamis AGT Medium (Thermo Fisher Scientific^® ) supplemented with 8 mM GlutaMAX, 1 % antibiotic antimycotic solution, and 1 % Anti-Clumping Agent (Thermo Fisher Scientific) for a complete medium [1].We prepared our cell culture inoculum by cultivating the cells in single-use, baffled bottom shake flasks (VWR) at a 20 % max fill volume. We first thawed the cells from a previously cryopreserved vial using a ThawSTAR^® CFT2 Automated Thawing System (BioLife Solutions) and inoculated a 125 mL flask at a seeding density of 0.3 × 10⁶ cells/mL. After that, we incubated the flask in the New Brunswick S41i CO₂ Incubator Shaker at 37 °C. Shaking speed was set to 125 rpm and CO₂ concentration to 8 %.

Routine passaging and flask culture scale-up
We allowed the freshly seeded cells to grow for a few passages to acclimate them to the new environment prior to scale-up. The CHO cell line’s optimal passage schedule was every other day. After monitoring cell growth and viability, we then started to increase the culture volume from 125 mL to 250 mL, and finally, to 1 L shake flasks. During these scale-up steps, the flask inoculation density, percentage fill and other parameters remained the same [2]. Our target inoculation density for all bioreactors in this study was 0.3 × 10⁶ cells/mL. The shake flask culture was transferred into a 1 L addition bottle that was prepared and sterilized by a steam sterilizer. This addition bottle was then used to inoculate all bioreactors via tube welding (Terumo BCT).

Bolus addition using the Bioprocess Autosampler
The Bioprocess Autosampler was designed to automate any bolus addition that might be required during a process. With the right scripting, it can be scheduled in accordance with the user’s workflow, based on a specific time or process parameter changes to trigger the bolus addition [2]. However, also real time bolus addition is possible when needed by simply pressing the “add bolus” button under the sampling tab in the DASware control software. In this study, we used the real time bolus addition to add Antifoam Emulsion C (Merck) to our vessels when needed. Once the “add bolus” button is pressed, the bolus addition is scheduled and will start as soon as the Bioprocess Autosampler is free to do so. For our bolus addition, we chose to add 5 mL of Antifoam Emulsion C to all our vessels when needed.
The estimated dead volume of the bolus addition line going into the vessel was 5 mL, which potentially prevents the full bolus volume from reaching the culture. Therefore, after the bolus is added the Bioprocess Autosampler features the function of flushing the tubing with sterile air in order to push the remaining dead volume out of the tubing and into the bioreactor, thus ensuring consistency with each bolus addition step. This way, the more precise automatic bolus addition saves on costly additives, like activating or growth factors. Lastly, the scripting opportunity enables fully automated and situation-specific treatment of the cells, minimizing the need for manual work or being present in the lab at non-working hours.

Automated sampling and analytics
Two samples were taken automatically from each bioreactor daily by the Bioprocess Autosampler, one in the morning and one in the evening. During the week, the samples were stored in the Peltier-cooled sample storage at 4 °C for 24 hours before being analyzed. Samples that were taken over the weekend were stored in the Peltier-cooled sample storage at 4 °C for 72 hours before being analyzed. The cell density, viability, metabolites, biproducts (NH₃) and hMAb concentrations were measured.
We measured cell density and viability (via the trypan blue exclusion method) using a Vi-CELL^® XR Viability Analyzer (Beckman Coulter). We confirmed our pH values offline using an Orion Star 8211 pH-meter (Thermo Fisher Scientific). If necessary, we used the offline pH measurement to re-standardize the pH calibration on the controller, preventing any discrepancies between online and offline measurements. We measured glucose, ammonia, lactate, and hMAb IgG using a CEDEX^® Bio Analyzer (Roche) [1].

Metabolic profile of the growing cells
The metabolic profile of the cell culture from all four bioreactors is shown in Figure 5 as the mean +/- standard deviation. Cells were actively growing and depleted initially supplied glucose by day 6. Lactate levels remained under 2 g/L for the duration of the run. Ammonia increased every day to a toxic level of 10.9 mmol/L by day 7 (Figure 5).

Cell growth comparisons
The cell growth curves and viability for all four vessels are shown in Figure 6. All vessels peaked around 13 × 10⁶ cells/mL between day 5 and 6. Viability remained high (90 % or above) until day 5 (Figure 6) before ammonia reached toxic levels (Figure 5). All vessels showed similar IgG yields by the end of the runs. Each culture produced a maximum of around 190 mg/L of IgG by day 6 (Figure 7).

Metabolite and product stability study
To determine whether sample storage temperature would influence metabolite readings and the stability of the CHO cell line products, a metabolite and product stability study was done by storing samples in the Peltier-cooled sample storage of the Bioprocess Autosampler at 4 °C. For that, the same CHO cell line was grown in a total of 3 × 1 L baffled bottom Erlenmeyer shake flasks in a New Brunswick S41i CO₂ Incubator Shaker for 7 days as a batch culture. The reason for the use of Erlenmeyer shake flasks was to demonstrate an easy-to-perfom test for users to evaluate if the Bioprocess Autosampler fits their needs. With that simple procedure the customer can determine prior to purchasing the device whether their cell line or cell products are suitable for storage at 4 °C for maximum 72 hours, for example by storing in a refrigerator.Each 1 L flask was inoculated at 0.3 × 10⁶ cells/mL. We took a 15 mL sample from each flask every day. The 15 mL sample was then split into three 15 mL conical tubes to a final sample volume of 5 mL per tube. These conical tubes were then used to measure the following timepoints: fresh sample, or samples stored in the Bioprocess Autosampler’s Peltier-cooled sample storage at 4 °C for either 24 hours or 72 hours. All fresh samples were frozen immediately at -80 °C, while the others were frozen after the Peltier-cooled sample storage storage period.
The metabolites glucose, ammonia, and lactate, as well as the produced human monoclonal antibody (hMAb) . Immunoglobulin G (IgG) were measured. Metabolites and antibodies were analyzed at the same time by thawing the previously frozen samples and then measuring by using the CEDEX Bio Analyzer. This resulted in three data points per timepoint. Figures 8-11 show the average metabolite or antibody concentration +/- standard deviation. All metabolites and antibodies analyzed resulted in stable measurements, providing confidence in sample storage.

In this study, we used the Eppendorf Bioprocess Autosampler for automatic sampling of bioreactors, thus simplifying the cell culture process as the lab technician does not need to be present in the lab to take manual samples. A total of four bioreactors were run in a batch culture mode and controlled by the DASGIP Parallel Bioreactor System. Samples were taken on the weekends and outside of business hours, then stored at 4 °C for later analysis. This provides more flexibility during the process and avoids gaps in sampling at critical timepoints.

We demonstrated how the DASGIP Parallel Bioreactor System in combination with the Eppendorf Bioprocess Autosampler can increase productivity and efficiency, eliminating the need for manual sampling altogether. In this study, we were able to achieve a peak cell density of 13 × 10⁶ cells/mL and peak IgG production of 192 mg/L by day 6 of culture. The cell culture metabolites and monoclonal antibodies that were produced in this study remained stable during 24 and 72 hour-storage in the Bioprocess Autosampler Peltier-cooled sample storage at 4 °C. Furthermore, the Bioprocess Autosampler’s seamless integration into DASware control 6 software allows for ease of use.

Taken together, the automated sampling and bolus additions capabilities greatly increase productivity by reducing the manual workload, freeing up time to spend on other tasks while still expanding critical insights in the process.

[1] Suttle A. and Sha M. (2020). Next Generation Bioprocess Controller: SciVario twin. The flexible controller for all your bioprocess needs. Eppendorf Application Note 432.
[2] Bioprocess Autosampler Operating Manual (2024). Eppendorf, Inc.
[3] DASware Control 6 Operating Manual (2024). Eppendorf, Inc.

Eppendorf SE reserves the right to modify its products and services at any time. This application note is subject to change without notice. Although prepared to ensure accuracy, Eppendorf SE assumes no liability for errors, or for any damages resulting from the application or use of this information. Viewing the application note alone cannot as such provide for or replace reading and respecting the current version of the operating manual.