June 15, 2016 (Vol. 36, No. 12)

Caspar Demuth Head of Measuring and Sensor Technology ZHAW Zurich University of Applied Sciences
Iris Poggendorf Biotechnology Lecturer ZHAW Zurich University of Applied Sciences
Ruma Lüthi Student ZHAW Zurich University of Applied Sciences
Marlene Frank Product Manager Hamilton Bonaduz
Kelsey McNeel Market Segment Manager Hamilton Company

Online Viable Cell Density Measurements on Microcarriers

The monitoring of viable cells within a bioreactor is an important step in ensuring an optimal growth process. Quantification of viable cells after added stimuli, for example, can be used to determine the effect of the stimulus on the growth of cells in a process.

Traditional offline cell counting, also called grab sampling, involves extraction from the bioreactor and time-consuming analysis. Extraction of a sample is a tedious procedure and bears the risk of contaminating the entire batch (Figure 1). These manual offline measurements can also be inaccurate.

By the end of this high-risk, time-consuming procedure, the cell count obtained will likely no longer accurately represent the contents of the bioreactor because processing has continued throughout the analysis time. For all these reasons, online determination of viable cell density delivers better results. It not only comes with a much lower risk, but also provides real-time data for more accurate analysis and decisions.


Figure1. Offline sampling: a time-consuming procedure with many manual steps.

Quantification of Cells on Microcarriers

Suspension cells are relatively simple to analyze because the cells are separated from each other and are not anchored. Simple dilution and staining are often sufficient sample preparation for suspension cells. Analysis of cells adhered on microcarriers, however, requires significant effort.

Microcarriers offer the substantial advantage of maximizing the cultivation of anchorage-dependent cells, such as baby hamster kidney (BHK) cells, in a fixed-volume bioreactor. However, offline quantification of these cells requires detachment of the cells from the microcarrier by trypsinization, an enzymatic procedure that affects the cell physiology and viability.

Online measurements of cell count are often based on optical properties such as absorbance. This type of measurement is not compatible with microcarrier systems due to the significant optical density and light-scattering properties of the microcarriers. In order to utilize the significant advantages of both online measurements and microcarrier cell growth, a new type of measurement parameter is required.

The Incyte sensor (Hamilton Company) is based on permittivity readings (capacitance per unit area). In an alternating electrical field viable cells behave like capacitors. The stored charge of the cells is measured by the Incyte sensor. This signal correlates with the viable cell density because dead cells have leaky membranes that cannot store a charge (Figure 2). Microcarriers also do not store charge from an alternating electric field. Below we describe an experiment to demonstrate that this sensor accurately measures viable cell density in a microcarrier cultivation.


Figure 2. The Incyte measurement principle. In an alternating electrical field, viable cells behave like small capacitors. The charge from these small capacitors is measured by the sensor and reported as permittivity (capacitance per area).

Robust Sensing for Process Applications

Aeration and agitation are common in cell growth processes. In order to ensure that aeration and agitation would not interfere with the permittivity measurement, the Incyte sensor was used to measure the growth of BHK cells on microcarriers within the range of typical process parameters in phosphate buffered saline (PBS). Agitation was set to either 100 rpm or 400 rpm, and measurements were performed for two hours.

No significant influence of the agitation was observed. The aeration was varied between 0.5 L/min and 2 L/min for 1 and 2 hours. Again, no significant effect was seen on the measurement of viable cell density. Additionally, an experiment was performed under culture conditions with 3 g/L microcarrier (Cytodex 3) with no cells present. No influence in the permittivity reading was recorded, confirming that the presence of microcarriers does not disturb permittivity readings.

Online Measurement of Cell Viability

A typical cell culture process was initiated for comparison of offline measurements with permittivity-based online measurements. BHK cells were cultivated in a Biostat B 2 L reactor (Sartorius) with a culture volume of 1 L. Cultivation medium was DMEM, F-12 HAM with stable glutamine source, and 10% FBS. Glutamine was known to be the limiting C-source and therefore used as feed.

Culture conditions were set to 37°C, 120 rpm, pH 7.2, and 40% oxygen saturation. Cytodex 3 was the defined microcarrier and was used at 3 g/L. The Incyte sensor was installed for the permittivity measurement, with a measurement frequency of 300 kHz and a background frequency at 10 MHz (used to zero out matrix effects). The offline samples were analyzed with the Nucleo Counter NC100 (ChemoMetec) for viable and total cell counts.

In Figure 3A the permittivity (in pF/cm) determined online is compared to the offline cell counts (viable cells/mL). As aforementioned, offline measurements are prone to errors. This tendency is represented graphically by the significant standard deviation seen in the offline measurements. That being said, the overall trend line of offline measurements (blue line) closely matches the online measurements (red line).

Interestingly, an outlier was found at 54 hours of process time. This outlier shows very low deviation. However, it fits neither the expectation based on all offline measurement, nor the online signal. Such an outlier found in a real process would likely lead to a false interpretation of cell growth, further causing a wrong feed or harvesting decision.

Online and offline data from the duration of this experiment were correlated (blue data in Figure 3B). The R2 value of this data was calculated as 80.6% due to changes in cell heterogeneity in the plateau phase of the growth curve. When only the exponential growth phase, viz. 30–70 hours, is taken into account (red data in Figure 3B), the R2 value was calculated to be 99.4%. It should be noted that the outlier data at 54 hours were not used for these correlations. This strong, almost ideal, correlation between online data and offline data allows the prediction of the viable cell density based on permittivity monitoring for future processes, particularly during the exponential growth phase.

To support the measurements of online and offline cell growth, the concentrations of glutamine, ammonium, glucose, and lactate were analyzed using the Bioprofile 100 Plus (Novo Biochemical). As expected, glutamine and glucose concentrations decreased during the growth phase of the cells (Figure 3C). Lactate and ammonium concentrations increased throughout the entire experiment. This was as expected because they are both metabolic waste products. Glutamine needed to be supplemented twice, at around 55 and 70 hours, to continue cell growth.


Figure 3. (A) Overlay of online (permittivity, pF/cm) and offline measurements (viable cells/mL) of cell viability. (B) Correlation of offline (viable cells/mL) and online data (permittivity, pF/cm). The blue correlation represents all phases of cell growth, whereas the red correlation represents the exponential phase only. (C) Concentration of metabolites over time. Glutamine and glucose decrease; lactate and ammonia increase.

Conclusions

The Incyte sensor enables measurement of the most important process parameter in cell culture, the viable cell density, even in the case of anchorage-dependent cells on microcarriers. By employing a sensor based on permittivity it is possible to analyze viable cell density of mammalian cells on microcarriers in real-time. Online sensing reduces data inaccuracies, workload, sample consumption, and analysis time. This is especially true for cells grown on microcarriers because the Incyte sensor allows cells to remain attached to their anchor. Incyte can be used for reliable understanding of bioprocesses in real time for more advanced process control.

Caspar Demuth is head of measuring and sensor technology department, Iris Poggendorf is lecturer, biotechnology, and Ruma Lüthi is a student at ZHAW Zurich University of Applied Sciences, ICBT Institute of Chemistry and Biotechnology, department of life sciences and facility management. Marlene Frank is product manager, cell density at Hamilton Bonaduz, and Kelsey McNeel ([email protected]) is market segment manager at Hamilton Company.

Previous articleTeva Ends MPC Platform Collaboration with Mesoblast
Next articleNanotechnology Tool Allows Detailed Study of Proteins in Living Cells