Analyze cell growth in bioreactors

Raman spectroscopy has a wide range of applications as it can detect subtle changes in intricate aqueous systems, including the analysis of biopharmaceutical processes such as cell growth in bioreactors.

The Raman spectrometer can monitor biopharmaceutical production processes in real-time, in situ, and non-destructively when it is used as a continuous process analyzer of these intricate chemical systems. Raman spectroscopy has become a powerful process analytical tool due to its capacity to identify changes in a wide variety of metabolites during bioreactor processes.

MarqMetrix^® Bioreactor BallProbe^® with TouchRaman^® immersion technology

Reusable optical probes for Raman spectroscopy analysis provide advantages such as increased process repeatability and reliability by lowering run-to-run variability. There are many different probes that can be used with the Thermo Fisher™ MarqMetrix™ All-In-One Process Raman Analyzer

The MarqMetrix™ All-In-One Process Raman Analyzer can be used with the MarqMetrix Bioreactor BallProbe, which is made to meet the needs of the bioprocess industry. These probes can handle sterility procedures such as offline autoclaving, are quick and simple to swap out, and are also easy to connect.

Thermo Scientific™ HyPerforma™ DynaDrive™ Single-Use Bioreactor (S.U.B.), for perfusion cell culture applications

The most recent development in SUB technology, the DynaDrive Single-Use Bioreactor (SUB), provides enhanced performance and scalability for large-volume bioproduction. The cuboid-shaped tank has several significant advantages over earlier S.U.B. designs, including better mass transfer and mixing capabilities and increased scalability.

This application note explains how to integrate the MarqMetrix™ All-In-One Process Raman Analyzer system with the 500L HyPerforma DynaDrive bioreactor to carry out in-line measurements of crucial process parameters (CPPs).

This integrated system enabled the development of precise prediction models for many parameters and metabolites using continuously generated spectral data throughout a cell growth culture run.

Materials and methods

Cell culture and feeding strategy

Cell culture was carried out in a 500 L HyPerforma DynaDrive S.U.B. with a working volume of roughly 320 L of cell culture medium and 0.5*106 cells/mL inoculated at 36.5 °C, pH = 6.9+/–0.3, and 50% DO). CO₂ gassing and sodium carbonate additions, as necessary, were used to regulate the pH level.

Starting on day three, a two-step feeding procedure was used every day to feed the cells as they grew in a chemically defined medium. Using the initial volume as a starting point, the first feed media was added at 4% by weight, and the second feed media at 0.4%.

On day six, the temperature changed to 33 °C. After 14 days, the run was over. To block stray light, the bioreactor was covered. For in-line, real-time spectral Raman data generation following autoclaving, the MarqMetrix™ All-In-One Process Raman Analyzer Bioreactor optical ball probe was inserted into the HyPerforma DynaDrive S.U.B.

Figure 1. 500 L Thermo Scientific HyPerforma DynaDrive S.U.B. for cell culture applications. Image Credit: Thermo Fisher Scientific-Handheld Elemental & Radiation Detection

MarqMetrix™ All-In-One Process Raman Analyzer measurements

The optical bioreactor ball probe of the MarqMetrix™ All-In-One Process Raman Analyzer was directly submerged in the bioreactors (500 L) D when measurements were made using the MarqMetrix™ All-In-One Process Raman Analyzer system. Each Raman spectrum was produced using an average of 20 measurements, a 3 second integration/exposure time, and a 450-mW laser.

The MarqMetrix™ All-In-One Process Raman Analyzer and offline instrument analysis used a timestamp match to determine the total acquisition time per data spectra, which was two minutes.

Chemometrics, model building

Models were built using independent data from various MarqMetrix™ All-In-One Process Raman Analyzer instruments, probes, and bioreactor types. Each chemometric model was developed using 45 samples per bioreactor from the training datasets. After reviewing the spectral data, cosmic ray-related outlier spectral spikes were eliminated.

Figure 2. Thermo Fisher™ MarqMetrix™ All-In-One Process Raman Analyzer. Image Credit: Thermo Fisher Scientific-Handheld Elemental & Radiation Detection

The spectra were pre-processed to remove the baseline and increase signal-to-noise after the spectral region of interest was chosen. The Savitzky Golay filter with derivatives, Automatic Whitaker Smoothing, Extended Multiplicative Scatter Correction, SNV, and mean centering were just a few of the pre-processing methods that were tested.

The most effective pre-processing methods varied depending on the particular parameter of interest that was modeled. For each property of interest, partial least squares (PLS) models were made, and cross-validation was used to assess how well each model had been optimized.

During the bioreactor culture run, common metabolites such as glucose, lactate, glutamine, glutamate, TCD, and VCD were produced.

Results

In this study, a fed-batch CHO cell culture process was subjected to continuous in-line Raman spectroscopy. The offline analytical data collected for the relevant parameters and the in-line spectral data were correlated.

Before using Raman spectroscopy to monitor process parameters, chemometric models must be built using a set of externally calibrated data (independent offline data). Bioreactor samples were taken every day and analyzed for comparison to judge the reliability of the MarqMetrix™ All-In-One Process Raman Analyzer predicted values.

For each parameter, the root mean square error of calibration (RMSEC), root mean square error of cross validation (RMSECV), and root mean square error of prediction (RMSEP) were calculated. To determine the RMSECV, which is used to build the model, the error was averaged based on the model’s prediction.

The model is tested using the RMSEP against “new” data that it has never seen before. Each PLS model’s R² coefficient of variation was noted. The value is used to calculate the percentage of the Y variable’s variation that the model predictors (X variables) can account for.

It is significant to note that more precise and reliable predictive chemometric models were created when several sizable, independent data sets from bioreactor runs of the same CHO culture process were combined. Five distinct datasets from earlier bioreactor runs were combined for this study to train a sizable chemometric model.

The spectral data collected during this HyPerforma DynaDrive S.U.B bioreactor run was then subjected to the calibration model. The data show that the model was able to predict this new dataset accurately, and those model predictions were highly correlated with data measurements obtained offline for many metabolites, as shown in Table 1.

Table 1. Correlation of model prediction with offline data analysis. Source: Thermo Fisher Scientific-Handheld Elemental & Radiation Detection

Figure 3. Thermo Scientific HyPerforma DynaDrive S.U.B. Chemometric Model Plots- Comparison of Raman Model vs Offline Analytical Data for Important Bioreactor Parameters. Image Credit: Thermo Fisher Scientific-Handheld Elemental & Radiation Detection

Conclusions

Utilizing the reusable MarqMetrix Bioreactor BallProbe with TouchRaman immersion technology, the MarqMetrix™ All-In-One Process Raman Analyzer provides precise in-line, real-time measurements of the vital process parameters glucose, glutamine, and lactate as well as total and viable cell densities in the 500 L HyPerforma DynaDrive S.U.B.

The robustness of the model, when applied to the parameters in Table 1, is demonstrated by the correlation analysis, which demonstrates excellent agreement between the model prediction data and the offline analytical data.

About Thermo Fisher Scientific – Handheld Elemental & Radiation Detection

Founded more than 25 years ago, we are the pioneer in portable x-ray fluorescence (XRF). With more than 35,000 instruments installed worldwide, we maintain offices in Munich, Germany, and Hong Kong, in addition to our headquarters outside Boston, Massachusetts, USA.

A culture of innovation and a distinguished history of breakthrough achievements have defined our Thermo Scientific portable XRF analyzers since we introduced the first handheld XRF instrument in 1994. Many “firsts” have followed, including the:

• First use of miniaturized x-ray tubes in one-piece handheld analyzers
• First and only isotope-based handheld analyzer that never requires source replacement
• First handheld analyzer equipped with an He purge for direct analysis of Mg, Al, Si, P, and S in metal alloys
• First handheld small-spot XRF analyzer
• First handheld analyzer to feature a 50kV x-ray tube
• First geometrically optimized large area drift detector

The company has been awarded numerous patents and has received many honors and awards, including an unprecedented three R&D 100 Awards…in 1995 for the Thermo Scientific Niton XL-300, in 2003 for the Niton® XLi and XLt Series, and, most recently, in 2008 for the Niton XL3t Series.

• Handheld and mobile Thermo Scientific portable XRF analyzers, from our value-leading Niton XL2 instrument to our Niton FXL field x-ray lab, are engineered for accurate and reliable on-site elemental analysis and ease of use. These nondestructive analyzers are used for a variety of applications, including:
• Identification and analysis of metal alloys, including precious metal & jewelry analysis
• Mining and exploration
• Testing for lead, cadmium, and other hazardous substances in toys, jewelry, apparel, and other consumer goods
• Compliance testing for RoHS, WEEE, and ELV regulations
• Environmental analysis – site assessment, monitoring, and clearance testing
• Lead-paint inspection

We are committed to making the highest performance, easiest to use, most economical portable XRF analyzer in the world…giving you the tools that help you work more productively than ever before…and enabling you to make the world healthier, cleaner, and safer.

Sponsored Content Policy: News-Medical.net publishes articles and related content that may be derived from sources where we have existing commercial relationships, provided such content adds value to the core editorial ethos of News-Medical.Net which is to educate and inform site visitors interested in medical research, science, medical devices and treatments.