Potential of Mid-Infrared Spectroscopy for On-Line Monitoring of Mammalian Cell Culture Medium Components

Abstract

This study proposed a methodology to evaluate the potential of midinfrared (MIR) spectroscopy as a process analytical technology (PAT) tool for in situ (in-line) monitoring of cell culture media constituents, paving the way for on-line bioprocess monitoring and control of mammalian cell cultures. The methodology included a limit of detection (LOD) analysis and external influence investigation in addition to the calibration model development. The LOD analysis in the initial step provided a detailed procedure by which to evaluate the monitoring potential of the instrument of choice, for the application in question. The external influence study highlighted the potential difficulties when applying this technique to a typical mammalian cell culture. A comparative investigation between a fixed conduit immersion probe and flexible fiber-optic immersion probe was also carried out. Limitations associated with the use of MIR spectroscopy in the cell culture environment were also examined. A preliminary investigation, on components typically found in mammalian cell cultures, involving spectral characterization and limit of detection analysis was completed. It was evident at this initial stage that glutamine, could not be accurately detected at levels typically found in a mammalian cell culture medium. Results for glucose and ammonia, however, proved promising. A seven-concentration-level experimental design was used, and partial least squares regression employed, to develop calibration models. Optimized model results echoed the results of the preliminary analysis with the percentage error of prediction for glucose as low as 6.03% with the fixed conduit probe and glutamine having a higher error of 63.06% for the same probe. Comparison of the model results obtained from both probes supported the fixed conduit as the more accurate of the two probes for this experimental setup. The effect of external influences on the MIR spectra and hence the concentrations predicted by the model were also examined. These were subjected to statistical analysis to determine the significance of the effect. This study demonstrates that MIR spectroscopy as a PAT tool has limited potential for mammalian cell culture monitoring due to low concentrations of analytes present and outlines a method to allow the system to be evaluated.