Prediction of filamentous process performance attributes by CSL quality assessment using mid-infrared spectroscopy and chemometrics

Abstract Every biopharmaceutical production process aims for control strategies to achieve process robustness in order to ensure consistent product quality. Process variability can origin from process parameters, the biological nature as well as from high lot-to-lot variability of raw materials. In filamentous processes raw materials with very complex matrices, such as corn steep liquor (CSL), are used, which are especially challenging to characterize. In this study, CSL was characterized in detail for its ingredients presenting an overall composition of its matrix of 50 analyzed components (19 amino acids, 5 organic acids, 8 reducing sugars, 7 water-soluble vitamins and 11 trace elements/minerals) in order to facilitate analytical reduction to fingerprinting methods FT-MIR was evaluated as fast and non-destructive spectroscopic fingerprinting method for adequate assessment of CSL quality. Feasibility of this method was shown by the correlation of certain bands in the spectra to substance groups present in CSL, such as the Amide I and II band and amino acids, respectively. Additionally, applicability of FT-MIR could be shown for classification of different CSL lots differing in provider and corn quality as well as for predictability of process performance attributes. The latter was demonstrated on a fed-batch filamentous fungi process for the production of antibiotics. By multivariate data analysis, it could be shown that CSL quality assessment via FT-MIR can be used for the prediction of maximal biomass generated in the process, with a correlation coefficient R 2 of 0.964, as well as for the prediction of an unwanted impurity. The combination of a fast and easy method for CSL quality assessment and correlations of this quality with process performance attributes may facilitate the establishment of a risk-based acceptance criteria for raw material quality release of CSL. As CSL is a frequent used raw material, we believe that this method will also be useful for other processes and that CSL quality assessment is of high relevance in academia and industry.