Improving reliability of Raman spectroscopy for mAb production by upstream processes during bioprocess development...

Area of Science:

• Biotechnology and bioprocess engineering
• Analytical chemistry
• Spectroscopy

Background:

• Raman spectroscopy offers potential for bioprocess understanding and monitoring.
• Challenges exist in maintaining performance when bioprocess conditions (clone, media, scale) change.
• Existing methods may require extensive data for calibration.

Purpose of the Study:

• To develop strategies for balancing information content from multiple mammalian cell cultivations for PAT monitoring.
• To propose a pragmatic approach for Process Analytical Technology (PAT) monitoring applicable to both development and manufacturing stages.
• To demonstrate the impact of local models on calibration performance for Raman spectroscopy in bioprocess development.

Main Methods:

• Combined risk-assessment techniques with local and general model development for Raman spectroscopy calibrations.
• Investigated the effects of process scale, media composition, and cell line on Raman spectra.
• Utilized thirty-five cultivations across four CHO cell line types, eight clones, and four scales (2L to 10,000L) in fed-batch and perfusion modes.

Main Results:

• Achieved good predictive power from Raman spectroscopy as a PAT tool despite variations in cultivation conditions.
• Demonstrated that local models, specifically Clone-based models, significantly improve calibration performance.
• Showcased that Clone-based local models require fewer batches (3-9) compared to traditional approaches for accurate calibrations.

Conclusions:

• A pragmatic PAT monitoring approach using Raman spectroscopy can be effectively developed in parallel with bioprocess development.
• Local modeling strategies, particularly Clone-based models, enhance the accuracy and efficiency of Raman spectroscopy in bioprocessing.
• This study provides a blueprint for developing robust PAT approaches in diverse bioprocess settings.