Recent advances in culture medium design for enhanced production of monoclonal antibodies in CHO cells: A comparative study of machine learning and systems biology approaches

Abstract

The production of monoclonal antibodies (mAbs) using Chinese Hamster Ovary (CHO) cells has revolutionized the treatment of numerous diseases, solidifying their position as a cornerstone of the biopharmaceutical industry. However, achieving maximum mAb production while upholding strict product quality standards remains a significant hurdle. Optimizing cell culture media emerges as a critical factor in this endeavor, requiring a nuanced understanding of the complex interplay of nutrients, growth factors, and other components that profoundly influence cellular growth, productivity, and product quality. Significant strides have been made in media optimization, including techniques such as media blending, one factor at a time, and statistical design of experiments approaches. The present review provides a comprehensive analysis of the recent advancements in culture media design strategies, focusing on the comparative application of systems biology (SB) and machine learning (ML) approaches. The applications of SB and ML in optimizing CHO cell culture medium and successful examples of their use are summarized. Finally, we highlight the immense potential of integrating SB and ML, emphasizing the development of hybrid models that leverage the strengths of both approaches for robust, efficient, and scalable optimization of mAb production in CHO cells. This review provides a roadmap for researchers and industry professionals to navigate the complex landscape of mAb production optimization, paving the way for developing next-generation CHO cell culture media that drive significant improvements in yield and productivity.

Keywords: Chinese hamster ovary cells; Genome-scale modeling; Hybrid modeling; Machine learning; Medium optimization; Monoclonal antibody production.

Figure 1.

Overview of Approaches to Media Optimization and their limitations. Basic methods for optimizing media include media blending and OFAT. While media blending combines components in a trial-and-error approach, OFAT focuses on adjusting individual factors one at a time. DoE methods offer a systematic approaches to explore and optimize media components by evaluating individual characteristics and their interactions. Advanced techniques like SB and ML offer great potential for optimizing cell culture media by providing data-driven insights into complex biological systems.

Figure 2.

Timeline of published GEMs for CHO cell lines.

Figure 3:

Types of CBM and ML Integration Methods. (A) ML predicts parameters using fluxomics data generated from simulations of metabolic models. (B) ML predicts parameters by integrating fluxomics with multi-omics data from high-throughput analytics. (C) ML can use multi-omics data to predict and improve metabolic models and fluxomics data.