Multi-cell line kinetic modelling of CHO fed-batch production runs to assess clone variability

This paper presents the Multi Cell-line Kinetic Model (MCKM), a novel generalised kinetic mechanistic model specifically tailored for Ambr15™ fed-batch cultivations of multiple Chinese Hamster Ovary (CHO) cell lines producing different recombinant monoclonal antibodies (mAbs). Unlike traditional models that requires multiple culture runs per regression, the MCKM extracts a complete set of 13 kinetic parameters from just a single culture run per cell line, regressing upon merely 49 data points. The MCKM thus captures inter-clonal heterogeneity, whilst also mathematically describing the lactate switch of CHO cells. By assigning a distinct kinetic profile to each cell line from a single culture run, the MCKM provides deep insights into metabolic variability, enabling data-driven cell line selection (CLS) and improving the efficiency of cell line development (CLD). This is illustrated in a case study that uses linear discriminant analysis to study the metabolic differences in highly productive and unstable cell lines. The MCKM successfully simulated 656 cell culture runs across 157 unique clonal CHO cell lines, studied across multiple passage generations, each recombinant for one of three distinct mAbs, achieving high accuracy in biomass and mAb titer (R2>0.90). The MCKM serves as a versatile tool for in-silico simulation of Ambr15™ production runs across various CHO cell lines and mAbs. When integrated with machine learning, it facilitates cell line kinetic predictions and can also aid in identifying critical process parameters, biomarkers, optimising media and feeding strategies, and reducing experimental workload.