Dynamic Simulation and Optimisation of a Fed-Batch Bioreactor for mAbs Under Varying Substrate and Feed Flow Conditions

A previously published, experimentally validated fed-batch reactor model for production of monoclonal antibodies (mAbs) via a fermentation process (which includes substrate consumption by cells, resulting in biomass growth and product formation) is employed here (the model is based on nonlinear ODEs from first principles and parameter estimation for a mammalian cell culture process, on the basis of experimental data). Dynamic simulation is first performed to visualise responses. Then, a dynamic optimisation problem is formulated, aiming to maximise the product (antibody) yield. Dynamic optimisation is then performed via the dynamic model by solving the nonlinear programming (NLP) using the interior point optimisation (IPOPT) solver in MATLAB and Python, and the AP Monitor libraries. Feed rate manipulations of glucose and glutamine (two key substrates fed to the bioreactor) are carried out to compute optimal control trajectories, and relevant comparisons are presented. Maximum mAb titers achieved a range between 0.09 and 0.31 μM, depending on parameter values.