(235h) Towards Multivariable Optimization and Control for a Perfusion Bioreactor System in Tissue Engineering

In recent years, the field of tissue engineering (TE) has provided a sustainable improvement towards life quality as well as a decrease in both the social and economic cost of healthcare and life expectancy. TE is focused on growing cells that can be used to regenerate or repair tissue or organ function. For cell cultivation, bioreactor systems need to have tissue engineered grafts which are viable, have uniform cell distribution and growth. The use of bioreactor systems over static cultivation has brought great improvements regarding tissue quality. For this, the proper conditions for cultivations have to be given in such a way that they mimic the in vivo environment (Schmid, Schwarz et al. 2018). The use of a perfusion bioreactor will provide an even distribution of the cells on stable scaffolds and facilitates an optimal feed of nutrients. Moreover it can successfully remove toxic metabolites found in the cell culture (Coletti, Macchietto et al. 2006). In comparison with conventional processes, perfusion strategies could result in ten to a hundred times higher product yields.

This work is focused on setting the foundation for the development of advanced control strategies that will facilitate manufacturing for any type of cell culture using a continuous perfusion bioreactor. The developed control strategies are capable of optimizing the cell density production while minimizing the consumption of oxygen and glucose and fulfilling the constraints. Moreover, this will facilitate reduced waste metabolites production and lead to an enhanced productivity.

So far, the common practice in the literature is to estimate offline the optimal feeding using high fidelity mathematical models without closing the loop. The optimal feeding profile is determined considering a base calculation on the cells need for nutrients (Kiparissides, Koutinas et al. 2011). The goal of this work and in general from a process engineering perspective, the objective of any modelling attempt is to close the loop. Hitherto there have only been several studies in literature presenting the potential of using model-based control and optimization for perfusion bioreactors (Nascu, Sebastia-Saez et al. 2021), there is still a lot of work to be done in developing and implementing these control strategies.

For this work a comprehensive mathematical model of convection and diffusion in a perfusion bioreactor, combined with cell growth kinetics, is developed and implemented using Computational Fluid Dynamics (with the commercial software COMSOL Multiphysics v5.5). The model describes the spatio-temporal evolution of glucose concentration, oxygen concentration, lactate concentration and cell density within a 3D polymeric scaffold. One of the main issues in the perfusion bioreactor process modelling is that so far all the mathematical models found in literature show discrepancies between the models as well as discrepancies between the values of the parameters. Moreover, for some of the parameters, no values could be found for the case of the perfusion reactor (López-Meza, Araíz-Hernández et al. 2016, Nokhbatolfoghahaei, Bohlouli et al. 2020). The first step of this work presents a global sensitivity analysis (SA) and simulations of a perfusion bioreactor process (Nascu, Chen et al. 2021, Nascu, Sebastia-Saez et al. 2021). The use of sensitivity analysis in the biomedical field has been very beneficial especially for the assessment of the robustness of complex biological and biomedical models and in uncertainty quantification. Moreover it is an important field that brings forward vast benefits especially for model development and control (Razavi, Jakeman et al. 2021). Assessing the impact of the lack of knowledge regarding the model inputs on the predicted outputs of the model is an important step.

This first step will set the foundations for the design of multivariable control algorithms that will facilitate manufacturing for a cell growth process using a continuous perfusion bioreactor. Having an in depth understanding of the model, the interactions between the inputs and the outputs is essential in the optimal design of multivariable control algorithms. Since the high fidelity mathematical model developed and implemented using Computational Fluid Dynamics (CFD) is too complex, a simplified model was approximated and used for the design of the PID controllers. The high fidelity model was used to calibrate the simplified model and will also be used to test the performances of the developed strategies. The use of closed loop controllers in tissue manufacturing will lead to a decrease in production costs and bioreactor operating time as well as an increase in cell density.

Using closed loop control, the control system optimizes the medium feeding quantity for the bioreactor. Using hierarchic control can lead to the optimization of the response of the culture density, the setpoint values can be calculated at the optimization level. Some of the most important advantages of using a controller is that it is able to deal with model uncertainties as well as rejecting disturbances and taking the process back to the desired setpoint values. This work illustrates how using model-based control approaches greatly improves the time and resource utilization during bioreactor operation. The algorithms are designed to work with different types of cell cultures and deal with any disturbances that might appear in the process. Moreover, these strategies are designed to be able to work with different manipulated and controlled variables depending on the needs of the process.

To test how the controllers perform as well as the limitations, the developed control strategies are tested for changes in the parameters, varying operating targets as well as different process disturbances. Furthermore, to test the robustness of the control strategies, they will be tested for different cell cultures. By standardizing and automating tissue manufacturing, production costs and time could be reduced when using closed-loop controlled bioreactors systems. This will lead to a wider use of engineered tissue, it will guarantee consistency of product quality as well as of the time spent to produce the product.

Coletti, F., S. Macchietto and N. Nlvassore (2006). Mathematical modelling of three-dimensional cell cultures in perfusion bioreactors. Part II. Computer Aided Chemical Engineering. W. Marquardt and C. Pantelides, Elsevier. 21: 1699-1704.

Kiparissides, A., M. Koutinas, C. Kontoravdi, A. Mantalaris and E. N. Pistikopoulos (2011). "‘Closing the loop’ in biological systems modeling — From the in silico to the in vitro." Automatica 47(6): 1147-1155.

López-Meza, J., D. Araíz-Hernández, L. M. Carrillo-Cocom, F. López-Pacheco, M. d. R. Rocha-Pizaña and M. M. Alvarez (2016). "Using simple models to describe the kinetics of growth, glucose consumption, and monoclonal antibody formation in naive and infliximab producer CHO cells." Cytotechnology 68(4): 1287-1300.

Nascu, I., T. Chen, W. Du and I. Nascu (2021). Global Sensitivity Analysis for the input parameters of a Perfusion Bioreactor System in Tissue Engineering. 2021 25th International Conference on System Theory, Control and Computing (ICSTCC).

Nascu, I., D. Sebastia-Saez, T. Chen and W. Du (2021). A combined computational-fluid-dynamics model and control strategies for perfusion bioreactor systems in tissue engineering. IFAC-PapersOnLine.

Nascu, I., D. Sebastia-Saez, T. Chen and W. L. Du (2021). "A combined computational-fluid-dynamics model and control strategies for perfusion bioreactor systems in tissue engineering." IFAC-PapersOnLine: 1-6.

Nokhbatolfoghahaei, H., M. Bohlouli, K. Adavi, Z. Paknejad, M. Rezai Rad, M. M. khani, N. Salehi-Nik and A. Khojasteh (2020). "Computational modeling of media flow through perfusion-based bioreactors for bone tissue engineering." Proceedings of the Institution of Mechanical Engineers, Part H: Journal of Engineering in Medicine 234(12): 1397-1408.

Razavi, S., A. Jakeman, A. Saltelli, C. Prieur, B. Iooss, E. Borgonovo, E. Plischke, S. Lo Piano, T. Iwanaga, W. Becker, S. Tarantola, J. H. A. Guillaume, J. Jakeman, H. Gupta, N. Melillo, G. Rabitti, V. Chabridon, Q. Duan, X. Sun, S. Smith, R. Sheikholeslami, N. Hosseini, M. Asadzadeh, A. Puy, S. Kucherenko and H. R. Maier (2021). "The Future of Sensitivity Analysis: An essential discipline for systems modeling and policy support." Environmental Modelling & Software 137: 104954.

Schmid, J., S. Schwarz, R. Meier-Staude, S. Sudhop, H. Clausen-Schaumann, M. Schieker and R. Huber (2018). "A Perfusion Bioreactor System for Cell Seeding and Oxygen-Controlled Cultivation of Three-Dimensional Cell Cultures." Tissue Eng Part C Methods 24(10): 585-595.