Modeling and simulation of mammalian cell culture in hollow fiber bioreactors

Abstract

In vitro culturing of mammalian cells is indispensable for the production of a large number of valuable products with potential medical applications. The fastidious nature of these cells, however, imposes several restrictions on the scale-up of the laboratory-scale reactors used for the production of these cells. In an attempt to overcome these limitations, several new technologies have emerged over the years. One of these employs a hollow fiber bioreactor in which a membrane is used for the separation of the cells from the moving fluid stream. The hollow fiber reactors offer several advantages over conventional reactors and have been successfully used to obtain high cell densities.

Experimental investigations reported in the literature have shown the presence of a recirculatory flow across the fiber membranes in the hollow fiber reactors, in addition to the primary flow occurring through the fibers. The recirculatory flow (also referred to as secondary flow) has significant implications for the substrate and metabolite transfer and hence the growth of the cells. A thorough understanding of the hydrodynamics, substrate and product transport, and kinetics of cell growth and product formation forms a prerequisite to an efficient scale-up of these reactors. The intricate design of these reactors, along with the necessity to preserve a stringent micro-environment for proper cell growth, however, makes experimental measurements of all the quantities of interest extremely tedious. Thus, mathematical models are necessary to picture the intercoupled phenomena of fluid flow, nutrient and metabolite transfer, and cell growth in these reactors. In the present work, an endeavor has been made to study this complex interdependence between fluid flow and cell culture in a hollow fiber bioreactor based on single fiber studies.

To describe the flow in a hollow fiber during cell growth, a model has been formulated and solved analytically. The results show the presence of recirculatory flow in the fiber. The dependence of the flow field on the operating pressure drop, fiber dimensions, and membrane properties have been looked into. The results of the single fiber studies have been used to find the optimal fiber spacing in a multifiber module so that the secondary flow is maximized. Since the mammalian cells are shear-sensitive, a shear stress analysis has been carried out. The model has been extended to incorporate spatial variation in permeabilities and has been solved analytically. The results illustrate the change in the flow profile in the fiber in the presence of such variations.

The models for cell culture have been formulated to describe the growth of both anchorage-dependent and anchorage-independent cells. The model equations have been solved numerically using the finite volume method. The numerical scheme has been coupled with the analytical solutions for fluid flow to incorporate the changes in the fluid flow that occur during cell growth. The growth of the anchorage-dependent cells has been studied for two different kinetics. It has been shown that the reactor performance depends on the design of the fiber, membrane properties, operating conditions, and the type of cell lines used. The results show that the dependence of the reactor performance on various parameters is influenced by the substrate consumption and cell growth kinetics. The effect of the recirculatory flow has also been seen to be kinetics-dependent.

The culturing of the anchorage-independent mammalian cells has been studied considering a complex kinetics with dual substrate limitation, inhibitory metabolite formation, death of cells, and monoclonal antibody production. The effect of the secondary flow has been studied. The dependence of the performance of the reactor on the various operating, design, and growth parameters has been investigated. The results show that oxygen acts as the major limiting substrate in the culturing of these cells.

The cell growth studies for a single fiber have been used to find the optimal fiber spacing in a multifiber module so that the total cell count for the module is maximized. The optimal fiber spacing has also been obtained for maximum production of monoclonal antibody, using anchorage-independent cells. Various operating strategies have been shown to improve the reactor performance. The increase in the lumen pressure drop, inlet substrate concentration, and reversal of lumen flow at specified intervals of time have been shown to improve the reactor performance. Operating two short reactors in parallel instead of using a long, single one has also been shown to yield more cells and more quantities of monoclonal antibody.