Glucose diffusivity in tissue engineering membranes and scaffolds: implications for hollow fibre membrane bioreactor

Unlike thin tissues (e.g., skin) which has been successfully grown, growing thick tissues (e.g., bone and muscle) still exhibit certain limitations due to lack of nutrients (e.g., glucose and oxygen) feeding on cells in extracapillary space (ECS) region, or also known as scaffold in an in vitro static culture. The transport of glucose and oxygen into the cells is depended solely on diffusion process which results in a condition where the cells are deprived of adequate glucose and oxygen supply. This condition is termed as hypoxia and leads to premature cell death. Hollow fibre membrane bioreactors (HFMBs) which operate under perfusive cell culture conditions, have been attempted to reduce the diffusion limitation problem. However, direct sampling of glucose and oxygen is almost impossible; hence noninvasive methods (e.g., mathematical models) have been developed in the past. These models have defined that the glucose diffusivity in cell culture medium (CCM) is similar to the diffusivity in water; thus, they do not represent precisely the nutrient transport processes occurring inside the HFMB. In this research, we define glucose as our nutrient specie due to its limited published information with regard to its diffusivity values, especially one that corresponds to cell/tissue engineering (TE) experiments. A series of well-defined diffusion experiments are carried out with TE materials of varying pore size and shapes imbibed in water and CCM, namely, cellulose nitrate (CN) membrane, polyvinylidene fluoride (PVDF) membrane, poly(L-lactide) (PLLA) scaffold, poly(caprolactone) (PCL) scaffold and collagen scaffold. A diffusion cell is constructed to study the diffusion of glucose across these materials. The glucose diffusion across cell-free membranes and scaffolds is investigated first where pore size distribution, porosity and tortuosity are determined and correlated to the effective diffusivity. As expected, the effective diffusivity increases correspondingly with the pore size of the materials. We also observe that the effective glucose diffusivity through the pores of these materials in CCM is smaller than in water. Next, we seeded human osteoblast cells (HOSTE85) on the scaffolds for a culture period of up to 3 weeks. Similar to the first series of the diffusion experiments, we have attempted to determine the effective glucose diffusivity through the pores of the scaffolds where cells have grown at 37°C. The results show that cell growth changes the morphological structure of the scaffolds, reducing the effective pore space which leads to reduced effective diffusivity. In addition, the self-diffusion of glucose in CCM and water has also been determined using a diaphragm cell method (DCM). The results have shown that the glucose diffusivity in CCM has significantly reduced in comparison to the water diffusivity which is due to the larger dynamic viscosity of CCM. The presence of other components and difference in fluid properties of CCM may also contribute to the decrease. We finally employ our experimentally deduced effective diffusivity and self-diffusivity values into a mathematical model based on the Krogh cylinder assumption. The glucose concentration is predicted to be the lowest near the bioreactor outlet, or in the scaffold region, hence this region becomes a location of interest. The governing transport equations are non-dimensionalised and solved numerically. The results shown offer an insight into pointing out the important parameters that should be considered when one wishes to develop and optimise the HFMB design