Computational fluid modeling and performance analysis of a bidirectional rotating perfusion culture system

Osteogenic Activity on NaOH-Etched Three-Dimensional-Printed Poly-ɛ-Caprolactone Scaffolds in Perfusion or Spinner Flask Bioreactor

Bioreactor systems, for example, spinner flask and perfusion bioreactors, and cell-seeded three-dimensional (3D)-printed scaffolds are used in bone tissue engineering strategies to stimulate cells and produce bone tissue suitable for implantation into the patient. The construction of functional and clinically relevant bone graft using cell-seeded 3D-printed scaffolds within bioreactor systems is still a challenge. Bioreactor parameters, for example, fluid shear stress and nutrient transport, will crucially affect cell function on 3D-printed scaffolds. Therefore, fluid shear stress induced by spinner flask and perfusion bioreactors might differentially affect osteogenic responsiveness of pre-osteoblasts inside 3D-printed scaffolds. We designed and fabricated surface-modified 3D-printed poly-ɛ-caprolactone (PCL) scaffolds, as well as static, spinner flask, and perfusion bioreactors to determine fluid shear stress and osteogenic responsiveness of MC3T3-E1 pre-osteoblasts seeded on the scaffolds in the bioreactors using finite element (FE)-modeling and experiments. FE-modeling was used to quantify wall shear stress (WSS) distribution and magnitude inside 3D-printed PCL scaffolds within spinner flask and perfusion bioreactors. MC3T3-E1 pre-osteoblasts were seeded on NaOH surface-modified 3D-printed PCL scaffolds, and cultured in customized static, spinner flask, and perfusion bioreactors up to 7 days. The scaffolds' physicochemical properties and pre-osteoblast function were assessed experimentally. FE-modeling showed that spinner flask and perfusion bioreactors locally affected WSS distribution and magnitude inside the scaffolds. The WSS distribution was more homogeneous inside scaffolds in perfusion than in spinner flask bioreactors. The average WSS on scaffold-strand surfaces ranged from 0 to 6.5 mPa for spinner flask bioreactors, and from 0 to 4.1 mPa for perfusion bioreactors. Surface modification of scaffolds by NaOH resulted in a surface with a honeycomb-like pattern and increased surface roughness (1.6-fold), but decreased water contact angle (0.3-fold). Both spinner flask and perfusion bioreactors increased cell spreading, proliferation, and distribution throughout the scaffolds. Perfusion, but not spinner flask bioreactors more strongly enhanced collagen (2.2-fold) and calcium deposition (2.1-fold) throughout the scaffolds after 7 days compared with static bioreactors, likely due to uniform WSS-induced mechanical stimulation of the cells revealed by FE-modeling. In conclusion, our findings indicate the importance of using accurate FE models to estimate WSS and determine experimental conditions for designing cell-seeded 3D-printed scaffolds in bioreactor systems. Impact Statement The success of cell-seeded three-dimensional (3D)-printed scaffolds depends on cell stimulation by biomechanical/biochemical factors to produce bone tissue suitable for implantation into the patient. We designed and fabricated surface-modified 3D-printed poly-ɛ-caprolactone (PCL) scaffolds, as well as static, spinner flask, and perfusion bioreactors to determine wall shear stress (WSS) and osteogenic responsiveness of pre-osteoblasts seeded on the scaffolds using finite element (FE)-modeling and experiments. We found that cell-seeded 3D-printed PCL scaffolds within perfusion bioreactors more strongly enhanced osteogenic activity than within spinner flask bioreactors. Our results indicate the importance of using accurate FE-models to estimate WSS and determine experimental conditions for designing cell-seeded 3D-printed scaffolds in bioreactor systems.

Inlet flow rate of perfusion bioreactors affects fluid flow dynamics, but not oxygen concentration in 3D-printed scaffolds for bone tissue engineering: Computational analysis and experimental validation

Fluid flow dynamics and oxygen-concentration in 3D-printed scaffolds within perfusion bioreactors are sensitive to controllable bioreactor parameters such as inlet flow rate. Here we aimed to determine fluid flow dynamics, oxygen-concentration, and cell proliferation and distribution in 3D-printed scaffolds as a result of different inlet flow rates of perfusion bioreactors using experiments and finite element modeling. Pre-osteoblasts were treated with 1 h pulsating fluid flow with low (0.8 Pa; PFF^low) or high peak shear stress (6.5 Pa; PFF^high), and nitric oxide (NO) production was measured to validate shear stress sensitivity. Computational analysis was performed to determine fluid flow between 3D-scaffold-strands at three inlet flow rates (0.02, 0.1, 0.5 ml/min) during 5 days. MC3T3-E1 pre-osteoblast proliferation, matrix production, and oxygen-consumption in response to fluid flow in 3D-printed scaffolds inside a perfusion bioreactor were experimentally assessed. PFF^high more strongly stimulated NO production by pre-osteoblasts than PFF^low. 3D-simulation demonstrated that dependent on inlet flow rate, fluid velocity reached a maximum (50-1200 μm/s) between scaffold-strands, and fluid shear stress (0.5-4 mPa) and wall shear stress (0.5-20 mPa) on scaffold-strands surfaces. At all inlet flow rates, gauge fluid pressure and oxygen-concentration were similar. The simulated cell proliferation and distribution, and oxygen-concentration data were in good agreement with the experimental results. In conclusion, varying a perfusion bioreactor's inlet flow rate locally affects fluid velocity, fluid shear stress, and wall shear stress inside 3D-printed scaffolds, but not gauge fluid pressure, and oxygen-concentration, which seems crucial for optimized bone tissue engineering strategies using bioreactors, scaffolds, and cells.

Fluid Dynamic Modeling to Support the Development of Flow-Based Hepatocyte Culture Systems for Metabolism Studies

Accurate prediction of metabolism is a significant outstanding challenge in toxicology. The best predictions are based on experimental data from in vitro systems using primary hepatocytes. The predictivity of the primary hepatocyte-based culture systems, however, is still limited due to well-known phenotypic instability and rapid decline of metabolic competence within a few hours. Dynamic flow bioreactors for three-dimensional cell cultures are thought to be better at recapitulating tissue microenvironments and show potential to improve in vivo extrapolations of chemical or drug toxicity based on in vitro test results. These more physiologically relevant culture systems hold potential for extending metabolic competence of primary hepatocyte cultures as well. In this investigation, we used computational fluid dynamics to determine the optimal design of a flow-based hepatocyte culture system for evaluating chemical metabolism in vitro. The main design goals were (1) minimization of shear stress experienced by the cells to maximize viability, (2) rapid establishment of a uniform distribution of test compound in the chamber, and (3) delivery of sufficient oxygen to cells to support aerobic respiration. Two commercially available flow devices – RealBio^® and QuasiVivo^® (QV) – and a custom developed fluidized bed bioreactor were simulated, and turbulence, flow characteristics, test compound distribution, oxygen distribution, and cellular oxygen consumption were analyzed. Experimental results from the bioreactors were used to validate the simulation results. Our results indicate that maintaining adequate oxygen supply is the most important factor to the long-term viability of liver bioreactor cultures. Cell density and system flow patterns were the major determinants of local oxygen concentrations. The experimental results closely corresponded to the in silico predictions. Of the three bioreactors examined in this study, we were able to optimize the experimental conditions for long-term hepatocyte cell culture using the QV bioreactor. This system facilitated the use of low system volumes coupled with higher flow rates. This design supports cellular respiration by increasing oxygen concentrations in the vicinity of the cells and facilitates long-term kinetic studies of low clearance test compounds. These two goals were achieved while simultaneously keeping the shear stress experienced by the cells within acceptable limits.

Numerical investigation of dynamic microorgan devices as drug screening platforms. Part II: Microscale modeling approach and validation

The authors have previously reported a rigorous macroscale modeling approach for an in vitro 3D dynamic microorgan device (DMD). This paper represents the second of a two-part model-based investigation where the effect of microscale (single liver cell-level) shear-mediated mechanotransduction on drug biotransformation is deconstructed. Herein, each cell is explicitly incorporated into the geometric model as single compartmentalized metabolic structures. Each cell's metabolic activity is coupled with the microscale hydrodynamic Wall Shear Stress (WSS) simulated around the cell boundary through a semi-empirical polynomial function as an additional reaction term in the mass transfer equations. Guided by the macroscale model-based hydrodynamics, only 9 cells in 3 representative DMD domains are explicitly modeled. Dynamic and reaction similarity rules based on non-dimensionalization are invoked to correlate the numerical and empirical models, accounting for the substrate time scales. The proposed modeling approach addresses the key challenge of computational cost towards modeling complex large-scale DMD-type system with prohibitively high cell densities. Transient simulations are implemented to extract the drug metabolite profile with the microscale modeling approach validated with an experimental drug flow study. The results from the author's study demonstrate the preferred implementation of the microscale modeling approach over that of its macroscale counterpart.