Identification of the metabolic parameters of hepatocytes for the generation of scalable in vitro models

Nowadays, a critical aspect in tissue engineering is the need to provide engineered tissues an adequate nutrient supply, in particular oxygen, consumed by the cells as it is one of the main actors in cellular respiration. This is particularly crucial for three-dimensional (3D) in vitro constructs. In fact, since they lack a vascular network, oxygen transport is only delivered through passive diffusion, which is limited because of the low oxygen solubility within the culture medium. The dynamics of oxygen diffusion and consumption can be modeled thanks to the second Fick’s equation, coupled with the Michaelis-Menten (MM) kinetics, a nonlinear functional able to describe the behaviour of cells competing/collaborating for the oxygen present in the environment. Specifically, the characteristic parameters of kinetics are: the rate of cell oxygen consumption, OCR, and the MM constant, km. These two parameters are characteristic of a specific cell type and depend on different variables (e.g., cell density, temperature). This thesis aims at identifying the parameters characterizing the oxygen consumption of hepatocytes cultured in monolayer. Specifically, their dependence on different cell densities was evaluated, keeping the temperature constant. The work was carried out in four different stages: 1. Since the kinetics parameters depend on the temperature (in accordance with the Arrhenius equation, a lower temperature corresponds to a lower kinetic constant and, therefore, to a slower consumption), a heating chamber was purposely designed and fabricated for keeping the cells at a constant physiological temperature of 37 °C. The optimal chamber dimensions were defined based on finite element (FE) simulations, taking into account some design constraints. In fact, the heating chamber needs to be used in different experimental setups, thus it must be easy to disassemble and reassemble, minimizing the temperature losses. 2. In parallel, I defined the experimental setup for the in vitro oxygen measurements through FE simulations of oxygen transport and consumption. Specifically, I modelled different surface cell densities (3.13e8, 3.13e7, 3.13e6, 3.13e5 [cell/m^2]), cultured in monolayer in a 96 well plate with increasing volumes (i.e. heights: 10.1, 8.18, 5.45, 2.43 [mm]) of normoxic medium. A no-flux condition has been imposed on the upper surface of the well. Under this hypothesis, the system can reach a stationary state, characterized by a complete consumption of the oxygen initially administered through the medium, thus allowing the identification of both parameters. 3. The experiments were carried out on human hepatocytes (Hepg2) seeded in monolayer onto an oxygen sensor spot (Pyroscience) in a 96 well plate, for monitoring the decrease of the oxygen concentration over time. After overnight incubation and sensor 2-point calibration using normoxic and hypoxic medium, respectively, the well was placed in the heating chamber set at 37 °C. Mineral oil -which is impermeable to oxygen and creates a barrier to environmental oxygen diffusion- was overlaid to fresh medium before starting the tests. 4. The oxygen concentration profiles obtained were analyzed for deriving the consumption parameters. Specifically, an iterative procedure was developed, that minimizes the mean square error (MSE), calculated between the experimental oxygen concentration profiles and those estimated by simulating the dynamics of the corresponding configuration, through the Matlab Partial Derivative Equation (PDE) toolbox. Despite that, during the experiments, the hypoxic steady state predicted by the simulations was never reached, probably due to the mineral oil, which is not able to perfectly seal the well. Thus, the only identifiable parameter was the consumption rate per cell (OCR), that can be easily identified analysing the oxygen profiles during the first time points of the experiments. Statistical analyses carried out using a non-parametric test confirm that the OCR is not a constant for a given cell type but depends on the cell density in culture. These preliminary results demonstrated that, unlike what stated in the literature, the consumption parameters cannot be assumed as constants. Given the importance of oxygen consumption in tissues, this study provides a rigorous methodology to quantify the cell oxygen consumption kinetics, that is critical for the development of physiologically relevant engineered tissues