You are here:
Agent-Based Computational Modeling of Cell Culture: Understanding Dosimetry In Vitro as Part of In Vitro to In Vivo Extrapolation
Citation:
Conolly, R., W. Cheng, C. Eklund, J. Samet, P. Wages, K. Lavrich, AND S. Bhattacharya. Agent-Based Computational Modeling of Cell Culture: Understanding Dosimetry In Vitro as Part of In Vitro to In Vivo Extrapolation. SOT 55th Annual Meeting and ToxExpo, New Orleans, LA, March 13 - 17, 2016.
Impact/Purpose:
Quantitative characterization of cellular dose in vitro is needed for alignment of doses in vitro and in vivo. Parallel development of PBPK and virtual tissue models of in vivo biology would support a quantitative, biologically-based approach to the alignment of cellular doses in vitro and in vivo.
Description:
Quantitative characterization of cellular dose in vitro is needed for alignment of doses in vitro and in vivo. We used the agent-based software, CompuCell3D (CC3D), to provide a stochastic description of cell growth in culture. The model was configured so that isolated cells assumed a “fried egg shape” but became increasingly cuboidal with increasing confluency. The surface area presented by each cell to the overlying medium varies from cell-to-cell and is a determinant of diffusional flux of toxicant from the medium into the cell. Thus, dose varies among cells for a given concentration of toxicant in the medium. Computer code describing diffusion of H2O2 from medium into each cell and clearance of H2O2 was calibrated against H2O2 time-course data (25, 50, or 75 uM H2O2 for 60 min) obtained with the Amplex Red assay for the medium and the H2O2-sensitive fluorescent reporter, HyPer, for cytosol. Cellular H2O2 concentrations peaked at about 5 min and were near baseline by 10 min. The model predicted a skewed distribution of surface areas, with between cell variation usually 2 fold or less. Predicted variability in cellular dose was in rough agreement with the variation in the HyPer data. These results are preliminary, as the model was not calibrated to the morphology of a specific cell type. Future work will involve morphology model calibration against human bronchial epithelial (BEAS-2B) cells. Our results show, however, the potential of agent-based modeling to provide a rich description of individual cell dosimetry, with average dose and variability statistics readily derivable. Parallel development of PBPK and virtual tissue models of in vivo biology would support a quantitative, biologically-based approach to the alignment of cellular doses in vitro and in vivo. This abstract does not necessarily reflect any specific policy of the US EPA.