Describing in-vitro cell proliferation and transformation with cellular automata

The importance of in-vitro carcinogenesis tests is growing, either for health risk assessments or for screening candidate drugs. Although these systems are simpler than their in-vivo counterparts, their outcomes are nonetheless the result of the interaction of several nonlinear processes. Therefore modelling their behaviour may significantly improve our understanding of these tests. A dynamical model is introduced, which describes the growth of cell cultures (coupling metabolism with proliferation) and the birth of "transformed" cells (which give rise to malignant cell clusters) under the action of a carcinogen. By averaging over the space variable, a simpler (ordinary differential equation) model is obtained, and its behaviour is compared with that of a cellular automata model which preserves space dependence and locality of interactions. It is shown that the latter may describe important phenomena which are hidden by averaging over the whole space. Experimental data are interpreted on the basis of the model, pointing to the role of a previously overlooked experimental variable. These results provide a further indication of the usefulness of cellular automata in modelling complex biological systems.