The Influence of Cellular Characteristics on the Evolution of Shape Homeostasis

Abstract

The importance of individual cells in a developing multicellular organism is well known, but precisely how the individual cellular characteristics of those cells collectively drive the emergence of robust, homeostatic structures is less well understood. For example, cell communication via a diffusible factor allows for information to travel across large distances within the population, and cell polarization makes it possible to form structures with a particular orientation, but how do these processes interact to produce a more robust and regulated structure? In this study we investigate the ability of cells with different cellular characteristics to grow and maintain homeostatic structures. We do this in the context of an individual-based model where cell behavior is driven by an intracellular network that determines the cell phenotype. More precisely, we investigated evolution with 96 different permutations of our model, where cell motility, cell death, long-range growth factor (LGF), short-range growth factor (SGF), and cell polarization were either present or absent. The results show that LGF has the largest positive influence on the fitness of the evolved solutions. SGF and polarization also contribute, but all other capabilities essentially increase the search space, effectively making it more difficult to achieve a solution. By perturbing the evolved solutions, we found that they are highly robust to both mutations and wounding. In addition, we observed that by evolving solutions in more unstable environments they produce structures that were more robust and adaptive. In conclusion, our results suggest that robust collective behavior is most likely to evolve when cells are endowed with long-range communication, cell polarisation, and selection pressure from an unstable environment.

PhysiBoSS: a multi-scale agent-based modelling framework integrating physical dimension and cell signalling

“…The choice of appropriate formalism depends on the modelling scope (e.g. see the summary of different models used to study shape homeostasis in Gerlee et al , 2017). At the level of the cell population, some previous modelling efforts have been done with cellular automaton models (e.g.…”

Towards Self-organized Control: Using Neural Cellular Automata to Robustly Control a Cart-pole Agent

“…To search among the wide rule spaces, genetic algorithms have been frequently used to find CA that exhibited a predefined behavior. This enables the design of CA that robustly grows a shape, in effect exhibiting homeostasis [21]. Another work was able to develop a targeted shape and maintain it despite damage [22].…”

PhysiBoSS: a multi-scale agent based modelling framework integrating physical dimension and cell signalling

“…[8][9][10]. The choice for the appropriate formalism depends on the modelling scope (see for example the summary of different models used in shape homoeostasis in [11]). Some previous modelling efforts have been done with models of cellular automata [12], with the development of Potts models [13,14], or agent-based models [15,16], but these approaches have typically focused on the emergent multicellular dynamics after assigning simple, microenvironment-driven single-cell phenotypes, rather than including both the intracellular events and how individual cellular alterations might affect the tumour.…”