P51. Bistable cell fate specification as a result of stochastic fluctuations and collective spatial cell behaviour

Cell differentiation is traditionally seen as a sequence of genetically predetermined gene expression changes. This view has been challenged recently by the discovery of important stochastic fluctuations in gene expression and by the identification of the role these fluctuations may play in the cell fate decision. In particularly, it has been proposed that fully multipotent cells fluctuate slowly between states with varying likelihoods of differentiation and this spontaneous heterogeneity or “noise” is the central driving force behind multipotency. It has been proposed that the different phenotypic states correspond to attractor states in the “epigenetic landscape” defined by the network of genes. In our study we used myoblasts that express the differentiation marker CD56 considered as a sign of definitive fate commitment to the muscle pathway. We show experimentally that these cells fluctuate between two phenotypes. Using spatial distribution analysis of the cells in growing population, computer simulations of these observations and experimental test of the new predictions made by the model we show that the fluctuations follow a bistable dynamics driven by a microenvironment and phenotype dependent noise. We show that fluctuation between phenotypes is not a distinguishing feature of multipotent cells. Committed tissue cells (myoblasts, used in our study) also may fluctuate between different phenotypes. We suggest that phenotypic fluctuations are a general feature of any non-terminally differentiated cell, thus, reversion to a previous differentiation state is a normal event. The cellular microenvironment contributes actively to the fluctuations by increasing the noise level in cells with a given phenotype and decreasing it in others. Importantly, the microenvironment is created by the cells themselves as a consequence of their motility that creates random cell interactions. In this way each cell contributes to put together its own microenvironment that in turn stimulates it to fluctuate between the phenotypes until the most appropriate phenotypic state with low noise is found. This means that the attractor state is not cell intrinsic; rather it is specified by the joint action of the cell-intrinsic gene network and the network of cell-cell interactions. Both networks are subject to stochastic fluctuations.