The objective of this work is to obtain an in depth understanding of how embryonic stem cellsecreted signals contribute to their identity. We analyze the contribution of broad and specific signals present in the cell-secreted microenvironment using techniques that can easily be applied to studies of other cell types and signaling systems. Determining the effects of external signals produced endogenously by stem cells is important for understanding fundamental biological processes regarding cell communication and for implementing more sophisticated manipulation protocols for future clinical applications. Harnessing the ability of stem cells to generate specific cell types is necessary for many regenerative medicine and tissue engineering applications and would be enhanced by a more thorough understanding of the signaling pathways required to maintain stem cell self-renewal and to initiate an exit from the self-renewing state.

In this thesis, we describe work showing that mouse embryonic stem cell (mESC)-secreted signals are required to maintain self-renewal, as cells enter a primed, epiblast-like state of early differentiation when microfluidic perfusion is used to deplete soluble cell-secreted signals. We show that this phenotypic change can be used to our advantage for directed differentiation, and further demonstrate that remodeling the endogenous extracellular matrix halts the exit from the self-renewing state that occurs in mESCs growing under perfusion. Matrix remodeling is then shown to be both necessary and sufficient for maintaining mouse embryonic stem cell selfrenewal in the absence of other external cues, and we demonstrate a method for assessing the relative contributions of soluble versus matrix-based cues.

Together, our data indicate the importance of mESC-secreted factors in contributing to cell survival, self-renewal, and differentiation in normal cultures. Beyond furthering our understanding of intrinsic signaling mechanisms, this information can be used to devise better culture systems for directed differentiation of pluripotent cells. In addition, the techniques developed and implemented here for assessing the contributions of endogenous signals can all be applied generally to any adherent cell type for studies of how the cell-secreted microenvironment contributes to signaling processes and ultimately to cell phenotype.