Abstract
The p53 transcription factor is a master regulator of cellular responses to stress that is commonly inactivated in diverse cancer types. Despite decades of research, the mechanisms by which p53 impedes tumorigenesis across vastly different cellular contexts requires further investigation. The bulk of research has been completed using in vitro studies of cancer cell lines or in vivo studies in mouse models, but much less is known about p53 action in diverse non-transformed human tissues. Here, we investigated how different cellular states modify the p53 transcriptional program in human cells through a combination of computational analyses of publicly available large-scale datasets and in vitro studies using an isogenic system consisting of induced pluripotent stem cells (iPSCs) and two derived lineages. Analysis of publicly available mRNA expression and genetic dependency data demonstrated wide variation in terms of expression and function of a core p53 transcriptional program across various tissues and lineages. To monitor the impact of cell differentiation on the p53 transcriptome within an isogenic cell culture system, we activated p53 by pharmacological inhibition of its negative regulator MDM2. Using cell phenotyping assays and genome wide transcriptome analyses, we demonstrated that cell differentiation confines and modifies the p53 transcriptional network in a lineage-specific fashion. Although hundreds of p53 target genes are transactivated in iPSCs, only a small fraction is transactivated in each of the differentiated lineages. Mechanistic studies using small molecule inhibitors and genetic knockdowns revealed the presence of two major regulatory mechanisms contributing to this massive heterogeneity across cellular states: gene silencing by epigenetic regulatory complexes and constitutive transactivation by lineage-specific transcription factors. Altogether, these results illuminate the impact of cell differentiation on the p53 program, thus advancing our understanding of how this tumor suppressor functions in different contexts.