In the human body, the behavior of cells is choreographed by their response to hundreds of proteins or small molecule ligands. Ligands regulate growth, proliferation, migration, differentiation and even cell death. We want to understand how ligands exert their control over the behavior of human cells, and how the state of each cell – its genetic make-up and biochemical state – affects its response to these ligands.
Our current focus is on the cell death/survival decision induced by treating cancer cells with Tumor Necrosis Factor (TNF), a pro-inflammatory ligand. A fascinating aspect of the response to TNF is its variability: all cells express TNF receptor I, but in some TNF promotes survival or differentiation and in others, it promotes apoptosis. Even in clonal populations of cancer cells, some cells die, some survive. Because TNF induces a complex signaling network that includes pro-survival gene transcription by NF-kB, stress kinase and pro-apoptotic caspase activation, each cell must weigh many signals before committing to survival or apoptosis. Our goal is to identify the main determinants of the TNF-induced cellular response and to understand how cells “compute”, or integrate these many signals to make a life/death decision.
To investigate the TNF-induced life/death decision in human cancer cells, we use a combination of experimental and computational approaches. Our computational models formalize the known – or hypothesized – biochemistry of the TNF-induced signaling network and we use them to predict the effect of genetic or biochemical perturbations to the signaling network. Relying primarily on single-cell methods, we measure protein levels and signaling responses in treated cells in end-point or live-cell assays to test model predictions and the underlying hypotheses. The TNF-signaling network intersects with many others and we hope that by learning how cells compute the signals in this network, we can also predict cellular responses in many other contexts.