Cell shape as a window into cell state

The shape of an adherent cell spread on a surface depends upon the biophysical properties of the cytoskeleton. However these are controlled by biochemical circuits and ultimately by gene expression. While shape of a single cell is dynamic, enough evidence has accumulated that cells in the same state adopt shapes that are, in a statistical sense, similar to each other, and distinguishable from cells in different states. If we could understand the major determinants of cell shape, we may be able to infer aspects of cell state merely by observing cell shape. We have imaged thousands of cells in different experimental conditions and have developed a large number of morphological parameters to quantify cell shape and cytoskeletal morphology. Using these we show that quantifiers of cell shape and cytoskeletal texture can be used to discriminate between different cell states. Projections of the data to lower-dimensional shape space can be used to distinguish between similar and dissimilar changes in shape. Pharmacological drugs that perturb the cytoskeletal can help to identify some of the biochemical circuits that control cell shape. Linking results of our experiments and data analysis with previous molecular biology studies on cytoskeletal proteins that affect cell shape, coupled with mathematical modeling, allow us to deconstruct some of the determinants of cell shape.