How cells regulate their shape has mainly been studied using standard genetic and biochemical approaches, revealing roles for many individual genes and signalling proteins. Now, by combining genetic screening with detailed automated measurements of cell morphology, a recent study has begun to reveal how these components act in networks to bring about changes in cell shape.

Bakal, Aach and colleagues screened for cell-shape changes in Drosophila melanogaster BG-2 cells for a total of 249 RNAi or overexpression treatments, corresponding to genes that might be expected to regulate cell morphology, as well as controls. An essential feature of the screen was the use of software that carried out automated imaging and quantitative measurement of 149 features of cell morphology, such as aspects of geometry and cellular protrusions, making the screen rapid and efficient. Altogether, more than 12,600 cells were screened in this way.

To get biologically meaningful information from these measurements, the authors took seven of the treatment conditions that gave particularly distinctive cell morphologies (for example, very flat cells with lots of protrusions). They then trained neural networks to distinguish between these phenotypes using defining sets of the 149 morphological measurements. By applying these networks to all cells in the screen, the authors derived quantitative morphological signatures for each treatment condition that reflected the level of similarity in cell shape with the seven training phenotypes. The treatment conditions could then be clustered into groups, revealing genes that give similar morphological phenotypes when their functions are disrupted.

Analysing these 'phenoclusters' provided insights into the organization of the genetic networks that regulate cell shape. For example, one cluster that corresponded to cells that were particularly round and generally lacking protrusions contained both genes with known functions in the formation of adherence junctions and others that function downstream of these junctions in cell spreading. This finding highlighted the involvement of two distinct pathways that have coordinated roles in the formation of protrusions. For another cluster, the cellular phenotype suggested problems with forming lamellipodia, which is a step downstream of adherence junction formation and the initial extension of protrusions. So, this type of screen can be used to identify sets of genes that are involved in temporal hierarchies of events during cell-shape changes.

The low cost and high efficiency of this method makes it a promising approach for further investigations into the gene networks that regulate cell shape.

Author: Louisa Flintoft