Abstract

A High Content Imaging Platform to Study Spatio-temporal Signaling Networks during Neuronal Differentiation

Olivier Pertz, Professor, University Of Basel

Differentiating neurons extends neuronal processes that will later become the axons and dendrites that connect the adult brain. This process, called "neurite outgrowth", depends on highly dynamic and co-ordinated regulation of actin and microtubule cytoskeleton, adhesion and trafficking dynamics. However, the different proteins regulating this process have been studied one by one in a wide variety of model systems, precluding a systems level view of this process. An additional caveat is that neurite outgrowth is a highly dynamic process, that consists of multiple stochastic events such as neurite initiation, branching and protrusion/retraction cycles that will be missed in classic steady-state neurite outgrowth assays. To address these problems, we report on a high content imaging patform that allows to study neurite outgrowth dynamics in response to a large amount of perturbations. A computer vision approach then allows to extract a large number of features that describe neurite morphodynamics. Finally, a machine learning-based algorithm allows to select relevant phenotypic features that are pertinent to each molecular perturbation. This automated platform was used to dissect a potential Rho GTPase interactome using an RNA interference approach, and revealed modular signaling units that regulate specific morphodynamic phases of the neurite outgrowth process (inititation, protrusion, retraction, branching, ...). The computational efficiency of our platform allowed to analyze 3 TB of datasets, representing a total of 20000 timelapse movies. Exploring neurite morphodynamic phenotypes clearly allows for a better understanding of this process, which should give novel insight for the rational targeting of a number of pathological conditions.