Predicting the Morphology of Class IV Neurons from the Dynamics of Dendritic Growth in Drosophila

Abstract

The development of neurons into their stereotyped morphology is paramount to their function. The class IV neurons of the Drosophila melanogaster larva are sensory neurons that develop a complex, highly branched dendritic arbor sensitive to mechanical stimuli. The fully-developed dendritic tree results from a multitude of stochastic processes including dendritic tip growth, branching and self-avoidance. Previous studies have identified key molecular players involved in these processes, which include actin, microtubules, molecular motors, Golgi outposts, endosomes and the Down Syndrome cell adhesion molecule (DSCAM). However, it is yet unknown how these dendritic processes can produce the observed morphologies of the class IV neurons. Here, we formulated an agent-based model of dendritic growth that takes inputs from our previous measurements and analysis of the tip dynamics and branching process. The rules of the agent are based on the observed local behavior of dendritic tips: 1) Branches elongate stochastically by visiting three dynamical states, i.e, growing, paused and shrinking; 2) New branches are created at a rate that is proportional to the local density of dendrites; 3) Branches stop growing when they collide with another branch, which mimics contact-based retraction. Using these rules constrained by experimental observations, we show that the model recapitulates the morphology of class IV neurons. This allows us to bridge the scales of class IV dendritic growth by connecting the local tip dynamics (∼1 µm) to the global tree growth (∼1 mm). In summary, our results establish a mechanistic approach to understand how the small-scale growth dynamics of class IV neurons can shape the large-scale morphology of the dendritic tree.