Molecular Control of Neuron Differentiation: From Chromatin to Connectivity

As neurons differentiate and assemble into circuits, they form the most complex and diverse structures of any cell type. Disruption of this process leads to neurodevelopmental and neuropsychiatric disorders, and 1 in 54 children develops one or both intellectual disability and autism spectrum disorder. Our goal is to understand the regulatory control systems of neuron differentiation; by doing this, we will understand the fundamental mechanisms that create form and function in nervous systems and learn how to harness them to understand nervous system disease and nervous system repair.

We study mechanisms that generate and pattern neuronal dendrite and axon arbors and specify their connectivity. To do this, we utilize the nervous system of Drosophila, which allows us to visualize and manipulate live neurons in their natural environment and repeatedly return to the same neuron for quantitative cell biology. We are further developing human cortical precursor and neuron models in culture. Our studies will provide new and unexpected mechanistic insights into creating connectivity and suggest mechanisms dysregulated in neurodevelopmental and neuropsychiatric disorders.

Dendrite and axon arbors connect a neuron to its partners; they pattern to support each neuron's precise connectivity and computational requirements. Genetic programs encode the connectivity of these arbors, and mutations that disrupt these programs lead to the cognitive and behavioral impairments seen in neurodevelopmental and neuropsychiatric disorders. However, the number of specific neuron-to-partner connections in the brain exceeds the number of genes by many orders of magnitude; how is the complex connection map encoded?

Differentiating neurons go through a series of cell biological behaviors to put the arbor into the correct position to find accurate partners and to generate the right pattern for computing information. Thus, genetic programs encode nervous system connectivity by controlling the spatial and temporal dynamics of the neuron differentiation program. To study this, we have developed sophisticated live imaging approaches to follow neuron differentiation in vivo. We employ temporal control over the expression of transgenic reagents in the neuron or surrounding tissues, make precise mutations in genes of interest, and insert tags to track the localization and interactions of endogenous proteins.

Our experimental strategy is to ask how transcriptional and genomic controls regulate the expression and operation of cell biological cytoskeleton effector interaction networks. Then we follow the operation of these networks by in vivo mapping of neuron differentiation and show how the operation of these networks controls arbor pattern formation over time.

Transcriptional control over dendrite arbor pattern. Neurons of the same type share primary arbor pattern parameters, which are critical for circuit assembly. We used Drosophila body wall-situated nociceptive, thermosensitive, and proprioceptive sensory neurons as models to make a series of pioneering discoveries revealing the logic by which transcription factor codes specify these arbor pattern parameters. These have become central to the textbook model of how connectivity is genetically encoded.

Competition between two distinct microtubule generation pathways sets the frequency at which dendrite branches form. The activity of transcription factors that control connectivity is expression level-dependent. Therefore, manipulating the level of these factors is a potent tool for discovering the cell biological effector networks involved in the genetic encoding of connectivity because these manipulations will lead to quantitative changes in the level or distribution of effector activity in the differentiating cell. This approach revealed Centrosomin (Cdk5Rap2 in humans) and Augmin (HAUS in humans). These two distinct microtubule generation pathways create the mitotic spindle6. This study found that they are reutilized after the cell has exited from proliferation to drive postmitotic differentiation. This finding was the first identification that these factors act in neurons, and many international studies have followed. The two factors compete to position and orient microtubule generation events in the differentiating dendrite arbor. Because of this, control over Centrosomin levels during arbor outgrowth sets the frequency at which polymerizing microtubules invade and stabilize newly forming dendrite branches. At a conceptual level, the study demonstrated a route through which the neuron translates genetic encoding of connectivity by organizing dynamic control of local arbor cell biological processes to create arbor patterns.

Atypical myosin targets growth cone-generated microtubules to regulate dendrite arbor differentiation. This project followed the entire sequence of dendrite outgrowth in vivo at the level of the whole arbor (Fig. 1A, B). Then, it identified which individual cytoskeletal remodeling events within that sequence contribute to the final arbor pattern (Fig. 1C–E). Next, it discovered the molecular mechanism of these cytoskeletal remodeling events. Finally, it showed how the events are regulated to control arbor pattern formation (Fig. 1F). My laboratory created a new experimental model for in vivo imaging of the sequence of dendrite arbor differentiation — the pupal c4da neuron. We utilized these new technologies for groundbreaking in vivo live imaging analyses of the function of neurodevelopmental disorder predisposition gene families in dendrite differentiation. These studies revealed a position of repetitive microtubule generation in the growth cone of extending dendrites (Fig. 1D). During branching, growth cone filopodia extend their actin filament core toward the center of the growth cone. Filaments that extend far enough capture and guide polymerizing microtubules into the filopodia, creating branches (Fig. 1E). Quantifying the actin filament dynamics and the consequent microtubule targeting allowed us to build a new model of branch formation. It showed how the branching process is regulated. The actin motor protein Myosin6 is a central regulator of this process; it stabilizes a set of actin filaments, allowing them to lengthen. Control over Myosin6 levels regulates branching frequency and arbor pattern (Fig. 1F).

The position of this microtubule generation center is predictable, and its molecular processes are quantifiable. In future projects, we will use it as one in vivo model in a new research theme designed to understand the operation of microtubule generation events in the neuron.

Control over axon arbor targeting through changing cellular responses to a repetitively-used signaling pathway. Even if they are generated from the same neuronal precursor, closely related neurons create axon arbors that target different partners. Using Drosophila olfactory receptor neurons as a model, this project showed how Notch pathway signaling activity regulates the output of the neuron differentiation program to create differences in axon targeting. Notch pathway activation leads to Notch binding and regulating gene targets. The study showed how controlling the chromatin structure of Notch target genes in the newly born neuron changes axon targeting. This finding provides a new framework for understanding the genetic encoding of connectivity. In a new research theme, we are asking whether related processes function in human cortical precursors as they switch on the neuron differentiation program.