Brain cells come in a staggering variety of shapes and sizes. So how do neurons grow to be the shapes they are? It has been thought that neural activity at the sites where neurons communicate (synapses) helps mold cell shape. However, experimentally testing this idea in vivo has proven tricky. A recent study by Marco Tripoli and co-workers from the University of Cambridge set out to test with a new level of precision and rigor how synaptic activity shapes neuronal growth using the unique and powerful advantages of Drosophila as a model organism.
First, the team characterized how known fly motor neuron morphology changes during embryonic life. They wanted to know whether any morphological changes are correlated with the onset of synaptic input to these neurons. Indeed, they found that the cellular arbors of identified fly neurons grow linearly for a time, then suddenly slow their growth. And this slow down happens right around the time when synapses start bombarding the cells with excitatory input.
To test whether there is a causal link between arbor growth and synaptic activity, the team used genetic tools to increase and decrease synaptic input from upstream (pre-) synaptic partners. They found that when pre-synaptic activity was abolished, the neurons grew longer; conversely, when they genetically increased the number of pre-synaptic connections, the neuron trees grew less and stayed closer to home. So the neurons seemed to grow in search of synaptic activity.
The researchers then went in and counted synaptic sites on the growing neurons. Each neuron appeared to match its own growth to the availability of pre-synaptic sites. It's hard not to anthropomorphize: it's almost as if the neurons `knew' how many synaptic connections they needed to make, and then adjusted their own size until they filled the quota.
The team also found that initially, arbor length was inhibited by the presence of a synapse – regardless of whether the synapse was silent or active. Furthermore, areas farther away from synaptic sites showed more growth. This suggests that early remodeling in this neuron is a `locally grown' operation based on local physical contact, but not activity of the synapse.
Finally, the team wanted to determine what molecules inside the cells could be involved in signaling `grow' or `don't grow'. Protein kinase A (PKA) has been implicated in modulating physiological responses to changing synaptic input. To test whether PKA also plays a role in structural responses, the team went back to their genetic toolbox, pulled out a way to up regulate and down regulate PKA function and found that neurons with reduced PKA function have trouble remodeling themselves, and vice versa. All of these results suggest that, through PKA signaling, neurons are able to adjust their shape as they grow to maintain an optimal number of connections, even in the face of varying synaptic input.
One thing that makes this paper powerful is the elegant use of genetic tools. A lot of the experiments are conceptually straightforward, but extremely tough to actually pull off in vivo. Drosophila genetics makes these experiments possible. This study is also important because it calls out for future work by computational neuroscientists. Clearly, the next step is to try to link structural features to function in the growing neuron arbor. Computer simulations would be the best way to do this, and the detailed 3D reconstructions of neurons in this study are calling out to computer scientists to `come model me'. In future years, coupling the Drosophila genetic toolbox with the theoretician's computational toolbox could provide powerful insights into how local activity in the distant branches of cellular arbors shapes how neurons grow.
