Representative three-dimensional serial reconstructions of two mossy fibre synaptic complexes taken before (top) and during (bottom) the stereotyped pruning of infrapyramidal bundle mossy fibres, which originate from the dentate gyrus of the mouse hippocampus. Before stereotyped pruning (top), a large mossy fibre presynaptic terminal (green) is seen associating with several postsynaptic densities (yellow) on multiple dendritic spines (purple) of a CA3 pyramidal cell. During stereotyped pruning, the synaptic complex is much smaller and less complex in shape (bottom) as a presynaptic terminal (green) is seen to associate with a single postsynaptic density (yellow) on a small dendritic shaft (purple) of a CA3 pyramidal cell. Image courtesy of H.-J. Cheng, Center for Neuroscience, University of California, Davis, USA.

Collateral branch connections of axons are issued during development to ensure that all potential postsynaptic targets are covered. However, the exact cellular mechanisms underlying the pruning of inappropriate connections continue to elude developmental neuroscientists. Writing in The Journal of Neuroscience, Liu, Low and colleagues shed light on how these connections are eliminated.

It takes hundreds of thousands of synaptic connections to create the intricate circuitry of the CNS, and, to ensure that all synaptic targets are covered, some of the early connections are imprecise and axonal projections are often too long. How this rough draft of the CNS is subsequently refined is still not fully understood, but it seems that different neuronal populations use various processes to prune and remodel axons.

Focusing on the hippocampal mossy fibre system, Liu et al. investigated pruning in the synaptic complex, which is formed by mossy fibres of the infrapyramidal bundle that transiently make contact with basal dendrites of CA3 pyramidal cells. The authors considered three potential pruning methods: axon retraction (axon contents are recycled); Wallerian degeneration (axon fibres are atrophic and destroyed after injury); and axosome shedding (axonal contents are discharged from the terminal and the remnants are engulfed by glial cells).

Immunocytochemistry and electron microscopy analysis showed that a transient contact was formed between the mossy fibre boutons and the basal dendrites. But, as pruning progressed, both the size and complexity of the transient synaptic complex diminished. Because no axonal degeneration or glial cell engulfment was detected, the authors concluded that axonal retraction was occurring during stereotyped pruning, a process that is thought to be triggered by plexin A3 signalling.

It has been suggested that plexin A3 signalling initially disrupts the development of the synaptic complex, and it is this disruption that elicits stereotyped pruning. So, the authors reasoned that in the absence of plexin A3 signalling, the mossy fibre synaptic complexes would continue to acquire more complex structures. Indeed, they found that in plexin A3 and neuropilin 2 (also required for stereotyped pruning) knockout mice, the mossy fibre boutons continued to mature.

The purpose of pruning is thought to be the fine-tuning of neural circuitry rather than the correction of errors. How pruning maintains appropriate connections while removing inappropriate ones remains to be resolved. However, the results reported by Liu, Low and colleagues provide insights into the molecular events that underlie the refinement of neural circuitry.