In many vertebrates, neural circuits undergo substantial reorganization in early postnatal life. This reorganization strengthens some neuron-to-neuron connections, while other neuronal connections are weakened by synapse elimination. It is likely that these changes are related to early electrical activity in the nervous system. Hence, this phenomenon is likely one of the main strategies by which a vertebrate nervous system is modified in long-lasting ways in response to experience.
These synaptic rearrangements have been best described in the peripheral nervous system, where the relative simplicity of the circuits and their better accessibility have made analysis less difficult than in the brain. As reported in Cell Reports this week, Alyssa Wilson and her colleagues in the Lichtman lab have applied a new observational method to circumvent the challenges of analyzing synaptic reorganization in the brain, and have generated a detailed description of how neurons in the mouse cerebellum reorganize their connections during the first postnatal week. Wilson et al. reconstructed the connections between neurons using serial electron microscopy in 3- and 7-day-old mouse cerebella. Because synapse reorganization is a process that is conserved across mice, the authors were able to learn more about it by making quantitative comparisons of synaptic connectivity patterns at each time point they reconstructed (each sample came from a distinct mouse).
Wilson and her colleagues found that the reorganization process proceeds by a highly selective strengthening, through synapse addition, of connections between individual inputs (“climbing fibers”) and a small subset of their target neurons (“Purkinje cells’”). Between postnatal days 3 and 7, the most strongly connected pairs of inputs and targets get progressively stronger to an extent that suggests a positive feedback mechanism. Interestingly, the loss of synapses that eventually prunes away weak connections is not occurring at this early stage, even at low rates. Thus synapse addition precedes the synapse elimination phase, and input strengthening by one input does not require weakening by another. The electron microscopy approach also allowed a comparison of the individual synaptic sites associated with very strong and weak connections. The authors found that they were indistinguishable. Thus, a change in the absolute number of synapses connecting neurons (as opposed to the size of the synapses) is the way these inputs get stronger. These mechanisms are quite unlike what has been found in the peripheral nervous system, where synapse removal and synapse addition occur concurrently and are intimately related.
Because many inputs and target cells could be analyzed at the same time, it was also possible to compare the behavior of different neurons in the same sample. The authors found that branches of single climbing fibers tended to work together, forming strong connections on the same few Purkinje cell targets and maintaining weak connections with their other Purkinje cell targets. These patterns were recapitulated by all the climbing fibers and Purkinje cells that the authors analyzed.
Finally, Wilson and her colleagues were able to quantify the number of input climbing fibers within a local region of cerebellar cortex (~100 μm on a side). They found that the numbers of climbing fibers and target Purkinje cells were similar. This signifies that initial over-innervation by extra branching is actually economical, because each of the inputs is in principle able to retain a connection with one Purkinje cell after rewiring is complete, so that none of them needs to have sent branches there in vain. These kinds of insights could only be formulated using the connectomic approach, which allows for a “big picture” analysis of synaptic reorganization.