After more than five years of heroic work, EMBL Rome PhD student Laetitia Weinhard in the Gross Lab has finally completed her massive imaging study of microglia. Published last week in BioRxiv, the work uses correlated light and electron microscopy (CLEM) as well as time-lapse light sheet imaging to find out whether microglia eat synapse during brain development. Microglia have always held a special fascination as invaders in the brain with unknown function. They are born from blood precursors in the early embryo and then migrate and take up long-term residency in the brain where they have been shown to rapidly move about their long processes, presumably to survey the surrounding tissue – but for what reason remained a mystery. As cousins of macrophages, many researchers proposed that they pluck off and eat synapses as an essential step in the pruning of connections during early circuit refinement. But no one had seen them phagocytose synapses until Laetitia looked more closely.
Laetitia made several observations. First, she found that microglia do not engulf dendritic spines as had been previously proposed – neither electron microscopy nor time-lapse fluorescence imaging provided any evidence for the phagocytosis of postsynaptic components. Instead, microglia eat presynaptic structures, biting off, or “nibbling”, small pieces of boutons and axons. In all cases, however, Laetitia found that a functional synapse was left behind – why, she asked, do microglia nibble boutons if not to eliminate them? A clue came from her second observation – microglia make synapses grow and rearrange. Half of the time that microglia contacted synapses, the spine head sent out a thin process, or filopodia, to greet them. In one particularly dramatic case that Laetitia captured by electron microscopy, 15 spines extended filopodia toward a single microglia process as it was nibbling a synaptic bouton (see image above). This observation suggested that nibbling and filopodia extension might be linked and Laetitia was able to find several cases where microglia elicited filopodia that then reached out toward a nearby bouton, helping its spine to switch from its old bouton to the new one. Such spine switching is thought to underlie the formation of double synapses in which one bouton releases neurotransmitter onto two spines. If these spines are located on the same dendritic shaft, then this sort of multi-synapse bouton (MSB) supports more effective synaptic connectivity. A key maturation step in the postnatal development of neural circuits is the formation of MSBs, and earlier work by Yang Zhan in our lab had shown that deficits in neuron-microglia signaling lead to a failure to form MSBs and global deficits in functional brain connectivity. Could microglia then, be nibbling synapses as a way to make them stronger, rather than weaker?