Learning depends in part on new synapse formation. But neurons make few synapses in the absence of astrocytes, which seem to secrete a synaptogenic factor. Christopherson et al. now show that this synaptogenic activity is thrombospondin (TSP)-1 and –2.
To find the synaptic helpers, the authors fractionated astrocyte-conditioned medium. From this soup, TSPs were both necessary and sufficient for neurons to form synapses in vitro. TSPs are extracellular matrix proteins that alter cell adhesions by binding to other matrix proteins or to membrane receptors. It is not clear how TSPs build synapses, but they boost synaptic protein localization. TSPs may activate signaling pathways via receptors on the neuronal cell body, or they may act more locally to reorganize synaptic proteins. There are many known TSP receptors; identification of the relevant ones should help to resolve this question.
The TSP-induced synapses looked normal, but they lacked functional AMPA receptors on the postsynaptic side. As functional synapses are made in the presence of live astrocytes, the findings suggest that the missing effect is due to a second, unidentified, astrocyte-derived factor.
TSP is around at the right time and place to regulate synaptogenesis in the developing brain. The authors found that TSP expression was strong in the postnatal mouse brain but was turned off in adults. Mutant mice lacking TSP-1 and -2 were missing 40% of the synaptic connections of their wild-type counterparts. Humans express much more TSP than do other primates. Perhaps this difference is one reason why Earth is not the planet of the apes.