Synapses are crucial nodes of information transfer in the brain. The pattern of these connections, which outnumber neurons by 100 to 1, is central to brain function, but little has been known about the basic molecular mechanisms that allow this intricate network to wire together. A tantalizing clue was presented by the study of purified retinal ganglion cells (RGCs) in culture, which form few synapses unless they are cultured in the presence of medium conditioned by exposure to glial cells (glial-conditioned medium; GCM). This observation indicated that a factor secreted by glial cells might be limiting for neuronal synapse formation. Now Mauch et al. report that they have identified the crucial glial-derived factor, and surprisingly it seems to be cholesterol.

The authors took two approaches to find the glial-derived factor. First, they performed partial purification of the synapse-inducing activity from GCM to establish the molecular weight and chromatographic properties of the active fraction. Then they ran two-dimensional silver-stained gels of RGCs grown in the absence or presence of GCM to identify proteins that transferred from the glia to the neurons. One prominent silver-stained spot was seen to correlate with exposure to glial cells. Microsequencing by nanospray mass spectrometry showed the protein to be the lipoprotein carrier ApoE, the molecular weight and chromatographic profile of which are consistent with the partially purified active fraction from GCM.

Despite this correlation, when the authors added purified ApoE back into RGC cultures, the protein had no synapse-inducing activity. However, reasoning that the function of ApoE is to deliver cholesterol to cells, the authors next tested the effects of adding purified cholesterol to the cultures, and found that cholesterol alone recapitulated all of the synapse-inducing effects of GCM. In addition, they found that the cholesterol content of RGCs rose when they were cultured with GCM, indicating that the synthesis of cholesterol within neurons might be limiting for synaptogenesis.

Although compelling, these data raise a number of questions. Is cholesterol synthesis really limiting for synaptogenesis in vivo, when the RGCs are in a far more enriched environment? What is the role of glial-derived ApoE, as the in vitro addition of cholesterol alone can induce synapses? Cholesterol is an important structural component of both presynaptic vesicles and the caveolae-like lipid rafts where many postsynaptic receptors cluster, but ApoE can also activate a large number of intracellular signalling pathways through interactions with the low density lipoprotein (LDL) family of receptors. Nonetheless, this elegant assay is supported by the observation of Ullian et al. that synaptogenesis of RGCs in vivo correlates in time with gliogenesis, indicating that further study in animals could define a role for lipoprotein modulation of synaptogenesis in the brain.