Science 14 December 2007:
Vol. 318. no. 5857, pp. 1710 - 1711

Immune Molecules Prune Synapses in Developing Brain
Greg Miller
The complement cascade is part of the body's innate immune defense: a protein work crew whose duties include tagging bacteria and other bad guys for elimination. A new study suggests that complement proteins may have a surprising yet analogous function in the developing brain, tagging unwanted synapses for removal. The work also hints that these proteins may promote synapse loss in early stages of neurodegenerative disease.




"It's a pretty provocative finding," says Greg Lemke, a neurobiologist at the Salk Institute for Biological Studies in San Diego, California. "This is part of a growing body of evidence that many molecules of the immune system have a second set of jobs in the brain," says Lisa Boulanger, a neurobiologist at the University of California, San Diego.
The new study, which appears in the 14 December issue of Cell, began as an attempt to determine whether neural support cells called astrocytes have a role in refining synaptic connections between neurons during development, says senior author Ben Barres of Stanford University in Palo Alto, California. Postdoc Beth Stevens and colleagues used gene chips to look for changes in gene expression in neurons from the developing retinas of rats when the neurons were cultured with astrocytes.

To their surprise, astrocytes spurred the neurons to crank out a complement protein called C1q, which elsewhere in the body kicks off a cascade of chemical events that culminates in the destruction of an intruding cell. In experiments with mice, the researchers found that C1q concentrations in the retina and brain peaked a week or so after birth and dropped dramatically as mice matured. The peak coincided with the period when unwanted synapses are pruned. More intriguing, C1q seemed to concentrate at puny, immature-looking synapses in the developing nervous system.

When the researchers examined the brains of mice lacking a functional C1q gene, they found that development had gone awry in the lateral geniculate nucleus, a relay station in the brain that receives synaptic inputs directly from retinal neurons. In normal mice, geniculate neurons initially receive inputs from both eyes and then prune them so that they only receive input from one eye or the other. In the mutant mice, geniculate neurons maintained extraneous inputs from both eyes into adulthood.

That's a striking finding, Boulanger says: "When you get rid of these proteins that we thought just functioned in the immune system, it disrupts a very specific event that we think is involved in making the precise, final connections in the developing visual system." Many questions remain, however. Barres suspects that complement proteins mark unwanted synapses for removal by microglia, immune cells in the brain. More work is needed to demonstrate that, Boulanger says, and to figure out why only certain synapses are flagged for removal.

Finally, Barres and colleagues collaborated with Simon John's group at the Jackson Laboratory in Bar Harbor, Maine, to investigate whether C1q might have a role in synapse loss in a mouse model of glaucoma. Compared to normal adult mice, adult glaucoma mice exhibited elevated C1q levels: The protein accumulates at retinal synapses early in the disease, even before synapses disappear and neurons die off.

Synapse loss precedes cell death in Alzheimer's and other neurodegenerative diseases, Barres notes. He speculates that drugs that block the complement cascade may forestall neurodegeneration in a number of disorders. It's an exciting idea, says Monica Vetter, a neurobiologist at the University of Utah in Salt Lake City: "There's good evidence that these complement components are upregulated in other diseases."