Sekar and McCarroll soon launched a series of collaborative experiments with Beth Stevens to determine how C4A might be involved in synapse editing, and how synapse editing, in turn, might be linked to schizophrenia. They were joined by Mike Carroll, an immunologist, who had long studied C4’s role in immune diseases. In human-brain tissues and in neurons cultured in flasks, they found that the C4 protein accumulated abundantly at synapses; in mice, this accumulation occurred almost precisely at the time that pruning begins. Mice that lacked the C4 gene underpruned the synapses in parts of their brain, which suggested a direct connection between the gene and pruning. Sekar confirmed that the C4A was more abundant in the brains of schizophrenic patients than in normal brains. The increase in C4A levels in schizophrenic patients was most significant in the parts of the brain involved in cognition, planning, and thinking, the functions that are most impaired in people with schizophrenia, and less noticeable in parts of the brain that control balance, posture, and speech, aspects that remain relatively intact in those with the disease.

A magnificently simple theory began to convulse out of the results. Perhaps C4A, like the other immunological factors that Stevens had identified in synapse pruning, marks neuronal synapses destined to be eliminated during normal brain development. During the maturation of the brain, microglia recognize these factors as tags and engulf the tagged synapses. Variations in the C4A gene cause different amounts of the C4A protein to be expressed in the human brain. The overabundance of C4A protein in some people contributes to an excessively exuberant pruning of synapses—thereby decreasing the number of synapses in the brain, which would explain the well-established fact that schizophrenic patients tended to have fewer neuronal connections. That the symptoms of schizophrenia break loose during the second and third decades of life makes sense, in retrospect: adolescence and early adulthood are periods when synaptic pruning reaches a climax in the regions of the brain that govern planning and thinking.

Schizophrenia, as McCarroll put it, “may be a disease of overpruning.” Synapses that should have been preserved get cut, like a garden that has been sheared back too aggressively in the winter. “The C4A paper is one of the most important papers in schizophrenia in our times, because it identifies a pathway and provides a mechanism,” Lieberman said. “It opens a black box. Now we have to figure out how overpruned synapses cause all the diffuse symptoms of the disease—the psychosis, the cognitive collapse, the emotional emptiness, and the withdrawal.”

Runs in the Family New findings about schizophrenia rekindle old questions about genes and identity.