The result links schizophrenia risk to a problem with a normal process that happens in adolescence and early adulthood, when disease symptoms often appear. That age range is when the brain trims back the number of specialized places on brain cells where the cells signal each other, called synapses. The new work suggests a connection to schizophrenia when this process gets out of hand, deleting too many synapses.
"It's like you have a gardener who was supposed to prune the bushes and just got overactive," Cuthbert observed. "You end up with bushes that are pruned way too much."
The result doesn't mean over-pruning causes schizophrenia on its own. It could promote the disease in combination with other factors in the brain, McCarroll said.
The work began with a genetic investigation. Previous analysis of the human DNA indicates over 100 places that influence the risk of getting schizophrenia, but detailed biological explanations for those influences are very rare. The new work identified a risk gene and found evidence for the over-pruning idea.
Drawing on DNA data from 28,799 people with schizophrenia and 35,986 people without it, the researchers found that a gene called C4 can raise a person's risk by about 30 percent over that of the general population.
The gene comes in several forms, and researchers examining brain tissue found evidence that the forms that pose the most risk of schizophrenia were also the most active in the brain. In lab mice, they found that the gene plays a key role in pruning synapses.
The study doesn't directly demonstrate that excessive pruning of synapses plays a role in schizophrenia, but the idea makes sense, McCarroll said. It ties together previous observations, among them that schizophrenia most often develops during youth and that patients' brains show unusually few synapses, he said.
Dr. Kenneth Kendler, a schizophrenia genetics expert at Virginia Commonwealth University in Richmond who didn't participate in the project, said the work presents an impressive array of results. The evidence that C4 can raise schizophrenia risk is strong, he said. The proposal that it does so through excessive pruning of synapses is "plausible and interesting, but not yet fully convincing," he said.
"We don't yet know (whether) their hypothesis is completely true," Kendler said, but the work is still "a pretty big deal."
If it's true, scientists can think about finding drugs that would intervene, McCarroll said. They might be useful to give when young people show symptoms that suggest they may be on the road to developing schizophrenia, he said. And even after the diagnosis, such drugs might keep the disease from getting worse, he said.
But any such treatments are years away, he cautioned.