Unusually high levels of the signaling peptide BDNF, or brain-derived neurotrophic factor, have been detected in blood samples from children with autism. Barbara Hempstead of Weill Medical College at Cornell University and her colleagues propose that BDNF may also be over-expressed in the brains of these children, causing neurological defects that lead to the disorder.

A well-characterized microdeletion in the 22q11.2 chromosomal region is known to be associated with schizophrenia. There is growing evidence that duplications at this locus are also associated with autism. Joseph Gogos and his colleagues at Columbia University in New York have been studying the chain of molecular and cellular events arising from copy number variants (CNVs) at this locus, which may lead to altered brain connectivity.

The relationship between genetics and autism is not always straightforward, but some autism spectrum disorders are known to be caused by defects in a single gene. These simpler cases give researchers the opportunity to create animal models with the genetic defect and use them to test hypotheses about the mechanisms at work in autism.

Current genetic evidence links autism with defects in signaling between brain cells, or neurons, particularly at specialized regions of the neurons known as synapses. Bernardo Sabatini and his colleagues at Harvard Medical School plan to study how neuron activity affects synapse formation, focusing on the roles of three autism-associated proteins, which are members of the Shank, neurexin and neuregulin families.

Most of the genes known to be associated with autism seem to regulate how neurons communicate with each other at structures called synapses. Li-Huei Tsai and her colleagues at the Massachusetts Institute of Technology study the role of CDK5, an enzyme involved in synapse activity, and its influence on SHANK3, another protein known to be important in autism.

Many children with autism have unusually high numbers of synapses, or connections between neurons, particularly in the cortex, which may result from overgrowth and a disruption of neuronal pruning during childhood. Pruning and reshaping of neurons pares down the number of synapses in the brain while eliminating inappropriate synapses that lead to over-connectivity between brain regions, and possibly inappropriate learning, behavior and seizures. David Sulzer and his colleagues at Columbia University hypothesize that autism-associated mutations in the tuberous sclerosis gene, TSC, can cause over-connectivity when the target of TSC, the mTOR pathway, interferes with normal neuronal pruning.

The neurobiology behind the classic signs of autism — repetitive behaviors, impaired social interactions and language deficits — is largely unknown. Based on their studies on a mouse model of obsessive-compulsive disorder, Guoping Feng and his colleagues at Duke University propose that an imbalance between two circuits of neurons may underlie the repetitive behaviors associated with autism.

Autism arises in early childhood, during a period of intense learning when many of the brain’s connections are modified by experience. Eric Kandel, Yun-Beom Choi and Craig Bailey of Columbia University are using animal models of memory formation to investigate how two autism-associated proteins — neuroligin and neurexin — regulate this process.

Like autism, Rett syndrome arises in young children with a progressive loss of skills such as speech and control of movements, and is frequently accompanied by mental retardation and seizures. Mutations in the gene MECP2 are known to cause Rett syndrome. Josh Huang and his colleagues at Cold Spring Harbor Laboratory plan to study how the MECP2 gene regulates brain circuitry — information that may have implications for both Rett syndrome and autism.

Genetic studies have linked several new genes to increased risk of autism. The challenge, however, is in understanding how mutations in these genes disrupt neural development and cause the disorder. Ted Dawson and his colleagues at Johns Hopkins University plan to undertake this task for the signaling protein known as the MET receptor.