Autism is a neurodevelopmental disorder for which diagnosis is based on three domains of behavioral symptoms: (1) abnormal social interactions, (2) impaired communication, and (3) repetitive behaviors. Currently, the only treatments that effectively improve these core symptoms are behavioral interventions implemented at early ages (Vismara and Rogers, 2010). Although pharmacological treatments are available for associated symptoms, including self-injury, tantrums, aggression, and seizures, considerable research is needed to discover pharmacological targets for the diagnostic domains.

Mouse models of autism spectrum disorders provide translational research tools for understanding the causes of autism spectrum disorders and for developing treatments (Ehninger et al, 2008; Silverman et al, 2010). We are interested in the mechanisms that underlie improvements in autism-relevant behavioral phenotypes in genetic mouse models. Given the effectiveness of early behavioral therapies for reducing symptoms in autism, we reasoned that behavioral interventions might similarly rescue social and/or repetitive abnormalities in mouse models. To test this hypothesis, we used an inbred strain of mice, BTBR T+tf/J (BTBR), which displays low sociability on multiple social tasks, reduced ultrasonic vocalizations and behavioral responses to olfactory social cues, and high-repetitive self-grooming. BTBR were given opportunities to interact with C57BL/6J (B6), an inbred strain with high-sociability and low-repetitive behaviors.

First, we attempted to model behavioral interventions given by caretakers to young autistic children. Newborn BTBR were cross-fostered with B6 mothers. Cross-fostering produced no significant effects on social or repetitive behaviors in B6 and BTBR mice, either at juvenile or adult ages. B6 and BTBR raised by dams of the opposite strain showed behaviors similar to those raised by foster dams of the same strain and those raised by their biological mothers (Yang et al, 2007). This finding is consistent with the rejection of the early ‘refrigerator mother’ explanation of autism.

Second, we modeled peer interventions in older children and adolescents with autism (Reichow and Volkmar, 2010). Juvenile BTBR were reared with juvenile B6, beginning at weaning. BTBR who lived with B6 cagemates during juvenile ages developed high sociability as adults, whereas control BTBR who lived with BTBR cagemates continued to show social deficits (Yang et al, 2011).

Third, we are now engaged in understanding the specific behaviors occurring between BTBR and B6 juveniles in their shared home cages, which might lead to improved sociability in BTBR adults. Video recordings of home cages during the dark phase, when mice are awake and interactive, are being scored on measures including social investigation, proximity states, aggressive interactions, and activity levels. Preliminary observations suggest that BTBR housed with B6 cagemates receive more social investigation than BTBR housed with BTBR cagemates. It is possible that increased exposure to social solicitation behaviors as juveniles may be facilitating the adult sociability seen in BTBR reared with B6 cagemates.

Animal models of autism will need to meet the standard criteria of face validity (analogous symptoms, such as social deficits), construct validity (analogous causes, such as genetic mutations), and predictive validity (analogous responses to treatments). Evidence that an early behavioral intervention rescued adult sociability in BTBR mice gives credence to the predictive value of this mouse model.