Behavioral and neuroanatomical effects of prenatal exposure to valproic acid in the mouse : relevance to autism spectrum disorders

Abstract

Valproic acid (VPA) is a broad-spectrum anticonvulsant and antiepileptic drug and widely used in many neurological conditions and psychiatric disorders as well as in cancer and HIV treatment. However, despite all its many benefits, VPA also has side effects. It is a strong fetal teratogen that can induce congenital malformation and neurodevelopmental problems. Case reports and population studies have revealed that prenatal exposure to VPA is associated with a higher risk of autism in postnatal life. Animal models also have confirmed that VPA can induce autistic-like features in rodents. Yet, there are some questions remaining unanswered by existing animal studies of prenatal VPA exposure. The embryotoxicity of a drug is not only determined by its own chemical, physiological or pharmacological properties, but also on the dose and the time in development that the exposure happens. The majority of studies investigating the behavioral, neuroanatomical and physiological impact of VPA in animals have examined early gestational exposure to relatively high doses that can cause significant malformation or loss of offspring. Thus, although there are behavioral alterations considered similar to autistic symptoms in humans, these are found in the offspring that have survived a very toxic insult, leading to problems of interpretation. Low dose exposure has not been widely studied, nor has the impact of VPA exposure late in gestation. Moreover, how the age and sex of offspring influences the phenotypic outcome has rarely been considered. Therefore, in the present series of studies, a battery of behavioral tests and in vivo magnetic resonance imaging (MRI) was used to investigate the postnatal consequences of prenatal exposure to lower doses of VPA in mice in early and late gestation. The effects of VPA were examined in female and male mice at juvenile and adult ages. The main findings were, that low doses of VPA in early or late gestation cause no physical malformation and no gross neurological functional impairments, but induce behavioral abnormalities and neuroanatomical differences related to autism. Generally, VPA-treated mice exhibited lower motor activities and higher anxiety levels in the open field test; dislike of novelty in the novel abject exploration test; higher startle response and sensorimotor gating differences; decreased responses to non-social and social odors in the olfactory test; and volumetric changes in brain structures similar to those found in autism. However, the timepoint of exposure, dose of VPA, sex and age of testing influenced the phenotypic outcome. Although largely neglected in previous studies, late gestation exposure to VPA elicited an autistic phenotype. Surprisingly, given the male bias in autism, female mice were often more ‘sensitive’ to VPA. Although the present studies had some limitations, these experiments confirmed that low dose VPA in pregnancy could trigger behavioral abnormalities and brain anatomical differences in mice that resembled a range of features of autism. Importantly, these behaviors were unconfounded by ‘gross’ neurological or physical abnormalities. Further studies to investigate the cellular mechanisms underlying the low dose VPA phenotype will therefore be helpful to shed light on possible causal pathways with specific relevance to autism.