Animal models of bipolar disorder.

Animal models of human diseases should meet three sets of criteria: construct validity, face validity, and predictive validity. To date, several putative animal models of bipolar disorder have been reported. They are classified into various categories: pharmacological models, nutritional models, environmental models, and genetic models. None of them, however, totally fulfills the three validity criteria, and thus may not be useful for drug development. Mounting evidence suggests that mitochondrial dysfunction has a role in bipolar disorder. To test whether accumulation of mtDNA deletions in the brain can cause bipolar disorder, we generated transgenic mice with neuron-specific expression of mutant Polg (D181A). These mice showed altered diurnal activity rhythm and periodic activity change associated with the estrous cycle. These phenotypes were worsened by administration of a tricyclic antidepressant, but improved after lithium treatment. This mouse model of bipolar disorder potentially fulfills the three validity criteria, and therefore might be used for future drug development studies.