Abstract
Bipolar disorder (BD) is a highly heritable mood disorder with intermittent episodes of mania and depression. Lithium is the first-in-line medication to treat BD, but it is only effective in a subset of individuals. Large-scale human genomic studies have repeatedly linked the ANK3 gene (encoding ankyrin-G, AnkG) to BD. Ank3 knockout mouse models mimic BD behavioral features and respond positively to lithium treatment. We investigated cellular phenotypes associated with BD, including dendritic arborization of pyramidal neurons and spine morphology in two models: (1) a conditional knockout mouse model which disrupts Ank3 expression in adult forebrain pyramidal neurons, and (2) an AnkG knockdown model in cortical neuron cultures. We observed a decrease in dendrite complexity and a reduction of dendritic spine number in both models, reminiscent of reports in BD. We showed that lithium treatment corrected dendrite and spine deficits in vitro and in vivo. We targeted two signaling pathways known to be affected by lithium using a highly selective GSK3β inhibitor (CHIR99021) and an adenylate cyclase activator (forskolin). In our cortical neuron culture model, CHIR99021 rescues the spine morphology defects caused by AnkG knockdown, whereas forskolin rescued the dendrite complexity deficit. Interestingly, a synergistic action of both drugs was required to rescue dendrite and spine density defects in AnkG knockdown neurons. Altogether, our results suggest that dendritic abnormalities observed in loss of function ANK3 variants and BD patients may be rescued by lithium treatment. Additionally, drugs selectively targeting GSK3β and cAMP pathways could be beneficial in BD.