Antecedent enhancer activity predicts future susceptibility to seizures in mice.

Wide variation of responses to identical stimuli presented to genetically inbred mice suggests the hypothesis that stochastic non-genetic variation, such as in chromatin state or enhancer activity during neurodevelopment, can mediate such phenotypic differences. However, this hypothesis is largely untested since capturing pre-existing molecular states requires non-destructive, longitudinal recording. Therefore, we tested the potential of Calling Cards (CC) to record transient neuronal enhancer activity during postnatal development in mice, and thereby associate such non-genetic variation with a subsequent phenotypic presentation - degree of seizure response to the pro-convulsant pentylenetetrazol. We show that recorded differences in enhancer activity at 243 loci predict a severe vs. mild response, and that these are enriched near genes associated with human epilepsy. We also validated pharmacologically a seizure-modifying role for two previously unassociated genes, Htr1f and Let7c. This proof-of-principle supports using CC broadly to discover predisposition loci for other neuropsychiatric traits and behaviors. Finally, as human disease is also influenced by non-genetic factors, similar epigenetic predispositions are possible in humans.