Deciphering H3K4me3 broad domains associated with gene-regulatory networks and conserved epigenomic landscapes in the human brain

Regulators of the histone H3-trimethyl lysine-4 (H3K4me3) mark are significantly associated with the genetic risk architecture of common neurodevelopmental disease, including schizophrenia and autism. Typical H3K4me3 is primarily localized in the form of sharp peaks, extending in neuronal chromatin on average only across 500–1500 base pairs mostly in close proximity to annotated transcription start sites. Here, through integrative computational analysis of epigenomic and transcriptomic data based on next-generation sequencing, we investigated H3K4me3 landscapes of sorted neuronal and non-neuronal nuclei in human postmortem, non-human primate and mouse prefrontal cortex (PFC), and blood. To explore whether H3K4me3 peak signals could also extend across much broader domains, we examined broadest domain cell-type-specific H3K4me3 peaks in an unbiased manner with an innovative approach on 41+12 ChIP-seq and RNA-seq data sets. In PFC neurons, broadest H3K4me3 distribution ranged from 3.9 to 12 kb, with extremely broad peaks (~10 kb or broader) related to synaptic function and GABAergic signaling (DLX1, ELFN1, GAD1, IGSF9B and LINC00966). Broadest neuronal peaks showed distinct motif signatures and were centrally positioned in prefrontal gene-regulatory Bayesian networks and sensitive to defective neurodevelopment. Approximately 120 of the broadest H3K4me3 peaks in human PFC neurons, including many genes related to glutamatergic and dopaminergic signaling, were fully conserved in chimpanzee, macaque and mouse cortical neurons. Exploration of spread and breadth of lysine methylation markings could provide novel insights into epigenetic mechanism involved in neuropsychiatric disease and neuronal genome evolution.