Bridging between Mouse and Human Enhancer-Promoter Long-Range Interactions in Neural Stem Cells, to Understand Enhancer Function in Neurodevelopmental Disease

Abstract

Non-coding variation in complex human disease has been well established by genome-wide association studies, and is thought to involve regulatory elements, such as enhancers, whose variation affects the expression of the gene responsible for the disease. The regulatory elements often lie far from the gene they regulate, or within introns of genes differing from the regulated gene, making it difficult to identify the gene whose function is affected by a given enhancer variation. Enhancers are connected to their target gene promoters via long-range physical interactions (loops). In our study, we re-mapped, onto the human genome, more than 10,000 enhancers connected to promoters via long-range interactions, that we had previously identified in mouse brain-derived neural stem cells by RNApolII-ChIA-PET analysis, coupled to ChIP-seq mapping of DNA/chromatin regions carrying epigenetic enhancer marks. These interactions are thought to be functionally relevant. We discovered, in the human genome, thousands of DNA regions syntenic with the interacting mouse DNA regions (enhancers and connected promoters). We further annotated these human regions regarding their overlap with sequence variants (single nucleotide polymorphisms, SNPs; copy number variants, CNVs), that were previously associated with neurodevelopmental disease in humans. We document various cases in which the genetic variant, associated in humans to neurodevelopmental disease, affects an enhancer involved in long-range interactions: SNPs, previously identified by genome-wide association studies to be associated with schizophrenia, bipolar disorder, and intelligence, are located within our human syntenic enhancers, and alter transcription factor recognition sites. Similarly, CNVs associated to autism spectrum disease and other neurodevelopmental disorders overlap with our human syntenic enhancers. Some of these enhancers are connected (in mice) to homologs of genes already associated to the human disease, strengthening the hypothesis that the gene is indeed involved in the disease. Other enhancers are connected to genes not previously associated with the disease, pointing to their possible pathogenetic involvement. Our observations provide a resource for further exploration of neural disease, in parallel with the now widespread genome-wide identification of DNA variants in patients with neural disease.

Hooked Up from a Distance: Charting Genome-Wide Long-Range Interaction Maps in Neural Cells Chromatin to Identify Novel Candidate Genes for Neurodevelopmental Disorders

“…The identified connected enhancers were mostly active (14/15 connected enhancers tested) in transgenic zebrafish and mouse, directing the expression of reporter genes to the developing forebrain; this indicated the ability of RNApolII ChIA-PET to identify functional enhancers active in the brain in vivo . Further, the vast majority of the connected enhancers, identified in NSC by RNApolII ChIA-PET, had a counterpart in epigenetic enhancers carrying active enhancer marks in the adult brain [ 31 ], as defined by overlap between connected RNApolII ChIA-PET enhancers and epigenetic enhancers (candidate cis-regulatory elements, cCREs) previously defined by Li et al [ 32 ]. Out of about 10,000 mouse interactions involving epigenetic enhancers (on one “anchor”) and gene promoters (on the other “anchor”), about 7500 were conserved in humans [ 31 ].…”

“…Further, the vast majority of the connected enhancers, identified in NSC by RNApolII ChIA-PET, had a counterpart in epigenetic enhancers carrying active enhancer marks in the adult brain [ 31 ], as defined by overlap between connected RNApolII ChIA-PET enhancers and epigenetic enhancers (candidate cis-regulatory elements, cCREs) previously defined by Li et al [ 32 ]. Out of about 10,000 mouse interactions involving epigenetic enhancers (on one “anchor”) and gene promoters (on the other “anchor”), about 7500 were conserved in humans [ 31 ]. For these interactions, both the enhancer and promoter regions could be re-mapped onto a syntenic region of the human genome; in addition, most human enhancers and promoters carried epigenetic marks of activity (H3K27Ac; H3K4me1) also in human neural cells, and were connected in interaction maps [ 31 ].…”

“…Out of about 10,000 mouse interactions involving epigenetic enhancers (on one “anchor”) and gene promoters (on the other “anchor”), about 7500 were conserved in humans [ 31 ]. For these interactions, both the enhancer and promoter regions could be re-mapped onto a syntenic region of the human genome; in addition, most human enhancers and promoters carried epigenetic marks of activity (H3K27Ac; H3K4me1) also in human neural cells, and were connected in interaction maps [ 31 ]. The identified connected enhancers were subsequently overlapped with DNA sequence variants associated to NDD, of both the SNV and CNV type [ 31 ].…”

“…For these interactions, both the enhancer and promoter regions could be re-mapped onto a syntenic region of the human genome; in addition, most human enhancers and promoters carried epigenetic marks of activity (H3K27Ac; H3K4me1) also in human neural cells, and were connected in interaction maps [ 31 ]. The identified connected enhancers were subsequently overlapped with DNA sequence variants associated to NDD, of both the SNV and CNV type [ 31 ]. The overlap with CNV identified in ASD patients showed that microdeletions could overlap with enhancers, connected, in turn, to distant genes (not involved, themselves, in the deletion).…”

“…Further, many SNV, previously associated with SCZ, bipolar disorder (BD) and intelligence, were found to overlap with connected enhancers identified by ChIA-PET; while some were connected to genes, already identified as NDD contributors by mutations in the gene’s protein coding region, others were connected to genes active in the developing nervous system and in neurons, but not previously involved in NDD pathogenesis. Significantly, while some of these genes had already emerged in the studies discussed above, mapping interactions with HiC, about half of the genes were newly identified by the ChIA-PET maps [ 31 ], pointing to the usefulness of different, complementary mapping methods for a thorough charting of chromatin interactions relevant for NDD.…”