Variant-to-function analysis of the childhood obesity chr12q13 locus implicates rs7132908 as a causal variant within the 3’ UTR of FAIM2 (snATAC-Seq)

The ch12q13 obesity locus is among the most significant childhood obesity loci identified in genome-wide association studies. This locus resides in a non-coding region within FAIM2; thus, the underlying causal variant(s) presumably influence disease susceptibility via an influence on cis-regulation within the genomic region. We implicated rs7132908 as a putative causal variant at this locus leveraging a combination of our inhouse 3D genomic data, public domain datasets, and several computational approaches. Using a luciferase reporter assay in human primary astrocytes, we observed allele-specific cis-regulatory activity of the immediate region harboring rs7132908. Motivated by this finding, we went on to generate isogenic human embryonic stem cell lines homozygous for either rs7132908 allele with CRISPR-Cas9 homology-directed repair to assess changes in gene expression due to genotype and chromatin accessibility throughout a differentiation to hypothalamic neurons, a key cell type known to regulate feeding behavior. We observed that the rs7132908 obesity risk allele influenced the expression of FAIM2 along with other genes, decreased the proportion of neurons produced during differentiation, up-regulated cell death gene sets, and conversely down-regulated neuron differentiation gene sets. We have therefore functionally validated rs7132908 as a causal obesity variant which temporally regulates nearby effector genes at the ch12q13 locus and influences neurodevelopment and survival. To investigate the cis-regulatory effects of rs7132908 genotype on chromatin accessibility, we established H9 ESC lines homozygous for the rs7132908 obesity risk A allele with CRISPR-Cas9 homology-directed repair while the parent line is homozygous for the non-risk G allele. We performed single-nucleus ATAC-seq on hypothalamic cells derived from these ESCs after 40 days of differentiation. We identified cell types found in our data using paired single-nucleus RNA-seq profiles and then detected significantly differentially accessible peaks due to rs7132908 genotype in each cell type. We also identified transcription factor motifs significantly enriched in each differentially accessible peak.