Comparative Analyses of Chromatin Landscape in White Adipose Tissue Suggest Humans May Have Less Beigeing Potential than Other Primates

Abstract

Humans carry a much larger percentage of body fat than other primates. Despite the central role of adipose tissue in metabolism, little is known about the evolution of white adipose tissue in primates. Phenotypic divergence is often caused by genetic divergence in cis-regulatory regions. We examined the cis-regulatory landscape of fat during human origins by performing comparative analyses of chromatin accessibility in human and chimpanzee adipose tissue using rhesus macaque as an outgroup. We find that many regions that have decreased accessibility in humans are enriched for promoter and enhancer sequences, are depleted for signatures of negative selection, are located near genes involved with lipid metabolism, and contain a short sequence motif involved in the beigeing of fat, the process in which lipid-storing white adipocytes are transdifferentiated into thermogenic beige adipocytes. The collective closing of many putative regulatory regions associated with beigeing of fat suggests a mechanism that increases body fat in humans.

Keywords: adipose; chromatin accessibility; comparative genomics; primates.

Fig. 1.

—Detection of species-specific OCR state changes. (A) Principal component analysis of OCRs in human, chimpanzee, and rhesus macaque adipose. Note that intraspecific variation is much smaller than interspecific variation. A common OCR state is depicted in (B), where the x-axis represents chromosome coordinates and the y-axis represents OCR intensity from MACS2. Human-specific OCR state changes (red dash) to increased accessibility (C) and to decreased accessibility (D) from ancestral state (i.e., rhesus macaque accessibility). Chimpanzee-specific OCR state changes (red dash) to increased accessibility (E) and to decreased accessibility (F) from ancestral state. Genomic coordinates of the OCR of interest are listed.

Fig. 2.

—Species-specific OCR groups are enriched for cis-regulatory functions. Species-specific OCR groups enrichment (Fisher’s exact, **P < 0.01, ***P < 0.001) for promoters (A) enhancers (B), and adipose eQTLs (C). GREAT enrichment bubble plot (D) with labeled GO terms for bubbles containing at least 25 genes.

Fig. 3.

—Branch-specific positive selection as detected by phyloP. Percentage of OCRs under (A) human and (B) chimpanzee branch-specific positive selection (top panels). Percentage of OCRs under (A) human and (B) chimpanzee branch-specific negative selection (bottom panels, Fisher’s exact test, ***P < 0.001).

Fig. 4.

—Human-decreased OCRs are associated with NFIA. We used MEME-ChIP to identify motifs that are enriched in human-decreased OCRS (A) and gkm-SVM to distinguish species-specific OCRs from null common OCRS. Shown are the receiver/operating curve for human-decrease (B). We scanned OCRs for expanded NFIA (C) and PPARG (D) motifs. We compared NFIA motifs in adipose (E) and fibroblasts (F) OCRs and PPARG motifs in adipose (G), and fibroblasts (H).