Repeated mutation of a developmental enhancer contributed to human thermoregulatory evolution

Abstract

Humans sweat to cool their bodies and have by far the highest eccrine sweat gland density among primates. Humans' high eccrine gland density has long been recognized as a hallmark human evolutionary adaptation, but its genetic basis has been unknown. In humans, expression of the Engrailed 1 (EN1) transcription factor correlates with the onset of eccrine gland formation. In mice, regulation of ectodermal En1 expression is a major determinant of natural variation in eccrine gland density between strains, and increased En1 expression promotes the specification of more eccrine glands. Here, we show that regulation of EN1 has evolved specifically on the human lineage to promote eccrine gland formation. Using comparative genomics and validation of ectodermal enhancer activity in mice, we identified a human EN1 skin enhancer, hECE18. We showed that multiple epistatically interacting derived substitutions in the human ECE18 enhancer increased its activity compared with nonhuman ape orthologs in cultured keratinocytes. Repression of hECE18 in human cultured keratinocytes specifically attenuated EN1 expression, indicating this element positively regulates EN1 in this context. In a humanized enhancer knock-in mouse, hECE18 increased developmental En1 expression in the skin to induce the formation of more eccrine glands. Our study uncovers a genetic basis contributing to the evolution of one of the most singular human adaptations and implicates multiple interacting mutations in a single enhancer as a mechanism for human evolutionary change.

Keywords: Engrailed 1; eccrine gland; human evolution; regulatory evolution; sweat.

Fig. 1.

Identification of Engrailed 1 candidate enhancer ECE18. (A) Strategy to identify putative developmental En1 enhancers. Staining for En1 mRNA (purple) in mouse volar limb skin at P2.5 is shown The basal keratinocyte layer (arrow) and eccrine placode (double arrow) are shown. (B) The location in mm10 of ECEs tested in vivo (vertical gray lines). ECE18 is in red and boxed and other positive ECEs are highlighted in orange. (C) Representative images of mouse, chimpanzee, and hECE18 transgenic P2.5 volar limb stained with anti-GFP antibody. eGFP (black arrow) is visualized using HRP-DAB coupled immunohistochemistry. (D) Sequence alignment and evolutionarily features of ECE18, which contains four conserved elements by phastCons (blue rectangles) and overlaps the human accelerated regions HACNS56, HAR19, and HAR80 (collectively called 2xHAR20).

Fig. 2.

Repeated mutation of ECE18 produced human-specific gains in enhancer activity. (A) Fold induction of luciferase reporter activity relative to empty vector (Control) by ECE18 orthologs in cultured human keratinocytes. (B) Schematic of ECE18 variants tested in C. Derived human bases at A and B are highlighted in red, and the positions of the remaining derived human-specific base substitutions are shown as dashed vertical lines. (C) Localization of hECE18 enhancer activity. The fold change normalized luciferase activities of full-length hECE18, chimpanzee ECE18 (cECE18), human 2xHAR20, hECE18 fragment (Frag) A and B, and mutant hECE18 in which the indicated derived base is replaced by the ancestral ape base are plotted. Mutated base (Mut). All 10 derived human substitutions are mutated to ancestral ape bases in hECE18^MutA−J. (D) Species alignment and predicted binding affinities for SP1 at SP1^A and SP1^B. Derived human bases at A and B are in red. Dots indicate identity to human base. Blue arrows indicate the location of ChIP-qPCR primers used in E. (E) Enrichment of SP1 by ChIP-qPCR at hECE18 interval containing SP1^A and SP1^B motifs (SP1^{A, B}) in human keratinocytes. Human γ-globin (HBG2) promoter is used as a negative control, and IgG or SP1 enrichment over the input for each set of primers is shown. Mean enrichment is reported across three independent experiments (line). (F) Fold induction of luciferase reporter by hECE18^FragA upon deletion of SP1^A and SP1^B. (G) Fold induction of luciferase reporter by mouse ECE18^FragA (mECE18^FragA) in human keratinocytes upon knock-in of human SP1^A and SP1^B motifs. Normalized Firefly luciferase activity is plotted as the fold change relative to Control (empty reporter vector). Firefly luciferase values are normalized to Renilla luminescence. The dots represent an individual biological replicate. Median (line), box (bounds 25 to 75%) and whiskers (min and max) plotted. Significance by one-way ANOVA. Tukey-adjusted P values are reported in heatmaps. Assays are performed in human GMA24F1A keratinocytes.

Fig. 3.

hECE18 positively regulates Engrailed 1 to promote eccrine gland formation. (A) Fold change normalized EN1 mRNA by qRT-PCR in human GMF24F1A cultured keratinocytes upon dCAS9-KRAB–mediated hECE18 repression. The schematic of strategy and relative positions of EN1, SP1^A and SP1^B, and Cr1 and Cr2 gRNA targets is shown. The fold change calculated relative to dCAS9-KRAB transduction alone is shown. (B) Fold luciferase induction relative to empty reporter vector (Control) by ECE18 orthologs in primary mouse keratinocytes. (C) Strategy to generate hECE18 knock-in (hECE18^KI) mouse model. (D) Fold change in En1 mRNA by qRT-PCR in P2.5 volar forelimb skin of wild-type (^+/+), hECE18^KI heterozygote (hECE18^KI/+), and hECE18^KI homozygote (hECE18^KI/hECE18^KI) mice relative to wild type. (E) Normalized ratio of C57BL/6J:FVB/N En1 allelic expression in volar forelimb skin of wild-type (C57BL/6J/FVB/N) and hECE18^KI (C57BL/6J^(hECE18KI)/FVB/N) hybrid mice. The ratios were normalized to genomic DNA allelic ratio. (F) Representative stained epidermal preparations of volar forelimb skin from En1^KO/+;^+/+ and En1^KO/+;+/hECE18^KI adult mice. The number of eccrine glands in the IFP area (outlined) and excluding the footpads (FP, circled) were quantified in analyses in G. Hair follicle (HF, *), eccrine gland (EG). (G) Quantification of IFP eccrine glands in En1 ^KO/+;^+/+ and En1 ^KO/+;+/hECE18^KI mice. Each point represents the average number of eccrine glands in the IFP across both forelimbs of a mouse. (A, B, G) Median (line), 25 to 75% percentiles (box bounds) and min and max (whiskers) are plotted. (D and E) Genotype mean is shown as a line and each point represents a single biological sample of pooled volar skin from both forelimbs of at least three mice. (A, B, D) significance assessed by one-way ANOVA and Tukey-adjusted P values are reported. (E and G) Significance assessed by a two-tailed t test. ****P < 0.0001, **P < 0.01, *P < 0.05. (+) wild- type allele. (KO) knockout. (KI) knock-in. The dots represent an individual biological replicate.

Fig. 4.

A genetic basis for the evolution of increased eccrine gland density in human skin. ECE18 is a developmental enhancer of the anthropoid Engrailed 1 locus. Over the course of human evolution, the ECE18 enhancer accumulated multiple activity-enhancing single nucleotide mutations, particularly at the SP1 sites SP1^A and SP1^B (boxed and bold), that collectively induced the specification of more eccrine glands in human skin by potentiating developmental EN1 expression. Derived human substitutions relative to other apes underlying higher activity of hECE18 are in red.