Genetic effects on the skin methylome in healthy older twins

Abstract

Whole-skin DNA methylation variation has been implicated in several diseases, including melanoma, but its genetic basis has not yet been fully characterized. Using bulk skin tissue samples from 414 healthy female UK twins, we performed twin-based heritability and methylation quantitative trait loci (meQTL) analyses for >400,000 DNA methylation sites. We find that the human skin DNA methylome is on average less heritable than previously estimated in blood and other tissues (mean heritability: 10.02%). meQTL analysis identified local genetic effects influencing DNA methylation at 18.8% (76,442) of tested CpG sites, as well as 1,775 CpG sites associated with at least one distal genetic variant. As a functional follow-up, we performed skin expression QTL (eQTL) analyses in a partially overlapping sample of 604 female twins. Colocalization analysis identified over 3,500 shared genetic effects affecting thousands of CpG sites (10,067) and genes (4,475). Mediation analysis of putative colocalized gene-CpG pairs identified 114 genes with evidence for eQTL effects being mediated by DNA methylation in skin, including in genes implicating skin disease such as ALOX12 and CSPG4. We further explored the relevance of skin meQTLs to skin disease and found that skin meQTLs and CpGs under genetic influence were enriched for multiple skin-related genome-wide and epigenome-wide association signals, including for melanoma and psoriasis. Our findings give insights into the regulatory landscape of epigenomic variation in skin.

Keywords: DNA methylation; QTLs; gene expression; heritability; skin.

Figure 1

Heritability of the skin DNA methylome (A) Proportion of DNA methylation variance explained for each tested CpG site by narrow-sense heritability (a²), common environmental effects between co-twins (c²), and environmental effects not shared between co-twins (e²). Y axis is cut-off at 25,000-count. (B) Cumulative distribution of narrow-sense heritability for CpG sites in different genetic contexts. UTR, untranslated region; TSS1500, 200—1,500 bp upstream of gene transcription start site (TSS); TSS200, $<$200 bp upstream of gene TSS. (C) Cumulative distribution of narrow-sense heritability in variable CpG sites (SD $>$ 0.05) vs. non-variable CpG sites. (D) Combined r² of all independent skin meQTL effects per CpG vs. estimated narrow-sense heritability in skin (purple). The number of CpG sites in each binned group is shown in gray.

Figure 2

Genetic effects on the whole-skin methylome (A) Distance between lead cis-meQTL and CpG vs. strength of association of the meQTL effect. (B) Most associated cis-meQTL association between variant rs1253098 (chr14:59945536) and cg01543583 (chr14:59947673). (C) Most associated trans-meQTL association between the variant rs10744202 (chr12:125800244) and cg03923277 (chr12:104359732). (D) Enrichment and depletion of meQTL effects in different genomic contexts. Green indicates significant (FDR $<$ 0.05) enrichment, purple indicates significant depletion. CpG islands based on UCSC refGene annotations (CpG shore, $<$2 kbp from CpG island; CpG shelf, 2–4 kbp from CpG island; open sea, $>$4 kbp from CpG island), and gene context is based on refGene annotations (TSS200, $<$200 bp from transcription start site (TSS); TSS1500, 200–1,500 bp from TSS). (E) Enrichment and depletion of meQTL effects in transcription factor binding sites (TFBSs) in suprapubic skin and genomic states in leg skin generated using an 18-state ChromHMM model from the Roadmap Epigenomics Consortium. White asterisks indicate significant depletion, black asterisks indicate significant enrichment. Suprapubic skin TFBS and leg skin chromatin state annotations were obtained from the EpiMap Repository (TssA, active TSS; TssBiv, bivalent/poised TSS; TssFlnk, flanking active TSS; TssFlnkU, flanking TSS upstream; TssFlnkD, flanking TSS downstream; Tx, strong transcription; TxWk, weak transcription; EnhA1, active enhancers 1; EnhA2, active enhancers 2; EnhG1, genic enhancers 1; EnhG2, genic enhancers 2; EnhBiv, bivalent enhancer; EnhWk, weak enhancer; ReprPC, repressed polycomb; ReprPCWk, weak repressed polycomb; ZNF_Rpts, ZNF genes and repeats; Quies, quiscent chromatin; Het, heterochromatin; see Kundaje et al.⁵⁹).

Figure 3

Colocalization and mediation between cis-CpGs and cis-eGenes (A) The proportion of cis-CpGs and cis-eGenes associated with at least one colocalizing QTL and the proportion of cis-meQTLs and cis-eQTLs that colocalize. (B) The proportion of cis-CpGs and cis-eGenes with at least one colocalized QTL that also show evidence of mediation via SEM or SME. (C) Proportion of effect mediated versus significance of the mediation effect (-log10(Sobel test p value). Only results with a positive proportion of the effect mediated are shown. Mediation analyses under the SME model tended to have a larger proportion of effect mediated than results from the SEM model. Purple and yellow dashed lines indicate the FDR 5% threshold for SEM and SME Sobel mediation tests, respectively. (D) SME and SEM pathways.

Figure 4

Enrichment of skin meQTL-CpG effects in loci from published epi(genome)-wide association studies of skin-related traits Left panel shows enrichment of cis-CpGs (CpG sites associated with a meQTL in cis) to overlap signals from skin-phenotype EWASs. Right panel shows enrichment of cis-meQTLs to be in GWAS signals for skin-related phenotypes. Green indicates significant (FDR $<$ 0.05) enrichment. GWAS enrichment analyses are performed using all SNPs associated with a CpG site in cis. Studies of the same phenotype have been pooled together. Biological sample types for EWASs is denoted in the EWAS trait label (s, skin; o, other). SCC, squamous cell carcinoma; CSS, cutanous systemic sclerosis; PhenoAge AA, PhenoAge age acceleration; DNAm intrinsic AA, DNAm intrinisic age acceleration; GrimAge AA, GrimAge age acceleration; DNAm Hannum, Hannum DNAm age acceleration; CM, cutaneous melanoma; MCID; multiple chronic inflammatory disease; SJS, Stevens-Johnson syndrome; TEN, toxic epidermal necrolysis; Allergy 1, asthma, hay fever, or eczema; Allergy 2, age of onset of asthma, hay fever, and/or eczema; Allergy 3, multivariate analysis of asthma, hay fever, and/or eczema.

Figure 5

Three examples of mediating SNP-CpG-gene triplets Methylated CpG sites are represented by a filled black circle, and un-methylated CpG sites are represented by an empty circle. (A) Hypothetical mechanism by which cg05215272 mediates an eQTL effect on ALOX12 expression where ALOX12 has been implicated in melanoma, vitiligo, psoriasis, and atopic dermatitis. (B) Hypothetical mechanism by which cg21460582 mediates an eQTL effect on CSPG4 expression, which is strongly associated in melanoma. (C) Hypothetical mechanism by which CDK10 expression mediates a meQTL effect on cg00001687.