A systematic review of skin ageing genes: gene pleiotropy and genes on the chromosomal band 16q24.3 may drive skin ageing

Abstract

Skin ageing is the result of intrinsic genetic and extrinsic lifestyle factors. However, there is no consensus on skin ageing phenotypes and ways to quantify them. In this systematic review, we first carefully identified 56 skin ageing phenotypes from multiple literature sources and sought the best photo-numeric grading scales to evaluate them. Next, we conducted a systematic review on all 44 Genome-wide Association Studies (GWAS) on skin ageing published to date and identified genetic risk factors (2349 SNPs and 366 genes) associated with skin ageing. We identified 19 promising SNPs found to be significantly (p-Value < 1E-05) associated with skin ageing phenotypes in two or more independent studies. Here we show, using enrichment analyses strategies and gene expression data, that (1) pleiotropy is a recurring theme among skin ageing genes, (2) SNPs associated with skin ageing phenotypes are mostly located in a small handful of 44 pleiotropic and hub genes (mostly on the chromosome band 16q24.3) and 32 skin colour genes. Since numerous genes on the chromosome band 16q24.3 and skin colour genes show pleiotropy, we propose that (1) genes traditionally identified to contribute to skin colour have more than just skin pigmentation roles, and (2) further progress towards understand the development of skin pigmentation requires understanding the contributions of genes on the chromosomal band 16q24.3. We anticipate our systematic review to serve as a hub to locate primary literature sources pertaining to the genetics of skin ageing and to be a starting point for more sophisticated work examining pleiotropic genes, hub genes, and skin ageing phenotypes.

Figure 1

Workflow for screening through publications to include in the review.

Figure 2

(a) SNPs associated with the phenotype: wrinkles, site: whole face. This is a phenotype under Category A Phenotype—skin wrinkling and sagging-related phenotypes. Pooled Odds Ratio and the 95% Confidence Interval are calculated from both studies which reported significant associations between the SNP and the phenotype. An inconsistency index (I² index) value ≥ 50% and a Het p-Value < 0.05 are considered statistically significant for heterogeneity. Begg’s funnel plots and Egger’s test were used to assess publication bias; Begg’s funnel plots were only drawn for meta-analysis of ≥ 2 studies or cohorts and Egger’s test was only performed for meta-analysis of ≥ 3 studies or cohorts to estimate asymmetry of data; an Egger’s test p-Value < 0.05 was considered as that publication bias possibly exist. (b) SNPs associated with the phenotype: skin type—inability to tan, site: whole body. This is a phenotype under Category B Phenotype—skin pigmentation-related phenotypes. Pooled Odds Ratio and the 95% Confidence Interval are calculated from both studies which reported significant associations between the SNP and the phenotype. An inconsistency index (I² index) value ≥ 50% and a Het p-Value < 0.05 are considered statistically significant for heterogeneity. Begg’s funnel plots and Egger’s test were used to assess publication bias; Begg’s funnel plots were only drawn for meta-analysis of ≥ 2 studies or cohorts and Egger’s test was only performed for meta-analysis of ≥ 3 studies or cohorts to estimate asymmetry of data; an Egger’s test p-Value < 0.05 was considered as that publication bias possibly exist. (c) SNPs associated with the phenotype: skin colour, site: upper inner arm. This is a phenotype under Category B Phenotype—skin pigmentation-related phenotypes. Pooled Odds Ratio and the 95% Confidence Interval are calculated from both studies which reported significant associations between the SNP and the phenotype. An inconsistency index (I² index) value ≥ 50% and a Het p-Value < 0.05 are considered statistically significant for heterogeneity. Begg’s funnel plots and Egger’s test were used to assess publication bias; Begg’s funnel plots were only drawn for meta-analysis of ≥ 2 studies or cohorts and Egger’s test was only performed for meta-analysis of ≥ 3 studies or cohorts to estimate asymmetry of data; an Egger’s test p-Value < 0.05 was considered as that publication bias possibly exist. (d) Forest plot for all skin ageing phenotypes associated with rs258322 under the additive allele model. Pooled Odds Ratios (OR) and 95% Confidence Intervals (CI) under the random effects model were computed for each phenotype with ≥ 2 studies or cohorts. Effect alleles were streamlined as allele G; in cases where the primary literature reports the effect allele to be allele A, reciprocals of the OR and CI reported in the primary literature were taken where appropriate to ensure compatibility. (e) Begg’s funnel plots and Egger’s test p-Values for rs258322 and skin ageing.

Figure 2

(a) SNPs associated with the phenotype: wrinkles, site: whole face. This is a phenotype under Category A Phenotype—skin wrinkling and sagging-related phenotypes. Pooled Odds Ratio and the 95% Confidence Interval are calculated from both studies which reported significant associations between the SNP and the phenotype. An inconsistency index (I² index) value ≥ 50% and a Het p-Value < 0.05 are considered statistically significant for heterogeneity. Begg’s funnel plots and Egger’s test were used to assess publication bias; Begg’s funnel plots were only drawn for meta-analysis of ≥ 2 studies or cohorts and Egger’s test was only performed for meta-analysis of ≥ 3 studies or cohorts to estimate asymmetry of data; an Egger’s test p-Value < 0.05 was considered as that publication bias possibly exist. (b) SNPs associated with the phenotype: skin type—inability to tan, site: whole body. This is a phenotype under Category B Phenotype—skin pigmentation-related phenotypes. Pooled Odds Ratio and the 95% Confidence Interval are calculated from both studies which reported significant associations between the SNP and the phenotype. An inconsistency index (I² index) value ≥ 50% and a Het p-Value < 0.05 are considered statistically significant for heterogeneity. Begg’s funnel plots and Egger’s test were used to assess publication bias; Begg’s funnel plots were only drawn for meta-analysis of ≥ 2 studies or cohorts and Egger’s test was only performed for meta-analysis of ≥ 3 studies or cohorts to estimate asymmetry of data; an Egger’s test p-Value < 0.05 was considered as that publication bias possibly exist. (c) SNPs associated with the phenotype: skin colour, site: upper inner arm. This is a phenotype under Category B Phenotype—skin pigmentation-related phenotypes. Pooled Odds Ratio and the 95% Confidence Interval are calculated from both studies which reported significant associations between the SNP and the phenotype. An inconsistency index (I² index) value ≥ 50% and a Het p-Value < 0.05 are considered statistically significant for heterogeneity. Begg’s funnel plots and Egger’s test were used to assess publication bias; Begg’s funnel plots were only drawn for meta-analysis of ≥ 2 studies or cohorts and Egger’s test was only performed for meta-analysis of ≥ 3 studies or cohorts to estimate asymmetry of data; an Egger’s test p-Value < 0.05 was considered as that publication bias possibly exist. (d) Forest plot for all skin ageing phenotypes associated with rs258322 under the additive allele model. Pooled Odds Ratios (OR) and 95% Confidence Intervals (CI) under the random effects model were computed for each phenotype with ≥ 2 studies or cohorts. Effect alleles were streamlined as allele G; in cases where the primary literature reports the effect allele to be allele A, reciprocals of the OR and CI reported in the primary literature were taken where appropriate to ensure compatibility. (e) Begg’s funnel plots and Egger’s test p-Values for rs258322 and skin ageing.

Figure 3

Visual representation of the pleiotropic diversity of genes on the chromosomal band 16q24.3. Each gene shown in this figure is associated with morphologically distinct phenotypes from different phenotypic categories (e.g., Phenotype Category A, B, C, and D colour coded in yellow, blue, orange, and green respectively).

Figure 4

Functional enrichment analysis of skin ageing genes. Each gene shown in this figure is associated with two or more morphologically distinct phenotypes from different phenotypic categories (e.g., Phenotype Category A, B, C, and D). Nodes coloured red represent genes on the chromosomal band 16q24.3 and nodes coloured green represent skin colour genes. Edges (i.e., lines) linking the node indicate evidence of enrichment between the two connected nodes.

Figure 5

Network enrichment of skin ageing hub proteins. Only one strong network (i.e., Network score ≥ 4) is identified for each key skin ageing phenotype. Red gene nodes are genes on the chromosomal band 16q24.3 and yellow gene nodes are skin colour genes. Edges linking the node indicate evidence of enrichment between the two connected nodes.

Figure 6

Functional enrichment of skin ageing genes reveals that skin colour genes are pleiotropic for morphologically diverse (Category A and B) skin ageing phenotypes.

Figure 7

Functional enrichment of skin ageing genes reveals that genes on the chromosomal band 16q24.3 are pleiotropic for morphologically diverse (Category A, B, C, and D) skin ageing phenotypes.

Figure 8

Gene expression of the pleiotropic and hub gene SPIRE2 is downregulated in non-infant skin, aged skin, and melanoma skin, when compared against expression levels in infant skin, young skin, and non-melanoma control skin respectively. Melanoma is a skin ageing phenotype. Expression ratio is the ratio of the average expression level in the cases (e.g., aged skin tissue) against the average expression level of the reference group (e.g., young skin tissue). Ratios above 1 indicate higher expression levels in the cases group while ratios below 1 indicate lower expression levels in the cases group.