Transcriptome analysis of human ageing in male skin shows mid-life period of variability and central role of NF-κB

Abstract

Age is well-known to be a significant factor in both disease pathology and response to treatment, yet the molecular changes that occur with age in humans remain ill-defined. Here, using transcriptome profiling of healthy human male skin, we demonstrate that there is a period of significantly elevated, transcriptome-wide expression changes occurring predominantly in middle age. Both pre and post this period, the transcriptome appears to undergo much smaller, linear changes with increasing age. Functional analysis of the transient changes in middle age suggest a period of heightened metabolic activity and cellular damage associated with NF-kappa-B and TNF signaling pathways. Through meta-analysis we also show the presence of global, tissue independent linear transcriptome changes with age which appear to be regulated by NF-kappa-B. These results suggest that aging in human skin is associated with a critical mid-life period with widespread transcriptome changes, both preceded and proceeded by a relatively steady rate of linear change in the transcriptome. The data provides insight into molecular changes associated with normal aging and will help to better understand the increasingly important pathological changes associated with aging.

Figure 1. Expression of the transcriptome of human skin across the lifespan.

(a) 174 genes undergo a gradual, progressive increase in expression with age. (b) The entire transcriptional profile of normal human skin from 98 individuals relative to age groups. (c) 2969 genes demonstrate a statistically significant transient peak in expression at age 30–45. (d) 556 genes show an opposing transient decrease in expression at age 30–45.

Figure 2

(a) Number of genes significantly upregulated in each age group. (b) Number of genes significantly downregulated in each age group. (c) Interaction network of common factors regulating or regulated by genes identified as upregulated at age 70+. (d) Interaction network of common factors involved in the metabolism of those genes identified as upregulated at age 70+. (e) qPCR results for a subset of genes identified as upregulated with age and implicated in the stress and/or apoptotic response. Results indicate relative expression in age 70+ compared with age 19–29. Error bars represent standard error of the mean. (*P ≤ 0.05).

Figure 3

(a) Interaction network of common factors involved in the expression of genes peaking in expression at age 30–45. (b) Interaction network of common factors in the protein modification of genes peaking in expression at age 30–45. (c) Metabolite profile of genes peaking in expression at age 30–45. (d) Interaction network of common factors regulating or regulated by genes identified as downregulated at age 30–45.

Figure 4

(a) qPCR results for a subset of genes identified as transiently upregulated at age 30–45. Results indicate relative expression at age 30–45 compared with age 19–29. Error bars represent standard error of the mean. (*P ≤ 0.05) (b) qPCR results for a subset of genes identified as transiently downregulated at age 30–45. Results indicate relative expression at age 19–29 compared with age 30–45. Error bars represent standard error of the mean. (*P ≤ 0.05) (c) Interaction network of common factors regulating or regulated by genes identified as upregulated with a regression meta-analysis across the lifespan in male data. (d) Interaction network of common factors regulating or regulated by genes identified as upregulated at age 70+ in a meta-analysis of males aged 70+ compared with age 18–29.