Age-associated epigenetic drift: implications, and a case of epigenetic thrift?

Abstract

It is now well established that the genomic landscape of DNA methylation (DNAm) gets altered as a function of age, a process we here call 'epigenetic drift'. The biological, functional, clinical and evolutionary significance of this epigenetic drift, however, remains unclear. We here provide a brief review of epigenetic drift, focusing on the potential implications for ageing, stem cell biology and disease risk prediction. It has been demonstrated that epigenetic drift affects most of the genome, suggesting a global deregulation of DNAm patterns with age. A component of this drift is tissue-specific, allowing remarkably accurate age-predictive models to be constructed. Another component is tissue-independent, targeting stem cell differentiation pathways and affecting stem cells, which may explain the observed decline of stem cell function with age. Age-associated increases in DNAm target developmental genes, overlapping those associated with environmental disease risk factors and with disease itself, notably cancer. In particular, cancers and precursor cancer lesions exhibit aggravated age DNAm signatures. Epigenetic drift is also influenced by genetic factors. Thus, drift emerges as a promising biomarker for premature or biological ageing, and could potentially be used in geriatrics for disease risk prediction. Finally, we propose, in the context of human evolution, that epigenetic drift may represent a case of epigenetic thrift, or bet-hedging. In summary, this review demonstrates the growing importance of the 'ageing epigenome', with potentially far-reaching implications for understanding the effect of age on stem cell function and differentiation, as well as for disease prevention.

Figure 1.

Tissue-independent age-associated DNAm signatures. (A) Three gene modules derived using Illumina 27K arrays from whole blood with promoters undergoing significant changes in DNAm with age, demonstrating consistency across different tissue types (independent whole-blood data sets, brain, skin and buccal cells) (25). (B) FZD2/WNT-signalling interactome module from West et al. (25) with age-associated directional DNAm changes as derived from the Heyn et al.'s (17) whole-blood data (Illumina 450K arrays). Middle panel shows the validation of the modularity, i.e. the interactome hotspot nature of the module (25) in the 450K data set, and lower panel shows the overall consistency between the 27 and 450K sets, especially for those gene promoters undergoing significant hypermethylation with age. (C) Validation of the directional age-associated DNAm changes (t-statistics) between the whole-blood (WB) training set (27K) and the independent whole-blood 450K data set of Hannum et al. (18).

Figure 2.

Putative effects of epigenetic drift. In normal ageing, whereby an individual is not significantly exposed to disease risk factors and does not have an unfavourable genotype, the deregulation of DNAm happens only gradually and possibly in a linear fashion, as demonstrated by highly accurate age-predictive linear models (18). In contrast, an individual exposed to risk factors, either environmental or genetic, may experience an aggravated or premature ageing profile, characterized by an abnormally higher deregulation of DNAm patterns, increasing the risk of age-related diseases like cancer or diabetes. One can further hypothesize that individuals with a favourable genotype (longevity genes) and with a healthy lifestyle may preserve a more intact epigenome and hence experience longevity. Reprogramming of aged cells into iPSCs and regeneration of differentiated cells may provide a mechanism for epigenetic rejuvenation. In addition to epigenetic drift, telomere shortening has been associated with ageing, age-associated stem cell dysfunction and disease risk factors.