Epigenetic genome-wide association methylation in aging and longevity

Abstract

The aging phenotype is the result of a complex interaction between genetic, epigenetic and environmental factors. Evidence suggests that epigenetic changes (i.e., a set of reversible, heritable changes in gene function or other cell phenotype that occurs without a change in DNA sequence), may affect the aging process and may be one of the central mechanisms by which aging predisposes to many age-related diseases. The total number of altered methylation sites increases with increasing age, such that they could serve as marker for chronological age. This article systematically highlights the advances made in the field of epigenomics and their contribution to the understanding of the complex physiology of aging, lifespan and age-associated diseases.

Aging has emerged as a major global public health issue [1]. Worldwide, the number of individuals older than 65 years of age are projected to increase from 420 million in the year 2000 to as many as 973 million by 2030 [2]. In the USA, this same age group presently consists of 36 million individuals and could increase to nearly 80 million by 2050 [3]. Aging is generally accompanied by age-associated chronic diseases, diminishing quality of life and placement of additional burden on the healthcare system. These factors necessitate identifying biological components that influence human aging and age-related diseases. Both genome-wide association studies and candidate gene approaches have provided valuable information on the role of DNA in aging and disease risk. These approaches, however, do not provide information on differential genetic expression due to developmental or epigenetic changes. The latter relates to a series of histone modifications and DNA methylations, which in turn affect gene expression. Examples of epigenetic regulation of genome architecture and gene expression span across the evolutionary lineage, including sirtuins [4], p66Shc [5], monozygotic twins [6], cloned animals [7] and the maternal effect of agouti mice [8]. Inheritance of epigenetic modifications has been reported in a variety of taxa, ranging from plants [9], yeast [10] and flies [11,12] as well as vertebrates such as mice [13–15] and humans [16]. In general, DNA hypermethylation leads to gene silencing, and DNA hypomethylation endorses gene activation. DNA hypermethylation causing silencing of genes involved in the cell cycle, apoptosis, detoxification and cholesterol metabolism [17] has been reported. Methylation changes may lead to alterations in gene expression and thus contribute to the phenotype of decreased incidence or delayed onset of age-related diseases. In addition, epigenetic changes in specific gene regions have been associated with cancer risk [18]. Epigenetic modification has been linked to many genetic syndromes such as Prader–Willi [19], Angelman [20], Beckwith–Wiedemann [21,22] and Rett syndromes [23,24]. The agouti mouse model provides an example of epigenetic variability and inheritance of a pleiotropic trait with pathological effects [8].

Technological advances in the field

The concept of performing an unbiased approach to reveal age-associated methylation changes and to quantify it using high-throughput machinery has only recently been explored only recently through the use of newly developed high-throughput technologies. Genome-wide DNA methylation and its various effects have been investigated using high-throughput DNA methylation-profiling techniques. In the majority of epigenomic studies, BeadArray™ (Illumina) platform (with bisulfite-treated DNA) has been utilized to analyze a moderate number of CpGs across the genome in a large number of samples [25]. Another strategy has been the use of microarrays with restriction enzyme or affinity-based enrichment methods to provide relative levels of methylation across the genome [25]. More recently, next-generation sequencing-based methods have been implemented to assess tens of millions of DNA fragments, allowing for detection of DNA methylation across the entire genome, including interspersed repeat sequences that are inaccessible using microarrays. Sequencing has several advantages over array platforms, particularly since array restricts the profiling data to specific annotations and content [25]. A variety of platforms have been developed to increase the number of CpGs for which methylation can be assessed. Table 1 reviews the most current DNA methylation platforms. There is an increasing demand for understanding of the genome-wide methylation status via high-throughput techniques for assessment of methylation profiling. This review describes the challenges of applying epigenomic unbiased screening, and reports on the most recent studies carried out in this area.

Table 1.

Epigenomic platforms.

Approach	Coverage of CpGs (%)	Resolution	Platform	Tools	Method
Candidate	0.1–1	Single db	Sequencing	Bs-seq [69] RRBS [70] Array capture [71] Padlock probes [72,73]
		Single db	Bead arrays	Golden Gate [74]
Genome wide	1–10	~50 bp	Targeted arrays	CpG island microarrays [75] Promoter microarrays [76] Tiling microarrays [77,78] SNP arrays [79] Restriction enzymes [80] Microarrays and McrBC [81] Miseq	HELP assay [82] Microarrays and methyl sensitive MeDIP [75] MIRA [83]
		Single db	Bead arrays	Infinium [74]
Whole genome	10–100	1–100 bp (depends on the tool)	Sequencing	Roche – 454 [84] Illumina – Solexa (MeDIP-seq) [75,85] Applied Biosystems – SOLiD [86] Bs-Seq [87–89] Nanopore-seq [90] HiSeq2000,1000	HELPtag assay [91]

Bs-seq: Bisulfite DNA sequencing; HELP: HpaII tiny fragment enrichment by ligation-mediated PCR; MeDIP-seq: Methylated DNA immunoprecipitation sequencing; MIRA: Methylated CpG island recovery assay; RRBS: Reduced representation bisulfite sequencing; SOLiD: Sequencing by oligo ligation detection.

Epigenomic modification with age-empirical support

■ Humans

Evidence in literature suggests that epigenetic changes occur as a function of age in many tissues and may serve as a marker of chronological age. An insight into the epigenetic changes that occur with aging was provided by Bocker et al.; the detailed methylation profile of CD34⁺ hematopoietic progenitor cells demonstrated that epigenetic changes occur with aging and there is de novo methylation of Polycomb chromatin genes [26]. Using a different type of blood cell, CD4⁺, and utilizing the Infinium HumanMethylation27 BeadChip platform, 360 CpG sites (labeled as aging-associated differentially methylated regions) that were either hypo- or hyper-methylated with age have been identified. In total, 60% of the hypermethylation sites were replicated in CD14⁺ of independent cohorts. The aging-associated differentially methylated regions signature, especially hyper-methylation of chromatin domain promoters, has been replicated in buccal cells [27].

Epigenetic changes in several CpG loci, mostly in CpG islands, assessed by Infinium HumanMethylation27 BeadChip were associated with age in different parts of 387 humans (1–102 years old) brains. This central effect of methylation, especially in genes associated with DNA binding and transcription regulation, reemphasizes the importance of methylation in the mechanism of aging [28].

Using the powerful tool of homozygote twins, Bocklandt et al. expanded on an observation that was first reported by Fraga et al [6]. Studying the epigenetic changes with age in 21–55-year-old homozygous twins, they showed that 88 methylation sites, representing nearly 80 genes, demonstrate significant changes with age. The association of those loci with age were further replicated in independent cohort aged 18–70 years old [29] and a regression model built based on this observation could predict an individual’s age with an accuracy of 5.2 years [29]. Changes in epigenomic modification such as methylation can vary substantially between tissues and through the aging process. Christensen et al. examined a panel of 1413 autosomal CpG loci in 217 nonpathologic human tissues from ten anatomic sites and demonstrated that age-associated epigenomic changes occur with age in multiple tissues, and that the methylation pattern with aging can predict the tissue of origin [30]. Aiming to detect an aging signature, Koch and Wagner used five different cell types from four tissues, a total of 110 samples. The commercial DNA methylation platform initially employed to screen contained 27,578 CpG sites. Five genes demonstrated methylation changes with age across the board. These genes were further validated in eight more cell types representing eight tissues, in a total of 766 samples leading authors to conclude that this assay can serve as a predictor of donor age [31]. Using the same platform, Bork et al. studied mesenchymal stromal cells (MSCs) in old (52–83 years of age) and young (21–50 years of age) donors [32]. Significant methylation with age has observed in homeobox genes and genes involved in cell differentiation. Further validation of the epigenome using an independent technology (pyrosequencing) resulted in the discovery that six selected CpG sites demonstrated the same trend. These results further support the role of epigenetics in regulating replicative senescence and aging [32]. In a longitudinal study performed over 8 years in 1097 subjects aged 55–92 years old, the Alu elements demonstrated decline in methylation with age, while the LINE-1 repetitive elements did not change dramatically, emphasizing the role of methylation loss with cellular senescence and aging [33].

Gentilini et al. screened the HUMARA locus for heterozygosity to test the hypothesis that this locus is relevant for lifespan [34]. Using 50 female centenarians and three groups of controls, authors screened 1085 CpG sites across the X chromosome on top of the HUMARA locus for methylation changes, and found no difference between the groups. They concluded that although skewing of X-chromosome inactivation has been observed with aging, there were no associated epigenetic modifications [34].

■ Animal models

There are very few animal studies that have assessed global methylation changes with age. Genomic methylation changes were demonstrated with age using the HELP assay in liver and visceral adipose tissues from young and old rats. These methylation changes were validated with an independent technology (luminometric methylation assays) showing that these changes are tissue dependent. While the pattern of methylation and expression of some of the genes were similar in both the tissues, subsets of the genes that are associated with metabolism and metabolic regulation were differentially expressed with age [35].

miRNA & longevity

miRNAs are small ncRNAs that were initially discovered in Caenorhabditis elegans and since reported across the animal kingdom. In humans, thousands of miRNAs have been demonstrated in a variety of tissues with major impact on transcription and translational repression or gene silencing. The role of miRNAs in aging was demonstrated recently in C. elegans and in mice [36,37]. miRNAs affect gene expression during the aging process in mice and modulate senescence in human cell lines [38]. Studies in C. elegans and mice have resulted within some important observations, such as: miRNAs work in groups (packs) by coordinating and regulating gene expression/silencing resulting in age-dependent disease states or alternatively with longevity [39]; inherited epigenetic effects in miRNA loci lead to changes in gene expression that modulate longevity [40]; and miRNAs that target members of the insulin/IGF-1 pathway (a known target for genetic disruption that leads to life extension) can predict up to 47% of lifespan differences [36]. This observation on the role of IGF-1 was further supported by Liang et al. in studies in long-lived mutant mice, higher expression of three miRNAs altered IGF-1 signaling that in turn promotes long-lived phenomenon [41]; and de Lencastre et al. demonstrated that miRNAs could affect lifespan through disruption of multiple loci that are not necessarily associated with the insulin/IGF-1 pathway. Some loci illustrate positive effects on lifespan, promoting longevity, and some however demonstrate the opposite effect leading to a shorter lifespan [42]. Such observations are also reported by Ugalde et al.; altered expression of two miRNAs promoted progeroid phenotype in a mouse model for a progeria syndrome through the effect on key components of the DNA-damage response pathways [43]. Human studies are limited; however, a genome-wide miRNA screen for differential expression between long-lived individuals and controls revealed that 10% of the miRNA microarray (863 miRNAs) demonstrated significant alterations in expression, of which only 16 were upregulated in the exceptional long-lived individuals. Most of these differentially expressed miRNAs have been associated with genes linked to major age-associated diseases, suggesting under-regulation of key genes by miRNAs could promote longevity in humans [44].

The role of the epigenomic modification in aging & age-related diseases

Aging is a complex physiological process that results in compromise of biological functions, increased susceptibility to age-related diseases and eventually death [45]. It is well recognized that human aging and longevity are influenced by both genetic and environmental factors. Inherited genetic mutations and polymorphisms resulting in alterations in gene function can explain some features of aging and age-related diseases [46]. However, in addition to inherited genetic factors, aging is influenced by the gradual accumulation of molecular alterations after birth. Environmentally induced perturbations in the epigenetic processes that involve alterations of gene expression without a change in DNA sequence can determine different aspects of aging, as well as etiology and pathogenesis of age-related diseases. Epigenetic changes can specifically play a role in the modulation of aging processes and healthy life extension [47]. In particular, promoter DNA methylation changes and associated gene silencing are the epigenetic changes seen in age-related diseases [46].

Epigenetic mechanisms have been established to play a major role in aging at both cellular and organism level [48,49]. In addition to DNA methylation, one of the well-characterized epigenetic processes, several types of histone modifications have been demonstrated to occur both globally and at gene-specific loci during aging [49]. Other epigenetic processes include histone modifications of Polycomb group proteins, chromosomal position effects and methylation of ncRNAs [49–52].

Common human age-related diseases are accompanied by a loss of genomic DNA methylation. Progressive loss of genomic DNA methylation has been demonstrated throughout the human genome [33], in mice and in cell lines [53–55], although this decrease may be tissue and/or gene specific [56–58]. Age-related epigenetic changes have also been demonstrated in sperm cells; however, the direction of change over time appears to be gene specific [59]. Within pairs, differences in DNA methylation are greater in older than in younger monozygotic twins [6].

DNA methylation dynamics can inf luence brain function. 5-hydroxymethylcytosine (5-hmC) is a newly described epigenetic modification generated by the oxidation of 5-methylcytosine by the ten–eleven translocation family of enzymes [60–62]. An increase of 5-hmC with age in the mouse brain as well as an age- and gene-expression level related enrichment of 5-hmC in genes implicated in neurodegeneration have been demonstrated. Many 5-hmC-regulated regions are dynamically modified during neurodevelopment and aging [61], suggesting that 5-hmC may play an important role in the etiology and course of age-related neurodegenerative disorders [60].

With aging, there is a decrease in long-term synaptic plasticity, especially long-term potentiation (LTP), manifesting as cognitive decline. This decrease in LTP is linked to histone acetylation, and BDNF/trkB signaling. Indeed, treatment with histone deacetylase inhibitors or a neurotrophin receptor B agonist restores LTP in the hippocampus of old animals. These studies suggest that epigenetic changes may play a significant role in age-related diseases [63].

Conclusion & future perspective: exciting directions in epigenomic research in aging

Nutritional epigenetics has emerged as a novel mechanism underlying gene–diet interactions, elucidating the modulatory role of nutrition in aging and age-related disease development. Nutrients can regulate the placement of these epigenetic modifications [64–67]. Nutrients involved in one-carbon metabolism, namely folate, vitamin B₁₂, vitamin B₆, riboflavin, methionine, choline and betaine, are involved in DNA methylation by regulating levels of the universal methyl donor S-adenosylmethionine and methyltransferase inhibitor S-adenosylhomocysteine. The effects of folate on DNA methylation patterns have recently been investigated in prenatal and early-postnatal life and aging. Folate exposure in the intrauterine environment and during the aging process may have profound effects on DNA methylation with significant functional ramifications [68]. In addition to folate and vitamin B₁₂, other nutrients and related compounds such as retinoic acid, resveratrol, curcumin, sulforaphane and tea polyphenols can modulate epigenetic patterns by altering the levels of S-adenosylmethionine and S-adenosylhomocysteine or directing the enzymes that catalyse DNA methylation and histone modifications [64]. Indeed, sirtuin 1 may mediate some of the effects of dietary restriction through effects on DNA methylation [66].

The field of epigenomics is exciting, new and is rapidly evolving. Although a nascent field, significant progress and advancement of technology offer the promise of better understanding of the role of epigenetics in the complex process of aging, age-related diseases and, therefore, lifespan.

Executive summary.

• ▪Epigenetic changes occur with age.
• ▪Advances in the field of epigenetics have increased our understanding of the association between disease states and methylation patterns.
• ▪Technology to detect global epigenetic changes has made tremendous strides.
• ▪Unique ‘epigenetic signatures’ may be detected in different tissues.
• ▪The environment can contribute to the diversity in individual epigenetic changes.
• ▪New epigenetic mechanisms such as miRNAs play a role in gene regulation.
• ▪Nutritional epigenomics is an exciting and emerging field with translational potential.
• ▪Evidence is accumulating that the presumed ‘junk DNA’ is indeed functional with a role in epigenetics.
• ▪Finally, altered epigenetic patterns could play a role in cell maintenance, delay age-associated diseases and improve life expectancy.