Cell and tissue type independent age-associated DNA methylation changes are not rare but common

Epigenetic ageing clocks: statistical methods and emerging computational challenges

Over the past decade, epigenetic clocks have emerged as powerful machine learning tools, not only to estimate chronological and biological age but also to assess the efficacy of anti-ageing, cellular rejuvenation and disease-preventive interventions. However, many computational and statistical challenges remain that limit our understanding, interpretation and application of epigenetic clocks. Here, we review these computational challenges, focusing on interpretation, cell-type heterogeneity and emerging single-cell methods, aiming to provide guidelines for the rigorous construction of interpretable epigenetic clocks at cell-type and single-cell resolution.

A Sex-Specific Minimal CpG-Based Model for Biological Aging Using ELOVL2 Methylation Analysis

Significant deviations between chronological and biological age can signal the early risk of chronic diseases, driving the need for tools that accurately determine biological age. While DNA methylation-based clocks have demonstrated strong predictive power for biological aging determination, their clinical application is limited by several barriers including high costs, the need to analyze hundreds of methylation sites using sophisticated platforms and the lack of standardized measurement tools and protocols. In this study, we developed a multivariate linear model using the analysis of eight CpGs within the promoter region of the very long chain fatty acid elongase 2 gene (ELOVL2). The model generated predicts biological age with a mean absolute error (MAE) of 5.04, providing a simplified, cost-effective alternative to more complex methylation-based clocks. Additionally, we identified sex-specific biological clocks, achieving MAEs of 4.37 for males and 5.38 for females, highlighting sex-related molecular differences in the methylation of this gene during aging. Our minimal CpG-based clock offers a practical solution for estimating biological age, with potential applications in clinical practice for assessing age-related disease risks and providing personalized healthcare interventions.

An in-depth understanding of the role and mechanisms of T cells in immune organ aging and age-related diseases

T cells play a critical and irreplaceable role in maintaining overall health. However, their functions undergo alterations as individuals age. It is of utmost importance to comprehend the specific characteristics of T-cell aging, as this knowledge is crucial for gaining deeper insights into the pathogenesis of aging-related diseases and developing effective therapeutic strategies. In this review, we have thoroughly examined the existing studies on the characteristics of immune organ aging. Furthermore, we elucidated the changes and potential mechanisms that occur in T cells during the aging process. Additionally, we have discussed the latest research advancements pertaining to T-cell aging-related diseases. These findings provide a fresh perspective for the study of T cells in the context of aging.

Cell-type specific epigenetic clocks to quantify biological age at cell-type resolution

The ability to accurately quantify biological age could help monitor and control healthy aging. Epigenetic clocks have emerged as promising tools for estimating biological age, yet they have been developed from heterogeneous bulk tissues, and are thus composites of two aging processes, one reflecting the change of cell-type composition with age and another reflecting the aging of individual cell-types. There is thus a need to dissect and quantify these two components of epigenetic clocks, and to develop epigenetic clocks that can yield biological age estimates at cell-type resolution. Here we demonstrate that in blood and brain, approximately 39% and 12% of an epigenetic clock's accuracy is driven by underlying shifts in lymphocyte and neuronal subsets, respectively. Using brain and liver tissue as prototypes, we build and validate neuron and hepatocyte specific DNA methylation clocks, and demonstrate that these cell-type specific clocks yield improved estimates of chronological age in the corresponding cell and tissue-types. We find that neuron and glia specific clocks display biological age acceleration in Alzheimer's Disease with the effect being strongest for glia in the temporal lobe. Moreover, CpGs from these clocks display a small but significant overlap with the causal DamAge-clock, mapping to key genes implicated in neurodegeneration. The hepatocyte clock is found accelerated in liver under various pathological conditions. In contrast, non-cell-type specific clocks do not display biological age-acceleration, or only do so marginally. In summary, this work highlights the importance of dissecting epigenetic clocks and quantifying biological age at cell-type resolution.

MinLinMo: a minimalist approach to variable selection and linear model prediction

Generating prediction models from high dimensional data often result in large models with many predictors. Causal inference for such models can therefore be difficult or even impossible in practice. The stand-alone software package MinLinMo emphasizes small linear prediction models over highest possible predictability with a particular focus on including variables correlated with the outcome, minimal memory usage and speed. MinLinMo is demonstrated on large epigenetic datasets with prediction models for chronological age, gestational age, and birth weight comprising, respectively, 15, 14 and 10 predictors. The parsimonious MinLinMo models perform comparably to established prediction models requiring hundreds of predictors.

A comprehensive map of the aging blood methylome in humans

BACKGROUND

During aging, the human methylome undergoes both differential and variable shifts, accompanied by increased entropy. The distinction between variably methylated positions (VMPs) and differentially methylated positions (DMPs), their contribution to epigenetic age, and the role of cell type heterogeneity remain unclear.

RESULTS

We conduct a comprehensive analysis of > 32,000 human blood methylomes from 56 datasets (age range = 6-101 years). We find a significant proportion of the blood methylome that is differentially methylated with age (48% DMPs; FDR < 0.005) and variably methylated with age (37% VMPs; FDR < 0.005), with considerable overlap between the two groups (59% of DMPs are VMPs). Bivalent and Polycomb regions become increasingly methylated and divergent between individuals, while quiescent regions lose methylation more uniformly. Both chronological and biological clocks, but not pace-of-aging clocks, show a strong enrichment for CpGs undergoing both mean and variance changes during aging. The accumulation of DMPs shifting towards a methylation fraction of 50% drives the increase in entropy, smoothening the epigenetic landscape. However, approximately a quarter of DMPs exhibit anti-entropic effects, opposing this direction of change. While changes in cell type composition minimally affect DMPs, VMPs and entropy measurements are moderately sensitive to such alterations.

CONCLUSION

This study represents the largest investigation to date of genome-wide DNA methylation changes and aging in a single tissue, providing valuable insights into primary molecular changes relevant to chronological and biological aging.

On epigenetic stochasticity, entropy and cancer risk

Epigenetic changes are known to accrue in normal cells as a result of ageing and cumulative exposure to cancer risk factors. Increasing evidence points towards age-related epigenetic changes being acquired in a quasi-stochastic manner, and that they may play a causal role in cancer development. Here, I describe the quasi-stochastic nature of DNA methylation (DNAm) changes in ageing cells as well as in normal cells at risk of neoplastic transformation, discussing the implications of this stochasticity for developing cancer risk prediction strategies, and in particular, how it may require a conceptual paradigm shift in how we select cancer risk markers. I also describe the mounting evidence that a significant proportion of DNAm changes in ageing and cancer development are related to cell proliferation, reflecting tissue-turnover and the opportunity this offers for predicting cancer risk via the development of epigenetic mitotic-like clocks. Finally, I describe how age-associated DNAm changes may be causally implicated in cancer development via an irreversible suppression of tissue-specific transcription factors that increases epigenetic and transcriptomic entropy, promoting a more plastic yet aberrant cancer stem-cell state. This article is part of a discussion meeting issue 'Causes and consequences of stochastic processes in development and disease'.

Pan-primate studies of age and sex

Age and sex have a profound effect on cytosine methylation levels in humans and many other species. Here we analyzed DNA methylation profiles of 2400 tissues derived from 37 primate species including 11 haplorhine species (baboons, marmosets, vervets, rhesus macaque, chimpanzees, gorillas, orangutan, humans) and 26 strepsirrhine species (suborders Lemuriformes and Lorisiformes). From these we present here, pan-primate epigenetic clocks which are highly accurate for all primates including humans (age correlation R = 0.98). We also carried out in-depth analysis of baboon DNA methylation profiles and generated five epigenetic clocks for baboons (Olive-yellow baboon hybrid), one of which, the pan-tissue epigenetic clock, was trained on seven tissue types (fetal cerebral cortex, adult cerebral cortex, cerebellum, adipose, heart, liver, and skeletal muscle) with ages ranging from late fetal life to 22.8 years of age. Using the primate data, we characterize the effect of age and sex on individual cytosines in highly conserved regions. We identify 11 sex-related CpGs on autosomes near genes (POU3F2, CDYL, MYCL, FBXL4, ZC3H10, ZXDC, RRAS, FAM217A, RBM39, GRIA2, UHRF2). Low overlap can be observed between age- and sex-related CpGs. Overall, this study advances our understanding of conserved age- and sex-related epigenetic changes in primates, and provides biomarkers of aging for all primates.

The Cutting Edge of Epigenetic Clocks: In Search of Mechanisms Linking Aging and Mental Health

Individuals with psychiatric disorders are at increased risk of age-related diseases and early mortality. Recent studies demonstrate that this link between mental health and aging is reflected in epigenetic clocks, aging biomarkers based on DNA methylation. The reported relationships between epigenetic clocks and mental health are mostly correlational, and the mechanisms are poorly understood. Here, we review recent progress concerning the molecular and cellular processes underlying epigenetic clocks as well as novel technologies enabling further studies of the causes and consequences of epigenetic aging. We then review the current literature on how epigenetic clocks relate to specific aspects of mental health, such as stress, medications, substance use, health behaviors, and symptom clusters. We propose an integrated framework where mental health and epigenetic aging are each broken down into multiple distinct processes, which are then linked to each other, using stress and schizophrenia as examples. This framework incorporates the heterogeneity and complexity of both mental health conditions and aging, may help reconcile conflicting results, and provides a basis for further hypothesis-driven research in humans and model systems to investigate potentially causal mechanisms linking aging and mental health.

A meta-analysis of immune-cell fractions at high resolution reveals novel associations with common phenotypes and health outcomes

BACKGROUND

Changes in cell-type composition of tissues are associated with a wide range of diseases and environmental risk factors and may be causally implicated in disease development and progression. However, these shifts in cell-type fractions are often of a low magnitude, or involve similar cell subtypes, making their reliable identification challenging. DNA methylation profiling in a tissue like blood is a promising approach to discover shifts in cell-type abundance, yet studies have only been performed at a relatively low cellular resolution and in isolation, limiting their power to detect shifts in tissue composition.

METHODS

Here we derive a DNA methylation reference matrix for 12 immune-cell types in human blood and extensively validate it with flow-cytometric count data and in whole-genome bisulfite sequencing data of sorted cells. Using this reference matrix, we perform a directional Stouffer and fixed effects meta-analysis comprising 23,053 blood samples from 22 different cohorts, to comprehensively map associations between the 12 immune-cell fractions and common phenotypes. In a separate cohort of 4386 blood samples, we assess associations between immune-cell fractions and health outcomes.

RESULTS

Our meta-analysis reveals many associations of cell-type fractions with age, sex, smoking and obesity, many of which we validate with single-cell RNA sequencing. We discover that naïve and regulatory T-cell subsets are higher in women compared to men, while the reverse is true for monocyte, natural killer, basophil, and eosinophil fractions. Decreased natural killer counts associated with smoking, obesity, and stress levels, while an increased count correlates with exercise and sleep. Analysis of health outcomes revealed that increased naïve CD4 + T-cell and N-cell fractions associated with a reduced risk of all-cause mortality independently of all major epidemiological risk factors and baseline co-morbidity. A machine learning predictor built only with immune-cell fractions achieved a C-index value for all-cause mortality of 0.69 (95%CI 0.67-0.72), which increased to 0.83 (0.80-0.86) upon inclusion of epidemiological risk factors and baseline co-morbidity.

CONCLUSIONS

This work contributes an extensively validated high-resolution DNAm reference matrix for blood, which is made freely available, and uses it to generate a comprehensive map of associations between immune-cell fractions and common phenotypes, including health outcomes.

The immune factors driving DNA methylation variation in human blood

Epigenetic changes are required for normal development, yet the nature and respective contribution of factors that drive epigenetic variation in humans remain to be fully characterized. Here, we assessed how the blood DNA methylome of 884 adults is affected by DNA sequence variation, age, sex and 139 factors relating to life habits and immunity. Furthermore, we investigated whether these effects are mediated or not by changes in cellular composition, measured by deep immunophenotyping. We show that DNA methylation differs substantially between naïve and memory T cells, supporting the need for adjustment on these cell-types. By doing so, we find that latent cytomegalovirus infection drives DNA methylation variation and provide further support that the increased dispersion of DNA methylation with aging is due to epigenetic drift. Finally, our results indicate that cellular composition and DNA sequence variation are the strongest predictors of DNA methylation, highlighting critical factors for medical epigenomics studies.

Aging the brain: multi-region methylation principal component based clock in the context of Alzheimer's disease

Alzheimer's disease (AD) risk increases exponentially with age and is associated with multiple molecular hallmarks of aging, one of which is epigenetic alterations. Epigenetic age predictors based on 5' cytosine methylation (DNAm), or epigenetic clocks, have previously suggested that epigenetic age acceleration may occur in AD brain tissue. Epigenetic clocks are promising tools for the quantification of biological aging, yet we hypothesize that investigation of brain aging in AD will be assisted by the development of brain-specific epigenetic clocks. Therefore, we generated a novel age predictor termed PCBrainAge that was trained solely in cortical samples. This predictor utilizes a combination of principal components analysis and regularized regression, which reduces technical noise and greatly improves test-retest reliability. To characterize the scope of PCBrainAge's utility, we generated DNAm data from multiple brain regions in a sample from the Religious Orders Study and Rush Memory and Aging Project. PCBrainAge captures meaningful heterogeneity of aging: Its acceleration demonstrates stronger associations with clinical AD dementia, pathologic AD, and APOE ε4 carrier status compared to extant epigenetic age predictors. It further does so across multiple cortical and subcortical regions. Overall, PCBrainAge's increased reliability and specificity makes it a particularly promising tool for investigating heterogeneity in brain aging, as well as epigenetic alterations underlying AD risk and resilience.

Lipid metabolism dysfunction induced by age-dependent DNA methylation accelerates aging

Epigenetic alterations and metabolic dysfunction are two hallmarks of aging. However, the mechanism of how their interaction regulates aging, particularly in mammals, remains largely unknown. Here we show ELOVL fatty acid elongase 2 (Elovl2), a gene whose epigenetic alterations are most highly correlated with age prediction, contributes to aging by regulating lipid metabolism. We applied artificial intelligence to predict the protein structure of ELOVL2 and the interaction with its substrate. Impaired Elovl2 function disturbs lipid synthesis with increased endoplasmic reticulum stress and mitochondrial dysfunction, leading to key aging phenotypes at both cellular and physiological level. Furthermore, restoration of mitochondrial activity can rescue age-related macular degeneration (AMD) phenotypes induced by Elovl2 deficiency in human retinal pigmental epithelial (RPE) cells; this indicates a conservative mechanism in both human and mouse. Taken together, we revealed an epigenetic-metabolism axis contributing to aging and illustrate the power of an AI-based approach in structure-function studies.

Making sense of the ageing methylome

Over time, the human DNA methylation landscape accrues substantial damage, which has been associated with a broad range of age-related diseases, including cardiovascular disease and cancer. Various age-related DNA methylation changes have been described, including at the level of individual CpGs, such as differential and variable methylation, and at the level of the whole methylome, including entropy and correlation networks. Here, we review these changes in the ageing methylome as well as the statistical tools that can be used to quantify them. We detail the evidence linking DNA methylation to ageing phenotypes and the longevity strategies aimed at altering both DNA methylation patterns and machinery to extend healthspan and lifespan. Lastly, we discuss theories on the mechanistic causes of epigenetic ageing.

Cell-type heterogeneity: Why we should adjust for it in epigenome and biomarker studies

Most studies aiming to identify epigenetic biomarkers do so from complex tissues that are composed of many different cell-types. By definition, these cell-types vary substantially in terms of their epigenetic profiles. This cell-type specific variation among healthy cells is completely independent of the variation associated with disease, yet it dominates the epigenetic variability landscape. While cell-type composition of tissues can change in disease and this may provide accurate and reproducible biomarkers, not adjusting for the underlying cell-type heterogeneity may seriously limit the sensitivity and precision to detect disease-relevant biomarkers or hamper our understanding of such biomarkers. Given that computational and experimental tools for tackling cell-type heterogeneity are available, we here stress that future epigenetic biomarker studies should aim to provide estimates of underlying cell-type fractions for all samples in the study, and to identify biomarkers before and after adjustment for cell-type heterogeneity, in order to obtain a more complete and unbiased picture of the biomarker-landscape. This is critical, not only to improve reproducibility and for the eventual clinical application of such biomarkers, but importantly, to also improve our molecular understanding of disease itself.

Epigenetics and first-episode psychosis: A systematic review

BACKGROUND

Schizophrenia has a large disease burden globally. Early intervention in psychosis, and therefore a decreased duration of untreated psychosis, has a positive clinical impact. There are several recognized risk factors for psychosis, including trauma history and substance use. This systematic review examined the literature for studies related to epigenetic changes in first-episode psychosis, with the goal of identifying future research directions.

METHODS

A literature review was conducted from inception to October 3, 2021 using MedLine/PubMed, Web of Science, and PsycInfo searches with the keywords ("first-episode schizophrenia" OR "first-episode psychosis" OR "drug-naive schizophrenia" OR "drug-naive psychosis") AND (epigenetic OR methylation OR hydroxymethylation OR "histone modification" OR "miRNA") as well as a search of the bibliography of the identified papers.

RESULTS

Seventeen studies that examined various portions of the genome were included in this systematic review. There were two studies that showed hypomethylation at the LINE-1 portion of the genome and two that showed hypermethylation at LINE-1. Additionally, two studies showed hypomethylation specifically at the GRIN2B promoter (part of LINE-1).

CONCLUSIONS

Although sample sizes were small, these studies provide evidence for epigenetic alterations in early psychosis. Further research in this area is warranted for more definitive epigenetic correlations.

Aging biomarkers and the brain

Quantifying biological aging is critical for understanding why aging is the primary driver of morbidity and mortality and for assessing novel therapies to counter pathological aging. In the past decade, many biomarkers relevant to brain aging have been developed using various data types and modeling techniques. Aging involves numerous interconnected processes, and thus many complementary biomarkers are needed, each capturing a different slice of aging biology. Here we present a hierarchical framework highlighting how these biomarkers are related to each other and the underlying biological processes. We review those measures most studied in the context of brain aging: epigenetic clocks, proteomic clocks, and neuroimaging age predictors. Many studies have linked these biomarkers to cognition, mental health, brain structure, and pathology during aging. We also delve into the challenges and complexities in interpreting these biomarkers and suggest areas for further innovation. Ultimately, a robust mechanistic understanding of these biomarkers will be needed to effectively intervene in the aging process to prevent and treat age-related disease.

DNA methylation and histone variants in aging and cancer

Aging-related diseases such as cancer can be traced to the accumulation of molecular disorder including increased DNA mutations and epigenetic drift. We provide a comprehensive review of recent results in mice and humans on modifications of DNA methylation and histone variants during aging and in cancer. Accumulated errors in DNA methylation maintenance lead to global decreases in DNA methylation with relaxed repression of repeated DNA and focal hypermethylation blocking the expression of tumor suppressor genes. Epigenetic clocks based on quantifying levels of DNA methylation at specific genomic sites is proving to be a valuable metric for estimating the biological age of individuals. Histone variants have specialized functions in transcriptional regulation and genome stability. Their concentration tends to increase in aged post-mitotic chromatin, but their effects in cancer are mainly determined by their specialized functions. Our increased understanding of epigenetic regulation and their modifications during aging has motivated interventions to delay or reverse epigenetic modifications using the epigenetic clocks as a rapid readout for efficacity. Similarly, the knowledge of epigenetic modifications in cancer is suggesting new approaches to target these modifications for cancer therapy.

The genomic loci of specific human tRNA genes exhibit ageing-related DNA hypermethylation

The epigenome has been shown to deteriorate with age, potentially impacting on ageing-related disease. tRNA, while arising from only ˜46 kb (<0.002% genome), is the second most abundant cellular transcript. tRNAs also control metabolic processes known to affect ageing, through core translational and additional regulatory roles. Here, we interrogate the DNA methylation state of the genomic loci of human tRNA. We identify a genomic enrichment for age-related DNA hypermethylation at tRNA loci. Analysis in 4,350 MeDIP-seq peripheral-blood DNA methylomes (16-82 years), identifies 44 and 21 hypermethylating specific tRNAs at study-and genome-wide significance, respectively, contrasting with none hypomethylating. Validation and replication (450k array and independent targeted Bisuphite-sequencing) supported the hypermethylation of this functional unit. Tissue-specificity is a significant driver, although the strongest consistent signals, also independent of major cell-type change, occur in tRNA-iMet-CAT-1-4 and tRNA-Ser-AGA-2-6. This study presents a comprehensive evaluation of the genomic DNA methylation state of human tRNA genes and reveals a discreet hypermethylation with advancing age.

Temporal changes in DNA methylation and RNA expression in a small song bird: within- and between-tissue comparisons

Background

DNA methylation is likely a key mechanism regulating changes in gene transcription in traits that show temporal fluctuations in response to environmental conditions. To understand the transcriptional role of DNA methylation we need simultaneous within-individual assessment of methylation changes and gene expression changes over time. Within-individual repeated sampling of tissues, which are essential for trait expression is, however, unfeasible (e.g. specific brain regions, liver and ovary for reproductive timing). Here, we explore to what extend between-individual changes in DNA methylation in a tissue accessible for repeated sampling (red blood cells (RBCs)) reflect such patterns in a tissue unavailable for repeated sampling (liver) and how these DNA methylation patterns are associated with gene expression in such inaccessible tissues (hypothalamus, ovary and liver). For this, 18 great tit (Parus major) females were sacrificed at three time points (n = 6 per time point) throughout the pre-laying and egg-laying period and their blood, hypothalamus, ovary and liver were sampled.

Results

We simultaneously assessed DNA methylation changes (via reduced representation bisulfite sequencing) and changes in gene expression (via RNA-seq and qPCR) over time. In general, we found a positive correlation between changes in CpG site methylation in RBCs and liver across timepoints. For CpG sites in close proximity to the transcription start site, an increase in RBC methylation over time was associated with a decrease in the expression of the associated gene in the ovary. In contrast, no such association with gene expression was found for CpG site methylation within the gene body or the 10 kb up- and downstream regions adjacent to the gene body.

Conclusion

Temporal changes in DNA methylation are largely tissue-general, indicating that changes in RBC methylation can reflect changes in DNA methylation in other, often less accessible, tissues such as the liver in our case. However, associations between temporal changes in DNA methylation with changes in gene expression are mostly tissue- and genomic location-dependent. The observation that temporal changes in DNA methylation within RBCs can relate to changes in gene expression in less accessible tissues is important for a better understanding of how environmental conditions shape traits that temporally change in expression in wild populations.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12864-020-07329-9.

A DNA methylation age predictor for zebrafish

Changes in DNA methylation at specific CpG sites have been used to build predictive models to estimate animal age, predominantly in mammals. Little testing for this effect has been conducted in other vertebrate groups, such as bony fish, the largest vertebrate class. The development of most age-predictive models has relied on a genome-wide sequencing method to obtain a DNA methylation level, which makes it costly to deploy as an assay to estimate age in many samples. Here, we have generated a reduced representation bisulfite sequencing data set of caudal fin tissue from a model fish species, zebrafish (Danio rerio), aged from 11.9-60.1 weeks. We identified changes in methylation at specific CpG sites that correlated strongly with increasing age. Using an optimised unique set of 26 CpG sites we developed a multiplex PCR assay that predicts age with an average median absolute error rate of 3.2 weeks in zebrafish between 10.9-78.1 weeks of age. We also demonstrate the use of multiplex PCR as an efficient quantitative approach to measure DNA methylation for the use of age estimation. This study highlights the potential further use of DNA methylation as an age estimation method in non-mammalian vertebrate species.

Age-related DNA methylation changes are sex-specific: a comprehensive assessment

The existence of a sex gap in human health and longevity has been widely documented. Autosomal DNA methylation differences between males and females have been reported, but so far few studies have investigated if DNA methylation is differently affected by aging in males and females. We performed a meta-analysis of 4 large whole blood datasets, comparing 4 aspects of epigenetic age-dependent remodeling between the two sexes: differential methylation, variability, epimutations and entropy. We reported that a large fraction (43%) of sex-associated probes undergoes age-associated DNA methylation changes, and that a limited number of probes show age-by-sex interaction. We experimentally validated 2 regions mapping in FIGN and PRR4 genes and showed sex-specific deviations of their methylation patterns in models of decelerated (centenarians) and accelerated (Down syndrome) aging. While we did not find sex differences in the age-associated increase in epimutations and entropy, we showed that the number of probes having an age-related increase in methylation variability is 15 times higher in males compared to females. Our results can offer new epigenetic tools to study the interaction between aging and sex and can pave the way to the identification of molecular triggers of sex differences in longevity and age-related diseases prevalence.

A cell-type deconvolution meta-analysis of whole blood EWAS reveals lineage-specific smoking-associated DNA methylation changes

Highly reproducible smoking-associated DNA methylation changes in whole blood have been reported by many Epigenome-Wide-Association Studies (EWAS). These epigenetic alterations could have important implications for understanding and predicting the risk of smoking-related diseases. To this end, it is important to establish if these DNA methylation changes happen in all blood cell subtypes or if they are cell-type specific. Here, we apply a cell-type deconvolution algorithm to identify cell-type specific DNA methylation signals in seven large EWAS. We find that most of the highly reproducible smoking-associated hypomethylation signatures are more prominent in the myeloid lineage. A meta-analysis further identifies a myeloid-specific smoking-associated hypermethylation signature enriched for DNase Hypersensitive Sites in acute myeloid leukemia. These results may guide the design of future smoking EWAS and have important implications for our understanding of how smoking affects immune-cell subtypes and how this may influence the risk of smoking related diseases.

Smoking-associated DNA methylation changes in whole blood have been reported by many EWAS. Here, the authors use a cell-type deconvolution algorithm to identify cell-type specific DNA methylation signals in seven EWAS, identifying lineage-specific smoking-associated DNA methylation changes.

A HIERARCHICAL MODEL FOR THE CONTROL OF EPIGENETIC AGING IN MAMMALS

Regulatory mechanisms range from a single level of control in simple metazoans to multi-level hierarchical control networks in higher animals. Organismal regulation encompasses homeostatic and circadian networks that are interconnected, with no documented exceptions. The epigenetic clock is a highly accurate biomarker of age in humans, defined by a mathematical algorithm based on the methylation of a subset of age-related CpG sites on DNA. Experimental evidence suggests the existence of an underlying regulatory mechanism. By analogy with other integrative systems as the neuroendocrine-immune network and the circadian clocks, a hierarchical organization in the control of the ticking rate of the epigenetic clock is hypothesized here. The hierarchical organization of the neuroendocrine, immune and circadian systems is briefly reviewed. This is followed by a brief review of the epigenetic clock at cell level. Finally, different lines of indirect evidence, consistent with the existence of a central pacemaker controlling the ticking rate of the epigenetic clock at organismal level are discussed. The concluding remarks put the hierarchical model proposed for the control of the clock into an evolutionary perspective. Within this perspective, the present hypothesis is intended as a conceptual outline based on designs consistently favored by evolution in higher animals.

A comparison of epigenetic mitotic-like clocks for cancer risk prediction

Background

DNA methylation changes that accrue in the stem cell pool of an adult tissue in line with the cumulative number of cell divisions may contribute to the observed variation in cancer risk among tissues and individuals. Thus, the construction of epigenetic “mitotic” clocks that can measure the lifetime number of stem cell divisions is of paramount interest.

Methods

Building upon a dynamic model of DNA methylation gain in unmethylated CpG-rich regions, we here derive a novel mitotic clock (“epiTOC2”) that can directly estimate the cumulative number of stem cell divisions in a tissue. We compare epiTOC2 to a different mitotic model, based on hypomethylation at solo-WCGW sites (“HypoClock”), in terms of their ability to measure mitotic age of normal adult tissues and predict cancer risk.

Results

Using epiTOC2, we estimate the intrinsic stem cell division rate for different normal tissue types, demonstrating excellent agreement (Pearson correlation = 0.92, R² = 0.85, P = 3e−6) with those derived from experiment. In contrast, HypoClock’s estimates do not (Pearson correlation = 0.30, R² = 0.09, P = 0.29). We validate these results in independent datasets profiling normal adult tissue types. While both epiTOC2 and HypoClock correctly predict an increased mitotic rate in cancer, epiTOC2 is more robust and significantly better at discriminating preneoplastic lesions characterized by chronic inflammation, a major driver of tissue turnover and cancer risk. Our data suggest that DNA methylation loss at solo-WCGWs is significant only when cells are under high replicative stress and that epiTOC2 is a better mitotic age and cancer risk prediction model for normal adult tissues.

Conclusions

These results have profound implications for our understanding of epigenetic clocks and for developing cancer risk prediction or early detection assays. We propose that measurement of DNAm at the 163 epiTOC2 CpGs in adult pre-neoplastic lesions, and potentially in serum cell-free DNA, could provide the basis for building feasible pre-diagnostic or cancer risk assays. epiTOC2 is freely available from 10.5281/zenodo.2632938

Weighted elastic net for unsupervised domain adaptation with application to age prediction from DNA methylation data

MOTIVATION

Predictive models are a powerful tool for solving complex problems in computational biology. They are typically designed to predict or classify data coming from the same unknown distribution as the training data. In many real-world settings, however, uncontrolled biological or technical factors can lead to a distribution mismatch between datasets acquired at different times, causing model performance to deteriorate on new data. A common additional obstacle in computational biology is scarce data with many more features than samples. To address these problems, we propose a method for unsupervised domain adaptation that is based on a weighted elastic net. The key idea of our approach is to compare dependencies between inputs in training and test data and to increase the cost of differently behaving features in the elastic net regularization term. In doing so, we encourage the model to assign a higher importance to features that are robust and behave similarly across domains.

RESULTS

We evaluate our method both on simulated data with varying degrees of distribution mismatch and on real data, considering the problem of age prediction based on DNA methylation data across multiple tissues. Compared with a non-adaptive standard model, our approach substantially reduces errors on samples with a mismatched distribution. On real data, we achieve far lower errors on cerebellum samples, a tissue which is not part of the training data and poorly predicted by standard models. Our results demonstrate that unsupervised domain adaptation is possible for applications in computational biology, even with many more features than samples.

AVAILABILITY AND IMPLEMENTATION

Source code is available at https://github.com/PfeiferLabTue/wenda.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

DNA methylation age calculators reveal association with diabetic neuropathy in type 1 diabetes

BACKGROUND

Many CpGs become hyper or hypo-methylated with age. Multiple methods have been developed by Horvath et al. to estimate DNA methylation (DNAm) age including Pan-tissue, Skin & Blood, PhenoAge, and GrimAge. Pan-tissue and Skin & Blood try to estimate chronological age in the normal population whereas PhenoAge and GrimAge use surrogate markers associated with mortality to estimate biological age and its departure from chronological age. Here, we applied Horvath's four methods to calculate and compare DNAm age in 499 subjects with type 1 diabetes (T1D) from the Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications (DCCT/EDIC) study using DNAm data measured by Illumina EPIC array in the whole blood. Association of the four DNAm ages with development of diabetic complications including cardiovascular diseases (CVD), nephropathy, retinopathy, and neuropathy, and their risk factors were investigated.

RESULTS

Pan-tissue and GrimAge were higher whereas Skin & Blood and PhenoAge were lower than chronological age (p < 0.0001). DNAm age was not associated with the risk of CVD or retinopathy over 18-20 years after DNAm measurement. However, higher PhenoAge (β = 0.023, p = 0.007) and GrimAge (β = 0.029, p = 0.002) were associated with higher albumin excretion rate (AER), an indicator of diabetic renal disease, measured over time. GrimAge was also associated with development of both diabetic peripheral neuropathy (OR = 1.07, p = 9.24E-3) and cardiovascular autonomic neuropathy (OR = 1.06, p = 0.011). Both HbA1c (β = 0.38, p = 0.026) and T1D duration (β = 0.01, p = 0.043) were associated with higher PhenoAge. Employment (β = - 1.99, p = 0.045) and leisure time (β = - 0.81, p = 0.022) physical activity were associated with lower Pan-tissue and Skin & Blood, respectively. BMI (β = 0.09, p = 0.048) and current smoking (β = 7.13, p = 9.03E-50) were positively associated with Skin & Blood and GrimAge, respectively. Blood pressure, lipid levels, pulse rate, and alcohol consumption were not associated with DNAm age regardless of the method used.

CONCLUSIONS

Various methods of measuring DNAm age are sub-optimal in detecting people at higher risk of developing diabetic complications although some work better than the others.

DNA methylation and histone acetylation changes to cytochrome P450 2E1 regulation in normal aging and impact on rates of drug metabolism in the liver

Aging is associated with reduced liver function that may increase the risk for adverse drug reactions in older adults. We hypothesized that age-related changes to epigenetic regulation of genes involved in drug metabolism may contribute to this effect. We reviewed published epigenome-wide studies of human blood and identified the cytochrome P450 2E1 (CYP2E1) gene as a top locus exhibiting epigenetic changes with age. To investigate potential functional changes with age in the liver, the primary organ of drug metabolism, we obtained liver tissue from mice aged 4-32 months from the National Institute on Aging. We assayed global DNA methylation (5-methylcytosine, 5mC), hydroxymethylation (5-hydroxymethylcytosine, 5hmC), and locus-specific 5mC and histone acetylation changes around mouse Cyp2e1. The mouse livers exhibit significant global decreases in 5mC and 5hmC with age. Furthermore, 5mC significantly increased with age at two regulatory regions of Cyp2e1 in tandem with decreases in its gene and protein expressions. H3K9ac levels also changed with age at both regulatory regions of Cyp2e1 investigated, while H3K27ac did not. To test if these epigenetic changes are associated with varying rates of drug metabolism, we assayed clearance of the CYP2E1-specific probe drug chlorzoxazone in microsome extracts from the same livers. CYP2E1 intrinsic clearance is associated with DNA methylation and H3K9ac levels at the Cyp2e1 locus but not with chronological age. This suggests that age-related epigenetic changes may influence rates of hepatic drug metabolism. In the future, epigenetic biomarkers could prove useful to guide dosing regimens in older adults.

DNA methylation clocks as a predictor for ageing and age estimation in naked mole-rats, Heterocephalus glaber

The naked mole-rat, Heterocephalus glaber (NMR), the longest-lived rodent, is of significance and interest in the study of biomarkers for ageing. Recent breakthroughs in this field have revealed ‘epigenetic clocks’ that are based on the temporal accumulation of DNA methylation at specific genomic sites. Here, we validate the hypothesis of an epigenetic clock in NMRs based on changes in methylation of targeted CpG sites. We initially analysed 51 CpGs in NMR livers spanning an age range of 39-1,144 weeks and found 23 to be significantly associated with age (p<0.05). We then built a predictor of age using these sites. To test the accuracy of this model, we analysed an additional set of liver samples, and were successfully able to predict their age with a root mean squared error of 166 weeks. We also profiled skin samples with the same age range, finding a striking correlation between their predicted age versus their actual age (R=0.93), but which was lower when compared to the liver, suggesting that skin ages slower than the liver in NMRs. Our model will enable the prediction of age in wild-caught and captive NMRs of unknown age, and will be invaluable for further mechanistic studies of mammalian ageing.

Impaired lipid metabolism by age-dependent DNA methylation alterations accelerates aging

Epigenetic alterations and metabolic dysfunction are two hallmarks of aging. However, the mechanism of how their interaction regulates aging, particularly in mammals, remains largely unknown. Here we show ELOVL fatty acid elongase 2 (Elovl2), a gene whose epigenetic alterations are most highly correlated with age prediction, contributes to aging by regulating lipid metabolism. Impaired Elovl2 function disturbs lipid synthesis with increased endoplasmic reticulum stress and mitochondrial dysfunction, leading to key accelerated aging phenotypes. Restoration of mitochondrial activity can rescue age-related macular degeneration (AMD) phenotypes induced by Elovl2 deficiency in human retinal pigmental epithelial (RPE) cells. We revealed an epigenetic-metabolism axis contributing to aging and potentially to antiaging therapy.

DNA methylation aging clocks: challenges and recommendations

Epigenetic clocks comprise a set of CpG sites whose DNA methylation levels measure subject age. These clocks are acknowledged as a highly accurate molecular correlate of chronological age in humans and other vertebrates. Also, extensive research is aimed at their potential to quantify biological aging rates and test longevity or rejuvenating interventions. Here, we discuss key challenges to understand clock mechanisms and biomarker utility. This requires dissecting the drivers and regulators of age-related changes in single-cell, tissue- and disease-specific models, as well as exploring other epigenomic marks, longitudinal and diverse population studies, and non-human models. We also highlight important ethical issues in forensic age determination and predicting the trajectory of biological aging in an individual.

Early-life DNA methylation profiles are indicative of age-related transcriptome changes

Background

Alterations to cellular and molecular programs with brain aging result in cognitive impairment and susceptibility to neurodegenerative disease. Changes in DNA methylation patterns, an epigenetic modification required for various CNS functions are observed with brain aging and can be prevented by anti-aging interventions, but the relationship of altered methylation to gene expression is poorly understood.

Results

Paired analysis of the hippocampal methylome and transcriptome with aging of male and female mice demonstrates that age-related differences in methylation and gene expression are anti-correlated within gene bodies and enhancers. Altered promoter methylation with aging was found to be generally un-related to altered gene expression. A more striking relationship was found between methylation levels at young age and differential gene expression with aging. Highly methylated gene bodies and promoters in early life were associated with age-related increases in gene expression even in the absence of significant methylation changes with aging. As well, low levels of methylation in early life were correlated to decreased expression with aging. This relationship was also observed in genes altered in two mouse Alzheimer’s models.

Conclusion

DNA methylation patterns established in youth, in combination with other epigenetic marks, were able to accurately predict changes in transcript trajectories with aging. These findings are consistent with the developmental origins of disease hypothesis and indicate that epigenetic variability in early life may explain differences in aging trajectories and age-related disease.

The effect of age on DNA methylation in whole blood among Bangladeshi men and women

BACKGROUND

It is well-known that methylation changes occur as humans age, however, understanding how age-related changes in DNA methylation vary by sex is lacking. In this study, we characterize the effect of age on DNA methylation in a sex-specific manner and determine if these effects vary by genomic context. We used the Illumina HumanMethylation 450 K array and DNA derived from whole blood for 400 adult participants (189 males and 211 females) from Bangladesh to identify age-associated CpG sites and regions and characterize the location of these age-associated sites with respect to CpG islands (vs. shore, shelf, or open sea) and gene regions (vs. intergenic). We conducted a genome-wide search for age-associated CpG sites (among 423,604 sites) using a reference-free approach to adjust for cell type composition (the R package RefFreeEWAS) and performed an independent replication analysis of age-associated CpGs.

RESULTS

The number of age-associated CpGs (p < 5 x 10- 8) were 986 among men and 3479 among women of which 2027(63.8%) and 572 (64.1%) replicated (using Bonferroni adjusted p < 1.2 × 10- 5). For both sexes, age-associated CpG sites were more likely to be hyper-methylated with increasing age (compared to hypo-methylated) and were enriched in CpG islands and promoter regions compared with other locations and all CpGs on the array. Although we observed strong correlation between chronological age and previously-developed epigenetic age models (r ≈ 0.8), among our top (based on lowest p-value) age-associated CpG sites only 12 for males and 44 for females are included in these prediction models, and the median chronological age compared to predicted age was 44 vs. 51.7 in males and 45 vs. 52.1 in females.

CONCLUSIONS

Our results describe genome-wide features of age-related changes in DNA methylation. The observed associations between age and methylation were generally consistent for both sexes, although the associations tended to be stronger among women. Our population may have unique age-related methylation changes that are not captured in the established methylation-based age prediction model we used, which was developed to be non-tissue-specific.

Screening for genes that accelerate the epigenetic aging clock in humans reveals a role for the H3K36 methyltransferase NSD1

BACKGROUND

Epigenetic clocks are mathematical models that predict the biological age of an individual using DNA methylation data and have emerged in the last few years as the most accurate biomarkers of the aging process. However, little is known about the molecular mechanisms that control the rate of such clocks. Here, we have examined the human epigenetic clock in patients with a variety of developmental disorders, harboring mutations in proteins of the epigenetic machinery.

RESULTS

Using the Horvath epigenetic clock, we perform an unbiased screen for epigenetic age acceleration in the blood of these patients. We demonstrate that loss-of-function mutations in the H3K36 histone methyltransferase NSD1, which cause Sotos syndrome, substantially accelerate epigenetic aging. Furthermore, we show that the normal aging process and Sotos syndrome share methylation changes and the genomic context in which they occur. Finally, we found that the Horvath clock CpG sites are characterized by a higher Shannon methylation entropy when compared with the rest of the genome, which is dramatically decreased in Sotos syndrome patients.

CONCLUSIONS

These results suggest that the H3K36 methylation machinery is a key component of the epigenetic maintenance system in humans, which controls the rate of epigenetic aging, and this role seems to be conserved in model organisms. Our observations provide novel insights into the mechanisms behind the epigenetic aging clock and we expect will shed light on the different processes that erode the human epigenetic landscape during aging.

Growth disrupting mutations in epigenetic regulatory molecules are associated with abnormalities of epigenetic aging

Germline mutations in fundamental epigenetic regulatory molecules including DNA methyltransferase 3 alpha (DNMT3A) are commonly associated with growth disorders, whereas somatic mutations are often associated with malignancy. We profiled genome-wide DNA methylation patterns in DNMT3A c.2312G > A; p.(Arg771Gln) carriers in a large Amish sibship with Tatton-Brown–Rahman syndrome (TBRS), their mosaic father, and 15 TBRS patients with distinct pathogenic de novo DNMT3A variants. This defined widespread DNA hypomethylation at specific genomic sites enriched at locations annotated as genes involved in morphogenesis, development, differentiation, and malignancy predisposition pathways. TBRS patients also displayed highly accelerated DNA methylation aging. These findings were most marked in a carrier of the AML-associated driver mutation p.Arg882Cys. Our studies additionally defined phenotype-related accelerated and decelerated epigenetic aging in two histone methyltransferase disorders: NSD1 Sotos syndrome overgrowth disorder and KMT2D Kabuki syndrome growth impairment. Together, our findings provide fundamental new insights into aberrant epigenetic mechanisms, the role of epigenetic machinery maintenance, and determinants of biological aging in these growth disorders.

Epigenetic clock analysis of human fibroblasts in vitro: effects of hypoxia, donor age, and expression of hTERT and SV40 largeT

Aging is associated with a genome-wide change of DNA methylation (DNAm). "DNAm age" is defined as the predicted chronological age by the age estimator based on DNAm. The estimator is called the epigenetic clock. The molecular mechanism underlining the epigenetic clock is still unknown. Here, we evaluated the effects of hypoxia and two immortalization factors, hTERT and SV40-LargeT (LT), on the DNAm age of human fibroblasts in vitro. We detected the cell division-associated progression of DNAm age after >10 population doublings. Moreover, the progression of DNAm age was slower under hypoxia (1% oxygen) compared to normoxia (21% oxygen), suggesting that oxygen levels determine the speed of the epigenetic aging. We show that the speed of cell division-associated DNAm age progression depends on the chronological age of the cell donor. hTERT expression did not arrest cell division-associated progression of DNAm age in most cells. SV40LT expression produced inconsistent effects, including rejuvenation of DNAm age. Our results show that a) oxygen and the targets of SV40LT (e.g. p53) modulate epigenetic aging rates and b) the chronological age of donor cells determines the speed of mitosis-associated DNAm age progression in daughter cells.