A slanted view of the epigenetic clock

The founder of the epigenetic clock technique was interviewed for MIT Technology Review:

“We need to find ways to keep people healthier longer,” he says. He hopes that refinements to his clock will soon make it precise enough to reflect changes in lifestyle and behavior.”


The journalist attempted to dumb the subject down “for the rest of us” with distortions such as the headline. The varying correlation of epigenetic age to chronological age was somewhat better reported in the story:

“The epigenetic clock is more accurate the younger a person is. It’s especially inaccurate for the very old.”

The journalist inappropriately used luck as a synonym for randomness/stochasticity:

“He estimates that about 40% of the ticking rate is determined by genetic inheritance, and the rest by lifestyle and luck.”

A third example of less-than-straightforward journalism started with:

“Such personalization raises questions about fairness. If your epigenetic clock is ticking faster through no fault of your own..”

Were MIT Technology Review readers unable to comprehend a straightforward story on the epigenetic clock? What was the purpose of slants and distortions in an introductory article?

https://www.technologyreview.com/s/612256/want-to-know-when-youre-going-to-die/ “Want to know when you’re going to die?”

Advertisements

Epigenetic transgenerational inheritance of ovarian disease

This 2018 Washington rodent study investigated ovarian disease in F3 great-granddaughters caused by their F0 great-grandmothers’ exposures to DDT or vinclozolin while pregnant:

“Two of the most prevalent ovarian diseases affecting women’s fertility and health are Primary Ovarian Insufficiency (POI) and Polycystic Ovarian Syndrome (PCOS). POI is characterized by a marked reduction in the primordial follicle pool of oocytes and the induction of menopause prior to age 40. POI currently affects approximately 1% of female population. While genetic causes can be ascribed to a minority of patients, around 90% of POI cases are considered idiopathic, with no apparent genetic link nor known cause.

PCOS is a multi-faceted disease that affects 6-18% of women. It is characterized by infrequent ovulation or anovulation, high androgen levels in the blood, and the presence of multiple persistent ovarian cysts.

For both PCOS and POI other underlying causes such as epigenetic transgenerational inheritance of disease susceptibility have seldom been considered. Epigenetic transgenerational inheritance is defined as “the germline transmission of epigenetic information and phenotypic change across generations in the absence of any continued direct environmental exposure or genetic manipulation.” Epigenetic factors include:

  • DNA methylation,
  • Histone modifications,
  • Expression of noncoding RNA,
  • RNA methylation, and
  • Alterations in chromatin structure.

The majority of transgenerational studies have examined sperm transmission of epigenetic changes due to limitations in oocyte numbers for efficient analysis.

There was no increase in ovarian disease in direct fetal exposed F1 [grandmothers] or germline exposed F2 [mothers] generation vinclozolin or DDT lineage rats compared to controls.

The transgenerational molecular mechanism is distinct and involves the germline (sperm or egg) having an altered epigenome that following fertilization may modify the embryonic stem cells epigenome and transcriptome. This subsequently impacts the epigenetics and transcriptome of all somatic cell types derived from these stem cells.

Therefore, all somatic cells in the transgenerational [F3] animal have altered epigenomes and transcriptomes and those sensitive to this alteration will be susceptible to develop disease. The F3 generation can have disease while the F1 and F2 generations do not, due to this difference in the molecular mechanisms involved.

The epimutations and gene expression differences observed are present in granulosa cells in the late pubertal female rats at 22-24 days of age, which is long before any visible signs of ovarian disease are detectable. This indicates that the underlying factors that can contribute to adult-onset diseases like PCOS and POI appear to be present early in life.

Ancestral exposure to toxicants is a risk factor that must be considered in the molecular etiology of ovarian disease.”


1. The study highlighted a great opportunity for researchers of any disease that frequently has an “idiopathic” diagnosis. It said a lot about research priorities that “around 90% of POI cases are considered idiopathic, with no apparent genetic link nor known cause.”

It isn’t sufficiently explanatory for physicians to continue using categorization terminology from thousands of years ago. Science has progressed enough with measured evidence to discard the “idiopathic” category and express probabilistic understanding of causes.

2. One of this study’s coauthors made a point worth repeating in The imperative of human transgenerational studies: What’s keeping researchers from making a significant difference in their fields with human epigenetic transgenerational inheritance studies?

3. Parts of the study’s Discussion section weren’t supported by its evidence. The study didn’t demonstrate:

  • That “all somatic cells in the transgenerational animal have altered epigenomes and transcriptomes”; and
  • The particular “molecular mechanisms involved” that exactly explain why “the F3 generation can have disease while the F1 and F2 generations do not.”

https://www.tandfonline.com/doi/abs/10.1080/15592294.2018.1521223 “Environmental Toxicant Induced Epigenetic Transgenerational Inheritance of Ovarian Pathology and Granulosa Cell Epigenome and Transcriptome Alterations: Ancestral Origins of Polycystic Ovarian Syndrome and Primary Ovarian Insuf[f]iency” (not freely available)

The epigenetic clock now includes skin

The originator of the 2013 epigenetic clock improved its coverage with this 2018 UCLA human study:

“We present a new DNA methylation-based biomarker (based on 391 CpGs) that was developed to accurately measure the age of human fibroblasts, keratinocytes, buccal cells, endothelial cells, skin and blood samples. We also observe strong age correlations in sorted neurons, glia, brain, liver, and bone samples.

The skin & blood clock outperforms widely used existing biomarkers when it comes to accurately measuring the age of an individual based on DNA extracted from skin, dermis, epidermis, blood, saliva, buccal swabs, and endothelial cells. Thus, the biomarker can also be used for forensic and biomedical applications involving human specimens.

The biomarker applies to the entire age span starting from newborns, e.g. DNAm of cord blood samples correlates with gestational week.

Furthermore, the skin & blood clock confirms the effect of lifestyle and demographic variables on epigenetic aging. Essentially it highlights a significant trend of accelerated epigenetic aging with sub-clinical indicators of poor health.

Conversely, reduced aging rate is correlated with known health-improving features such as physical exercise, fish consumption, high carotenoid levels. As with the other age predictors, the skin & blood clock is also able to predict time to death.

Collectively, these features show that while the skin & blood clock is clearly superior in its performance on skin cells, it crucially retained all the other features that are common to other existing age estimators.”

http://www.aging-us.com/article/101508/text “Epigenetic clock for skin and blood cells applied to Hutchinson Gilford Progeria Syndrome and ex vivo studies”


An introduction to the study highlighted several items:

“Although the skin-blood clock was derived from significantly less samples (~900) than Horvath’s clock (~8000 samples), it was found to more accurately predict chronological age, not only across fibroblasts and skin, but also across blood, buccal and saliva tissue. A potential factor driving this improved accuracy in blood could be related to the approximate 18-fold increase in genomic coverage afforded by using Illumina 450k/850k beadarrays.

It serves as a roadmap for future clock studies, pointing towards the importance of constructing tissue or cell-type specific epigenetic clocks, to more accurately measure biological aging in the given tissue/cell-type, and therefore with the potential to be more informative of disease-risk or the success of disease interventions in the tissue or cell-type of interest.”

http://www.aging-us.com/article/101533/text “Epigenetic clocks galore: a new improved clock predicts age-acceleration in Hutchinson Gilford Progeria Syndrome patients”

Organ epigenetic memory

This 2018 Japanese review subject was the relationships of organ memory and non-communicable diseases:

“Organ memory is the engraved phenotype of altered organ responsiveness acquired by a time-dependent accumulation of organ stress responses. This phenomenon is known as “metabolic memory” or “legacy effect,” which is similar to neuronal and immune memory.

Not only is the epigenetic change of key genes involved in the formation of organ memory but the alteration of multiple factors, including low molecular weight energy metabolites, immune mediators, and tissue structures, is involved as well. These factors intercommunicate during every stress response and carry out incessant remodeling in a certain direction in a spiral fashion through positive feedback mechanisms.

The systematic review revealed that each intervention type, that is:

  • Glucose lowering,
  • Blood pressure lowering, or
  • LDL-cholesterol lowering,

possessed unique characteristics of the memory phenomenon. Most of the observational periods of these studies lasted for > 10 years. Memory phenomenon was suggested to last for a long time and is thought to have a considerable effect on the clinical course of NCDs [non-communicable diseases].

Organs cannot possess consciousness, so it might not be appropriate to consider whether a recalling process exists in organs. However, the properties of organs are incessantly altered by external stimuli loaded on organs as if it is updating.

It is clinically important to investigate whether organ memory can be updated by our behaviors. Once organ memory is established in an organ, organ memory in each organ can influence one another and affect organ memory in a different organ.

Epigenome-modification enzymes, such as histone deacetylases and DNA methyltransferases, and transcription factors seem to be essential for the epigenetic regulation of gene expression, which is involved in the generation of organ memory. Cellular metabolism can epigenetically modulate the expression of genes that are related to the progression of diseases.”


The reviewers asserted:

“Organs cannot possess consciousness, so it might not be appropriate to consider whether a recalling process exists in organs.”

Memory studies don’t require this consciousness to investigate even the brain organ’s areas and functions. Researchers observe memory by measuring stimulus/response items like neuron activation and various levels of behavior. Consciousness is an emergent property.

Regarding recall: An organ’s “engraved phenotype of altered organ responsiveness” may not have recall itself, but it doesn’t have a separate existence apart from its body. An organ can’t be removed from its body for very long and still be part of its body.

When an organ is in its normal state as part of a body, it has access to recall-like functions via the “inter-organ communication of organ memory.” The review also mentioned:

“Organ memory in each organ can influence one another and affect organ memory in a different organ.

Evolution didn’t support unnecessary duplication for a kidney’s memory to include recall because it’s part of a body that includes a brain that has recall. Evolution didn’t duplicate functions of a kidney’s memory in a brain, either.

https://www.nature.com/articles/s41440-018-0081-x “Organ memory: a key principle for understanding the pathophysiology of hypertension and other non-communicable diseases” (not freely available)

Flawed epigenetic measurements of behavioral experiences

This 2018 New York rodent study not only wasted resources but also speciously attempted to extrapolate animal study findings to humans:

“While it is clear that behavioral experience modulates epigenetic profiles, it is less evident how the nature of that experience influences outcomes and whether epigenetic/genetic “biomarkers” could be extracted to classify different types of behavioral experience.

Male and female mice were subjected to either:

  • a Fixed Interval (FI) schedule of food reward, or
  • a single episode of forced swim followed by restraint stress, or
  • no explicit behavioral experience

after which global expression levels of two activating (H3K9ac and H3K4me3) and two repressive (H3K9me2 and H3k27me3) post-translational histone modifications (PTHMs), were measured in hippocampus (HIPP) and frontal cortex (FC).

A random subset of 5 of the 12 animals from each sex/behavioral experience group were used for these analyses. FC and HIPP were dissected from each of those 5 brains and homogenized for subsequent analyses. Thus, sample size for PTHM expression levels was n = 5 for each region/sex/behavioral treatment group and all PTHM expression level analyses utilized the homogenized tissue.

The specific nature of the behavioral experience differentiated profiles of PTHMs in a sex- and brain region-dependent manner, with all 4 PTHMs changing in parallel in response to different behavioral experiences. Global PTHMs may provide a higher-order pattern recognition function.”


The researchers knew or should have known that measuring “global expression levels” in “homogenized tissue” of “n = 5” subjects was flawed, and they did it anyway. They acknowledged some of the numerous study design defects with qualifiers such as:

“Even though these were global levels of histone modifications (and thus not indicative of changes at specific genes or sites on genes)..

As FS-RS behavioral experience was completed before FI behavioral experience, a longer overall post-behavior experience time (approximately 1 week) elapsed for this group, resulting in some differences in overall timing between these experiences and global PTHM assessment. However, extending the duration of the FS-RS experience (i.e., repeated exposures) would also have led to habituation..”

Did they purposely make these mistakes because of the “biomarkers” paradigm?

What would they have found if they had followed their judgments and training to design a better study? Experience-dependent histone modifications that differed by gender and brain region was certainly a promising research opportunity.

As for extrapolating the cited animal study findings to humans? Ummm..NO!

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6060276/ “Different Behavioral Experiences Produce Distinctive Parallel Changes in, and Correlate With, Frontal Cortex and Hippocampal Global Post-translational Histone Levels”

Prenatal programming of human HPA axis development

This 2017 UC Irvine human review subject provided details of how fetal hypothalamic-pituitary-adrenal components and systems develop, and how they are epigenetically changed by the mother’s environment:

“The developmental origins of disease or fetal programming model predicts that intrauterine exposures have life-long consequences for physical and psychological health. Prenatal programming of the fetal hypothalamic-pituitary-adrenal (HPA) axis is proposed as a primary mechanism by which early experiences are linked to later disease risk.

Development of the fetal HPA axis is determined by an intricately timed cascade of endocrine events during gestation and is regulated by an integrated maternal-placental-fetal steroidogenic unit. Mechanisms by which stress-induced elevations in hormones of maternal, fetal, or placental origin influence the structure and function of the emerging fetal HPA axis are discussed.

Human gestational physiology and fetal HPA axis development differ even from that of closely related nonhuman primates, thereby limiting the generalizability of animal models. This review will focus solely on studies of prenatal stress and fetal HPA axis development in humans.”


Every time I read a prenatal study I’m in awe of all that has to go right, and at the appropriate time, and in sequence, for a fetus to be undamaged. Add in what needs to happen at birth, during infancy, and throughout early childhood, and it seems impossible for a human to escape epigenetic damage.


1. The reviewers referenced human research performed with postnatal subjects, as well as animal studies, despite the disclaimer:

This review will focus solely on studies of prenatal stress and fetal HPA axis development in humans.”

This led to blurring of what had been studied or not with human fetuses regarding the subject.

2. The reviewers uncritically listed many dubious human studies that had both stated and undisclosed severe limitations on their findings. It’s more appropriate for reviewers to offer informed reviews of cited studies, as Sex-specific impacts of childhood trauma summarized with cortisol:

“Findings are dependent upon variance in extenuating factors, including but not limited to, different measurements of:

  • early adversity,
  • age of onset,
  • basal cortisol levels, as well as
  • trauma forms and subtypes, and
  • presence and severity of psychopathology symptomology.”

3. It would have been preferable had the researchers stayed with their stated intention and critically reviewed only a few dozen studies with solid evidence of the review title: “Developmental origins of the human hypothalamic-pituitary-adrenal axis.” Let other reviews cover older humans, animals, and questionable evidence.

I asked the reviewers to provide a searchable file so that their work could be better used as a reference.

https://www.researchgate.net/publication/318469661_Developmental_origins_of_the_human_hypothalamic-pituitary-adrenal_axis “Developmental origins of the human hypothalamic-pituitary-adrenal axis” (registration required)

Hidden hypotheses of epigenetic studies

This 2018 UK review discussed three pre-existing conditions of epigenetic genome-wide association studies:

“Genome-wide technology has facilitated epigenome-wide association studies (EWAS), permitting ‘hypothesis-free’ examinations in relation to adversity and/or mental health problems. Results of EWAS are in fact conditional on several a priori hypotheses:

  1. EWAS coverage is sufficient for complex psychiatric problems;
  2. Peripheral tissue is meaningful for mental health problems; and
  3. The assumption that biology can be informative to the phenotype.

1. CpG sites were chosen as potentially biologically informative based on consultation with a consortium of DNA methylation experts. Selection was, in part, based on data from a number of phenotypes (some medical in nature such as cancer), and thus is not specifically targeted to brain-based, stress-related complex mental health phenotypes.

2. The assumption is often that distinct peripheral tissues are interchangeable and equally suited for biomarker detection, when in fact it is highly probable that peripheral tissues themselves correspond differently to environmental adversity and/or disease state.

3. Analyses result in general statements such as ‘neurodevelopment’ or the ‘immune system’ being involved in the aetiology of a given phenotype. Whether these broad categories play indeed a substantial role in the aetiology of the mental health problem is often hard to determine given the post hoc nature of the interpretation.”


The reviewers mentioned in item #2 the statistical flaw of assuming that measured entities are interchangeable with one another. They didn’t mention that the problem also affected item #1 methodologies of averaging CpG methylation measurements in fixed genomic bins or over defined genomic regions, as discussed in:

The reviewers offered suggestions for reducing the impacts of these three hypotheses. But will doing more of the same, only better, advance science?

Was it too much to ask of researchers whose paychecks and reputations depended on a framework’s paradigm – such as the “biomarker” mentioned a dozen and a half times – to admit the uselessness of gathering data when the framework in which the data operated wasn’t viable? They already knew or should have known this.

Changing an individual’s future behavior even before they’re born provided one example of what the GWAS/EWAS framework missed:

“When phenotypic variation results from alleles that modify phenotypic variance rather than the mean, this link between genotype and phenotype will not be detected.”

DNA methylation and childhood adversity concluded that:

“Blood-based EWAS may yield limited information relating to underlying pathological processes for disorders where brain is the primary tissue of interest.”

The truth about complex traits and GWAS added another example of how this framework and many of its paradigms haven’t produced effective explanations of “the aetiology of the mental health problem”

“The most investigated candidate gene hypotheses of schizophrenia are not well supported by genome-wide association studies, and it is likely that this will be the case for other complex traits as well.”

Researchers need to reevaluate their framework if they want to make a difference in their fields. Recasting GWAS as EWAS won’t make it more effective.

https://www.sciencedirect.com/science/article/pii/S2352250X18300940 “Hidden hypotheses in ‘hypothesis-free’ genome-wide epigenetic associations”