Age is simply a state of mind. You’re only as old as you think you are. Age is just a number.
The adages pile up and yet, for many, it is impossible to look beyond the noises our bodies make simply by rising from the couch, the number of trips to the restroom each night and the extra time it takes to focus on words on a page or fleeting thoughts in our heads.
We monitor our diet. We exercise. We reduce stress.
We are careful about environmental impacts, spending a fortune on air purifiers, water filters and organic foods.
We cannot (yet) change our genetics and familial legacies, but at a molecular level, these choices can have an impact. Although it may be impossible to alter our chronological age, we may yet be able to modulate our biological age through efforts reflected in epigenetic changes in our genetic machinery.
Time flies
That the epigenetic profiles of organisms change quite naturally over time is nothing new, according to Keith Booher, epigenetics service and product manager at Zymo Research.
“Developmental biologists have known for a long time that DNA methylation changes dramatically from the moment of conception,” he explains. “A fertilized single-celled zygote will have most of its DNA methylation pretty much erased.”
As the cells divide and the embryo grows during development, he continues, the patterns are then re-established. Thus, although every single cell in the body has the exact same genetic information, organs and tissues have very diverse functions.
“I think what we’re seeing now with [Steve] Horvath’s work is that this developmental process never really stops,” Booher enthuses.
In an epigenetic analysis of almost 8,000 samples covering 51 tissue and cell types, as well as almost 6,000 cancer samples, UCLA’s Steve Horvath identified a subset of 353 specific DNA methylation markers that accurately predicted the age of the sample donor. Horvath found that the markers could be divided into two subgroups based on how they correlated with age.
“The 193 positively and 160 negatively correlated CpGs get hypermethylated and hypomethylated with age, respectively,” he wrote back in 2013.
Horvath saw the potential for DNAm age, as he called it, as a surrogate to monitor rejuvenation therapies. But perhaps more important, he suggested, was its applicability to a wide variety of cell and tissue types.
“Since it allows one to contrast the ages of different tissues from the same subject, it can be used to identify tissues that show evidence of accelerated age due to disease (for example, cancer),” he concluded.
To a large extent, Horvath’s predictions came true as his epigenetic clock model, as well as those developed by others, have become fundamental tools of age-related research and have even started to enter the direct-to-consumer market.
Last summer, for example, Epimorphy started offering myDNAge as a biological aging resource to consumers, in a model similar to the more familiar 23andMe mutational analysis resource. Epimorphy licensed the technology from Zymo Research, which continues to provide its research-only DNAge test to scientists performing studies with human samples, as well as mice.
“In terms of function, age-associated CpGs in humans and mice seem to be enriched in genes that are involved in morphogenesis and development,” Wolfgang Wagner of the University Hospital of the RWTH Aachen recently wrote. “However, in both species age-associated DNAm changes are not generally reflected at the gene expression level—and thus the biological relevance remains largely unclear.”
A complicating factor in aging-associated epigenetic studies, as with so many other pursuits, is the translation of results from in-vitro and animal studies to humans.
“Given the timespan of human generations and events involving manipulation of a real human environment, these studies are extremely low in number,” suggests Terry Kelly, director of R&D at Active Motif. A second major challenge, she continues, is an inability to employ canonical epigenetic features in the face of psychosocial, lifestyle and nutritional confounders.
“A new term—the epigenetic landscape—has being coined to include address of non-canonical features (e.g., non-coding or nc-RNAs) that interconnect environmental change with actual biochemical regulatory networks,” Kelly says. “This is well documented for nutrition and has translated well into clinical studies.”
Despite having been studied for upward of 50 years, she continues, it is only now that the understanding of human epigenetic regulation is starting to account for life course variation, health span (the period over which an individual experiences no major health issues) and the newly emerging awareness of the role of the microbiome in human health.
Disease, not age
As the concept of health span would indicate, however, aging itself is not a disease state—Western society’s fixation on and fear of it notwithstanding. It is, however, a predominant risk factor for a variety of medical conditions.
As Paul Shiels of the Glasgow Ageing Research Network explains, human aging is associated with chronic inflammation—also known as inflammaging—a proven risk factor for morbidity and mortality.
“Many age-related morbidities present with an underlying component of low-grade inflammation, though its etiology remains undetermined,” he continues. “Typically, it manifests with an increased frequency of age-related complications, such as vascular stiffening, osteoporosis, muscle wasting, depression, cognitive dysfunction and frailty.”
This results in a loss of physical capability and physiological function, Shiels notes.
“However, the degree of inter-individual variation in individuals of the same chronological age is substantial and the causes of this variation are multifactorial,” he says.
Shiels differentiates this chronological aging from biological aging.
“A growing body of evidence has indicated that social, psychological, lifestyle and nutritional risk factors influence the trajectory of age-related health and age-related morbidities, such as chronic kidney disease, by acting either independently, cumulatively or synergistically with an individual’s genetics, and in particular epigenetics, thus determining health span,” he presses.
Shiels offers examples of recent evidence that suggests epigenetic regulation of nutrient-sensing pathways and nutritional differences tied to socioeconomic position can differentially affect the aging process, and in particular, age-related genomic hypomethylation and inflammatory status.
The link between epigenetic profiles and disease seems to be something of a balance between chronology and environmental exposure.
Gene expression change can be quite responsive to environmental inputs, Booher says, but the epigenetic responses are more robust. The dynamics of a DNA methylation mark just couldn’t respond that fast, in his opinion.
Cambridge Epigenetix CEO Jason Mellad concurs, suggesting that the dynamics of epigenetic changes really comes down to the types of questions you are asking.
“If you want to identify epigenetic marks that tell me whether I have cancer or not, you’re looking for something that’s fairly robust, that is consistently going to say yes, this cell is cancerous,” he explains. “If you’re asking the question of 'well, what happens if I’ve exposed a cell to a toxic stimulus in the first couple of hours,' you’re going to sequence marks that might be more dynamic.”
Mellad offers the example of the exposure of cells to carcinogens on a regular basis and how this might lead to epigenetic changes that prime those cells to become cancerous. Specifically, he points to a recent paper by Stephen Baylin of the Sidney Kimmel Comprehensive Cancer Center and colleagues.
In brief, the researchers bathed human bronchial epithelial cells in cigarette smoke condensate and then performed genome-wide DNA-methylation and gene-expression analysis to understand the impact of exposure.
“What they found is, if you transfected these cells after 15 months of this liquid smoke treatment with oncogenes versus the controls, the cells that had been exposed to the carcinogen become tumors when you inject them into immunocompromised mice,” Mellad recounts. “They were able to identify epigenetic changes due to the liquid smoke exposure, which actually made the cells more likely to become cancerous when that mutation occurred.”
Aside from cancer, increased risk of neurodegenerative disease is probably most associated with aging, and there is an increasing body of evidence to suggest that many neuropathologies have some connection to changes in epigenetic patterning.
Recently, the University of Exeter’s Jonathan Mill and colleagues looked for links between histone acetylation patterns and Alzheimer’s disease (AD).
“Building on the hypothesis that epigenomic dysregulation is important in the etiology and progression of AD neuropathology, we and others recently performed the first genome-scale cross-tissue analyses of DNA methylation in AD, identifying robust DNA methylation differences associated with AD neuropathology across multiple independent human post-mortem brain cohorts,” the authors explained. “To date, however, no study has systematically examined other types of regulatory genomic modifications in AD.”
To fill this gap, the researchers performed chromatin immunoprecipitation followed by sequencing (ChIP-Seq) to determine the levels of a specific histone modification—H3K27ac—in post-mortem brain samples from AD patients and controls, age 58 to 97 years. Of the more than 180,000 peaks identified, 4,162 were AD-associated differentially acetylated, 60 percent of those hypo-acetylated.
The researchers then probed deeper to see if any of the differentially acetylated peaks were associated with genes known to be involved in AD pathology, and they discovered hyperacetylation near the tau and presenilin genes, as well as hypoacetylation near the amyloid precursor protein gene.
“Given that histone-acetylation modifiers are amongst the most promising target pharmacological treatments of AD, the identification of altered H3K27ac in AD is important, giving clues as to which genes and pathways may be involved,” the authors wrote.
That said, the researchers were quick to acknowledge the limitations of this study, not the least of which was the breadth of epigenetic mechanisms they did not examine, noting: “Although profiling H3K27ac can provide relatively robust information about transcriptional activity, it represents only one of perhaps ~100 post-translational modifications occurring at >60 histone amino-acid residues regulating genomic function.”
Another challenge in these studies is separating the role of genetic and environmental influences from epigenetic ones. To some extent, the problem is addressed by studies involving monozygotic twins, where the genetics are identical and at least the early environmental exposure is similar.
Late last year, Aarhus University’s Anna Starnawska and colleagues published their search for correlations between epigenetic patterns and cognitive decline in middle-aged monozygotic twins recruited by The Danish Twin Registry.
“Among various age-related changes taking place in the brain and multiple factors influencing them, there is a growing body of evidence for a crucial role of DNA methylation in memory formation, maintenance, synaptic plasticity and also cognitive functioning,” the authors explained.
As examples, they pointed to deficiencies in DNA methyltransferases (DNMTs) that were linked to defects in synaptic plasticity, learning and memory in mice, as well as deficiencies in Tet-1—an enzyme that oxidizes 5-methylcytosine (5mC)—associated with impaired spatial learning and memory, also in mice.
The researchers undertook an epigenome-wide association study of decade-spanning changes in cognitive abilities and DNA methylation patterns in blood samples from 243 pairs of monozygotic twins.
The group identified numerous CpG sites where cognitive function changes were associated with differences in DNA methylation patterns. Of particular note were altered patterns in AGBL4, a gene linked to neuronal survival, and SORBS1, a gene previously implicated in Alzheimer’s disease and ischemic stroke.
In defining the limitations of their findings, however, the authors were quick to caution: “Further research is warranted to answer the question if the associated differences in DNA methylation levels are causal of cognitive functioning or are a consequence of any disease’s pathological process.”
Methyl mania
Although epigenetic modification comes in many forms, much of the focus of the last few decades has been on the profiles and changing patterns of DNA methylation.
Awareness of “DNA methylation came first and thus has had longer to be adopted by disease-based researchers,” explains Kelly. “In addition, to study DNA methylation, one only needs DNA, which is relatively stable independent of how samples are stored.”
“Preparation of high-quality chromatin requires a relatively high level of expertise,” she continues. “There are hundreds of histone [post-translational modifications] with as many proteins regulating them.”
“As a result, chromatin regulation is more combinatorial, and scientists often have to look at multiple features to understand the biological relevance.”
As the interest in chromatin and nucleosomes grows, Kelly continues, so will the availability of disease-based material for its study.
“Histone modification, nucleosome assembly, chromatin organization in 3D in general, these are all important aspects of epigenetics,” echoes Booher. “These modifications cross-talk among one another.”
“So, DNA methylation can be a signal to lay down DNA methyl-binding proteins, and there’s a sequential process to organizing the chromatin in a certain way based on the patterns of those DNA methylation marks and methyl-binding proteins that help to regulate gene expression,” he continues.
That said, Booher acknowledges that the DNA methylation mark is the most abundant in the cell, and so from that perspective, it’s the most important epigenetic modification.
DNA methylation analysis is very much at the heart of Cambridge Epigenetix, which early on recognized a challenge with traditional bisulfite sequencing chemistry, according to Mellad.
“Basically, there are multiple types of DNA modifications, and bisulfite chemistry cannot distinguish between some of the major ones that are functionally distinct,” he explains.
“We’re particularly interested in one mark called 5-hydroxymethylcytosine (5hmC), which is generated as a result of enzymatic oxidation of the methylcytosine (5mC) mark itself,” Mellad continues. “DNA methyltransferase adds a methyl group to cytosine and then another enzyme called TET comes and oxidizes that methyl group to hydroxyl, or it can go to formylC, which is an aldehyde, or carboxyC, which is a carboxylic acid.”
According to Mellad, 5hmC is highly associated with active enhancers and active gene expression, whereas 5mC is the opposite and is highly associated with closed chromatin and repressed gene expression.
“You can imagine, you have a data set and you can’t distinguish between the marks; you’re trying to correlate your epigenetic data to your gene-expression data, but things aren’t adding up quite the way you’d expect,” he offers. “You scratch your head. It causes difficulties and it really hampers progress in the field.”
To address this problem, company founder Shankar Balasubramanian developed oxidative bisulfite sequencing (oxBS-Seq), which allowed researchers to distinguish the different methylation types by selectively oxidizing 5hmC to 5-formylcytosine and then to uracil, while leaving 5mC unchanged. The technology became the foundation of the company’s TrueMethyl platform.
Because Cambridge Epigenetix is more interested in leveraging their expertise in the diagnostic space, it signed an agreement last summer to integrate oxBS-Seq with the next-generation sequencing portfolio of NuGEN Technologies.
“After I was appointed as CEO last year, one of the first decisions I made was to no longer sell kits directly to the research-use-only market,” Mellad explains. “So, I was looking for a partner who was of the right size so that we have global reach but not so large that it made a partnership challenging; the right partner who had an innovative set of technologies in their own right, but were synergized with ours.”
NuGEN and its portfolio were the perfect match.
The same process applies as Cambridge Epigenetix looks for pharma partners, he adds.
“We’re really focused on the liquid biopsy diagnostic tests we’re developing, but at the same time, we’re open to working with different partners, particularly in the pharma and biotech world, who want to avail themselves of our platform.”
From Booher’s perspective, the need to differentiate between 5mC and 5hmC really depends on the nature of the system you’re examining.
“In some cases, that difference may just not matter, whether it’s methylated or hydroxymethylated,” he says. “You may just need to know whether it is modified. For other applications, that may not be good enough. Not only is it modified, but I need to know which modification.”
Booher adds that Zymo hasn’t really seen much interest in the distinction between the two methylation types.
Regardless, a better awareness of the epigenetic profiles and molecular links of these patterns to age-related disease opens the possibility of developing epigenetics-focused diagnostics and therapeutics.
Rx/Dx opportunities
In a 2016 review, Weill Cornell Medicine’s Sangita Pal and Jessica Tyler discussed the opportunities for anti-aging intervention, specifically targeting epigenetic pathways and processes.
They described dietary or calorie restriction (DR or CR) as the most accessible means toward lifespan extension, having demonstrated profound effects in various organisms up to non-human primates. They acknowledged, however, that CR adherence represents a problem in humans. Thus, Pal and Tyler suggested that the search was on for CR mimetics, therapeutic molecules that produced the same outcome as CR.
“Given that the sirtuins promote longevity in diverse species, they have attracted considerable interest as drug targets,” they wrote, highlighting resveratrol.
“Resveratrol and synthetic STACs [sirtuin-activating compounds] have been shown to induce physiological changes and gene expression changes that are similar to those that are induced by CR and to extend life span in many model organisms,” the authors noted. “Promisingly, the synthetic STACs SRT1720 and SRT2104 extend the life span of obese mice and protect against age-related changes in multiple tissues.”
Similarly, Pal and Tyler pointed to metformin, noting that diabetes is now considered an age-related disease. In particular, they highlighted a recent study of more than 180,000 people that showed diabetes patients treated with metformin not only lived longer than other diabetes patients, but also healthy control subjects.
“Further exploration is needed to fully understand how metformin mechanistically extends life span in humans,” they cautioned. “However, given that metformin is approved by the U.S. Food and Drug Administration for use in humans, commercially available and relatively cheap, it has the potential to be revolutionary for increasing human health span and life span.”
Similarly, the Bellvitge Biomedical Research Institute’s Maria Berdasco and colleagues recent reviewed the therapeutic opportunities for epigenetic interventions in treating human brain disorders.
“The idea of effective stimulation of neuronal production by using epidrugs is highly attractive, and although in its infancy, it is supported by several lines of evidence,” the researchers suggested. “Interestingly, pharmacological inhibition of HDAC [histone deacetylase] activity alters neuronal differentiation. It has been reported that treatments with trichostatin A or valproic acid induced neuronal differentiation in adult progenitor cells.”
“Pharmacological manipulation of chromatin complexes is also an alternative,” they continued. “The histone-interacting BET bromodomain proteins are downregulated during neurogenesis from [neuronal stem precursor cells], and the use of a bromodomain-selective inhibitor results in an increase in neuronal differentiation.”
They also noted that in animal and human post-traumatic stress disorder, several studies have demonstrated success in extinguishing fear responses through enhancement in histone acetyltransferase (HAT) activity, reduction in HDAC activity and changes in DNA methylation mediated by TET enzymes.
Given the ubiquity of drugs and clinical candidates targeting many of these pathways in the treatment of cancer, the number of trials testing these compounds in age-related conditions outside of oncology can only be expected to grow.
Mellad, however, also sees opportunities arising from the use of epigenetic tools to change the way we approach therapeutic decisions.
“We now have the epigenetic sequencing tools available to measure what’s happening in a system—and not only can we start to predict who will and won’t respond to drug A, we might be able to specifically alter the epigenome of the non-responders to make them respond,” he explains. “Or we might be able to target the epigenome of the disease patients themselves to directly treat the disease.”
As an example, he describes the MGMT promoter methylation assay, which is run as a standard of care in glioblastoma. MGMT is an enzyme that repairs DNA damage.
“If your promoter is methylated, studies have shown that you are less likely to express the enzyme, which means that if you treat your brain cancer with a drug like temozolomid, which is an alkylating DNA-damaging agent, then the cell won’t be able to repair the damage and it’s more likely to be susceptible,” Mellad continues. “If your promoter is not methylated, chances are you’re expressing this enzyme and you’ll be resistant to the chemotherapy.”
Booher likewise points to work being done at Exact Sciences and MDxHealth to produce diagnostic tests based on epigenetic patterns.
Exact’s Cologuard platform relies, in part, on the detection of oncogenic mutations as well as specific methylation markers via the company’s Quantitative Allele-specific Real-time Target and Signal Amplification (QuARTS) technology.
For its part, MDxHealth uses what it describes as real-time methylation-specific PCR to identify epigenetic patterns in specific CpG islands. Its lead product is the ConfirmMDx test for prostate cancer, but it is also pursuing other tests in urology such as the AssureMDx test in bladder cancer and early-stage work in kidney cancer. In a study described last summer, researchers suggested that AssureMDx could spare three-quarters of hematuria patients highly invasive cystoscopy-based diagnoses.
Unfortunately, in parallel to the therapeutic opportunities to ameliorate or reverse age-related diseases are the age-accelerating impacts of therapies for other conditions such as cancer.
In a recent review, the Mayo Clinic’s Shahrukh Hashmi and colleagues examined premature aging in cancer survivors.
“The Childhood Cancer Survivor Study (CCSS) compared the health status of 9,353 adult long-term survivors of childhood cancer diagnosed between 1970 and 1986 to the health status of a randomly selected cohort of the survivors’ siblings,” the authors wrote. “Survivors were more likely to report poor general health, poor mental health, activity limitations and functional impairment.
“Additionally, 20-year-old survivors of childhood cancer were found to have the same cumulative incidence of severe, life-threatening and fatal chronic health conditions as 50-year-old siblings.”
Similar patterns have also been seen in adult cancer survivor populations, they continued, including evidence of premature neurocognitive decline, osteoporosis, sexual dysfunction and chronic fatigue.
“While aging prematurely is a better alternative to dying prematurely, a better understanding of what drives this process presents an opportunity for improvement,” the researchers argued.
In their analysis, they pointed to drugs that induced cytotoxicity via CpG hypermethylation, noting that although this effect is advantageous to cancer treatment, it may also be a negative side effect to the otherwise healthy cells in these patients.
On the other side were drugs that led to DNA hypomethylation—e.g., azacitidine, decitabine and hydralazine—that might also accelerate aging in healthy cells or possibly promote new secondary cancers.
Again, the authors did not advocate for the termination of the cancer treatments so much as recognition of the potential harm and understanding its possible mechanisms underlying. Such awareness, they argued, would offer an opportunity to either develop therapeutic alternatives or treatments to ameliorate the effects.
Thus, as the tools to identify epigenetic changes expand, so too do the variety and impact of the changes, demanding an even greater arsenal for an ever-widening audience.
“We want to turn epigenetics from a field that’s been relatively niche into a mainstream activity,” offers Mellad. “I envision a future in which every sequencing run gives you not only genetic information, but also epigenetic because it’s so important.”
--------------------------------------------------------------
Fountain of youth
The study of age-related diseases has always been problematic, according to Paolo Garagnani and colleagues at Università di Bologna.
“The need for large cohorts of patients at different ages or for long-term follow-up studies, and the limited availability for certain tissues, made these studies particularly challenging,” the researchers wrote in a recent review.
As with so many other fields, however, the advent of induced pluripotent stem cells (iPSCs) was expected to revolutionize the field of cell-based disease modeling, and did facilitate the study of disorders such as Alzheimer’s and Parkinson's disease.
More recently, however, questions have arisen as to how well these induced cells reflect the molecular profiles of the patient tissues they are meant to model.
“Cell reprogramming exerts a profound remodelling of DNAm [DNA methylation] profiles,” the authors wrote. “All the somatic adult cell lines possess a specific DNAm landscape that acts as an epigenetic fingerprint being indicative for their type, age and fate.
“Most (but not all) of these tissue-specific DNAm signatures are altered to favour the establishment of an human embryonic stem cell-like (hESC) DNAm landscapes.”
According to Garagnani’s group, two studies independently looked at reprogramming-associated changes to DNA methylation age using two different epigenetic clocks. The studies suggested that the predicted DNA methylation age of iPSCs derived from tissues of different chronological ages were reset from the donor’s epigenetic age to effectively zero. When those same iPSCs were differentiated back to their parental cell type, however, their DNAm age did not increase and only minor changes were noted with extended culturing.
This idea of induction leading to stem cell de-aging makes complete sense to Zymo Research’s epigenetics service and product manager, Keith Booher.
“If you want to take whatever age the person is and dedifferentiate their cells to a pluripotent state, you’re resetting that epigenetic program,” he says. “If you’re looking at aging disease models, then that could be a potentially huge confounding factor.”
Although he does not consider himself to be expert in iPSCs, Cambridge Epigenetix CEO Jason Mellad wonders if transdifferentiation directly from one cell type to another without going through a pluripotent intermediate might not help maintain age-related epigenetic marks.
“I can envision a world where you force the fibroblast to turn into, say, a cardiomyocyte directly, without first taking it all the way back to an iPSC and then differentiating,” he says.
Chun-Li Zhang and colleagues at UT Southwestern tested this very hypothesis in the neurodegenerative space, publishing their findings late last year.
“Aging is by far the most critical risk factor in many forms of late-onset neurodegenerative disorders,” they wrote. “Thus, it is particularly important to recapitulate age-related characteristics by using neurons that reflect the age at which the disease develops.”
They noted that neurons differentiated from iPSCs were generally known to be immature and often required long-term culture to develop mature functional properties. As well, they recounted that many studies suggested “iPSC-derived neurons only recapitulated early biochemical phenotypes without mimicking the late-onset severe neurodegenerative hallmarks of the disease.”
To avoid these issues, the researchers examined fibroblasts and their respective iPSCs, spanning two donor ages: young (age 3 months to 3 years) and old (age 53 to 71 years).
They noted that several age-related markers prevalent in the donor cells were greatly decreased in the iPSCs, highlighting that the cells had experienced a degree of de-aging.
The researchers then devised a single protocol that would allow them to generate motor neurons (MNs) from both donor fibroblasts (Fib-iMNs) and iPSCs (iPSC-MNs), ensuring all cells experienced the same conditions.
Examining the resulting MNs, they found that regardless of whether the iPSCs came from young or old donors, the resulting iPSC-MNs were indistinguishable in terms of aging-related factors. In complete contrast, however, the epigenetic profiles of Fib-iMNs clearly correlated with the ages of the donor fibroblasts.
“Our results also raise concerns on using iPSCs to model age-dependent neurodegenerations,” the authors concluded. “Such concerns have already drawn attention, as strategies are emerging to induce aging-like features in iPSC-derived cells, for example, by telomerase manipulation, progerin treatment or prolonged culture.
“Clearly, a cell culture system maintaining aging features such as our directly reprogrammed MNs will greatly facilitate our understanding of neurodegeneration and therapeutic identification and validation.”
Like the Horvath age clock itself, Garagnani and colleagues were quick to highlight, it remains unclear as to whether erasing age-associated DNA methylation patterns is functionally associated with specific physiological changes in a cell.
Adding to the complexity, however, is the idea that even when dedifferentiated to an embryonic state, the stem cells retain some epigenetic memory of their original state, says Mellad.
“So, for example, if you took a fibroblast or a T cell or a neuron and you ran them all through the same iPSC protocol, there appears to still be a shadow of epigenetic information that the cell can remember that it originally was a different cell type. It’s not completely erased,” he explains. “It’s not enough to say 'I’ll just take any cell type and I can make my iPSCs.'”
“It may turn out that a fibroblast-derived iPSC is very different from a blood-derived iPSC,” Mellad opines.
Thus, it seems, the iPSC fountain of youth may yet offer some intractable epigenetic wrinkles.
