Measures of biological age based on epigenetic marks, protein levels, transcriptomic profiles, and similar collections of biological data are proliferating rapidly. The first epigenetic clock, a weighted combination of DNA methylation status at numerous CpG sites, is barely a decade old. The results correlate quite tightly with chronological age, but it was quickly established that people with epigenetic ages greater than chronological age tend to exhibit a greater risk of mortality and presence of age-related disease, and vice versa. More clocks followed, and the diversity of data used to generate these assessments of age increased along the way.
All of these approaches to measuring the burden of age suffer the same issue: they are disconnected from the well established causative mechanisms of aging, from cellular senescence to mitochondrial dysfunction. It is near entirely unknown as to how the specifics of the epigenome, proteome, or transcriptome used in these clocks are determined by mechanisms of aging. The clocks produce an outcome, but there is no way to predict in advance how the outcome will change in response to specific interventions, or whether such changes are in any way an accurate reflection of the impact of an intervention on aging.
For example, perhaps some clocks are largely measures of inflammatory status and downstream effects of chronic inflammation. Interventions that reduce inflammation would produce impressive results, while others would not. But inflammation is only one aspect of aging. There are other mechanisms that are just as important. Similarly, the clocks all have their quirks. The original epigenetic clock is insensitive to exercise, for example. It does not distinguish between fit and sedentary twins, which makes little sense given what we know of the power of exercise to influence the course of long-term health and aging. Further, epigenetic aging doesn’t correlate well with loss of telomere length.
The open access paper I’ll point out today is a different example of the quirky nature of epigenetic clocks. Researchers have found that heart tissue consistently produces younger epigenetic ages than assessments carried out in white blood cells, using a clock based on a much smaller number of CpG sites than the original epigenetic clock. The question in all such studies is the degree to which it reflects a real phenomenon – i.e. that the heart ages more slowly than the immune system – versus being an artifact of the clock, resulting from tissue-specific interactions between processes of aging and the epigenetic regulation of cellular metabolism.
People do not age at the same rate, and some of us age much more dramatically than others. Genetic and environmental factors can contribute to biological aging, which means that people may be affected differently, appearing younger or older than their birth date may predict. Consequently, age, when measured chronologically, may not be a reliable indicator of the rate of physiological breakdown of the body or organs. Indeed, individual organ systems, cells, organelles, and molecules within individuals may age at significantly different rates. Therefore, it can be postulated that even the heart may have a different aging profile to the body.
The advent of epigenome-wide high-throughput sequencing analyses has led to a successful identification of a large number of genomic sites highly associated with age. Age-predicting models have been developed and validated for an accurate “biological age” estimation. An “epigenetic clock” has been created, with unprecedented accuracy for DNAmAge estimation with an average error of only 3.6 years. Such models were based on DNA mainly derived from blood circulating leucocytes as they represent an easily available source. In this study, we applied a well studied prediction model developed on data from five CpG sites, to increase the practicability of these tests.
We have determined the biological age of the heart, specifically of the right atrium (RA) and left atrium (LA), and of peripheral blood leucocytes, by measuring the mitotic telomere length (TL) and the non-mitotic epigenetic age (DNAmAge). We found that DNAmAge, of both atrial tissues (RA and LA), was younger in respect to the chronological age (-12 years). Furthermore, no significant difference existed between RA and LA, suggesting that, although anatomically diverse and exposed to different physiological conditions, different areas of the heart had the same epigenetic non-mitotic age. Furthermore, the epigenetic age of both RA and LA, was even younger than that of the blood (-10 years).
In the present study, we demonstrated that biological age of the heart did not reflect the donor’s chronological age, while blood tracked these modifications. This would suggest that while blood is more susceptible to epigenetic changes induced by the interaction of advancing age and environmental factors, the heart is affected by these factors to a lower extent. It could be also postulated that the presence of stem cells in the cardiac muscle may explain why human heart tissue tends to have a lower DNAmAge. In fact, stem cells are found in relatively large numbers within myocardial tissue and show a DNAmAge close to zero. However, further investigation is required to elucidate the role of cardiac stem cells in determining epigenetic age of cardiac tissue and to fully understand its discrepancy with chronological age.