For four years, biostatistician Steve Horvath (Steve Horvath) from the University of California at Los Angeles has identified a set of 353 methylation zones that correlate best with the chronological age of a person. Genes localised in these regions are activated and inactivated at specific stages of life. Computer analysis of DNA samples (from blood, skin, or even urine) can determine a person’s age with an accuracy of about 2 years. There are two reasons why Horvath’s proposed clocks are of some importance. Firstly, they are the best existing method for determining a person’s biological ageing and provide objective results for assessing the effectiveness of anti-ageing interventions. Suppose we have a promising new drug and we want to know whether it makes a person younger. Before the Horvath Clock, it had to be given to thousands of people and they had to wait decades to see that fewer of them would die than people who didn’t take the drug. The Horvath watch shortens the waiting period considerably. The drug can be given to a small number of people and their age can be estimated by methylation before and after the intervention. Only a few dozen people can take the drug for two years to get a reasonably accurate idea of its efficacy. Second, there is evidence and theoretical support for the idea that the methylation sites identified by Horvath are not just markers of aging, but trigger its mechanisms. This means that if we can find a way to get inside the cell nucleus and alter methylation profiles on chromosomes, we could impact the root cause of aging. (Before we get too enthusiastic: ‘gene therapy’ was invented about 20 years ago but is still under development; we need ‘epigenetic therapy,’ which still doesn’t exist but is technically possible using genetically modified viruses and CRISPR technology.) Below is a recording of two talks Horvath gave at the US National Institutes of Health offices in 2016 in Maryland and last month in Los Angeles. In 2012-2013, 3 papers were published describing the idea that the underlying cause of aging in humans and other higher animals is an epigenetic programme. Genes are activated and inactivated at different stages of life, which triggers the processes of growth, development and aging, forming a continuous sequence. (The fourth article, authored by Favoured, expressed a similar idea. It focuses on the role of a single transcription factor, the regulator of expression of the gene encoding the mammalian target protein of rapamycin, mTOR, and sweeps away the conclusion that, in terms of natural selection, aging is the preferred outcome.) This powerful hypothesis simultaneously provides explanations for evolutionary and metabolic issues. It provides recommendations for anti-aging research – even though epigenetics has proven to be such a complex field that practical modification of gene expression change sequences requires a great deal of basic work. Without the knowledge of experts who have long worked on such theoretical treatises, Steve Horvath had already laboured to calibrate and evaluate the effectiveness of the aging clock and published basic results by the end of 2013. One important property of Horvath’s clock is that it predicts with greater accuracy than chronological age which individuals will develop age-related diseases and who will die earlier than their peers. Even though the watch is based on an algorithm that makes the output age as close as possible to the chronological age, the result is guaranteed to provide more information than the chronological age. ‘In developing the watch, chronological age was used as an approximation of biological age.’ People whose ‘methylation age’ is greater than their chronological age are more likely to have poorer health and die more quickly than people whose methylation age is less than their chronological age.’ Horvath has put his methodology and computer programme in the public domain. Based on Horvath’s clock, a California company last year began offering a commercial DNA methylation age-determination test. The test requires sending a blood or urine sample to Zymo Research.

Horvath tells us that he arrived at using an elimination process, starting with four candidates for the aging clock:

1. Telomere length;
2. gene expression profile;
3. proteomic data;
4. DNA methylation.

Telomere length This indicator was learnt to be estimated quickly and cheaply more than 10 years ago, but its correlation with chronological age (and with mortality) is not strong enough to be used as a biological clock. Gene expression profile Which genes are transcribed in human RNA at specific times? This is determined by RNA isolation, and the resulting data are highly specific to a particular tissue. In other words, they vary depending on which part of the body is being studied. Proteomic data After transcription, genes are translated into proteins. Some of these proteins remain inside the cell, while others circulate through the body. Inexpensive CHIP technology allows the levels of different proteins to be estimated with reasonable accuracy. DNA methylation This is easier to measure than (2) and (3). Methylation is one of the many mechanisms for controlling gene expression, providing the most persistent effects. Horvath has found that certain sets of DNA methylation regions are characteristic of a particular age, regardless of which human organ the material for analysis comes from.

Many genes have so-called promoters adjacent to them, which are DNA regions that store temporary information about whether a gene is activated or inactivated. Promoter regions contain a repeating sequence of nucleotide bases C-G-C-C-G-C-G-C-G-C. It is called a CpG island (where ‘p’ means only that C is joined to G on the same chain, not on a complementary chain, where C is always paired to G). ‘C’ stands for “cytosine” and the cytosine molecule can be modified by attaching an additional methyl group (CH3-) to give 5-methyl-cytosine. The cell contains molecular agents whose function is to selectively attach methyl groups to certain sites of DNA and remove them from other sites. The essence of this is that methylated cytosine is a tag signalling ‘do not transcribe the adjacent gene’. Removal of the methyl groups is a signal to restart transcription. Methyltransferase enzymes are localised in specific regions of the genome and function to activate and inactivate genes. Methylation can be transient. There is evidence of the existence of circadian cycles of methylation. It can also be quite long-lasting. Methylation profiles can persist for decades. They can be copied during cell division and passed on to the organism’s descendants as part of their epigenetic inheritance. However, inherited methylation regions are an exception, as most of the genome is re-programmed shortly after birth. During the formation of oocytes and spermatozoa, methylation profiles are formed that ensure their pluripotency.

Using a standard statistical algorithm, Horvath identified the 363 CpG regions that correlated most strongly with chronological age, regardless of which part of the body the cells were isolated from. The same algorithm ensured that the 353 numbers were multiplied by the values of the methylation levels of each region, then added up all the values. The resulting number is not a direct measure of age, in the last step of the evaluation a table (empirically constructed curve) is used to determine the age corresponding to the number. This curve is a rough representation of the function before it is transformed into a measure of age. It should be noted that the methylation profile changes at a high rate during the first five years of life, which gradually decreases during the growth phase and levels off to a steady decline after about 18 years of age. Although the Horvath clock was developed independently of the tissue from which the DNA was isolated, some variation is possible. The strongest variations are characteristic of the breast, which ages faster than the rest of the body, and the brain, which ages much more slowly. Blood and bone tissue are characterised by slightly accelerated ageing; sperm and eggs are at ‘zero age’ regardless of a person’s age. The placentas of women of any age are also zero-aged. Similarly, induced pluripotent stem cells (derived from the 4 Yamanaka factors) are zero age. At the same time, similar exposure can turn differentiated cells of one type into cells of another type, such as skin cells into neurons. This has no effect on epigenetic age. Liver cells tend to be older than the rest of the body in overweight people and younger in underweight people. This pattern does not appear to extend to other tissues. For example, the age by methylation of fat cells in obese people does not exceed the corresponding figure for the rest of the body. And, perhaps unexpectedly, weight loss does not normalise the methylation age of liver cells (at least not over a 9-month follow-up period in one of the studies investigating this issue). A number of studies have found correlations between methylation age and the risk of various diseases and mortality. In such studies, all environmental factors, including smoking, obesity, physical activity, occupational hazards, etc., are adjusted for in the analyses. Together, these exposures are referred to as ‘externalities’. The results showed that exposure to such factors increases methylation age and, independently of this, methylation age correlates with intrinsic (genetic) factors that influence longevity. Horvath estimates that genotype is responsible for 40 per cent of the variability in methylation age, providing a discrepancy with chronological age. The methylation age of males is slightly higher than that of females. This is evident already by the age of two years. Delayed menopause corresponds to a younger methylation age. The level of cognitive function has an inverse correlation with age by brain methylation. Speaking to Horvath at the same conference, Jim Watson stated that there are many supplements and drugs that can slow Horvath’s clock. He devotes his talk to metformin, which Watson says affects epigenetics through a mechanism completely unrelated to lowering blood sugar levels (the reason metformin is prescribed to tens of millions of people with diabetes). There is a very interesting clue: a small number of children never develop and grow and continue to look like babies until they are 20 or possibly more years old. These children have a normal methylation age. Whatever is blocking their growth is not altering the methylation process of their DNA. Does this mean that there are other epigenetic mechanisms, more powerful than methylation, regulating growth and development? Or do children with this syndrome have normal epigenetic development, but something downstream of gene expression is blocking their growth? In contrast, Hutchinson-Gilford progeria is caused by a defect in the LMNA gene that causes premature aging and death that occurs earlier than adulthood. According to Horvath’s clock, children with this disease have a normal methylation age. Radiation, like smoking and oxidation by environmental factors, accelerates aging of the body. This is independent of age at methylation, which is not affected by radiation. Neither smoking nor exposure to radiation affects epigenetic age. HIV also accelerates the aging process and does not affect age at methylation. Methylation age and telomere age are correlated with chronological age, but they predict mortality and morbidity independent of chronological age. At the same time, the two indices do not correlate with each other. In other words, the data provided by the methylation clock and telomere length measurements complement each other, and their combined use provides better performance in predicting age-related decline that will occur in the future than their use separately. Diet has a weak effect on age at methylation. Diets very high in carbohydrate and very low in protein are markedly worse. In addition to this, evidence in favour of the ‘golden mean’ comes from two approaches, namely the protein-depleted Ornish diet and the type of diets that include the Zone diet and the Atkins diet. The evidence is not strong enough to be unequivocally conclusive, but it does point to the possible efficacy of these approaches. It is also known that the epigenetic clock does not work in malignant tumours.

The watch was originally optimised to track chronological age, but it randomly provides much more information. In the second phase of his work, Horvath set out to learn how to accurately track biological age. He used blood samples preserved since the 1990s and matched them to medical histories and death certificates to identify the methylation regions most strongly correlated with age-associated clinical outcomes. As a result, he invented a phenotypic clock called DNAm phenoAge. They use 523 methylation regions to predict:

• mortality from all causes;
• mortality from cardiovascular disease;
• the development of lung disease;
• the development of cancer;
• development of diabetes mellitus;
• decline in physical strength;
• decline in cognitive function.

Epigenetic clock is adapted to work efficiently with the most accessible cells – skin and blood. Sufficient number of epithelial cells for DNAm test can be painlessly scraped from the mucous membrane of the oral cavity (inner surface of the cheeks).

Several research groups have begun experimenting with transfusions of plasma from young donors as a possible method of rejuvenation. Horvath describes promising observations: sometimes older people develop a variant of leukaemia that requires transfusions of blood and bone marrow (containing stem cells that give rise to new blood cells) from a donor. It has been reported that after such therapy, the patient’s blood shows age by methylation of the donor rather than the recipient himself.

In different individuals, the methylation age is higher or lower than the chronological age by an average of 2 years. 40% of this variability is due to heredity. Some common gene variants can make the epigenetic clock go faster or slower. The most significant genetic variants link telomere aging to age by methylation. The faster a person’s epigenetic clock goes, the longer their telomeres. The slower the epigenetic clock, the shorter the telomeres. This is alluded to in the genetic theory of aging known as the antagonistic pleiotropy theory. In 1957, George Williams suggested that aging-inducing genes must have positive and detrimental effects at the same time. This would serve to explain why natural selection allowed aging to emerge despite the fact that it reduces fitness. According to Williams, ‘nature had no choice but to accept the genes that cause aging, since there was no other way to obtain the positive effects of those same genes’ (which, according to his assumption, favours fertility). In the author’s understanding, the theory of antagonistic pleiotropy does not describe a situation of ‘forced selection.’ Rather, aging is an important factor in maintaining the health of society and Mother Nature is faced with the dilemma of how to maintain aging despite natural selection directed against it at the individual level. Aging is so important to society that the goal of evolution has been to preserve it despite the short-term temptation for natural selection to favour the survival of those who live longer and have a better chance of leaving offspring. According to the author’s hypothesis, evolution gave rise to pleiotropy to solve this problem. The telomerase/epigenetic clock is an example of this. There is no physical necessity for there to be a relationship between telomeric and epigenetic aging, but evolution has developed a seesaw relationship between the two, making it difficult to eliminate the aging process through the accumulation of certain mutations. This also applies to the latest refutation of the existence of a relationship between telomerase and cancer. At first glance, it was questionable whether genetic variants that increase telomere length could be associated with an increased risk of certain types of cancer. We have a clue: genetic variants that increase telomere length simultaneously accelerate the epigenetic aging programme. The particular cancer most closely associated with increased levels of telomerase is melanoma, a malignancy least sensitive to age than other types of cancer. Melanoma tends to develop at an earlier age than other skin tumours. Therefore, it is conceivable that other pleiotropic relationships with mechanisms specifically associated with melanoma development will also be found for genetic variants that provide longer telomeres.

Evidence to date suggests that programmed methylation is an important, but not the only, trigger of the aging process. Smoking affects life expectancy but does not alter age by methylation. Reduction in body weight has a positive effect on life expectancy but no effect on age at methylation. Most interestingly, there are children who do not develop or age prematurely due to genetic defects, yet have a normally progressive methylation age. Why does radiation age the body without shifting the methylation clock? Perhaps the accumulation of damage is part of the aging phenotype, although one would like to believe that the body would retain the ability to repair this damage even late in life if reprogrammed to do so. Why does AIDS speed up the aging clock? Perhaps the immune system is the central signalling mechanism of the ageing process. Thus, aging is ‘methylation plus’. Plus what? Not just ‘methylation plus damage’: although we can shorten our lives with radiation or smoking, we cannot lengthen them by avoiding toxins. ‘Methylation plus other epigenetic programmes’ is the first thing that comes to mind. ‘Methylation plus mitochondrial status’ is the second thing that comes to mind. Methylation occurs within the nucleus, and the cytoplasm of the cell appears to retain the information independently and is even capable of reprogramming the state of the nucleus, as evidenced by the results of parabiosis experiments. There is also evidence in favour of the existence of a combination of ‘methylation plus telomere shortening’. Author: Evgenia Ryabtseva, Eternal Youth portal.