Revealing the Longevity Code: Research on Epigenetic Characteristics of Over Centenarians, Predicting Healthy Longevity

At present, with the continuous extension of human life expectancy, the population with a long lifespan of 100 years is no longer rare. Centenarians (people 100 years old or above) and super centenarians (people 110 years old or above) are excellent examples of studying healthy and long life. There are constantly researchers conducting scientific explorations related to longevity from a scientific perspective, hoping to find the secret to longevity.

THE LANCET Health Longevity published a heavyweight study showing that based on non gender specific epigenetic clock, the epigenetic age of centenarians and over centenarians accelerates to negative. Whole epigenome association analysis further indicates that 163 centenarians and supercentennials participate in anti-inflammatory reactions, which contributes to healthy TGF- β signal related genes represent advanced (or older) epigenetic states at the CpG locus.

In this cross-sectional study, researchers recruited 421 healthy non centenarians, as well as 94 centenarians and super centenarians, evaluated the participants' whole blood DNA methylation profiles, calculated the individual's epigenetic age, and evaluated their age acceleration.

The results showed:

1. The epigenetic age of longevity elderly people is younger than their actual age

DNA methylation profiles of 757,461 CpG sites were obtained from non centenarians. Gender specific principal component analysis did not find any outliers, so researchers developed an epigenetic clock using data from non centenarians. 144 blood samples were collected from 94 centenarians and super centenarians, and DNA methylation profiles of 779,753 CpG sites were obtained. The epigenetic age of centenarians and over centenarians was calculated using an epigenetic clock trained on data from 336 participants (Figure 2A). Except for a centenarian, all participants had an epigenetic age that was younger than their actual age, with an accelerated negative epigenetic age (Figure 2B).

The degree and direction of longitudinal changes in epigenetic age vary among 26 multi sample individuals, with 21 individuals (81%) having an increase in epigenetic age with actual age. For these 26 subjects, the average annual change in epigenetic age was 0.703 years (SD 0.839), which showed a significant difference from 0 (t-test p=0.0003). For non centenarians, the average annual variation in epigenetic age is 0.851 years old. The estimated epigenetic aging rate (p=0.38, double sample t-test) based on longitudinal data of centenarians and over centenarians is not significantly slower than that of non centenarians.

2. Significantly increased repeatability of age related CpG sites

Gender specific EWAS identified 3731 female CpG sites and 7525 male CpG sites related to age. 1217 females [33%] showed a positive correlation, 2514 females [67%] showed a negative correlation, 1071 males [14%] showed a positive correlation, and 6454 females [86%] showed a negative correlation (Figure 3A-D). Out of 2109 (408 [19%] positive and 1701 [81%] negative) age related CpG sites in both sexes, except for three (<1%), all CpG sites were replicated through non gender specific EWAS adjusted for gender and cell composition, confirming that these age CpG relationships are not affected by cell composition or gender. For the combination method of single linear regression and analysis of variance, it is known that the reproducibility of age related CpG sites significantly increases. These CpG sites also exhibit constant methylation or demethylation with age, and no significant deviation trend was found in any specific age group.

3. Different DNA methylation states at different sites

Among the 408 CpG sites positively correlated with age among centenarians and supercentennials, 295 (72%) showed an expected increase in DNA methylation sites compared to non centenarians (i.e., based on actual age, there was no significant difference from expected). The DNA methylation profiles of the remaining 113 (28%) CpG sites were lower than expected. On the contrary, out of 1701 CpG sites negatively correlated with age, 1094 (64%) centenarians and over centenarians showed a decrease in DNA methylation with actual age. Therefore, for centenarians and super centenarians, 113 posHypo and 444 negHypo CpG sites have younger DNA methylation states, while 163 negHypo CpG sites have older DNA demethylation states.

In the posHypo, negHyper, and negHypo CpG sites, 6-20% were significantly associated with adjacent single nucleotide polymorphism genotypes (methylation quantitative trait sites CpG sites). Compared to posHypo, age negatively correlated CpG sites (negHyper and negHypo) are more likely to be influenced by genotype. Among these methylation quantitative trait sites, 8 negHyperCpG related single nucleotide polymorphisms have been reported as whole genome association research signals. Enrichment analysis showed that the posHypo CpG site was enriched in the cell adhesion related group. The proportion of oncogenes in adjacent CpG site is too high. On the contrary, the advantage of annotating the posHypo CpG site to oncogenes is relatively low.

In the analysis of the protein-protein interaction network of the negHyper gene, six networks were formed. The largest network consists of 51 proteins, with CD44 antigen having the greatest centrality. This central protein participates in multiple functions, among which cancer-related functions are widely shared among neighboring proteins (Human Protein Atlas Database 20). The central proteins at the other end of the network (such as AUTS2, CNTNAP2, ITPR1, and NRXN1) and their neighboring proteins are associated with many neuropsychiatric disorders, such as attention deficit hyperactivity disorder, ataxia, autism spectrum disorder, bipolar disorder, intellectual disability, and schizophrenia. For the posHypo gene, only one pair of proteins belonging to the pro-calmodulin β cluster was found to interact, which plays a role in neuronal cell adhesion.

4. TGF-β signal pathway triggers anti-inflammatory response

Protein-protein interaction analysis of demethylation enhancing genes in centenarians and over centenarians shows that all three proteins are associated with TGF-β signal pathway related: SMAD7 inhibits TGF-β signal pathway; ACVR1 activates SMAD7; TAB2 activates TAK1, which is in TGF-β release from receptor complex under stimulation and transmit downstream TGF-β signal. SMAD7 vs. TGF-β anti inflammatory activity is crucial, and by interacting with another protein, SMAD7 can trigger inhibition of pro-inflammatory gene expression. The screening of the iMETHYL expression quantitative trait methylation database showed a significant negative correlation between the two negHypo CpG sites labeled with SMAD7 and the expression of SMAD7 in CD4+T cells. Therefore, in centenarians and over centenarians, decreased DNA methylation at these CpG sites is associated with upregulation of SMAD7.

The results of this study indicate that the epigenetic ages of centenarians and super centenarians are significantly lower than their actual ages. There is a connection between specific epigenetic states and health and longevity. Some epigenetic characteristics of centenarians and super centenarians remain in a young state, while others remain in an elderly state. This discovery may provide an epigenetic basis for healthy longevity, which is related to avoiding or delaying age-related disease outbreaks and cognitive decline. On the contrary, accelerated (late) demethylation of CpG sites was also detected in centenarians and over centenarians. Knowledge based analysis suggests that some demethylated CpG sites can affect the anti-inflammatory cytokine TGF-β activity. Anti-inflammatory responses such as those mediated by TGF-β and other cytokines are critical for healthy aging and longevity.