Abstract
Aging is a complex, multifaceted process characterised by biological decline and ultimately, death. Biological sex has a clear influence on the aging process, as evidenced by the lifespan advantage possessed by females in a wide range of mammalian species despite the physiological burden of pregnancy and lactation. Male hormones are a likely driver of this effect, considering castration – even in humans – generally provides longevity extension. Progressing age is associated with various cellular changes, including alterations in DNA methylation – a stable yet adaptable DNA modification that affects gene expression and preserves cellular identity. These modifications accumulate at specific sites within the genome in such a predictable manner that they form the basis of epigenetic clocks – molecular predictors capable of estimating age using only DNA methylation data.
In this thesis, I developed an epigenetic clock for sheep, the first of its kind for livestock species. This epigenetic predictor used DNA methylation data from 185 CpG sites and could estimate chronological age with a median absolute error of only 5.1 months. Because my dataset included castrated sheep, I had the unique opportunity to investigate the effect of castration on the intrinsic aging rate. Consistent with previous reports of lifespan extension following castration, adult castrated sheep exhibited a delayed epigenetic aging rate compared to intact males. Significantly, I identified a subset of CpG dinucleotides that become hypomethylated with advancing age specifically in intact males, while these same sites are “feminised” in castrated males – which I named androgen-sensitive differentially methylated probes (asDMPs). I then uncovered a plausible mechanism for the existence of such sites by observing that many of these sites are bound by the androgen receptor (AR).
Next, I utilised these asDMPs to build a novel epigenetic predictor called the androgen clock, capable of estimating cumulative androgen exposure, i.e., the time since the onset of puberty in male sheep. To meet the aims of this tool being cost-effective and accessible, I used bisulfite amplicon-based techniques to capture methylation data at just one site, cg21524116, mapping to MKLN1 in sheep. Remarkably, this predictor estimated androgen exposure with a median absolute error of 4.3 months in sheep, an impressive achievement for a single CpG site. I also illustrated the broad applicability of the androgen clock by testing it across methylation profiling platforms and migrating it into a new species: mice.
I used a chronic 5α-dihydrotestosterone (DHT) treatment mouse model to confirm that the age-associated demethylation seen in intact males was indeed due to the presence of androgens rather than another testicular factor. Significantly, the administration of DHT for only 13 weeks resulted in an acceleration of the androgen clock by an average of 1.71 years compared to saline-treated controls, demonstrating the ability of the androgen clock to detect dosage effects.
In summary, this thesis successfully utilised and adapted existing epigenetic clock technology to construct a novel predictor and identify androgen-dependent erosion of the epigenetic landscape. The androgen clock has potential applications in healthcare to enhance the diagnosis and monitoring of androgen-related conditions, in agricultural settings for the detection of hormonally altered meats, as well as in sports for testing for the abuse of performance-enhancing androgenic steroids. The androgen clock significantly broadens the capability of epigenetic clocks are encourages re-evaluation of the biological information epigenetics can hold.