Aging and Rejuvenation Converge on the Same Epigenetic Switch in Mouse Skin

In a finding that could reshape how scientists think about growing old, researchers at Johns Hopkins University and the Salk Institute have mapped, at single-base resolution, the epigenetic changes that occur in mouse skin during aging and during a remarkable process that partially reverses them. The study, published in Molecular Systems Biology, reveals that aging and rejuvenation converge on the same set of genomic targets: genes controlled by the Polycomb repressive complex 2, or PRC2, a master regulator of cellular identity. The work provides the most comprehensive picture yet of what happens to DNA methylation when an old tissue is coaxed back toward a younger state, and it points to a specific chromatin machinery that may sit at the heart of both aging and its reversal.

The experimental system builds on a technique known as partial reprogramming. Cells can be reset to a more youthful, stem-like state by transiently expressing four transcription factors, Oct4, Sox2, Klf4, and c-Myc, collectively called OSKM. Full reprogramming wipes out cellular identity entirely, but short, repeated pulses of OSKM avoid this hazard while restoring youthful characteristics. In this study, the team used mice engineered to carry a single copy of an OSKM cassette that could be switched on with a drug. Starting at fifteen months of age, roughly the mouse equivalent of late middle age, the animals received long-term cyclic OSKM treatment until they reached twenty-two months. Previous work had shown that the skin of these treated animals looked and behaved more like young skin, with thicker epidermis, greater proliferative capacity, reduced fibrosis after injury, and dampened inflammatory gene expression.

What remained unknown was the epigenetic mechanism behind this rejuvenation. Earlier assessments relied on DNA methylation arrays that interrogate fewer than seven hundred CpG sites, most of them chosen for clock-like behavior rather than functional importance. The new study went far beyond that, deploying whole-genome bisulfite sequencing to profile roughly fifteen million CpG dinucleotides across young skin, old untreated skin, and old skin treated with OSKM. That is more than four orders of magnitude greater coverage than array-based epigenetic clocks provide, allowing the researchers to examine the entire methylome rather than a curated snapshot.

To interpret this torrent of data, the team used an information-theoretic framework that treats the epigenome as a potential energy landscape, in the spirit of statistical physics. Rather than simply asking whether average methylation differs between groups, the approach, implemented in a pipeline called informME, computes two quantities in 150-base-pair windows across the genome: the mean methylation level, and the normalized methylation entropy, a Shannon-entropy measure of how disordered or stochastic methylation states are within each window. It also calculates the Jensen-Shannon distance, an information-theoretic measure of how discordant the full probability distributions of methylation states are between two conditions. This matters because two tissues can have identical average methylation yet radically different underlying variability, and that variability may carry crucial biological information about cellular identity and its erosion.

The results were striking. Old untreated skin showed a genome-wide shift toward hypomethylation and a marked increase in methylation entropy, meaning the epigenome had become more random and disordered. Old skin treated with OSKM moved in the opposite direction: global methylation levels rose and entropy fell, resembling the profiles of young animals. Principal component analysis cleanly separated the three groups, with the first axis capturing age-related differences and the second capturing the effects of the reprogramming treatment. A cell-type deconvolution analysis argued against the possibility that these shifts merely reflected changes in the proportions of different skin cell types; instead, aging and rejuvenation appeared to shift the methylation state of the tissue itself.

The most consequential discovery emerged when the researchers ranked genes by their epigenetic discordance, the Jensen-Shannon distance across promoter and gene-body regions. Genes disrupted by aging and genes disrupted by aging but then reversed by reprogramming overlapped far more than chance would predict, with an odds ratio of nearly seventeen. In total, 661 genes showed age-related epigenetic changes that were reversed by partial reprogramming. When the team performed over-representation analysis on the most discordant genes, the top five gene sets in both the aging and rejuvenation comparisons were all annotated as targets of PRC2, bound by its core subunits EED and SUZ12 or marked by its catalytic product, the histone modification H3K27me3. In other words, the epigenetic storm of aging, and the calm that follows rejuvenation, both concentrate on the same Polycomb-regulated territory of the genome.

PRC2 is a chromatin complex that deposits H3K27me3, a repressive mark characteristic of facultative heterochromatin that keeps developmental genes poised and cellular identity intact. Within PRC2-bound regions, the researchers found that old skin gained DNA methylation and entropy relative to young skin, and that OSKM treatment specifically reversed these changes. Poised promoters, regions carrying both the repressive H3K27me3 mark and the activating H3K4me3 mark, showed the greatest discordance of any chromatin state, gaining methylation and entropy with age in a reversible fashion. This is notable because the rest of the genome tends toward hypomethylation with age, making poised promoters the exception that highlights how vulnerable these plastic, identity-defining regions are to epigenetic drift.

To connect DNA methylation to chromatin structure directly, the team developed improved protocols for isolating epidermal nuclei and performing native chromatin immunoprecipitation, a technique with a better signal-to-noise ratio for heterochromatin than conventional cross-linked methods. Using Drosophila chromatin as a spike-in standard for quantitative comparison, they found widespread loss of H3K27me3 in aged epidermis, with nearly forty-five percent of peaks significantly reduced at a false discovery rate below 0.05, many of them at gene promoters. Genes losing the mark were enriched for neuronal developmental pathways, hinting that aged skin cells drift away from their proper identity toward a neuroectodermal precursor-like state. The overlap between genes with altered H3K27me3 and genes with altered DNA methylation was statistically robust, and intersecting all three criteria, methylation discordance during aging, discordance reversed by reprogramming, and H3K27me3 changes, yielded a core set of one hundred PRC2-regulated genes.

Among these core genes were several with compelling connections to tissue maintenance. Sox11, involved in epidermal development, wound repair, and skin fibrosis, lost promoter H3K27me3 with age while its methylation entropy rose, both reversed by OSKM. Lin28b, a recognized regulator of physiological aging in heart, blood, and testis that promotes tissue repair when overexpressed, showed the same convergent pattern. The analysis also flagged genes such as Rara, a retinoic acid receptor known to counter skin aging; Erbb4, linked to keratinocyte proliferation; Fgfr1, a regulator of wound healing; and Cbx8, a Polycomb component implicated in senescence. Separately, the team identified 4,707 large blocks of age-related DNA hypomethylation in epidermis, and found that over eighty-seven percent of them fell within H3K9me2-marked LOCK domains, vast repressive chromatin regions averaging 1.3 megabases that anchor genome architecture. The LOCKs themselves barely changed with age, but they appear to define the boundaries within which methylation is lost, and both H3K27me3 sites and LOCKs showed increased methylation entropy during aging despite being mutually exclusive territories.

The implications extend beyond skin. The findings give concrete molecular content to the idea that aging represents a loss of epigenetic information: the regions most sensitive to aging are precisely the bivalent, plastic branch points that define what a cell is, and their methylation becomes noisier as the energy landscape of the epigenome flattens. The parallels with cancer are hard to ignore, since tumor epigenomes likewise combine large hypomethylated blocks with focal hypermethylation at Polycomb targets. Most provocatively, the study suggests that PRC2 is not merely a bystander but a potential lever: if restoring youthful methylation patterns at Polycomb targets is how partial reprogramming rejuvenates tissue, then modulating PRC2 activity directly might one day offer a route to reversing age-related decline without the risks of expressing reprogramming factors. The hundred-gene core set identified here now provides a concrete molecular readout for detecting, and perhaps one day correcting, the epigenetic noise of aging.

Subject of Research: Epigenetic convergence of aging and partial reprogramming on PRC2 target genes in mouse skin

Article Title: Convergence of aging- and rejuvenation-related epigenetic alterations on PRC2 targets

Article References: Convergence of aging- and rejuvenation-related epigenetic alterations on PRC2 targets. (n.d.). https://doi.org/10.1038/s44320-026-00195-9

Image Credits: AI Generated

DOI: 10.1038/s44320-026-00195-9

Keywords: aging, rejuvenation, partial reprogramming, PRC2, DNA methylation, H3K27me3, epigenetic entropy, OSKM, skin, LOCKs, chromatin, epigenetics