The 'information theory of aging' is a bold idea resting almost entirely on mice

The "information theory of aging" proposes that we grow old not mainly because our DNA sequence decays but because our cells lose the epigenetic instructions telling them which genes to switch on — and that, in principle, those instructions can be restored [s1]. It is an elegant and much-promoted idea, but the evidence behind it is drawn almost entirely from yeast and mice; no human study has shown that resetting the epigenome slows ageing or extends life [s1][s2].

The theory is worth taking seriously because the underlying observation is real. As cells age, the chemical marks on DNA that control gene activity — chief among them DNA methylation — drift from their youthful pattern. That drift is regular enough to build "epigenetic clocks": a landmark 2013 study developed a DNA-methylation predictor of age from 8,000 samples spanning 51 human tissues and cell types, and found it read close to zero in embryonic and induced pluripotent stem cells, correlated with how many times cells had divided, and ran fast in every one of 20 cancer types examined, by an average of 36 years [s3]. Something about the epigenome tracks biological time. The leap the theory makes is from that correlation to causation — and then to reversibility.

What the mouse experiments actually showed

The strongest experimental case comes from a 2023 study using a system its authors called ICE, for "inducible changes to the epigenome" [s1]. By engineering mice so that harmless DNA breaks could be triggered and then faithfully repaired, the researchers separated the act of repair from any lasting change to the DNA sequence. Repeated repair alone, they reported, pushed the animals toward aging on physiological, cognitive and molecular measures — eroding the epigenetic landscape, causing cells to lose their specialised identity, and advancing the DNA-methylation clock — and some of those changes could be reversed by switching on reprogramming genes [s1]. The authors concluded the results were "consistent with the information theory of aging," which holds that lost epigenetic information is a reversible cause of ageing [s1].

The reversal tool is partial cellular reprogramming. In a 2020 study, delivering three of the four Yamanaka factors — Oct4, Sox2 and Klf4 — to nerve cells in the mouse eye restored youthful DNA-methylation patterns, regrew injured nerve fibres, and reversed vision loss in a mouse model of glaucoma and in old mice; the effect depended on enzymes that strip methylation marks [s2]. Taken together, the two studies are a genuinely impressive demonstration that mammalian tissue keeps a "backup copy" of its youthful settings that can, under laboratory conditions, be accessed.

Why the skepticism is warranted

Impressive in mice is not the same as true in people, and several gaps should temper the enthusiasm this work attracts.

First, it is a hallmark, not the mechanism. The most cited catalogue of ageing biology lists epigenetic alterations as just one of twelve interconnected hallmarks — alongside genomic instability, telomere attrition, mitochondrial dysfunction, cellular senescence and the rest — with no single one crowned as the master cause [s4]. A theory that elevates epigenetic information to the driver of ageing is making a much stronger claim than the consensus framework supports.

Second, correlation is doing heavy lifting. Epigenetic clocks predict chronological age well, but predicting age is not the same as measuring the thing that causes ageing, and the clocks themselves are less reliable than their marketing suggests — the same sample can read years apart on repeat testing. An intervention that moves a clock backward has not thereby been shown to make an organism younger in any way that matters.

Third, the human evidence is essentially absent. Reprogramming carries a real hazard: push it too far and cells dedifferentiate completely, which is how you grow a teratoma rather than a younger tissue. That is why the field works with brief, partial reprogramming — and why translation is slow. The first human test of epigenetic reprogramming is an early-stage eye study designed to check safety in a handful of patients, not to reverse ageing. There is, as yet, no clinical result showing that any of this extends human healthspan.

What to watch

The information theory of aging is a testable hypothesis that has produced striking mouse data and a plausible mechanism for why the epigenome might store recoverable youth [s1][s2]. It is not an established account of human ageing, and it sits within a broader hallmarks framework that treats it as one contributor among many [s4]. The honest position is that the biology is exciting and the clinical claims are premature. The things to watch are whether partial reprogramming can be delivered safely to whole organs, whether it does anything beyond moving a methylation clock, and whether any of it survives contact with a randomised human trial — the bar every longevity idea eventually has to clear, and that this one has not yet approached.