A new model links brain aging to disrupted daily DNA repair scheduling in neurons and proposes a molecular switch that drives the cycle
A new model links brain aging to disrupted daily DNA repair scheduling in neurons and proposes a molecular switch that drives the cycle
On September 17, GeroScience published a theoretical model connecting two well-known but previously separate hallmarks of brain aging: the accumulation of DNA damage in neurons and the weakening of circadian rhythms. The authors propose the enzyme tyrosyl-tRNA synthetase (TyrRS) as the switch governing both processes and explain why both break down together during aging and in Alzheimer's disease. The paper was written by a team that is developing a drug based on this same model.
TyrRS has two jobs. Normally it handles protein synthesis: it loads the amino acid tyrosine onto the appropriate transfer RNA. When free tyrosine runs low (at night, during fasting), a fraction of TyrRS translocates to the nucleus and engages three genome repair processes. It activates PARP1, the enzyme that detects DNA strand breaks. It keeps transposable-element loci silenced, preventing them from triggering inflammation. And it induces the gene LIN9, which is part of the DREAM complex. During the day, DREAM represses 67 DNA repair genes, including BRCA1; at night it releases them for repair.
The daily amplitude of these three processes matters more than their average level. With age, two factors squeeze this system: circulating tyrosine rises (by 15–25% according to cohort data), narrowing the nighttime window in which TyrRS can act, while circadian rhythms weaken as neurons in the suprachiasmatic nucleus, the brain's master clock, die off. Together, these forces freeze the cascade in an intermediate state: still active, but no longer swinging between its peak and its trough.
This reframes older data. In Alzheimer's brain tissue, average DREAM activity is elevated, and previous work interpreted this as deeper suppression of DNA repair. The authors propose a different reading: the daily oscillation has flattened, and any tissue measurement captures a frozen mean rather than the oscillation itself. If the model is correct, sustained-release formulations are more likely to cause harm than benefit, because they flatten the very oscillation that needs restoring. What is needed instead is a short-acting dose timed to the sleep phase. That phase can already be measured: the HairTime test calculates a person's internal time from the activity of 17 clock genes in the root of a single plucked hair. The authors align their model with neuroscientist Maiken Nedergaard's work on sleep rhythms that drive the brain's nightly clearance of amyloid and tau, and they propose repairing both processes together.
The same logic addresses a longstanding paradox in Alzheimer's treatment: lecanemab and donanemab clear amyloid plaques yet deliver only modest cognitive benefit. One earlier attempt to work around this mismatch used CAR-T immune cells reprogrammed to attack plaques directly, instead of repeated antibody infusions. The TyrRS model offers a third path: if the DNA repair clock is broken, amyloid clearance alone is insufficient regardless of the method.
The model can be tested without human subjects: in exceptionally long-lived species such as the naked mole-rat and Myotis bats (mouse-eared bats), the daily amplitude of DNA repair gene oscillation should be the greatest, while the baseline average should be the lowest across species.
The paper contains no original experiments. It synthesizes three findings published separately by three groups. In 2015, co-author Sajish Mathew initiated this line of work by showing that TyrRS activates PARP1; the convergence of all three findings on a single enzyme is what the paper cites as the model's principal empirical support. William Smith founded Functional Longevity Labs; Mathew, Seiden, and Ingram serve as advisors, though only Mathew and Seiden (in addition to Smith) hold equity in the company. The company has already disclosed a preclinical candidate, FLL-001, at a conference.