Editorial NexoScience
A blood sample from a 101-year-old is, in a chemical sense, an old object. Its DNA carries tiny tags that pile up and shift over a lifetime, so steadily that a computer can read them and guess the donor’s age to within a few years. So what happens to that record if you force the blood to become a different kind of cell altogether?
It gets wiped, mostly, but slowly. Scientists at the University of Bonn in Germany took immature blood cells from 15 donors, from newborns to a 101-year-old, and turned them straight into neural stem cells, the kind that go on to make brain cells. The cells’ molecular age fell hard. Early on, the new stem cells read at about 13% of their donor’s real age, and after longer growth at about 5%. The work, led by Lea Jessica Berg and published in 2026 in the journal Aging Cell, shows a way to reverse aging in human cells in a lab dish. It is not a therapy. What makes it useful is the pace: the reset unfolds over weeks, slow enough to watch.
The method skips a step everyone else takes. The classic way to rewind a cell is to reprogram it into an induced pluripotent stem cell, a cell that can become almost any tissue. That route erases the age marks completely and fast. Turning skin cells directly into neurons, by contrast, keeps the donor’s age, earlier work found. The Bonn team sat in between. They grew red blood cell precursors from an ordinary blood draw, then used a harmless virus to switch on just two genes, SOX2 and cMYC. Within about 12 days, cells carrying neural stem cell markers were already showing up, without ever passing through the pluripotent stage.
How an epigenetic clock reads a cell’s age
The clock is written in chemical tags. An epigenetic clock looks at DNA methylation, small chemical tags stuck onto DNA that don’t change the genetic code but do change which genes get used. Some of those tags shift in a predictable way as we age. In 2013 and again in 2018, the geneticist Steve Horvath built formulas that turn the pattern into an age estimate, and the Bonn team used the 2018 version. Think of it as reading someone’s age from the wear on their shoes rather than from their birth certificate. It works well. It’s still wear, not the date.
The check passed in blood. In the starting blood cells, the clock reading tracked each donor’s real age closely, a sign the tool was doing its job. The team also made a second set of stem cells from the same samples the long way, through the pluripotent stage. Those all read as embryonic, whatever the donor’s age. The directly converted cells came close behind. At low passage (a passage is each time cells are split into fresh dishes), they kept a little over one-eighth of the donor’s age, and later only about a twentieth.
Why the epigenetic clock reset took weeks
The reset crawled. The researchers followed three cell lines from donors aged 50 to 101 over time. It took about 50 days of conversion before most of the age marks were gone. Single lines got down to about 20% of the donor’s age only after around seven weeks. Classic reprogramming reaches a full reset in roughly 20 days, according to earlier studies. The cells here looked like neural stem cells long before their clock caught up. “Unexpectedly,” the authors wrote, the de-aging process is “largely protracted.”
Cell division wasn’t the engine. The team had a suspect. Reprogrammed stem cells and these new neural stem cells both divide a lot, while directly made neurons don’t, and only the dividing ones lose their age. So the researchers slowed division on purpose, in cells from donors aged 81 to 86, using two chemicals, glycerol and thymidine. The cells divided far less, yet their clocks fell about as much as untreated cells. Then they pushed most of the cells to become neurons, which stop dividing. After 20 days as neurons, the age reading still didn’t differ significantly from dividing cells. You know what? That rules out the simplest explanation, that the marks just get diluted each time DNA is copied.
Old and young donors stopped looking different. The team also tracked gene activity. In the starting blood, 1,871 genes differed between the youngest and the oldest donors. By day 14 of conversion that gap had shrunk to 20 genes, and to 12 in the established stem cells. The young group was only three newborns, though, and the authors warn that this limits the comparison.
Other aging checks came back clean. Clock readings are one thing, so the researchers also looked for classic signs of old cells in finished lines from donors aged 0, 87 and 101. They checked the scaffold inside the cell nucleus, the cell’s recycling system (autophagy), DNA damage, waste from the mitochondria and genes tied to senescence, the state in which worn-out cells stop dividing. On all of these, the cells from very old donors looked like the fully rewound ones. Their identity as brain stem cells took weeks to settle as well. Genes for growing nerve fibers and building connections were still climbing late in the process.
What these young-again cells can’t tell us yet
This is a dish, not a treatment. Nobody got younger here, and no cells went into a person. DNA methylation age is a proxy for biological age, and the authors write that it “does certainly not capture the full picture.” There’s also a measurement problem they flag themselves. Truly aged neural stem cells don’t exist to serve as a reference, so the change of cell type could itself shift the clock reading. The Horvath clock is known to read lower in nerve tissue than in blood, which could inflate how much rewinding seems to happen. Oxygen is another confounder: the cells grew in low oxygen until day 35, with brief swings during handling, then in normal air. The authors can’t rule out that this helped. The groups were small. Two authors co-founded companies, one of which offers epigenetic testing, the paper discloses.
The effect may not belong to one recipe. When the authors ran the same clocks on public data from another lab, which had turned middle-aged skin cells into a different kind of early nerve stem cell with four genes, the resulting cells read as infants or toddlers. SOX2 was the only gene the two recipes shared. That makes it worth a closer look, the authors suggest, though cMYC has been tied to aging too.
Why a slow reset matters for aging research
Slow is the point. “Rejuvenation that extends over such a long period of time is ideal as an experimental model,” Oliver Brüstle, director of the Institute of Reconstructive Neurobiology at the University Hospital of Bonn and a co-author of the study, said in a statement from the University of Bonn. With weeks to work in, researchers can add a drug or block a gene and see whether the clock speeds up or stalls. The authors pitch it as an alternative to “partial” reprogramming, where cells get only a short dose of rewinding factors. For brain research the stakes are plain. Age is the biggest risk factor for diseases like Alzheimer’s, as the team notes, and our roundup of what the evidence says about brain health keeps running into the same fact.
The next step is one cell at a time. Lines from the same donor didn’t all de-age the same way. The authors call for fine-grained time courses that follow single cells, to see whether some clones rewind faster and take over the dish, and whether the cells brush close to the pluripotent state on the way. The clock has been read in the dish. Now it has to be read cell by cell.
Source: Berg, L. J., Franzen, J., Buneß, A., Konang, R., Sheng, C., Peitz, M., et al. “Protracted Fate Acquisition and Epigenetic De-Aging During Induced Neural Stem Cell Conversion of Human Blood Cells.” Aging Cell (2026). DOI: 10.1111/acel.70751
This article summarizes published scientific research for general information. It is not medical advice and does not replace consultation with a qualified health professional.
