Aging reversed: old eyes saw again

Researchers at Harvard Medical School switched on three of the four Yamanaka genes inside the eyes of aged mice; optic nerves regrew and lost vision returned.

Four genes that turn a cell back into an embryo

In 2006, Kazutoshi Takahashi and Shinya Yamanaka at Kyoto University reported in Cell that four genes were enough to turn an adult mouse skin cell back into something resembling an embryonic stem cell. The genes were Oct3/4, Sox2, c-Myc and Klf4. Fibroblasts carrying all four, grown under embryonic stem cell conditions, acquired the shape, growth behaviour and marker genes of embryonic cells. Injected under the skin of immunodeficient mice, they formed tumours containing tissues from all three germ layers. Injected into early embryos, they contributed to the developing mouse. Takahashi and Yamanaka named them induced pluripotent stem cells. The four genes became known as the Yamanaka factors.

Full reprogramming erases a cell's identity. A retinal neuron run through the complete process stops being a retinal neuron. It also resets the cell's epigenetic age, the pattern of methyl marks on DNA that drifts with time. In 2013 Steve Horvath showed that 353 such sites, read across 8,000 samples from 51 healthy human tissues and cell types, predicted chronological age in nearly all of them, and that the clock reads close to zero in embryonic and induced pluripotent stem cells. Full reprogramming turned the clock back to zero and destroyed the cell's identity in the same stroke, which raised a question the field spent the next decade on: whether the age reset could be separated from the identity reset.

Partial reprogramming in a living animal

The first in vivo answer came from the Salk Institute in 2016. Alejandro Ocampo, Juan Carlos Izpisua Belmonte and colleagues switched all four factors on in short, repeated cycles in mice carrying a mutation that causes premature ageing. The cyclic exposure improved cellular and physiological hallmarks of ageing and prolonged the animals' lifespan. In older wild-type mice, the same short-term expression improved recovery from metabolic disease and from muscle injury. The paper framed epigenetic dysregulation as a driver of mammalian ageing rather than a bystander, and it showed that reprogramming could be dosed.

The approach carried a known hazard. Among the four factors, c-Myc is an oncogene, and continuous expression of all four in mice had produced tumours in earlier work. Any attempt to apply the method to a tissue for weeks or months would have to deal with that.

Three of the four, in the eye

In December 2020, Yuancheng Lu, David Sinclair and colleagues at Harvard Medical School published the paper this story rests on, in Nature. They dropped c-Myc entirely, noting that it reduces mouse lifespan and is not required to begin reprogramming, and expressed only Oct4, Sox2 and Klf4, abbreviated OSK. In cultured fibroblasts from old mice, five days of OSK expression shifted the mRNA profile toward a youthful pattern without inducing Nanog, a marker of pluripotency, and without apparent loss of cell identity.

The team chose the eye as a test tissue because it is part of the central nervous system and because retinal ganglion cells, the neurons whose axons form the optic nerve, lose the ability to regenerate within days of birth. An adeno-associated virus carrying the three genes as a single polycistron was injected into the vitreous body of the eye. Expression reached roughly 40 percent of retinal ganglion cells and was controlled by doxycycline in the drinking water, so that it could be turned on or off.

Safety was tested first. Mice given the construct systemically at 5 months and at 20 months of age were induced continuously for 10 to 18 months with no increase in tumour incidence and no negative effect on overall health. In the retina, 15 months of continuous OSK expression produced no tumours and no structural change.

Crushed optic nerves regrew

The first functional test was an optic nerve crush, an injury after which adult retinal ganglion cells normally die and their axons do not regrow. Two weeks after viral delivery the nerve was crushed. In 12-month-old mice, OSK doubled the survival of retinal ganglion cells, matching the effect seen in 1- and 3-month-old animals. Axons regenerated through the crush site. When induction was extended to 12 to 16 weeks, regenerating and sprouting fibres extended more than 5 millimetres along the nerve into the optic chiasm, where the nerve meets the brain.

The regeneration depended on active DNA demethylation. When the enzymes TET1 or TET2 were knocked down or deleted in the retinal ganglion cells, the gain in survival and regeneration disappeared. The epigenetic age of the cells, calculated with a DNA methylation clock, was pushed toward a younger value by the treatment. The authors read this as evidence that the injured neurons still held a record of their younger epigenetic state, and that OSK gave them access to it.

Vision lost to glaucoma and to age came back

Two further models carried the result from axons to sight. In the first, intraocular pressure was raised for 21 days by injecting microbeads into the front chamber of the eye, a standard model of glaucoma that kills retinal ganglion cells. Three to four weeks after the microbead injection, after damage had already occurred, the viral construct was delivered and OSK was induced for four weeks. Treated eyes recovered about half of the visual acuity they had lost, measured by the optomotor response, a behavioural test in which a mouse tracks moving stripes. The pattern electroretinogram, an electrical readout specific to retinal ganglion cells, improved in parallel.

In the second model there was no injury at all. Mice aged 3 and 11 months were treated and induced for four weeks. By one year of age, untreated mice had lost a significant fraction of visual acuity and ganglion cell electrical activity. OSK restored both. RNA sequencing of purified ganglion cells from 12-month-old mice identified 464 genes whose expression had changed with age relative to 5-month-old animals; about 90 percent of them returned to youthful levels after treatment, with no increase in the number of ganglion cells or in axon density. The cells had not been replaced. They had been reset. Both TET1 and TET2 were required for the restoration. In 18-month-old mice, the restoration of acuity was not seen.

From a mouse eye to a clinical trial

The 2020 paper is one of the reasons partial reprogramming became the most heavily financed idea in ageing research. On January 19, 2022, Altos Labs launched with 3 billion dollars in committed funding and a mission to restore cell health and resilience, with Yamanaka as senior scientific adviser and Izpisua Belmonte leading one of its institutes. Sinclair is not named in that announcement. His own company, Life Biosciences, took the OSK eye approach forward as ER-100, an intravitreal injection of the same three factors. On January 28, 2026, the company announced that the U.S. Food and Drug Administration had cleared its investigational new drug application for ER-100 in open-angle glaucoma and non-arteritic anterior ischaemic optic neuropathy. On June 9, 2026, it announced the first patient dosed in a Phase 1 trial.

What the story does and does not show

The 2020 paper showed, in mice, that three of the four Yamanaka genes could make injured optic nerve axons regrow, double the survival of the neurons that carry them, restore about half of the visual acuity lost in a glaucoma model, and reverse the age-related loss of vision in 12-month-old animals. The gene expression and DNA methylation data support the interpretation that the cells were returned to a younger epigenetic state rather than replaced.

The experiments were performed in one tissue of one species, with group sizes typically of 4 to 10 eyes or nerves per condition. The restoration of acuity was not observed in 18-month-old mice, so the result does not show that the method works at every age. It does not show that any mouse lived longer, and it does not show that vision in any human has been improved; the first human trial began dosing in mid 2026 and has reported no efficacy results. The safety data cover up to 18 months of expression in mice, with no tumours, and no longer. The story is a demonstration that youthful epigenetic information survives in old cells and can be read back out, measured in a mouse eye.

The clock was not stopped. It was turned back.