An enzyme called Setd8 keeps retinal stem cells flexible, and losing it collapses eye development

Stem Cell Reports, February 2026. DOI: 10.1016/j.stemcr.2025.102789

The mammalian retina has no second chances. Once development is complete, the stem-like cells that built every photoreceptor and supporting neuron transform into a different cell type entirely. They become Muller glia, structural support cells with no capacity to regenerate lost neurons. If disease or injury destroys retinal cells after that point, the damage is permanent.

Understanding what keeps those progenitor cells in their flexible, neuron-producing state, and what eventually ends that flexibility, is central to any future attempt at retinal regeneration. A study from the Nara Institute of Science and Technology (NAIST) in Japan, published in Stem Cell Reports, has identified one of the key molecular players: an enzyme called Setd8.

Chromatin as a control switch

Retinal progenitor cells (RPCs) carry the same DNA as every other cell in the body. What distinguishes them is which genes are accessible for reading and which are locked away. This accessibility is controlled by chromatin, the complex of DNA and structural proteins that packages genetic material. When chromatin is "open," genes can be expressed. When it is "closed," they are silenced.

The research team, led by Associate Professor Taito Matsuda, suspected that specific epigenetic modifications, chemical changes to chromatin proteins that do not alter the DNA sequence itself, were responsible for maintaining the open chromatin state that defines RPCs. To test this, they isolated mouse RPCs at different developmental stages and analyzed both gene expression and chromatin accessibility using genome-wide sequencing.

The analyses converged on Setd8, an enzyme that adds a methyl group to histone H4 at a specific position (lysine 20). This modification, known as H4K20me1, is associated with maintaining chromatin in an accessible state. Setd8 was highly active in RPCs during the period when they were producing neurons and declined as the cells transitioned toward becoming Muller glia.

What happens without the gatekeeper

To confirm Setd8's role, the researchers engineered mice whose RPCs lacked the enzyme. The consequences were severe. Without Setd8, RPCs showed reduced proliferation, meaning fewer cell divisions and therefore fewer daughter cells available to become neurons. DNA damage increased. Cell death rose.

The physical result was a thinner retina with fewer of the later-developing neuron types. These late-born neurons depend on progenitor cells remaining active and flexible deep into development. Without Setd8, the progenitor cells lost their identity prematurely, shutting down the developmental program before it was complete.

Further molecular analysis revealed the mechanism. Loss of Setd8 caused widespread closing of chromatin regions that are normally open in RPCs. Genes involved in maintaining progenitor identity and DNA repair were downregulated. In effect, without the enzyme to keep the genome accessible, the cells could no longer read the instructions they needed to remain progenitors.

Regenerative implications, eventually

Matsuda framed the findings in terms of regenerative medicine and ophthalmology. If researchers can understand and manipulate the epigenetic mechanisms that maintain progenitor identity, they might eventually be able to reactivate neuron production in damaged retinas. Muller glia, which RPCs naturally become, are an obvious target: if these cells could be coaxed back into a progenitor-like state, they might be able to replace lost photoreceptors.

This is a long-term aspiration, not an imminent therapy. The study identifies one piece of the epigenetic machinery, but retinal progenitor identity is maintained by a network of interacting factors, not a single enzyme. Manipulating Setd8 alone is unlikely to be sufficient for regeneration. And the work was done in mice; human retinal biology shares many features but also has important differences.

Retinal diseases, including age-related macular degeneration and retinitis pigmentosa, are among the leading causes of blindness worldwide, and their prevalence is growing as populations age. Any approach that could restore even partial neuronal regeneration in the retina would represent a fundamental shift in treatment.

But that shift requires understanding far more about how progenitor cells maintain and lose their identity. The identification of Setd8 as a key gatekeeper is one step in that direction.