How Lab-Grown Stem Cells Could Change the Way We Fight Diabetic Blindness

Every year, millions of people with diabetes watch their eyesight slowly fade as tiny blood vessels inside their eyes break down. For decades, the cells that line those vessels have been nearly impossible to study, being scarce, difficult to obtain, and quick to lose their defining traits once removed from the eye. A team at Duke University has now grown those cells in the lab from human stem cells, then used them to help revascularize damaged retinas in mice.

In the study, published in Nature Biomedical Engineering, these lab-grown cells, called iRECs, do more than resemble the real thing under a microscope. Injected into mice with damaged retinas, they wove themselves into the eye’s existing vessel network and helped restore blood flow to starved tissue. Building on that result, the team used the same cells to construct a miniature, working model of the eye’s inner blood barrier on a chip, a small device that could help researchers study diabetic retinopathy and develop new treatments.

How Scientists Turned Stem Cells into Retinal Blood Vessel Cells

Deep in the eye sits its own version of the blood-brain barrier, a tightly controlled wall of cells that keeps the retina protected and stable. This barrier breaks down in diabetic retinopathy, a leading cause of blindness in adults. Studying exactly how and why that breakdown happens has been maddeningly difficult, because the cells involved are so hard to get and so quick to deteriorate in the lab.

To solve that problem, the Duke team essentially replicated the chemical signals the body uses during fetal development to build the retina’s blood supply. Retinal blood vessels begin forming around the fourth month of pregnancy and continue maturing for months after birth, guided by a precise sequence of biological instructions.

Central to their approach was a protein called Norrin and its partner, a surface receptor called Frizzled4 that sits on retinal blood vessel cells. When this connection is disrupted during development, the retina’s vessels fail to form properly. Using that same biological switch, the researchers steered stem cells toward a retinal identity, adding two other compounds to help the cells develop the tight, selective barrier properties that make retinal vessel cells unique.

Resulting iRECs carried the genetic and protein markers of real retinal vessel cells, including a junction protein found in eye vessel cells but absent from similar cells in the brain. They were selective about what they let pass through, just as real retinal vessel cells must be. They could shuttle glucose through a channel critical for fueling the retina’s intense energy demands, and they showed activity from a pump that helps clear harmful substances from the eye. In lab tests, iRECs took up glucose at five times the baseline rate of a non-retina-specific comparison cell type, a sign the team read as confirmation that the cells had taken on a genuine retinal identity.

Injecting Lab-Grown Retinal Cells into Damaged Eyes

To test whether iRECs could actually help heal a damaged eye, the researchers used a mouse model designed to mimic advanced diabetic eye disease. Newborn mice are exposed to high oxygen levels, then returned to normal air; the resulting stress causes their retinal blood vessels to deteriorate and then regrow in an abnormal, chaotic pattern.

Researchers injected iRECs directly into the eyes of these mice and examined the retinas five days later, at the point when damage peaks. Compared with eyes injected with a saltwater solution, iREC-treated eyes showed significantly less vessel loss and far less abnormal regrowth. Cross-sectional imaging confirmed that the injected cells had not simply clustered near the injection site; they had physically integrated into the existing vessel network and joined working, blood-carrying structures within the host retina.

A tracer dye injected into the bloodstream then showed the restored vessels were doing their job: iREC-treated retinas leaked far less than control eyes, a sign the rebuilt barrier was holding. Vessel networks in treated eyes also more closely matched the size and structure of healthy retinal capillaries, evidence that the cells were helping normalize the architecture of the blood supply rather than merely filling gaps.

Modeling Diabetic Eye Disease on a Chip

Beyond their potential as a therapy, iRECs turned out to be a powerful tool for studying the disease itself. Exposed to conditions that mimic diabetic retinopathy, high glucose paired with low oxygen, the cells behaved in ways that matched the clinical reality of the disease. Proteins that normally hold vessel cells tightly together began to scatter away from cell borders, the barrier weakened measurably over several days, and three-dimensional vessel networks grown from iRECs became shorter, narrower, and more fragmented. Granular deposits also appeared within those networks, a form of surface damage to vessel walls documented in patients with diabetes.

Crucially, iRECs reacted far more strongly to these diabetic conditions than non-retina-specific vessel cells derived from the same stem cells. Their takeaway is straightforward: disease modeling is only as accurate as the cells being used, and the wrong cell type can mask the very responses researchers are trying to study.

In parallel, the team built a miniature model of the eye’s inner blood barrier inside a small chip threaded with tiny channels of gel, where iRECs self-assembled into a network of blood-carrying microvessels. They also derived retinal-specific support cells, called pericytes, from stem cells using the same Norrin signaling approach. Pericytes wrap around vessel walls and help maintain barrier integrity, and when these lab-grown pericytes were added to the chip alongside iRECs, the barrier tightened further, reaching levels consistent with those measured in living tissue.

What a Renewable Supply of Retinal Cells Could Mean for Patients

Perhaps the deepest problem this work addresses is as much logistical as scientific. Human retinal vessel cells are scarce, difficult to obtain ethically, and quick to lose their specialized character in culture. That scarcity has long bottlenecked research into diseases affecting millions of people. Derived from stem cells, iRECs offer a renewable and reproducible source of retinal-specific cells that hold onto their identity across repeated use.

According to the authors, patient-specific iRECs could one day be grown from a person’s own cells, opening a path toward disease models that reflect an individual’s genetic background and, potentially, toward cell therapies tailored to individual patients. For the millions of people living with diabetes who face progressive vision loss, that pairing of sharper research tools and a possible route to repair is a concrete reason for optimism.

---