Lab-Grown Retinas Reveal How the Foveola Gets Its Cone Cell Pattern During Fetal Development

The foveola is a tiny pit at the center of the retina, less than half a millimeter across, and it is responsible for approximately half of everything you see. It contains no blue cone cells - only red and green cones packed at the highest density found anywhere in the retina. This distribution is the basis of human sharp, high-resolution color vision. For 30 years, the prevailing explanation was that the blue cones initially present in that central region simply migrate away as the eye develops, leaving the red-green dominated foveola behind.

New data from Johns Hopkins University suggests that migration explanation is likely wrong. Instead, a team led by associate professor of biology Robert J. Johnston Jr. used lab-grown retinal organoids - small tissue clusters grown from fetal cells - to watch foveal development unfold over months and found something different: blue cone cells do not move. They convert.

The study was published in Proceedings of the National Academy of Sciences.

The Problem with Studying Human Foveal Development

The foveola and the broader foveal region pose a longstanding challenge for developmental biology. Mice, zebrafish, and most other model organisms used in eye research do not have a fovea - they have more uniform cone distributions across the retina. This means that the specific developmental mechanisms shaping the primate fovea cannot be studied in standard laboratory animals. Human tissue is difficult to obtain and cannot be manipulated experimentally. The field has been stuck with a 30-year-old model based on limited observational data.

Retinal organoids changed that. By growing retinal tissue from pluripotent stem cells derived from fetal tissue and monitoring the organoids over several months in culture, the Johnston lab can watch human retinal development - including foveal formation - in a system amenable to experimental intervention. They can add or block signaling molecules, follow individual cells over time, and test causal hypotheses in ways that were not possible with cadaveric or biopsy tissue alone.

A Two-Step Conversion Process

The organoid observations revealed a clear temporal pattern in the foveola. Between weeks 10 and 12 of retinal development, a sparse population of blue cone cells is present at the center. By week 14, those blue cones have disappeared - and the foveola's characteristic red-green cone dominance has been established. The cells did not leave. They changed identity.

Two molecular signals drive that conversion in sequence. First, retinoic acid - a derivative of vitamin A - is broken down in the foveola region, reducing its local concentration. Retinoic acid normally promotes blue cone identity, so its degradation suppresses the formation of new blue cones in the central region. This limits blue cone numbers but does not convert those already present.

The second step involves thyroid hormones. After retinoic acid sets the initial pattern by limiting blue cone production, thyroid hormones act on the remaining blue cone cells in the foveola and drive them to convert into red and green cones. The two signals operate sequentially, not simultaneously.

"First, retinoic acid helps set the pattern. Then, thyroid hormone plays a role in converting the leftover cells," Johnston said. "That's very important because if you have those blue cones in there, you don't see as well."

Challenging the Migration Model

The conversion finding directly challenges the prior model. For three decades, the field assumed that cell fate is fixed once cone identity is established - blue cones become blue cones permanently and those in the fovea simply relocate. The new data shows that cells can be redirected from one cone fate to another by the right combination of molecular signals at the right developmental window.

Johnston acknowledged the caveat clearly: "The main model in the field from about 30 years ago was that somehow the few blue cones you get in that region just move out of the way, that these cells decide what they're going to be, and they remain this type of cell forever. We can't really rule that out yet, but our data supports a different model. These cells actually convert over time, which is really surprising."

The inability to completely rule out migration reflects the limits of organoid systems - cells in culture do not perfectly replicate the physical and signaling environment of the developing eye in the body. Confirming cell conversion as the dominant mechanism will require corroborating evidence from other experimental approaches, ideally including primate tissue when available.

Path Toward Therapies for Vision Loss

Macular degeneration, the leading cause of irreversible vision loss in older adults, destroys precisely the cells studied here - the photoreceptors of the central retina. There is currently no treatment that restores lost photoreceptors. Cell replacement therapy - transplanting healthy photoreceptors to replace damaged or lost ones - is a long-term goal in retinal medicine, but it requires producing the right type of photoreceptor in sufficient numbers and with correct identity.

Understanding how cone fate is specified during normal development provides the blueprint for doing that deliberately. If retinoic acid degradation and thyroid hormone signaling specify red and green cone identity during foveal development, those same pathways may be manipulable in organoid systems to produce populations enriched for the specific cone types needed for therapeutic transplantation.

"The goal with using this organoid technology is to eventually make an almost made-to-order population of photoreceptors," said co-author Katarzyna Hussey, a former doctoral student in Johnston's lab. "A big avenue of potential is cell replacement therapy to introduce healthy cells that can compensate for cells that have been lost."

That clinical application remains distant - producing transplant-ready photoreceptors that survive, integrate, and function in a diseased retina is a substantially harder problem than characterizing development in a lab dish. But the developmental insights established here define the molecular targets and the developmental logic that any therapeutic approach will need to replicate.