As a particularly accessible part of the central nervous system, the mouse retina has emerged as a favorite model for analyzing the development, structure and function of neural circuits. As a model for human vision, however, it suffers from a severe drawback. Humans and most other primates see objects clearly only when they look straight at them, because only then does their image falls on a specialized central portion of the retina called the fovea. The fovea mediates our high acuity vision as well as much of our chromatic (color) vision. Although it occupies only ~1% of the retinal surface, the fovea provides over 50% of visual input to the cerebral cortex, which is ultimately responsible for interpreting what we see. Moreover, diseases that selectively affect this region, such as macular degeneration and diabetic macular edema, which are leading causes of blindness in the developed world, selectively affect the macula. (The macula is a slightly larger region than the fovea, with the fovea at its center.) The problem is that among mammals, only primates have a fovea. One cannot, therefore, study critically important aspects of visual function or dysfunction in mice or other widely-used model organisms.
Although structural and functional specializations of the fovea have been documented by many groups, the molecular mechanisms that underlie these specializations remain almost entirely unknown. In their new paper, Peng, Karthik and others from the Sanes lab set about to redress this balance. They used high throughput methods for characterizing the genes expressed by single cells (called single cell RNA-seq or scRNA-seq), and applied them to the retina of the macaque (Macaca fascicularis), a widely used primate model in clinical vision studies. They analyzed around 165,000 cells, divided roughly equally between fovea and peripheral retina. Using computational methods, they divided these into 65-70 cell types per region, thereby generating the first cell atlas of the primate retina – lists of the cell types that comprise the retina and of the genes that each type expresses. They then used the atlas to ask whether foveal specializations resulted from different cell types in the two regions, different proportions of the same cell types, or different genes expressed by corresponding types. They found that the two regions have similar, though not identical “parts lists” but with different proportions and many differentially expressed genes. Among the genes that distinguish foveal cells from their peripheral counterparts are many that hint at bases for structural and functional specializations. Peng et al. also used their data to map expression of nearly 200 genes that have been implicated in blinding diseases, and found that some involved in macular degeneration and diabetic macular edema are selectively expressed by foveal cells – an intriguing hint about why these diseases affect the fovea. Finally, they showed that many of the genes and differences they found are shared with human retina, supporting the relevance of their findings to human vision. In summary, their work provides a foundation that many labs can use as they strive to understand how humans see and why some go blind.