The compound eye of the fruit fly, Drosophila melanogaster, is an excellent model system for studying such diverse topics as tissue determination, compartment boundary establishment, cell fate specification, cell proliferation and apoptosis, cell and planar polarity, signal transduction and cell-cell communication. The retina is a particularly good experimental model in part because it contains a limited number cell types that follow a precise and stereotyped mode of development. Additionally, thirty years of work has produced a detailed survey of the eye's cellular, molecular and morphological development. Eye development begins during embryogenesis when two groups of cells located at the dorso-lateral regions of the head express the Pax6 gene eyeless (Figure 1). The roughly 20 cells that initially express eyeless (and several other eye specification genes) will proliferate and organize themselves into a monolayer epithelial sheet called the eye-antennal imaginal disc. This disc will form the template from which the adult eye is created.

The eye imaginal discs of 1st instar larvae contain cells that are, to our inspection, completely unpatterned and undifferentiated. As the larvae undergo consecutive rounds of molting and growth the eye imaginal disc undergoes dramatic increases in cell proliferation and will eventually reach a size of nearly 20,000 cells. Overt patterning of the retina begins at the posterior margin of the eye disc when a wave of differentiation initiates and sweeps across the epithelium much like a wave sweeps across the ocean. The leading edge of this mobile compartment boundary is called the morphogenetic furrow. As the furrow passes cells are organized into unit eyes or ommatidia (Figure 2).Within a developing ommatidium lie twenty cells of which eight are photoreceptors and twelve are non-neuronal accessory cells. The eight photoreceptors are the first to be specified. Each ommatidium undergoes a stereotyped series of cell determination events: the R8 is specified first followed by the stepwise addition of the R2/5 pair, the R3/4 pair, the R1/6 pair and finally the R7 neuron. The non-neuronal cone and pigment cells are added to each unit eye after the specification of the photoreceptors. A revealing feature of ommatidial construction is the assembly line-like development of each unit eye: all steps of the assembly process can be observed in a single eye imaginal disc (Figure 3).

During pupal eye development the clusters of photoreceptor neurons undergo dramatic morphogenetic movements that results in their rearrangement into asymmetrical trapezoid elements, a feature of the adult ommatidium. Each unit eye is an exact duplicate of its neighbor with the only observable difference being that the chiral form of ommatidia in the dorsal half of the retina is a mirror image of the chiral form found in the ventral half of the eye. The two chiral forms meet at the equator, which is the boundary between the dorsal and ventral regions of the retina (Figure 4).

Above the photoreceptors lie four cone cells that secrete the overlying pseudocone and lens. Photons of light that strike the external surface of the eye are directed through the lens and pseudocone towards the photoreceptor cells. Surrounding the photoreceptors are several classes of pigment cells that optically insulate one ommatidium from another. The pigment cells also function in the packaging of the unit eyes. They are arranged around the photoreceptor core and build the hexagonal array of the adult compound eye (Figure 5).

A decade of experimentation has identified a group of nuclear factors that form the eye specification or retinal determination cascade. Additionally, a number of signaling cascades have been shown to transmit instructions from the cell surface to members of this network. A wealth of genetic, molecular and biochemical data suggests that these nuclear factors and signaling pathways are part of a complicated and interwoven regulatory network. Each nuclear factor and each signaling cascade has functional orthologs in all seeing animals including vertebrates and, more importantly, several human retinal disorders are attributed to mutations within the human orthologs of the fly genes (Figure 6). These results and observations have suggested that these genes play a crucial role in eye specification in all seeing animals.

Excitingly, loss-of-function mutant phenotypes and forced expression assays have yielded many spectacular results. For example, removal of any member of the core eye specification cascade leads to severe if not total loss of retinal tissue. On the other hand, forced expression of these factors results in the redirection of non-retinal tissues towards an eye fate (Figure 7). Even more exciting is the observation that expression of several mammalian orthologs can (1) rescue loss-of-function fly mutations and (2) induce the formation of ectopic eyes. These results further the argument that the members of the retinal determination network sit atop the hierarchy of genes that regulate eye development in all seeing animals. Additionally, these findings have caused a profound rethinking of the origins of the eye and it is now accepted by most that the eye has had a monophyletic origin.

Figure 6:
Retinal Determination Genes: Development and Disease