Development of the rabbit retina, III: Differential retinal growth, and density of projection neurons and interneurons.

To provide a quantitative description of postnatal retinal expansion in rabbits, a new procedure was developed to map the retinae, which cover the inner surface of hemispheres or parts of rotation ellipsoids, in situ, onto a single plane. This method, as well as the known distribution of Müller cells per unit retinal surface area, were used to estimate the redistribution of specific subpopulations of Müller cells within different topographic regions of the retinae. Müller cells are known to exist as a stable population of cells 1 week after birth and can therefore be used as "markers" for determining tissue expansion. Our results show that differential retinal expansion occurs during development. Peripheral retinal regions expand at least twice as much as the central ones. Furthermore, there is a greater vertical than horizontal expansion. This differential retinal expansion leads to a corresponding redistribution of 5-hydroxytryptamine (5-HT) accumulating amacrine cells. Differential retinal expansion, however, does not account for all of the changes in the centro-peripheral density gradient of cells in the ganglion cell layer (GCL)--mostly retinal ganglion cells--during postnatal development. The changes in the ganglion cell layer were evaluated in Nissl-stained wholemount retinal preparations. Additionally, the difference between expansion-related redistribution of cells in the GCL and Müller cells was confirmed in wholemount preparations where Müller cells (identified as vimentin positive) and cells in the GCL (identified by fluorescent supravital dyes) were simultaneously labeled. It is assumed that many of the ganglion cells within the retinal center are not translocated during retinal expansion, possibly because their axons are fixed. In contrast, 5-HT accumulating amacrine cells--which are interneurons without a retinofugal axon--display a passive redistribution together with the surrounding retinal tissue.

Development of the rabbit retina: II. Müller cells.

Müller (glial) cells of the rabbit retina were stained with antibodies against the intermediate filament protein vimentin in retinal wholemounts from various developmental stages. Both the density of stained profiles and the mean diameter of these profiles were measured, with the microscope focus in the inner plexiform layer of the retinae. Within this retinal layer, every Müller cell possesses one stout vitread process; thus counts of the stained profiles allow an estimation of their number. After postnatal day (P) 9, the total number of stained cells was slightly above 4 million per retina; for the adult rabbit retina, this agrees well with earlier data obtained by our group based on another method, as well as with published data from other groups. We suggest that after P 9, only Müller cells are stained, and this population is numerically stable. In contrast, neonatal retinae contained significantly more stained profiles. This indicates that either the total number of Müller cells is reduced by "physiological cell death" or that additional cells are stained neonatally. We discuss why we favour the second possibility. After P 9, two peculiarities occur in the Müller cell population: (1) their density decreases gradually, to a greater extent in the retinal periphery than in the center (i.e., in the "visual streak"), and (2) Müller cell diameters increase, again more in the periphery than in the center. We argue that differential retinal expansion leads to dispersion of the pre-existing cell population and allows for widening of the Müller cell processes. We conclude that Müller cells can be used postnatally in the rabbit retina as "landmarks" of expansion.