DEVELOPMENTAL BRAIN RESEARCH ELSEVIER
DevelopmentalBrain Research96 (1996) 290-294
Genesis of the frog retinal pigment epithelium L.D. Beazley *, M. Tennant, T.L. Tomlin, J.M. Preuss, L.-A. Coleman, S.A. Dunlop Department of Zoology, University of WesternAustralia, Nedlands, WA 6907, Australia
Accepted 25 June 1996
To examine cell generation in the frog retinal pigment epithelium (RPE), representative developmental stages from tail-bud to adulthood received a single injection of [3H]thymidine. Animals were killed either 24 h or several weeks later; eyes were sectioned and processed by standard autoradiographic procedures and viewed by epi-polarised illumination. The distribution of [3H]thymidine-labelled cells indicated that the RPE is formed throughout life, including in adulthood, by cell addition at the ciliary margin, matching the pattern for the neural retina. In addition, a very small number of peripapillary RPE cells underwent division but only in the adult. Keywords: Amphibia; Cell division; Retinal pigment epithelium; [3H]Thymidine
The vertebrate retina consists of two closely apposed tissues, namely the pigment epithelium (RPE) and the overlying neural retina . These are both of neural origin being formed from the outer and inner layers of the invaginating optic cup respectively . Both are essential for normal vision. Functions of the RPE include providing nutrients to the neural retina, removing the debris of shed outer segments of photoreceptors and absorbing stray light within the eye . Moreover, the RPE can trans-differentiate into a variety of tissues; these include the neural retina and lens in both adult Urodeles [1,5] and embryonic chick [ 15,16]. Studies of retinal development have largely concentrated on the neural retina rather than the RPE. For example, in frogs, autoradiographic studies have revealed that cells are added to the neural retina throughout life at the ciliary margin [3,14]. The progressive build-up of cells is at least partially responsible for increases in retinal area . Here we provide the results of a complementary study to establish patterns of cell generation in the frog RPE. Local frogs Litoria moorei and Limnodynastes dorsalis were collected under licence. Pre-metamorphic stages were maintained in glass tanks containing aerated spring water and were fed with Biorel fish food; post-metamorphic animals were housed in tanks containing damp gravel and were fed with mealworms two or three times a week. Animals were maintained at 22 + 2°C.
* Corresponding author. Fax: (61) (9) 380 1029.
We examined six representative stages. The four premetamorphic stages were tail-bud, early and mid-larval and metamorphic climax, equivalent to stages 35, 42, 53 and 66 for Xenopus laevis ; juveniles were 2 - 3 months post metamorphosis and adults at least 1 year post metamorphosis. Two series were conducted. In the first (the 'pulse kills'), animals were killed 24 h after injection to reveal the site of [3H]thymidine incorporation. In the second series (the 'pulse leaves'), most animals were killed at the next developmental stage selected for analysis; a minority were injected at both mid-larval and metamorphic climax stages and killed as juveniles. Adults were kept for 10 weeks before being killed. Five L. moorei and five L. dorsalis were studied at each time point for both series; as exceptions, two adults of each species were used per series. Some L. dorsalis had been used previously to analyse genesis of the neural retina . Under brief anaesthesia (immature stages: immersion in 0.1% tricaine methosulphonate; adults: inhalation of Halothane), animals were injected intraperitoneally using a Hamilton syringe with 2 - 1 0 /zCi [3H]thymidine (specific activity 888 B q / m m o l , Amersham, the dose increasing with age) and allowed to recover. Subsequently animals were overdosed (immature stages: immersion in 1% tricaine methosulphonate; adults: Saffan, 1 m g / k g i.p.), tissue dissected and fixed in buffered formalin (pH 7.4) for at least a week. At tail-bud and early larval stages, whole heads were processed for sectioning; at later stages, eyes were orientated by a dorsal cut or a suture thread before
0165-3806/96/$15.00 Copyright © 1996 Elsevier Science B.V. All rights reserved. Pll S0165-3806(96)00125-3
LD. Beazley et al. / Developmental Brain Research 96 (1996) 290-294
being removed and the cornea and lens dissected away. Tissue was wax-embedded and sectioned at 7/xm, parallel to either the naso-temporal or dorso-ventral axis of the eye. Tissue was processed autoradiographically using NTB emulsion (Kodak), exposed for 10-14 days at 4°C and developed in D19 before staining with cresyl violet or haematoxylin and eosin and mounting in Depex. Tissue was examined ~tt 400 X or 1000 × magnification using a standard light microscope with an epi-polarising attachment . The autoradiographic grains of reduced silver appear white, distinguishing them from the underlying black melanin (Fig. 1A,C,E, Fig. 2A,C). To restrict our analysis to ce,lls which had undergone their final DNA replication within hours of the injection, we
included only those nuclei overlain by half or more of the maximal count of silver grains . The microscope was connected to a MD.II Digitiser and Hewlett Packard Plotter, allowing positions of labelled cells to be recorded along the length of each section. Every tenth section was analysed and maps reconstructed as Mercator's projections (Fig. 3A-K). Results were similar for both species and are described together (summarised in Fig. 3L). In animals which received a single injection and were examined after 24 h ('pulse kills'), labelled RPE cells were present at every stage examined. At all stages except the adult, labelled cells were confined to the ciliary margin (Fig. 1A-D, Fig. 3A-E). In the adult, the vast majority of labelled cells
Fig. 1. Cross-sections of the ciliary margins of the tail-bud (A,B), mid-larval (C,D) and adult stages (E,F) of L. dorsalis, each injected with [3H]thymidine 24 h before being killed. Tissue is viewed by epi-polarised illumination (A,C,E) and by bright (B,D,F) field. The ganglion cell layer is uppermost in each case. The location of [3H]thymidine-labelled RPE cells is indicated (single arrows) in each micrograph. A [3H]thymidine-labelled cell of the neural retina is also indicated (double arrows) to aid orientation. Cresyl violet, section thickness 7/zm, scale bar is 20 /~m throughout.
L.D. Beazley et al. / Developmental Brain Research 96 (1996) 290-294
were located at the ciliary margin but a few were also seen in peripapillary retina (Fig. 1E,F, Fig. 2A,B, Fig. 3F). As expected, label in the neural retina was seen exclusively at the ciliary margin in all stages including adults (Fig. 1). For animals with longer survival times ('pulse leaves'), labelled RPE cells formed an annulus centred on the optic nerve head and at a distance from the ciliary margin (Fig. 2C,D, Fig. 3G-I,K), coincident with labelling in the neural retina. The younger the animal at the time of injection, the nearer the annulus to the optic nerve head. In adults, the annulus was located only 5 0 - 1 0 0 /~m from the ciliary margin. Presumably cells lying central to the annulus were
generated before the [3H]thymidine injection and those peripheral thereafter. When [3H]thymidine was injected at both at mid-larval and metamorphic climax stages, two concentric annuli were seen in the juvenile retina (Fig. 3J), zone between the labelled annuli presumably being generated in the period between injections. On close inspection, the most centrally located of the labelled RPE cells at the ciliary margin of the short survival series (Fig. 1A,C,E,) and in the annuli for the longer surviving animals (Fig. 2C) were seen to lie slightly peripheral to the leading edge of labelled cells of the neural retina; as reported previously [3,14], the leading
Fig. 2. Cross-sections of the peripapillary retina of L. moorei (A,B) and mid-nasal retina of L. dorsalis (C,D) showing the RPE (towards the bottom) overlain by neural retina (outer nuclear layer in A,B; inner and outer nuclear layers in C,D). A,B: The adult was injected with [3H]thymidine24 h before being killed (i.e. the 'pulse kill' series). C,D: The juvenile was injected with [3H]thymidineat the mid-larval stage (i.e. the 'pulse leave' series). Tissue is viewed by epi-polarised illumination (A,C) and by bright (B,D) field. The location of [3H]thymidine-labelledRPE cells is indicated (single arrows) in each micrograph. The location of two pigment granules in RPE nticrovillae is indicated (double arrows) to aid orientation in B. A [3H]thymidine-labelledcell in the inner nuclear layer of the neural retina is indicated (double arrows) to aid orientation in D. Cresyl violet, section thickness 7/xm, scale bar is 20 /xm throughout.
L.D. Beazley et al. / Deoelopmental Brain Research 96 (1996) 290-294
edge for the neural retina was in the inner nuclear layer. From this observation, we conclude that RPE cells tend to exit the germinal zone sl:ightly later than the most advanced cells of the neural retina. Overall, our results indicate that the RPE is generated throughout life, including adulthood, and that the generation takes place almost exclusively at the ciliary margin. As a result, the first generated RPE cells come to lie adjacent to the optic nerve: head and the youngest at the periphery. Moreover, the similar patterns in the RPE and the neural retina suggest that the two retinal compartments are subject to similar developmental signals, or possibly a common one.
Similar to our results for frogs, antibody labelling of newly post-mitotic RPE cells suggests that generation of the Urodele RPE is limited to the ciliary margin . We predict that a comparable pattern will be found in fish. However, species such as the goldfish may possess additional foci of RPE generation away from the ciliary margin, to accommodate ‘hot spots’ of rod cell generation within the established neural retina . By contrast, cell generation in mammalian RPE and neural retina are largely complete early in life with differing patterns in the two retinal compartments [2,4,13]. The results presented here imply that the frog RPE grows almost exclusively by cell addition at the ciliary
. : : . . i.
.* ‘. : : :
: 0. f . . .*
.. :. PULSE-LEAVES
Fig. 3. A-K: Maps of [3H]thymidir~e-labelled RPE cells. The left column (A-F) shows maps for ‘pulse-kill’ animals (i.e. those killed after 24 h); the right column (G-K) shows maps for ‘pulse leaves’, in other words those injected with [3H]thymidine at one stage and raised to another (G-1,K) or injected at 2 stages and raised to a t hird (J). Maps are for L. dorsalis except for the adult, which is L.moorei. 35, 42 and 53 are tail-bud, early and mid-larval stages respectively, MC = metamorphic climax, Juv = juvenile. Every tenth section is represented. The optic nerve head is shown as a circle. The scale below diagram K refers to A-K, that below diagram F to F alone, D,V,N,T = dorsal, ventral, nasal and temporal respectively. L: Synoptic diagram to indicate the experimental design and sequence Iof RPE generation relative to the optic nerve head (ONH). Approximate ages are shown in days (d) or weeks (wks) since fertilisation; yr = year. For ‘pulse kills’, the arrows indicate the stage when [3H]thymidine was injected and the animal examined. For most ‘pulse leaves’, [3H]thymidine was injected at the time indicated by the end of the tail of the arrow and the animal was killed at the time indicated by the head of the arrow. The double-headed arrow refers to animals injected at the mid-larval and metamporphic climax stages and killed as a juvenile; each tail indicates the time of a [3H]thymidine injection, the foremost arrowhead represents the time tissue was examined.
L.D. Beazley et al. /" Developmental Brain Research 96 (1996) 290-294
margin. Moreover, the comparable location in the long survival series of labelled cells of the RPE and neural retina indicates that cells born simultaneously remain in apposition thereafter. In the neural retina, a stretching apart of existing elements contributes to increases in retinal area . The area of the RPE must therefore be similarly enhanced by a stretching of established cells. In post-natal rodent retina, increases in RPE cell area are associated with an extensive wave of acytokinetic division to convert the majority to a multi-nucleate status . Acytokinetic division may also underlie the small minority of peripapillary RPE cells which were found to incorporate [3H]thymidine in our adult animals. A search for multi-nucleate cells would address this issue.
  
Acknowledgements This research was funded by the National Health and Medical Research Council (NH&MRC). S.A.D. is a Research Fellow, NH&MRC. The experiments conformed to the guidelines of N H & M R C and the Animal Welfare Committee of the University of Western Australia. Frogs were collected with permission from the Department of Conservation and Land Management.
References   Abe, S.-I. and Eguchi, G., An analysis of differentiative capacity of pigmented epithelial cells of adult newt iris in clonal cell culture, Dev. Growth. Differ., 19 (1977) 309-317.  Bodenstein, L. and Sidman, R.L., Growth and development of the mouse retinal pigment epithelium. 1. Cell and tissue morphometrics and topography of mitotic activity, Dev. Biol., 121 (1987) 192-204.  Coleman, L.-A., Dunlop, S.A. and Beazley, L.D., Patterns of cell
division during visual streak formation in the frog Limnodynastes dorsalis, J. Embryol. Exp. Morphol., 83 (1984) 119-135. Fleming, P.A., Braekevelt, C.R., Harman, A.M. and Beazley, L.D., Retinal pigment epithelium and photoreceptor maturation in a wallaby, the quokka, J. Comp. Neurol., in press. Gaze, R.M. and Watson, W.E., Cell division and migration in the brain after optic nerve lesions. In: Wolstenholme, G.E.W. and O'Connor, M. (Eds.), Growth of the Nervous System, Ciba Foundation Symposium, 1968, pp. 53-67. Johns, P.R. and Femald, R.D., Genesis of rods in the teleost fish retina, Nature, 293 (1981) 141-142. Klein, L.R., MacLeish, P.R. and Wiesel, T.N., Immunolabelling by a newt retinal pigment epithelium antibody during retinal development and regeneration, J. Comp. Neurol., 293 (1990) 331-339. Mann, I.D., The Development of the Human Eye, British Medical Association, London, 1964. Nieuwkoop, P.D. and Faber, J., Normal Table of Xenopus laevis (Daudin). North-Holland, Amsterdam. Panova, I.G. and Stroeva, O.G., Retinal pigment epithelium: pattern of proliferative activity and its regulation by intraocular pressure in postnatal rats, J. Embryol. Exp. Morphol., 83 (1975) 271-293. Rakic, P., Kinetics of proliferation and latency between final division and onset of differentiation of cerebellar stellate neurons and basket neurons, J. Comp. Neurol., 147 (1973) 523-546. Rapaport, D.H., Herman K.G. and LaVail, M.M., Epi-polarisation and incident light microscopy readily resolve an autoradiograph or heavy metal label from obscuring background or a second label, J. Neurosci. Methods, 41 (1992) 231-238. Rapaport, D.H., Rakic, P., Yasamura, D. and LaVail, M.M., Genesis of the retinal pigment epithelium in the Macaque monkey, J. Comp. Neurol., 363 (1995) 359-376. Straznicky, K. and Gaze, R.M., The growth of the retina in Xenopus laevis: An autoradiographic study, J. Embryol. Exp. Morphol., 26 (1971) 67-79. Tsunematsu, Y. and Coulombre, A.J., Demonstration of transdifferentiation of neural retina from pigmented retina in culture, Dev. Growth Differ., 23 (1981) 297-311. Yasuda, K.G., Eguchi, G. and Okada, T.S., Age-dependent changes in the capacity of transdifferentiation of retinal pigment cells as revealed by clonal culture, Cell. Differ., 10 (1981) 3-11. Zinn, K.M. and Marmot, M.F. (Eds.) Retinal Pigment Epithelium. Harvard University Press, Boston, MA, 1979.