Control of vertebrate retinal cell production

EXPERIMENTAL

NEUROLOGY

115,65-68

(19%)

Control of Vertebrate Retinal Cell Production AND~EASF.MACKAND~USSELLD.FERNALD' Institute

of Neuroscience,

University

of Oregon,

INTRODUCTION Regeneration of vertebrate sensory cells can be considered an extension and elaboration of the process of cellular repair. As such, cellular repair, replacement, and system regeneration are related, with the former an essential process in every cell and the latter thought to be restricted to a limited range of organisms and specific cell types. Two general rules can be stated based on data on the regeneration of sensory systems (7). First, the likelihood that regeneration of a sensory system might occur increases with the proximity of the receptor surface to the environment. Second, the likelihood that regeneration will occur decreases along a continuum from cold- to warm-blooded vertebrates (Table 1). An example of these rules is the renewal process in vertebrate rod photoreceptors. As in all cells, molecular repair occurs continuously in photoreceptors, but is particularly evident because of their striking structural polarity. Young (12) first demonstrated that photoreceptor outer segments are renewed in an organized fashion, with disk membranes in the outer segment of rod photoreceptors assembled at the outer segment base, disof

Oregon

97403

placed outward by new disks, and eventually shed at the tip. This renewal process is regulated at the molecular level by light and a circadian rhythm (3, 9). Under certain circumstances, degeneration of photoreceptors can be delayed in the presence of basic fibroblast growth factor (bFGF) although there is no evidence that photoreceptor cells can be replaced (2). Just as cellular renewal requires the proper extracellular environment, so will any receptor replacement. Discovery of these processes and their cellular and molecular bases must be analyzed in a system where the actors are accessible. For retinal renewal, in uiuo systems prohibit many manipulations necessary to discover the nature of cellular interactions and to assay intercellular communication. In vitro systems, while allowing necessary access to the tissue may be so different from the normal state that the insights gained are ultimately not applicable. For these reasons, we have developed a retinal slice preparation in which normal developmental processes can be traced and putative regulatory mechanisms studied. Our motivation is to understand how cell division and cell fate are determined in the retina in order to understand retinal repair.

Regeneration of vertebrate sensory cells can be seen as an extension and elaboration of the process of cellular repair and to understand repair requires knowledge of how cell division and cell fate are determined. To approach these problems, we have developed a slice culture for the teleost retina. Cells continue to divide in the same pattern in this slice culture as they do in uiuo as demonstrated with r3H]thymidine labeling. Moreover, cells which divided in culture became retinal cell phenotypes as identified with monoclonal antibodies. Some presumptive rod progenitors in the outer nuclear layer in the center of the retina were also labeled cone-specific, possibly as a regeneration response. Those data add to the evidence that cell fate is determined by the environment. This slice preparation will be a useful model system for analyzing putative environmental cues responsible for guiding cell proliferation and differentiation in the fish retina.

’ Present address: Neuroscience Program & Department chology, Stanford University, Stanford, CA 94305.

Eugene,

MATERIALS

AND METHODS

Retinal slices were prepared from 12- to 14-day-old juvenile African cichlid fish (Huplochromis burtoni) using the methods described in Mack and Fernald (8). Briefly, following anesthesia, eyes were dissected out of the animal and the cornea, lens, and sclera were removed before embedding the eyecup in 3% agarose. Subsequently, the embedded eyecup was sectioned on a vibratome at 150 pm and agarose (0.7%) was used to stabilize the slices in a 35-mm culture dish. Slices were cultured in 2 ml of supplemented L-15 medium (Flow Lab), (10 mA4 glucose, 25 miI4 taurine, 5 mglml gentamicin, 5 mM Hepes, 5 pglml insulin-tr~sferrin-selenite; all Sigma) and 3% fish embryo extract prepared from homogenates of 4- to 5-day old embryos. The covered dishes were kept in a humidified chamber under light/dark and temperature conditions which mimic those in which the animals are kept. To establish a defined culture medium and to test the effect of embryo extract on proliferation, we replaced the embryo extract with basic fibroblast growth factor

Psy-

65 Ail

Copyright Q 1992 rights of reproduction

0014.4886192 $3.00 by Academic Press, Inc. in any form reserved.

66

MACK

TABLE

AND

1

Schematic Illustration of the Relationships among Cellular Renewal and Repair Processeswithin Vertebrate Phyla

‘?

has not been

studied

Note. When considering the repair to replacement sequence as a continuum, the general conclusion is that cold blooded vertebrates are able to replace sensory neural tissue in contrast to warm blooded vertebrates. In no cases, however, has the sensory system been tested to see whether it supports “normal” behavior. [After Fernald in CIBA, 19911.

FERNALD

(bFGF, 20 rig/ml medium). To show that cells proliferate in this slice culture, we used [3H]thymidine to tag newly dividing cells. [3H]thymidine was added to the culture medium (2.5 &i/ml medium, spec act 68 Ci/ mmol) for 12 h and the tissue then processed (see below). To demonstrate that retina-specific cell phenotypes are produced in the slice culture, we used monoclonal antibodies which label specific photoreceptor types in combination with [3H]thymidine. For these experiments, slices were allowed to survive for 2.5 days after the 12-h exposure to [3H]thymidine. Following these procedures, retinal slice cultures were fixed in 4% paraformaldehyde in PBS, for 1 h and processed for immunocytochemistry and/or autoradiography. To distinguish photoreceptor phenotypes, we used cone-specific (13) and rod-specific (14) monoclonal antibodies. Whole slices were incubated in an antibody for 2-3 days at 4°C (0.1% Triton X, 1% DMSO). This step was followed by incubation with secondary anti-mouse antibody and peroxidase anti-peroxidase complex (Jackson Immuno Res. Lab., Inc.) 12-24 h each. Antibody binding was visualized by diaminobenzidine reaction. Immediately after the antibody staining, the slices were dehydrated and embedded in plastic (Immunobed, Polysciences) and sectioned at 3 pm. Sections were then

FIG. 1. Photomicrograph of the proliferative area in a 3-pm section through the margin of a retinal slice stained with cresyl violet. This slice was cultured for 24 h in defined medium containing fibroblast growth factor (see Materials and Methods). After 12 h in culture, [3H]thymidine was added to the culture medium to label dividing cells located in the margin and in the outer nuclear layer (ONL, arrows). The slice was subsequently fixed, embedded in immunobed, and processed for autoradiography. INL, inner nuclear layer. Scale bar = 50 pm.

CONTROL

OF

VERTEBRATE

RETINAL

CELL

PRODUCTION

67

FIG. 2. Photomicrograph of a 3-pm plastic section through the central portion of a retinal slice stained with a rhodopsin-specific antibody. The slice was exposed to [3H]thymidine during the first 12 h in culture, fixed after 3 days in culture, and afterward processed for immunocytochemistry and autoradiography. Many cells in the outer nuclear layer (ONL) show antibody binding (arrowheads). A cell in the outer nuclear layer is labeled by the antibody and silver grains (arrows). (a) Focus on the cell bodies, (b) same frame with focus on the silver grains. INL, inner nuclear layer; PE, pigment epithelium. Scale bar = 25 pm.

processed for autoradiography days later.

(15) and developed

3-5

RESULTS

We have shown previously divisions in the organotypic

(8) that the pattern of cell retinal slice mimic those

found from comparable experiments in vivo (4). As in normal tissue, virtually all cells in the peripheral growth zone and scattered cells in the outer nuclear layer (ONL) incorporated [3H]thymidine in slice-cultured retinas. When embryo extract was replaced with bFGF, the distribution of proliferating cells in the slice culture was unchanged (Fig. 1). However, in slices incubated in this

FIG. 3. Photomicrographs of an oblique 3-pm section through the outer and inner nuclear layer of a retinal slice close to the proliferative area. The slice was exposed to [3H]thymidine during the first 12 h in culture, fixed after 3 days in culture, and afterward processed for immunocytochemistry and autoradiography. The tissue was stained with a cone-specific antibody. (a) Focus on the immunostained cell bodies; (b) same frame with focus on silver grains indicating cells which divided while the slice was in culture. A cell is double labeled by antibody binding and silver grains (arrows). This indicates that this cell in the outer nuclear layer expressed the cone phenotype in culture. Scale bar = 25 Wm.

68

MACK

AND

defined medium, relatively few cells were labeled, suggesting that additional factors are necessary for normal growth. To ascertain the phenotype of cells produced in the slice, we used photoreceptor-specific antibodies to label rods and cones. In the outer nuclear layer, we found cells labeled with both [3H]thymidine and the rod-specific antibody (Fig. 2) demonstrating that some cells had divided in culture and begun expressing opsin. Although rod outer segments often degenerate after a few days in culture, cell bodies and some processes are nonetheless labeled after a 2.5-day culture period. These processes can be identified as rod outer segments. We cannot distinguish if these are old or regenerated outer segment discs. Cells which were double labeled with cone-specific antibodies and [3H]thymidine can also be found in the retinal slice culture. Cone-specific antibody staining was restricted to the cells in the ONL and the marginal area of the slices. Cone photoreceptors typically retain their inner and outer segments and processes with pedicle in the slice culture. Double-labeled cells were located mostly in the marginal zones, but were also occasionally found in the ONL (Fig. 3).

FERNALD

molecular cues are needed to modify proliferation and differentiation in retinal tissue. The slice culture described here is a first step toward developing an accessible model system with which to answer these questions. ACKNOWLEDGMENTS We thank Dr. Adamus for the rod-specific antibody, the monoclonal antibody facility for the cone antibody, and Tom Kasten and Matthew Fry for technical support. Supported by Grant EY 05051 to R.D.F. and training Grant GM 07257 to A.M.

REFERENCES 1.

2.

3. FERNALD,

R. D., R. MCDONALD, AND J. KORENBROT. 1987. Light-dark cycle of opsin mRNA production in toads and fish. Invest. Ophthal. Vis. Sci. Suppl. 28(3): 184.

4. FERNALD,

DISCUSSION Two salient features distinguish the retinas of coldblooded vertebrates (fish and amphibia) from those of warm-blooded vertebrates: (i) Retinas continue to grow after embryonic development, not only by stretching but also by generation of new neurons and photoreceptors (reviewed in (4)); (ii) Retinas can regenerate after chemical or mechanical damage, including new connections to higher brain centers (1, 11). We have shown that in the organotypic slice culture system, neuronal progenitors continue to divide and differentiate. Since progenitor cells in the outer nuclear layer of the differentiated fish retina normally produce only rods in uiuo, these cells have been termed rod precursors (5,lO). Our finding that some of these cells express a cone phenotype in vitro suggests two conclusions: (a) rod precursors are not committed to the rod phenotype but are able to take on other cell fates. (b) In the slice culture system, this change in cell fate may be a regenerative reaction to damage caused by the preparation of the slice. Within the tissue slice, progenitor cells can be challenged with different environmental cues and the influence on cell fate studied. Since rod progenitors in fish are a continuing source of new photoreceptors, it may be possible to modify their fate using regulatory reagents. It remains to be discovered what cellular and

ATTARDI, D. G., AND R. W. SPERRY. 1963. Preferential selection of central pathways by regenerating optic nerve fibers. Exp. Neural. 7: 46-64. FAKTOROVICH, E. G., R. H. STEINBERG, D. YASUMURA, M. T. MATTHES, AND M. M. LAVAIL. 1990. Photoreceptor degeneration in inherited retinal dystrophy delayed by basic fibroblast growth factor. Nature 347: 83-86.

5.

R. D. 1989. Retinal Rod Neurogenesis. Development of the Vertebrate Retina (B. L. Finlay and D. R. Sengelaub, Eds.), pp. 31-42. Plenum, New York. FERNALD, R. D. 1989. Fish vision. In Development of the Vertebrate Retina (B. L. Finlay and D. R. Sengelaub, Eds.), pp. 247265. Plenum, New York.

7. FERNALD, CIBA

R. D. Foundation.

1991.

Vertebrate

Sensory

System

Renewal,

8. MACK, 9.

A. F., AND R. D. FERNALD. 1991. Thin slices of teleost retina continue to grow in culture. J. Neurosci. Methods 36: 195-202. KORENBROT, J. I., AND R. D. FERNALD. 1989. Circadian rhythm and light regulate opsin mRNA in rod photoreceptors. Nature

337: 454-457. 10.

RAYMOND, P. A. 1985. The unique origin of rod photoreceptors in the teleost retina. Trends Neurosci. 8: 12-17.

11.

WOLBURG, H. 1981. Axonal transport, degeneration and regeneration in the visual system of the goldfish. Adu. Anat. Embryol. Cell Biol. 67. YOUNG, R. W. 1967. The renewal of photoreceptor outer segments. J. Cell Biol. 33: 61-72. HAGEDORN, M., AND R. D. FERNALD. 1991. Retinal growth and cell addition during embryogenesis in the teleost, Haplochromk burtoni, submitted for publication. ADAMUS, G., A. ARENDT, Z. S. ZAM, J. H. MCDOWELL, AND P. A. HARGRAVE. 1988. Use of peptides to select for anti-rhodopsin antibodies with desired amino acid sequence specificities. PeptideRes. 1:42-47. FERNALD, R. D. 1983. Neural basis of visual pattern recognition. In Advances in Vertebrate Neuroethology (J.-P. Ewart, R. R. Capranica, and D. J. Ingle, Eds.), pp. 569-580. Plenum, New York.

12. 13.

14.

15.