 |
||||
|
||||
|
||||
|
||||
Previous Article | Next Article
The Journal of Neuroscience, September 1, 2001, 21(17):6745-6757
A Mutation of Early Photoreceptor Development, mikre oko, Reveals Cell-Cell Interactions Involved in the Survival and Differentiation of Zebrafish Photoreceptors
Geoffrey Doerre and Jarema Malicki Department of Ophthalmology, Harvard Medical School, Boston, Massachusetts 02114
ABSTRACT
TOP
ABSTRACT
INTRODUCTION
To gain insight into mechanisms involved in photoreceptor development, we characterized a zebrafish mutation in the mikre oko locus that produces early loss of photoreceptor cells. mikre oko photoreceptors lose their elongated morphology at the time of wild-type outer segment formation and undergo cell death within a few days. To investigate whether this phenotype involves cell-cell interaction defects, we performed analysis of genetically mosaic animals. Interactions of mikre oko photoreceptors with wild-type cells rescue several aspects of the mutant phenotype. When placed in a wild-type environment, mikre oko photoreceptor cells retain elongated morphology and survive longer. Moreover, although mutant mikre oko photoreceptor outer segments develop only infrequently and are usually disorganized, mikre oko cone and rod cells in mosaic retinas develop robust outer segments that closely resemble the wild type. In contrast to the outer segments, the proximal regions of mikre oko photoreceptor cells, including their inner segments, the nuclear regions, and the synaptic termini, retain the mutant appearance. mikre oko outer segment rescue is not mediated by interactions with the retinal pigment epithelium. These studies demonstrate that the differentiation of outer segments is surprisingly independent from the more proximal photoreceptor cell features and that outer segment development includes retinal pigment epithelium-independent cell-cell interactions.
Key words: retina; photoreceptor; outer segment; cell-cell signaling; genetics; zebrafish
INTRODUCTION
The vertebrate neural retina is an evolutionarily highly conserved sensory organ that develops from a single neuroepithelial sheet into a complex, laminar structure comprising primarily six neuronal and one glial cell type (Ramon y Cajal, 1893
; Rodieck, 1973
The genetic mechanisms that are involved in vertebrate photoreceptor development are largely unknown. Current models propose that both environmental cues and the ability of pluripotent progenitor cells to respond to them change over time, resulting in an ordered appearance of different retinal cell types (Cepko et al., 1996
Cell-cell interactions have been demonstrated to play important roles in CNS development. Analysis of genetically mosaic animals containing intermingled wild-type and mutant cell clones provides a productive approach to search for evidence of cell-cell signaling events. One of the first studies of this type in the retina demonstrated that the photoreceptor degeneration phenotype in the RCS rat originates in the retinal pigment epithelium (RPE) (Mullen and LaVail, 1976
The zebrafish has been recently established as a genetic model of vertebrate eye development (for review, see Malicki, 2000a
MATERIALS AND METHODS
Animals. Zebrafish (Danio rerio) were kept on a 14/10 hr light/dark cycle according to standard procedures (Westerfield, 1994
Histology. Mutant and wild-type siblings were collected at appropriate time points, fixed in 4% (w/v) paraformaldehyde in PBST (PBS, 0.1% Tween 20) for 2 hr at room temperature or overnight at 4°C, washed in PBST for 10 min, and dehydrated in 10 min washes with a graded ethanol series (30, 50, 75, 85, and 95%, and two times 100%). After overnight infiltration in JB-4 resin (Polysciences, Warrington, PA), embryos were embedded according to manufacturer's directions, sectioned at 3 µm, and dried on a hotplate. Sections were collected on Superfrost Plus slides (Fisher Scientific, Houston, TX), stained with methylene blue-azure II for 10 sec (Humphrey and Pittman, 1974
Electron microscopy. Embryos were fixed in 4% (w/v) paraformaldehyde/2% glutaraldehyde in 75 mM phosphate buffer, pH 7.2, for 2 hr on ice, decapitated with a razor blade, and post-fixed with 2% osmium tetroxide in 50 mM phosphate buffer, pH 7.2. After being washed in 50 mM maleate buffer, pH 5.9, for 15 min on ice, specimens were stained with 2% uranyl acetate in maleate buffer for 15 min on ice. Then, specimens were dehydrated in 10 min steps through a graded ethanol series as above, infiltrated in a graded Epon-propylene oxide series for several hours to overnight, and embedded in Epon (Polysciences). After thin sectioning, sections were transferred to formvar-coated grids (Fisher Scientific) and poststained with lead citrate. Analysis was performed using Phillips EM410 and CM10 electron microscopes.
Immunohistochemistry. Embryos were fixed in 4% (w/v) paraformaldehyde, washed in PBS, infiltrated in 30% sucrose in PBS, and embedded using TBS freezing medium (Triangle Biomedical Sciences, Durham, NC). Cryosections (20 µm) were heated to 80°C for 2 min, air-dried for 60 min, washed in PBS, and blocked for 1 hr with 10% goat serum in PBD (0.5% Triton X-100, 0.1% Tween 20, 1% DMSO in PBS). Sections were incubated in primary antibody in block solution for 2-5 hr at room temperature or overnight at 4°C. The following antibodies and dilutions were used: Fret 43 (1:100; University of Oregon Stock Center, Eugene, OR), Ret 11 (1:200; University of Oregon Stock Center), Zn8 (1:25; University of Oregon Stock Center), anti-rhodopsin (1:1000), anti-UV opsin (1:1000), anti-green opsin (1:500), anti-red and anti-blue opsin (1:200; gifts of Tom Vihtelic and David Hyde, University of Notre Dame), and anti-carbonic anhydrase II (1:250; gift from Paul Linser, Whitney Laboratory, St. Augustine, FL). After being washed three times in PBD, sections were exposed to the appropriate combinations of FITC-, Cy3-, or Cy5-conjugated secondary antibodies (1:500; Jackson ImmunoResearch, West Grove, PA) for 1 hr, washed three times in PBD, mounted in 50% glycerol, 2% propyl gallate (Sigma, St. Louis, MO) in 0.2 M Tris-HCl, pH8.0, and analyzed using a Leica TCS 4D confocal microscope. Z-series of mosaic clones (1 µm) were recorded and analyzed using Scanware 5.1 (Leica, Nussloch, Germany) and Photoshop software.
Quantitation of cell survival. Embryos were fixed in 4% (w/v) paraformaldehyde in PBST for 2 hr at room temperature or overnight at 4°C and decapitated. Embryonic heads were washed in PBST and water for 10 min each and treated with prechilled acetone for 5 min at
20°C. Subsequently, embryos were washed once in water and then in PBST for 5 min each, incubated with 200 µg/ml collagenase (C0773; Sigma) in PBST for 30 min at room temperature, fixed again (10 min, 4% PFA at room temperature), and washed three times for 5 min each in PBST. After being blocked with 10% goat serum in PBST for 1 hr at room temperature, heads were incubated overnight at 4°C using opsin or Fret 43 antisera at the dilutions specified above. Washing and secondary antibody detection were performed as described above with the appropriate fluorophores. To evaluate the distribution and intensity of staining, specimens were dehydrated, infiltrated, embedded in JB-4 resin, and sectioned at 3 µm as described above, with the following modifications: the percentage of JB-4 catalyst that was added was decreased to 0.7%, and sections were air-dried and shielded from light to protect fluorophore stability. Sections were immersed for 15 min in 0.2 µg/ml Hoechst 33258 (Molecular Probes, Eugene, OR) in PBD, washed three times for 5 min each with water, and mounted in glycerol-propyl gallate as described above. Alternatively, frozen sections were stained with Fret 43 and anti-opsin antibodies as described above. Nuclear staining was observed on a Zeiss Axioscope microscope equipped with a mercury arc lamp and recorded using a SPOT digital camera as described above. Photoreceptor-specific staining was analyzed by confocal microscopy as described above.
Mosaic analysis. Blastomere transplantations were performed as previously described (Ho and Kane, 1990
Donor-derived cells were detected using either HRP staining or immunofluorescence. For HRP detection, donor and host embryos were fixed in 4% (w/v) paraformaldehyde in PBST for 2 hr at room temperature or overnight at 4°C, washed in PBST and water for 10 min each, and treated with prechilled
Alternatively, host embryos were cryosectioned and processed for Fret43, Ret11, or anti-carbonic anhydrase immunofluorescence as described above, with the additional modification that dextran tracer signal was detected using Cy3-conjugated streptavidin (1:500; Jackson ImmunoResearch) during the secondary antibody step. Confocal analysis of mosaic clones was performed as above.
RESULTS
mikre oko affects early stages of photoreceptor development
In the vertebrate eye, photoreceptor cells form a layer of radially elongated cells that are adjacent to the retinal pigment epithelium and separated from the inner nuclear layer by a synaptic outer plexiform layer (OPL). The photoreceptor cell layer is readily apparent in histological sections of wild-type zebrafish retinas at 3 and 5 dpf (Fig. 1A,C,E). Histological analysis of mikre oko retinas demonstrates a severe defect of the PRCL already at 3 dpf. Fewer elongated cells are seen in the periphery of the retina at 3 dpf (Fig. 1B), as compared with a wild-type PRCL at equivalent stage. At 3 dpf, there is still some evidence of a discontinuous OPL, and a number of morphologically distinguishable photoreceptor cells are present in the center of the retina of most animals (Fig. 1B, arrowhead; data not shown). Although most of the inner retina appears relatively normal, morphologically distinguishable horizontal cells are largely absent. Additionally, there appears to be an increased number of densely staining pyknotic nuclei throughout both the INL and the outer retina (data not shown). By 5 dpf, the central retina of mok no longer contains any morphologically distinguishable photoreceptors (Fig. 1D,F). Although no morphological defects outside the retina are obvious in mok, mutant larvae do not survive past 10 dpf.
View larger version (119K):[in this window]
[in a new window]
Figure 1. Histological analysis of mikre oko development. Nomarski images of transverse sections through central retinas at or near the optic nerve at indicated time points. Retinas in C-F have been treated with PTU to inhibit pigment granule formation in the RPE. A, C, Wild-type retinas display a continuous, uniform layer of elongated photoreceptor cells at both 3 and 5 dpf (arrowheads). B, mok retinas at 3 dpf lack a distinct PRCL. Although elongated cells are sometimes seen in the central retina (arrowhead), cells along the periphery appear rounded or dead (arrows). Horizontal cells are not distinguishable. D, By 5 dpf, all surviving mok PRCL cells have adopted a rounded morphology (arrows). E, F, Enlargements of the central retina seen in C and D, close to the optic nerve. E, In wild-type retinas, tangentially elongated horizontal cells, radially elongated photoreceptor cell bodies, and the outer segments (arrows) are clearly visible. F, In mok retinas, on the other hand, few elongated nuclei (arrowheads) and no outer segment outlines are apparent. In A-D, dorsal is upward, RPE is right. In E and F, dorsal is left, RPE is upward. * indicates the optic nerve. rpe, Retinal pigment epithelium; prcl, photoreceptor cell layer; inl, inner nuclear layer; opl, outer plexiform layer; ipl, inner plexiform layer; gcl, ganglion cell layer; d, dpf. Scale bar: A-D, 100 µm; E, F, 20 µm.
To precisely determine the timing of photoreceptor loss, a time course of immunohistochemical staining was performed with Fret 43, an antibody specific for double cones, the most prevalent photoreceptor subtype in early larval retinas (Larison and Bremiller, 1990
View larger version (75K):[in this window]
[in a new window]
Figure 2. Pattern and time course of mikre oko photoreceptor loss. Confocal images of transverse retinal sections near the optic nerve immunostained with Fret 43 antibody for double cones (green) and with anti-carbonic anhydrase for Mueller glia (red) at indicated time points. A, C, E, G, Wild-type retinas exhibit a uniform photoreceptor staining throughout the PRCL at all stages. Mueller glia extend radial processes spanning the entire thickness of the retina. B, The mok PRCL already displays photoreceptor loss in the periphery at 3 dpf (arrows), whereas some elongated cells persist in the central PRCL (arrowhead). Mueller glia also appear fewer in number, and their processes are absent or disorganized. D, F, Photoreceptor loss increases in severity (arrows), and all cells adopt a rounded shape by 4 dpf. Glial staining is decreased. H, By 7 dpf, Fret 43 immunoreactivity frequently is absent in mok, and few carbonic anhydrase-positive cells can be found. Dorsal is upward, and RPE is right in all panels. * indicates optic nerve. d, dpf. Scale bar, 100 µm.
To evaluate whether cell loss occurs at the same rate for all types of photoreceptor cells, the percentage of opsin- and Fret 43-expressing cells remaining in mutant retinas at 7 dpf was assayed by immunohistochemistry. At 7 dpf, the number of wild-type photoreceptor cells accounts for ~19% of the total number of retinal cells (169 of 881, n = 5) (Table 1). Based on antibody staining, ~40% (80 of 199) of these photoreceptor cells are Fret 43-positive, 22% (43 of 199) are UV opsin-positive, 17% (34 of 199) are blue-opsin positive, and 21% (42 of 199) are rod opsin-positive (Fig. 3, Table 1). In contrast to the wild type, in mok retinas at this stage, the PRCL is no longer morphologically distinguishable, and on average only 3% (6 of 199) (Table 1) of cells are positive for photoreceptor-specific markers. Cell loss of different photoreceptor categories exhibits small differences only. The Fret 43-positive double cones and UV cones appear to deteriorate at a somewhat faster rate than rods and blue cones (Table 1). In parallel to photoreceptor loss, 7 dpf mok retinas suffer a 35% decrease in total cell number relative to the wild type (Table 1). This is more than the 20% decrease that would be expected if only the normal photoreceptor population were absent, and these differences are likely caused by INL abnormalities (see below). The time course and pattern of photoreceptor loss observed with Fret 43 antibody has been confirmed with blue, red, UV, and rod opsin in situ probes (Fig. 3E-H). Thus, it appears that both rods and cones are initially formed. However, their differentiation becomes aberrant early in development as evidenced by the abnormal shape of opsin- and Fret 43-positive cells, as well as their decreased number. The rate of photoreceptor loss is approximately the same for all photoreceptor types including both rods and cones, suggesting that the primary defects are not confined to a single photoreceptor type or to a subset of photoreceptor types.
View this table: [in this window]
[in a new window]
Table 1. Cell survival in mikre oko retinas
View larger version (45K):[in this window]
[in a new window]
Figure 3. Early differentiation of individual mikre oko photoreceptor types. A-D, Confocal images of transverse retinal sections immunostained with anti-rhodopsin antibody (green) and anti-carbonic anhydrase antibody to Mueller glia (red) at 60 hpf (A, B) and 72 hpf (C, D). Zn-8 antibody to ganglion cells and optic nerve (green) was included as an internal control to monitor the stage of early embryos. A, C, In wild-type retinas, rhodopsin localizes mostly to the photoreceptor outer segments (arrowheads). B, D, In mok, rhodopsin is mislocalized throughout the PRCL (arrowheads), even appearing near the INL. Also, note the reduction of glial staining (red in B). E-H, Bright-field images of transverse retinal sections showing in situ hybridization signal obtained with rod opsin (E, F) or red opsin (G, H) probes at 72 hpf. E, G, Wild-type PRCL displays an even distribution of opsin expression in elongated photoreceptor cells (arrowheads). F, H, The mutant PRCL displays discontinuities, and opsin-expressing cells have an abnormal, rounded morphology (arrowheads). I-L, Confocal images of transverse retinal sections (ventral retina only) immunostained with anti-blue opsin antibody (green) and Fret 43 (red) (I, J), or with anti-UV opsin antibody (green) and anti-rod opsin antibody (red) (K, L). I, K, Wild-type rod and cone visual pigments localize to the apical (outer segment) portion of the PRCL (arrowheads), whereas Fret 43 stains the entire bodies of double cones. J, L, The mok PRCL displays discontinuities, defective photoreceptor morphology, and abnormal distribution of opsin immunoreactivity (arrowheads). In all panels, dorsal is upward, lens is left. * indicates the optic nerve. Scale bar: A-H, 100 µm; I-L, 20 µm.
The subcellular distribution of photoreceptor visual pigments in mok was examined with antisera specific to opsin genes. Mutant immunostaining patterns differed noticeably from wild-type retinas. At 60 and 72 hpf, wild-type rod opsin is concentrated at the apical ends of photoreceptors in developing outer segments (Fig. 3A,C). In mok retinas, on the other hand, already at 72 hpf, rod opsin staining is concentrated in foci frequently mislocalized relative to the apical photoreceptor termini (Fig. 3D). Similar distribution patterns are observed for blue, UV, red, and green opsins (Fig. 3J-L; data not shown). Simultaneous mislocalization of visual pigments in rods and several cone types is another indication that the mok mutation affects mechanisms common to the whole photoreceptor cell class and not to individual photoreceptor types.
mikre oko affects the formation of both photoreceptor outer segments and synaptic termini
To gain a better understanding which aspects of photoreceptor morphogenesis are affected in mikre oko, electron microscopic analysis was performed at 3 and 5 dpf, revealing a variety of structural defects. In wild-type zebrafish retina, the majority of inner and outer segments develop between 60 and 72 hpf (Branchek and Bremiller, 1984
View larger version (132K):[in this window]
[in a new window]
Figure 4. Ultrastructure of wild-type and mikre oko photoreceptor cells. A, Electron micrograph of a wild-type retina at 3 dpf reveals elongated nuclei in the PRCL, prominent outer segments adjacent to the RPE (arrowheads), and ellipsoid assemblies of mitochondria in the inner segments in between (asterisks). B, At 3 dpf, mok retinas have few to no outer segments (arrowheads) or ellipsoids (asterisks). The photoreceptor nuclei display elongated morphology toward the central retina (left), but become increasingly disorganized toward the periphery (right). Additionally, numerous cell corpses are present (p). C, D, Enlargements of the PRCL apical region at 3 dpf. C, In wild-type outer segments, a stack of membranous discs is clearly discernible. D, mok outer segments are found much less frequently, are smaller, and appear disorganized. E, F, Enlargements of the PRCL synaptic region at 3 dpf. E, Wild-type photoreceptor termini form characteristic invaginations occupied by postsynaptic processes of bipolar and horizontal cells. A synaptic ribbon is indicated (arrowhead). F, mok synaptic termini are poorly differentiated, and no synaptic ribbons are observed. G, H, Enlargements of PRCL apical region at 5 dpf. G, Wild-type photoreceptors form well differentiated inner (asterisks) and outer segments (arrowheads). H, mok outer segments, when present, are severely disorganized, and inner segments are not obvious. Arrowheads indicate outer segments. Asterisks indicate inner segments. RPE is upward in all panels. Scale bars, 1 µm.
Cell-cell interactions are involved in the development of elongated morphology of photoreceptor cells
The mutant photoreceptor phenotypes could be caused by intrinsic defects in the photoreceptor cells themselves, or by defective cell-cell or cell-extracellular matrix interactions. To distinguish between these possibilities, genetically mosaic animals were constructed using the blastomere transplantation technique. In this method, cells are removed from a tracer-labeled donor embryo and placed into unlabeled hosts (Ho and Kane, 1990
Blastomere transplantations were performed to produce four classes of mosaics: wild-type embryos containing wild-type donor cell clones, wild-type embryos containing mutant donor cell clones, mutant embryos containing wild-type donor clones, and mutant embryos containing mutant donor clones. In these experiments, donor cells were labeled using biotinylated dextran and detected using avidin-HRP conjugate. Representative clones are shown in Figure 5, A, C, E, and G. Transplants between genotypically wild-type embryos produce retinal clones in which donor cells contribute to all cell layers of 7 dpf retina (Fig. 5A). Approximately 20% of donor cells contribute to PRCL cells (152 of 783 cells in 17 independent clones) and develop elongated morphology characteristic of photoreceptor cells. Similarly, mutant mikre oko cells transplanted into a wild-type host contribute to all layers of the retina, and ~9% of them (44 of 472 cells in 17 independent clones) form morphologically distinguishable, elongated photoreceptors surviving at least until 7 dpf (Fig. 5C). In contrast, 0% (0 of 102 cells in five independent clones) of wild-type cells developing in mutant PRCL display elongated morphology at this time (Fig. 5E). Similarly, as expected on the basis of histological studies, mutant donor cells in a mutant environment do not develop elongated morphology (0 of 103 cells in eight independent clones at 5 and 7 dpf) (Figs. 5G, 6F). Thus, the mok photoreceptor morphology phenotype displays a strong cell-nonautonomous component. Because the percentage of mutant mok photoreceptor cells in clones developing in a wild-type environment is lower than that found in wild-type to wild-type transplants, the mok phenotype appears to combine cell-autonomous components as well. In a wild-type PRCL environment, numerous mok mutant photoreceptors retain an elongated morphology until at least 7 dpf, more than twice as long as photoreceptors in unmanipulated mok PRCL, which completely lose elongated morphology by 4 dpf (compare Fig. 2D to 5C,D). These results indicate that the development of elongated photoreceptor morphology involves cell-cell interactions and that these interactions are disrupted in mok.
View larger version (75K):[in this window]
[in a new window]
Figure 5. Morphology and survival of photoreceptor cells in mikre oko mosaic animals. Examples of four possible genotypic combinations of mosaic retinas at 7 dpf are visualized with two different detection methods. A, C, E, G, Bright-field images of mosaic retinas processed with HRP-avidin to detect donor-derived cells (brown). B, D, F, H, Confocal images of mosaic retinas immunostained with a mix of all anti-opsin antibodies (green) and Cy3-streptavidin to detect tracer (red). A, B, Wild-type donor cells contribute to all retinal layers in wild-type hosts, including the PRCL (arrows). C, D, Elongated morphology of mok photoreceptor cells is rescued in wild-type environment (arrows). E, Wild-type cells do not form elongated photoreceptors next to the RPE of mok host embryos. F, Wild-type donor cells express opsins in mok retinas (arrows). G, H, Mutant mok cells do not form morphologically distinguishable photoreceptors in a mutant background and, with rare exceptions, do not express opsin genes (arrows). Insets in the bottom left of B and D show twofold magnifications of the areas enclosed in boxes. The approximate position of the outer plexiform layer is indicated with dashes. Note that mikre oko photoreceptor cells in the wild-type environment extend from the outer plexiform layer to the opsin-positive outer segment layer. Donor-host genotypes are indicated at the bottom of each panel. Dorsal is upward, and RPE is right in all panels. m, Mutant; w, wild type. Scale bar, 10 µm.
View larger version (124K):[in this window]
[in a new window]
Figure 6. Outer segment rescue in mikre oko photoreceptors at 3, 5, and 7 dpf. PTU treatment inhibits the appearance of pigmentation and permits the observation of outer segment morphology. A, Wild-type donor-derived cells form elongated photoreceptors in a 3 dpf wild-type host PRCL, extending outer segment rudiments (arrowheads) toward the RPE (bracket). B, mok donor cells form elongated photoreceptors in a 3 dpf wild-type host PRCL. Although their inner segment regions appear abnormally thin, the outer segments of these cells closely resemble the wild type. Inset shows an enlargement of a mok photoreceptor cell in a wild-type retina. The horizontal arrow indicates an abnormal constriction in the inner segment region of this cell. Such constrictions are not seen in wild-type cells. C, At 5 dpf, wild-type donor cells display distinct outer segments (arrowheads) and robust inner segments. D, mok donor cells elaborate both cone (thin arrowhead) and rod (wide arrowhead) outer segments. The inner segment regions of rescued mok photoreceptor cells are frequently abnormally thin. This is evident in the case of the rod cell indicated with a wide arrowhead. E, F, Neither wild-type nor mok donor cells form elongated photoreceptors at 5 dpf when placed in mutant environment, and there is no evidence of outer segments. G, H, I, Enlargements of HRP-stained donor photoreceptors in PTU-treated retinas at 5 dpf. G, Wild-type photoreceptors in wild-type environment display robust inner and outer segments (arrowheads) and well developed synaptic pedicles (vertical arrow). Both rod (H) and cone (I) mok outer segments are rescued in wild-type environment (arrowheads), whereas more proximal cell regions deteriorate and no pedicles are distinguishable (vertical arrows). Panel I is a composite of two images taken in different planes of focus. Horizontal arrows indicate abnormal constrictions in the inner segment region. In wild-type cells, such constrictions are found only between the nucleus and the synaptic terminus. J, K, Confocal reconstructions of donor photoreceptors at 7 dpf. J, Tracer-containing wild-type donor cells (red) in a wild-type host PRCL. Outer segments (arrowheads) and a pedicle (vertical arrow) are indicated. K, mok donor cells adopt an elongated shape (red), yet their inner segment regions (horizontal arrow) and synaptic termini (vertical arrows) are grossly abnormal. RPE is upward in all panels. m, Mutant; w, wild type. Scale bar: A-F, 8 µm; G-K, 4 µm.
Cell-cell interactions are involved in the survival of photoreceptor cells
The survival of photoreceptor cells in mosaic retinas was quantitated as a ratio of photoreceptors to the total number of cells in donor clones. Because in mutant retinas photoreceptor cells lose elongated morphology and cannot be distinguished by morphological criteria, we stained cryosections of mosaic retinas with a mix of antibodies directed to all five zebrafish opsins to determine the photoreceptor survival rate in mutant eyes. When wild-type clones of donor cells develop in wild-type host retinas, ~21% form photoreceptors (in 18 clones inspected, 134 of 653 cells formed photoreceptors at 7 dpf) (Fig. 5B, Table 2). These results are in agreement with the previous studies in which we used HRP-avidin to detect donor-derived photoreceptor cells as well as with the estimates based on Hoechst-stained zebrafish retinas (Table 1). As expected, mutant mikre oko photoreceptors developing in the mutant environment survive at a low rate of ~8.2% (47 of 576 cells in 18 clones) (Fig. 5H, Table 2). To investigate whether the survival of mutant cells depends on interactions with the surrounding tissue, in the same experiments we determined the survival rates of mutant photoreceptor cells in wild-type environment. Approximately 22% (83 photoreceptors of 380 cells total, 16 independent clones) of mok mutant donor cells form photoreceptors in wild-type hosts, consistent with the results obtained in previous experiments (Fig. 5D, Table 2). This represents a dramatic increase compared with the survival rate of the mutant cells developing in the mutant environment. These results demonstrate that mok photoreceptor survival is enhanced by interactions with a wild-type environment, indicating that the mok gene is directly or indirectly involved in cell-cell interactions that promote photoreceptor cell survival. Surprisingly, in contrast to the loss of the elongated morphology, this cell-nonautonomous effect is not obvious in wild-type to mutant transplants. Approximately 16% of wild-type cells express opsin in a mutant environment (54 of 330 cells, 12 independent clones) (Fig. 5F, Table 2). One interpretation of the increased survival of wild-type versus mutant cells in a mutant environment is that wild-type cells are providing a rescuing cell-nonautonomous component through interactions with each other. More likely, this observation indicates that although mok-dependent cell-cell interactions are sufficient to promote the survival of mutant photoreceptor cells, they are not necessary for the survival of wild-type cells.
View this table: [in this window]
[in a new window]
Table 2. Survival of photoreceptors in mosaic retinas RPE-independent cell-cell interactions are involved in the development of photoreceptor outer segments
The close physical association of photoreceptors to the RPE obscures photoreceptor outer segments. To study inner and outer segment development in mosaic animals, after blastomere transplantations, host animals were raised in the presence of 1-phenyl-2-thiourea (PTU), a compound that inhibits pigment accumulation in the RPE (Westerfield, 1994
View this table: [in this window]
[in a new window]
Table 3. Photoreceptor morphology in mosaic retinas What cells in the environment are responsible for the rescue of the mok outer segment phenotype? Mosaic analysis of RCS rat retina as well as other embryological studies suggests that the RPE may be involved in this process (Hollyfield and Witkovsky, 1974
View larger version (144K):[in this window]
[in a new window]
Figure 7. Effect of RPE cells on photoreceptor development in mosaic mikre oko retinas at 5 dpf. A, B, In a wild-type host, mok donor-derived RPE cells (B) have no deleterious impact on outer segment morphology (arrowheads) in adjacent photoreceptor cells when compared with wild-type donor-derived RPE cells (A, arrowheads). C, Wild-type donor RPE cells do not have a rescuing effect on mok putative photoreceptors, which retain a rounded morphology even when positioned directly adjacent to a wild-type cell. D, Similarly, putative mutant photoreceptor cells in a mok host retina remain rounded when positioned adjacent to a mok RPE donor cell. Donor-host genotypes are indicated at the bottom of each panel. RPE is upward in all panels. m, Mutant; w, wild type. Scale bar, 10 µm.
Mueller glia and horizontal cells are affected in mikre oko retina
Horizontal cells are clearly distinguishable in sections of wild-type retinas as a layer of elongated cells at the ventricular edge of the inner nuclear layer oriented tangentially to the curvature of the outer plexiform layer (Fig. 1A,C,E). These cells are not apparent or are severely reduced in mikre oko at 3 dpf (Fig. 1B,D,F). Because presently there are no reliable molecular markers to establish horizontal cell identity in embryonic zebrafish, it is uncertain whether this INL phenotype reflects a loss of the normally elongated horizontal cell morphology or a true absence of these cells. The morphological phenotype of horizontal cells appears to be cell-nonautonomous because mutant cells in wild-type environment display a normal shape and are present at a normal frequency relative to other cell types. At 7 dpf, in wild-type clones in a wild-type environment, ~8% of cells develop as horizontal neurons (34 of 412 cells, 12 clones). A similar fraction of horizontal cells, 9% (22 of 238 in 8 clones), is present in mok clones developing in wild-type environment (Fig. 7A,B).
Mueller cells are the major glial component of the retina, forming extensions toward the inner and outer limiting membranes (Dowling, 1987
DISCUSSION
mikre oko mutation affects early aspects of photoreceptor differentiation
In the zebrafish retina, a layer of photoreceptor cells first becomes morphologically distinguishable between 40 and 50 hpf (Schmitt and Dowling, 1999
A loss of outer segments shortly after their appearance has only rarely been observed in vertebrate mutants of photoreceptor development. Several reports of mutations affecting zebrafish eye development describe the occurrence of irregular or deformed outer segments (Malicki et al., 1996
In other vertebrate model systems, early defects in outer segment development have been observed only infrequently. In the mouse, in which outer segments appear between postnatal day 9 and 10 (White et al., 1993
mikre oko affects mechanisms common to the development of all photoreceptor types
The zebrafish retina features five types of photoreceptor cells, each characterized by a specific morphology, spectral sensitivity, and visual pigment expression (for review, see Malicki, 1999
Although the molecular basis of signal transduction appears to be similar in rods and cones, many components of the signal transduction machinery in these cell types are encoded by separate, although highly homologous, genes. This is true, for example, for the visual pigments themselves, transducin, cGMP phosphodiesterase, cGMP-gated Na/K channel, and arrestin (Lerea et al., 1986
Cell-cell interactions influence survival and differentiation of photoreceptor cells
Cell-nonautonomy has been demonstrated in the case of several mutations affecting development of vertebrate photoreceptor cells. A mosaic mouse expressing a rhodopsin Pro344 to Ser mutant transgene in a subset of cells suffers uniform degeneration across the entire chimeric retina, not only in transgene-expressing photoreceptors but also in adjacent normal photoreceptors (Huang et al., 1993
In all of the above experiments, the presence of mutant photoreceptor cells has been shown to cause a decreased survival of wild-type cells. Whether the opposite is true and the mutant cells survive longer when surrounded by wild-type cells has not been investigated. Our experiments demonstrate, for the first time, that mutant photoreceptor cells display increased survival when surrounded by wild-type cells. When placed in wild-type environment, mok photoreceptor cells survive until 7 dpf, and their survival rate is >2.5 times higher than that of mutant cells in mutant environment. These results are consistent with the observations in mammals indicating that photoreceptor survival involves cell-nonautonomous components. Previous investigations of cell-nonautonomous photoreceptor defects were limited, however, to the evaluation of cell survival (Huang et al., 1993
Outer segments differentiate in the absence of normal inner segment and nuclear regions
An unexpected result of these experiments is that cell-cell interactions are sufficient to rescue the development of mikre oko outer segments. Although this effect is not completely penetrant, the size and shape of rescued rod and cone outer segments closely resemble the wild type. This is in sharp contrast to the mutant mok retinas in which only very few, abnormally shaped photoreceptor outer segments persist until 5 dpf. The rescued outer segments enlarge from 3 to 5 dpf in the absence of properly developed inner segments and nuclear regions, indicating that their development takes place largely independently from these cell features. This is surprising, given that the outer segment metabolism is very active and is presumed to depend on the continuous transport of polypeptides, including rhodopsin, from other cell regions (for review, see Sung and Tai, 2000
The rescue of outer segment is RPE-independent
Both genetic and embryological studies provide evidence indicating that the retinal pigment epithelium plays an essential role in the differentiation and maintenance of the photoreceptor outer segments. Surgical removal of the RPE results in an almost complete absence of photoreceptor outer segment development (Hollyfield and Witkovsky, 1974
mikre oko mutant phenotype is more pronounced in retinal periphery
Analysis of zebrafish photoreceptor mutants revealed several patterns of photoreceptor loss (for review, see Malicki, 2000a
FOOTNOTES Received Dec. 11, 2000; revised June 1, 2001; accepted June 7, 2001.
This work was supported by awards from the March of Dimes Birth Defects Foundation, Research to Prevent Blindness, and National Institutes of Health (NIH) Grant RO1 EY11882-01A1 to J.M. G.D. was supported by NIH Training Grants T32-EY07145 and T32-AG00222. We thank Paul Linser, Tom Vihtelic, and David Hyde for providing useful reagents, as well as Xiangyun Wei and Zac Pujic, members of the Malicki laboratory, for sharing their technical expertise. For insightful comments on this manuscript, we thank Drs. Thaddeus Dryja, Connie Cepko, Tiansen Li, Francesca Pignoni, Seth Blackshaw, Stephanie Hagstrom, and Zac Pujic.
Correspondence should be addressed to Dr. Jarema Malicki, Department of Ophthalmology, Harvard Medical School, 243 Charles Street, Boston, MA 02114. E-mail: jarema_malicki{at}meei.harvard.edu.
REFERENCES
- Becker TS, Burgess SM, Amsterdam AH, Allende ML, Hopkins N (1998) not really finished is crucial for development of the zebrafish outer retina and encodes a transcription factor highly homologous to human Nuclear Respiratory Factor-1 and avian Initiation Binding Repressor. Development 125:4369-4378[Abstract].
- Biel M, Seeliger M, Pfeifer A, Kohler K, Gerstner A, Ludwig A, Jaissle G, Fauser S, Zrenner E, Hofmann F (1999) Selective loss of cone function in mice lacking the cyclic nucleotide-gated channel CNG3. Proc Natl Acad Sci USA 96:7553-7557
[Abstract/Free Full Text] .- Blanks JC, Mullen RJ, LaVail MM (1982) Retinal degeneration in the pcd cerebellar mutant mouse. II. Electron microscopic analysis. J Comp Neurol 212:231-246[ISI][Medline].
- Bok D, Hall M (1971) The role of the pigment epithelium in the etiology of inherited retinal dystrophy of the rat. J Cell Biol 49:664-682
[Abstract/Free Full Text] .- Bonigk W, Altenhofen W, Muller F, Dose A, Illing M, Molday RS, Kaupp UB (1993) Rod and cone photoreceptor cells express distinct genes for cGMP-gated channels. Neuron 10:865-877[ISI][Medline].
- Branchek T, Bremiller R (1984) The development of photoreceptors in the zebrafish, Brachydanio rerio. I. Structure. J Comp Neurol 224:107-115[ISI][Medline].
- Brockerhoff SE, Dowling JE, Hurley JB (1998) Zebrafish retinal mutants. Vision Res 38:1335-1339[ISI][Medline].
- Burrill J, Easter S (1995) The first retinal axons and their microenvironment in zebrafish cryptic pioneers and the pretract. J Neurosci 15:2935-2947[Abstract].
- Cepko C, Austin C, Yang X, Alexiades M, Ezzeddine D (1996) Cell fate determination in the vertebrate retina. Proc Natl Acad Sci USA 93:589-595
[Abstract/Free Full Text] .- Chen S, Wang Q, Nie Z, Sun H, Lennon G, Copeland N, Gilbert D, Jenkins N, Zack D (1997) Crx, a novel Otx-like paired-homeodomain protein, binds to and transactivates photoreceptor cell-specific genes. Neuron 19:1017-1030[ISI][Medline].
- Dowling J (1987) In: The Retina. Cambridge, MA: Belknap.
- Drummond IA, Majumdar A, Hentschel H, Elger M, Solnica-Krezel L, Schier AF, Neuhauss SC, Stemple DL, Zwartkruis F, Rangini Z, Driever W, Fishman MC (1998) Early development of the zebrafish pronephros and analysis of mutations affecting pronephric function. Development 125:4655-4667[Abstract].
- Dryja T, Li T (1995) Molecular genetics of retinitis pigmentosa. Hum Mol Genet 4:1739-1743[Abstract].
- Dryja T, McGee T, Reichel E, Hahn L, Cowley G, Yandell D, Sandberg M, Berson E (1990) A point mutation of the rhodopsin gene in patients with auto
