THE JOURNAL OF COMPAFWI'IW NEUROLOGY 301530-92 (1990)
Further Study of the Outward Displacement of Retinal Ganglion Cells During Optic Nerve Regeneration,With a Note on the Normal Cells of Dogiel in the Adult Frog ERIC L. SINGMAN AND FRANK SCALIA Department of Anatomy and Cell Biology, State University of New York, Health Science Center, Brooklyn, New York 11203
ABSTRACT In a previous study we observed massive retinal ganglion cell death in adult Rana pipiens after periods of optic nerve regeneration, and reported that large numbers of the surviving cells had become displaced bodily into the inner plexiform layer of the affected eye (Scalia et al.: Brain Research 344:267-280,1985). The outwardly displaced cells could be identified as retinal ganglion cells because they could be back-filled with horseradish peroxidase (HRP) injected into the regenerated optic nerve. Quantitative observations on the abnormal outward displacement of ganglion cells are reported here. Parallel observations on normally displaced ganglion cells (cells of Dogiel) are also reported to clarify the distinctions between these two classes of cells. For the present work, injections of HRP of varying size were placed in the optic tectum bilaterally in 3 normal frogs and 9 frogs sustaining unilateral optic nerve regeneration. Most injections were centered at loci mapping the middle region of the nasal retina. The retinas were examined as flat-mounts and in-section. In 8 other frogs sustaining optic nerve regeneration, the HRP was administered bilaterally directly to the optic nerves in the orbit. Ganglion cells were labeled by retrograde transport of the HRP in the retinal ganglion cell layer in both the normal and affected eyes in areas topographically isomorphic with the tectal areas subtended by the injections. In the normal eyes, the orthotopic ganglion cells formed a strict monolayer, and virtually no cells existed in the inner plexiform layer. In the retinas sustaining optic nerve regeneration, the retinal ganglion cells abnormally displaced into the inner plexiform layer were also labeled topographically in correspondance with the injection sites. The abnormally displaced cells comprised 5.5% of the total population of surviving neurons in the retinal ganglion cell and inner plexiform layers. The mean outward dislocation of the displaced cells, as measured in one frog surviving optic nerve crush for 8 weeks, was 69.9 2 2.4% of the distance across the inner plexiform layer, which itself was uniformly 14.3 % 0.39 bm thick. Cells of Dogiel, which were embedded within the inner nuclear layer, were also labeled when the injections of HRP spread to include the area of representation of the optic disc. The labeled cells were restricted to a dorsal, peripapillary locus capping the optic disc. Therefore, some cells of Dogiel project to the tectum normally, but only from the central retina. Some possible mechanisms behind the abnormal postoperative displacement of the ganglion cells are discussed. Key words: cell movement, cell survival,Runu pipiens, visual pathways, tectum opticum
Since the movements of the growth cone during the regeneration of a transected axon may be thought of as the renewal of cell motility mechanisms by the axon terminus, the potential for renewed motility by other parts of the neuron, or by the neuron as a whole, may be important in o 1990 WILEY-LJSS, INC.
the study of axon regeneration. There is evidence of dendrite sprouting in mature neurons undergoing axon regeneration (Sumner and Watson, '71; Purves, '75; Yawo, '87). Accepted July 18,1990.
GANGLION CELL DISPLACEMENT The present paper documents an instance of cell body translocation that may signify renewal of motility associated with the perikaryon. We reported previously that many ganglion cells disappear from the retina after periods of optic nerve regeneration in the frog, Rana pipiens, and that other ganglion cells become displaced abnormally into the inner plexiform layer (Scalia et al., '85). While other laboratories have confirmed the loss of ganglion cells during optic nerve regeneration in the frog (Humphrey and Beazley, '85; Stelzner and Strauss, '86; Beazley et al., '86; Humphrey, '87, '88; Stelzner and Strauss, '88; Sheard and Beazley, '88; Humphrey et al., '89; Dunlop et al., '89), none have included in their reports any observations on nor mention of the ganglion cell displacement. Since our original observations were made on flatmounted retinas, in which estimates of depth may involve uncertainties, we have now reexamined the phenomenon of cell dislocation in sectioned material. In sections taken perpendicular to the plane of the retina, it is easier to determine whether the inner plexiform and orthotopic ganglion cell layers remain uniform after nerve section, and whether any ganglion cells are normally present in the inner plexiform layer of the frog. With such observations as background, any abnormal displacement of ganglion cells taking place during optic nerve regeneration would be readily demonstrable. The cells of Dogiel, which exist in most if not all vertebrate species (see Ramon y Cajal, '721, are ganglion cells normally displaced from the retinal ganglion cell layer. In the frog, their cell bodies lie among the amacrine cells in the inner nuclear layer and their dendrites extend, in an inverted sense, vitread into the inner plexiform layer. The cells of Dogiel represent only 1 4 % of the normal ganglion cell population in Rana pipiens, and are most numerous along the central (suprapapillary)section of the horizontal meridian of the retina (Frank and Hollyfield, '87). At least some of the cells of Dogiel project their axon to the nucleus of the basal (accessory) optic root (Montgomery et al., '€411, but their possible contribution to other parts of the frog's primary visual pathway has not been studied. In the course of our studies on optic nerve regeneration in Rana pipiens, we have made additional observations on the central projection of both the normal cells of Dogiel and the abnormally displaced ganglion cells. Our results show that only those cells of Dogiel clustered around the optic disc project to the tectum, and that the abnormally displaced ganglion cells, which appear in all parts of the retina, are part of the overall postoperative population of tectally projecting cells. Partial results of this study were reported earlier (Singman and Scalia, '89). This paper is the first of a series of communications on the organization of the retinotectal system in the normal frog and in frogs sustaining optic nerve regeneration (Singman and Scalia, '90a,b).
Abbreviations
D GCL INL IPL N OD T
dorsal quadrant of retina ganglion cell layer inner nuclear layer inner plexiform layer nasal quadrant of retina optic disc temporal quadrant of retina
81
Fig. 1. Photomicrographs of sections of nasal retina taken from the right (normal)eye of frog R137, showing the normal paucity of ganglion cells in the inner plexiform layer (IPL), and the arrangement of orthotopic ganglion cells in a monolayer. The ganglion cells presented in the micrographs are back-filled with HRP applied to the optic nerve. The sections are lightly counterstained with cresyl violet. A typical cell of Dogiel is labeled by the HRP (arrow) in A (bar = 40 pm). The soma is located in the inner nuclear layer (INL) and a dendrite can be seen extendingvitread into the IPL. The same cell is labeled (arrow)in B in a lower power view (bar = 150 Fm). No other cell of Dogiel is apparent in this section, nor in the longer section shown in C (magnification as in
B).
MATI3RIALSANDMETHODS All the frogs studied were adult Rana pipiens (Northern variety, 7-9 cm body length). To mark retinal ganglion cells that project to the optic tectum, horseradish peroxidase (HRP) was injected into the tectal plate in normal frogs (N = 3) and in frogs that were undergoing or had completed unilateral optic nerve regeneration ( N = 9). Optic nerve regeneration had been initiated earlier by crushing the left optic nerve intraorbitally with fine forceps while under MS-222 anaesthesia. The tectum was exposed widely by removing the overlying bone and meninges under MS-222 anaesthesia. With the aid of a micromanipulator to control placement, the injections were made by inserting a pellet of pure HRP (300-500 pm diameter) dried onto the end of a glass micropipette (50 pm tip diameter) into the tectum and holding it there until the pellet dissolved. Often, a second and third HRP pellet were introduced in the same way at
82
E.L. SINGMAN AND F. SCALIA
Fig. 2. Photomicrographs of sections of nasal retina taken from the left (affected) eye of the same frog (R137)used for Figure 1, showing the abnormal displacement of ganglion cells (arrows) into the IPL. Ganglion cells are back-filled with HRP applied to the optic nerve. The orthotopic ganglion cells of the GCL continue to form a monolayer after
optic nerve regeneration, and the IPL, although reduced, remains of uniform thickness. The sections on the left were counterstained with cresyl violet to show the layers of the retina. Counterstaining was omitted on the right to emphasize the HRP-label. Bar = 45 pm.
the same site t o increase the degree of endocytotic uptake. A plate of dental acrylic was then placed over the cranial defect, and the skin was sutured closed. The injections were made bilaterally in mirror-symmetric locations. In most
cases, the injection was aimed for the region of tectum to which the middle part of the nasal retina projects. In a smaller number of cases, we injected the tectal hemisphere either anterior, lateral or medial to the site of representa-
GANGLION CELL DISPLACEMENT
Fig. 3. Photomicrographs of sections taken from the left (affected) retina of frog R141,showing that the abnormally displaced ganglion cells (arrows) project to the optic tectum. This frog survived 17 weeks following nerve crush. The sections were lightly stained with cresyl violet. Bar = 25 +m.
83
E.L. SINGMAN AND F. SCALIA
84
TABLE 2. Proportion of SurvivingNeurons That Undergo Displacement
TABLE 1. Distance of Ganglion Cell Displacement During Optic Nerve Regeneration' Mean values (N = 219) ~~
Displaced Cells per 10,000pm2
Percentage of total neurons
2.2 ? 0.37 1.9 t 0.38 2.6 t 0.49 1.7 t 0.44 1.7 ? 0.28 2.3 ? 0.46 2.5 ? 0.47 1.5 i 0.37 2.2 t 0.50 Mean = 2.1 f 0 25
5.6 -t 1.1 9.3 i 1.1 6.8 i 1.3 3.7 i 0.92 4.8 2 0.90 4.2 t 0.93 5.3 t 0.99 3.5 t 1.1 6.5 i 1.5 Mean = 5.5 t 1.2
~
14.3 5 0.39 Fm 5.8 5 0.20 pm 7.1 ? 0.41 Km (S.D. = 3.1) 10.0 2 0.45 km
Width of IPL (A) Cell diameter (Bf Distance from cell center to outer border of GCL tC! Assumed cell displacement (D) Percent oftheoretical maxunum I s placement (% !
69.9 i 2.4%
'The diameters and positions within the IPL of 219 abnormally displaced &an cellslocated in the nasal quadrant of the affeded (left) retina of the frog used for Figure 2 (R137) were measured. Each parameter presented in this table is descrild s c h e m a t i d y in Figure 4. For all mean values tabulated, the accompanyingconfidence intervals are taken at the 9 5 6 level (conf. int. = 1 9 6 x SE).
R141 R145 R120 R121 R138 R147 R148 R152 R151
17 18 35 39 52 75 79 94 101
< . . ... . . . .
.... .. .'
..., ...
C
MAX.
I PL
0
Fig. 4. The parameters measured for Table 1are shown diagrammatically. Measurements included the width of the IPL (A) at the location of an abnormally displaced ganglion cell (stippled), the diameter (B) of the abnormally displaced cell, taken perpendicular to the plane of the GCL, and the perpendicular distance ( C ) from the center of the abnormally displaced cell to the outer border of the GCL. One displaced cell (at right) is shown flattened against the inner surface of the IPL.
Assuming that the abnormally displaced cells were originally located in the GCL, then the degree of displacement (D)for any cell would be C + Bi2. Since none of the abnormally displaced ganglion cells has been observed to cross the inner border of the IPL, then the maximum displacement (MAX.) is taken as A. The darkened cell within the INL represents a typical cell of Dogiel.
tion of the optic disc. The retinas of both eyes were then studied after periods of retrograde transport of 3-5 days. To study retinal ganglion cell distributions without regard for their sites of projection, HRP was delivered to the optic nerve instead of to the tectum in some cases (N = 8). This was done by placing a pellet of HRP against the cut
surface of the optic nerve transected intraorbitally (Scalia and Colman, '74). Retinal ganglion cells projecting anywhere in the brain become heavily back-filled by this method. At the time of sacrifice, the frogs were first dark-adapted, then they were anesthetized with MS-222 and placed on ice.
GANGLION CELL DISPLACEMENT
85
T39 Left Retina
.J \
)
R138 Normol (Right) Retina
.h
. . . D. . .. . . . .. .. .. ........... . . . ... .. ..... .0 . OD
T54 Right Retina
\
n
R138 Affected (Left) Retina
. . ....- . . . . . .
'i
Fig. 5. The topographic distribution of the 100 km x 100 km sampling frames (filled squares) is depicted on tracings of the retinas. To enhance their visibility, the squares are drawn larger than scale. Upper left: Left retina of normal frog T39. Three days prior to sacrifice, this frog received a large injection of HRP into a caudomedial locus of its right tectal hemisphere (see Fig. 7). Lower left: Right retina of normal frog T54. Five days prior to sacrifice, this frog received
.. .. .. .. .. . .. . . . . . . . . .. . .. . .. . ' .OD. I . . . . . . . . . . . . . . . .. . .. . .. . .. . .
- . .
a large injection of HRP into a rostrolateral locus of its left tectal hemisphere (see Fig. 8). Upper right: Normal (right) retina of regenerate frog R138. Five days prior to sacrifice, large injections of HRP were made into bilaterally symmetric, caudolateral loci of the frog's tectal hemispheres (see Fig. 9). Lower right: Left (affected) retina of frog R138 (see Fig. 10).
E.L. SINGMAN AND F. SCALIA
86 In semidarkness, the frogs were exsanguinated by perfusion through the conus arteriosus with frog Ringer's (pH 7.3). The eyes were immediately removed, and the carcass was perfused with a fixative (4% glutaraldehyde in 0.15 M phosphate buffer, pH 7.3) to save the brain. After marking the dorsal, ventral, and temporal poles of the eyes with small puncture-holes, the retinas were dissected out into frog Ringer's solution containing 4% dextrose. The pigment cell layer (and any attached choroid coat) was removed with a forceps, andor with gentle brushing, and the retina was brought up onto a glass slide. Relief cuts were made to allow flattening of the retina, and the flattened specimen was fixed in ascending steps of glutaraldehyde concentrations, while under light pressure from a glass coverslip. After fixation in 4% glutaraldehyde for 2 hours, the retinas were washed in phosphate buffer overnight and incubated for HRP activity by the nickel-cobalt intensified diaminobenzidine system of Adams ('81). Subsequently, the retinas were mounted on glass slides for LM analysis as flat-mounts, or the retinas were sectioned frozen at 12-25 km perpendicular to the plane of flattening, and the sections were mounted on slides in serial order. When counterstaining was required, cresyl violet was used. In some cases, flat-mounted retinas that had been previously examined were removed from their slides, sectioned as above, and reexamined in-section. Ganglion cells were examined in the flat-mounted retinas for the presence or absence of HRP, and the numbers of HRP-labeled and unlabeled cells counted were recorded with the assistance of a microscope-pantograph system (Microcode, Boeckeler Instruments) and a video-image processing system (Image-Pro, Media Cybernetics, Inc.) connected to an IBM personal computer. The population densities of the labeled and unlabeled cells were analyzed by plotting these data onto tracings of the retinas with a computer-based graphics system (SURFER, Golden Software, Inc.). Two sampling methods were used: 1)Cells were counted at 1 , 0 0 0 ~in 10,000 pm2 (100 pm x 100 pm) windows systematically distributed over the retinal areas corresponding to the tectal injection sites. Each of these distributed counting frames was carefully examined through depth to detect possible cell-labeling in the inner plexiform or inner nuclear layer. 2 ) Cells of Dogiel projecting to the tectum were counted by scanning the retina at 1 , 0 0 0 ~in adjacent vertically oriented rasters over two 2-mm-wide strips centered on the optic disc and on the area of maximal retrograde label in the appropriate retinal sector. The rasters containing the optic disc extended the full vertical diameter of the retina. Since all the labeled cells of Dogiel found near the optic disc were located well within a 2 mm square, the search for cells of Dogiel in the area of maximal labeling was limited to a square of the same size.
RESULTS Degree and frequency of cell displacement In all the normal retinas examined as flat-mounts, in which ganglion cells had been back-filled with HRP applied to the optic nerve, the labeled ganglion cells were confined to a uniform monolayer (the retinal ganglion cell layer), except for the small populations of cells of Dogiel in the
Fig. 6. A photomicrograph focused upon the vitreal surface of the INL of the wholemounted left retina from frog T39,showing six cells of Dogiel (at a-0 labeled by an injection of HRP to the right tectal hemisphere. The cell at f could not be focused with the others. The clarity of the cells is generally not optimal because their focal plane is deep within the thick whole-mounted retina.
inner nuclear layer, and the rare occurrences' of cell placements within the inner plexiform layer. This can be verified in the sectioned material (Fig. 11, and is consistent with earlier descriptions (Jacobson, '62; Ramon y Cajal, '72; Frank and Hollyfield, '87). Cells of Dogiel were observed in the retina flat-mounts from the normal frogs and from either eye of the frogs sustaining unilateral optic nerve regeneration. Their cell bodies were located within the inner surface of the inner nuclear layer, and their dendrites formed an inverted tree, directed vitreally. Characteristically, only the dendrites of the cells of Dogiel extended into the inner plexiform layer; their cell bodies were always submerged fully within the confines of the inner nuclear layer (Fig. 1). Their location in the inner nuclear layer serves to distinguish the cells of Dogiel of the frog from the abnormally displaced ganglion cells observed in this species after optic nerve regeneration. In contrast, the abnormal displacement of ganglion cells into the inner plexiform layer was observed only in the retinas that were undergoing or had completed optic nerve regeneration. When observed in the sectioned material (Figs. 2, 31, the dendrites of the abnormally displaced cells were usually tangentially spread, or else they were directed away from the vitreous body, in the manner of the orthotopic ganglion cells (cells residing in the ganglion cell layer proper), but unlike the cells of Dogiel, whose dendrites extend toward the vitreous body. The degree of displacement varied widely across the thickness of the inner plexiform layer. Any abnormally displaced cells present at 'Over a large sample of neurons counted in the retinal ganglion cell- and inner plexiform-layers in normal frog T39, without regard to the presence or absence of the HRP-label,only 17 out of 5367 neurons (0.32%)were located in the inner plexiform layer.
87
GANGLION CELL DISPLACEMENT the border of the inner nuclear layer were usually flattened against its inner surface. Sometimes the flattening was very marked, suggesting the presence of a planar barrier against displacement across the interface between the inner plexiform and inner nuclear layers. Since the inner plexiform layer remains of uniform thickness and the orthotopic ganglion cells continue to form a monolayer after nerve section (Figs. 2, 3), it was possible to measure the outward displacement of the ectopic ganglion cells. Table 1 (and see Fig. 4)gives measurements of the cell displacement in one frog (R137). This frog was making prey-catching responses using its affected (left) eye when sacrificed at 8 weeks after nerve crush. Three days prior to sacrifice, both of its optic nerves were transected and HRP was instilled into the peripheral stump to label its retinal ganglion cells (see Materials and Methods). The mean width of the inner plexiform layer in the nasal quadrant of the affected retina was 14.3 _" 0.39 pm, and the mean distance from the center of the abnormally displaced cells to the scleral edge of the ganglion cell layer was 7.1 r 0.41 pm. Assuming that the abnormally displaced ganglion cells had migrated outward from the ganglion cell layer, the mean cell displacement would be 10.0 5 0.45 pm. Taking the maximum theoretical displacement as the distance to the inner border of the inner nuclear layer, then the mean traverse of the abnormally displaced cells would be 69.9 ? 2.4%of the maximum theoretical displacement. This takes many of the translocated cells more than half-way across the inner plexiform layer at a survival period of only 8 weeks. In the corresponding quadrant of the normal retina of this frog, the mean width of the inner plexiform layer was 18.7 ? 0.36 Km (N = 345 measurements taken along 18 sections at 100 Km intervals), indicating that the thickness of the inner plexiform layer had been reduced by approximately 24% in this specimen after optic nerve section. The increase in the number of neurons that can be observed in the inner plexiform layer after initiating optic nerve regeneration is shown in Table 2. The cell-counts for this table were taken from flat-mounted retinas (e.g., Fig. 5) of frogs that had received a tectal injection of HRP. On average, the abnormally displaced cells represented 5.5%of the number of surviving neurons in the combined ganglion cell-inner plexiform layer in the areas counted. This percentage was stable over nine frogs surviving widely varying periods since nerve section.
Contrastingtectal projections of the abnormallydisr>laced sandioncells and the Tectal injections of HRP labeled many ganglion cells in the normal retinas and in the retinas sustaining optic nerve regeneration. The HRP-labeled orthopic ganglion cells and the HRP-labeled abnormally displaced ganglion cells (Fig. 3) were localized in regions of the retina corresponding to the sites of tectal injection.* The population densities of labeled cells were greatest in the areas representing the centers of the injection sites. Since the injections subtended large areas of the tectum, the labeling frequencies in the
'Several laboratories (Scalia et al., '85; Stelzner and Strauss, '86; Beazley et al., '86; Frank and Hollyfield, '87) have estimated with different methods that 12-16% of the neurons in the ganglion cell layer of R. pipiens are displaced amacrine cells. We therefore adjusted our counts of total neurons downward by our own previous estimate of the displaced amacrine population (16%)to obtain the number of ganglion cells in the ganglion cell layer.
retina attained a maximum over a fairly level plateau (broken by minor peaks and depressions) that extended over a significant portion of the retina. In contrast, cells of Dogiel (Fig. 6) were labeled from the tectum only in those specimens in which the plateau of maximal labeling extended near the optic disc. Then, the labeled cells of Dogiel (i.e., ganglion cells in the inner nuclear layer3)were concentrated in a horseshoe-shaped area capping the optic disc on its nasal, dorsal, and temporal sides, independent of the location of the center of the injection site (Figs. 7-10). Differential cell counts are given in Table 3. Few or none of the cells of Dogiel were labeled in the centers of the areas of maximal labeling. In two animals (T55 and R120) in which tectal injections did not involve the area of representation of the optic disc, no cells of Dogiel were labeled.
DISCUSSION Cell displacement The present study confirms and extends our earlier report that large numbers of ganglion cells may be found in the inner plexiform layer in the retina of R a m pipiens after a period of optic nerve regeneration (Scalia et al., '85), whereas very few neurons are normally found in the inner plexiform layer of frogs (Frank and Hollyfield, '87; Straznicky and Straznicky, '88). Only a few specimens were available at the time of our previous report on ganglion cell displacement, and we did not know that the large numbers of ectopic cells observed at 8-12 months survival appear during the stages of optic nerve regeneration. We have presently counted the numbers of abnormally displaced cells for the first time and studied their degree of displacement in sectioned material. Any question that the appearance of cell displacement may be an illusory consequence of irregularity in ganglion cell alignment combined with shrinkage of the inner plexiform layer after cell loss can be dispelled by reference to Figures 2 and 3. The measurements given in Table 1 indicate that although most of the abnormally displaced cells come to lie roughly halfway across the inner plexiform layer, there is a large degree of variability. Some cells were found fully displaced and flattened against the inner face of the inner nuclear layer. Therefore, we are convinced of the reality of the cell displacement, and urge consideration of its mechanism and functional significance. It was originally suggested that the abnormally displaced cells may be redifferentiated amacrine cells (Scalia et al., '85). Although we have no further information on that particular question, some additional alternatives can be
Yt should be noted that the cellular layers of the retina in the frog remain parallel to each other as they approach the optic disc, and the optic nerve head penetrates the retina without disturbing the regularity of the layers. The cells of Dogiel counted in the peripapillary region are clearly localized within the inner border of the inner nuclear layer.
Figs. 7-10 (see pages 88-91). A Expanded view of the optic disc, showing the topographic locations of tectally projecting cells of Dogiel (one dot = one cell). B: Schematic drawing of the corresponding teetum, showing the HRP-injection site (filled region). C: Tracing of the corresponding retina, showing the region of maximal HRP labeling as well as the field of labeled cells of Dogiel. The orientation of the optic disc is the same in both A and C. Rectangular boxes outline the areas subtended by the distributed sampling frames.
88
E.L. SINGMAN AND F. SCALIA
T39 Left Retina
A
C Fig. 7. The left retina of normal frog T39.
considered here. One suggestion is that the displaced cells might be migrated cells of Dogiel. This seems unlikely because there are more abnormally displaced cells than cells of Dogiel, which account for only 1 4 % of the ganglion cell population, and because the cells of Dogiel suffer rates of cell death comparable to that of the population of ganglion cells in the ganglion cell layer (Dunlop et al., '89). Another possibility is that there may be ganglion cell mitogenesis from a precursor population of cells capable of mitotic activity and neuronal differentiation. This is made unlikely by the observations of Bohn and Reier ('82) and Pedalina and Beazley ('861, who used tritiated-thymidine autoradiography to show that there is no resurgence of mitotic activity after optic nerve section in the retinas of adult Xenopus laevis or Hyla moorei, respectively. Perhaps the simplest explanation is that the ganglion cells of the ganglion cell layer are the source of the displaced neurons. Two mechanisms for the displacement of ganglion cells from the ganglion cell layer may be suggested. One is that ganglion cells retract their dendrites, losing all or part of their synaptic attachments and become migratory for a period during which they journey across the inner plexiform layer. Subsequently, they would reestablish their dendrites and presynaptic connections. This mechanism may be examined by further morphological study of retinal ganglion cells during the early stages of optic nerve regeneration. While there is no precedent for the resumption of cellular motility by adult neurons, such an event would not be inconsistent with the widely held idea that mature
neurons capable of axon regeneration undergo a partial reversion to earlier developmental states after axotomy. Admittedly, migration away from the vitreous is the reverse of the direction ganglion cells migrate during development. The hypothesized loss of presynaptic input, which would seem to be a necessary adjunct to cell migration, is not inconsistent with earlier observations on the loss of at least some part of the synaptic scale from motorneurons (e.g., Sumner and Sutherland, '73) and other central neurons (Chen et al., '77) surviving axotomy. Loss and reacquisition of synaptic input by retinal ganglion cells during optic nerve regeneration would suggest a remarkable capacity for reconstruction of functionally useful synaptic microcircuitry, since it is reported that all electrophysiologically defined classes of retinal ganglion cells reappear after nerve regeneration (Maturana et al., '59; Keating and Gaze, '70). The other mechanism we can suggest to explain the ganglion cell displacement involves an intracytoplasmic nuclear migration. Specifically, it is possible that the nucleus of a certain proportion of the ganglion cells undergoing axon regeneration may move into a primary dendrite and migrate some distance within the dendrite shaft towards the inner nuclear layer. During this migration, perinuclear organelles might be transported along with the nucleus, resulting in the outward (sclerad) transposition of the perikaryon, and the transformation of the trailing portion of the cell into part of the axon. The distance over which the nucleus is capable of moving may be constrained by the branching pattern of the dendrite system. The
GANGLION CELL DISPLACEMENT
89
A
T54 Right Retina
C Fig. 8.
The right retina of normal frog T54.
nucleus may move only a short distance in cells whose major dendrites branch low (vitreally) in the inner plexiform layer, but may move entirely across in those cells whose primary dendrite bifurcates only after reaching nearly to the inner nuclear layer. The nuclear migration suggested here is reminiscent of the nuclear translocations that take place in the developing neural tube during mitosis (Sauer, '351, specifically, the proluminal phase of nuclear migration immediately preceding cell division. In the adult retina, movement from within the ganglion cell layer toward the photoreceptor layer is equivalent to movement toward the ventricular lumen of the developing eye. This explanation is also consistent with the concept that neurons undergoing axon regeneration experience a partial reversion to a developmentally juvenile biology. In this case, certain processes normally leading to cell-division may be abortively reinstated after axon section in some neurons.
Cells ofDogiel We had previously reported that cells of Dogiel do not project to the optic tectum in the frog (Singman and Scalia, '89); we now present evidence to the contrary. We had originally overlooked the tectally projecting population of cells of Dogiel because they are not distributed evenly throughout the retina. The population of tectally projecting cells of Dogiel is restricted to a horseshoe-shaped locus surrounding the optic disc nasally, temporally, and dorsally. Their axons apparently terminate near the tectal
representation of the optic disc, because they are labeled only when tectal HRP injections spread to involve the peripapillary representation area. Their presence as a dorsal cap around the optic disc and their absence on its ventral side recalls the early topography of the developing retina, in which the area at the dorsal end of the optic disc is the first part of the retina to develop (Grant et al., '80). This suggests that these cells may be among the developmentally oldest ganglion cells in the frog retina. The projections of cells of Dogiel in the frog are not completely understood. Frank and Hollyfield ('87) estimated that cells of Dogiel comprise approximately 5% of the overall population of ganglion cells in a horizontally oriented streak situated just dorsal to the optic disc. Elsewhere in the retina, the cells of Dogiel represented less than 1% of the overall population of ganglion cells. Montgomery et al. ('81) showed that some cells of Dogiel project to the nucleus of the basal (accessory) optic root (nBOR) in the frog, although these cells comprised the smallest portion of the ganglion cells projecting to this structure. The cells of Dogiel projecting to nBOR were widely distributed throughout the retina, although their frequency in the peripapillary region was not specifically noted. It would be interesting to know whether the tectally projecting cells of Dogiel and those projecting to the nBOR occupy separate spatial domains. Cells of Dogiel are found in most if not all vertebrate species. They generally comprise less than 5% of the overall
E.L. SINGMAN AND F. SCALIA
90
R138 Normal (Right) Retina
A .-
OD
L
TECTUM
C Fig. 9. The normal (right) retina of regenerate frog R138. Only the injection site on the left tectal hemisphere is shown.
population of retinal ganglion cells, although the proportion is as great as 15%in the mudpuppy (Arkin and Miller, '88) and 43%in the larval sea lamprey (Rubinson and Cain, '89). The relative distributions and projections of cells of Dogiel vary widely between species. Cells of Dogiel are most heavily concentrated in the retinal periphery in the newt (Ball and Dickson, '831,the ray (Collin, '88), the larval sea lamprey (Rubinson and Cain, '89), the chick (Prada et al., '89), the mouse (Drager and Olsen, '81), the rat (Buhl and Dann, '88; Linden, '87), the rabbit (Robson and Hollander, '84), and the wallaby (Coleman et al., '87). In contrast, most cells of Dogiel reside in the central retina in the pigeon (Hayes and Holden, '831,and in the frogs Rana pipiens (Frank and Hollyfield, '87) and Hyla moorei (Dunlop et al., '89). In the mudpuppy (Arkin and Miller, '88) and the frog Xenopus laevis (Straznicky and Straznicky, '881, the cells of Dogiel are distributed fairly evenly across the retina. It has been shown that cells of Dogiel comprise the only cells projecting to nBOR of the pigeon and that they apparently do not project to the tectum in this species (Fite et al., '81; Hayes and Holden, '83).In contrast, it has been reported that the tectally projecting cells of Dogiel in the chicken are restricted to the temporal hemi-retina, which contains the optic disc in this species, and are most concentrated near the optic disc itself (Heaton et al., '79).
Although the cells of Dogiel do not comprise the exclusive projection to nBOR in the turtle, many do make this projection, but, as in the pigeon, they apparently do not project to the optic tectum (Reiner, '80). In mammals such as the rabbit (Oyster et al., '80), the rat (Buhl and Dann, '88), and the cat (Farmer and Rodieck, '821, however, the cells of Dogiel do project to the superior colliculus, and also to the dorsal lateral geniculate nucleus. The cells of Dogiel projecting to the monkey's dorsal lateral geniculate nucleus are restricted to the peripapillary region (Bunt and Minckler, '77). These projections appear to be more prominent than the projections to the medial terminal nucleus (MTN), the homologue of nBOR (for review see Simpson, '84). Specifically, it has been reported that no cells of Dogiel project to the MTN in the rabbit and cat.
ACKNOWJXDGMENTS We thank Ms. Suri Roca, Ms. Jan Prokosch, and Ms. Gloria Scott for their technical assistance. We also thank Mr . Vincent Garofalo for his expert photographic assistance. This study was supported by a USPHS Research Grant (EY 05284) from the National Eye Institute and was submitted in partial fulfillment of requirements for the degree of Doctor of Philosophy.
GANGLION CELL DISPLACEMENT
91
R138 Affected (Left) Retina
A
OD
0
1 rnrn
\
C Fig. 10. T h e left (affected) retina of frog R138.Only t h e injection site on t h e right hemisphere is shown.
TABLE 3. Location of Cells of Dogie1That Project to the Tectuml
Frog T39 T54 R138
R138" T554 ~ 1 2 0 ~
No. of distrihuted samples 52 30 73 116 35 36
Overall area
sampled2 7.9 mm2 15 mm2 Entire retina Entire retina 17 mm2 3.6 mm2
Adams, J.C. (1981) Heavy metal intensification of DAB-based HRP reaction product. J. Histochem. Cytochem. 29:775. Arkn, M.S., and R.F. Miller (1988)Mudpuppy retinal ganglion cell morphology revealed by an HRP impregnation technique which provides Golgilike staining. J. Comp. Neurol. 270t185-208. Ball, A.K., and D.H. Dickson (1983) Displaced amacrine and ganglion cells in the newt retina. Exp. Eye Res. 36t199-213. Beazley, L.D., J.E. Darby, and V.H. Perry (1986) Cell death in the retinal ganglion cell layer during optic nerve regeneration for the frog Rana pipiens. Vision Res. 26:543-556. Bohn, R.C., and P.J. Reier (1982) Anomalous axonal outgrowth at the retina
Retinal sector containing maximal label Nasal Temporal Dorsotempral Dorsotemporal Ventronasal
Nasal
RGC
cDt
cDu
4,508 2,914 4,890 2,683 1,572 2,549
2 2 0
232 93 100
0 0 0
143
0 0
caused by injury to the optic nerve or tectal ablation in adult Xenopus. J. Neurocytol. 1It2 11-234. Buhl, E.H., and J.F. Dann (1988) Morphological diversity of displaced retinal ganglion cells in the rat: a lucifer yellow study. J. Camp. Neurol. 269t210-218. Bunt, A.H., and D.S. Minckler (1977) Displaced ganglion cells in the retina of the monkey. Invest. Ophthalmol. Vis. Sci. 16t95-98. Chen, D., W.W. Chambers, and C.N. Liu (1977) Synaptic displacement in intracentral neurons of Clarke's nucleus following axotorny in the cat. Exp. Neurol. 57t1026-1041. Coleman, L.A., A.M. Harman, and L.D. Beazley (1987) Displaced retinal ganglion cells in the wallaby Setoncx bruchyurus. Vision Res. 27:12691277.
92 Collin, S.P. (1988) The retina of the shovel-nosed ray, Rhinobatos batillurn (Rhinobatidae): morphology and quantitative analysis of the ganglion, amacrine and bipolar cell populations. Exp. Biol. 47:195-207. Drager, U.C., and J.F. Olsen (1981)Ganglion cell distribution in the retinaof the mouse. Invest. Ophthalmol. Vis. Sci. 20285-293. Dunlop, S.A., M.F. Humphrey, and L.D. Beazley (1989) Survival of displaced retinal ganglion cells after optic nerve regeneration in the frog Hyla rnoorei. SOC. Neurosci. Abstracts 15:872. Farmer, S.G., and R.W. Rodieck 11982) Ganglion cells in the cat accessory optic system: Morphology and retinal topography. J. Comp. Neurol. 205: 190-198. Fite, K.V., N. Brecha, H.J. Karten, and S.P. Hunt (1981) Displaced ganglion cells and the accessory optic system of the pigeon. J. Comp. Neurol. 1955279-288. Frank, B.D., and J.G. Hollyfield (1987) Retinal ganglion cell morphology in the frog, Ranapipzens. J. Comp. Neurol. 266,413-434. Grant, P., E. Rubin, and C. Cima (1980) Ontogeny of the retina and optic nerve in Xenopus laeuis. I. Stages in the early development of the retina. J. Comp. Neurol. 189~593-613. Hayes, B.P., and A.L. Holden (1983) The distribution of displaced ganglion cells in the retina of the pigeon. Exp. Brain Res. 49:181-188. Heaton, M.B., I.M. Alvarez, and J.E. Crandall (1979) The displaced ganglion cell in the avian retina: developmental and comparative considerations. Anat. Embryol. (Berl.) 155:161-178. Humphrey, M.F. (1987) Effect of different optic nerve lesions on retinal ganglion cell death in the frog Ranapipiens. J. Comp. Neurol. 266:209219. Humphrey, M.F. (1988) A morphometric study of the retinal ganglion cell response to optic nerve severance in the frog Rana pzpiens. J. Neurocytol. 17293-304. Humphrey, M.F., and L.D. Beazley (1985) Retinal ganglion cell death during optic nerve regeneration in the frog Hyla moorei. J. Comp. Neurol. 236:382402. Humphrey, M.F., J.E. Darby, and L.D. Beazley (1989) Prevention of optic nerve regeneration in the frog Hyla rnoorei transiently delays the death of some ganglion cells. J. Comp. Neurol. 279:187-198. Jacobson, M. (1962) The representation of the retina on the optic tectum of the frog. Correlation between retino-tectal magnification factor and retinal ganglion cell count. Q. J. Exp. Physiol. 47:170-178. Keating, M.J., and R.M. Gaze (1970) The depth distribution ofvisual units in the contralateral optic tectum following regeneration of the optic nerve in the frog. Brain Res. 21:197-206. Linden, R. 11987) Displaced ganglion cells in the retina of the rat. J. Comp. Neurol. 258:138-143. Maturana, H.R., J.Y. Lettvin, W.S. McCulloch, and W.H. Pitts (1959) Physiological evidence that cut optic nerve fibers in the frog regenerate to their proper places in the tectum. Science 130:1709, 1710. Montgomery, N., K.V. Fite, and L. Bengston (1981) The accessory optic system of Rana pzpiens: Neuroanatomical connections and intrinsic organization. J. Comp. Neurol. 203:595-612. Oyster, C.W., J.I. Simpson, E.S. Takahashi, and R.E. Soodak (1980) Retinal ganglion cells projecting to the rabbit accessory optic system. J. Comp. Neurol. 190~49-61. Pedalina, R., and L.D. Beazley 11986) A tritiated thymidine study of cell division in the retina during optic nerve regeneration for the frog Hyla rnoorei. Neurosci. Lett. 1Suppl.l23:S70.
E.L. SINGMAN AND F. SCALIA Prada, F.A., C.E. Chmielewski, M.E. Dorado, C. Prada, and J.M. GenisGalvez (1989) Displaced ganglion cells in the chick retina. Neurosci. Res. 6:329-339. Purves, D. (1975) Functional and structural changes in mammalian sympathetic neurones following interruption of their axons. J. Physiol. (Lond.) 252t429-463. Ramon y Cajal, S. (1972) The Structure of the Retina. Springfield, 11,: Charles C. Thomas. Reiner, A. (1980) A projection of displaced ganglion cells and giant ganglion cells to the accessory optic nuclei in turtle. Brain Res. 204:403409. Robson, J.A., and H. Hollander (1984) Displaced ganglion cells in the rabbit retina. Invest. Ophthalmol. Vis. Sci. 25~1376-1381. Rubinson, K., and H. Cain (1989) Neural differentiation in the retina of the larval sea lamprey (Petromyzon marinus). Visual Neurosci. 3t241-248. Sauer, F.C. 11935) Mitosis in the neural tube. J. Comp. Neurol. 62377-405. Scalia, F., V. Arango, and E.L. Singman (1985) Loss and displacement of ganglion cells after optic nerve regeneration in adult Ranapzpiens. Brain Res. 344.267-280. Scalia, F., and D.R. Colrnan (1974) Aspects of the central projections of the optic nerve in the frog as revealed by anterograde migration of horseradish peroxidase. Brain Res. 79t496-504. Sheard, P.W., and L.D. Beazley (1988) Retinal ganglion cell death is not prevented by application of tetrodotoxin during optic nerve regeneration in the frog Hyla rnoorei. Vision Res. 28:461-470. Simpson, J.I. (1984) The accessory optic system. Annu. Rev. Neurosci. 7: 13-41. Singman, E.L., and Scalia, F. (1989) Cells of Dogie1 and ganglion cell displacement during optic nerve regeneration. SOC.Neurosci. Abstr. 15:1208. Singman, E.L., and F. Scalia (1990a) A quantitative study of the tectally projecting retinal ganglion cells in the adult frog. I. The size of the contralateral and ipsilateral projections. (Submitted. Singman, E.L., and F. Scalia (1990b) A quantitative study of the tectally projecting retinal ganglion cells in the adult frog. 11. Cell survival and functional recovery after optic nerve transection. (Submitted.) Stelzner, D.J., and J.A. Strauss (1986) A quantitative analysis of frog optic nerve regeneration: Is retrograde ganglion cell death or collateral axonal loss related to selective reinnervation? J. Comp. Neurol. 245:83-106. Stelzner, D.J., and J.A. Strauss (1988) Increase in ganglion cell size after optic nerve regeneration in the frog, Runapipiens. Exp. Neurol. 100~210215. Straznicky, C., and I.T. Straznicky (1988) Morphological classification of retinal ganglion cells in adult Xenopus laeuzs. Anat. Embryol. (Berl.) 178:143-153. Surnner, B.E.H., and F.I. Sutherland (1973) Quantitative electron microscopy on the injured hypoglossal nucleus in the rat. J. Neurocytol. 2: 3 15-328. Sumner, B.E.H., and W.E. Watson (1971) Retraction and expansion of the dendritic tree of motor neurones of adult rats induced in uiuo. Nature 233973-275. Yawo, H. (1987) Changes in the dendritic geometry of mouse superior cervical ganglion cells following postganglionic axotomy. J. Neurosci. 7:3703-3711.
Further Study of the Outward Displacement of Retinal Ganglion Cells During Optic Nerve Regeneration,With a Note on the Normal Cells of Dogiel in the Adult Frog ERIC L. SINGMAN AND FRANK SCALIA Department of Anatomy and Cell Biology, State University of New York, Health Science Center, Brooklyn, New York 11203
ABSTRACT In a previous study we observed massive retinal ganglion cell death in adult Rana pipiens after periods of optic nerve regeneration, and reported that large numbers of the surviving cells had become displaced bodily into the inner plexiform layer of the affected eye (Scalia et al.: Brain Research 344:267-280,1985). The outwardly displaced cells could be identified as retinal ganglion cells because they could be back-filled with horseradish peroxidase (HRP) injected into the regenerated optic nerve. Quantitative observations on the abnormal outward displacement of ganglion cells are reported here. Parallel observations on normally displaced ganglion cells (cells of Dogiel) are also reported to clarify the distinctions between these two classes of cells. For the present work, injections of HRP of varying size were placed in the optic tectum bilaterally in 3 normal frogs and 9 frogs sustaining unilateral optic nerve regeneration. Most injections were centered at loci mapping the middle region of the nasal retina. The retinas were examined as flat-mounts and in-section. In 8 other frogs sustaining optic nerve regeneration, the HRP was administered bilaterally directly to the optic nerves in the orbit. Ganglion cells were labeled by retrograde transport of the HRP in the retinal ganglion cell layer in both the normal and affected eyes in areas topographically isomorphic with the tectal areas subtended by the injections. In the normal eyes, the orthotopic ganglion cells formed a strict monolayer, and virtually no cells existed in the inner plexiform layer. In the retinas sustaining optic nerve regeneration, the retinal ganglion cells abnormally displaced into the inner plexiform layer were also labeled topographically in correspondance with the injection sites. The abnormally displaced cells comprised 5.5% of the total population of surviving neurons in the retinal ganglion cell and inner plexiform layers. The mean outward dislocation of the displaced cells, as measured in one frog surviving optic nerve crush for 8 weeks, was 69.9 2 2.4% of the distance across the inner plexiform layer, which itself was uniformly 14.3 % 0.39 bm thick. Cells of Dogiel, which were embedded within the inner nuclear layer, were also labeled when the injections of HRP spread to include the area of representation of the optic disc. The labeled cells were restricted to a dorsal, peripapillary locus capping the optic disc. Therefore, some cells of Dogiel project to the tectum normally, but only from the central retina. Some possible mechanisms behind the abnormal postoperative displacement of the ganglion cells are discussed. Key words: cell movement, cell survival,Runu pipiens, visual pathways, tectum opticum
Since the movements of the growth cone during the regeneration of a transected axon may be thought of as the renewal of cell motility mechanisms by the axon terminus, the potential for renewed motility by other parts of the neuron, or by the neuron as a whole, may be important in o 1990 WILEY-LJSS, INC.
the study of axon regeneration. There is evidence of dendrite sprouting in mature neurons undergoing axon regeneration (Sumner and Watson, '71; Purves, '75; Yawo, '87). Accepted July 18,1990.
GANGLION CELL DISPLACEMENT The present paper documents an instance of cell body translocation that may signify renewal of motility associated with the perikaryon. We reported previously that many ganglion cells disappear from the retina after periods of optic nerve regeneration in the frog, Rana pipiens, and that other ganglion cells become displaced abnormally into the inner plexiform layer (Scalia et al., '85). While other laboratories have confirmed the loss of ganglion cells during optic nerve regeneration in the frog (Humphrey and Beazley, '85; Stelzner and Strauss, '86; Beazley et al., '86; Humphrey, '87, '88; Stelzner and Strauss, '88; Sheard and Beazley, '88; Humphrey et al., '89; Dunlop et al., '89), none have included in their reports any observations on nor mention of the ganglion cell displacement. Since our original observations were made on flatmounted retinas, in which estimates of depth may involve uncertainties, we have now reexamined the phenomenon of cell dislocation in sectioned material. In sections taken perpendicular to the plane of the retina, it is easier to determine whether the inner plexiform and orthotopic ganglion cell layers remain uniform after nerve section, and whether any ganglion cells are normally present in the inner plexiform layer of the frog. With such observations as background, any abnormal displacement of ganglion cells taking place during optic nerve regeneration would be readily demonstrable. The cells of Dogiel, which exist in most if not all vertebrate species (see Ramon y Cajal, '721, are ganglion cells normally displaced from the retinal ganglion cell layer. In the frog, their cell bodies lie among the amacrine cells in the inner nuclear layer and their dendrites extend, in an inverted sense, vitread into the inner plexiform layer. The cells of Dogiel represent only 1 4 % of the normal ganglion cell population in Rana pipiens, and are most numerous along the central (suprapapillary)section of the horizontal meridian of the retina (Frank and Hollyfield, '87). At least some of the cells of Dogiel project their axon to the nucleus of the basal (accessory) optic root (Montgomery et al., '€411, but their possible contribution to other parts of the frog's primary visual pathway has not been studied. In the course of our studies on optic nerve regeneration in Rana pipiens, we have made additional observations on the central projection of both the normal cells of Dogiel and the abnormally displaced ganglion cells. Our results show that only those cells of Dogiel clustered around the optic disc project to the tectum, and that the abnormally displaced ganglion cells, which appear in all parts of the retina, are part of the overall postoperative population of tectally projecting cells. Partial results of this study were reported earlier (Singman and Scalia, '89). This paper is the first of a series of communications on the organization of the retinotectal system in the normal frog and in frogs sustaining optic nerve regeneration (Singman and Scalia, '90a,b).
Abbreviations
D GCL INL IPL N OD T
dorsal quadrant of retina ganglion cell layer inner nuclear layer inner plexiform layer nasal quadrant of retina optic disc temporal quadrant of retina
81
Fig. 1. Photomicrographs of sections of nasal retina taken from the right (normal)eye of frog R137, showing the normal paucity of ganglion cells in the inner plexiform layer (IPL), and the arrangement of orthotopic ganglion cells in a monolayer. The ganglion cells presented in the micrographs are back-filled with HRP applied to the optic nerve. The sections are lightly counterstained with cresyl violet. A typical cell of Dogiel is labeled by the HRP (arrow) in A (bar = 40 pm). The soma is located in the inner nuclear layer (INL) and a dendrite can be seen extendingvitread into the IPL. The same cell is labeled (arrow)in B in a lower power view (bar = 150 Fm). No other cell of Dogiel is apparent in this section, nor in the longer section shown in C (magnification as in
B).
MATI3RIALSANDMETHODS All the frogs studied were adult Rana pipiens (Northern variety, 7-9 cm body length). To mark retinal ganglion cells that project to the optic tectum, horseradish peroxidase (HRP) was injected into the tectal plate in normal frogs (N = 3) and in frogs that were undergoing or had completed unilateral optic nerve regeneration ( N = 9). Optic nerve regeneration had been initiated earlier by crushing the left optic nerve intraorbitally with fine forceps while under MS-222 anaesthesia. The tectum was exposed widely by removing the overlying bone and meninges under MS-222 anaesthesia. With the aid of a micromanipulator to control placement, the injections were made by inserting a pellet of pure HRP (300-500 pm diameter) dried onto the end of a glass micropipette (50 pm tip diameter) into the tectum and holding it there until the pellet dissolved. Often, a second and third HRP pellet were introduced in the same way at
82
E.L. SINGMAN AND F. SCALIA
Fig. 2. Photomicrographs of sections of nasal retina taken from the left (affected) eye of the same frog (R137)used for Figure 1, showing the abnormal displacement of ganglion cells (arrows) into the IPL. Ganglion cells are back-filled with HRP applied to the optic nerve. The orthotopic ganglion cells of the GCL continue to form a monolayer after
optic nerve regeneration, and the IPL, although reduced, remains of uniform thickness. The sections on the left were counterstained with cresyl violet to show the layers of the retina. Counterstaining was omitted on the right to emphasize the HRP-label. Bar = 45 pm.
the same site t o increase the degree of endocytotic uptake. A plate of dental acrylic was then placed over the cranial defect, and the skin was sutured closed. The injections were made bilaterally in mirror-symmetric locations. In most
cases, the injection was aimed for the region of tectum to which the middle part of the nasal retina projects. In a smaller number of cases, we injected the tectal hemisphere either anterior, lateral or medial to the site of representa-
GANGLION CELL DISPLACEMENT
Fig. 3. Photomicrographs of sections taken from the left (affected) retina of frog R141,showing that the abnormally displaced ganglion cells (arrows) project to the optic tectum. This frog survived 17 weeks following nerve crush. The sections were lightly stained with cresyl violet. Bar = 25 +m.
83
E.L. SINGMAN AND F. SCALIA
84
TABLE 2. Proportion of SurvivingNeurons That Undergo Displacement
TABLE 1. Distance of Ganglion Cell Displacement During Optic Nerve Regeneration' Mean values (N = 219) ~~
Displaced Cells per 10,000pm2
Percentage of total neurons
2.2 ? 0.37 1.9 t 0.38 2.6 t 0.49 1.7 t 0.44 1.7 ? 0.28 2.3 ? 0.46 2.5 ? 0.47 1.5 i 0.37 2.2 t 0.50 Mean = 2.1 f 0 25
5.6 -t 1.1 9.3 i 1.1 6.8 i 1.3 3.7 i 0.92 4.8 2 0.90 4.2 t 0.93 5.3 t 0.99 3.5 t 1.1 6.5 i 1.5 Mean = 5.5 t 1.2
~
14.3 5 0.39 Fm 5.8 5 0.20 pm 7.1 ? 0.41 Km (S.D. = 3.1) 10.0 2 0.45 km
Width of IPL (A) Cell diameter (Bf Distance from cell center to outer border of GCL tC! Assumed cell displacement (D) Percent oftheoretical maxunum I s placement (% !
69.9 i 2.4%
'The diameters and positions within the IPL of 219 abnormally displaced &an cellslocated in the nasal quadrant of the affeded (left) retina of the frog used for Figure 2 (R137) were measured. Each parameter presented in this table is descrild s c h e m a t i d y in Figure 4. For all mean values tabulated, the accompanyingconfidence intervals are taken at the 9 5 6 level (conf. int. = 1 9 6 x SE).
R141 R145 R120 R121 R138 R147 R148 R152 R151
17 18 35 39 52 75 79 94 101
< . . ... . . . .
.... .. .'
..., ...
C
MAX.
I PL
0
Fig. 4. The parameters measured for Table 1are shown diagrammatically. Measurements included the width of the IPL (A) at the location of an abnormally displaced ganglion cell (stippled), the diameter (B) of the abnormally displaced cell, taken perpendicular to the plane of the GCL, and the perpendicular distance ( C ) from the center of the abnormally displaced cell to the outer border of the GCL. One displaced cell (at right) is shown flattened against the inner surface of the IPL.
Assuming that the abnormally displaced cells were originally located in the GCL, then the degree of displacement (D)for any cell would be C + Bi2. Since none of the abnormally displaced ganglion cells has been observed to cross the inner border of the IPL, then the maximum displacement (MAX.) is taken as A. The darkened cell within the INL represents a typical cell of Dogiel.
tion of the optic disc. The retinas of both eyes were then studied after periods of retrograde transport of 3-5 days. To study retinal ganglion cell distributions without regard for their sites of projection, HRP was delivered to the optic nerve instead of to the tectum in some cases (N = 8). This was done by placing a pellet of HRP against the cut
surface of the optic nerve transected intraorbitally (Scalia and Colman, '74). Retinal ganglion cells projecting anywhere in the brain become heavily back-filled by this method. At the time of sacrifice, the frogs were first dark-adapted, then they were anesthetized with MS-222 and placed on ice.
GANGLION CELL DISPLACEMENT
85
T39 Left Retina
.J \
)
R138 Normol (Right) Retina
.h
. . . D. . .. . . . .. .. .. ........... . . . ... .. ..... .0 . OD
T54 Right Retina
\
n
R138 Affected (Left) Retina
. . ....- . . . . . .
'i
Fig. 5. The topographic distribution of the 100 km x 100 km sampling frames (filled squares) is depicted on tracings of the retinas. To enhance their visibility, the squares are drawn larger than scale. Upper left: Left retina of normal frog T39. Three days prior to sacrifice, this frog received a large injection of HRP into a caudomedial locus of its right tectal hemisphere (see Fig. 7). Lower left: Right retina of normal frog T54. Five days prior to sacrifice, this frog received
.. .. .. .. .. . .. . . . . . . . . .. . .. . .. . ' .OD. I . . . . . . . . . . . . . . . .. . .. . .. . .. . .
- . .
a large injection of HRP into a rostrolateral locus of its left tectal hemisphere (see Fig. 8). Upper right: Normal (right) retina of regenerate frog R138. Five days prior to sacrifice, large injections of HRP were made into bilaterally symmetric, caudolateral loci of the frog's tectal hemispheres (see Fig. 9). Lower right: Left (affected) retina of frog R138 (see Fig. 10).
E.L. SINGMAN AND F. SCALIA
86 In semidarkness, the frogs were exsanguinated by perfusion through the conus arteriosus with frog Ringer's (pH 7.3). The eyes were immediately removed, and the carcass was perfused with a fixative (4% glutaraldehyde in 0.15 M phosphate buffer, pH 7.3) to save the brain. After marking the dorsal, ventral, and temporal poles of the eyes with small puncture-holes, the retinas were dissected out into frog Ringer's solution containing 4% dextrose. The pigment cell layer (and any attached choroid coat) was removed with a forceps, andor with gentle brushing, and the retina was brought up onto a glass slide. Relief cuts were made to allow flattening of the retina, and the flattened specimen was fixed in ascending steps of glutaraldehyde concentrations, while under light pressure from a glass coverslip. After fixation in 4% glutaraldehyde for 2 hours, the retinas were washed in phosphate buffer overnight and incubated for HRP activity by the nickel-cobalt intensified diaminobenzidine system of Adams ('81). Subsequently, the retinas were mounted on glass slides for LM analysis as flat-mounts, or the retinas were sectioned frozen at 12-25 km perpendicular to the plane of flattening, and the sections were mounted on slides in serial order. When counterstaining was required, cresyl violet was used. In some cases, flat-mounted retinas that had been previously examined were removed from their slides, sectioned as above, and reexamined in-section. Ganglion cells were examined in the flat-mounted retinas for the presence or absence of HRP, and the numbers of HRP-labeled and unlabeled cells counted were recorded with the assistance of a microscope-pantograph system (Microcode, Boeckeler Instruments) and a video-image processing system (Image-Pro, Media Cybernetics, Inc.) connected to an IBM personal computer. The population densities of the labeled and unlabeled cells were analyzed by plotting these data onto tracings of the retinas with a computer-based graphics system (SURFER, Golden Software, Inc.). Two sampling methods were used: 1)Cells were counted at 1 , 0 0 0 ~in 10,000 pm2 (100 pm x 100 pm) windows systematically distributed over the retinal areas corresponding to the tectal injection sites. Each of these distributed counting frames was carefully examined through depth to detect possible cell-labeling in the inner plexiform or inner nuclear layer. 2 ) Cells of Dogiel projecting to the tectum were counted by scanning the retina at 1 , 0 0 0 ~in adjacent vertically oriented rasters over two 2-mm-wide strips centered on the optic disc and on the area of maximal retrograde label in the appropriate retinal sector. The rasters containing the optic disc extended the full vertical diameter of the retina. Since all the labeled cells of Dogiel found near the optic disc were located well within a 2 mm square, the search for cells of Dogiel in the area of maximal labeling was limited to a square of the same size.
RESULTS Degree and frequency of cell displacement In all the normal retinas examined as flat-mounts, in which ganglion cells had been back-filled with HRP applied to the optic nerve, the labeled ganglion cells were confined to a uniform monolayer (the retinal ganglion cell layer), except for the small populations of cells of Dogiel in the
Fig. 6. A photomicrograph focused upon the vitreal surface of the INL of the wholemounted left retina from frog T39,showing six cells of Dogiel (at a-0 labeled by an injection of HRP to the right tectal hemisphere. The cell at f could not be focused with the others. The clarity of the cells is generally not optimal because their focal plane is deep within the thick whole-mounted retina.
inner nuclear layer, and the rare occurrences' of cell placements within the inner plexiform layer. This can be verified in the sectioned material (Fig. 11, and is consistent with earlier descriptions (Jacobson, '62; Ramon y Cajal, '72; Frank and Hollyfield, '87). Cells of Dogiel were observed in the retina flat-mounts from the normal frogs and from either eye of the frogs sustaining unilateral optic nerve regeneration. Their cell bodies were located within the inner surface of the inner nuclear layer, and their dendrites formed an inverted tree, directed vitreally. Characteristically, only the dendrites of the cells of Dogiel extended into the inner plexiform layer; their cell bodies were always submerged fully within the confines of the inner nuclear layer (Fig. 1). Their location in the inner nuclear layer serves to distinguish the cells of Dogiel of the frog from the abnormally displaced ganglion cells observed in this species after optic nerve regeneration. In contrast, the abnormal displacement of ganglion cells into the inner plexiform layer was observed only in the retinas that were undergoing or had completed optic nerve regeneration. When observed in the sectioned material (Figs. 2, 31, the dendrites of the abnormally displaced cells were usually tangentially spread, or else they were directed away from the vitreous body, in the manner of the orthotopic ganglion cells (cells residing in the ganglion cell layer proper), but unlike the cells of Dogiel, whose dendrites extend toward the vitreous body. The degree of displacement varied widely across the thickness of the inner plexiform layer. Any abnormally displaced cells present at 'Over a large sample of neurons counted in the retinal ganglion cell- and inner plexiform-layers in normal frog T39, without regard to the presence or absence of the HRP-label,only 17 out of 5367 neurons (0.32%)were located in the inner plexiform layer.
87
GANGLION CELL DISPLACEMENT the border of the inner nuclear layer were usually flattened against its inner surface. Sometimes the flattening was very marked, suggesting the presence of a planar barrier against displacement across the interface between the inner plexiform and inner nuclear layers. Since the inner plexiform layer remains of uniform thickness and the orthotopic ganglion cells continue to form a monolayer after nerve section (Figs. 2, 3), it was possible to measure the outward displacement of the ectopic ganglion cells. Table 1 (and see Fig. 4)gives measurements of the cell displacement in one frog (R137). This frog was making prey-catching responses using its affected (left) eye when sacrificed at 8 weeks after nerve crush. Three days prior to sacrifice, both of its optic nerves were transected and HRP was instilled into the peripheral stump to label its retinal ganglion cells (see Materials and Methods). The mean width of the inner plexiform layer in the nasal quadrant of the affected retina was 14.3 _" 0.39 pm, and the mean distance from the center of the abnormally displaced cells to the scleral edge of the ganglion cell layer was 7.1 r 0.41 pm. Assuming that the abnormally displaced ganglion cells had migrated outward from the ganglion cell layer, the mean cell displacement would be 10.0 5 0.45 pm. Taking the maximum theoretical displacement as the distance to the inner border of the inner nuclear layer, then the mean traverse of the abnormally displaced cells would be 69.9 ? 2.4%of the maximum theoretical displacement. This takes many of the translocated cells more than half-way across the inner plexiform layer at a survival period of only 8 weeks. In the corresponding quadrant of the normal retina of this frog, the mean width of the inner plexiform layer was 18.7 ? 0.36 Km (N = 345 measurements taken along 18 sections at 100 Km intervals), indicating that the thickness of the inner plexiform layer had been reduced by approximately 24% in this specimen after optic nerve section. The increase in the number of neurons that can be observed in the inner plexiform layer after initiating optic nerve regeneration is shown in Table 2. The cell-counts for this table were taken from flat-mounted retinas (e.g., Fig. 5) of frogs that had received a tectal injection of HRP. On average, the abnormally displaced cells represented 5.5%of the number of surviving neurons in the combined ganglion cell-inner plexiform layer in the areas counted. This percentage was stable over nine frogs surviving widely varying periods since nerve section.
Contrastingtectal projections of the abnormallydisr>laced sandioncells and the Tectal injections of HRP labeled many ganglion cells in the normal retinas and in the retinas sustaining optic nerve regeneration. The HRP-labeled orthopic ganglion cells and the HRP-labeled abnormally displaced ganglion cells (Fig. 3) were localized in regions of the retina corresponding to the sites of tectal injection.* The population densities of labeled cells were greatest in the areas representing the centers of the injection sites. Since the injections subtended large areas of the tectum, the labeling frequencies in the
'Several laboratories (Scalia et al., '85; Stelzner and Strauss, '86; Beazley et al., '86; Frank and Hollyfield, '87) have estimated with different methods that 12-16% of the neurons in the ganglion cell layer of R. pipiens are displaced amacrine cells. We therefore adjusted our counts of total neurons downward by our own previous estimate of the displaced amacrine population (16%)to obtain the number of ganglion cells in the ganglion cell layer.
retina attained a maximum over a fairly level plateau (broken by minor peaks and depressions) that extended over a significant portion of the retina. In contrast, cells of Dogiel (Fig. 6) were labeled from the tectum only in those specimens in which the plateau of maximal labeling extended near the optic disc. Then, the labeled cells of Dogiel (i.e., ganglion cells in the inner nuclear layer3)were concentrated in a horseshoe-shaped area capping the optic disc on its nasal, dorsal, and temporal sides, independent of the location of the center of the injection site (Figs. 7-10). Differential cell counts are given in Table 3. Few or none of the cells of Dogiel were labeled in the centers of the areas of maximal labeling. In two animals (T55 and R120) in which tectal injections did not involve the area of representation of the optic disc, no cells of Dogiel were labeled.
DISCUSSION Cell displacement The present study confirms and extends our earlier report that large numbers of ganglion cells may be found in the inner plexiform layer in the retina of R a m pipiens after a period of optic nerve regeneration (Scalia et al., '85), whereas very few neurons are normally found in the inner plexiform layer of frogs (Frank and Hollyfield, '87; Straznicky and Straznicky, '88). Only a few specimens were available at the time of our previous report on ganglion cell displacement, and we did not know that the large numbers of ectopic cells observed at 8-12 months survival appear during the stages of optic nerve regeneration. We have presently counted the numbers of abnormally displaced cells for the first time and studied their degree of displacement in sectioned material. Any question that the appearance of cell displacement may be an illusory consequence of irregularity in ganglion cell alignment combined with shrinkage of the inner plexiform layer after cell loss can be dispelled by reference to Figures 2 and 3. The measurements given in Table 1 indicate that although most of the abnormally displaced cells come to lie roughly halfway across the inner plexiform layer, there is a large degree of variability. Some cells were found fully displaced and flattened against the inner face of the inner nuclear layer. Therefore, we are convinced of the reality of the cell displacement, and urge consideration of its mechanism and functional significance. It was originally suggested that the abnormally displaced cells may be redifferentiated amacrine cells (Scalia et al., '85). Although we have no further information on that particular question, some additional alternatives can be
Yt should be noted that the cellular layers of the retina in the frog remain parallel to each other as they approach the optic disc, and the optic nerve head penetrates the retina without disturbing the regularity of the layers. The cells of Dogiel counted in the peripapillary region are clearly localized within the inner border of the inner nuclear layer.
Figs. 7-10 (see pages 88-91). A Expanded view of the optic disc, showing the topographic locations of tectally projecting cells of Dogiel (one dot = one cell). B: Schematic drawing of the corresponding teetum, showing the HRP-injection site (filled region). C: Tracing of the corresponding retina, showing the region of maximal HRP labeling as well as the field of labeled cells of Dogiel. The orientation of the optic disc is the same in both A and C. Rectangular boxes outline the areas subtended by the distributed sampling frames.
88
E.L. SINGMAN AND F. SCALIA
T39 Left Retina
A
C Fig. 7. The left retina of normal frog T39.
considered here. One suggestion is that the displaced cells might be migrated cells of Dogiel. This seems unlikely because there are more abnormally displaced cells than cells of Dogiel, which account for only 1 4 % of the ganglion cell population, and because the cells of Dogiel suffer rates of cell death comparable to that of the population of ganglion cells in the ganglion cell layer (Dunlop et al., '89). Another possibility is that there may be ganglion cell mitogenesis from a precursor population of cells capable of mitotic activity and neuronal differentiation. This is made unlikely by the observations of Bohn and Reier ('82) and Pedalina and Beazley ('861, who used tritiated-thymidine autoradiography to show that there is no resurgence of mitotic activity after optic nerve section in the retinas of adult Xenopus laevis or Hyla moorei, respectively. Perhaps the simplest explanation is that the ganglion cells of the ganglion cell layer are the source of the displaced neurons. Two mechanisms for the displacement of ganglion cells from the ganglion cell layer may be suggested. One is that ganglion cells retract their dendrites, losing all or part of their synaptic attachments and become migratory for a period during which they journey across the inner plexiform layer. Subsequently, they would reestablish their dendrites and presynaptic connections. This mechanism may be examined by further morphological study of retinal ganglion cells during the early stages of optic nerve regeneration. While there is no precedent for the resumption of cellular motility by adult neurons, such an event would not be inconsistent with the widely held idea that mature
neurons capable of axon regeneration undergo a partial reversion to earlier developmental states after axotomy. Admittedly, migration away from the vitreous is the reverse of the direction ganglion cells migrate during development. The hypothesized loss of presynaptic input, which would seem to be a necessary adjunct to cell migration, is not inconsistent with earlier observations on the loss of at least some part of the synaptic scale from motorneurons (e.g., Sumner and Sutherland, '73) and other central neurons (Chen et al., '77) surviving axotomy. Loss and reacquisition of synaptic input by retinal ganglion cells during optic nerve regeneration would suggest a remarkable capacity for reconstruction of functionally useful synaptic microcircuitry, since it is reported that all electrophysiologically defined classes of retinal ganglion cells reappear after nerve regeneration (Maturana et al., '59; Keating and Gaze, '70). The other mechanism we can suggest to explain the ganglion cell displacement involves an intracytoplasmic nuclear migration. Specifically, it is possible that the nucleus of a certain proportion of the ganglion cells undergoing axon regeneration may move into a primary dendrite and migrate some distance within the dendrite shaft towards the inner nuclear layer. During this migration, perinuclear organelles might be transported along with the nucleus, resulting in the outward (sclerad) transposition of the perikaryon, and the transformation of the trailing portion of the cell into part of the axon. The distance over which the nucleus is capable of moving may be constrained by the branching pattern of the dendrite system. The
GANGLION CELL DISPLACEMENT
89
A
T54 Right Retina
C Fig. 8.
The right retina of normal frog T54.
nucleus may move only a short distance in cells whose major dendrites branch low (vitreally) in the inner plexiform layer, but may move entirely across in those cells whose primary dendrite bifurcates only after reaching nearly to the inner nuclear layer. The nuclear migration suggested here is reminiscent of the nuclear translocations that take place in the developing neural tube during mitosis (Sauer, '351, specifically, the proluminal phase of nuclear migration immediately preceding cell division. In the adult retina, movement from within the ganglion cell layer toward the photoreceptor layer is equivalent to movement toward the ventricular lumen of the developing eye. This explanation is also consistent with the concept that neurons undergoing axon regeneration experience a partial reversion to a developmentally juvenile biology. In this case, certain processes normally leading to cell-division may be abortively reinstated after axon section in some neurons.
Cells ofDogiel We had previously reported that cells of Dogiel do not project to the optic tectum in the frog (Singman and Scalia, '89); we now present evidence to the contrary. We had originally overlooked the tectally projecting population of cells of Dogiel because they are not distributed evenly throughout the retina. The population of tectally projecting cells of Dogiel is restricted to a horseshoe-shaped locus surrounding the optic disc nasally, temporally, and dorsally. Their axons apparently terminate near the tectal
representation of the optic disc, because they are labeled only when tectal HRP injections spread to involve the peripapillary representation area. Their presence as a dorsal cap around the optic disc and their absence on its ventral side recalls the early topography of the developing retina, in which the area at the dorsal end of the optic disc is the first part of the retina to develop (Grant et al., '80). This suggests that these cells may be among the developmentally oldest ganglion cells in the frog retina. The projections of cells of Dogiel in the frog are not completely understood. Frank and Hollyfield ('87) estimated that cells of Dogiel comprise approximately 5% of the overall population of ganglion cells in a horizontally oriented streak situated just dorsal to the optic disc. Elsewhere in the retina, the cells of Dogiel represented less than 1% of the overall population of ganglion cells. Montgomery et al. ('81) showed that some cells of Dogiel project to the nucleus of the basal (accessory) optic root (nBOR) in the frog, although these cells comprised the smallest portion of the ganglion cells projecting to this structure. The cells of Dogiel projecting to nBOR were widely distributed throughout the retina, although their frequency in the peripapillary region was not specifically noted. It would be interesting to know whether the tectally projecting cells of Dogiel and those projecting to the nBOR occupy separate spatial domains. Cells of Dogiel are found in most if not all vertebrate species. They generally comprise less than 5% of the overall
E.L. SINGMAN AND F. SCALIA
90
R138 Normal (Right) Retina
A .-
OD
L
TECTUM
C Fig. 9. The normal (right) retina of regenerate frog R138. Only the injection site on the left tectal hemisphere is shown.
population of retinal ganglion cells, although the proportion is as great as 15%in the mudpuppy (Arkin and Miller, '88) and 43%in the larval sea lamprey (Rubinson and Cain, '89). The relative distributions and projections of cells of Dogiel vary widely between species. Cells of Dogiel are most heavily concentrated in the retinal periphery in the newt (Ball and Dickson, '831,the ray (Collin, '88), the larval sea lamprey (Rubinson and Cain, '89), the chick (Prada et al., '89), the mouse (Drager and Olsen, '81), the rat (Buhl and Dann, '88; Linden, '87), the rabbit (Robson and Hollander, '84), and the wallaby (Coleman et al., '87). In contrast, most cells of Dogiel reside in the central retina in the pigeon (Hayes and Holden, '831,and in the frogs Rana pipiens (Frank and Hollyfield, '87) and Hyla moorei (Dunlop et al., '89). In the mudpuppy (Arkin and Miller, '88) and the frog Xenopus laevis (Straznicky and Straznicky, '881, the cells of Dogiel are distributed fairly evenly across the retina. It has been shown that cells of Dogiel comprise the only cells projecting to nBOR of the pigeon and that they apparently do not project to the tectum in this species (Fite et al., '81; Hayes and Holden, '83).In contrast, it has been reported that the tectally projecting cells of Dogiel in the chicken are restricted to the temporal hemi-retina, which contains the optic disc in this species, and are most concentrated near the optic disc itself (Heaton et al., '79).
Although the cells of Dogiel do not comprise the exclusive projection to nBOR in the turtle, many do make this projection, but, as in the pigeon, they apparently do not project to the optic tectum (Reiner, '80). In mammals such as the rabbit (Oyster et al., '80), the rat (Buhl and Dann, '88), and the cat (Farmer and Rodieck, '821, however, the cells of Dogiel do project to the superior colliculus, and also to the dorsal lateral geniculate nucleus. The cells of Dogiel projecting to the monkey's dorsal lateral geniculate nucleus are restricted to the peripapillary region (Bunt and Minckler, '77). These projections appear to be more prominent than the projections to the medial terminal nucleus (MTN), the homologue of nBOR (for review see Simpson, '84). Specifically, it has been reported that no cells of Dogiel project to the MTN in the rabbit and cat.
ACKNOWJXDGMENTS We thank Ms. Suri Roca, Ms. Jan Prokosch, and Ms. Gloria Scott for their technical assistance. We also thank Mr . Vincent Garofalo for his expert photographic assistance. This study was supported by a USPHS Research Grant (EY 05284) from the National Eye Institute and was submitted in partial fulfillment of requirements for the degree of Doctor of Philosophy.
GANGLION CELL DISPLACEMENT
91
R138 Affected (Left) Retina
A
OD
0
1 rnrn
\
C Fig. 10. T h e left (affected) retina of frog R138.Only t h e injection site on t h e right hemisphere is shown.
TABLE 3. Location of Cells of Dogie1That Project to the Tectuml
Frog T39 T54 R138
R138" T554 ~ 1 2 0 ~
No. of distrihuted samples 52 30 73 116 35 36
Overall area
sampled2 7.9 mm2 15 mm2 Entire retina Entire retina 17 mm2 3.6 mm2
Adams, J.C. (1981) Heavy metal intensification of DAB-based HRP reaction product. J. Histochem. Cytochem. 29:775. Arkn, M.S., and R.F. Miller (1988)Mudpuppy retinal ganglion cell morphology revealed by an HRP impregnation technique which provides Golgilike staining. J. Comp. Neurol. 270t185-208. Ball, A.K., and D.H. Dickson (1983) Displaced amacrine and ganglion cells in the newt retina. Exp. Eye Res. 36t199-213. Beazley, L.D., J.E. Darby, and V.H. Perry (1986) Cell death in the retinal ganglion cell layer during optic nerve regeneration for the frog Rana pipiens. Vision Res. 26:543-556. Bohn, R.C., and P.J. Reier (1982) Anomalous axonal outgrowth at the retina
Retinal sector containing maximal label Nasal Temporal Dorsotempral Dorsotemporal Ventronasal
Nasal
RGC
cDt
cDu
4,508 2,914 4,890 2,683 1,572 2,549
2 2 0
232 93 100
0 0 0
143
0 0
caused by injury to the optic nerve or tectal ablation in adult Xenopus. J. Neurocytol. 1It2 11-234. Buhl, E.H., and J.F. Dann (1988) Morphological diversity of displaced retinal ganglion cells in the rat: a lucifer yellow study. J. Camp. Neurol. 269t210-218. Bunt, A.H., and D.S. Minckler (1977) Displaced ganglion cells in the retina of the monkey. Invest. Ophthalmol. Vis. Sci. 16t95-98. Chen, D., W.W. Chambers, and C.N. Liu (1977) Synaptic displacement in intracentral neurons of Clarke's nucleus following axotorny in the cat. Exp. Neurol. 57t1026-1041. Coleman, L.A., A.M. Harman, and L.D. Beazley (1987) Displaced retinal ganglion cells in the wallaby Setoncx bruchyurus. Vision Res. 27:12691277.
92 Collin, S.P. (1988) The retina of the shovel-nosed ray, Rhinobatos batillurn (Rhinobatidae): morphology and quantitative analysis of the ganglion, amacrine and bipolar cell populations. Exp. Biol. 47:195-207. Drager, U.C., and J.F. Olsen (1981)Ganglion cell distribution in the retinaof the mouse. Invest. Ophthalmol. Vis. Sci. 20285-293. Dunlop, S.A., M.F. Humphrey, and L.D. Beazley (1989) Survival of displaced retinal ganglion cells after optic nerve regeneration in the frog Hyla rnoorei. SOC. Neurosci. Abstracts 15:872. Farmer, S.G., and R.W. Rodieck 11982) Ganglion cells in the cat accessory optic system: Morphology and retinal topography. J. Comp. Neurol. 205: 190-198. Fite, K.V., N. Brecha, H.J. Karten, and S.P. Hunt (1981) Displaced ganglion cells and the accessory optic system of the pigeon. J. Comp. Neurol. 1955279-288. Frank, B.D., and J.G. Hollyfield (1987) Retinal ganglion cell morphology in the frog, Ranapipzens. J. Comp. Neurol. 266,413-434. Grant, P., E. Rubin, and C. Cima (1980) Ontogeny of the retina and optic nerve in Xenopus laeuis. I. Stages in the early development of the retina. J. Comp. Neurol. 189~593-613. Hayes, B.P., and A.L. Holden (1983) The distribution of displaced ganglion cells in the retina of the pigeon. Exp. Brain Res. 49:181-188. Heaton, M.B., I.M. Alvarez, and J.E. Crandall (1979) The displaced ganglion cell in the avian retina: developmental and comparative considerations. Anat. Embryol. (Berl.) 155:161-178. Humphrey, M.F. (1987) Effect of different optic nerve lesions on retinal ganglion cell death in the frog Ranapipiens. J. Comp. Neurol. 266:209219. Humphrey, M.F. (1988) A morphometric study of the retinal ganglion cell response to optic nerve severance in the frog Rana pzpiens. J. Neurocytol. 17293-304. Humphrey, M.F., and L.D. Beazley (1985) Retinal ganglion cell death during optic nerve regeneration in the frog Hyla moorei. J. Comp. Neurol. 236:382402. Humphrey, M.F., J.E. Darby, and L.D. Beazley (1989) Prevention of optic nerve regeneration in the frog Hyla rnoorei transiently delays the death of some ganglion cells. J. Comp. Neurol. 279:187-198. Jacobson, M. (1962) The representation of the retina on the optic tectum of the frog. Correlation between retino-tectal magnification factor and retinal ganglion cell count. Q. J. Exp. Physiol. 47:170-178. Keating, M.J., and R.M. Gaze (1970) The depth distribution ofvisual units in the contralateral optic tectum following regeneration of the optic nerve in the frog. Brain Res. 21:197-206. Linden, R. 11987) Displaced ganglion cells in the retina of the rat. J. Comp. Neurol. 258:138-143. Maturana, H.R., J.Y. Lettvin, W.S. McCulloch, and W.H. Pitts (1959) Physiological evidence that cut optic nerve fibers in the frog regenerate to their proper places in the tectum. Science 130:1709, 1710. Montgomery, N., K.V. Fite, and L. Bengston (1981) The accessory optic system of Rana pzpiens: Neuroanatomical connections and intrinsic organization. J. Comp. Neurol. 203:595-612. Oyster, C.W., J.I. Simpson, E.S. Takahashi, and R.E. Soodak (1980) Retinal ganglion cells projecting to the rabbit accessory optic system. J. Comp. Neurol. 190~49-61. Pedalina, R., and L.D. Beazley 11986) A tritiated thymidine study of cell division in the retina during optic nerve regeneration for the frog Hyla rnoorei. Neurosci. Lett. 1Suppl.l23:S70.
E.L. SINGMAN AND F. SCALIA Prada, F.A., C.E. Chmielewski, M.E. Dorado, C. Prada, and J.M. GenisGalvez (1989) Displaced ganglion cells in the chick retina. Neurosci. Res. 6:329-339. Purves, D. (1975) Functional and structural changes in mammalian sympathetic neurones following interruption of their axons. J. Physiol. (Lond.) 252t429-463. Ramon y Cajal, S. (1972) The Structure of the Retina. Springfield, 11,: Charles C. Thomas. Reiner, A. (1980) A projection of displaced ganglion cells and giant ganglion cells to the accessory optic nuclei in turtle. Brain Res. 204:403409. Robson, J.A., and H. Hollander (1984) Displaced ganglion cells in the rabbit retina. Invest. Ophthalmol. Vis. Sci. 25~1376-1381. Rubinson, K., and H. Cain (1989) Neural differentiation in the retina of the larval sea lamprey (Petromyzon marinus). Visual Neurosci. 3t241-248. Sauer, F.C. 11935) Mitosis in the neural tube. J. Comp. Neurol. 62377-405. Scalia, F., V. Arango, and E.L. Singman (1985) Loss and displacement of ganglion cells after optic nerve regeneration in adult Ranapzpiens. Brain Res. 344.267-280. Scalia, F., and D.R. Colrnan (1974) Aspects of the central projections of the optic nerve in the frog as revealed by anterograde migration of horseradish peroxidase. Brain Res. 79t496-504. Sheard, P.W., and L.D. Beazley (1988) Retinal ganglion cell death is not prevented by application of tetrodotoxin during optic nerve regeneration in the frog Hyla rnoorei. Vision Res. 28:461-470. Simpson, J.I. (1984) The accessory optic system. Annu. Rev. Neurosci. 7: 13-41. Singman, E.L., and Scalia, F. (1989) Cells of Dogie1 and ganglion cell displacement during optic nerve regeneration. SOC.Neurosci. Abstr. 15:1208. Singman, E.L., and F. Scalia (1990a) A quantitative study of the tectally projecting retinal ganglion cells in the adult frog. I. The size of the contralateral and ipsilateral projections. (Submitted. Singman, E.L., and F. Scalia (1990b) A quantitative study of the tectally projecting retinal ganglion cells in the adult frog. 11. Cell survival and functional recovery after optic nerve transection. (Submitted.) Stelzner, D.J., and J.A. Strauss (1986) A quantitative analysis of frog optic nerve regeneration: Is retrograde ganglion cell death or collateral axonal loss related to selective reinnervation? J. Comp. Neurol. 245:83-106. Stelzner, D.J., and J.A. Strauss (1988) Increase in ganglion cell size after optic nerve regeneration in the frog, Runapipiens. Exp. Neurol. 100~210215. Straznicky, C., and I.T. Straznicky (1988) Morphological classification of retinal ganglion cells in adult Xenopus laeuzs. Anat. Embryol. (Berl.) 178:143-153. Surnner, B.E.H., and F.I. Sutherland (1973) Quantitative electron microscopy on the injured hypoglossal nucleus in the rat. J. Neurocytol. 2: 3 15-328. Sumner, B.E.H., and W.E. Watson (1971) Retraction and expansion of the dendritic tree of motor neurones of adult rats induced in uiuo. Nature 233973-275. Yawo, H. (1987) Changes in the dendritic geometry of mouse superior cervical ganglion cells following postganglionic axotomy. J. Neurosci. 7:3703-3711.
