0306-4522/9255.00+ 0.00 Pergamon Press plc 0 1991IBRO
NeuroscienceVol. 46, No. 2, pp. 419429, 1992 Printed in Great Britain
NEUROGENESIS
IN THE RETINAL GANGLION LAYER OF THE RAT
CELL
B. E. REP.sE*t$ and R. J. COLELLO$§ tNeuroscience
Research Institute and Department of Psychology, University of California at Santa Barbara, U.S.A. $Department of Human Anatomy, University of Oxford, U.K.
Abstract-The present study has examined the birthdates of neurons in the retinal ganglion cell layer of the adult rat. Rat fetuses were exposed to tritiated thymidine in urero to label neurons departing the mitotic cycle at different gestational stages from embryonic days 12 through to 22. Upon reaching adulthood, rats were either given unilateral injections of horseradish peroxidase into target visual nuclei in order to discriminate (1) ganglion cells from displaced amacrine cells, (2) decussating from non-decussating ganglion cells, and (3) alpha cells from other ganglion cell types; or, their retinae were immunohistochemically processed to reveal the choline acetyltransferase-immunoreactive amacrine cells in the ganglion cell layer. Retinae were embedded flat in resin and cut enjim to enable reconstruction of the distribution of labelled cells. Retinal sections were autoradiographically processed and then examined for neurons that were both tritium-positive and either horseradish peroxidase-positive or choline acetyltransferase-positive. Tritium-positive neurons in the ganglion cell layer were present in rats that had been exposed to tritiated
thymidine on embryonic days ElkE22. Retinal ganglion cells were generated between El4 and E20, the ipsilaterally projecting ganglion cells ceasing their neurogenesis a full day before the contralaterally projecting ganglion cells. Alpha cells were generated from the very outset of retinal ganglion cell genesis, at E14, but completed their neurogenesis before the other cell types, by E17. Tritium-positive, horseradish peroxidase-negative neurons in the ganglion cell layer were present from El4 through to E22, and are interpreted as displaced amacrine cells. Choline acetyltransferase-positive displaced amacrine cells were generated between El6 and E20. Individual cell types showed a rough centroperipheral neurogenetic gradient, with the dorsal half of the retina slightly preceding the ventral half. These results demonstrate, first, that retinal ganglion cell genesis and displaced amacrine cell genesis overlap substantially in time. They do not occur sequentially, as has been commonly assumed. Second, they demonstrate that the alpha cell population of retinal ganglion cells and the choline acetyltransferaseimmunoreactive population of displaced amacrine cells are each generated over a limited time during the periods of overall ganglion cell and displaced amacrine cell genesis, respectively. Third, they show that the very earliest ganglion cells to be generated in the temporal retina have exclusively uncrossed optic axons, while the later cells to be generated therein have an increasing propensity to navigate a crossed chiasmatic course.
During retinal development in the rodent, neurons in the ganglion cell layer are said to be generated in a rough centroperipheral gradient, with large neurons being generated before small neurons at any retinal locus.5~29~30~37 The larger neurons are said to be ganglion cells, while the smaller neurons are considered to be displaced amacrine cells.29~37Such a precedence of ganglion cell genesis over displaced amacrine cell genesis has also been proposed for the cat.** In the rat, displaced amacrine cells migrate into the ganglion cell layer relatively late during retinal development,*’ consistent with their postulated late genesis. Soma size differences are also evident between the different retinal ganglion cell classes, and there is *To whom correspondence should be addressed at: the Neuroscience Research Institute, University of California, Santa Barbara, CA 93106, U.S.A. §Present address: Institute for Brain Research, University of Zurich, Switzerland. ChAT, choline acetyltransferase; CRF, corticotropin releasing factor; DA, dopamine; DAB, diaminobenzidine; E, embryonic day; HRP, horseradish peroxidase.
Abbreviations:
419
good evidence that the soma size of some retinal ganglion cell classes overlaps with displaced amacrine cell classes 13,15,18,19,21; se also 16.36339 Consequently, soma size is clearly an inadequate basis for a thorough discrimination of ganglion from displaced amacrine cells. Of the above studies, only two have employed the retrograde transport of horseradish peroxidase (HRP) as a means of positively identifying retinal ganglion cells following large injections of the tracer into visual nuclei;51” if the injections label the complete population of retinal ganglion cells, displaced amacrine cells within the ganglion cell layer can be identified by the absence of the label. Using this approach, Harman and Beazley” report that displaced amacrine cell genesis in the wallaby is contemporaneous with ganglion cell genesis, rather than sequential. In the cat, the different retinal ganglion cell classes, which are more readily differentiated according to their size, proliferate independently, though partially overlapping in time, with central retina preceding peripheral retina for any cell class.34,35The morphologically distinct ganglion cell classes of rodents may also be generated independently, but because there is
420
3. E.
REESE and
substantial overlap between their soma size distributions, and because of the difficulty in discriminating displaced amacrine from ganglion cells, the available evidence does not resolve the pattern of neurogenesis in the ganglion cell layer of rodents. In the mouse, genesis of retinal ganglion cells has been reported to proceed as one gradual centroperipheral wave,5 in which the temporally independent genesis of different cell classes may be concealed (e.g. Ref. 6). In order to resolve these issues, the present study has investigated the genesis of neurons in the retinal ganglion cell layer of the rat. We have made large, multiple, unilateral injections of HRP into the visual nuclei of adult rats which had been exposed to tritiated thymidine in utero on a specific embryonic day. We have used the presence of the HRP to positively identify axon-bearing neurons (ganglion cells) within the ganglion cell layer. Further, we have used the HRP to permit the identification of one type of retinal ganglion cell, the alpha cell,” in order to define its neurogenetic period and its topography of genesis relative to other HRP-positive cells. We have also used the presence of the tracer to determine the relative neurogenetic periods and the topography of genesis of the decussating from non-decussating retinal ganglion cells. As in the study by Harman and Beazley,” the absence of HRP labelling may be regarded as an indication that a given neuron in the ganglion cell layer is an axon-less interneuron (a displaced amacrine cell). We have therefore addressed the issue of whether displaced amacrine cell genesis follows, or coincides with, ganglion cell genesis. Because one may question whether the lack of HRP labelling is a reliable indication of displaced amacrine cells, we have also used an antibody to choline acetyltransferase (ChAT) in iittermates which had been exposed to tritiated thymidine in utero in order to positively identify one class of displaced amacrine cell within the ganglion cell layer.32
R. J.
COLELLO
ing. The latter caiendar day is called embryonic day 1 (El), the first day of gestation. Pregnant dams were anaesthetized with ether and given intra-peritoneal injections of 5 &i of [methy~,l’,2’-3H]thymidine (Amersham International; 126Ci/mmol) per gram of maternal body weight on the 12th. 13th, lrlth, 15th, l&h, 17th, l&h, 19th, 20th or 22nd day of gestation. Parturition routinely occurred on E23. Pups were reared by the dams until weaning, after which they were group-caged until adulthood. They were then used in one of two experiments, using HRP to reveal retinal ganglion cells, or using an antibody directed against ChAT to reveal this class of displaced amacrine cell (Table 1).
Rats were anaesthetized with a mixture of sodium pentobarbitone and chloral hydrate (30 mg/kg c L26mg!kg, respectively), placed in a stereotaxic headhoider, and then given six injections of 0.1 ~1 of 40% HRP dissolved in 2% dimethyl suiphoxide. The injections were placed stereotaxitally to include the optic tract and target optic nuclei of one hemisphere. Two days later, rats were heavily anaesthetizcd and returned to the headholder. and the dorsal margin of the limbus was then cauterized for later orientation of the retina. Rats were perfused through the heart with 50 ml of 0.9% saline followed by 100 ml of 1% paraformaldehyde in 0.1 M sodium phosphate buffer, pH 7.2, at 4°C. Eyes were removed immediately and a iarge radial cut was made extending from the dorsal margin of the retina to the optic disk following removal of the anterior chamber and lens. Retinae were dissected, mounted flat and postfixed for 1 h in 2.5% glutaraldehyde in buffer at 4°C. They were then rinsed in buffer and reacted for enzymatic activity using p-phenylenediamine + catechoi as the chromogen.” The rat was further perfused with 2OOml of 1.25% paraformaldehyde + 2.5% glutaraldehyde in buffer at 20°C. Brains were cut at 25 pm in the coronal plane and every fifth section was collected and hist~~emically and autoradiographical~y processed. Further details of the surgery, perfusion and histochemistry can be found elsewhere.27 Retinae were dehydrated and embedded flat in LR White resin2 and then cut in the plane of the retinal surface at 5 pm. Retinal sections were mounted from warm acetone onto slides, and then dried, coated with Ilford K-2 emulsion using the wire loop technique, I2stored in the dark at 4°C for t 2 weeks. develoned in Kodak D-19, stained with Methylene Blue and Azur i1 and then coverslipped. Retinae ipsilateral to the HRP injections were examined with a x 100 oil immersion objective for the presence of double-labelled neurons (cells which were both HRP-positive and heavily tritium-positive). The criterion for defining heavily labelled tritium-positive cells (henceforth “tritiumpositive”) was 50%. or greater, of the number of silver grains observed overlying the average of the three most heavily labelled cells in each retina. Retinae contralateral to the HRP injections were examined for the presence of tritium-positive cells, and were recorded as being either HRP-positive or HRP-negative. All double-Ia~lled cells in
Hooded rats of the Long-Evans strain, obtained from the breeding colony maintained in the Department of Experimental Psychology, Oxford, were used. Males were put with females from 400 p.m. until 9:00 a.m. the following mom-
Table 1. Number of retinae examined from rats that had been exposed to tritiated thymidine on different embryonic days Cohort age El2 Ef3 E14 El5 Ef6 El7 Et8 Ei9 B20
ChAT retinae 0
3 5 4 2 6 6 3 2
Ipsi HRP retinae 2 2 2 2 2 2 3 2 2
Contra HRP retinae 2 2 2 2 2 2 2
I ?8.
Neurogenesis in the rat’s retinal ganglion cell layer either retina were identified as being alpha or non-alpha (e.g. Fig. 1), based on the presence of a large soma and stout primary dendrites in the same 5 pm section.‘7*27 The distribution of tritium-positive cells was plotted on drawings of the retinal sections at a magnification of x 20 or x 26.5 with the aid of a drawing tube attached to the microscope. The complete retinal ganglion cell layer was then reconstructed using the pattern of blood vessels and retinal contours to align successive sections, as described by Walsh and Polley.”
have been seen in the ventral
421 retina of the ipsilateral
eye. This was never observed.
Tritium-positive cells were found in the retinal ganglion cell layer of rats that had been exposed to tritiated thymidine on embryonic days 14-22. None of the rats that had been exposed to tritiated thymidine on El2 or El3 in the HRP studies contained triti~-positive cells in the retina, despite detectable labelling within the brainstem of these very Choliw acetyitransferase experiments same animals. (One littermate from the ChAT study Rats were heavily anaesthetized, placed in the headholder, which had been exposed to tritiated thymidine on E 13 and the dorsal margin of the limbus of each eye was did contain a number of tritium-positive cells, none cauterized for later orientation. Rats were then removed of them double-labelled, in the central region of from the headholder and perfused through the heart with the retinal ganglion cell layer, presumably indicative 50 ml of 0.9% saline in 0.1 M sodium phosphate buffer of the variable maturity between fetuses. We shall fnH 7.2). followed bv 50 ml of either 1% or 2% ~arafo~ldehyde in thesame, at 4°C. Eyes were enucleated regard El4 as the beginning of neurogenesis for imm~iately and following removal of the anterior chamber neurons in the ganglion cell layer.) We were, of and lens, a large radial cut was made extending from the course, examining the retinal ganglion cell layer in the dorsal margin of the retina to the optic disk. Retinae were adult, and all of our observations concerning neurothen dissected, mounted flat and postfixed in 4% paraformaldehyde plus 0.05% glutaraldehyde in buffer at genesis are specific for those cells in the ganglion cell 4°C for 1 h. They were then left to sink in 30% sucrose at layer which survive into adulthood. We shall call the 4”C, and rinsed in a minimum of three changes of buffer population of tritium-positive cells from a rat that overnight. had received tritated thymidine on a given day a Retinae were shock-frozen in liquid nitrogen, rinsed in buffer (3 x 10 mm), and then incubated for 16 h at 4°C in rat “cohort” (e.g. an El6 cohort). monoclonal anti-ChAT antibody (Boeh~nger) at a dilution Cells in the retinal ganglion cell layer which were of 1: 10 dissolved in 0.04 M Tris-phosphate buffer (pH 7.8), positive for both tritiated thymidine and for HRP containing 0.2% carrageenan and 0.5% Triton X-100. They were found in rats that had been exposed to tritiated were then rinsed in several changes of Tris-phosphate buffer thymidine on El4E20 (e.g. Fig. Id). Ganglion cell for 1 h, and incubated for 1 h in rabbit anti-rat IgG (Sigma) at a dilution of 1: 50 in buffer (plus carrageenan and Triton, genesis begins in the dorsocentral retina, then spreads as above), at 20°C. After a rinse in buffer, retinae were to cover the entire retina on E16, and by E20 can be incubated in rat peroxidase-antiperoxidase (ICN Immunofound almost exclusively in the far ventral periphery. biologicals) at a dilution of 1: 75 at 20°C for I h. They were No double-labelled cells were found in rats that then washed in several changes of 0.1 M sodium phosphate buffer (pH 7.2, 20°C) for I h, and finally reacted for received injections of tritiated th~idine on E22. enzymatic activity using diamino~nzi~ne as the chroThis topographic pattern of neurogenesis across mogen (0.2% DAB plus O.Ot% H,O& Retinae were rinsed the dorsocentral to ventroperipheral retina conceals in phosphate buffer, dehydrated, embedded flat in LR White at least one independent neurogenetic gradient of a resin, sectioned at 5 pm in the plane tangential to the retinal specific retinal ganglion cell class, the alpha cell class. surface, and then mounted, processed for autoradiography and stained as above. The distribution of double-labelled Tritium-positive, HRP-positive cells were classed as cells (tritium-positive and ChAT-positive) was plotted on being alpha (e.g. arrowheads in Fig. 1) or non-alpha, reconstructed flat-mounts of the retinae, as above. and plotted separately in Figs 2-5. Consider first the genesis of the non-alpha retinal ganglion cells, summarized in the data from three cohorts for both the RESULTS contralateral (Fig. 2) and ipsilateral (Fig. 3) retinae. Genesis of the contralateral cells begins on El4 in a Genesis of retinal ganglion ceils broad zone of the dorsocentral retina (Fig. 2). On The large injections of HRP produced dense El.S, there is less of an apparent focus dorsocentrally, labelling across the complete contralateral retinal yet the distribution of double-labelled cells has not surface, and throughout the temporal crescent of the extended to the retinal periphery. Ganglion cell genipsilateral retina. The density of retrogradely labelled esis peaks on El6 and E17, and the complete retinal cells in the temporal crescent of the ipsilateral retina surface contains double-la~lled cells, with no clearly was about one quarter of that in the temporal discernible region of focus (Fig. 2). By E19, ganglion crescent of the contralateral retina, and the majority cell genesis is focused in the ventral half of the retina, of the cells were very heavily labelled. There was no and on E20, the last day to include double-labelled evidence for spread of the HRP across the midline, as cells, they are almost completely restricted to the judged from the histological sections through the ventral periphery (Fig. 2). brain, nor for the uptake of HRP by optic terminals Genesis of the non-alpha ipsiiateral cells (Fig. 3) in the visual nuclei opposite the side of the injection. likewise begins on El4 in the dorsocentral portion of Had the latter event occurred, particularly at the the temporal crescent, that part of the retina giving collicular midline, a blanket of lightly labelled cells, rise to the uncrossed projection.4 It spreads to further equal to the density of the crossed projection, should peripheral and ventral regions of the crescent by E17,
422
B.E.
REESE and
R.J.
COLELLO
Fig. I. Photomicrographs of HRP-labelled ganglion cells in retinal whole-mounts (a, c) and m tangential
sections(b, dj contralateral (a: b) and ipsilateral (c, d) to the injected hemisphere, Only the ganglion cells are visible in the whole-mounts (a, c), whereas the tangential sections of the retinal ganglion cell layer (b, d) have been subsequently processed for autoradiography and then stained with Methylene Blue and Azur II. Notice the large somata and radiate dendritic arbors characteristic of the alpha cells in the whole-mounts (arrowheads in a and c), and the co-localization of the tritiated thymidine label and the HRP label in one of the two alpha cells in d (arrowhead). a and c come from nasal and temporal retinae, respectively, while b and d both come from temporal retina. Notice the large, HRP-negative cell from the contralateral retina’s temporai crescent in b (arrowhead), presumably in an alpha cell projecting to the ipsilateral hemi~he~. b and d come from a rat that had been exposedto tritiated thymidine on El 5. Scale bar in c = 30pm for a and c, and = 20hm for b and d.
and a few ectopic, nasal cells are included in that cohort (Fig. 3). On E18, it is restricted to the most peripheral extreme of the crescent, and only a very few double-Ia~lled cells can be found on El9 (Fig. 3). No double-labelled cells were observed in rats that had received injections of tritiated thymidine on E20 (see also Ref. 26 for a more detailed description of the neurogenesis of the uncrossed projection).
In contrast, alpha cells are generated over a rela-
tively restricted portion of the period during which ganglion cells are being generated. This is shown jn Figs 4 and 5, again for the contralatera1 and ipsilatera1 retinae, respectively. Contralaterally projecting alpha cells are first generated dorsocentrally on E14; over the next three days, their topogenesis spreads to occupy progressively peripheral retinal loci so that by
423
Neurogenesis in the rat’s retinal ganglion cell layer E-20
E-17
temporal
Fig. 2. Distributions of double-la~li~ (HRP-positive, tritium-~sitive) non-alpha retina1 ganglion cells with crossed optic axons generated on different days of gestation. Dorsal is up in this and in all sub~quent retinal reconstructions. Each dot represents an individual cell.
El7 they are almost exclusively found in peripheral retina (Fig. 4). Ipsilaterally projecting alpha cells are generated during the same four day period, displaying similar dorsal before ventral, and central before peripheral gradients (Fig. 5). The lack of dorsal alpha cells in the
temporal crescent on El5 is surmising, since they are clearly present there on the preceding and subsequent days. Nevertheless, the uncrossed alpha cell population displays, to a rough approximation, the same topographic pattern of genesis as the non-alpha population, although restricted to a briefer duration.
E-17
,:_ .’ .: ;. .;,: 7..
temporal
@
nasal
:,..‘. .: 2. ?xt 8-y.‘.&.,&>..
Fig. 3. Distributions of double-labelled (HRP-positive, tritium-positive) non-alpha retinal ganglion cells with uncrossed optic axons generated on different days of gestation.
temporal
nasal
E-16
E-l?
Fig. 4. Distributions of double-labelled (HRP-positive, tritium-positive) alpha retinal ganglion crossed optic axons generated on different days of gestation.
cells
with
424
REESEand
B. E. E-14
R. J.
COLELLO E-15
temporal
Fig.
i
5.
Distributions of double-labelled (HRP-positive, tritium-positive) alpha retinal ganglion cells wtth uncrossed optic axons generated on different days of gestation.
0.6 -
‘-
P
d+
0.6 -
E ‘E $ $ 0.4 = $ z ._ $j 0.2 E
01
’
14
’
15
’
16
’
17
’
18
’
19
’
20
’
21
1
22
Embryonicday of 3H-thyexposure Fig. 6. Proportion of neurons in the ganghon cell layer generated on different embryonic days which were presumptive displaced amacrine cells. A ratio of 0.0 indicates that HRP-positive (ganglion) cells are generated exclusively, while a ratio of 1.0 indicates that HRP-negative (displaced amacrine) cells are generated exclusively. Six adjacent square millimetres were sampled from retina nasal and ventral to the optic disk, at eccentricities less than 50% of the retinal radius (and therefore outside the temporal crescent).
The large injections of HRP that we have employed were meant to retrogradely label all of the ganglion cells projecting to one hemisphere. Such injections have been estimated to label the large majority of such cellsi The number of unlabelled cells following such injections is also coincident with the number of neurons remaining in the retinal ganglion cell layer following transection of the optic nerve-the presumed displaced amacrine cell population.” Consequently, those cells which do not show HRP labelling in the present study should correspond to the population of displaced amacrine cells in the retina, with one exception: ganglion cells in the temporal crescent may project to either hemisphere, so lack of HRP labelling is not a reliable index of a displaced amacrine cell in this region (e.g. Fig. 1b). In the remainder of the retina contralateral to the injected hemisphere (that is, in nasal retina), HRP-negative cells in the ganglion cell layer should all be displaced amacrine cells. In every age group from El4 through to E22, we have observed tritium-positive, HRP-negative neurons in the nasal portion of the retinal ganglion cell layer. These cells comprise a small proportion of
Fig. 7. Photomicrographs of ChAT-immunoreactive amacrine cells in the retinal ganglion cell layer (a, c-f) and in the inner nuclear layer (b) in retinal whole-mounts (a, b) and in tangential sections through to the retinal ganglion cell layer (c-f). d-f are focused on the plane of the silver grains, showing the tritium label in the ganglion cell layer of rats exposed to [‘Hlthymidine on embryonic days 14, 17 and 20, respectively. Three tritium-positive, ChAT-positive double-labelled cells are indicated in e (arrowheads), but the silver grains are not readily apparent against the immunoreaction product in this reproduction. a-c have been photographed with Nomarski optics, and c is the same field as f, focused on the somata. Scale bar in f = 30 11m for a and h, and = 20 pm for c f.
Neurogenesis in the rat’s retinal ganglion cell layer
Fig. 7.
425
426
B. E. REESE and
the tritium-positive cells initially (Fig. 6), and are found only in the central retina. They extend out to the retinal periphery by E16, and reach their peak neurogenesis on E17. On E20 they can still be found at all retinal loci, albeit at reduced density, and by E22 they are found primarily in the ventral retina, comprising the only tritium-positive cells in the ganglion cell layer to have been generated on this day. This developmental increase in the genesis of presumptive displaced amacrine cells relative to ganglion cells is shown in Fig. 6. While it is clear that the population of retinal ganglion cells cease neurogenesis before the population of presumptive displaced amacrine ceils, and that they occupy a progressively smaller fraction of recently generated cells as development proceeds (Fig. 6), the neurogenetic periods for the two populations overlap substantially. All of the contralateral retinae displayed heavy HRP labelling across the complete retinal surface, so we do not believe lack of HRP labelling to be a consequence of inadequate uptake from a portion of the target visual nuclei. Further, the largest neurons in the ganglion cell layer were never HRP-negative in the retina contralateral to the injections (except in the temporal retina, e.g. Fig. lb), supporting the view that no portion of the retinotopic map had escaped labelling. The present results support the observations of Harman and Beazley,” in the wallaby, that ganglion cell genesis and displaced amacrine cell genesis are largely overlapping in time, rather than occurring in a primarily sequential fashion. Nevertheless, it could be argued that smaller retinal ganglion cells may be more difficult to label retrogradely, and that all of these tritium-positive, HRPnegative cells are in fact ganglion cells which had
R.J. COLELLO failed to be labelled. While an unlikely account for all of the tritium-positive, HRP-negative neurons, this may explain the higher proportion of displaced amacrine cells generated on El4 relative to the trend across the remaining days, shown in the graph of Fig. 6: if a few small retinal ganglion cells had not displayed detectable HRP labelling on E14, they would have been misclassified as displaced amacrine cells, thus incorrectly raising the ratio for this retina. An alternative approach would be to use a label which positively identifies displaced amacrine cells, or at least a sub-population of them. We have used a monoclonal antibody directed against ChAT to label one distinct class of displaced amacrine cell, the cholinergic amacrine cell, 32in retinae that had been exposed to tritiated thymidine on different embryonic days. ChAT-labelled retinal whole-mounts contained immunoreactive somas in two distinct retinal layers: one population was situated in the retinal ganglion cell layer (Fig. 7a), and the other, of similar density, was situated along the inner margin of the inner nuclear layer (Fig. 7b). Between each population were two thin sublaminae within the inner plexiform layer, containing the immunoreactive processes of the two respective populations, as described in the literature.‘ss32 Tritium-positive and ChAT-positive doublelabelled amacrine cells in the ganglion cell layer were observed in the retinae of rats that had been exposed to tritiated thymidine on E16-E20. Examples of tritium-positive and ChAT-positive cells in the ganglion cell layer are shown in Fig. 7d-f, from rats that had been exposed to [‘Hlthymidine on E14, El7 and E20, respectively, the El 7 sample including double-labelled cells (arrowheads). A single double-
temporal
nasal E-20
Fig.
8. Distributions
of double-labelled generated
tritium-positive) (ChAT-positive, on different days of gestation.
displaced
amacrine
cells
Neurogenesis in the rat’s retinal ganglion cell layer
labelled cell was observed in each of two rats exposed to [3H]thymidine on El5 The incidence of doublelabelled cells increased and peaked on El7 and E18, and then tapered off until E20. No double-la~lled cells were observed on E22. A similar time course for the genesis of ChAT-positive amacrine cells in the inner nuclear layer was alSO observed. As with the other populations of neurons in the ganglion cell layer, the topography of neurogenesis of the ChAT-positive ceils showed similar centroperiphera1 and dorsoventral trends (Fig. 8). The earliest generated cells were found in the central retina on E16, while at progressively later embryonic stages, neurogenesis of the ChAT-positive cells occurred in the peripheral and ventral retinal regions. On E19, double-labelled cells are most numerous in the ventral periphery, although there are a few found elsewhere across the retina. On E20, the last day of neurogenesis of the ChAT-positive cells, the few double-labelled cells are restricted to the ventral periphery. We have examined a total of 36 retinae in this study, and are confident about the neurogenetic period of the ChAT-positive cells. We are slightly less secure about the topographic information contained in Fig. 8 because although we have consistently obtained specific ChAT-positive labelling, we have not always obtained adequate labelling across the full extent of the ganglion cell layer in every retina. Consequently, while we are certain about the identification of double-labelled cells (e.g. those indicated in Fig. 8), regions of retina displaying no doublelabelled cells may be due to poor ChAT labelling. The four retinae shown in Fig. 8 contained the most thorough coverage of ChAT labelling across the ganglion cell layer, with only the El9 retina displaying a large portion of its dorsonasal region lacking ChAT labelling. DISCUSSION
The present study has demonstrated that the periods of genesis of one type of retinal ganglion cell, the alpha cell, and one type of displaced amacrine cell, the cholinergic cell, are limited in their durations relative to the genetic periods of the total populations of ganglion and displaced amacrine cells in the ganglion cell layer. This study also demonstrates that the neurogenetic periods of the populations of ganglion and displaced amacrine cells are largely overlapping, rather than being sequential. Displaced amacrime cell genesis and ganglion ceN genesis
Displaced amacrine cells in the rat’s retinal ganglion cell layer, identified by the absence of HRP labelling, are generated throughout the period of ganglion cell genesis, rather than following ganglion cell genesis. While the failure to label retrogradely some Optic axons may account for the lack of HRP labelling in a proportion of neurons (as should be
427
expected in the temporal retina), this seems an unlikely account for all of the tritium-positive, HRPnegative neurons in the nasal portion of the ganglion ccl layer. Further, our demonstration that one class of positively identified displaced amacrine cellsthe ChAT-immunoreactive cells-is generated wholly during the period of ganglion cell genesis shows conclusively that ganglion and displaced amacrine cell genesis substantially overlap in time and retinal locus. An identical temporal gradient was recently described for displaced amacrine cells in the rat which are immunoreactive for corticotropin releasing factor (CRF).@ Both this population of displaced amacrine cells, and the displaced cholinergic amacrine cells, have matching populations of cells in the inner nuclear layer.32,40 Those immunohistochemi~lly identified populations of amacrine cells in the inner nuclear layer are generated during the identical time period (Ref. 40 and present results), along with dop~ine-containing (DA) amacrine cells in the inner nuclear layer (E16E20).7 They may account for only a small proportion of the total population of amacrine cells, and future identification and birthdating of other types of amacrine cell may reveal displaced amacrine cell classes generated as early as El4 or as late as E22. The observation that displaced amacrine cells and ganglion cells are generated cont~poraneously is not incompatible with the late migration of displaced amacrine cells into the ganglion cell layer.” Zimmerman et aL4’ have shown that horizontal cells and ganglion cells which are generated at the same time do not necessarily migrate simultaneously, and the same may be true for the displaced amacrine cells, leading to their delayed arrival within the ganglion cell layer. Gang/ion ceil genesis andjibre order in the optic tract
Ganglion cell classes in the cat’s retina are generated as independent though temporally overlapping gradients in which cells in the central retina are, on average, generated before cells in the peripheral retina.34,35Most of the beta cells are generated before alpha cells, and much of the gamma cell population is generated at the end of the period of ganglion cell genesis, after the alpha cells. This temporai pattern of genesis is reflected in the adult organization of the cat’s optic tract, in which the beta axons are positioned deepest in the tract, alpha axons are situated superficial to the beta axons, and a very large number of the gamma axons are found sub-pially, at the tract’s superficial surface. *J’ Fibre order in the optic tract, then, is a chronological index of axonal arrival during development: the oldest axons are deepest in the tract, younger axons being positioned at progressively superficial positions.9~33 The adult rat’s optic tract likewise contains a segregation of axon classes, with some interesting differences from the fibre order found in the cat. First,
428
B. E.
REESE
and R. J.
the coarsest axons in the rat’s optic tract, believed to arise from the alpha cells, course through the very deepest parts of the tract,23 whereas in the cat, the beta axons separate the alpha axons from this border. Second, the population of uncrossed optic axons is densest in the deeper parts of the tract, and is scarce or absent along the pial surface,z4 whereas uncrossed optic axons in the cat can be found throughout most of the tract, including its sub-piai region.28 These two features of the rat’s optic tract led to two novel predictions about retinal ganglion cell genesis: (1) alpha cells should be generated from the very outset of ganglion ceil genesis, rather than at a relatively late stage, and (2) ipsiiaterally projecting retinal ganglion ceils should complete their genesis before the population of contralaterally projecting ganglion cells. The results of the present study confirm these two predictions. First, alpha ceils are generated during the first half of the total period of ganglion cell genesis, from embryonic day 14 through to day 17, while other ganglion cell classes continue to be generated thereafter, until embryonic day 20. No other cell type initiates its neurogenetic period prior to the alpha cells. And second, but for a very few cells born on Ei9, the population of uncrossed ganglion cells is almost completely generated by E 18, two days before the cessation of crossed ganglion cell genesis. As in carnivores then, an axon’s position in the rat’s optic tract reflects the relative time of genesis of its parent retinal ganglion ceil (see also Ref. 3). Ganglion cell genesis and the line of decussation
The present study replicates, in the rat, the observations by DrPgerS in the mouse that the earliest ceils
COLELLO
to be generated in the temporal crescent have uncrossed optic axons, while the latest cells to be generated in the crescent have crossed axons. These results are consistent with the observations of Bunt et al.’ and of Wikler et aL3’ that the crossed projection from temporal retina is a relatively late-developing feature of the rodent’s optic pathway. While neither the present study nor the study by Drgge? addresses the role of ceil death in creating this difference in the neurogenesis of temporal retina, in conjunction with these other studies,‘,3s they suggest that early generated cells from the temporal retina navigate into the ipsilaterai tract exclusively, whereas later generated cells show an increasing propensity to navigate a crossed course through the optic ehiasm, as in carnivores.25,28,35 Finally, it is interesting to note the discrepancy between the centropetipheral gradient of alpha ceil genesis and the distribution of alpha cells in adulthood. Like the population of non-aipha ceils, the alpha ceils initiate their genesis in dorsocentral retina; yet the peak density of the former ceils in the mature retina is situated dorsocentrally, whereas the peak density of the alpha cells is shifted into the far dorsotemporai periphery, upon the line of decussation.“a.‘7.27
acknowledgements-his research was supported by a ghnt from the Medical Research Council of the United Kingdom (PG 8324037). The experiments on animals were conducted under licensing and certification by the Home Office of the U.K. We thank Sue Mygal for her animal husbandry, Davina Hocking, Maurice Sanders, Mary Walker and especially Zillah Deusson for their assistance with histology, Gigi Raott! for her contribution to the analysis, and Gary Baker, John Mitrofanis and Ray Guillery for comments on the manuscript.
REFERENCES
1. Bunt S. M., Lund R. D. and Land P. W. (1983) Prenatal development of the optic projection in albino and hooded rats. Lteul Brain Res. 6, 149-168. 2. Colello R. J. (1988) Modifications of a method for preparing retinal wholemounts for sectioning. Stain Technol. 63,
183-188. 3 Colello R. J. and Guillery R. W. (1991) Observations on the early development of the optic nerve and tract of the mouse. .I. camp. NeuroZ.(in press). 4. Cowey A. and Perry V. H. (1979) The projection of the temporal retina in rats, studied by retrograde transport of horseradish peroxidase. Expi Brain Res. 35, 457464. 5. Driiger U. (1985) Birth dates of retinal ganglion cells giving rise to the crossed and uncrossed optic projections in the mouse. Proc. R. Sot. Land. B22q 57-77. 6. Driiger U. and Hofbauer A. (1984) Antibodies to heavy neurofilament subunit detect a subpopulation of damaged ganglion cells in retina. Nature 309, 624-626. . .. < .* 6a. Dreher B., Sefton A. J., Ni S. Y. K. and Nisbett G. (1985) The morphology, number, &stnl)utlon ana cemral projections of class I retinal ganglion cells in albino and hooded rats. Brain, Behao. Evol. 26, l&48. 7. Evans J. A. and Battelle B.-A. (1987) Histogenesis of dopamine-containing neurons in the rat retina. Expl Eye Res. 44, 407-414. 8. Guillery R. W., Polley E. H. and Torrealba F. (1982) The arrangement of axons according to fiber diameter in the optic tract of the cat. J. Neurosci. 2, 714-721. 9. Guillery R. W. and Walsh C. (1987) Changing glial organization relates to changing fiber order in the developing optic nerve of ferrets. J. cotnp. Neural. 265, 203-217. 10. Hanker J. S., Yates P. E., Metz C. B. and Rustioni A. J. (1977) A new, specific, and non-carcinogenic reagent for the demonstration of horseradish peroxidase. ~~stoche~. J. 9, 789-792. 11. Harman A. M. and Beaziey L. D. (1989) Generation of retinal cells in the wallaby, Sefonix brachyurus (quokka). Neuroscience 28, 219-232. 12. Jenkins E. C. (1972) Wire loop application of liquid emulsion to slides for autoradiography Technol. 47, zj-26.
in light microscopy. Stain
Neurogenesis
in the rat’s retinal
ganglion
cell layer
429
13. Linden R. and Esberard C. E. L. (1987) Displaced amacrine cells in the ganglion cell layer of the hamster retina. Vision Res. 27, 1071-1076. projections in rats. Brain Res. 272, 145-149. 14. Linden R. and Perry V. H. (1983) Massive retinotectal J., Maslim J. and Stone J. (1988) Catecholaminergic and cholinergic neurons in the developing retina of 15. Mitrofanis the rat. J. camp. Neural. 276, 343-359. J., Maslim J. and Stone J. (1989) Ontogeny of catecholaminergic and cholinergic cell distributions in the 16. Mitrofanis cat’s retina. J. camp. Neurol. zs9, 2288246. 17 Peichl L. (1989) Alpha and delta ganglion cells in the rat retina. J. camp. Neural. 286, 12&139. 18. Perry V. H. (1979) The ganglion cell layer of the retina of the rat: a Golgi study. Proc. R. Sot. Lond. B204, 363-375. 19 Perry V. H. (198 1) Evidence for an amacrine cell system in the ganglion cell layer of the rat retina. Neuroscience 5, 931-944. 20 Perry V. H., Henderson Z. and Linden R. (1983) Postnatal changes in retinal ganglion cell and optic axon populations in the pigmented rat. J. camp. Neural. 219, 356368. 21 Perry V. H. and Walker M. (1980) Amacrine cells, displaced amacrine cells and interplexiform cells in the retina of the rat. Proc. R. Sot. Lond. B208, 415431. 37 Polley E. H., Zimmerman R. P. and Fortney R. L. (1989) Neurogenesis and maturation of cell morphology in the development of the mammalian retina. In Deuelooment of the Vertebrate Retina (eds Finlav B. L. and Sengelaub D. R.), pp. 3-29. Plenum Press, New York. ‘ ” of axons according to diameter in the optic nerve and optic tract of the rat. 23 Reese B. E. (1987) The distribution Neuroscience 22, 1015-1024. 24 Reese B. E. (1987) The position of the crossed and uncrossed optic axons, and the non-optic axons, in the optic tract of the rat. Neuroscience 22, 102551039. 25 Reese B. E. and Baker G. E. (1990) The course of fibre diameter classes through the chiasmatic region in the ferret. Eur. J. Neurosci. 2, 3449. W. (1991) Cell survival in the mammalian retina’s ipsilateral projection is 26. Reese B. E., Colello R. J. and Thompson independent of birthdate. ARVO Abstr. 13, 925. 27 Reese B. E. and Cowey A. (1986) Large retinal ganglion cells in the rat: their distribution and laterality of projection. Expl Brain Res. 61, 375-385. 28 Reese B. E., Guillery R. W., Marzi C. A. and Tassinari G. (1991) The position of axons in the cat’s optic tract in relation to their retinal origin and their chiasmatic pathway. J. cony. Neural. 306, 539-553. 29. Sengelaub D. R., Dolan R. P. and Finlay B. L. (1986) Cell generation, death, and retinal growth in the development of the hamster retinal ganglion cell layer. J. camp. Neural. 246, 527-543. of mouse retina studied with thymidine-‘H. In Structure of the Eye (ed. Sm.&or 30. Sidman R. L. (1961) Histogenesis G. K.), pp. 487-506. Academic Press, New York. 31. Torrealba F., Guillery R. W., Eysel U., Polley E. H. and Mason C. A. (1982) Studies of retinal representations within the cat’s optic tract. J. camp. Neural. 211, 3777396. 32. Voigt T. (1986) Cholinergic amacrine cells in the rat retina. J. camp. Neural. 248, 19-35. 33. Walsh C. and Guillery R. W. (1985) Age-related fiber order in the optic tract of the ferret. J. Neurosci. 5, 3061-3069. 34. Walsh C. and Polley E. H. (1985) The topography of ganglion cell production in the cat’s retina. J. Neurosci. 5, 741-750. 35. Walsh C., Polley E. H., Hickey T. L. and Guillery R. W. (1983) Generation of cat retinal ganglion cells in relation to central pathways. Nature 302, 61 l-614. 36. Wassle H., Chun M. H. and Miiller F. (1987) Amacrine cells in the ganglion cell layer of the cat retina. J. camp. Neural. 265, 391-408. 37. Wikler K. C., Perez G. and Finlay B. L. (1989) Duration of retinogenesis: its relationship to retinal organization in cricetine rodents. J. camp. Neural. 285, 1577176. 38. Wikler K. C., Raabe J. I. and Finlay B. L. (1985) Temporal retina is preferentially represented in the early retinotectal projection in the hamster. Devl Brain Res. 21, 1522155. 39. Wong R. 0. L. and Hughes A. (1987) The morphology, number, and distribution of a large population of confirmed displaced amacrine cells in the adult cat retina. J. camp. Neural. 255, 159-177. 40. Zhang D. and Yeh H. H. (1990) Histogenesis of corticotropin releasing factor-like immunoreactive amacrine cells in the rat retina. Deul Brain Res. 53, 194199. 41. Zimmerman R. P., Policy E. H. and Fortney R. L. (1988) Cell birthdays and rate of differentiation of ganglion and horizontal cells of the developing cat’s retina. J. camp. Neural. 274, 77-90. A&.
(Accepted 26 June 1991)
NeuroscienceVol. 46, No. 2, pp. 419429, 1992 Printed in Great Britain
NEUROGENESIS
IN THE RETINAL GANGLION LAYER OF THE RAT
CELL
B. E. REP.sE*t$ and R. J. COLELLO$§ tNeuroscience
Research Institute and Department of Psychology, University of California at Santa Barbara, U.S.A. $Department of Human Anatomy, University of Oxford, U.K.
Abstract-The present study has examined the birthdates of neurons in the retinal ganglion cell layer of the adult rat. Rat fetuses were exposed to tritiated thymidine in urero to label neurons departing the mitotic cycle at different gestational stages from embryonic days 12 through to 22. Upon reaching adulthood, rats were either given unilateral injections of horseradish peroxidase into target visual nuclei in order to discriminate (1) ganglion cells from displaced amacrine cells, (2) decussating from non-decussating ganglion cells, and (3) alpha cells from other ganglion cell types; or, their retinae were immunohistochemically processed to reveal the choline acetyltransferase-immunoreactive amacrine cells in the ganglion cell layer. Retinae were embedded flat in resin and cut enjim to enable reconstruction of the distribution of labelled cells. Retinal sections were autoradiographically processed and then examined for neurons that were both tritium-positive and either horseradish peroxidase-positive or choline acetyltransferase-positive. Tritium-positive neurons in the ganglion cell layer were present in rats that had been exposed to tritiated
thymidine on embryonic days ElkE22. Retinal ganglion cells were generated between El4 and E20, the ipsilaterally projecting ganglion cells ceasing their neurogenesis a full day before the contralaterally projecting ganglion cells. Alpha cells were generated from the very outset of retinal ganglion cell genesis, at E14, but completed their neurogenesis before the other cell types, by E17. Tritium-positive, horseradish peroxidase-negative neurons in the ganglion cell layer were present from El4 through to E22, and are interpreted as displaced amacrine cells. Choline acetyltransferase-positive displaced amacrine cells were generated between El6 and E20. Individual cell types showed a rough centroperipheral neurogenetic gradient, with the dorsal half of the retina slightly preceding the ventral half. These results demonstrate, first, that retinal ganglion cell genesis and displaced amacrine cell genesis overlap substantially in time. They do not occur sequentially, as has been commonly assumed. Second, they demonstrate that the alpha cell population of retinal ganglion cells and the choline acetyltransferaseimmunoreactive population of displaced amacrine cells are each generated over a limited time during the periods of overall ganglion cell and displaced amacrine cell genesis, respectively. Third, they show that the very earliest ganglion cells to be generated in the temporal retina have exclusively uncrossed optic axons, while the later cells to be generated therein have an increasing propensity to navigate a crossed chiasmatic course.
During retinal development in the rodent, neurons in the ganglion cell layer are said to be generated in a rough centroperipheral gradient, with large neurons being generated before small neurons at any retinal locus.5~29~30~37 The larger neurons are said to be ganglion cells, while the smaller neurons are considered to be displaced amacrine cells.29~37Such a precedence of ganglion cell genesis over displaced amacrine cell genesis has also been proposed for the cat.** In the rat, displaced amacrine cells migrate into the ganglion cell layer relatively late during retinal development,*’ consistent with their postulated late genesis. Soma size differences are also evident between the different retinal ganglion cell classes, and there is *To whom correspondence should be addressed at: the Neuroscience Research Institute, University of California, Santa Barbara, CA 93106, U.S.A. §Present address: Institute for Brain Research, University of Zurich, Switzerland. ChAT, choline acetyltransferase; CRF, corticotropin releasing factor; DA, dopamine; DAB, diaminobenzidine; E, embryonic day; HRP, horseradish peroxidase.
Abbreviations:
419
good evidence that the soma size of some retinal ganglion cell classes overlaps with displaced amacrine cell classes 13,15,18,19,21; se also 16.36339 Consequently, soma size is clearly an inadequate basis for a thorough discrimination of ganglion from displaced amacrine cells. Of the above studies, only two have employed the retrograde transport of horseradish peroxidase (HRP) as a means of positively identifying retinal ganglion cells following large injections of the tracer into visual nuclei;51” if the injections label the complete population of retinal ganglion cells, displaced amacrine cells within the ganglion cell layer can be identified by the absence of the label. Using this approach, Harman and Beazley” report that displaced amacrine cell genesis in the wallaby is contemporaneous with ganglion cell genesis, rather than sequential. In the cat, the different retinal ganglion cell classes, which are more readily differentiated according to their size, proliferate independently, though partially overlapping in time, with central retina preceding peripheral retina for any cell class.34,35The morphologically distinct ganglion cell classes of rodents may also be generated independently, but because there is
420
3. E.
REESE and
substantial overlap between their soma size distributions, and because of the difficulty in discriminating displaced amacrine from ganglion cells, the available evidence does not resolve the pattern of neurogenesis in the ganglion cell layer of rodents. In the mouse, genesis of retinal ganglion cells has been reported to proceed as one gradual centroperipheral wave,5 in which the temporally independent genesis of different cell classes may be concealed (e.g. Ref. 6). In order to resolve these issues, the present study has investigated the genesis of neurons in the retinal ganglion cell layer of the rat. We have made large, multiple, unilateral injections of HRP into the visual nuclei of adult rats which had been exposed to tritiated thymidine in utero on a specific embryonic day. We have used the presence of the HRP to positively identify axon-bearing neurons (ganglion cells) within the ganglion cell layer. Further, we have used the HRP to permit the identification of one type of retinal ganglion cell, the alpha cell,” in order to define its neurogenetic period and its topography of genesis relative to other HRP-positive cells. We have also used the presence of the tracer to determine the relative neurogenetic periods and the topography of genesis of the decussating from non-decussating retinal ganglion cells. As in the study by Harman and Beazley,” the absence of HRP labelling may be regarded as an indication that a given neuron in the ganglion cell layer is an axon-less interneuron (a displaced amacrine cell). We have therefore addressed the issue of whether displaced amacrine cell genesis follows, or coincides with, ganglion cell genesis. Because one may question whether the lack of HRP labelling is a reliable indication of displaced amacrine cells, we have also used an antibody to choline acetyltransferase (ChAT) in iittermates which had been exposed to tritiated thymidine in utero in order to positively identify one class of displaced amacrine cell within the ganglion cell layer.32
R. J.
COLELLO
ing. The latter caiendar day is called embryonic day 1 (El), the first day of gestation. Pregnant dams were anaesthetized with ether and given intra-peritoneal injections of 5 &i of [methy~,l’,2’-3H]thymidine (Amersham International; 126Ci/mmol) per gram of maternal body weight on the 12th. 13th, lrlth, 15th, l&h, 17th, l&h, 19th, 20th or 22nd day of gestation. Parturition routinely occurred on E23. Pups were reared by the dams until weaning, after which they were group-caged until adulthood. They were then used in one of two experiments, using HRP to reveal retinal ganglion cells, or using an antibody directed against ChAT to reveal this class of displaced amacrine cell (Table 1).
Rats were anaesthetized with a mixture of sodium pentobarbitone and chloral hydrate (30 mg/kg c L26mg!kg, respectively), placed in a stereotaxic headhoider, and then given six injections of 0.1 ~1 of 40% HRP dissolved in 2% dimethyl suiphoxide. The injections were placed stereotaxitally to include the optic tract and target optic nuclei of one hemisphere. Two days later, rats were heavily anaesthetizcd and returned to the headholder. and the dorsal margin of the limbus was then cauterized for later orientation of the retina. Rats were perfused through the heart with 50 ml of 0.9% saline followed by 100 ml of 1% paraformaldehyde in 0.1 M sodium phosphate buffer, pH 7.2, at 4°C. Eyes were removed immediately and a iarge radial cut was made extending from the dorsal margin of the retina to the optic disk following removal of the anterior chamber and lens. Retinae were dissected, mounted flat and postfixed for 1 h in 2.5% glutaraldehyde in buffer at 4°C. They were then rinsed in buffer and reacted for enzymatic activity using p-phenylenediamine + catechoi as the chromogen.” The rat was further perfused with 2OOml of 1.25% paraformaldehyde + 2.5% glutaraldehyde in buffer at 20°C. Brains were cut at 25 pm in the coronal plane and every fifth section was collected and hist~~emically and autoradiographical~y processed. Further details of the surgery, perfusion and histochemistry can be found elsewhere.27 Retinae were dehydrated and embedded flat in LR White resin2 and then cut in the plane of the retinal surface at 5 pm. Retinal sections were mounted from warm acetone onto slides, and then dried, coated with Ilford K-2 emulsion using the wire loop technique, I2stored in the dark at 4°C for t 2 weeks. develoned in Kodak D-19, stained with Methylene Blue and Azur i1 and then coverslipped. Retinae ipsilateral to the HRP injections were examined with a x 100 oil immersion objective for the presence of double-labelled neurons (cells which were both HRP-positive and heavily tritium-positive). The criterion for defining heavily labelled tritium-positive cells (henceforth “tritiumpositive”) was 50%. or greater, of the number of silver grains observed overlying the average of the three most heavily labelled cells in each retina. Retinae contralateral to the HRP injections were examined for the presence of tritium-positive cells, and were recorded as being either HRP-positive or HRP-negative. All double-Ia~lled cells in
Hooded rats of the Long-Evans strain, obtained from the breeding colony maintained in the Department of Experimental Psychology, Oxford, were used. Males were put with females from 400 p.m. until 9:00 a.m. the following mom-
Table 1. Number of retinae examined from rats that had been exposed to tritiated thymidine on different embryonic days Cohort age El2 Ef3 E14 El5 Ef6 El7 Et8 Ei9 B20
ChAT retinae 0
3 5 4 2 6 6 3 2
Ipsi HRP retinae 2 2 2 2 2 2 3 2 2
Contra HRP retinae 2 2 2 2 2 2 2
I ?8.
Neurogenesis in the rat’s retinal ganglion cell layer either retina were identified as being alpha or non-alpha (e.g. Fig. 1), based on the presence of a large soma and stout primary dendrites in the same 5 pm section.‘7*27 The distribution of tritium-positive cells was plotted on drawings of the retinal sections at a magnification of x 20 or x 26.5 with the aid of a drawing tube attached to the microscope. The complete retinal ganglion cell layer was then reconstructed using the pattern of blood vessels and retinal contours to align successive sections, as described by Walsh and Polley.”
have been seen in the ventral
421 retina of the ipsilateral
eye. This was never observed.
Tritium-positive cells were found in the retinal ganglion cell layer of rats that had been exposed to tritiated thymidine on embryonic days 14-22. None of the rats that had been exposed to tritiated thymidine on El2 or El3 in the HRP studies contained triti~-positive cells in the retina, despite detectable labelling within the brainstem of these very Choliw acetyitransferase experiments same animals. (One littermate from the ChAT study Rats were heavily anaesthetized, placed in the headholder, which had been exposed to tritiated thymidine on E 13 and the dorsal margin of the limbus of each eye was did contain a number of tritium-positive cells, none cauterized for later orientation. Rats were then removed of them double-labelled, in the central region of from the headholder and perfused through the heart with the retinal ganglion cell layer, presumably indicative 50 ml of 0.9% saline in 0.1 M sodium phosphate buffer of the variable maturity between fetuses. We shall fnH 7.2). followed bv 50 ml of either 1% or 2% ~arafo~ldehyde in thesame, at 4°C. Eyes were enucleated regard El4 as the beginning of neurogenesis for imm~iately and following removal of the anterior chamber neurons in the ganglion cell layer.) We were, of and lens, a large radial cut was made extending from the course, examining the retinal ganglion cell layer in the dorsal margin of the retina to the optic disk. Retinae were adult, and all of our observations concerning neurothen dissected, mounted flat and postfixed in 4% paraformaldehyde plus 0.05% glutaraldehyde in buffer at genesis are specific for those cells in the ganglion cell 4°C for 1 h. They were then left to sink in 30% sucrose at layer which survive into adulthood. We shall call the 4”C, and rinsed in a minimum of three changes of buffer population of tritium-positive cells from a rat that overnight. had received tritated thymidine on a given day a Retinae were shock-frozen in liquid nitrogen, rinsed in buffer (3 x 10 mm), and then incubated for 16 h at 4°C in rat “cohort” (e.g. an El6 cohort). monoclonal anti-ChAT antibody (Boeh~nger) at a dilution Cells in the retinal ganglion cell layer which were of 1: 10 dissolved in 0.04 M Tris-phosphate buffer (pH 7.8), positive for both tritiated thymidine and for HRP containing 0.2% carrageenan and 0.5% Triton X-100. They were found in rats that had been exposed to tritiated were then rinsed in several changes of Tris-phosphate buffer thymidine on El4E20 (e.g. Fig. Id). Ganglion cell for 1 h, and incubated for 1 h in rabbit anti-rat IgG (Sigma) at a dilution of 1: 50 in buffer (plus carrageenan and Triton, genesis begins in the dorsocentral retina, then spreads as above), at 20°C. After a rinse in buffer, retinae were to cover the entire retina on E16, and by E20 can be incubated in rat peroxidase-antiperoxidase (ICN Immunofound almost exclusively in the far ventral periphery. biologicals) at a dilution of 1: 75 at 20°C for I h. They were No double-labelled cells were found in rats that then washed in several changes of 0.1 M sodium phosphate buffer (pH 7.2, 20°C) for I h, and finally reacted for received injections of tritiated th~idine on E22. enzymatic activity using diamino~nzi~ne as the chroThis topographic pattern of neurogenesis across mogen (0.2% DAB plus O.Ot% H,O& Retinae were rinsed the dorsocentral to ventroperipheral retina conceals in phosphate buffer, dehydrated, embedded flat in LR White at least one independent neurogenetic gradient of a resin, sectioned at 5 pm in the plane tangential to the retinal specific retinal ganglion cell class, the alpha cell class. surface, and then mounted, processed for autoradiography and stained as above. The distribution of double-labelled Tritium-positive, HRP-positive cells were classed as cells (tritium-positive and ChAT-positive) was plotted on being alpha (e.g. arrowheads in Fig. 1) or non-alpha, reconstructed flat-mounts of the retinae, as above. and plotted separately in Figs 2-5. Consider first the genesis of the non-alpha retinal ganglion cells, summarized in the data from three cohorts for both the RESULTS contralateral (Fig. 2) and ipsilateral (Fig. 3) retinae. Genesis of the contralateral cells begins on El4 in a Genesis of retinal ganglion ceils broad zone of the dorsocentral retina (Fig. 2). On The large injections of HRP produced dense El.S, there is less of an apparent focus dorsocentrally, labelling across the complete contralateral retinal yet the distribution of double-labelled cells has not surface, and throughout the temporal crescent of the extended to the retinal periphery. Ganglion cell genipsilateral retina. The density of retrogradely labelled esis peaks on El6 and E17, and the complete retinal cells in the temporal crescent of the ipsilateral retina surface contains double-la~lled cells, with no clearly was about one quarter of that in the temporal discernible region of focus (Fig. 2). By E19, ganglion crescent of the contralateral retina, and the majority cell genesis is focused in the ventral half of the retina, of the cells were very heavily labelled. There was no and on E20, the last day to include double-labelled evidence for spread of the HRP across the midline, as cells, they are almost completely restricted to the judged from the histological sections through the ventral periphery (Fig. 2). brain, nor for the uptake of HRP by optic terminals Genesis of the non-alpha ipsiiateral cells (Fig. 3) in the visual nuclei opposite the side of the injection. likewise begins on El4 in the dorsocentral portion of Had the latter event occurred, particularly at the the temporal crescent, that part of the retina giving collicular midline, a blanket of lightly labelled cells, rise to the uncrossed projection.4 It spreads to further equal to the density of the crossed projection, should peripheral and ventral regions of the crescent by E17,
422
B.E.
REESE and
R.J.
COLELLO
Fig. I. Photomicrographs of HRP-labelled ganglion cells in retinal whole-mounts (a, c) and m tangential
sections(b, dj contralateral (a: b) and ipsilateral (c, d) to the injected hemisphere, Only the ganglion cells are visible in the whole-mounts (a, c), whereas the tangential sections of the retinal ganglion cell layer (b, d) have been subsequently processed for autoradiography and then stained with Methylene Blue and Azur II. Notice the large somata and radiate dendritic arbors characteristic of the alpha cells in the whole-mounts (arrowheads in a and c), and the co-localization of the tritiated thymidine label and the HRP label in one of the two alpha cells in d (arrowhead). a and c come from nasal and temporal retinae, respectively, while b and d both come from temporal retina. Notice the large, HRP-negative cell from the contralateral retina’s temporai crescent in b (arrowhead), presumably in an alpha cell projecting to the ipsilateral hemi~he~. b and d come from a rat that had been exposedto tritiated thymidine on El 5. Scale bar in c = 30pm for a and c, and = 20hm for b and d.
and a few ectopic, nasal cells are included in that cohort (Fig. 3). On E18, it is restricted to the most peripheral extreme of the crescent, and only a very few double-Ia~lled cells can be found on El9 (Fig. 3). No double-labelled cells were observed in rats that had received injections of tritiated thymidine on E20 (see also Ref. 26 for a more detailed description of the neurogenesis of the uncrossed projection).
In contrast, alpha cells are generated over a rela-
tively restricted portion of the period during which ganglion cells are being generated. This is shown jn Figs 4 and 5, again for the contralatera1 and ipsilatera1 retinae, respectively. Contralaterally projecting alpha cells are first generated dorsocentrally on E14; over the next three days, their topogenesis spreads to occupy progressively peripheral retinal loci so that by
423
Neurogenesis in the rat’s retinal ganglion cell layer E-20
E-17
temporal
Fig. 2. Distributions of double-la~li~ (HRP-positive, tritium-~sitive) non-alpha retina1 ganglion cells with crossed optic axons generated on different days of gestation. Dorsal is up in this and in all sub~quent retinal reconstructions. Each dot represents an individual cell.
El7 they are almost exclusively found in peripheral retina (Fig. 4). Ipsilaterally projecting alpha cells are generated during the same four day period, displaying similar dorsal before ventral, and central before peripheral gradients (Fig. 5). The lack of dorsal alpha cells in the
temporal crescent on El5 is surmising, since they are clearly present there on the preceding and subsequent days. Nevertheless, the uncrossed alpha cell population displays, to a rough approximation, the same topographic pattern of genesis as the non-alpha population, although restricted to a briefer duration.
E-17
,:_ .’ .: ;. .;,: 7..
temporal
@
nasal
:,..‘. .: 2. ?xt 8-y.‘.&.,&>..
Fig. 3. Distributions of double-labelled (HRP-positive, tritium-positive) non-alpha retinal ganglion cells with uncrossed optic axons generated on different days of gestation.
temporal
nasal
E-16
E-l?
Fig. 4. Distributions of double-labelled (HRP-positive, tritium-positive) alpha retinal ganglion crossed optic axons generated on different days of gestation.
cells
with
424
REESEand
B. E. E-14
R. J.
COLELLO E-15
temporal
Fig.
i
5.
Distributions of double-labelled (HRP-positive, tritium-positive) alpha retinal ganglion cells wtth uncrossed optic axons generated on different days of gestation.
0.6 -
‘-
P
d+
0.6 -
E ‘E $ $ 0.4 = $ z ._ $j 0.2 E
01
’
14
’
15
’
16
’
17
’
18
’
19
’
20
’
21
1
22
Embryonicday of 3H-thyexposure Fig. 6. Proportion of neurons in the ganghon cell layer generated on different embryonic days which were presumptive displaced amacrine cells. A ratio of 0.0 indicates that HRP-positive (ganglion) cells are generated exclusively, while a ratio of 1.0 indicates that HRP-negative (displaced amacrine) cells are generated exclusively. Six adjacent square millimetres were sampled from retina nasal and ventral to the optic disk, at eccentricities less than 50% of the retinal radius (and therefore outside the temporal crescent).
The large injections of HRP that we have employed were meant to retrogradely label all of the ganglion cells projecting to one hemisphere. Such injections have been estimated to label the large majority of such cellsi The number of unlabelled cells following such injections is also coincident with the number of neurons remaining in the retinal ganglion cell layer following transection of the optic nerve-the presumed displaced amacrine cell population.” Consequently, those cells which do not show HRP labelling in the present study should correspond to the population of displaced amacrine cells in the retina, with one exception: ganglion cells in the temporal crescent may project to either hemisphere, so lack of HRP labelling is not a reliable index of a displaced amacrine cell in this region (e.g. Fig. 1b). In the remainder of the retina contralateral to the injected hemisphere (that is, in nasal retina), HRP-negative cells in the ganglion cell layer should all be displaced amacrine cells. In every age group from El4 through to E22, we have observed tritium-positive, HRP-negative neurons in the nasal portion of the retinal ganglion cell layer. These cells comprise a small proportion of
Fig. 7. Photomicrographs of ChAT-immunoreactive amacrine cells in the retinal ganglion cell layer (a, c-f) and in the inner nuclear layer (b) in retinal whole-mounts (a, b) and in tangential sections through to the retinal ganglion cell layer (c-f). d-f are focused on the plane of the silver grains, showing the tritium label in the ganglion cell layer of rats exposed to [‘Hlthymidine on embryonic days 14, 17 and 20, respectively. Three tritium-positive, ChAT-positive double-labelled cells are indicated in e (arrowheads), but the silver grains are not readily apparent against the immunoreaction product in this reproduction. a-c have been photographed with Nomarski optics, and c is the same field as f, focused on the somata. Scale bar in f = 30 11m for a and h, and = 20 pm for c f.
Neurogenesis in the rat’s retinal ganglion cell layer
Fig. 7.
425
426
B. E. REESE and
the tritium-positive cells initially (Fig. 6), and are found only in the central retina. They extend out to the retinal periphery by E16, and reach their peak neurogenesis on E17. On E20 they can still be found at all retinal loci, albeit at reduced density, and by E22 they are found primarily in the ventral retina, comprising the only tritium-positive cells in the ganglion cell layer to have been generated on this day. This developmental increase in the genesis of presumptive displaced amacrine cells relative to ganglion cells is shown in Fig. 6. While it is clear that the population of retinal ganglion cells cease neurogenesis before the population of presumptive displaced amacrine ceils, and that they occupy a progressively smaller fraction of recently generated cells as development proceeds (Fig. 6), the neurogenetic periods for the two populations overlap substantially. All of the contralateral retinae displayed heavy HRP labelling across the complete retinal surface, so we do not believe lack of HRP labelling to be a consequence of inadequate uptake from a portion of the target visual nuclei. Further, the largest neurons in the ganglion cell layer were never HRP-negative in the retina contralateral to the injections (except in the temporal retina, e.g. Fig. lb), supporting the view that no portion of the retinotopic map had escaped labelling. The present results support the observations of Harman and Beazley,” in the wallaby, that ganglion cell genesis and displaced amacrine cell genesis are largely overlapping in time, rather than occurring in a primarily sequential fashion. Nevertheless, it could be argued that smaller retinal ganglion cells may be more difficult to label retrogradely, and that all of these tritium-positive, HRPnegative cells are in fact ganglion cells which had
R.J. COLELLO failed to be labelled. While an unlikely account for all of the tritium-positive, HRP-negative neurons, this may explain the higher proportion of displaced amacrine cells generated on El4 relative to the trend across the remaining days, shown in the graph of Fig. 6: if a few small retinal ganglion cells had not displayed detectable HRP labelling on E14, they would have been misclassified as displaced amacrine cells, thus incorrectly raising the ratio for this retina. An alternative approach would be to use a label which positively identifies displaced amacrine cells, or at least a sub-population of them. We have used a monoclonal antibody directed against ChAT to label one distinct class of displaced amacrine cell, the cholinergic amacrine cell, 32in retinae that had been exposed to tritiated thymidine on different embryonic days. ChAT-labelled retinal whole-mounts contained immunoreactive somas in two distinct retinal layers: one population was situated in the retinal ganglion cell layer (Fig. 7a), and the other, of similar density, was situated along the inner margin of the inner nuclear layer (Fig. 7b). Between each population were two thin sublaminae within the inner plexiform layer, containing the immunoreactive processes of the two respective populations, as described in the literature.‘ss32 Tritium-positive and ChAT-positive doublelabelled amacrine cells in the ganglion cell layer were observed in the retinae of rats that had been exposed to tritiated thymidine on E16-E20. Examples of tritium-positive and ChAT-positive cells in the ganglion cell layer are shown in Fig. 7d-f, from rats that had been exposed to [‘Hlthymidine on E14, El7 and E20, respectively, the El 7 sample including double-labelled cells (arrowheads). A single double-
temporal
nasal E-20
Fig.
8. Distributions
of double-labelled generated
tritium-positive) (ChAT-positive, on different days of gestation.
displaced
amacrine
cells
Neurogenesis in the rat’s retinal ganglion cell layer
labelled cell was observed in each of two rats exposed to [3H]thymidine on El5 The incidence of doublelabelled cells increased and peaked on El7 and E18, and then tapered off until E20. No double-la~lled cells were observed on E22. A similar time course for the genesis of ChAT-positive amacrine cells in the inner nuclear layer was alSO observed. As with the other populations of neurons in the ganglion cell layer, the topography of neurogenesis of the ChAT-positive ceils showed similar centroperiphera1 and dorsoventral trends (Fig. 8). The earliest generated cells were found in the central retina on E16, while at progressively later embryonic stages, neurogenesis of the ChAT-positive cells occurred in the peripheral and ventral retinal regions. On E19, double-labelled cells are most numerous in the ventral periphery, although there are a few found elsewhere across the retina. On E20, the last day of neurogenesis of the ChAT-positive cells, the few double-labelled cells are restricted to the ventral periphery. We have examined a total of 36 retinae in this study, and are confident about the neurogenetic period of the ChAT-positive cells. We are slightly less secure about the topographic information contained in Fig. 8 because although we have consistently obtained specific ChAT-positive labelling, we have not always obtained adequate labelling across the full extent of the ganglion cell layer in every retina. Consequently, while we are certain about the identification of double-labelled cells (e.g. those indicated in Fig. 8), regions of retina displaying no doublelabelled cells may be due to poor ChAT labelling. The four retinae shown in Fig. 8 contained the most thorough coverage of ChAT labelling across the ganglion cell layer, with only the El9 retina displaying a large portion of its dorsonasal region lacking ChAT labelling. DISCUSSION
The present study has demonstrated that the periods of genesis of one type of retinal ganglion cell, the alpha cell, and one type of displaced amacrine cell, the cholinergic cell, are limited in their durations relative to the genetic periods of the total populations of ganglion and displaced amacrine cells in the ganglion cell layer. This study also demonstrates that the neurogenetic periods of the populations of ganglion and displaced amacrine cells are largely overlapping, rather than being sequential. Displaced amacrime cell genesis and ganglion ceN genesis
Displaced amacrine cells in the rat’s retinal ganglion cell layer, identified by the absence of HRP labelling, are generated throughout the period of ganglion cell genesis, rather than following ganglion cell genesis. While the failure to label retrogradely some Optic axons may account for the lack of HRP labelling in a proportion of neurons (as should be
427
expected in the temporal retina), this seems an unlikely account for all of the tritium-positive, HRPnegative neurons in the nasal portion of the ganglion ccl layer. Further, our demonstration that one class of positively identified displaced amacrine cellsthe ChAT-immunoreactive cells-is generated wholly during the period of ganglion cell genesis shows conclusively that ganglion and displaced amacrine cell genesis substantially overlap in time and retinal locus. An identical temporal gradient was recently described for displaced amacrine cells in the rat which are immunoreactive for corticotropin releasing factor (CRF).@ Both this population of displaced amacrine cells, and the displaced cholinergic amacrine cells, have matching populations of cells in the inner nuclear layer.32,40 Those immunohistochemi~lly identified populations of amacrine cells in the inner nuclear layer are generated during the identical time period (Ref. 40 and present results), along with dop~ine-containing (DA) amacrine cells in the inner nuclear layer (E16E20).7 They may account for only a small proportion of the total population of amacrine cells, and future identification and birthdating of other types of amacrine cell may reveal displaced amacrine cell classes generated as early as El4 or as late as E22. The observation that displaced amacrine cells and ganglion cells are generated cont~poraneously is not incompatible with the late migration of displaced amacrine cells into the ganglion cell layer.” Zimmerman et aL4’ have shown that horizontal cells and ganglion cells which are generated at the same time do not necessarily migrate simultaneously, and the same may be true for the displaced amacrine cells, leading to their delayed arrival within the ganglion cell layer. Gang/ion ceil genesis andjibre order in the optic tract
Ganglion cell classes in the cat’s retina are generated as independent though temporally overlapping gradients in which cells in the central retina are, on average, generated before cells in the peripheral retina.34,35Most of the beta cells are generated before alpha cells, and much of the gamma cell population is generated at the end of the period of ganglion cell genesis, after the alpha cells. This temporai pattern of genesis is reflected in the adult organization of the cat’s optic tract, in which the beta axons are positioned deepest in the tract, alpha axons are situated superficial to the beta axons, and a very large number of the gamma axons are found sub-pially, at the tract’s superficial surface. *J’ Fibre order in the optic tract, then, is a chronological index of axonal arrival during development: the oldest axons are deepest in the tract, younger axons being positioned at progressively superficial positions.9~33 The adult rat’s optic tract likewise contains a segregation of axon classes, with some interesting differences from the fibre order found in the cat. First,
428
B. E.
REESE
and R. J.
the coarsest axons in the rat’s optic tract, believed to arise from the alpha cells, course through the very deepest parts of the tract,23 whereas in the cat, the beta axons separate the alpha axons from this border. Second, the population of uncrossed optic axons is densest in the deeper parts of the tract, and is scarce or absent along the pial surface,z4 whereas uncrossed optic axons in the cat can be found throughout most of the tract, including its sub-piai region.28 These two features of the rat’s optic tract led to two novel predictions about retinal ganglion cell genesis: (1) alpha cells should be generated from the very outset of ganglion ceil genesis, rather than at a relatively late stage, and (2) ipsiiaterally projecting retinal ganglion ceils should complete their genesis before the population of contralaterally projecting ganglion cells. The results of the present study confirm these two predictions. First, alpha ceils are generated during the first half of the total period of ganglion cell genesis, from embryonic day 14 through to day 17, while other ganglion cell classes continue to be generated thereafter, until embryonic day 20. No other cell type initiates its neurogenetic period prior to the alpha cells. And second, but for a very few cells born on Ei9, the population of uncrossed ganglion cells is almost completely generated by E 18, two days before the cessation of crossed ganglion cell genesis. As in carnivores then, an axon’s position in the rat’s optic tract reflects the relative time of genesis of its parent retinal ganglion ceil (see also Ref. 3). Ganglion cell genesis and the line of decussation
The present study replicates, in the rat, the observations by DrPgerS in the mouse that the earliest ceils
COLELLO
to be generated in the temporal crescent have uncrossed optic axons, while the latest cells to be generated in the crescent have crossed axons. These results are consistent with the observations of Bunt et al.’ and of Wikler et aL3’ that the crossed projection from temporal retina is a relatively late-developing feature of the rodent’s optic pathway. While neither the present study nor the study by Drgge? addresses the role of ceil death in creating this difference in the neurogenesis of temporal retina, in conjunction with these other studies,‘,3s they suggest that early generated cells from the temporal retina navigate into the ipsilaterai tract exclusively, whereas later generated cells show an increasing propensity to navigate a crossed course through the optic ehiasm, as in carnivores.25,28,35 Finally, it is interesting to note the discrepancy between the centropetipheral gradient of alpha ceil genesis and the distribution of alpha cells in adulthood. Like the population of non-aipha ceils, the alpha ceils initiate their genesis in dorsocentral retina; yet the peak density of the former ceils in the mature retina is situated dorsocentrally, whereas the peak density of the alpha cells is shifted into the far dorsotemporai periphery, upon the line of decussation.“a.‘7.27
acknowledgements-his research was supported by a ghnt from the Medical Research Council of the United Kingdom (PG 8324037). The experiments on animals were conducted under licensing and certification by the Home Office of the U.K. We thank Sue Mygal for her animal husbandry, Davina Hocking, Maurice Sanders, Mary Walker and especially Zillah Deusson for their assistance with histology, Gigi Raott! for her contribution to the analysis, and Gary Baker, John Mitrofanis and Ray Guillery for comments on the manuscript.
REFERENCES
1. Bunt S. M., Lund R. D. and Land P. W. (1983) Prenatal development of the optic projection in albino and hooded rats. Lteul Brain Res. 6, 149-168. 2. Colello R. J. (1988) Modifications of a method for preparing retinal wholemounts for sectioning. Stain Technol. 63,
183-188. 3 Colello R. J. and Guillery R. W. (1991) Observations on the early development of the optic nerve and tract of the mouse. .I. camp. NeuroZ.(in press). 4. Cowey A. and Perry V. H. (1979) The projection of the temporal retina in rats, studied by retrograde transport of horseradish peroxidase. Expi Brain Res. 35, 457464. 5. Driiger U. (1985) Birth dates of retinal ganglion cells giving rise to the crossed and uncrossed optic projections in the mouse. Proc. R. Sot. Land. B22q 57-77. 6. Driiger U. and Hofbauer A. (1984) Antibodies to heavy neurofilament subunit detect a subpopulation of damaged ganglion cells in retina. Nature 309, 624-626. . .. < .* 6a. Dreher B., Sefton A. J., Ni S. Y. K. and Nisbett G. (1985) The morphology, number, &stnl)utlon ana cemral projections of class I retinal ganglion cells in albino and hooded rats. Brain, Behao. Evol. 26, l&48. 7. Evans J. A. and Battelle B.-A. (1987) Histogenesis of dopamine-containing neurons in the rat retina. Expl Eye Res. 44, 407-414. 8. Guillery R. W., Polley E. H. and Torrealba F. (1982) The arrangement of axons according to fiber diameter in the optic tract of the cat. J. Neurosci. 2, 714-721. 9. Guillery R. W. and Walsh C. (1987) Changing glial organization relates to changing fiber order in the developing optic nerve of ferrets. J. cotnp. Neural. 265, 203-217. 10. Hanker J. S., Yates P. E., Metz C. B. and Rustioni A. J. (1977) A new, specific, and non-carcinogenic reagent for the demonstration of horseradish peroxidase. ~~stoche~. J. 9, 789-792. 11. Harman A. M. and Beaziey L. D. (1989) Generation of retinal cells in the wallaby, Sefonix brachyurus (quokka). Neuroscience 28, 219-232. 12. Jenkins E. C. (1972) Wire loop application of liquid emulsion to slides for autoradiography Technol. 47, zj-26.
in light microscopy. Stain
Neurogenesis
in the rat’s retinal
ganglion
cell layer
429
13. Linden R. and Esberard C. E. L. (1987) Displaced amacrine cells in the ganglion cell layer of the hamster retina. Vision Res. 27, 1071-1076. projections in rats. Brain Res. 272, 145-149. 14. Linden R. and Perry V. H. (1983) Massive retinotectal J., Maslim J. and Stone J. (1988) Catecholaminergic and cholinergic neurons in the developing retina of 15. Mitrofanis the rat. J. camp. Neural. 276, 343-359. J., Maslim J. and Stone J. (1989) Ontogeny of catecholaminergic and cholinergic cell distributions in the 16. Mitrofanis cat’s retina. J. camp. Neurol. zs9, 2288246. 17 Peichl L. (1989) Alpha and delta ganglion cells in the rat retina. J. camp. Neural. 286, 12&139. 18. Perry V. H. (1979) The ganglion cell layer of the retina of the rat: a Golgi study. Proc. R. Sot. Lond. B204, 363-375. 19 Perry V. H. (198 1) Evidence for an amacrine cell system in the ganglion cell layer of the rat retina. Neuroscience 5, 931-944. 20 Perry V. H., Henderson Z. and Linden R. (1983) Postnatal changes in retinal ganglion cell and optic axon populations in the pigmented rat. J. camp. Neural. 219, 356368. 21 Perry V. H. and Walker M. (1980) Amacrine cells, displaced amacrine cells and interplexiform cells in the retina of the rat. Proc. R. Sot. Lond. B208, 415431. 37 Polley E. H., Zimmerman R. P. and Fortney R. L. (1989) Neurogenesis and maturation of cell morphology in the development of the mammalian retina. In Deuelooment of the Vertebrate Retina (eds Finlav B. L. and Sengelaub D. R.), pp. 3-29. Plenum Press, New York. ‘ ” of axons according to diameter in the optic nerve and optic tract of the rat. 23 Reese B. E. (1987) The distribution Neuroscience 22, 1015-1024. 24 Reese B. E. (1987) The position of the crossed and uncrossed optic axons, and the non-optic axons, in the optic tract of the rat. Neuroscience 22, 102551039. 25 Reese B. E. and Baker G. E. (1990) The course of fibre diameter classes through the chiasmatic region in the ferret. Eur. J. Neurosci. 2, 3449. W. (1991) Cell survival in the mammalian retina’s ipsilateral projection is 26. Reese B. E., Colello R. J. and Thompson independent of birthdate. ARVO Abstr. 13, 925. 27 Reese B. E. and Cowey A. (1986) Large retinal ganglion cells in the rat: their distribution and laterality of projection. Expl Brain Res. 61, 375-385. 28 Reese B. E., Guillery R. W., Marzi C. A. and Tassinari G. (1991) The position of axons in the cat’s optic tract in relation to their retinal origin and their chiasmatic pathway. J. cony. Neural. 306, 539-553. 29. Sengelaub D. R., Dolan R. P. and Finlay B. L. (1986) Cell generation, death, and retinal growth in the development of the hamster retinal ganglion cell layer. J. camp. Neural. 246, 527-543. of mouse retina studied with thymidine-‘H. In Structure of the Eye (ed. Sm.&or 30. Sidman R. L. (1961) Histogenesis G. K.), pp. 487-506. Academic Press, New York. 31. Torrealba F., Guillery R. W., Eysel U., Polley E. H. and Mason C. A. (1982) Studies of retinal representations within the cat’s optic tract. J. camp. Neural. 211, 3777396. 32. Voigt T. (1986) Cholinergic amacrine cells in the rat retina. J. camp. Neural. 248, 19-35. 33. Walsh C. and Guillery R. W. (1985) Age-related fiber order in the optic tract of the ferret. J. Neurosci. 5, 3061-3069. 34. Walsh C. and Polley E. H. (1985) The topography of ganglion cell production in the cat’s retina. J. Neurosci. 5, 741-750. 35. Walsh C., Polley E. H., Hickey T. L. and Guillery R. W. (1983) Generation of cat retinal ganglion cells in relation to central pathways. Nature 302, 61 l-614. 36. Wassle H., Chun M. H. and Miiller F. (1987) Amacrine cells in the ganglion cell layer of the cat retina. J. camp. Neural. 265, 391-408. 37. Wikler K. C., Perez G. and Finlay B. L. (1989) Duration of retinogenesis: its relationship to retinal organization in cricetine rodents. J. camp. Neural. 285, 1577176. 38. Wikler K. C., Raabe J. I. and Finlay B. L. (1985) Temporal retina is preferentially represented in the early retinotectal projection in the hamster. Devl Brain Res. 21, 1522155. 39. Wong R. 0. L. and Hughes A. (1987) The morphology, number, and distribution of a large population of confirmed displaced amacrine cells in the adult cat retina. J. camp. Neural. 255, 159-177. 40. Zhang D. and Yeh H. H. (1990) Histogenesis of corticotropin releasing factor-like immunoreactive amacrine cells in the rat retina. Deul Brain Res. 53, 194199. 41. Zimmerman R. P., Policy E. H. and Fortney R. L. (1988) Cell birthdays and rate of differentiation of ganglion and horizontal cells of the developing cat’s retina. J. camp. Neural. 274, 77-90. A&.
(Accepted 26 June 1991)
