selective uptake of tritiated glycine, gaba and d-aspartate by retinal cells in culture: a study...

Developmental Brain Research, 7 (1983) 107--120 107 Elsevier Biomedical Press

Research Reports

Selective Uptake of Tritiated Glycine, GABA and D-Aspartate by Retinal Cells in Culture: a Study using Autoradiography and

Simultaneous Immunofluorescense*

RICHARD BEALE and NEVILLE N. OSBORNE

University of Oxford, NuffieM Laboratory of Ophthalmology, Walton Street, Oxford 0)[2 6AW (U.K.)

(Accepted September 14th, 1982)

Key words: retina - - cell culture - - GABA - - glycine - - D-aspartate - - autoradiography - - cell identification

Autoradiographic studies on the uptake of 3 tritiated amino acids by cell cultures of rat retina are reported. At 0.2/zM [aH]- GABA is taken up with great selectivity by a minority of medium to large multipolar cells. At 2/~M [aH]glycine is taken up with slightly less selectivity by a similar proportion of multipolar cells; the proportions of cells labeled by either amino acid are additive which indicates that they are separate populations. In simultaneous autoradiographical and indirect immunofluorescence experiments, both of these cell classes were found to bind tetanus toxin and are therefore presumably neurons. At 0.4 #M [aH]D- aspartate is preferentially taken up by virtually all small spheroid cells which lack long processes (putative immature photoreceptors), as well as by a minority of larger multipolar cells. All 3 tritiated amino acids also label flattened cells which include retinal glia. The results are discussed in relation to similar experiments performed on undissociated retina and we conclude that these experiments allow us to identify and quantify immature photoreceptors and certain sub-types of amacrine cells in these cultures.

INTRODUCTION

Cell culture is a useful way to study the differ-

entiated properties o f CNS cells and their develop-

ment. The cells' dispersion limits the extent o f cell- cell interaction, promotes control o f the extracel-

lular environment, and permits microscopic exam- ination o f the individual living cells. However, for

these advantages to be fully realized, the individual

cell types present in the culture must be identifiable.

Three major cell types o f rat retinal cultures have been described and differentiated according to their binding o f tetanus toxin a, which is generally considered to be a neuronal markerS, 20. The most

numerous was a small spheroid cell having at most only a single short process. These cells share a lack

o f affinity for tetanus toxin with the photoreceptors o f adult retina, and we have tentatively identified them as immature photoreceptors. Larger spheroid cells bearing long processes bind tetanus toxin and so are presumably non-photoreceptor neurons. Cells

that flatten to the substrate and substantially in-

crease in number with time are mostly Miiller cells

since the majority bind the glial-specific Ran-2 anti-

body 1, but little if any anti-glial fibrillary acidic protein2, a.

As part o f our cont inuing aim to identify the

cell types in rat retinal cultures, we have examined

their capacity to take up exogenous tritiated GABA, glycine and aspartate. Several amino acids,

including these 3, are suspected to be neurotransmitters in the vertebrate retina 21. In particular, high-

affinity uptake systems, thought to be responsible for the termination o f transmitter action, have been described in rat retina for GABAI~, a° and glycine 24.

Rat retina also has temperature- and sodium-depen-

dent high-affinity uptake systems for aspartate and glutamate, which may be taken up by a c o m m o n mechanism 3e.

When these amino acids are presented to the retina at low concentrations, autoradiographical studies suggest that it concentrates them into dis-

Some of the results described in this paper have been previously reported in preliminary form 2.

0165-3806/83/0000-0000/$03.00 © 1983 Elsevier Science Publishers

108

crete cellular elements. Glycine is taken up by a sub- class of amacrine cells in all mammalian retinas including ratS,16,z8, za, but interplexiform cells accumulating glycinelS, z9 have not been described in mammalian retinas.

GABA is concentrated by Mailer and amacrine cells in most mammalian speciesS,10,17,~8, although in rat retina only Mailer cells have been firmly identified as GABA-accumulating 2z. Briiun and Ehinger 5 have suggested that intense Mailer cell uptake may obscure amacrine cell labeling in the tightly-packed retina in situ, and we hoped that the culture system, where the cells are well separated, would throw some light on this apparent anomaly in rat retina.

In order to differentiate firmly between Mailer and anaacrine cell uptake of GABA we have developed a technique for the simultaneous autoradiography of amino acid uptake and indirect immunofluorescence of tetanus toxin binding. This technique, based on the use of benzoquinone as a fixative, should have wide applicability in deter- mining whether selective uptake mechanisms coexist with antigens characteristic of a certain cell type.

Glutamate has been shown to be taken up by rod photoreceptors, Miiller cells, horizontal cells and some ganglion cells in mammalian retinas~, 1°,z4. Lain and Hollyfield 17 have studied the uptake of [3H]o-aspartate (a metabolically inert enantiomer of aspartate7, at) into human retina, and found that it labels rod photoreceptors as well as some cell bodies in the inner plexiform layer. We have used autoradiography to study the localization of uptake of this enantiomer in both whole rat retina and retinal cell cultures.

The work described here supports our previous identification ,~ of Mailer cells, photoreceptors and non-photoreceptor neurons and assigns identities to at least 2 large sub-classes of this last group.

MATERIALS AND METHODS

Cell cultures Cells were dissociated and cultured as described

previously z.

Solutions Incubations with tritiated amino acids were

carried out in a N-2-hydroxyethylpiperazine-N'-2-

ethanesulphonic acid (HEPES)-buffered minimal essential medium (MEM): Gibco 041-2370). In experiments investigating sodium dependancy this was replaced by a salt solution of composition (mM): Na+130; K+5.4; Ca2+2; Mg~+0.8; glucose 33; C1-139.4; SO2-0.8; HEPES 25, pH 7.4 with Tris base; or by a solution identical but for the equi- molar substitution of choline chloride for sodium chloride. For indirect immunofluorescence incubations the MEM was supplemented with 20 ~, normal lamb serum (Gibco). Washes were in a salt solution buffered by 20 mM HEPES that has been previously described 3. For indirect immunofluorescence assays this was supplemented with 5 ~.~, normal lamb serum.

A utoradiography on cultures Cultures on 13 mm diameter cover slips were

dipped in 4 successive beakers of wash solution, drained, and incubated with 40 #1 of MEM containing 0.2 #M GABA (Amersham TRK 527, 50 Ci/ mmol), 0.4 #M [aH]D-aspartate (Amersham TRK606, 10 Ci/mmol) or 2 #M [ZH]glycine (Amers- ham TRK71, 12.3 Ci/mmol) for 30 min at room temperature (20-24 °C). The radioactive solution was removed and the coverslips washed as before. They were then incubated in 1 ?/o glutaraldehyde in the wash solution for 5 rain at room temperature before being dipped successively in wash solution (4X), distilled water (1 x), 50~,,, ethanol (1 ×) and 100~ ethanol (2 ~ ). When dry the coverslips were attached to microscope slides with DPX (BDH) and coated with Ilford L4 emulsion diluted 1:1 (v/v) with water. After exposure for 6-8 days at --20 °C the emulsion was developed for 8 rain in llford 1D19. The coverslips bearing the cultures were mounted with a second coverslip using phosphate-buffered saline (PBS)-glycerol.

Simultaneous autoradiography and indirect immuno-

.fluorescence Cultures were processed for tetanus toxin im-

munofluorescence as described previously 3, with 0.2 #M [ZH]GABA or 4#M glycine being included in the last of the 3 30 rain incubations. After washing as above the cultures were fixed in benzoquinone (Aldrich) for 30 min at room temperature. The benzoquinone, which had been recrystallized fror: hot petroleum ether, was dissolved at 4 mg/ml in 17

109

mM sodium phosphate, pH 7.4 just before use, and replaced by a fresh solution after 15 min. Subse- quent processing was as described above except that the coverslips bearing the cultures were carefully detached and inverted onto a drop of PBS-glycerol on a clean slide after development of the emulsion.

Autoradiography oJ' [3 HJD. aspartate uptake by adult retina

Retinas were swiftly dissected from adult rats in ambient lighting, rinsed in MEM and incubated with 0.4-2/~M [aH]o-aspartate in oxygenated MEM for 30 min at 25 °C in a shaking water bath. Subse- quent processing for autoradiography was as previously described 25.

Metabolism of tritiated amino acids Adult retinas were incubated with 2 #M [3H]D-

aspartate, 0.2 #M [aH]GABA or 2/~M [3H]glycine as described above. They were then washed, homo- genized in 10 mM HC1 and centrifuged at 10,000 g for 10 min. The supernatant was lyophilized, re- dissolved in water :acetone (1:1 v/v) and spotted on Whatman no. 1 paper. After ascending chromato- graphy in n-butanol :acetic acid (9:1) saturated with water, more than 95 ~ of the radioactivity applied was recovered from the position of the corresponding authentic amino acid.

Photography Photographs on Tri-X film (Kodak)were taken

using a Zeiss Photomicroscope equipped with epi- fluorescence.

Cell counts and measurements Estimates of frequencies of cells taking up [SH]-

GABA or [aH]glycine were made by counting densely-labeled and total cell numbers in randomly chosen fields. Densely-labeled cells were defined as those in which more than half the cell soma was blackened by confluent silver grains. Frequencies of cells labeled by [ZH]D-aspartate were estimated by subtracting from the total those cells labeled less than or equal to the background labeling. At 6 days in vitro only spheroid cells were counted. At least 1000 cells were counted for each frequency estimate. Cell diameters were estimated using a calibrated ocular grid and expressed as the mean of 2 ortho- gonal measurements.

RESULTS

[aHJGABA uptake Fig. lc shows an autoradiograph of a 6d in vitro

retinal culture incubated with 0.2/~M [3H]GABA. Densely packed silver grains cover the somata and long processes of several multipolar cells, while cells flattened to the substrate are only moderately labeled. At 1 day in vitro the densely-labeled spheroid cells had many short processes radiating from the cell body; and a minority of the flattened cells (54 of 121 counted) were densely labeled (Fig. la, b) while others were only moderately labeled (not shown, but similar to fiat cell labeling in 1 day in vitro cultures incubated with [3H]glycine, Fig. 3a, b). Note that the great majority of spheroid cells are unlabeled in these autoradiographs and even seen to absorb fl- emissions from underlying flat cells (Fig. lb, c).

In cultures 6 days in vitro, all the labeled spheroid cells had 2 or more long branching processes. The most common morphology was a round soma with 4 or 5 processes emerging from it, although cells with triangular or fusiform somata were also present. Fig. 2 shows a selection of morphological types of GABA-labeled cells. Diameters of densely-labeled cell bodies in 6 days in vitro cultures prepared from 7 day rat retinas ranged from 9 to 16/~m with a mean of 11.2/~m (n = 30), but these measurements may overestimate the true diameters because of a lateral spread of fl-particles. The proportions of spheroid cells labeled by [3H]GABA in cultures from 3 separate dissections are given in Table I. Means of 2.7 ~ and 3.7 ~ were found for cultures 1 and 6 days in vitro respectively, prepared from 7- day-old rat pups. This compares with means of 12.7 ~'o and 9.0 ~ of cells labeled with tetanus toxin in separate experiments.

Although most autoradiography experiments were performed on cultures 1-6 days in vitro, cells densely labeled by [aH]GABA were found in cultures up to 21 days in vitro (Fig. 2d). In autoradiographs of cultures prepared from 1-day-old rat pups, cells of similar morphology were labeled by [aH]- GABA, although the proportion of densely-labeled spheroid cells appeared to be higher.

laHJglycine uptake Autoradiographs of cultures incubated with 2/zM

J I

O ~ ~

T~ B

Q a b "

Fig. 1. Autoradiography of [3H]GABA uptake. Cultures were incubated with 0.2 t~M [aH]GABA for 30 min at room temperature and then processed for autoradiography as described in the text. a and b show phase-contrast and bright-field views respectively of a culture 1 day in vitro. Most of the cells are clearly unlabeled. Silver grains densely cover 4 spheroid cells, 2 of which have short processes which are also clearly labeled. A large flattened cell is also densely labeled, c: bright-field view of a culture 6 days in vitro that has been very lightly counter-stained with methylene blue to make the unlabeled cells visible. Silver grains densely cover a minority of the spheroid cells. Most of these can be seen to have several long processes. The flattened cells are covered with moderate densities of silver grains. Scale bars : 50/~m.

111

Fig. 2. Autoradiography of [aH]GABA uptake. These are bright-field views of a selection of spatially isolated [aH]GABA-labeled cells, a -c : from cultures 6 days in vitro, d: from a culture 21 days in vitro. Scale bars ~ 50 #m.

112

TABLE I

Proportions of spheroM cells labeled by [3H]glycine, l 3H}GABA, [3H]D-aspartate and tetanus toxin in cell cultures 1 and 6 days in vitro prepared from 7- to 9-day Wistar rat retinas

Clones of flat cells, all of which were labeled by each tritiated amino acid were counted at 1 day in vitro to provide an estimate of their numbers in the original dissociate. Results are expressed as means ~ S.D. of percentages obtained from 3 separate dissections.

GABA G l y e i n e D-Aspartate Tetanus toxin Flat cells

1 day in vitro 2.7 ± 0.85 3.1 ± 0.83 88.4 _k 1.46 12.7 ~: 2.35 0.6 ~k 0.3 6 days in vitro 3.7 ± 0.17 3.4 :~: 1.32 87.9 ± 1.68 9.0 ± 2.70 --

glycine are shown in Fig. 3. A similar proportion of spheroid cells (means 3.1 and 3.4 ~ at 1 and 6 days in vitro respectively, Table I) were densely labeled by [3H]glycine, but the labeling was less selective than that displayed by [3H]GABA in that most of the other spheroid cells were lightly labeled. This is probably a consequence of the higher concentrations of glycine needed to densely label the spheroid cells which took up the most tritium. These cells, like those densely labeled by [3H]GABA were mostly multipolar, although they tended to have fewer and shorter processes; in some cells no process could be distinguished in autoradiographs. Cell soma diameters in cultures 6 days in vitro ranged from 7 to 13 #m with a mean of 9.6 #m (n ~- 30). Again these are possibly overestimates. Most of the neuritic fields were characterized by frequent branching, and these short branches gave a 'knob- bed' appearance to some cells. A particularly clear example of this morphology is shown in Fig. 3e, which also shows a cell with relatively smooth processes. Although the morphology of these cells appeared to be different from those labeled by [aH]- GABA, it was considered possible that the apparent differences were due to a greater efficiency of labeling by [3H]GABA. Therefore, in 2 experiments quadruplicate cultures were incubated with [aH]- GABA, [3H]glycine or a mixture of the two. Counts of these cultures showed that the proportions of densely labeled cells were additive (Table II) which indicates that the cells labeled by the 2 amino acids belonged to separate populations. Flat cells were lightly labeled by [aH]glycine.

[3 H]D-aspartate uptake

Autoradiographs of cultures incubated with 0.4 #M [3H]D-aspartate showed a quite different pattern

from those labeled by [3H]glycine or [3H]GABA (Fig. 4). The great majority of spheroid cells (means 87 and 8 8 ~ at 1 and 6 days in vitro respectively; Table I) were moderately labeled by this procedure. Of these labeled cells, most appeared to be small and lacking long processes, although some of the larger process-bearing cells did appear to be labeled (Fig. 4c-e). The detailed morphology of the labeled cells was difficult to determine because the selectivity of D-aspartate uptake was relatively low. Increasing the exposure time of the autoradiographs or the concen- tration of [3H]D-aspartate led to such an increased grain density (labeling of flat cells with this amino acid was relatively high) that the morphology of the underlying cells was further obscured. Similar results were obtained from cultures 1-21 days in

vitro. The proportion of spheroid cells labeled by [ZH]D-

aspartate agreed reasonably well with that of cells unlabeled by tetanus toxin in separate experiments (Table I) - - cells which we have tentatively identified as immature photoreceptors ~. The comparison indicates that only a small proportion of non-photoreceptor neurons take up [aH]o-aspartate in these cultures; direct counts of labeled multipolar cells suggested that they made up about 2 ~ of the total.

An autoradiograph of an adult rat retina incubated with 1.5 #M [3H]D-aspartate is shown in Fig. 4f, g. Moderately high densities of silver grains covered the outer nuclear and photoreceptor layers. Grains were also concentrated over numerous somata in the middle and inner parts of the inner nuclear layer. The inner plexiform layer was relatively free of silver grains. Labeling of the ganglion cell layer was restricted to the end feet of Miiller cells: neither ganglion cell bodies, nor bundles of their axons, were labeled. Similar results, but with

O - 7

Fig. 3. Autoradiography of [aH]glycine uptake. Cultures were incubated with 2 p M [aH]glycine for 30 min at room temperature and then processed for autoradiography as described in the text. a and b: phase-contrast and bright-field views respectively of a culture 1 day in vitro. Note that 4 cells are densely labeled, of which 2 have labeled processes. Several other cells including a flattened cell have moderate densities of silver grains above them, but would not be counted as [3H]glycine-labeled cells (see Methods). c and d: phase-contrast and bright-field views respectively of a culture 6 days in vitro. Flattened cells cover the majority of the substrate and are responsible for the apparent increase in background labeling. Six cells are densely labeled and have only short processes. Note that other cells, some of which appear to have processes in phase-contrast, are moderately labeled, e: a bright-field view of a culture 6 days in vitro, shows 2 densely labeled cells, one of which bears knobs on its soma and neurites, f: phase-contrast image of a culture 6 days in vitro incubated with [aH]glycine in the absence of sodium ions. Scale bars = 50/zm,

114

TABLE II

Percentages of cell labeled by [3H]glycine, IaHJGABA and a mixture of the two

Results are expressed as means ± S.D. from quadruplicate cultures. Means in column III significantly different from means in colums I and II by Students t-test.

1 H 111 GABA Glycine GABA + glycine SumofcolumnslandH

1 day in vitro 2.55 4- 0.17 2.75 4- 0.24 5.13 ± 0.39* 5.30 6 days in vitro 3.45 i 1.17 3.96 i 0.67 7.93 ± 1.57"* 7.41

* P < 0.01. ** P < 0.001.

less dense labeling were obtained f rom incubations

with 0.4 # M [3H]D-aspartate.

Variations in incubation medium composition

The incubat ion medium was usually a HEPES-

buffered M E M . This was chosen because it was

hoped that its higher amino acid content would

improve the selectivity o f uptake by inhibiting the

relatively non-specific low-affinity amino acid up-

take systems. The replacement o f the M E M with a

simple salt solution containing 130 m M NaC1 led to

essentially similar autoradiographs. However, re-

placement o f sodium chloride by choline chloride

led to the total abolit ion o f specific labeling by all 3 amino acids (e.g. Fig. 3f).

Combined autoradiography and immunofluorescence

In a previous study we showed that the larger

spheroid cells bearing long processes bound tetanus toxin and hence were presumably neurons 3. The

cells densely labeled by [aH]GABA and [aH]glycine

were o f this morphological type, but to make certain

that they bound tetanus toxin we wished to label them simultaneously by indirect immunofluores-

cence of tetanus toxin binding and autoradiography

of amino acid uptake. I t is not possible to simply

combine the 2 techniques because the fixatives

generally used in indirect immunofluorescence (such as formaldehyde and acid-alcohol) do not 'fix'

amino acids to cells, whereas glutaraldehyde, which does, causes autofluorescence and so destroys the

dark background necessary to localize specific fluo-

rescence. The fixative benzoquinone used to demonstrate substance P-immunoreact ive cells in retinal

cultures (unpublished observations) was therefore

used and found to fix amino acids into retinal cells,

a l though with substantially less efficiency than

glutaraldehyde. Cultures were therefore successively incubated

with tetanus toxin, rabbit anti-tetanus toxin and

rhodamine-conjugated goat anti-rabbit Ig with [aH]-

G A B A or [aH]glycine in the final incubation. They

were then fixed with benzoquinone and processed for autoradiography. Fig. 5a, b shows a pair o f

fluorescence and bright-field micrographs o f the same field o f a culture incubated with [aH]GABA. The large mult ipolar cell labeled by silver grains is

also labeled by rhodamine fluorescence, indicating

Fig. 4. Autoradiography of [aH]v-aspartate uptake. Cultures were incubated with 0.4 pM [3H]n-aspartate for 30 min at room temperature and then processed for autoradingraphy as described in the text. a and b: phase-contrast and bright-field views respectively of cells cultured for one day. Most ceUs are lightly labeled, but some cells bearing processes are unlabeled (outline arrows). Another process-bearing cell is clearly labeled (filled arrow), e and d: phase-contrast and bright-field views respectively of retinal cells cultured for 6 days. The apparently increased background labeling is due to flattened cells which almost entirely cover the substrate. Nearly all the smaller spheroid cells are moderately labeled, although some form tall aggregates and the overlying grains are not in the plane of focus of the bright-field picture. Most of the larger spheroid cells are unlabeled (outline arrows). One large labeled process-bearing cell is indicated with a filled arrow, e: a culture 6 days in vitro which has been lightly counter-stained with methylene blue and photographed using phase-contrast optics. Examples of labeled and unlabeled process-bearing spheroid cells are indicated by filled and outline arrows respectively. A gap in the flat cell layer is indicated by a dotted arrow, f and g: phase- contrast and bright-field views respectively of the same field of an adult retinal section prepared for autoradiography of [3HID- aspartate uptake as described in the text. Silver grains appear over almost all somatic profiles in the outer nuclear (on) layer except

115

for a few larger cell bodies (arrows) which may belong to cone photoreceptors. The outer part of the outer plexiform (op) layer is also labeled as are many cell bodies in the inner nuclear (in) layer. Labeling of the inner plexiform (ip) layer is sparse. The ganglion cell (gc) layer has a labeling pattern typical of Mfiller cell end-feet. Note that neither cell bodies nor bundles of ganglion cell axons (arrowed) are unlabeled. Scale bars : 50/~m.

- . " • . . , .

F

r l l

d ~

I

~-,, ~ . . - :- .

- j . -

. , ° ,

o : . . . -

. ' , i : , . :

Fig. 5. Simultaneous labeling using autoradiography and indirect ~mmunofluorescence. a and b: rhodamine fluorescence and bright-field images respectively of a culture 9 days in vitro incubated successively with tetanus toxin, rabbit anti-tetanus toxin, and goat anti-rabbit Ig conjugated to rhodamine plus 0.4 pM [aH]GABA. Note that although 3 cells are clearly labeled by tetanus toxin, only one is also labeled by [3H]GABA. c and d : a corresponding pair of micrographs from a sister culture treated in precisely the same way as that shown in a and b except for the substitution of 4/~M [aH]glycine for ['~H]GABA. Here 4 cells are clearly labeled by tetanus toxin, but only the 2 smallest are labeled by [aH]glyeine. Note the relative inefficiency of retention of [:~H]glycine by benzoquinone by comparison of d with Fig. 3. Scale bats -: 50 ore.

that it has both taken up [aH]GABA and bound tetanus toxin. An adjacent cell is labeled by rhodamine only. All cells densely labeled by [aH]GABA were also labeled by tetanus toxin.

A similar pair of micrographs is shown for a culture labeled by [3H]glycine and tetanus toxin. (Fig. 5c, d). Comparison with Fig. 2 shows the relative inefficiency of the fixation of this amino acid by benzoquinone relative to that by glutaraldehyde. Nevertheless, in most cases the dense to moderate labeling of a minority of small spheroid somata could be clearly distinguished from the light labeling of adjacent cells. Two small spheroid ceils in Fig. 5d were clearly labeled by both glycine and tetanus toxin, while the larger multipolar cell was only labeled by tetanus toxin. As yet we have been unable to satisfactorily double-label cultures with [3H]D- aspartate and tetanus toxin, due to the low efficiency of fixation of the amino acid by benzoquinone and its relatively low selectivity of uptake.

DISCUSSION

The experiments reported here were designed to identify cell types in cultures of dissociated rat retina. The 3 tritiated amino acids used have all been shown to be taken up selectively by certain cell types in mammalian retina, and our results show that comparable sub-populations can be identified in the retinal cultures. The selective labeling was probably by high-affinity uptake since the high concentrations of other amino acids in the incubation media should have blocked the less specific low- affinity uptake sites. The abolition of labeling in sodium-free medium is also consistent with high- affinity uptake, but kinetic experiments would be necessary to establish this point.

Cells labeled by [aH]glycine are likely to be identical to the cells accumulating glycine in rat retina. In cat retina these cells have been noted to have a morphology resembling the narrow field bistratified amacrine cells (see ref. 17). It is notable that the cells labeled by [3H]glycine in our cultures had small neuritic fields and sometimes displayed large knobs on the cell body and proximal dendrites similar to those reported for the narrow-field bistratified amacrine cell of rat retina described by Perry and Walker ~6.

117

Glycine-labeled cells had the morphology of neurons and bound tetanus toxin in double-labeling experiments. They comprised a substantial proportion of the tetanus-positive cells in cultures prepared from 7-day-old rats.

After incubations with 0.2/~M [~H]GABA, large multipolar cells with long neurites were densely labeled. These were clearly not the same class of cell that took up [ZH]glycine since the proportions of each cell type were additive. The GABA-labeled cells were shown to also bind tetanus toxin in double-labeling experiments and so were presumably neurons8, 2°. This finding is slightly surprising in view of previous reports that [3H]GABA only labeled Mi~ller cells in rat retina 5,~3, although glutamate decarboxylase, a marker for GABAergic neurons, has been since localized to a subpopulation of amacrine cells with somata in the middle and inner parts of the inner nuclear layer of rat retina3L The autoradiographic studies quoted above did in fact show labeled somata in these regions, but these could not be definitely identified, presumably because of the dense labeling over Miiller cells. Our results demonstrate that certain rat retinal neurons take up [aH]GABA from low concentrations and hence are similar to neurons in rabbit 1° and cat 28 retinas. The difficulty of demonstrating [3H]GABA uptake into neurons in whole rat retina seems likely to be due to a high intrinsic capacity by Miiller cells for [3H]GABA uptake relative to [3H]glycine uptake, since labeling of amacrine cells by [3H]glycine was not obscured by MOiler cell labeling. This is reflected in our cultures at one day in vitro when a large minority of flattened cells were densely labeled by [aH]GABA, whereas [3H]glycine only lightly labeled flattened cells. The lack of densely labeled cells after 6 days in vitro may represent a dediffer- entiation of the Miiller cells under culture conditions.

The [aH]GABA-labeled neurons probably rep- resented several different morphological types of amacrine cells. All seemed to have large neuritic fields compared to the neurons labeled by [3H]glycine. It is therefore likely that all are wide-field amacrine cells, but the only tentative identification of amacrine sub-class containing GABAergic cells that we can make is that of the wide-field diffuse amacrine cell (see ref. 26) which has the character-

118

istic dendritic swellings sometimes seen on GABA- labeled cells in these cultures (Fig. 5a, b). It is quite remarkable that the morphological differences between, for example, cells labeled by GABA and cells labeled by glycine seem to be retained in dispersed cell culture.

The uptake of [aH]GABA has been previously demonstrated in cell cultures of embryonic chick retina. Guerinot and Pessac 13 and de Pomerai and Carr 27 have demonstrated a high-affinity uptake system using biochemical techniques. In autoradiographic studies Hyndman and Adler 15 and de Pomerai and Carr z7 produced similar results to ours: flat cells were moderately labeled and a large subset of the cells of neuronal morphology were densely labeled. Identification of the labeled cell types in these chick cultures will require additional criteria since [3H]GABA is concentrated by both horizontal and amacrine cells in chick retina 19.

In our investigations on cells labeled by [3H]- GABA or [aH]glycine we developed a double-labeling technique for simultaneously demonstrating tetanus toxin binding by indirect immunofluorescence and amino acid uptake by autoradiography. This is not an ideal technique since the benzoquinone does not fix the amino acids into the cells with great efficiency. Nevertheless, it should prove possible to generalize this simple technique to other substances that can be localized by indirect immunofluorescence. One particular question that could be asked with this technique is whether pep- tide substances or transmitter-specific enzymes coexist with amino acid uptake systems in the same

cell. We examined [aH]o-aspartate uptake because this

amino acid, or glutamate has been shown to be selectively taken up by photoreceptors and Miiller cells in human, cat, monkey and rat retinasS, 17,aa, and we wished to see if it was also selectively taken up by the cell type we had tentatively identified as the immature photoreceptor on the basis of its lack of tetanus toxin binding a.

First we investigated [ZH]D-aspartate uptake into adult rat retina, since this had not been reported. The autoradiographic results were very similar to those reported by Lam and Hotlyfield 17 for human retina and Voaden et al.Z4 for [all]glutamate uptake into rat retina. Grains were concentrated over the

rod photoreceptor cells as well as certain cell bodies in the middle and inner parts of the inner nuclear layer. Labeling of the ganglion cell layer had the pattern of Miiller cell processes and it is likely that many of the labeled somata in the inner nuclear layer were those of Miiller cells. However, other labeled somata in the inner nuclear layer had the appearance of neurons.

In our cultures the small cells lacking long processes were labeled by [ZH]o-aspartate, and the mean proportion of these cells agreed fairly well with the proportion of cells not labeled by tetanus toxin. These results confirm that these cells are immature photoreceptors.

In addition to these small cells, we also observed a small proportion of multipolar cells labeled by [3H]o-aspartate. In view of our previous observa- tion that all multipolar cells bound tetanus toxin 3, these cells are likely to be neurons, and we are attempting to adapt the immunofluorescence-autoradiography double-labeling technique to settle this question. Tentative counts suggested that these cells comprised 2 0/o of the total. Identification of these o- aspartate-labeled multipolar cells must await further study, but the morphologies seen suggested that some amacrine cells might be labeled.

The significance of a selective uptake of [3H]o- aspartate is unclear. It may be that neurons dis- playing the selective uptake use glutamate or aspartate as transmitter 11, although there are examples of cells thought to be glutaminergic which do not take up the acidic amino acids 6. In skate retina, Wu and Dowling 3s have concluded from electrophysiologi- cal experiments that aspartate is the most likely candidate for the photoreceptor transmitter, and a light-inhibition of aspartate release from rat 21 and rabbit 2z retinas is consistent with the concept that release of the photoreceptor transmitter is continu- ous in the dark and inhibited by illumination 9. It would be interesting to see if the release of any amino acids from the cultured photoreceptors could

be detected. Reports of a selective aspartate or glutamate

uptake into neurons in cell cultures are relatively rare. [ZH]glutamate labels process-bearing cells in hippocampal cultures 35 but not in cultures of cere- bellum 6 or cerebral cortex a7. Hyndman and Adler 14 have demonstrated glutamate uptake into retinal

119

cultures prepared f rom early chick embryos, and find labeling of both process-bearing and non-process-bearing spheroid cells. The same authors have also reported cells labeled by [3H]GABA, but find none labeled by [3H]glycine uptakO 5, which has been shown to mature later than that of GABA in this tissue 4. Their cultures are therefore quite similar

to ours, but appear to contain the earlier developing retinal cells.

We decided to study rat retinal cultures because

the mammalian retina is of a 'simpler' type than others, and differs f rom non-mammalian retinas in its neurochemistry (for instance, GABA is taken up by amacrine cells, and not by horizontal cells, and the existence of serotonergic neurons has not been established). Also much work has been done on the development of the rat retinal projections and this is a problem that can be approached with some power in cell cultures.

These rat retinal cultures have now been exten- sively characterized. Glial cells, many of which are Miiller cells but with some astrocytes2, a comprise less than 1 ~ of the initial retinal dissociate, as judged by counts of flattened cells after 1 day in vitro (Table I). Immature photoreceptors do not bind tetanus toxin 8, but are selectively labeled by D- aspartate. They do not produce outer segments but a minority develop a single process, which may be analogous to the vitreal process ofphotoreceptors in vivo. Now that we have established that these cells

survive in culture, the way is open to examine what conditions are required for the growth of inner and outer segments. It would also be ol interest to examine the characteristics of the system responsible for their selective labeling and to see whether they can be shown to release aspartate or any other transmitter candidate.

This paper also establishes the presence of relatively large numbers of presumed GABAergic and glycinergic amacrine cells in these cultures, and

demonstrates that they are separate populations. We have also evidence for the presence of ganglion and horizontal cells as well as for amacrine cells containing substance P-like immunoreactivity (in prepara-

tion). With methods available for the simple identification of the retinal cell types it may be possible to discover factors which promote their survival, and methods for their purification for biochemical analysis and cellular interaction studies.

ACKNOWLEDGEMENTS

We are grateful to Dr. R. D. Thompson (Well-

come Labs, Beckenham) and Dr. Erika Abney (MRC Neuroimmunology Unit, University College, London) for gifts of tetanus toxin and rabbit anti- tetanus toxoid respectively. Saraswati Patel and Joy

Shirley provided excellent technical assistance. This work was supported by the Stifftung Volkswagen- werk.

REFERENCES

1 Bartlett, P. F., Noble, M. D., Pruss, R. M., Raft, M. C., Rattray, S. and Williams, C. A., Rat neural antigen-2 (Ran-2): a cell surface antigen on astrocytes, ependymal cells, Mfiller cells and leptomeninges defined by a monoclonal antibody, Brain Res., 204 (1981) 339-352.

2 Beale, R., Neuhoff, V. and Osborne, N. N., The character- isation of cell cultures of neonatal retina. In R. M. Clayton, J. Haywood, H. W. Redding and A. Wright (Eds.), Biology of Normal and Abnormal Retinas, Aca- demic Press, London, 1982, in press.

3 Beale, R., Nicholas, D., Neuhoff, V. and Osborne, N. N., The binding of tetanus toxin to retinal cells, Brain Res., 248 (1982) 141-149.

4 Bondy, S. C. and Purdy, J. C., Development of neurotransmitter uptake in regions of the chick brain, Brain Res., 119 (1977) 403--416.

5 Brfiun, A. and Ehinger, B., Uptake of certain possible neurotransmitters into retinal neurones of some mare-

mals, Exp. Eye. Res., 19 (1974) 435--447. 6 Currie, D. N. and Kelly, J. S., Glial versus neuronal

uptake of glutamate, J. exp. BioL, 95 (1981) 181-193. 7 Davies, L. P. and Johnston, G. A. R., Uptake and re-

lease of D- and L-aspartate by rat brain slices, J. Neuro- chem., 26 (1976) 1007-1014.

8 Dimpfel, W., Neale, J. H. and Habermann, E., 125I- labelled tetanus toxin as a neuronal marker in tissue cultures derived from embryonic CNS, Naunyn- Schmiedeberg's Arch. PharmacoL, 290 (1975) 329-333.

9 Dowling, J. E. and Ripps, H., Effect of magnesium on horizontal cell activity in the skate retina, Nature (Lond.), 242 (1973) 101-103.

10 Ehinger, B., Glial and neuronal uptake of GABA, glutamic acid, glutamine and glutathione in the rabbit retina, Exp. Eye Res., 25 (1977) 221-234.

11 Fonnum, F. and Malthe-Sorenssen, D., Localisation of glutamate neurons, In P. J. Roberts, J. Storm-Mat- hisen and G. A. R. Johnston (Ed.), Glutamate: Trans- mitter in the Central Nervous System, Wiley, Chichester, 1981.

120

12 Goodchild, M. and Neal, M. J., The uptake of [:~H]- 7-aminobutyric acid by the retina, Brit. J. Pharmacol., 47 (1973) 529-542.

13 Guerinot, F. and Pessac, B., Uptake of y-aminobutyric acid and glutamic acid decarboxylase activity in chick embryo neuroretinas in monolayer cultures, Brain Res., 162 (1979) 179-183.

14 Hyndman, A. G. and Adler, R., Analysis of glutamate uptake and monosodium glutamate toxicity in neural retina monolayer cultures, Develop. Brain Res., 2 (1982) 303-314.

15 Hyndman, A. G. and Adler, R., GABA uptake and release in purified neuronal and nonneuronal cultures from chick embryo retina, Develop. Brain Res., 3 (1982) 167 180.

16 Kong, Y. C., Fung, S. C. and Lam, D. M. K., Post- natal development of glycinergic neurons in the rabbit retina, J. comp. Neurol., 193 (1980) 1127-1135.

17 Lam, D. M. K. and Hollyfield, J. G., Localization of putative amino acid neurotransmitters in the human retina, Exp. Eye Res., 31 (1980) 729-732.

18 Marc, R. E. and Lam, D. M. K., Glycinergic path- ways in the goldfish retina, J. Neurosci., 1 (1980) 152- 165.

19 Marshall, J. and Voaden, M. J., An autoradiographic study of the cells accumulating [3H]-y-aminobutyric acid in the isolated retinas of pigeons and chickens, hivest. Ophthalmol., 13 (1974) 602-607.

20 Mirsky, R., Wendon, L. M. B., Black, P., Stolkin, C. and Bray, D., Tetanus toxin: a cell surface marker for neurones in culture, Brain Res., 148 (1978) 251-259.

21 Neal, M. J., Amino acid transmitter substances in the vertebrate retina, Gen. Pharmacol., 7 (1976) 321-332.

22 Neal, M. J., Collins, G. G. and Massey, S. C., Inhibi- tion of aspartate release from the retina of the anaes- thetised rabbit by stimulation with light flashes, Neu- rosci. Lett., 14 (1979) 241-245.

23 Neal, M. J. and Iversen, L. L., Autoradiographic localisation of [3H]GABA in rat retina, Nature New Biol., 235 (1972) 217-218.

24 Neal, M. J., Peacock, D. J. and White, R. D., Kinetic analysis of amino acid uptake by the rat retina in vitro, Brit. J. Pharmacol., 47 (1973) 656-657.

25 Osborne, N. N., Nesselhut, T., Nicholas, D. A., Patel, S. and Cuello, A. C., Serotonin-containing neurones in vertebrate retinas, J. Neurochem., 39 (1982) 1519-1528.

26 Perry, V. H. and Walker, M., Amacrine cells, displaced

amacrine cells and interptexiform cells in the retina of the rat, Proc. Roy. Soc. Lond. B., 208 (1980) 415-431.

27 Pomerai, D. I. and Carr, A., GABA and choline ac- cumulation by cultures of chick embryo neuroretinal cells, Exp. Eye Res., 34 (1982) 553-563.

28 Pourcho, R. G., Uptake of ['~H]glycine and [ZH]GABA by amacrine cells in the cat retina, Brain Res., 198 (1980) 333-346.

29 Rayborn, M. E., Sarthy, P. V., Hollyfield, J. G. and Lain, D. M. K., The emergence, Iocalisation and matur- ation of neurotransmitter systems during development of the retina in Xenopus laevis. II Glycine, J. eomp. Neurol., 195 (1980) 585-593.

30 Starr, M. S. and Voaden, M. J., The uptake of ~4C- l~-aminobutyric acid by the isolated retina of the rat, Vision Res., 12 (1972) 549-557.

31 Takagaki, G., Sodium and potassium ions and accumu- lation of labelled D-aspartate and GABA in crude synaptosomal fraction from rat cerebral cortex, J. Neurochem., 30 (1978) 47-56.

32 Vaughn, J. E., Famiglietti, E. V., Barber, R. P., Saito, K., Roberts, E. and Ribak, C. E., GABAergic amacrine cells in rat retina: immunocytochemical identification and synaptic connectivity, J. comp. Neurol., 197 (1981) t 13-127.

33 Voaden, M. J., Gamma-aminobutyric acid and glycine as retinal neurotransmitter. In S. L. Bonting (Ed.), Transmitters in the Visual Process, Pergamon, Oxford, 1976, pp. 107-125.

34 Voaden, M. J., Morjaria, B. and Oraedu, A. C. I., The Iocalisation and metabolism of glutamate, aspartate and GABA in the rat retina, Neurochem., 1 (1980) 151- 165.

35 Walker, C. R. and Peacock, J. H., Development of GABAergic function of dissociated hippocampal cultures from fetal mice, Develop. Brain Res., 2 (1982) 541- 555.

36 White, R. D. and Neal, M. J., The uptake of L-glutamate by the retina, Brain Res., 111 (1976) 79-93.

37 White, W. F., Snodgrass, S. R. and Dichter, M., Iden- tification of GABA neurons in rat cortical cultures by GABA uptake autoradiography, Brain Res., 190 (1980) 139-152.

38 Wu, S. M. and Dowling, J. E., L-Aspartate: evidence for a role in cone photoreceptor transmission in the carp retina, Proc. nat. Acad. Sci. U.S.A., 75 (1978) 5205-5209.

selective uptake of tritiated glycine, gaba and d-aspartate by retinal cells in culture: a study...

Documents