Brain Research, 564 (199l) 19-26 © 1991 Elsevier Science Publishers B.V. All rights reserved. 0006-8993/91/$03.50 ADONIS 0006899391171318
t9
BRES 17131
Localization of GAD- and GABA-like immunoreactivity in ground squirrel retina: retrograde labeling demonstrates GAD-positive ganglion cells Nidza Lugo-Garcia and Rosa E. Blanco Institute of Neurobiology and Department of Anatomy, University of Puerto Rico, San Juan, PR 00901 (Puerto Rico) (Accepted 18 June 1991)
Key words: Retina: Amacrine cell; Ganglion cell; Glutamic acid decarboxylase; 7-Aminobutyric acid; Immunohistochemistry; Ground squirrel
Glutamic acid decarboxylase (GAD)- and 7-aminobutyric acid (GABA)-like immunoreactivity was examined in the retina of the 13-lined ground squirrel (Spermophilus tridecemlineatus). Labeling was observed in the inner nuclear layer (INL), inner plexiform layer (IPL) and ganglion cell layer (GCL). The immunoreactive cell bodies in the inner third of the INL were 6-13 ~m in diameter and, because of their size and location it was considered that these were amacrine cells. Labeling in the IPL was concentrated in 5 bands corresponding to laminae la, lc, 2, 4 and 5. In the GCL a heterogeneous population of neurons exhibited GAD- and GABA-like immunoreactivity. The soma diameters of the GCL cells ranged from 5 to 17/zm. These may represent displaced amacrines and/or ganglion cells. To determine if any of the immunoreactive cells in the GCL were ganglion cells, double labeling experiments were performed using rhodamine latex microspheres ('beads') as retrograde neuronal tracers. Rhodamine beads were injected into the superior colliculus, and retinas with retrogradely labeled ganglion cells were subsequently incubated with the anti-GAD antiserum. These experiments revealed a small population of GAD-positive ganglion cells, setting a lower limit for the total number of GABAergic ganglion cells. INTRODUCTION
active horizontal and bipolar cells have also been rep o r t e d in rabbit, cat and macaque m o n k e y retinas 8'14'15'
The neurotransmitter y-aminobutyric acid ( G A B A ) has been detected in the retina of several v e r t e b r a t e species. G A B A e r g i c structures were originally localized in autoradiographs after uptake of [ 3 H ] G A B A or its agonist, [3H]muscimolt7A9'22. More recently, immunocy-
18. In addition, some m a m m a l i a n retinas (i.e. cat and macaque) have labeled interplexiform cells 8'14'18'2°.
tochemical methods have been e m p l o y e d to localize G A B A or glutamic acid decarboxylase ( G A D ) , the G A B A biosynthetic enzyme 32a. G A B A and/or G A D immunoreactivity has been observed in several cell types and in various sublaminae of the inner plexiform layer (IPL). In retinas where a consistent sublayering pattern in the IPL is seen, the number of labeled bands varies from 3 to 7. Widths and staining intensities vary, depending on the species 1'2'4'9' 13,~8 Immunoreactive amacrine cells have been identified in all vertebrate species studied 2a. G A B A has also been localized in horizontal and bipolar cells in several m a m m a l i a n and non-mammalian species. G A B A e r g i c horizontal cells have been described in goldfish, skate, frog, bird and turtle retinas, and G A B A e r g i c bipolar neurons have been found in frog retina 1'9'tl. I m m u n o r e -
G A B A and G A D immunoreactivity has also been rep o r t e d in the ganglion cell layer ( G C L ) 1'9'13'J5, however, many of the immunoreactive neurons in this layer are displaced amacrine cells. Stronger evidence for the existence of G A B A e r g i c ganglion cells has been obtained from studies of rabbit, rat and turtle retinas. Yu et al. 25 d e m o n s t r a t e d that in rabbit retina ganglion cells identified with a cell-specific monoclonal antibody were also G A B A immunoreactive. In turtle and rat, ganglion cells which had been retrogradely labeled with r h o d a m i n e B isothiocyanate or Diamino-yellow, respectively, were also immunoreactive after being subsequently incubated with a G A B A antiserum 6'9. In this study we examine the immunocytochemical localization of G A D and G A B A in the retina of a mammal with a highly d e v e l o p e d dichromat visual system, the 13-lined ground squirrel (Spermophilus tridecemlineatus). G A B A - l i k e and G A D - l i k e immunoreactivity was observed in the inner nuclear layer (INL), IPL and G C L ,
Correspondence: N. Lugo-Garcia, University of Puerto Rico, Institute of Neurobiology, 201 Boulevard del Valle, San Juan, PR 00901, Puerto Rico.
20 w i t h s o m e d i f f e r e n c e s in t h e l a b e l i n g p a t t e r n o f t h e t w o antibodies. We demonstrate
that a small population of
g a n g l i o n cells w h i c h p r o j e c t t o t h e s u p e r i o r c o l l i c u l u s a r e GABAergic, of GABAergic
setting a lower limit for the total number g a n g l i o n cells. S o m e o f o u r f i n d i n g s h a v e
p r e v i o u s l y b e e n p r e s e n t e d in a b s t r a c t f o r m t2. MATERIALS AND METHODS A total of 17 thirteen-lined ground squirrels (Spermophilus tridecemlineatus) were used. These were obtained from TLS Research, Illinois. The animals were given an overdose of Nembutal and then perfused through the heart with 6% dextran in 0.1 M phosphate buffer (pH 7.2) followed by 4% paraformaldehyde in 0.1 M phosphate buffer when the antibody directed against G A D was used, or 4% paraformaldehyde + 0.3% glutaraidehyde in the same buffer when the anti-GABA antiserum was used. The eyes were immediately removed and the anterior chamber dissected away. Great care was taken when removing the vitreous body so as not to damage the GCL. The eyecups were postfixed in the same fixative at 4 °C for 4-6 h and then placed overnight in a solution of 30% sucrose in 0.1 M phosphate buffer. The tissue was frozen in embedding medium consisting of 50% OCT compound + 50% Aquamount (Baxter). Transverse sections 10-15/~m thick were cut using a eryostat, then mounted on gelatin-covered slides. Some eyecups were flattened, frozen and cut with a freezing microtome into horizontal sections, 30/~m thick, others were flattened and processed as wholemounts. These were reacted as 'free-floating' sections in essentially the same manner as the cryostat sections.
lmmunofluorescent techniques The tissue sections were washed in phosphate buffer and then incubated with the primary antibody. The anti-GAD (no. 1440-4) 16 and the anti-GABA (Chemicon) antisera were both used at a dilution of 1:1000 in phosphate buffer containing 0.3% Triton X-100 and 0.05% sodium azide. They were incubated overnight at 4 °C. After phosphate buffer washes, the tissue was incubated in an FITC-conjugated secondary antibody diluted 1:100 in phosphate buffer containing 0.3% Triton X-100. After a final wash in the same buffer the sections were coverslipped in glycerin-carbonate buffer mounting medium.
Avidin-biotin complex (ABC) method The tissue sections which were to be processed with the avidinbiotin technique were incubated in the primary antibody and washed with phosphate buffer as described above. They were then incubated in biotinylated secondary antibody (Vector Labs.) for 60 min at room temperature. Sections which had been treated with the anti-GAD primary antibody were incubated with biotinylated antisheep IgG (raised in rabbit); those which had been treated with the anti-GABA antiserum were incubated with biotinylated antirabbit IgG (raised in goat). Following washes in phosphate buffer, the sections were incubated for 45 min in a complex consisting of biotin and avidin bound to horseradish peroxidase (HRP) in phosphate buffer. After washing in phosphate buffer, they were incubated in a freshly mixed solution of 0.05% diaminobenzidine and 0.4% hydrogen peroxide. The tissue was then washed, intensified in 0.05% osmium tetroxide for 20 s, dehydrated in graded alcohol solutions, cleared with xylene and mounted with Permount.
sence of obvious cytotoxicity or pholotoxicity*° The animals were anesthetized with Nembutal (40 mg/kg) and then placed in a small animal stereotaxic apparatus. The skin overlying the skull was reflected and a dental drill used to perform a craniotomy over the superior colliculus. Stereotaxic coordinates were used to reach the injection site. Approximately 1 /~1 of the undiluted rhodamine beads was pressure injected into each superior colliculus. Several injections were made at each site. After completion of the injection procedure the skin was sutured and a survival period of at least 10 days was allowed for transport of the beads to take place. The subjects were then given a lethal dose of Nembutal and perfused through the heart with a 0.l M phosphatc buffer wash followed by a fixative consisting of 4% paraformaldchyde in 0.1 M phosphate buffer (pH 7.2). The eyes were removed, postfixed, and processed as wholemounts for immunofluorescence. Retinas processed as wholemounts required longer incubation periods. Optimal results were obtained when they were incubated for 2-3 days in the primary antibody and for 2 days in the secondary. In one case (Fig. 5) better penetration of the antibodies was obtained by submitting the retinas to a freeze-thaw procedurc 7. As fixation with glutaraldehyde is required for the reaction with the anti-GABA antibody, and glutaraldehyde fixation is not recommended when using the rhodamine beads, only the anti-GAD antiserum was employed in the double labeling experiments, The brains of animals which had been injected with rhodamine beads were sectioned with a freezing microtome into 40-/~m-thick sections. The sections were examined microscopically to ascertain the location of the injection sites.
Analysis of retinal sections and wholemounts Preparations were examined with a Nikon mircroscope equipped for epifluorescence and differential interference contrast optics. Profiles of labeled cells were drawn with a camera lucida and the equivalent circle diameter of each of these cells was calculated with a Zeiss ZIDAS image analysis system. This system computes the area of an irregularly shaped object and converts it into a circle of the same area whose diameter is then calculated. In each retina 100-200 randomly selected labeled cells were measured and a histogram was constructed to illustrate the relation between cell size and frequency. Forty transverse sections from different and representative areas of the retina were used in cell counts to determine the number of GAD- or GABA-labeled cells in the 1NL and GCL.
RESULTS GAD-like and GABA-Iike immunoreactivity was obs e r v e d in t h e I N L , I P L a n d G C L o f t h e g r o u n d s q u i r r e l r e t i n a . S o m e d i f f e r e n c e s w e r e f o u n d in t h e s t a i n i n g p a t tern of the two antisera. Of particular interest was the presence of a small population of immunoreactive gang l i o n cells.
lmmunoreactivity in the I N L In retinas incubated with the GABA
o r t h e G A D an-
t i s e r a i m m u n o r e a c t i v e cell b o d i e s w e r e o b s e r v e d in t h e i n n e r p o r t i o n o f t h e I N L , i m m e d i a t e l y a d j a c e n t to t h e I P L . T h e n e u r o n s w e r e a r r a n g e d in 1 - 4 r o w s (Fig. 1 A ,
Retrograde labeling experiments We used rhodamine latex microspheres or 'beads' (LumaFluor Inc., New York) as a retrograde neuronal tracer in double-labeling experiments designed to confirm the presence of immunoreactive retinal ganglion cells. This particular retrograde tracer offers several distinct advantages; of particular importance to these experiments being its compatibility with immunocytochemical procedures, high stability, resistance to fading under illumination and the ab-
B), depending on the quality of the staining and indep e n d e n t o f e c c e n t r i c i t y . L a b e l e d cell b o d i e s r a n g e d f r o m 6 t o 1 3 / ~ m in d i a m e t e r w i t h a m e a n d i a m e t e r o f 8.8 ± 0.1 # m ( ± S . E . M . , n = 139; Fig. 2). T h e s e l a b e l e d cells r e s e m b l e d a m a c r i n e cells. T h e r e w a s a d i f f e r e n c e in t h e n u m b e r a n d s t a i n i n g in-
21
Fig. 1. Fluorescein-labeled transverse sections of ground squirrel retinas demonstrating GABA-Iike (A) and GAD-like (B) immunoreactivity in the inner third of the INL, IPL and GCL. GABA-positive cell bodies are more numerous and more intensely stained than GAD-positive cell bodies. Tile GAD-like immunoreactivity is mostly concentrated in neuronal processes. Scale bars: 20/tin. tensity of putative amacrine cell bodies labeled with each of the two antibodies. The m e a n n u m b e r of labeled neurons per transverse retinal section was 12 - 2 (n = 20) for G A D and 36 --- 7 (n = 19) for G A B A . It was also observed that G A B A - i m m u n o r e a c t i v e amacrines stained more intensely (Fig. 1A) and that when the antibody against G A B A was used the immunostain formed a ring
around the cell nucleus (Fig. 1B).
Immunoreactivity in the IPL Heavily labeled processes were observed in retinas incubated with either the G A D or G A B A antisera. These processes formed 5 bands in laminae la, lc, 2, 4 and 5 of the IPL. They were more clearly identified in
22
Immunoreactivity in the GCL
4,0.
odl
j
~0.
~
,.,
20.
b_ o
10,
o
.
o
.
.
.
.
.
5
L r-'N Il L K ~ GCL
.
.
10 SOMA SIZE Gum)
B~ 15
w 20
Fig. 2. Histograms of the soma (equivalent circle) diameter of GAD-immunoreactive cells in the [NL and C-CL. Note that a small percentage of the stamed cells in the GCL have a larger soma diameter. Based on a total of ]39 cells in the INL and 260 in the
GCL.
incubated with the antibody against G A B A (Fig. 1A). The 5 bands corresponded to bands 1, 2, 4, 10 and 11 of the 11 dark and fight bands d e , b e d by West 21 in the ground squirrel retina. The innermost layer was the most intensely stained. GAD-like immunoreactivity was also localized in strata across the IPL, although no sublayer was entirely devoid of staining (Fig. 1B).
In transverse cryostat sections a single row of G A B A or GAD-positive cell bodies was observed, immediately adjacent to the most strongly GABA-immtmoreactive inner IPL stratum (Fig. 1A). Again, more somata were stained with the anti-GABA antibody than the anti-GAD. The mean number of labeled cells per section was 16 -+ 2 (n = 19) for G A D and 37 -+ 7 (n = 19) for G A B A The G C L cell bodies expressing GAD-like and GABA-Iike immunoreactivity appeared to make up a heterogeneous population. This heterogeneity could more readily be appreciated in horizontal sections and wholemounts (Fig. 3). Diameters of immunoreactive neurons in the G C L ranged from 5 to 17 # m with a mean diameter of 8.1 - 0.1 (n = 260; Fig. 2). The presence of a small population of larger ceils in the G C L should be noted in Fig. 3. G A B A - and GAD-immunoreactive cells in the G C L could represent displaced amacrine cells or retinal ganglion cells. The existence of a small population of larger neurons in the G C L that were not seen in the IPL supports the idea that some retinal ganglion cells were also stained as well as amacrines. To test this possibility
Fig. 3. Retinal wholemount containing GAD-immunoreactive neurons in the GCL (ABC method), viewed with differential interference contrast optics. Note the large labeled cell in the center. The asterisk marks a blood vessel. Scale bar: 10/~m.
23 we carried out double labeling experiments. Rhodamine beads were injected into the superior colliculus of 14 animals (Fig. 4A). Contralateral retinas of the injected animals exhibited distinctive retrogradely labeled ganglion cells ranging in diameter from 5 to 17/~m. In all retinas examined, the highest density of bead-containing gan-
glion cells was around the visual streak. Retinas with the retrogradely labeled retinal ganglion cells were subsequently incubated with the anti-GAD antiserum, followed by incubation in a secondary antibody conjugated to FITC. U p o n comparing the distribution of the two fluorescent markers a very small number of GAD-posi-
Fig. 4. A: low power photomicrograph of a cross section through the ground squirrel mesencephalon, viewed with bright-field illumination. The arrow indicates the site where rhodamine-labeled latex microspheres ('beads') were injected into the superior colliculus (SC). Scale bar: 0.5 mm. B: photomicrograph of a retinal wholemount containing ganglion cells which are retrogradely labeled with rhodamine beads. Scale bar: 10/~m.
24
J.
neurons ._.
in the GCL, viewed .. .. _
with B fluorescein C.. c.
filter set which totally blocks rhodamir
.
n
. ,..
. .
.. IW and arrowhead).
,
Scale bar: 10 ym. D: photo-
r..
.
25 tive, retrogradely-labeled ganglion cells was observed, usually in regions of the retina in which the density of r h o d a m i n e - l a b e l e d cells was low (Fig. 5 A - E ) . The doubly labeled cells ranged from 10 to 14/~m in diameter and represented approximately 1-3% of G A D - i m m u n o reactive cells. For technical reasons our injections of rhodamine beads were limited to small areas of the superior colliculus, so it is not possible to determine whether the numerous o t h e r G A D - p o s i t i v e cells (Fig. 5) represent additional ganglion cells which project to other target areas or are simply displaced amacrine cells. It is likely that the doubly labeled cells seen in this study represent a lower limit to the total n u m b e r of G A B A e r gic ganglion cells in the ground-squirrel retina. DISCUSSION In the present study, G A B A - and G A D - i m m u n o r e active cells bodies were found in the I N L and G C L , while immunoreactive processes were localized in the IPL. In general, the results obtained with the a n t i - G A D and a n t i - G A B A antisera corresponded. Nevertheless, more cells were labeled with the latter, indicating that the G A D antiserum does not stain all G A B A - c o n t a i n ing cells. This p h e n o m e n o n has also been observed in retinas from animals as diverse as the frog, pigeon, chicken and rabbit 1'15. Several reasons have been put forward to explain this difference in immunostaining 15. O n e important consideration is that G A D is only present in neurons which are involved in G A B A synthesis, while G A B A may be found in cells involved in synthesis or transmitter uptake. In addition, as G A D tends to concentrate in axon terminals, the lightly labeled cell bodies may escape detection. The immunoreactive neurons in the I N L were limited to the inner third of this layer, where amacrine cell bodies are located. Because of their location and size, and the fact that G A B A e r g i c amacrines have been described in many other species 24 we can safely conclude that G A B A e r g i c amacrine neurons are also present in ground squirrel retina. West 21 described at least 5 subtypes of amacrines in the ground squirrel retina based on soma size and the pattern of arborization within the IPL; however, the extensive immunostaining of the IPL in our study makes it impossible to distinguish between the subtypes. It is almost certain that m o r e than one subtype
REFERENCES 1 Agardh, E., Bruun, A., Ehinger, B., Ekstrom, R, van Veen, T. and Wu, J.-Y., Gamma-aminobutyric acid- and glutamic acid decarboxylase-immunoreactive neurons in the retina of different vertebrates, J. Comp. Neurol., 258 (1987) 622-630. 2 Brandon, C., Lam, D.M.K and Wu, J.-Y., The 7-aminobutyric
was stained, because there were 5 immunoreactive bands and few, if any, retinal neurons arborize in more than 3 strata. L a b e l e d cells in the G C L occupied a larger size range. These putative G A B A e r g i c neurons could in theory be either displaced amacrines or true retinal ganglion cells. The double-labeling experiments show that at least 1-3% of them are ganglion cells. This n u m b e r represents a lower limit because our injections were confined to the superior colliculus. It is possible that very few ganglion cells project to this structure, and that after injections into other retinal target areas more of these neurons would be labeled. The existence of G A B A e r g i c cells in the G C L of different vertebrate species has been d e m o n s t r a t e d by immunohistochemical studies 1'~5. However, it was not conclusively d e m o n s t r a t e d that any of these immunoreactive cells were definitely ganglion cells. The existence of putative G A B A e r g i c ganglion cells was first d e m o n s t r a t e d in the rabbit by Yu et al. 25, who did double-labeling experiments using a ganglion cell specific monoclonal antibody in combination with G A B A immunohistochemistry. Similar results were later obtained in the turtle and rat by combining retrograde labeling and immunohistochemistry 6"9. O u r results further extend the range of species in which putative G A B A e r g i c ganglion cells are found, which suggests that they are also present in most, if not all, vertebrate retinas. A physiological role for G A B A e r g i c amacrines has been suggested by experiments in which picrotoxin (a G A B A antagonist) abolished the directional and orientational selectivity of certain retinal ganglion cells 5'23, possibly by the abolition of lateral inhibition. However, the physiological role of G A B A e r g i c ganglion cells has yet to be determined.
Acknowledgements. The authors wish to thank Christine Laerack, Ivonne Santiago and Clarissa del Cueto for technical assistance. We are indebted to Drs. T. Hughes and H. Karten for providing guidance during the early immunohistochemical experiments. We also thank Dr. E. Kicliter for help with an earlier version of the manuscript and Dr. J. Blagburn for his excellent comments and suggestions. Antiserum to GAD was provided through the Laboratory of Clinical Science, NIMH where it was developed under the supervision of Dr. I.J. Kopin with Drs. W. Oertel, D.W., Schmechel and M. Tappaz. This work was supported by ONR Grant N0001489-J-3070 to N.L.-G. and RCMI award RR-0351.
acid system in rabbit retina: localization by immunocytochemistry and autoradiography, Proc. Natl. Acad. Sci. U.S.A., 76 (1979) 3557-356l. 3 Brecha, N.C., Eldred, W.. Kuljis, R.D. and Karten, H.J., Identification and localization of biologically active peptides in the vertebrate retina. In N. Osborne and G. Chader (Eds.), Progress in Retinal Research, Vol. 3, Pergamon, New York, 1984, pp.
26 185-226. 4 Brecha, N.C. and Karten, H.J., Localization of biologically active peptides in the retina. In W.W. Morgan (Ed.), Retinal Transmitters and Modulators: Models for the Brain, CRC, Florida, 1985, pp. 93-118. 5 Caldwell, J.H., Daw, N.W. and Wyatt, H.J., Effects of picrotoxin and strychnine on rabbit retinal ganglion cells: lateral interactions for cells with more complex receptive fields, J. Physiol., 276 (1978) 277-298. 6 Caruso, D.M., Owczarzak, M,T., Goebel, D.J., Hazlett, J.C. and Pourcho, R.G., GABA-immunoreaetivity in ganglion cells of the rat retina, Brain Research, 476 (1989) 129-134. 7 Eldred, W.D., Zucker, C., Karten, H.J. and Yazulla, S., Comparison of fixation and penetration enhancement techniques for use in ultrastructural immunocytochemistry, J. Histochem. Cytochem., 31 (1983) 285-292. 8 Grunert, U. and Wassle, H., GABA-Iike immunoreactivity in the macaque monkey retina: a light and electron microscopic study, J. Comp. Neurol., 297 (1990) 509-524. 9 Hurd, II, L.B. and Eldred, W.D., Localization of GABA- and GAD-like immunoreaetivity in the turtle retina, Visual Neurosci., 3 (1989) 9-20. 10 Katz, L.C., Burkhalter, A. and Dreyer, W.J., Fluorescent latex microspheres as a retrograde marker for in vivo and in vitro studies of visual cortex, Nature, 310 (1984) 498-500. 11 Lain, D.M.K., Su, Y.Y.T., Swain, L., Marc, R.E., Brandon, C. and Wu, J.-Y., Immunocytochemical localization of L-glutamic acid decarboxylase in the goldfish retina, Nature, 278 (1979) 565-567. 12 Lugo-Gareia, N., Bianco, R.E., Hughes, T.E. and Karten, H.J., Localization of GAD-like and GABA-like immunoreactivity in the ground squirrel retina, Anat. Rec., 226 (1990) 60A. 13 Mariani, A.P. and Caserta, M.T., Electron microscopy of glutamate decarboxylase (GAD) immunoreactivity in the inner plexiform layer of the rhesus monkey retina, Z Neurocytol., 15 (1986) 645-655. 14 Mosinger, J.L., Yazulla, S. and Studholme, K.M., GABA-like immunoreactivity in the vertebrate retina: a species comparison,
Exp. Eye Res., 42 (1986) 631-644. 15 Mosinger, J. and Yazulla, S., Double-label analysis of GADand GABA-Iike immunoreactivity in the rabbit retina, Vision Res., 27 (1987) 23-30. 16 Oertel, W.H., Tappaz, M.L., Kopin, l.J., Ransom, D.H. and Schmechel, D.E., Production of an antiserum to rat brain glutamate (GAD)/cysteine sulfinate (CSD) decarboxylase, Brain Res. Bull., 5 (1980) 713-719. 17 Pourcho, R.G., Goebel, D.J. and McReynolds, J.S., Autoradiographic studies of [3H]glyeine, [3HIGABA, and [3H]muscimol uptake in the mudpuppy retina, Exp. Eye Res., 39 (t984) 6981. 18 Pourcho, R.G. and Owczarzak, M.T., Distribution of GABA immunoreactivity in the cat retina: a light- and electron-microscopic study, Visual Neurosci., 2 (1989) 425-435. 19 Voaden, M.J., Marshall, J. and Murani, N., The uptake of 3ny-aminobutyric acid and 3H-glycineby the isolated retina of the frog, Brain Research, 67 (1974) 115-132. 20 Wassle, H. and Chun, M.H., GABA-Iike immunoreactivity in the cat retina: light microscopy, J. Comp. NeuroL, 279 (1989) 43-54. 21 West, R.W., Light and electron microscopy of the ground squirrel retina: functional considerations, J. Comp. Neurol., 168 (1976) 355-378. 22 Wu, J.-Y., Brandon, C., Su, Y.Y. and Lam, D.M~K.~Immunocytochemical and autoradiographic localization of GABA system in the vertebrate retina, Mol. Cell Biochem., 39 (1981) 229-238. 23 Wyatt, H.J. and Daw, N.W., Specific effects of neurotransmitter antagonists on ganglion cells in rabbit retina, Science, 191 (1976) 204-205. 24 Yazulla, S., GABAergic mechanisms in the retina, In N. Osborne and G. Chader (Eds.), Progress in Retinal Research, Vol. 5, Pergamon, New York, 1986, pp. 1-52. 25 Yu, B.C.-Y., Watt, C.B., Lam, D.M.K. and Fry, K.R., GABAergic ganglion cells in the rabbit retina, Brain Research, 439 (1988) 376-382.
t9
BRES 17131
Localization of GAD- and GABA-like immunoreactivity in ground squirrel retina: retrograde labeling demonstrates GAD-positive ganglion cells Nidza Lugo-Garcia and Rosa E. Blanco Institute of Neurobiology and Department of Anatomy, University of Puerto Rico, San Juan, PR 00901 (Puerto Rico) (Accepted 18 June 1991)
Key words: Retina: Amacrine cell; Ganglion cell; Glutamic acid decarboxylase; 7-Aminobutyric acid; Immunohistochemistry; Ground squirrel
Glutamic acid decarboxylase (GAD)- and 7-aminobutyric acid (GABA)-like immunoreactivity was examined in the retina of the 13-lined ground squirrel (Spermophilus tridecemlineatus). Labeling was observed in the inner nuclear layer (INL), inner plexiform layer (IPL) and ganglion cell layer (GCL). The immunoreactive cell bodies in the inner third of the INL were 6-13 ~m in diameter and, because of their size and location it was considered that these were amacrine cells. Labeling in the IPL was concentrated in 5 bands corresponding to laminae la, lc, 2, 4 and 5. In the GCL a heterogeneous population of neurons exhibited GAD- and GABA-like immunoreactivity. The soma diameters of the GCL cells ranged from 5 to 17/zm. These may represent displaced amacrines and/or ganglion cells. To determine if any of the immunoreactive cells in the GCL were ganglion cells, double labeling experiments were performed using rhodamine latex microspheres ('beads') as retrograde neuronal tracers. Rhodamine beads were injected into the superior colliculus, and retinas with retrogradely labeled ganglion cells were subsequently incubated with the anti-GAD antiserum. These experiments revealed a small population of GAD-positive ganglion cells, setting a lower limit for the total number of GABAergic ganglion cells. INTRODUCTION
active horizontal and bipolar cells have also been rep o r t e d in rabbit, cat and macaque m o n k e y retinas 8'14'15'
The neurotransmitter y-aminobutyric acid ( G A B A ) has been detected in the retina of several v e r t e b r a t e species. G A B A e r g i c structures were originally localized in autoradiographs after uptake of [ 3 H ] G A B A or its agonist, [3H]muscimolt7A9'22. More recently, immunocy-
18. In addition, some m a m m a l i a n retinas (i.e. cat and macaque) have labeled interplexiform cells 8'14'18'2°.
tochemical methods have been e m p l o y e d to localize G A B A or glutamic acid decarboxylase ( G A D ) , the G A B A biosynthetic enzyme 32a. G A B A and/or G A D immunoreactivity has been observed in several cell types and in various sublaminae of the inner plexiform layer (IPL). In retinas where a consistent sublayering pattern in the IPL is seen, the number of labeled bands varies from 3 to 7. Widths and staining intensities vary, depending on the species 1'2'4'9' 13,~8 Immunoreactive amacrine cells have been identified in all vertebrate species studied 2a. G A B A has also been localized in horizontal and bipolar cells in several m a m m a l i a n and non-mammalian species. G A B A e r g i c horizontal cells have been described in goldfish, skate, frog, bird and turtle retinas, and G A B A e r g i c bipolar neurons have been found in frog retina 1'9'tl. I m m u n o r e -
G A B A and G A D immunoreactivity has also been rep o r t e d in the ganglion cell layer ( G C L ) 1'9'13'J5, however, many of the immunoreactive neurons in this layer are displaced amacrine cells. Stronger evidence for the existence of G A B A e r g i c ganglion cells has been obtained from studies of rabbit, rat and turtle retinas. Yu et al. 25 d e m o n s t r a t e d that in rabbit retina ganglion cells identified with a cell-specific monoclonal antibody were also G A B A immunoreactive. In turtle and rat, ganglion cells which had been retrogradely labeled with r h o d a m i n e B isothiocyanate or Diamino-yellow, respectively, were also immunoreactive after being subsequently incubated with a G A B A antiserum 6'9. In this study we examine the immunocytochemical localization of G A D and G A B A in the retina of a mammal with a highly d e v e l o p e d dichromat visual system, the 13-lined ground squirrel (Spermophilus tridecemlineatus). G A B A - l i k e and G A D - l i k e immunoreactivity was observed in the inner nuclear layer (INL), IPL and G C L ,
Correspondence: N. Lugo-Garcia, University of Puerto Rico, Institute of Neurobiology, 201 Boulevard del Valle, San Juan, PR 00901, Puerto Rico.
20 w i t h s o m e d i f f e r e n c e s in t h e l a b e l i n g p a t t e r n o f t h e t w o antibodies. We demonstrate
that a small population of
g a n g l i o n cells w h i c h p r o j e c t t o t h e s u p e r i o r c o l l i c u l u s a r e GABAergic, of GABAergic
setting a lower limit for the total number g a n g l i o n cells. S o m e o f o u r f i n d i n g s h a v e
p r e v i o u s l y b e e n p r e s e n t e d in a b s t r a c t f o r m t2. MATERIALS AND METHODS A total of 17 thirteen-lined ground squirrels (Spermophilus tridecemlineatus) were used. These were obtained from TLS Research, Illinois. The animals were given an overdose of Nembutal and then perfused through the heart with 6% dextran in 0.1 M phosphate buffer (pH 7.2) followed by 4% paraformaldehyde in 0.1 M phosphate buffer when the antibody directed against G A D was used, or 4% paraformaldehyde + 0.3% glutaraidehyde in the same buffer when the anti-GABA antiserum was used. The eyes were immediately removed and the anterior chamber dissected away. Great care was taken when removing the vitreous body so as not to damage the GCL. The eyecups were postfixed in the same fixative at 4 °C for 4-6 h and then placed overnight in a solution of 30% sucrose in 0.1 M phosphate buffer. The tissue was frozen in embedding medium consisting of 50% OCT compound + 50% Aquamount (Baxter). Transverse sections 10-15/~m thick were cut using a eryostat, then mounted on gelatin-covered slides. Some eyecups were flattened, frozen and cut with a freezing microtome into horizontal sections, 30/~m thick, others were flattened and processed as wholemounts. These were reacted as 'free-floating' sections in essentially the same manner as the cryostat sections.
lmmunofluorescent techniques The tissue sections were washed in phosphate buffer and then incubated with the primary antibody. The anti-GAD (no. 1440-4) 16 and the anti-GABA (Chemicon) antisera were both used at a dilution of 1:1000 in phosphate buffer containing 0.3% Triton X-100 and 0.05% sodium azide. They were incubated overnight at 4 °C. After phosphate buffer washes, the tissue was incubated in an FITC-conjugated secondary antibody diluted 1:100 in phosphate buffer containing 0.3% Triton X-100. After a final wash in the same buffer the sections were coverslipped in glycerin-carbonate buffer mounting medium.
Avidin-biotin complex (ABC) method The tissue sections which were to be processed with the avidinbiotin technique were incubated in the primary antibody and washed with phosphate buffer as described above. They were then incubated in biotinylated secondary antibody (Vector Labs.) for 60 min at room temperature. Sections which had been treated with the anti-GAD primary antibody were incubated with biotinylated antisheep IgG (raised in rabbit); those which had been treated with the anti-GABA antiserum were incubated with biotinylated antirabbit IgG (raised in goat). Following washes in phosphate buffer, the sections were incubated for 45 min in a complex consisting of biotin and avidin bound to horseradish peroxidase (HRP) in phosphate buffer. After washing in phosphate buffer, they were incubated in a freshly mixed solution of 0.05% diaminobenzidine and 0.4% hydrogen peroxide. The tissue was then washed, intensified in 0.05% osmium tetroxide for 20 s, dehydrated in graded alcohol solutions, cleared with xylene and mounted with Permount.
sence of obvious cytotoxicity or pholotoxicity*° The animals were anesthetized with Nembutal (40 mg/kg) and then placed in a small animal stereotaxic apparatus. The skin overlying the skull was reflected and a dental drill used to perform a craniotomy over the superior colliculus. Stereotaxic coordinates were used to reach the injection site. Approximately 1 /~1 of the undiluted rhodamine beads was pressure injected into each superior colliculus. Several injections were made at each site. After completion of the injection procedure the skin was sutured and a survival period of at least 10 days was allowed for transport of the beads to take place. The subjects were then given a lethal dose of Nembutal and perfused through the heart with a 0.l M phosphatc buffer wash followed by a fixative consisting of 4% paraformaldchyde in 0.1 M phosphate buffer (pH 7.2). The eyes were removed, postfixed, and processed as wholemounts for immunofluorescence. Retinas processed as wholemounts required longer incubation periods. Optimal results were obtained when they were incubated for 2-3 days in the primary antibody and for 2 days in the secondary. In one case (Fig. 5) better penetration of the antibodies was obtained by submitting the retinas to a freeze-thaw procedurc 7. As fixation with glutaraldehyde is required for the reaction with the anti-GABA antibody, and glutaraldehyde fixation is not recommended when using the rhodamine beads, only the anti-GAD antiserum was employed in the double labeling experiments, The brains of animals which had been injected with rhodamine beads were sectioned with a freezing microtome into 40-/~m-thick sections. The sections were examined microscopically to ascertain the location of the injection sites.
Analysis of retinal sections and wholemounts Preparations were examined with a Nikon mircroscope equipped for epifluorescence and differential interference contrast optics. Profiles of labeled cells were drawn with a camera lucida and the equivalent circle diameter of each of these cells was calculated with a Zeiss ZIDAS image analysis system. This system computes the area of an irregularly shaped object and converts it into a circle of the same area whose diameter is then calculated. In each retina 100-200 randomly selected labeled cells were measured and a histogram was constructed to illustrate the relation between cell size and frequency. Forty transverse sections from different and representative areas of the retina were used in cell counts to determine the number of GAD- or GABA-labeled cells in the 1NL and GCL.
RESULTS GAD-like and GABA-Iike immunoreactivity was obs e r v e d in t h e I N L , I P L a n d G C L o f t h e g r o u n d s q u i r r e l r e t i n a . S o m e d i f f e r e n c e s w e r e f o u n d in t h e s t a i n i n g p a t tern of the two antisera. Of particular interest was the presence of a small population of immunoreactive gang l i o n cells.
lmmunoreactivity in the I N L In retinas incubated with the GABA
o r t h e G A D an-
t i s e r a i m m u n o r e a c t i v e cell b o d i e s w e r e o b s e r v e d in t h e i n n e r p o r t i o n o f t h e I N L , i m m e d i a t e l y a d j a c e n t to t h e I P L . T h e n e u r o n s w e r e a r r a n g e d in 1 - 4 r o w s (Fig. 1 A ,
Retrograde labeling experiments We used rhodamine latex microspheres or 'beads' (LumaFluor Inc., New York) as a retrograde neuronal tracer in double-labeling experiments designed to confirm the presence of immunoreactive retinal ganglion cells. This particular retrograde tracer offers several distinct advantages; of particular importance to these experiments being its compatibility with immunocytochemical procedures, high stability, resistance to fading under illumination and the ab-
B), depending on the quality of the staining and indep e n d e n t o f e c c e n t r i c i t y . L a b e l e d cell b o d i e s r a n g e d f r o m 6 t o 1 3 / ~ m in d i a m e t e r w i t h a m e a n d i a m e t e r o f 8.8 ± 0.1 # m ( ± S . E . M . , n = 139; Fig. 2). T h e s e l a b e l e d cells r e s e m b l e d a m a c r i n e cells. T h e r e w a s a d i f f e r e n c e in t h e n u m b e r a n d s t a i n i n g in-
21
Fig. 1. Fluorescein-labeled transverse sections of ground squirrel retinas demonstrating GABA-Iike (A) and GAD-like (B) immunoreactivity in the inner third of the INL, IPL and GCL. GABA-positive cell bodies are more numerous and more intensely stained than GAD-positive cell bodies. Tile GAD-like immunoreactivity is mostly concentrated in neuronal processes. Scale bars: 20/tin. tensity of putative amacrine cell bodies labeled with each of the two antibodies. The m e a n n u m b e r of labeled neurons per transverse retinal section was 12 - 2 (n = 20) for G A D and 36 --- 7 (n = 19) for G A B A . It was also observed that G A B A - i m m u n o r e a c t i v e amacrines stained more intensely (Fig. 1A) and that when the antibody against G A B A was used the immunostain formed a ring
around the cell nucleus (Fig. 1B).
Immunoreactivity in the IPL Heavily labeled processes were observed in retinas incubated with either the G A D or G A B A antisera. These processes formed 5 bands in laminae la, lc, 2, 4 and 5 of the IPL. They were more clearly identified in
22
Immunoreactivity in the GCL
4,0.
odl
j
~0.
~
,.,
20.
b_ o
10,
o
.
o
.
.
.
.
.
5
L r-'N Il L K ~ GCL
.
.
10 SOMA SIZE Gum)
B~ 15
w 20
Fig. 2. Histograms of the soma (equivalent circle) diameter of GAD-immunoreactive cells in the [NL and C-CL. Note that a small percentage of the stamed cells in the GCL have a larger soma diameter. Based on a total of ]39 cells in the INL and 260 in the
GCL.
incubated with the antibody against G A B A (Fig. 1A). The 5 bands corresponded to bands 1, 2, 4, 10 and 11 of the 11 dark and fight bands d e , b e d by West 21 in the ground squirrel retina. The innermost layer was the most intensely stained. GAD-like immunoreactivity was also localized in strata across the IPL, although no sublayer was entirely devoid of staining (Fig. 1B).
In transverse cryostat sections a single row of G A B A or GAD-positive cell bodies was observed, immediately adjacent to the most strongly GABA-immtmoreactive inner IPL stratum (Fig. 1A). Again, more somata were stained with the anti-GABA antibody than the anti-GAD. The mean number of labeled cells per section was 16 -+ 2 (n = 19) for G A D and 37 -+ 7 (n = 19) for G A B A The G C L cell bodies expressing GAD-like and GABA-Iike immunoreactivity appeared to make up a heterogeneous population. This heterogeneity could more readily be appreciated in horizontal sections and wholemounts (Fig. 3). Diameters of immunoreactive neurons in the G C L ranged from 5 to 17 # m with a mean diameter of 8.1 - 0.1 (n = 260; Fig. 2). The presence of a small population of larger ceils in the G C L should be noted in Fig. 3. G A B A - and GAD-immunoreactive cells in the G C L could represent displaced amacrine cells or retinal ganglion cells. The existence of a small population of larger neurons in the G C L that were not seen in the IPL supports the idea that some retinal ganglion cells were also stained as well as amacrines. To test this possibility
Fig. 3. Retinal wholemount containing GAD-immunoreactive neurons in the GCL (ABC method), viewed with differential interference contrast optics. Note the large labeled cell in the center. The asterisk marks a blood vessel. Scale bar: 10/~m.
23 we carried out double labeling experiments. Rhodamine beads were injected into the superior colliculus of 14 animals (Fig. 4A). Contralateral retinas of the injected animals exhibited distinctive retrogradely labeled ganglion cells ranging in diameter from 5 to 17/~m. In all retinas examined, the highest density of bead-containing gan-
glion cells was around the visual streak. Retinas with the retrogradely labeled retinal ganglion cells were subsequently incubated with the anti-GAD antiserum, followed by incubation in a secondary antibody conjugated to FITC. U p o n comparing the distribution of the two fluorescent markers a very small number of GAD-posi-
Fig. 4. A: low power photomicrograph of a cross section through the ground squirrel mesencephalon, viewed with bright-field illumination. The arrow indicates the site where rhodamine-labeled latex microspheres ('beads') were injected into the superior colliculus (SC). Scale bar: 0.5 mm. B: photomicrograph of a retinal wholemount containing ganglion cells which are retrogradely labeled with rhodamine beads. Scale bar: 10/~m.
24
J.
neurons ._.
in the GCL, viewed .. .. _
with B fluorescein C.. c.
filter set which totally blocks rhodamir
.
n
. ,..
. .
.. IW and arrowhead).
,
Scale bar: 10 ym. D: photo-
r..
.
25 tive, retrogradely-labeled ganglion cells was observed, usually in regions of the retina in which the density of r h o d a m i n e - l a b e l e d cells was low (Fig. 5 A - E ) . The doubly labeled cells ranged from 10 to 14/~m in diameter and represented approximately 1-3% of G A D - i m m u n o reactive cells. For technical reasons our injections of rhodamine beads were limited to small areas of the superior colliculus, so it is not possible to determine whether the numerous o t h e r G A D - p o s i t i v e cells (Fig. 5) represent additional ganglion cells which project to other target areas or are simply displaced amacrine cells. It is likely that the doubly labeled cells seen in this study represent a lower limit to the total n u m b e r of G A B A e r gic ganglion cells in the ground-squirrel retina. DISCUSSION In the present study, G A B A - and G A D - i m m u n o r e active cells bodies were found in the I N L and G C L , while immunoreactive processes were localized in the IPL. In general, the results obtained with the a n t i - G A D and a n t i - G A B A antisera corresponded. Nevertheless, more cells were labeled with the latter, indicating that the G A D antiserum does not stain all G A B A - c o n t a i n ing cells. This p h e n o m e n o n has also been observed in retinas from animals as diverse as the frog, pigeon, chicken and rabbit 1'15. Several reasons have been put forward to explain this difference in immunostaining 15. O n e important consideration is that G A D is only present in neurons which are involved in G A B A synthesis, while G A B A may be found in cells involved in synthesis or transmitter uptake. In addition, as G A D tends to concentrate in axon terminals, the lightly labeled cell bodies may escape detection. The immunoreactive neurons in the I N L were limited to the inner third of this layer, where amacrine cell bodies are located. Because of their location and size, and the fact that G A B A e r g i c amacrines have been described in many other species 24 we can safely conclude that G A B A e r g i c amacrine neurons are also present in ground squirrel retina. West 21 described at least 5 subtypes of amacrines in the ground squirrel retina based on soma size and the pattern of arborization within the IPL; however, the extensive immunostaining of the IPL in our study makes it impossible to distinguish between the subtypes. It is almost certain that m o r e than one subtype
REFERENCES 1 Agardh, E., Bruun, A., Ehinger, B., Ekstrom, R, van Veen, T. and Wu, J.-Y., Gamma-aminobutyric acid- and glutamic acid decarboxylase-immunoreactive neurons in the retina of different vertebrates, J. Comp. Neurol., 258 (1987) 622-630. 2 Brandon, C., Lam, D.M.K and Wu, J.-Y., The 7-aminobutyric
was stained, because there were 5 immunoreactive bands and few, if any, retinal neurons arborize in more than 3 strata. L a b e l e d cells in the G C L occupied a larger size range. These putative G A B A e r g i c neurons could in theory be either displaced amacrines or true retinal ganglion cells. The double-labeling experiments show that at least 1-3% of them are ganglion cells. This n u m b e r represents a lower limit because our injections were confined to the superior colliculus. It is possible that very few ganglion cells project to this structure, and that after injections into other retinal target areas more of these neurons would be labeled. The existence of G A B A e r g i c cells in the G C L of different vertebrate species has been d e m o n s t r a t e d by immunohistochemical studies 1'~5. However, it was not conclusively d e m o n s t r a t e d that any of these immunoreactive cells were definitely ganglion cells. The existence of putative G A B A e r g i c ganglion cells was first d e m o n s t r a t e d in the rabbit by Yu et al. 25, who did double-labeling experiments using a ganglion cell specific monoclonal antibody in combination with G A B A immunohistochemistry. Similar results were later obtained in the turtle and rat by combining retrograde labeling and immunohistochemistry 6"9. O u r results further extend the range of species in which putative G A B A e r g i c ganglion cells are found, which suggests that they are also present in most, if not all, vertebrate retinas. A physiological role for G A B A e r g i c amacrines has been suggested by experiments in which picrotoxin (a G A B A antagonist) abolished the directional and orientational selectivity of certain retinal ganglion cells 5'23, possibly by the abolition of lateral inhibition. However, the physiological role of G A B A e r g i c ganglion cells has yet to be determined.
Acknowledgements. The authors wish to thank Christine Laerack, Ivonne Santiago and Clarissa del Cueto for technical assistance. We are indebted to Drs. T. Hughes and H. Karten for providing guidance during the early immunohistochemical experiments. We also thank Dr. E. Kicliter for help with an earlier version of the manuscript and Dr. J. Blagburn for his excellent comments and suggestions. Antiserum to GAD was provided through the Laboratory of Clinical Science, NIMH where it was developed under the supervision of Dr. I.J. Kopin with Drs. W. Oertel, D.W., Schmechel and M. Tappaz. This work was supported by ONR Grant N0001489-J-3070 to N.L.-G. and RCMI award RR-0351.
acid system in rabbit retina: localization by immunocytochemistry and autoradiography, Proc. Natl. Acad. Sci. U.S.A., 76 (1979) 3557-356l. 3 Brecha, N.C., Eldred, W.. Kuljis, R.D. and Karten, H.J., Identification and localization of biologically active peptides in the vertebrate retina. In N. Osborne and G. Chader (Eds.), Progress in Retinal Research, Vol. 3, Pergamon, New York, 1984, pp.
26 185-226. 4 Brecha, N.C. and Karten, H.J., Localization of biologically active peptides in the retina. In W.W. Morgan (Ed.), Retinal Transmitters and Modulators: Models for the Brain, CRC, Florida, 1985, pp. 93-118. 5 Caldwell, J.H., Daw, N.W. and Wyatt, H.J., Effects of picrotoxin and strychnine on rabbit retinal ganglion cells: lateral interactions for cells with more complex receptive fields, J. Physiol., 276 (1978) 277-298. 6 Caruso, D.M., Owczarzak, M,T., Goebel, D.J., Hazlett, J.C. and Pourcho, R.G., GABA-immunoreaetivity in ganglion cells of the rat retina, Brain Research, 476 (1989) 129-134. 7 Eldred, W.D., Zucker, C., Karten, H.J. and Yazulla, S., Comparison of fixation and penetration enhancement techniques for use in ultrastructural immunocytochemistry, J. Histochem. Cytochem., 31 (1983) 285-292. 8 Grunert, U. and Wassle, H., GABA-Iike immunoreactivity in the macaque monkey retina: a light and electron microscopic study, J. Comp. Neurol., 297 (1990) 509-524. 9 Hurd, II, L.B. and Eldred, W.D., Localization of GABA- and GAD-like immunoreaetivity in the turtle retina, Visual Neurosci., 3 (1989) 9-20. 10 Katz, L.C., Burkhalter, A. and Dreyer, W.J., Fluorescent latex microspheres as a retrograde marker for in vivo and in vitro studies of visual cortex, Nature, 310 (1984) 498-500. 11 Lain, D.M.K., Su, Y.Y.T., Swain, L., Marc, R.E., Brandon, C. and Wu, J.-Y., Immunocytochemical localization of L-glutamic acid decarboxylase in the goldfish retina, Nature, 278 (1979) 565-567. 12 Lugo-Gareia, N., Bianco, R.E., Hughes, T.E. and Karten, H.J., Localization of GAD-like and GABA-like immunoreactivity in the ground squirrel retina, Anat. Rec., 226 (1990) 60A. 13 Mariani, A.P. and Caserta, M.T., Electron microscopy of glutamate decarboxylase (GAD) immunoreactivity in the inner plexiform layer of the rhesus monkey retina, Z Neurocytol., 15 (1986) 645-655. 14 Mosinger, J.L., Yazulla, S. and Studholme, K.M., GABA-like immunoreactivity in the vertebrate retina: a species comparison,
Exp. Eye Res., 42 (1986) 631-644. 15 Mosinger, J. and Yazulla, S., Double-label analysis of GADand GABA-Iike immunoreactivity in the rabbit retina, Vision Res., 27 (1987) 23-30. 16 Oertel, W.H., Tappaz, M.L., Kopin, l.J., Ransom, D.H. and Schmechel, D.E., Production of an antiserum to rat brain glutamate (GAD)/cysteine sulfinate (CSD) decarboxylase, Brain Res. Bull., 5 (1980) 713-719. 17 Pourcho, R.G., Goebel, D.J. and McReynolds, J.S., Autoradiographic studies of [3H]glyeine, [3HIGABA, and [3H]muscimol uptake in the mudpuppy retina, Exp. Eye Res., 39 (t984) 6981. 18 Pourcho, R.G. and Owczarzak, M.T., Distribution of GABA immunoreactivity in the cat retina: a light- and electron-microscopic study, Visual Neurosci., 2 (1989) 425-435. 19 Voaden, M.J., Marshall, J. and Murani, N., The uptake of 3ny-aminobutyric acid and 3H-glycineby the isolated retina of the frog, Brain Research, 67 (1974) 115-132. 20 Wassle, H. and Chun, M.H., GABA-Iike immunoreactivity in the cat retina: light microscopy, J. Comp. NeuroL, 279 (1989) 43-54. 21 West, R.W., Light and electron microscopy of the ground squirrel retina: functional considerations, J. Comp. Neurol., 168 (1976) 355-378. 22 Wu, J.-Y., Brandon, C., Su, Y.Y. and Lam, D.M~K.~Immunocytochemical and autoradiographic localization of GABA system in the vertebrate retina, Mol. Cell Biochem., 39 (1981) 229-238. 23 Wyatt, H.J. and Daw, N.W., Specific effects of neurotransmitter antagonists on ganglion cells in rabbit retina, Science, 191 (1976) 204-205. 24 Yazulla, S., GABAergic mechanisms in the retina, In N. Osborne and G. Chader (Eds.), Progress in Retinal Research, Vol. 5, Pergamon, New York, 1986, pp. 1-52. 25 Yu, B.C.-Y., Watt, C.B., Lam, D.M.K. and Fry, K.R., GABAergic ganglion cells in the rabbit retina, Brain Research, 439 (1988) 376-382.
