Brain Research, 532 (1990) 197-202
197
Elsevier BRES 16008
Developmental immunoreactivity for GABA and GAD in the avian retina: possible alternative pathway for GABA synthesis Jan N. Hokoq, Ana L.M. Ventura, Patricia F. Gardino and Fernando G. De Mello Instituto de Biofisica Carlos Chagas Filho, Universidade Federal do Rio de Janeiro, Centro de CiOncias da Saade, Rio de Janeiro (Brazil)
(Accepted 15 May 1990) Key words: 7-Aminobutyric acid glutamate decarboxylase immunoreactivity; Development; Putrescine; Chicken retina
Although the distribution of GABAergic neurons in chick retina has been previously described by several investigators, the early appearance of these neurons has not been reported. In the present study immunohistochemical methods were used to localize GABAergic neurons with antisera to both GABA and its synthesizing enzyme, glutamate decarboxylase (GAD), in embryonic chick retina at several stages of development and beyond hatching. GABA-positive neuroblast-like cells were clearly detected in retinas as early as embryonic day 6. In contrast, GAD-containing cells were not observed in retinas until embryonic day 10. These findings indicated that immunocytochemically detectable amounts of GAD were not present in young GABAergic cells. Our data on the developmental appearance of GABA and GAD immunoreactivities are consistent with previous biochemical data for the development of GABA concentration and GAD activity in the chick retina. Together, these data suggest that retina cells from the early stages of development may synthesize GABA from an alternative pathway in which the most likely precursor is putrescine. INTRODUCTION Gamma-aminobutyric acid ( G A B A ) is generally regarded as a major inhibitory neurotransmitter in the central nervous system, including the retina of many vertebrate species, where its role is well established (see review in refs. 4, 17, 18, 36). A possible function for G A B A as a t r o p h i c or regulatory factor during neuronal development in vitro has also been suggested 7'14"28. At least 3 classes of G A B A neurons have been described in the retina of various species. G A B A is present predominantly in amacrine cells of mammalian retinas and amacrine and horizontal cells of non-mammalian retinas 3,12,15,35. More recently, G A B A interplexiform cells in the rat and rabbit retina have also been described19.20. 23. The existence of an 'exuberant' population of neurons stained by a n t i - G A B A serum which is not detected by a n t i - G A D serum has been reported in different areas of the nervous system by authors using different antiG A B A and a n t i - G A D sera 9,21'25'27'29. In the retina of several vertebrate species, GABA-immunoreactive amacrine cells are at least twice as numerous as G A D immunoreactive cells 1 and specifically in birds, G A B A and GAD-immunoreactive processes are distributed in different sublayers in the inner plexiform layer. In some regions of the developing nervous system
there is a marked disparity between the development of the content of endogenous G A B A and its biosynthetic enzyme, glutamic acid decarboxylase 6,z5. Inhibitory postsynaptic potentials that can be blocked by G A B A antagonists are also found earlier than the increase of GAD-activity 31 and, in other systems, postnatal development of G A D activity is taken as an index of G A B A e r g i c synaptogenesis 34. In addition to the classical pathway of G A B A synthesis from glutamate, G A B A can also be synthesized from putrescine in the central nervous system of different species 24. In early stages of chick retina development, a large proportion of G A B A present in the tissue comes from putrescine. This pathway is still present in mature tissue but is much less effective, accounting for less than 1% of the total G A B A 8. These observations raised the possibility that some retinal neurons might contain G A B A derived from the alternative pathway rather than G A B A produced by the classical, G A D - m e d i a t e d decarboxylation of glutamate. Putrescine is the precursor for G A B A in chick embryo brain as well 26. In the present study using an immunocytochemical approach we compared the appearance of G A B A and its synthetic enzyme G A D during development of the embryonic chick retina. Our data show that GAB,Aimmunoreactive cells are present early in development, at least 4 days prior to the appearance of G A D -
Correspondence: J.N. Hoko~, Departamento de Neurobiologia, Instituto de Biofisica Carlos Chagas F°, Universidade Federal do Rio de Janeiro, Centro de Cifncias da Safide, bloco G, 21944 Rio de Janeiro, Brazil.
0006-8993/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
198 immunoreactivity. The first GAD-containing neurons were detected during synapse formation in the inner plexiform layer. Our data strongly suggest that GABA in retinas from early embryonic stages comes from a source that utilizes putrescine as precursor.
MATERIALS AND METHODS Fertilized white Leghorn eggs were obtained from a local hatchery. Eyes from chick embryos ranging from 6 days of incubation 1° to various stages beyond hatching, were used. The eyes were enucleated, bisected with a razor, and the posterior pole immersed in fixative for 1 h. After several rinses in buffer, a piece of retina from the central region was dissected out from the sclera, and frozen sections 15-25/~m thick, were cut in the radial plane. The sections were mounted in gelatinized slides and incubated overnight with the primary antibody or kept frozen at -20 °C up to 2 months without loss of antigenicity. For GABA-immunoreactivity, retinas were fixed in 4% formaldehyde and 0.1% glutaraldehyde in 0.16 M phophate buffer. Retinal sections (15-25/~m) were incubated overnight at 4 °C with a rabbit antisera directed against a GABA/bovine serum albumin (BSA) conjugate 33 at a final dilution of 1:200. The GABA antisera were purified against glutamine, glutamate, aspartate, glycine, taurine, alanine and fl-alanine32. The primary reaction was visualized by the peroxidase-antiperoxidase (PAP) method 3°. Swine anti-rabbit Fab fragments (1:40) and rabbit PAP (1:80) (Dako) were used, followed by the diaminobenzidine reaction, used as a chromogen. Control sections were incubated either with normal rabbit serum or with the antibody preadsorbed with the amino acid conjugate GABA/ BSA 33. Immunocytochemical studies were then performed as described for the GABA-antisera-incubated sections. No immunoreactivity was detected in the controls. For GAD-immunoreactivity, retinas were fixed in 4% formaldehyde in 0.16 M phosphate buffer, pH 7.2. The sections were incubated overnight at 4 °C with GAD antiserum produced in sheep22 at a final dilution of 1:1000. Rabbit anti-sheep IgG, 1:150 (Sigma), and sheep PAP, 1:600 (Jackson), followed by diaminobenzidine (Sigma) as a chromogen, were used. Controls consisted of sections incubated either with normal sheep serum at a dilution of 1:1000 or with the preimmune serum 22 instead of the primary antisera and processed the same way, in parallel with the GAD antisera-exposed sections. No immunoreactivity was detected in the controls. Triton-X (0.25%, v/v) was used in all solutions to enhance permeability of the reagents. More concentrated primary antisera (1:400) was used for embryo retinas. More diluted primary antisera (1:2000) was better to visualize cell processes but not to label cell bodies. In 4 post-hatched animals 0.30 Id of colchicine 0.5% was injected on one of the eyes, for 24 h, in an attempt to label more cell bodies.
The sections were examined, studied and photographed under a light microscope.
RESULTS
In chick embryos, [3H]thymidine incorporation points to the posterior pole of the retina as the first region to complete mitotic activity 11. Thus, all the radial sections used for GABA- and GAD-immunoreactive studies, at different stages, were always taken from this specific region of the retina. GABA immunoreactivity was first detected in retinas at embryonic day 6 (E6). GABA was clearly detectable in a few cells neighboring the central retina, along a row 1/3 from the pigmented epithelium (Fig. 1A). On radial sections of 6-day-old embryo (E6) retinas, Coulombre 5 described the cells as elongated and forming an epithelium with nuclei arranged in several layers, except for the ganglion cell bodies, some of which were round and sent out axons. Thus, the GABA-positive cells detected at E6 could be either migrating neuroblasts or presumptive horizontal cell bodies. At higher magnification, some processes emerging from these cells could also be observed (Fig. 2D), possibly indicating horizontal cells undergoing a differentiation process. Unexpectedly, no GAD-immunoreactive cells could be detected at E6 (Fig.
1B). The next examined retina was at embryonic day 8 (E8). At this stage the retina is characterized by the appearance of inner and outer plexiform layers in the central region 5. The morphological aspect of cell bodies in the inner nuclear layer is still elongated, when compared to ganglion and horizontal cell somata. The immunohistochemistry for GABA showed labeled cell bodies in the neuroblastic layer adjacent to the inner plexiform layer (Fig. 1C). A row of cell bodies could also be distinguished along the presumptive location of horizontal cells, close to the prospective outer plexiform site (Fig. 2C). Near the central region of the retina, some GABA-positive cell somata were localized in the middle
Fig. 1. Light micrographs of 20/xm thick cryostat radial section from embryonic chick retinas at several stages of development and beyond hatching. A: retina section of the central region of an embryonic day 6 retina. GABA immunoreactivity is detected in neuroblast-like cells in a row close to the pigment epithelium. B: no immunoreactivity for GAD is detected in embryonic day 6 retina. C: retina section of the central region at embryonic day 8. GABA-immunoreactive cell bodies are localized in the inner nuclear layer, including horizontal and amacrine cells, and in the ganglion cell layer. Cells in migrating process can be seen in the middle of the inner plexiform layer. D: no immunoreactivity to GAD can be detected in day-8 embryonic retina. E: retina section of day-12 embryo retina incubated with GABA antiserum showing a profuse number of horizontal cell bodies, with process extended in only one stratum of the outer plexiform layer. Amacrine cell bodies and perikarya in the ganglion cell layer are also present. F: most of the GAD-containing cell bodies of day-12 embryonic retina are characterized as amacrine cells with processes extended in 3 well defined strata in the inner plexiform layer. Very weak label can be seen in horizontal cell bodies. G: GABA-immunoreactive labeling in mature chicken retina showing the immunoreactivity localized in horizontal cell bodies, several populations of amacrine cells and cell bodies in the ganglion cell layer. Labeled processes extended in the outer and inner strata of the outer plexiform layer and in the inner plexiform layer with no distinguishable layered organization. H: GAD-immunoreactivity in mature retina section is localized in a few cell bodies close to the vitreal side of the inner nuclear layer with processes extending in 3 distinct labeled bands in the inner plexiform layer. Weakly GAD-positive cell bodies are also labeled. Bar = 20/xm.
199 of the inner plexiform layer, resembling migrating cells. A few cells had processes extending toward the inner plexiform layer, these cell bodies can be displaced amacrine ceils in late migratory process, as suggested by
Layer and Vollmer 13. At the same stage (E8), GAD-positive neurons could still not be detected. Sections from 12-day-old embryo (El2) retinas already
....
~.7 L- . . . . .
-
~
:
~
....
-.--~
-
-~
- ~-L:~-i.~ZIZ
200 displayed the characteristic layered structure, except for the outer segments of the photoreceptor, which are not completely differentiated yet. At this stage, horizontal cell bodies at the outer portion of the inner nuclear layer had their processes extending along the outer plexiform layer. Thick processes from amacrine cells appeared and extended into the inner plexiform layer. By this time synaptogenesis begins in this layer5. GABA-immunoreactivity is detectable in El2 in the same types of cells as in the adult retina (Fig. 1E). Strongly reactive cell bodies of horizontal cells, with processes lying along the outer plexiform layer, seemed to be more numerous than in mature retina (Fig. 2B). Although weakly labeled, amacrine and ganglion cell somata were also present, resembling the adult pattern. A 3-strata pattern could be detected in the inner plexiform layer, which disappeared in older retinas (Fig. IE). At this stage (El2), GAD-containing amacrine cells were present along the inner portion of the inner nuclear layer. GAD-positive cytoplasm was concentrated at their vitreal side and processes of these cells extended radially along 3 laminae in the inner plexiform layer (Fig. 1F). Thus, GAD immunoreactivity in the chick retina is present at a later embryonic stage than GABA immunoreactivity. Finally, in posthatched chicken retinas, GABA immunoreactivity was localized in horizontal cells, several types of amacrine cells in the inner nuclear layer, and some perikarya in the ganglion cell layer (Fig. 1G), consistent with previous findings 1.
GABA-immunoreactive horizontal cells, ranging from 4 to 5 ,um in diameter, had their cell bodies lying along the outermost layer of the inner nuclear layer, with processes extending towards the sclera and spreading horizontally in the inner and outer strata of the outer plexiform layer (Fig. 2A). Although this label does not allow a precise definition of the dendrites as in silverimpregnated tissue, GABA-immunoreactive horizontal cells described here are likely to be of the H1 and H4 types, described for the pigeon retina 16. GABA-containing amacrine cell bodies, ranging from 4 to 9/zm in diameter, were along the inner margin and the vitreal half of the inner nuclear layer. Their processes in the inner plexiform layer extended from the outermost to the innermost layer, and were so numerous and intensely immunoreactive that no layered organization could be distinguished (Fig. 1G). In the ganglion cell layer, G A B A immunoreactive cell bodies were also of 4-9 ~zm size in diameter, corresponding to displaced amacrine cell bodies and possibly ganglion cell somata. No GABA-immunoreactive optic fibers were found in the nerve fiber layer. GAD immunoreactivity was observed in cell bodies in the inner nuclear layer and in the neuropil of the inner plexiform layer (Fig. 1H). Cell body sizes ranged from 5 to 10/xm in diameter, consisting of a thin rim of labeled cytoplasm, thickest toward the inner plexiform layer, and the nuclei were not immunoreactive. The inner plexiform layer consisted of 3 distinct labeled bands, corresponding to the outermost (13/xm thick), intermediate (9/~m thick) and innermost (20/xm) laminae, respectively. The intermediate band was the most intensely immunoreactive, followed by the outermost and the innermost. At any age up to El2, GAD-immunoreactivity was less intense than GABA-immunoreactivity in amacrine cell bodies. Although clearly positive for GABA, very weak labels were found for GAD in horizontal cell bodies (Fig. 1 E and F, and G and H). As expected from previous studies 1'3 ganglion cells were not GAD-immunoreactive in mature retinas. Occasionally, GAD-containing cell bodies were found in the ganglion cell layer, which by their soma diameter might correspond to displaced amacrine cells. No GAD- or GABA-immunoreactive interplexiform cells were found in chicken retina. DISCUSSION
Fig. 2. High magnification of GABA-immunoreactive horizontal cell population at mature retina (A); embryonic day 12 (B); embryonic day 8 (C) and embryonic day 6 (D). Bar = 50/xm.
The present study was undertaken to compare GABA and GAD immunoreactivity in embryonic chick retina. During development, GABA immunoreactivity was detected in neurons in the central retina as early as E6, in cells that display morphological characteristics of
201 migrating neuroblasts and/or committed neuroblasts at the end of their mitotic phase. In contrast, detectable levels of G A D were not found before El0-12, i.e., after a delay of 4 days. These data are unique since GAD, the synthesizing enzyme for G A B A , does not appear in neurons of younger preparations where GABA-positive neurons are clearly present. A few possibilities could explain this phenomenon: one is that the lack of immunoreactivity for G A D in cells from early-stage retinas could be due to the fact that the anti-GAD used in this work would not recognize a G A D isoform present in neuroblast-like cells of the mature tissue. This seems to be unlikely since our analysis was performed with a polyclonal antibody to GAD. This would substantially reduce the chances of a differential detection of GAD isoforms in this tissue. Another technical problem that might account for our data would be that if only trace amounts of G A D were present in the perikaryal cytoplasm of young neurons, our antibody would not be sensitive enough to detect the enzyme. This seems also to be questionable since the same antibody detects G A D immunoreactivity in amacrine cells of differentiated tissue, which display approximately the same intensity of reaction for G A B A as neuroblast like cells. Thus, since the immunoreactivity for G A B A seems to be similar in both cell populations (neuroblast and amacrine cells), one would expect that G A D (the limiting enzyme for G A B A synthesis) would be expressed in equivalent proportions in the two cell types. Finally, the third possible explanation for our observations would be that the G A B A present in neuroblastlike cells in the retina would come from an alternative route of synthesis in which putrescine would be the major precursor. This possibility is reinforced by the fact that G A B A can be synthesized from putrescine in several areas of the CNS 24, including the avian retina s. As mentioned above, during development the first GAD-containing neurons were detected between El0 and El2, thus prior to synapse formation in the inner plexiform layer 5. This observation is in agreement with a previous report 8. G A B A immunoreactivity, however, was detected in neurons in the central retina as early as E6. Previous reports 2,s have shown that ornithine decarboxylase activity, and consequently putrescine and other polyamines, are very high in undifferentiated cells with high mitotic activity or in cells at the end of their divisional cycle. After cells have withdrawn from the cell REFERENCES 1 Agardh, E., Bruun, A., Ehinger, B., Ekstrom, P., Van Veen, T. and Wu, J.Y., Gamma-aminobutyric acid- and glutamic acid decarboxylase immunoreactive neurons in the retina of different
cycle, this enzyme declines substantially as do putrescine and other polyamines. In the chick retina, the ornithine decarboxylase activity and putrescine content are extremely high in 6-day-old embryos and decline as the tissue differentiates. Concomitant with the decrease in putrescine content, retinal G A B A increases several-fold, preceding the increase in G A D activity by a few days. A substantial amount of G A B A in the early stages of development comes from putrescine s. The observation that G A B A is present in neuroblast-like cells, devoid of GAD immunoreactivity, strongly suggests that G A B A in these cells might be synthesized from putrescine. One major question that can be put forward, based on previous and present publication, is what would be the role of G A B A in retinal cells in embryos as early as the 6th day of incubation. Since several recent publications suggest a trophic role for G A B A in the CNS, an obvious idea would be that G A B A in the embryonic tissue would serve a t r o p h i c or controlling influence on developing retinal neurons. In cultured chick embryo retina cells the presence of G A B A in the incubating medium fully inhibits the expression of G A D activity during development, an effect that seems to be mediated, at least in part, by G A B A receptors 7. Since GAD-positive neurons in the retina are first detected at El0-12, when synapses start to form in the inner plexiform layer, it is possible that G A B A from an alternative source in the embryonic tissue could play a role in timing the precise moment during synaptogenesis in which committed GABAergic neurons would start to express GAD. Experiments from our laboratory, to be published elsewhere, show that the effect of G A B A in preventing the expression of G A D activity in cultured retina cells is reflected by a lack of immunoreactivity for G A D , suggesting that G A B A may be controlling the synthesis of the enzyme. If so, the neurotransmitter role for G A B A may be secondary to its metabolic or trophic function in the CNS of young developing animals.
Acknowledgements. We are grateful to Dr. Wenthold for his generous gift of antiserum to gamma-aminobutyric acid (#2) and GABA-BSA conjugate. Antiserum to GAD, as well as preimmune serum was provided through the laboratory of Clinical Science, NIMH, where it was developed under the supervision of Dr. Irwin J. Kopin with Drs. W. Oertel, D.E. Schmechel and M. Tappaz. We thank Dr. A.P. Mariani for critically reading this manuscript and Ms. Mair Medeiros for her technical assistance. This work was supported by grants from CNPq, FAPERJ and FINEP to J.N.H. and grants from FINEP, CNPq, FAPERJ and PEG to F.G.D.M.
vertebrates, J. Comp. Neurol. 258 (1987) 622-630. 2 Anderson, T.R. and Schamberg, S.M., Ornithine decarboxylase activity in developing rat brain, J. Neurochem., 19 (1972) 1471-1481. 3 Brandon, C., Retinal GABA neurons: localization in verte
202 brates species using an antiserum to rabbit brain glutamate decarboxylase, Brain Research, 344 (1985) 286-295. 4 Brecha, N., Retinal neurotransmitters: histochemical and biochemical studies. In P.C. Emson (Ed.), Chemical Neuroanatomy, Raven Press, New York, 1983, pp. 85-129. 5 Coulombre, A.J., Correlations of structural and biochemical changes in the developing retina of the chick, Am. J. Anat., 96 (1955) 153-188. 6 Coyle, J.T. and Enna, S.J., Neurochemical aspects of the ontogenesis of GABAergic neurons in the rat brain, Brain Research, 111 (1976) 119-133. 7 De Mello, F.G., GABA-mediated control of glutamate decarboxylase (GAD) in cell aggregate culture of chick embryo retina, Dev. Brain Res., 14 (1984) 7-13. 8 De Mello, EG., Bachrach, U. and Nirenberg, M., Ornithine and glutamic acid decarboxylase activities in the developing chick retina, J. Neurochem., 27 (1976) 847-851. 9 Fitzpatrick, D., Lund, J.S., Schmechel, D.E. and Towles, A.C., Distribution of GABAergic neurons and axons terminals in the macaque striate cortex, J. Comp. Neurol., 264 (1987) 73-91. 10 Hamburger, V. and Hamilton, H.L., A series of normal stages in the development of the chick embryo, J. Morphol., 88 (1951) 49-92. 11 Kahn, A.J., An autoradiographic analysis of the time of appearance of neurons in the developing chick neural retina, Dev. Biol., 38 (1974) 30-40. 12 Lam, D.M-K., Lasater, E.M. and Naka, K.-I., r-Aminobutyric acid: a neurotransmitter candidate for cone horizontal cells of the catfish retina, Proc. Natl. Acad. Sci. U.S.A., 75 (1978) 6310-6313. 13 Layer, P.G. and Vollmer, G., Lucifer yellow stains displaced amacrine cells of the chicken retina during embryonic development, Neurosci. Lett., 31 (1982) 99-104. 14 Madtes, R. and Redburn, D.A., GABA as atrophic factor during development, Life Sci., 33 (1983) 979-984. 15 Marc, R.E., Stell, W.K., Bok, D. and Lam, D.M.-K., GABAergic pathways in the goldfish retina, J. Comp. Neurol., 182 (1978) 221-246. 16 Mariani, A.P., Neuronal and synaptic organization of the outer plexiform layer of the pigeon retina, Am. J. Anat., 179 (1987) 25-39. 17 Massey, S.C. and Redburn, D.A., Transmitter circuits in the vertebrate retina, Prog. Neurobiol., 28 (1987) 55-96. 18 Morgan, W.W., GABA: a potential neurotransmitter in the retina. In W.W. Morgan (Ed.), Retinal Transmitters and Modulators: Models for the Brain, CRC Press, Boca Raton, FL, 1985, pp. 63-96. 19 Mosinger, J. and Yazulla, S., Co-localization of GAD-like immunoreactivity and 3H-GABA uptake in amacrine cells of rabbit retina, J. Comp. Neurol., 240 (1985) 396-406. 20 Nakamura, Y., McGuire, B.A. and Sterling, P., Interplexiform cells in cat retina: identification by uptake of 7-(3H)aminobutyric
acid serial reconstruction, Proc. Natl. Acad. Sci. U.S.A., 77 (1980) 658-661. 21 Newton, B.W. and Maley, B.E., A comparison of GABA-and GAD-like immunoreactivity within the area postrema of the rat and cat, J. Comp. Neurol., 255 (1987) 208-216. 22 Oertel, W.H., Schmechel, D.E., Tappaz, M,L. and Kopin, I.J., Production of a specific antiserum to rat brain glutamic acid decarboxylase by injections of an antigen-antibody complex, Neuroscience, 6 (1981) 2689-2700. 23 Pourcho, R.G., Autoradiographic localization of [3H]muscimol in the cat retina, Brain Research, 215 (1981) 187-199. 24 Seiler, N. and AI-Therib, M.J., Putrescine catabolism in mammalian brain, Biochem. J., 144 (1974) 29-35. 25 Seress, L. and Ribak, C.E., The development of GABAergic neurons in the rat hippocampal formation. An immunocytochemical study, Dev. Brain Res., 44 (1988) 197-209. 26 Sobue, K. and Nakajima, T., Changes in concentrations of polyamines and 7-aminobutyric acid and their formation in chick embryo brain during development, J. Neurochem., 1 (1978) 277-279. 27 Somogyi, P., Hodgson, A.J., Smith, A.D., Nunzi, M.G., Gorio, A. and Wu, J.Y., Different populations of GABAergic neurons in the visual cortex and hippocampus of cat contain somatostatin- or cholecystokinin-immunoreactive material, J. Neurosci., 4 (1984) 2590-2603. 28 Spoerri, EE., Neurotrophic effects of GABA in culture of embryonic chick brain and retina, Synapse, 2 (1988) 11-22. 29 Spreafico, R., De Biase, S., Frassoni, C. and Battaglia, G., A comparison of GAD- and GABA-immunoreactive neurons in the first somatosensory area (SI) of the rat cortex, Brain Research, 474 (1988) 192-196. 30 Sternberger, L.A., Immunocytochemistry, 2nd edn., John Wiley & Sons, Inc., New York, 1979. 31 Swann, J.W., Brady, R.J. and Martin, D.L., Postnatal development of GABA-mediated synaptic inhibition in rat hippocampus, Neuroscience, 28 (1989) 551-561. 32 Wenthold, R.J., Zempel, J.H., Parakkal, M.H., Reeks, K.A. and Altschuler, R.A., Immunocytochemical localization of GABA in the cochlear nucleus of the guinea pig, Brain Research, 380 (1986) 7-18. 33 Wenthold, R.J., Huie, D., Altschuler, R.A. and Reeks, K.A., Glycine immunoreactivity localized in the cochlear nucleus and superior olivary complex, Neuroscience, 22 (1987) 897-912. 34 Wong, P.T.H. and McGeer, E.G., Postnatal changes of GABAergic and glutamatergic parameters, Dev. Brain Res., 1 (1981) 519-529. 35 Wu, J.Y., Brandon, C., Su, Y.Y.T. and Lam, D.M.K., Immunocytochemical and autoradiographic localization of GABA system in the vertebrate retina, Mol. Cell. Biochem., 39 (1981) 229-238. 36 Yazulla, S., GABAergic mechanism in the retina, Prog. Ret. Res., 5 (1986) 1-52.
197
Elsevier BRES 16008
Developmental immunoreactivity for GABA and GAD in the avian retina: possible alternative pathway for GABA synthesis Jan N. Hokoq, Ana L.M. Ventura, Patricia F. Gardino and Fernando G. De Mello Instituto de Biofisica Carlos Chagas Filho, Universidade Federal do Rio de Janeiro, Centro de CiOncias da Saade, Rio de Janeiro (Brazil)
(Accepted 15 May 1990) Key words: 7-Aminobutyric acid glutamate decarboxylase immunoreactivity; Development; Putrescine; Chicken retina
Although the distribution of GABAergic neurons in chick retina has been previously described by several investigators, the early appearance of these neurons has not been reported. In the present study immunohistochemical methods were used to localize GABAergic neurons with antisera to both GABA and its synthesizing enzyme, glutamate decarboxylase (GAD), in embryonic chick retina at several stages of development and beyond hatching. GABA-positive neuroblast-like cells were clearly detected in retinas as early as embryonic day 6. In contrast, GAD-containing cells were not observed in retinas until embryonic day 10. These findings indicated that immunocytochemically detectable amounts of GAD were not present in young GABAergic cells. Our data on the developmental appearance of GABA and GAD immunoreactivities are consistent with previous biochemical data for the development of GABA concentration and GAD activity in the chick retina. Together, these data suggest that retina cells from the early stages of development may synthesize GABA from an alternative pathway in which the most likely precursor is putrescine. INTRODUCTION Gamma-aminobutyric acid ( G A B A ) is generally regarded as a major inhibitory neurotransmitter in the central nervous system, including the retina of many vertebrate species, where its role is well established (see review in refs. 4, 17, 18, 36). A possible function for G A B A as a t r o p h i c or regulatory factor during neuronal development in vitro has also been suggested 7'14"28. At least 3 classes of G A B A neurons have been described in the retina of various species. G A B A is present predominantly in amacrine cells of mammalian retinas and amacrine and horizontal cells of non-mammalian retinas 3,12,15,35. More recently, G A B A interplexiform cells in the rat and rabbit retina have also been described19.20. 23. The existence of an 'exuberant' population of neurons stained by a n t i - G A B A serum which is not detected by a n t i - G A D serum has been reported in different areas of the nervous system by authors using different antiG A B A and a n t i - G A D sera 9,21'25'27'29. In the retina of several vertebrate species, GABA-immunoreactive amacrine cells are at least twice as numerous as G A D immunoreactive cells 1 and specifically in birds, G A B A and GAD-immunoreactive processes are distributed in different sublayers in the inner plexiform layer. In some regions of the developing nervous system
there is a marked disparity between the development of the content of endogenous G A B A and its biosynthetic enzyme, glutamic acid decarboxylase 6,z5. Inhibitory postsynaptic potentials that can be blocked by G A B A antagonists are also found earlier than the increase of GAD-activity 31 and, in other systems, postnatal development of G A D activity is taken as an index of G A B A e r g i c synaptogenesis 34. In addition to the classical pathway of G A B A synthesis from glutamate, G A B A can also be synthesized from putrescine in the central nervous system of different species 24. In early stages of chick retina development, a large proportion of G A B A present in the tissue comes from putrescine. This pathway is still present in mature tissue but is much less effective, accounting for less than 1% of the total G A B A 8. These observations raised the possibility that some retinal neurons might contain G A B A derived from the alternative pathway rather than G A B A produced by the classical, G A D - m e d i a t e d decarboxylation of glutamate. Putrescine is the precursor for G A B A in chick embryo brain as well 26. In the present study using an immunocytochemical approach we compared the appearance of G A B A and its synthetic enzyme G A D during development of the embryonic chick retina. Our data show that GAB,Aimmunoreactive cells are present early in development, at least 4 days prior to the appearance of G A D -
Correspondence: J.N. Hoko~, Departamento de Neurobiologia, Instituto de Biofisica Carlos Chagas F°, Universidade Federal do Rio de Janeiro, Centro de Cifncias da Safide, bloco G, 21944 Rio de Janeiro, Brazil.
0006-8993/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
198 immunoreactivity. The first GAD-containing neurons were detected during synapse formation in the inner plexiform layer. Our data strongly suggest that GABA in retinas from early embryonic stages comes from a source that utilizes putrescine as precursor.
MATERIALS AND METHODS Fertilized white Leghorn eggs were obtained from a local hatchery. Eyes from chick embryos ranging from 6 days of incubation 1° to various stages beyond hatching, were used. The eyes were enucleated, bisected with a razor, and the posterior pole immersed in fixative for 1 h. After several rinses in buffer, a piece of retina from the central region was dissected out from the sclera, and frozen sections 15-25/~m thick, were cut in the radial plane. The sections were mounted in gelatinized slides and incubated overnight with the primary antibody or kept frozen at -20 °C up to 2 months without loss of antigenicity. For GABA-immunoreactivity, retinas were fixed in 4% formaldehyde and 0.1% glutaraldehyde in 0.16 M phophate buffer. Retinal sections (15-25/~m) were incubated overnight at 4 °C with a rabbit antisera directed against a GABA/bovine serum albumin (BSA) conjugate 33 at a final dilution of 1:200. The GABA antisera were purified against glutamine, glutamate, aspartate, glycine, taurine, alanine and fl-alanine32. The primary reaction was visualized by the peroxidase-antiperoxidase (PAP) method 3°. Swine anti-rabbit Fab fragments (1:40) and rabbit PAP (1:80) (Dako) were used, followed by the diaminobenzidine reaction, used as a chromogen. Control sections were incubated either with normal rabbit serum or with the antibody preadsorbed with the amino acid conjugate GABA/ BSA 33. Immunocytochemical studies were then performed as described for the GABA-antisera-incubated sections. No immunoreactivity was detected in the controls. For GAD-immunoreactivity, retinas were fixed in 4% formaldehyde in 0.16 M phosphate buffer, pH 7.2. The sections were incubated overnight at 4 °C with GAD antiserum produced in sheep22 at a final dilution of 1:1000. Rabbit anti-sheep IgG, 1:150 (Sigma), and sheep PAP, 1:600 (Jackson), followed by diaminobenzidine (Sigma) as a chromogen, were used. Controls consisted of sections incubated either with normal sheep serum at a dilution of 1:1000 or with the preimmune serum 22 instead of the primary antisera and processed the same way, in parallel with the GAD antisera-exposed sections. No immunoreactivity was detected in the controls. Triton-X (0.25%, v/v) was used in all solutions to enhance permeability of the reagents. More concentrated primary antisera (1:400) was used for embryo retinas. More diluted primary antisera (1:2000) was better to visualize cell processes but not to label cell bodies. In 4 post-hatched animals 0.30 Id of colchicine 0.5% was injected on one of the eyes, for 24 h, in an attempt to label more cell bodies.
The sections were examined, studied and photographed under a light microscope.
RESULTS
In chick embryos, [3H]thymidine incorporation points to the posterior pole of the retina as the first region to complete mitotic activity 11. Thus, all the radial sections used for GABA- and GAD-immunoreactive studies, at different stages, were always taken from this specific region of the retina. GABA immunoreactivity was first detected in retinas at embryonic day 6 (E6). GABA was clearly detectable in a few cells neighboring the central retina, along a row 1/3 from the pigmented epithelium (Fig. 1A). On radial sections of 6-day-old embryo (E6) retinas, Coulombre 5 described the cells as elongated and forming an epithelium with nuclei arranged in several layers, except for the ganglion cell bodies, some of which were round and sent out axons. Thus, the GABA-positive cells detected at E6 could be either migrating neuroblasts or presumptive horizontal cell bodies. At higher magnification, some processes emerging from these cells could also be observed (Fig. 2D), possibly indicating horizontal cells undergoing a differentiation process. Unexpectedly, no GAD-immunoreactive cells could be detected at E6 (Fig.
1B). The next examined retina was at embryonic day 8 (E8). At this stage the retina is characterized by the appearance of inner and outer plexiform layers in the central region 5. The morphological aspect of cell bodies in the inner nuclear layer is still elongated, when compared to ganglion and horizontal cell somata. The immunohistochemistry for GABA showed labeled cell bodies in the neuroblastic layer adjacent to the inner plexiform layer (Fig. 1C). A row of cell bodies could also be distinguished along the presumptive location of horizontal cells, close to the prospective outer plexiform site (Fig. 2C). Near the central region of the retina, some GABA-positive cell somata were localized in the middle
Fig. 1. Light micrographs of 20/xm thick cryostat radial section from embryonic chick retinas at several stages of development and beyond hatching. A: retina section of the central region of an embryonic day 6 retina. GABA immunoreactivity is detected in neuroblast-like cells in a row close to the pigment epithelium. B: no immunoreactivity for GAD is detected in embryonic day 6 retina. C: retina section of the central region at embryonic day 8. GABA-immunoreactive cell bodies are localized in the inner nuclear layer, including horizontal and amacrine cells, and in the ganglion cell layer. Cells in migrating process can be seen in the middle of the inner plexiform layer. D: no immunoreactivity to GAD can be detected in day-8 embryonic retina. E: retina section of day-12 embryo retina incubated with GABA antiserum showing a profuse number of horizontal cell bodies, with process extended in only one stratum of the outer plexiform layer. Amacrine cell bodies and perikarya in the ganglion cell layer are also present. F: most of the GAD-containing cell bodies of day-12 embryonic retina are characterized as amacrine cells with processes extended in 3 well defined strata in the inner plexiform layer. Very weak label can be seen in horizontal cell bodies. G: GABA-immunoreactive labeling in mature chicken retina showing the immunoreactivity localized in horizontal cell bodies, several populations of amacrine cells and cell bodies in the ganglion cell layer. Labeled processes extended in the outer and inner strata of the outer plexiform layer and in the inner plexiform layer with no distinguishable layered organization. H: GAD-immunoreactivity in mature retina section is localized in a few cell bodies close to the vitreal side of the inner nuclear layer with processes extending in 3 distinct labeled bands in the inner plexiform layer. Weakly GAD-positive cell bodies are also labeled. Bar = 20/xm.
199 of the inner plexiform layer, resembling migrating cells. A few cells had processes extending toward the inner plexiform layer, these cell bodies can be displaced amacrine ceils in late migratory process, as suggested by
Layer and Vollmer 13. At the same stage (E8), GAD-positive neurons could still not be detected. Sections from 12-day-old embryo (El2) retinas already
....
~.7 L- . . . . .
-
~
:
~
....
-.--~
-
-~
- ~-L:~-i.~ZIZ
200 displayed the characteristic layered structure, except for the outer segments of the photoreceptor, which are not completely differentiated yet. At this stage, horizontal cell bodies at the outer portion of the inner nuclear layer had their processes extending along the outer plexiform layer. Thick processes from amacrine cells appeared and extended into the inner plexiform layer. By this time synaptogenesis begins in this layer5. GABA-immunoreactivity is detectable in El2 in the same types of cells as in the adult retina (Fig. 1E). Strongly reactive cell bodies of horizontal cells, with processes lying along the outer plexiform layer, seemed to be more numerous than in mature retina (Fig. 2B). Although weakly labeled, amacrine and ganglion cell somata were also present, resembling the adult pattern. A 3-strata pattern could be detected in the inner plexiform layer, which disappeared in older retinas (Fig. IE). At this stage (El2), GAD-containing amacrine cells were present along the inner portion of the inner nuclear layer. GAD-positive cytoplasm was concentrated at their vitreal side and processes of these cells extended radially along 3 laminae in the inner plexiform layer (Fig. 1F). Thus, GAD immunoreactivity in the chick retina is present at a later embryonic stage than GABA immunoreactivity. Finally, in posthatched chicken retinas, GABA immunoreactivity was localized in horizontal cells, several types of amacrine cells in the inner nuclear layer, and some perikarya in the ganglion cell layer (Fig. 1G), consistent with previous findings 1.
GABA-immunoreactive horizontal cells, ranging from 4 to 5 ,um in diameter, had their cell bodies lying along the outermost layer of the inner nuclear layer, with processes extending towards the sclera and spreading horizontally in the inner and outer strata of the outer plexiform layer (Fig. 2A). Although this label does not allow a precise definition of the dendrites as in silverimpregnated tissue, GABA-immunoreactive horizontal cells described here are likely to be of the H1 and H4 types, described for the pigeon retina 16. GABA-containing amacrine cell bodies, ranging from 4 to 9/zm in diameter, were along the inner margin and the vitreal half of the inner nuclear layer. Their processes in the inner plexiform layer extended from the outermost to the innermost layer, and were so numerous and intensely immunoreactive that no layered organization could be distinguished (Fig. 1G). In the ganglion cell layer, G A B A immunoreactive cell bodies were also of 4-9 ~zm size in diameter, corresponding to displaced amacrine cell bodies and possibly ganglion cell somata. No GABA-immunoreactive optic fibers were found in the nerve fiber layer. GAD immunoreactivity was observed in cell bodies in the inner nuclear layer and in the neuropil of the inner plexiform layer (Fig. 1H). Cell body sizes ranged from 5 to 10/xm in diameter, consisting of a thin rim of labeled cytoplasm, thickest toward the inner plexiform layer, and the nuclei were not immunoreactive. The inner plexiform layer consisted of 3 distinct labeled bands, corresponding to the outermost (13/xm thick), intermediate (9/~m thick) and innermost (20/xm) laminae, respectively. The intermediate band was the most intensely immunoreactive, followed by the outermost and the innermost. At any age up to El2, GAD-immunoreactivity was less intense than GABA-immunoreactivity in amacrine cell bodies. Although clearly positive for GABA, very weak labels were found for GAD in horizontal cell bodies (Fig. 1 E and F, and G and H). As expected from previous studies 1'3 ganglion cells were not GAD-immunoreactive in mature retinas. Occasionally, GAD-containing cell bodies were found in the ganglion cell layer, which by their soma diameter might correspond to displaced amacrine cells. No GAD- or GABA-immunoreactive interplexiform cells were found in chicken retina. DISCUSSION
Fig. 2. High magnification of GABA-immunoreactive horizontal cell population at mature retina (A); embryonic day 12 (B); embryonic day 8 (C) and embryonic day 6 (D). Bar = 50/xm.
The present study was undertaken to compare GABA and GAD immunoreactivity in embryonic chick retina. During development, GABA immunoreactivity was detected in neurons in the central retina as early as E6, in cells that display morphological characteristics of
201 migrating neuroblasts and/or committed neuroblasts at the end of their mitotic phase. In contrast, detectable levels of G A D were not found before El0-12, i.e., after a delay of 4 days. These data are unique since GAD, the synthesizing enzyme for G A B A , does not appear in neurons of younger preparations where GABA-positive neurons are clearly present. A few possibilities could explain this phenomenon: one is that the lack of immunoreactivity for G A D in cells from early-stage retinas could be due to the fact that the anti-GAD used in this work would not recognize a G A D isoform present in neuroblast-like cells of the mature tissue. This seems to be unlikely since our analysis was performed with a polyclonal antibody to GAD. This would substantially reduce the chances of a differential detection of GAD isoforms in this tissue. Another technical problem that might account for our data would be that if only trace amounts of G A D were present in the perikaryal cytoplasm of young neurons, our antibody would not be sensitive enough to detect the enzyme. This seems also to be questionable since the same antibody detects G A D immunoreactivity in amacrine cells of differentiated tissue, which display approximately the same intensity of reaction for G A B A as neuroblast like cells. Thus, since the immunoreactivity for G A B A seems to be similar in both cell populations (neuroblast and amacrine cells), one would expect that G A D (the limiting enzyme for G A B A synthesis) would be expressed in equivalent proportions in the two cell types. Finally, the third possible explanation for our observations would be that the G A B A present in neuroblastlike cells in the retina would come from an alternative route of synthesis in which putrescine would be the major precursor. This possibility is reinforced by the fact that G A B A can be synthesized from putrescine in several areas of the CNS 24, including the avian retina s. As mentioned above, during development the first GAD-containing neurons were detected between El0 and El2, thus prior to synapse formation in the inner plexiform layer 5. This observation is in agreement with a previous report 8. G A B A immunoreactivity, however, was detected in neurons in the central retina as early as E6. Previous reports 2,s have shown that ornithine decarboxylase activity, and consequently putrescine and other polyamines, are very high in undifferentiated cells with high mitotic activity or in cells at the end of their divisional cycle. After cells have withdrawn from the cell REFERENCES 1 Agardh, E., Bruun, A., Ehinger, B., Ekstrom, P., Van Veen, T. and Wu, J.Y., Gamma-aminobutyric acid- and glutamic acid decarboxylase immunoreactive neurons in the retina of different
cycle, this enzyme declines substantially as do putrescine and other polyamines. In the chick retina, the ornithine decarboxylase activity and putrescine content are extremely high in 6-day-old embryos and decline as the tissue differentiates. Concomitant with the decrease in putrescine content, retinal G A B A increases several-fold, preceding the increase in G A D activity by a few days. A substantial amount of G A B A in the early stages of development comes from putrescine s. The observation that G A B A is present in neuroblast-like cells, devoid of GAD immunoreactivity, strongly suggests that G A B A in these cells might be synthesized from putrescine. One major question that can be put forward, based on previous and present publication, is what would be the role of G A B A in retinal cells in embryos as early as the 6th day of incubation. Since several recent publications suggest a trophic role for G A B A in the CNS, an obvious idea would be that G A B A in the embryonic tissue would serve a t r o p h i c or controlling influence on developing retinal neurons. In cultured chick embryo retina cells the presence of G A B A in the incubating medium fully inhibits the expression of G A D activity during development, an effect that seems to be mediated, at least in part, by G A B A receptors 7. Since GAD-positive neurons in the retina are first detected at El0-12, when synapses start to form in the inner plexiform layer, it is possible that G A B A from an alternative source in the embryonic tissue could play a role in timing the precise moment during synaptogenesis in which committed GABAergic neurons would start to express GAD. Experiments from our laboratory, to be published elsewhere, show that the effect of G A B A in preventing the expression of G A D activity in cultured retina cells is reflected by a lack of immunoreactivity for G A D , suggesting that G A B A may be controlling the synthesis of the enzyme. If so, the neurotransmitter role for G A B A may be secondary to its metabolic or trophic function in the CNS of young developing animals.
Acknowledgements. We are grateful to Dr. Wenthold for his generous gift of antiserum to gamma-aminobutyric acid (#2) and GABA-BSA conjugate. Antiserum to GAD, as well as preimmune serum was provided through the laboratory of Clinical Science, NIMH, where it was developed under the supervision of Dr. Irwin J. Kopin with Drs. W. Oertel, D.E. Schmechel and M. Tappaz. We thank Dr. A.P. Mariani for critically reading this manuscript and Ms. Mair Medeiros for her technical assistance. This work was supported by grants from CNPq, FAPERJ and FINEP to J.N.H. and grants from FINEP, CNPq, FAPERJ and PEG to F.G.D.M.
vertebrates, J. Comp. Neurol. 258 (1987) 622-630. 2 Anderson, T.R. and Schamberg, S.M., Ornithine decarboxylase activity in developing rat brain, J. Neurochem., 19 (1972) 1471-1481. 3 Brandon, C., Retinal GABA neurons: localization in verte
202 brates species using an antiserum to rabbit brain glutamate decarboxylase, Brain Research, 344 (1985) 286-295. 4 Brecha, N., Retinal neurotransmitters: histochemical and biochemical studies. In P.C. Emson (Ed.), Chemical Neuroanatomy, Raven Press, New York, 1983, pp. 85-129. 5 Coulombre, A.J., Correlations of structural and biochemical changes in the developing retina of the chick, Am. J. Anat., 96 (1955) 153-188. 6 Coyle, J.T. and Enna, S.J., Neurochemical aspects of the ontogenesis of GABAergic neurons in the rat brain, Brain Research, 111 (1976) 119-133. 7 De Mello, F.G., GABA-mediated control of glutamate decarboxylase (GAD) in cell aggregate culture of chick embryo retina, Dev. Brain Res., 14 (1984) 7-13. 8 De Mello, EG., Bachrach, U. and Nirenberg, M., Ornithine and glutamic acid decarboxylase activities in the developing chick retina, J. Neurochem., 27 (1976) 847-851. 9 Fitzpatrick, D., Lund, J.S., Schmechel, D.E. and Towles, A.C., Distribution of GABAergic neurons and axons terminals in the macaque striate cortex, J. Comp. Neurol., 264 (1987) 73-91. 10 Hamburger, V. and Hamilton, H.L., A series of normal stages in the development of the chick embryo, J. Morphol., 88 (1951) 49-92. 11 Kahn, A.J., An autoradiographic analysis of the time of appearance of neurons in the developing chick neural retina, Dev. Biol., 38 (1974) 30-40. 12 Lam, D.M-K., Lasater, E.M. and Naka, K.-I., r-Aminobutyric acid: a neurotransmitter candidate for cone horizontal cells of the catfish retina, Proc. Natl. Acad. Sci. U.S.A., 75 (1978) 6310-6313. 13 Layer, P.G. and Vollmer, G., Lucifer yellow stains displaced amacrine cells of the chicken retina during embryonic development, Neurosci. Lett., 31 (1982) 99-104. 14 Madtes, R. and Redburn, D.A., GABA as atrophic factor during development, Life Sci., 33 (1983) 979-984. 15 Marc, R.E., Stell, W.K., Bok, D. and Lam, D.M.-K., GABAergic pathways in the goldfish retina, J. Comp. Neurol., 182 (1978) 221-246. 16 Mariani, A.P., Neuronal and synaptic organization of the outer plexiform layer of the pigeon retina, Am. J. Anat., 179 (1987) 25-39. 17 Massey, S.C. and Redburn, D.A., Transmitter circuits in the vertebrate retina, Prog. Neurobiol., 28 (1987) 55-96. 18 Morgan, W.W., GABA: a potential neurotransmitter in the retina. In W.W. Morgan (Ed.), Retinal Transmitters and Modulators: Models for the Brain, CRC Press, Boca Raton, FL, 1985, pp. 63-96. 19 Mosinger, J. and Yazulla, S., Co-localization of GAD-like immunoreactivity and 3H-GABA uptake in amacrine cells of rabbit retina, J. Comp. Neurol., 240 (1985) 396-406. 20 Nakamura, Y., McGuire, B.A. and Sterling, P., Interplexiform cells in cat retina: identification by uptake of 7-(3H)aminobutyric
acid serial reconstruction, Proc. Natl. Acad. Sci. U.S.A., 77 (1980) 658-661. 21 Newton, B.W. and Maley, B.E., A comparison of GABA-and GAD-like immunoreactivity within the area postrema of the rat and cat, J. Comp. Neurol., 255 (1987) 208-216. 22 Oertel, W.H., Schmechel, D.E., Tappaz, M,L. and Kopin, I.J., Production of a specific antiserum to rat brain glutamic acid decarboxylase by injections of an antigen-antibody complex, Neuroscience, 6 (1981) 2689-2700. 23 Pourcho, R.G., Autoradiographic localization of [3H]muscimol in the cat retina, Brain Research, 215 (1981) 187-199. 24 Seiler, N. and AI-Therib, M.J., Putrescine catabolism in mammalian brain, Biochem. J., 144 (1974) 29-35. 25 Seress, L. and Ribak, C.E., The development of GABAergic neurons in the rat hippocampal formation. An immunocytochemical study, Dev. Brain Res., 44 (1988) 197-209. 26 Sobue, K. and Nakajima, T., Changes in concentrations of polyamines and 7-aminobutyric acid and their formation in chick embryo brain during development, J. Neurochem., 1 (1978) 277-279. 27 Somogyi, P., Hodgson, A.J., Smith, A.D., Nunzi, M.G., Gorio, A. and Wu, J.Y., Different populations of GABAergic neurons in the visual cortex and hippocampus of cat contain somatostatin- or cholecystokinin-immunoreactive material, J. Neurosci., 4 (1984) 2590-2603. 28 Spoerri, EE., Neurotrophic effects of GABA in culture of embryonic chick brain and retina, Synapse, 2 (1988) 11-22. 29 Spreafico, R., De Biase, S., Frassoni, C. and Battaglia, G., A comparison of GAD- and GABA-immunoreactive neurons in the first somatosensory area (SI) of the rat cortex, Brain Research, 474 (1988) 192-196. 30 Sternberger, L.A., Immunocytochemistry, 2nd edn., John Wiley & Sons, Inc., New York, 1979. 31 Swann, J.W., Brady, R.J. and Martin, D.L., Postnatal development of GABA-mediated synaptic inhibition in rat hippocampus, Neuroscience, 28 (1989) 551-561. 32 Wenthold, R.J., Zempel, J.H., Parakkal, M.H., Reeks, K.A. and Altschuler, R.A., Immunocytochemical localization of GABA in the cochlear nucleus of the guinea pig, Brain Research, 380 (1986) 7-18. 33 Wenthold, R.J., Huie, D., Altschuler, R.A. and Reeks, K.A., Glycine immunoreactivity localized in the cochlear nucleus and superior olivary complex, Neuroscience, 22 (1987) 897-912. 34 Wong, P.T.H. and McGeer, E.G., Postnatal changes of GABAergic and glutamatergic parameters, Dev. Brain Res., 1 (1981) 519-529. 35 Wu, J.Y., Brandon, C., Su, Y.Y.T. and Lam, D.M.K., Immunocytochemical and autoradiographic localization of GABA system in the vertebrate retina, Mol. Cell. Biochem., 39 (1981) 229-238. 36 Yazulla, S., GABAergic mechanism in the retina, Prog. Ret. Res., 5 (1986) 1-52.
