Development of GABAergic synaptic connections in vivo and in cultures from the rat superior colliculus

95

Developmental Brain Research, 52 (1990) 95-111 Elsevier BRESD 51027

Development of GABAergic synaptic connections in vivo and in cultures from the rat superior colliculus S.S. Warton*, M. Perouansky and R. Grantyn Department of Neurophysiology, Max Planck Institutefor Psychiatry, Martinsried (E R. G.) (Accepted 12 September 1989)

Key words: Rat superior colliculus; Cell culture; 7-Aminobutyric acid uptake; Glutamic acid decarboxylase immunostaining; Ultrastructure of GABAergic synapse; 7-Aminobutyric acid-A receptor; Chloride current; Synaptogenesis

Synaptic activity in the superficial (i.e. visual) layer of the superior colliculus was investigated with intracellular microelectrodes using a preparation of the isolated superfused tectum from neonatal rat. It was found that by postnatal day 9 (i.e. before eye opening) the majority of neurons in the superficial gray layer (SGS, stratum griseum superficiale) were already capable of generating Cl--dependent inhibitory postsynaptic potentials (IPSPs) in response to intracollicular stimulation. Properties and development of GABAergic synaptic connections were further characterized in a dissociated cell culture from the SGS. The cultures were prepared from E21 rat embryos and studied between 1 and 38 days in vitro (DIV). ~,-[3H]aminobutyric acid ([3H]GABA) uptake served to identify GABAergic neurons and to estimate their relative density. Axon terminals were labeled by indirect immunostaining for glutamic acid decarboxylase (GAD) and examined with light (LM) and electron microscopy (EM). Responsiveness to exogenous and endogenous GABA was investigated by recording ionic currents with patch clamp techniques. [3H]GABA uptake-positive neurons constituted about 40% of the whole cellular population dissociated from the SGS of E21 rats. After 2 weeks in culture, [aH]GABA uptake was observed in 45-60% of the cells with neuronal features. The relative number of GAD-immunoreactive neuronal perikarya ranged from 28 to 39%, after 2 weeks in vitro. Responsiveness to exogenous GABA was found in all freshly plated neurons. Release of GABA could be demonstrated after 2 DIV by recording spontaneous bicucuiline-sensitive Cl- currents. These currents had the characteristics of GABA A receptor-mediated synaptic currents. However, even as late as DIV 6, very few vesicle-containing axonal terminals apposing postsynaptic specializations were revealed with EM. GAD-labeled puncta became clearly visible only after DIV 10-12. Between DIV 14 and 21, the intensity of immunostaining and the density of GAD-labeled synaptic contacts increased, reaching a maximum around DIV 28. GAD-positive puncta covered both neurons and non-neuronal cells. At the level of EM, GAD-positive terminals were shown to establish synaptic contacts with neuronal somata and processes, forming in the majority of cases (22 out of 32 stained terminals) symmetrical contacts. It is concluded that in the SGS of the rat superior colliculus GABAergic neurons and GABA A receptors are present before birth. In dissociated cell cultures ionic currents can be generated in reponse to endogenous GABA before axonal terminals of GABAergic neurons fully mature. Finally, our experiments show that visual activity is not a prerequisite for the formation of GABAergic synapses between neurons of the SGS.

INTRODUCTION G l u t a m i c acid d e c a r b o x y l a s e ( G A D ) is present in nerve endings 15 where it catalyzes the synthesis of the inhibitory neurotransmitter ),-aminobutyric acid ( G A B A ) . G A B A e r g i c neurons can be identified by revealing the presence of either G A D 4°'~s'49'58'76 or G A B A 25'65 and by demonstrating a high-affinity [ 3 H ] G A B A uptake 13'34'71. The superficial gray layer of the mammalian superior colliculus (SC) contains a substantial amount of G A B A and enzymes related to its synthesis or degradation 39"5°'sl, due to a high number of G A B A e r g i c perikarya and terminals 44' 45,52. Mize44 showed that about 45% of the neurons in the stratum griseum superficiale (SGS) of the feline SC are G A B A - i m m u n o r e a c t i v e . Small to medium size granular

neurons constitute the majority of stained cells. In addition, a class of horizontal neurons in the upper SGS of the cat SC is likely to be G A B A e r g i c 43'44. This particular neuronal type has been described in a variety of species 33'56'70'72 and has been shown to receive retinal afferents which may form synaptic triades with presynaptic dendrites 36,42'69'7°. Since enucleation and visual cortex ablation had little effect on G A D and G A B A levels in the SC 16, and none of the extrinsic afferents to the SGS have so far been found to stain for G A D (cf. refs. 21,27), one can assume that all GAD-positive terminals in the SGS are issued by intrinsic G A B A e r g i c neurons. Houser et al. 13 used quantitative electron microscopical techniques to demonstrate that the vast majority of G A D positive terminals establish symmetrical contacts with post-

* Present address: Department of Anatomy and Human Biology, University of Western Australia, Nedlands, W.A. 6009, Australia. Correspondence: R. Grantyn, Department of Neurophysiology, Max Planck Institute for Psychiatry, Am Klopferspitz 18A, 8033 Martinsried, ER.G. 0165-3806/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)

96 synaptic structures. O n l y 1.6% of all G A D - p o s i t i v e terminals form asymmetrical contacts. This figure was, h o w e v e r , significantly increased by partial deafferentation after eye r e m o v a l in j u v e n i l e rats, suggesting that asymmetrical sites m a y be o c c u p i e d or f o r m e d by G A B A e r g i c terminals in the post-lesion period. It is not k n o w n w h e t h e r this structural reorganization of synaptic inputs, as o b s e r v e d in young animals, influences G A B A e r g i c synaptic transmission. I n t r a c e l l u l a r r e c o r d i n g f r o m S G S - n e u r o n s ex v i v o w e r e u s e d to i n v e s t i g a t e w h e t h e r

GABAergic

inhibition is

f u n c t i o n a l b e f o r e eye o p e n i n g . A culture of S G S - d e r i v e d cells, p r e p a r e d by dissecting the superficial gray layer in p e r i n a t a l rats 54, p r o v e d to be a c o n v e n i e n t e x p e r i m e n t a l m o d e l to e x a m i n e t h e p r o p e r t i e s of G A B A e r g i c synapses in (1) t h e t o t a l a b s e n c e of extrinsic afferents to S G S n e u r o n s , a n d (2) f o l l o w i n g c o m p l e t e de n o v o f o r m a t i o n of G A B A e r g i c synaptic contacts. [ 3 H ] G A B A a u t o r a d i o g r a p h y was u s e d to e s t i m a t e the f r a c t i o n of G A B A e r g i c n e u r o n s in S G S - d e r i v e d cultures. By a p p l y i n g an antib o d y against G A D ,

G A B A e r g i c t e r m i n a l s w e r e charac-

t e r i z e d at b o t h t h e light and e l e c t r o n m i c r o s c o p i c a l levels. P a t c h c l a m p r e c o r d i n g of synaptic activity in c u l t u r e d SGS-derived neurons

was p e r f o r m e d

presence

nature

and q u a n t a l

to test for the

of G A B A e r g i c

synaptic

activity in t h e s e cultures. MATERIALS AND METHODS

lntracellular recording from in vivo developing neurons in the isolated superfused SC Neonatal rats, 9-28 days of age, were anesthetized with ketamine (50 mg/kg, i.m.) or pentobarbital (80 mg/kg, i.p.). The animals were then transcardially perfused with cooled oxygenated standard Krebs-Ringer solution before dissecting the brain to remove the dorsal mesencephalon and pretectum. This tissue was transferred to a recording chamber and continuously superfused with standard salt solution containing (in mM): NaCI 124, KCI 3.5, KHzPO 4 1.2, CaCI 2 2.4, MgSO 4 1.3, NaHCO 3 26, glucose 10. The solution was equilibrated with 95% 02 and 5% CO 2 to achieve a pH of 7.4. All recordings were taken at temperatures between 26 and 32 °C. The isolated SGS remained viable for more than 8 h. The tissue preservation was usually excellent if the pieces were cut off at a depth of no more than 500 /~m below the surface of SC. Microelectrodes filled with a solution of 10% horseradish peroxidase (HRP) (Sigma, Type VI) in 0.25 M KCI and 0.1 M Tris (pH 7.6) were used to record the antidromic, orthodromic, or depolarization-induced activity of various neuron types in the SGS (Fig. 1). An array of stimulating microelectrodes was placed in the optical layer of the rostral SC for the antidromic activation of rostrally projecting neurons and for the orthodromic activation of synaptic activity. Some of the neurons were injected with HRP and reconstructed, as described previously~9'2°. Cell cultures Cultures were prepared at embryonic day 21 (E21). Fig. 2A shows the SC of an E21 rat, and the level at which the SGS was dissected to prepare the cultures. A typical cross-section indicates the portion of SC that remained after collecting the SGS (Fig. 2B). The cellular yield was about 200,000 cells per brain. Cells were plated on collagen-coated eoverslips at an initial density of 250,000 cells/cm2. Cultures were maintained for 4-5 weeks in minimal Eagle's medium

(MEM) supplemented with 5% fetal call Sel'Uln ~lnci5 ~ [|t)15¢ sofun] in the presence of 30 ug/ml insulin. After 4 daw m vitro (DIV) :~n antimitotic agent ( 10 ~ M cytosine arabinofurano~ide lriphosphate, ARA-CTP) was added to the cultures for 24 i~. Application ~1 ARA-CTP was repeated each week to prevent the appearance of oligodendrocytes and the formation of myelin. After 3 weeks m vitro SGS-derived cultures formed a two-dimensional neuropil-like network (Fig. 2C) and displayed GABAergic synaptic activity. Some of the electrophysiological features of these collieular cultures have been described previously2"~:4 In order to further characterize the cell populati~m in mature cultures, labeling experiments were performed between DIV 14 and 21. Two neuronal markers were utilized, namely tetanus toxin 55 and ~v-[lzSI]conotoxin, a radioactive blocker of neuronal calcium channels 9. Fibroblasts, in addition to neurons, were labeled with the antibody MRC OX-7 against the cell surface glycoprotein Thy I.t ~s. The presence of the glial fibrillar acidic protein (GFAP) was used to identify cells as astrocytes 5s. The results of these experiments (to be published elsewhere) aided the assignment of particular cell geometries to specific cellular phenotypes in these mixed cultures. Electrophysiological measurements were performed exclusively on phase bright cells with neuronal geometry, as the ones shown in Fig. 2C,D. These were the cells staining with the neuronal markers tetanus toxin and w[~25I}conotoxin. Flat cells, such as those labeled with anti-Thy 1.1 (fibroblasts) or GFAP (astrocytes) were rejected for recording.

[3H]GABA uptake autoradiography Cultures were incubated for 20-30 min in 0.1/~M [3H]GABA (25 /~Ci/mM; New England Nuclear), rinsed in ice-cold HEPESbuffered saline and fixed in 2.5% glutaraldehyde. The culturebearing coverslips were coated with Kodak NTB2 emulsion and exposed for 8 or 16 days. After developing the emulsion, a few samples were counterstained with 0.1% Cresyl violet, dehydrated by graded alcohols, cleared with xylene and mounted on glass slides. These preparations were used to visualize labeled neuronal processes in well isolated cells. Since, in older cultures, [3H]GABA uptake-positive axons could interfere with identification of underlying non-labeled dendrites it was also necessary to examine unstained preparations with phase optics (Figs. 4 and 6B). No attempt was made to quantify labeling density. However, for cell counting care was taken to use material labeled under identical conditions (i.e. same GABA concentrations, incubation and exposure times). Five different E21 cultures, between DIV 0 and 27, were processed for uptake of [3H]GABA. GA D-immunohistochemistrv GABAergic axonal terminals were identified by G A D immunostaining using a sheep anti-GAD serum 48'49. For this purpose the cultures were fixed for 30 min with 4% formaldehyde in 0.1 M Tris-buffered saline (TBS). After 3 washes with TBS, cultures were preincubated in 10% normal rabbit serum for 30 rain. After another wash, the sheep anti-GAD serum was applied at a dilution of t:3000 for 16 h at 4 °C. In every experiment control sister cultures were processed identically, but were not exposed to the anti-GAD serum. Following a rinse with TBS. the cultures were exposed for 60 min to the second antibody {rabbit anti-sheep serum, dilution of 1:50), again thoroughly rinsed with TBS. incubated in the peroxidaseantiperoxidase complex (1:100) from Sigma, and then treated with a solution containing 0.05% diaminobenzidine hydrochtoride and 0.01% hydrogen peroxide for 15 min to visualize the immunoperoxidase reaction product. For light microscopic (LM) investigation, samples were quickly dehydrated, cleared and mounted. After this treatment GAD-immunoreactive (GAD-IR) structures appeared dark brown and could be distinguished from pate yellow GADnegative structures. Although some of the neuronal somata stood out quite clearly (Figs. 5 and 6A, solid arrows) it was felt that GAD-immunoreactivity is less reliable as a basis for counts of GABAergic neurons than the uptake of [3H]GABA. In contrast, GAD-IR axon terminals were strongly labeled (Fig. 5).

97

Cell counts

The records were taken at postnatal day 21. A t this stage

In order to determine the proportion of GABAergic neurons present in culture, [3H]GABA uptake-positive neurons and GADIR perikarya were counted using a magnification of x32. A total of 3 dishes (about 25 randomly chosen fields) were examined at each age. When evaluating the [3H]GABA uptake, phase contrast optics was used to distinguish labeled flat (presumably non-neuronal)cells in cultures at DIV 1. Partially labeled neurons were also rejected, as such a labeling pattern could be due to contacts from GABAergic axons (see Fig. 6B,C). The counts represented in Fig. 11 were obtained from a set of cultures at various times in vitro but processed together for [3H]GABA uptake.

of d e v e l o p m e n t all n e u r o n s , including those with geom-

Electron microscopy Specimens selected for further electron microscopic (EM) analysis were treated with 0.1% osmium tetroxide for 30 min, dehydrated and embedded in a mixture of epon-araidite. Semithin resin sections were analyzed at LM level, reembedded and cut to ultrathin sections. The latter were stained with uranyl acetate and lead citrate and examined with a Zeiss electron microscope. Ten different cultures, between DIV 1 and 38, were prepared for GAD immunohistochemistry, 4 of them for subsequent EM (DIV 2, 10, 15, 20 and 38). As GAD immunohistochemistry reduces the quality of tissue preservation, untreated cultures fixed at DIV 3, 6, 7, 12, 15, 17 and 24 were directly processed for EM, to demonstrate the fine structure of synaptic connections during the course of synaptogenesis. The fraction of GAD-positive synapses was estimated at DIV 20 and 38. For this purpose, GAD-positive and GAD-negative synapses were counted in electron micrographs at a magnification of 9.600 or 12.600.

etries corresponding to the putative G A B A e r g i c phenotype (Fig. 3F), generated large O - - d e p e n d e n t IPSPs. F u r t h e r experiments were therefore performed on younger rats, down to an age of P9. In 16 out of 21 n e u r o n s tested before eye o p e n i n g (days P9-12) CI-d e p e n d e n t IPSP, as those shown in Fig. 1B, could be elicited by intracollicular stimulation, indicating that a G A B A e r g i c network is already u n d e r formation before the visual pathway is set into function. After D I V 14 IPSPs were found in all (n = 25) intrasomatically impaled

\

A

50pm

Patch clamp recording Responses to exogenous GABA, as well as spontaneous CIcurrents (Fig. 10) were recorded with patch electrodes in the whole cell configuration, as described previously22'54. Briefly, electrodes were filled with a solution containing (in mM): N-methylglucamine 120, TEA-CI 20, EGTA 8, CaCI2 0.2, MgCI2 0.2, glucose 10, ATP 2, cAMP 0.25, HEPES 10. The pH was adjusted to 7.3 with HCI. Superfusion solutions were applied through a 6-barrel pipette placed in the vicinity of the investigated cells. The standard superfusion solution contained (in mM): NaCI 120, KCI 3, CaCI2 1.5-10, MgCI2 1, glucose 10, HEPES 10. In some experiments GABAergic synaptic activity was stimulated by superfusion with 30/~M kainate. All experiments were carried out in the presence of 1 /tM tetrodotoxin, at room temperature (24 °C).

B

C j~

V

D 50mV ,

RESULTS

Intracellular recording from in vivo developing neurons in the isolated superfused superior colliculus The presence of synaptic activity in various n e u r o n types in the superficial gray layer of the neonatal rat SC was investigated, using a preparation of the isolated superfused tectum z°. Fig. 1 shows a representative set of recordings from an antidromically activated projection n e u r o n . The soma-dendritic profile of this n e u r o n was reconstructed after injection of H R P (Fig. 1A). The lower trace of Fig. 1B presents an inhibitory postsynaptic potential (IPSP), as activated with low stimulating currents via an microelectrode in the rostral SC. Intraceilular chloride injection reversed the IPSP. Increase in stimulation intensity activated an antidromic action potential prior to the inverted IPSP (Fig. 1B,C, top traces).

5

1mV

10ms

20 ms

100ms

Fig. 1. Somadendritic profile (A) and electrical activity (B-D) of a rostrally projecting neuron in the SGS at postnatal day 21 after development in vivo. A: reconstruction of somadendritic area and proximal axon (Ax) of an HRP-filled neuron in horizontal CS section. Inset: projection of the dendritic field onto the surface of the rat SC. B, upper 3 traces: responses to intracollicular microstimulation at decreasing (from top to bottom) intensity. An antidromically evoked action potential (truncated) is followed by an inverted IPSP. The IPSP prior to intracellular CI loading is shown in the bottom trace of B. C, upper trace: same as upper trace in B, but record taken at lower amplification. C, bottom trace: extracellular field potential corresponding to the upper trace in B,C. D: repetitive firing evoked by injection of depolarizing current pulses.

98

/

50,um Fig. 2. SGS-derived dissociated cell cultures. A: frontal section through the mesencephalon of an E21 rat embryo after staining with Cresyl violet. The superficial gray layer has a thickness of about 300 pm and can be dissected at the indicated level. B: preparation as in A, after cutting off the SGS for cell culture. C: dissociated cell culture after three weeks in vitro. D: large SGS-derived neuron, two weeks after plating. E: photomicrograph of an antidromically identified and HRP-stained projection neuron in the SGS of a 2-week-old rat (horizontal section). It can be seen that the size of larger neurons in SGS-derived cultures corresponds to that of larger multipolar projection cells in the brain. Calibration in E applies also to C, D. SGS neurons. It could therefore be predicted that" G A B A e r g i c synaptic connections are also developing in vitro, provided that the dissociation procedure does not significantly alter the cellular composition in culture, as compared to the SGS in vivo.

Autoradiographic identification of GABAergic neurons Cultures from the SGS of E21 rat embryos comprised a mixed population of cells (see section Cell cultures in Methods). After a few days in vitro phase bright cells staining with neuronal markers grew on top of a m o n o l a y e r formed by various non-neuronal cells, including Thy 1-positive fibroblasts and GFAP-positive astroglial cells. However, with the protocol used, [ 3 H ] G A B A uptake-labeling was low over the flat non-neuronal cells. Only some protoplasmic astrocytes were strongly [3HI-

G A B A uptake-positive (Fig. 3D). Photomicrographs of labeled multipolar and bipolar neurons in a mature SGS-derived culture ( D I V 16) are presented in Fig. 3 A - C , E. Most [ 3 H ] G A B A uptake-positive neurons had 'specialized' peripheral branching patterns, i.e. tufty or wavy dendrites. The geometry of some bipolar cells in vitro (Fig. 3E) closely resembled that of horizontal cells in the normally developing SGS (Fig. 3F). The diameter of [ 3 H ] G A B A uptake-positive neuronal perikarya (average of the maximal diameter and the perpendicular to the maximal diameter t9) ranged from 10 to 25 pm. This closely corresponds to the soma size of HRP-stained SGS neurons ex vivo. Counts of [ 3 H ] G A B A uptake-labeled neurons were performed at D I V 0 . 2 , 6, 16, 24 and 27 (Fig. 11). Fig. 4 presents photomicrographs from some of these cul-

99

'A,I.,

"

.:.

" "

,.:~,.~,

:

~'~,'-_-.....~

..-

. - . .

•.

.

-,

,:-

"-?.!. "?..:/ .

.•,

~ 1 ~ 1

-.~-..i' r.' ....

-.-.

/

".'

.

~

;

:"

::.' ;'.:,,

.

~"

,

- ,,,'

~::. :.. ....

.,~,~. ~.,. :' ,.~. •

~V ~.

.. -.

,

.

•~ j,,

"

"

"

.

/

':.-

~..7,,-.-,,

~

.

'

" .

.

..,

"

'

-'" ": "

•

'.:.~.':4.%

. ' "

.....~ .,~

a

':"

.

..

-

.'

:

;'

,'

~

.

.

.~,

''

"'

'

t "~

~" "

.

. "

'' t'

.

,

,

.

,"-

.

.

i

.

"'_':,,

'

.

.:

.~,. . .

....

.'

..

--

.~ "~ "

~.,, "3"'¸

.~/~'.",~

~,:

'

•~%'~;

.:

i --"

....

• .

:-

'.

,'~r~ .;'

-

•

;." : : / ~

-~-"% ~'~i~:--~"~"-

"

'

"

. . -

.

-

:~-~'~r",

•

•

•

:~.; ..;~;,

:..~,...:~,..,,... ~ ..:,

. ~.V'r'lk~':,.~¢,~-~':."L

" •

:

:. z;-'~.:.,~.f~S,j',~,*'.';.-.~

.,~"t_'~.f~

..

" '..., ""-~ :.~:

.:, "'~

'

~~-~'.< '"

":

".'7:-,';. .... . . . .

-r

.~.-':Jl~.'~" '"

.

"

:"~,

,_" .......

: l , " b

. v~.jr

"

-"

::

":'~-

D

.

-~7.

,...'. ~::

"

.~,.

,

•

... "- -

i

~, •

-

.~-~.,,

;~'. ~

.

" " :"~'." "'.*.~-,'.'~'. "'~-" ~...~Ii

..

-.~:::.f:':-.:i. ' . . . .

"~,2,;,.;' r

.

-"

..~.~<

..

.

"

..

.

.

". '<-';A

"

"

"

~

.

: :-,

"

a.,

"l

" " "

"

.":~ "-

r~ ,,.: , '

.

;~.~. -,

Fig. 3. Autoradiographic profiles of [3H]GABA uptake-positive neurons in a SGS-derived culture after 16 DIV. A-C: neurons. D: astroglial cell• E: bipolar neuron in culture. F: neuron of similar geometry as E, but from the in vivo grown SGS at postnatal day 24 (reconstruction after intracellular HRP-labeling in the sagittal plane). Calibration in F applies also to A-E. a, axon.

tures. O n e a n d the same field is shown with bright-field and with p h a s e contrast illumination to give an impression of the relative density of n o n - l a b e l e d cells (light arrows) and of strongly l a b e l e d neurons (heavy arrows). Since all but the clearly n o n - l a b e l e d neurons were r e g a r d e d as [ 3 H ] G A B A uptake-positive our counting

p r o c e d u r e p r o b a b l y tends to o v e r e s t i m a t e the relative n u m b e r of G A B A e r g i c n e u r o n s in these cultures. Ind e e d , the n u m b e r of [ 3 H ] G A B A uptake-positive neurons was always higher than the n u m b e r of G A D - p o s i t i v e p e r i k a r y a (Fig. 11). In m a t u r e cultures (after a b o u t 3 weeks in vitro) 4 5 - 6 0 % of all n e u r o n s displayed [3H]-

100

.-A 11~:;~~'~

\

\'\ ~5

C "4"

E Ax

dc

t

%

.i

J Fig. 4. Photomicrographs showing [3H]GABA uptake in SGS-derived cultures after various times in wtro. On the lelt: bright-field illumination. On the right: phase contrast illumination. Solid arrows: examples of [3H]GABA uptake-positive neurons. Light arrows: examples of [3H]GABA uptake-negative neurons. The multipolar or bipolar [3H]GABA uptake-positive neurons seen after 6-24 DIV are derived from initially spherical cells, as indicated by solid arrows in A, The larger labeled cell in A (empty arrow head) represents a neuron which retained the initial dendrites. These GABAergic neurons were already well developed at birth and not converted to spherical shape by the dissociation procedure. Such cells died off within the first 48 h in vitro. Ax, axon; GC, growth cone of a [3H]GABA uptake-positive axon.

G A B A uptake• A t D I V 0, the fraction of [3H]GABA uptake-positive n e u r o n s was 40%. The latter result supports the previous n o t i o n 35 that the G A B A e r g i c p h e n o t y p e is present in the SGS before birth. It is also clear that this rather large proportion of G A B A e r g i c n e u r o n s in SGS-derived cultures remains stable over the in vitro study period.

Identification of GAD-positive perikarya and terminals The G A B A - s y n t h e s i z i n g enzyme G A D was located in both perikarya and axonal terminals• Fig. 5 presents two photomicrographs from cultures which had been processed for G A D after nearly 3 weeks in vitro. N e u r o n a l perikarya that were regarded as G A D - I R , are indicated with solid arrows. Fig. 7C shows that ultrastructure of a

101 ~ll II

b~ iI.

o

|

"it .

t

-

|

.

r,>

m /

*it

I ,w

+

t

• B

m

i

•

,I

tk

•'

'

+

i..

--

"l

.

il

'

@

7"

~

.

i

#

-~' i

{

.

<

#

'i

.

o,:°

3.

-

Fig. 5. Photomicrographs of SGS-derived cultures after indirect immunostaining for GAD. GAD-positive perikarya are marked by arrows. Small dark puncta represent axonal terminals. Two larger unlabeled neurons (asterisks) in A and B are densely covered with GAD-positive axonal terminals. Cultures after 20 DIV,

102

.$ b

P

0

4

Fig. 6. Photomicrographs showing sister cultures at DIV 15 after processing for GAD-IR (A) and for [3H]GABA uptake (B,C). A, center: a large GAD-negative neuron is covered with few GAD-positive puncta (light arrow) which presumably represent axonal terminals. Note also the presence of dark GAD-IR perikarya in the vicinity (heavy arrows). B, between white arrows: a [3H]GABA uptake-positivc axon climbs alon~ a weakly labeled large neuron (phase contrast illumination) C: same under bright-field illumination.

Fig. 7. Light and electron microscopical images of GAD-positive terminals in SGS-derived cultures. A: semithin epon-araldite section (1-2 /~m thick). Counterstaining with Methylene blue. The large neuron (n) in the center is almost completely covered with GAD-positive terminals. Also covered are glial cells (gl). Two smaller neurons are nearly devoid of GAD-positive puncta. Culture after 20 DIV. B: EM image of a GAD-positive terminal contacting a large GAD-negative dendrite. Same neuron as A, center. C: collicular neurons at DIV 10. A GAD-positive neuronal perikaryon is marked by the arrow. The immunochemical reaction is weak at this time in vitro. D: EM image of a GAD-positive terminal contacting a GAD-negative dendrite. The reaction product gives only a light deposit around the vesicles. Vesicle size is 30-50 nm. Culture after 15 DIV.

104

Fig. 8. Electron micrographs showing vesicle-containing terminal profiles in cultures at various times in vitro. B: 6 DIV. D: 12 DIV. A,C,E: 24 DIV. Empty arrowheads indicate symmetrical contacts; filled arrowheads, non-symmetrical contacts, Note that few vesicles arc present in the terminal of the youngest culture shown in (B).

105

Fig. 9. Electron micrographs after immunostaining for GAD. A: one GAD-positive (3) and two GAD-negative (1,2) terminals contact the proximal dendrite (d) of a large neuron. Note that the GAD-positive terminal (3) forms a non-symmetrical contact. B: two vesicle-containing terminals contacting a perikaryal profile. The GAD-positive terminal (4) forms a symmetrical contact, and the GAD-negative terminal (5) forms a non-symmetrical contact. C: a GAD-positive terminal (6) contacts both a dendrite (d) and a vesicle-containing GAD-negative terminal. D: GAD-positive vesicle-containing neuronal process growing on top of a glial cell (gl). Cultures after 15 (A,D) and 20 (B,C) DIV. Bars in B - D = 0.5 ~m.

106 G A D - I R neuronal perikaryon. Although it was felt that evaluation of the number of G A D - I R perikarya gives a less reliable eatimate of the fraction of GABAergic neurons in SGS-derived cultures, such counts were performed and the results are given for comparison (Fig. 11). After two weeks in vitro the number of G A D - I R perikarya ranged from 28 to 39%. G A D - I R axonal terminals appear as dark brown puncta covering the surface of both G A D - I R positive and G A D - I R negative somata and dendrites. In Figs. 5 and 6A neurons tagged with an asterisk are densely covered with GAD-positive puncta. In semithin sections similar multipolar neurons were seen to receive GAD-positive boutons along the entire visible soma-dendritic profile (Fig. 7A). Subsequent processing for EM confirmed that these boutons show, in fact, synaptic specializations (Fig. 7B). Both terminaux and en passage types of vesiclecontaining boutons were labeled with the anti-GAD serum. G A D - I R was located along the presynaptic membrane, on synaptic vesicles, ribosomes, mitochondria and on neurofilaments (Figs. 7D and 9A-C). A study addressing in detail the principles governing the distribution of G A D - I R axon terminals is already in progress. Here, it is relevant to mention that the density of GAD-positive terminals substantially varied from cell

to cell. Densely covered neurons were found next to neurons bearing very few synaptic terminals (Figs. 5 and 6A). However, G A D - I R innervation did not spare any type of neuron, among them also the characteristic bipolar cells. The boutons were mostly issued by axons and contacted not only neurons, but also cells with clearlv non-neuronal geometry (denoted by 'gt' in Figs. 7A and 9D). The cellular phenotype of these non-neuronal cells has not been determined. It is unlikely that such contacts represent true synapses, as no synaptic specialization has been detected at the EM level. However, the possibility that the vesicles seen in some of the houtons on non-neuronal cells contain G A B A which is released by depolarization cannot be excluded. In this regard, it should be noted that GABA-activated C1- channels have recently been seen in glial cells3~. The development of G A D - I R lagged the development of [3H]GABA uptake. Fig. 6 shows for comparison the profiles of two multipolar neurons contacted by G A B A ergic axon terminals in sister cultures at DIV 15. One was processed for G A D - I R (A), and the other for [3H]G A B A uptake (B). It can be seen that after about two weeks in vitro, terminal G A D - I R is still weak, while strongly labeled axons in contact with labeled or nonlabeled somata and dendrites are present in autoradio-

B

A

C 50pA

100 pA

l

I 10s

~

20pAt •

20

+30 .

I

KA

GABA 10 pM

-80

[]

IIJ Ill

I

KA + Bic

I

,

]

pA

f~

f,

00 p A

20 pA

50 p A

O l()s

10 S

20 m s

Fig. 10. Patch-clamp recording from SOS-derived neurons in culture.A: 4 h afterplating. B,C: 44 h afterplating. As usually,E21 rat embryos

were chosen for dissecting the superficial gray layer. A: current response to superfusion with l0 pM GABA of a freshly plated SOS-derived neuron. Heavy horizontal bars indicate the period of drug application. Numbers by the traces refer to holding voltage. (Vh in mV), The I-V characteristic of the GABA-activated current reflects the CI- concentration gradient. The C1- equilibrium potential was set at -10 mV. At this Vh the GABA-activated current is zero. B,C: bicuculline-sensitive, CI--dependent synaptic activity is already present after 2 DIV. Single IO~ABA)SVNare shown in the top and bottom traces of C. B: during kainate (KA)-induced depolarization (30 pM KA) the frequency of ICI(GABA)SYNincreases. C: GABAergic synaptic activity is blocked by 25 pM bicuculline methiodide (Bic).

107 grahs of cultures from DIV 6 onwards (Fig. 6B,D).

(1,2,5). The fraction of GAD-positive terminals forming clearly symmetrical contacts was estimated as 69% (22 out of 32). This fraction is, thus, lower than that seen in the partially deafferented SGS, and much lower than in the normally developing rat SGS 26. In cultures at DIV 20 and 38 (pooled together) 53% (n = 244) of all synapses were GAD-IR.

The ultrastructure of GAD-positive terminals Fig. 8 shows the ultrastructure of synaptic arrangements formed by vesicle-containing terminals and postsynaptic dendrites in cultures of different age. These cultures were not processed for GAD and therefore retained their fine structure better. In D and E quite a few vesicle-containing terminals are seen to contact heavily impregnated postsynaptic contours (full arrows). The criteria for identification of synapses in vitro were similar to those in vivo 4'11'73. We required the presence of (1) presynaptic and postsynaptic structures with specialized membranes along a synaptic cleft, and (2) at least 4 synaptic vesicles in the presynaptic terminal in close proximity to the synaptic junction. Both symmetrical and non-symmetrical types of synapses were found. Synaptic contacts were regarded as symmetrical if the postsynaptic density had a thickness of less than 25 nm (Fig. 8A). Such contacts constituted 38% of all (n = 120) terminals in the untreated material. Fig. 9A,B shows the photomicrographs of dendrites bearing both GAD-positive and GAD-negative terminals. The former did not differ from the latter with respect to terminal size, vesicle shape or vesicle diameter. The average width of GADpositive terminals in cultures after 20-28 DIV ranged from 0.8 to 2.1/~m (mean 1.16/~m; S.D. 0.31; n = 32). The vesicles of GAD-positive terminals were polymorphic, with diameters between 30 to 50 nm. Fig. 9 shows also that GAD-positive terminals participated in contacts of both the symmetrical (4,6) and non-symmetrical (3) type. The same was true for GAD-negative terminals

Responsiveness to GABA Patch clamp experiments were carried out on neurons from mature cultures (between DIV 14 and 21) to find out whether the axosomatic and axodendritic contacts seen in our EM material can, in fact, be regarded as functional synapses. It was found that, in mature cultures, nearly all neurons displayed spontaneous bicuculline-sensitive C1- currents. In some cases, application of a single depolarizing pulse was sufficient to demonstrate the quantal -elease of endogenous GABA, as indicated by the generation of discrete bicuculline-sensitive CIcurrents (Ic,(~ABA)SVN)" The decay of evoked and spontaneous lCI(GABA)SVNwas fitted by a single exponential, the time constant of decay being about 25 ms (range: 10.8-38.3 ms) at a holding voltage o f - 5 0 mV. The unitary slope conductance of the Ia(GABA)SVN at maximum amplitudes ranged from 20 to 100 pS at negative holding voltages. The frequency of spontaneous ICI(GABA)SVN could be increased by superfusing the cells with a salt solution containing excitatory amino acids (Fig. 10B), and reduced by addition of 0.5 mM Cd 2÷ or 10 mM Co 2+, as well as other blockers of voltageactivated Ca 2+ channels. Tetrodotoxin, a blocker of voltage activated Na ÷ channels, had little effect on the

In vivo: Evoked I PSPcl

Response to exogenous GABA lOO-

,,

% cels

,,

r--, ,, ~/..~, ,.-., Fi/l_l

.,.-., .~

~

[]

[]

S~:~nt. P Ic) (GABA)

/0

0

0 Birth

5

I'0

GAD-R

I'5

2'0

2'5

DIV

3b

Eye opening

Fig. 11. Development of the GABAergic system in the SGS and in SGS-deriveddissociated cell cultures. The hatched horizontal bar indicates the time after birth when GABAergic Cl--dependent IPSPs could be recorded with intracellular microelectrodes from SGS neurons after development in vivo.

108 frequency of spontaneous 1C|(GABA)SYN"Fig. 10C shows that the generation of IC~(~ABA)SVN was completely blocked by 25 /~M of bicuculline methiodide. These properties of lC|(OAnA)SVN indicate that the spontaneous synaptic activity of cultured SGS-derived neurons is mediated by a G A B A A receptor 5'6~. In the presence of bicuculline, faster decaying, presumably excitatory synaptic currents, appeared occasionally in older cultures, but the respective transmitter substances have not yet been identified. Superfusion of SGS-derived neurons (n = 35) with salt solutions containing exogenous G A B A (5-25/~M) activated large currents that reversed at the Cl- equilibrium potential, the latter usually being set between -5 and 0 mV (symmetrical chloride concentrations). These currents were again blocked with bicuculline methiodide and picrotoxin.

Development of GABAergic terminals and responsiveness to GABA Since GABAergic IPSPs were found before eye opening (P9-12), it was of interest to use younger cultures to test for the capability of SGS-derived neurons to generate GABA-activated C1- currents. Fifteen freshly dissociated E21 neurons were exposed to G A B A superfusion (10/~M), and all of them responded as shown in Fig. 10A. Spontaneous bicuculline-sensitive synaptic activity was observed from D I V 2 onwards (Fig. 10B). Out of 10 neurons tested after 48 h in vitro, 6 generated spontaneous ICI(GABA)SYN"At this stage, G A D staining was still absent in the terminals. The first weakly stained GAD-positive puncta were seen with LM after 10-12 DIV. Between D I V 14 and 21, the number of G A D positive puncta and the intensity of their staining increased significantly, reaching a maximum around DIV 28. After D I V 28, the G A D - I R network was still intact and spontaneous GABA-release was significant, but the total number of neurons per dish quite clearly declined after D I V 28. At EM level, vesicle-containing terminals were detected by D I V 10. Fig. 8B represents the youngest axodendritic contact in our material of 244 contacts from cultures between D I V 3 and 31. The specimen was taken from a culture at D I V 6 and was not treated for G A D . Since careful examination of many sections obtained from cultures at D I V 3, 6 and 7 failed to reveal synaptic specializations at the EM level, it seems possible that chemically or structurally immature terminals may already be capable of activating GABA-currents which in all aspects resemble synaptic GABA-currents. Fig. 11 summarizes the results concerning the development of GABAergic synaptic connections in SGS-derived cultures.

DISCUSSION The results lead to the conclusion that neurons from the rat SGS establish extensive functional GABAergic connections in the absence of visual activity (SGS during normal development) and in the absence of retinal afferents (SGS-derived culture). The connections were characterized utilizing both morphological and electrophysiological criteria. We shall briefly discuss to what extent the features of the in vivo developing superficial gray layer of the rat superior colliculus were reproduced in culture, and which factors may have influenced the observed pattern of GABAergic termination~ In vivo studies have shown that in newborn rats, retinal axons have already invaded the SGS ~'':~s, and the precursors of intrinsic GABAergic neurons have completed their last mitotic cycle 2'46. G A B A immunostaining revealed labeled neurons in the tectum at El835, while [3H]GABA uptake is expressed even earlier, i.e. by El5 (R. Grantyn, unpublished). Although synaptogenesis starts in the upper SC on day E173s~ the density of vesicle-containing synaptic contacts in the SGS remains low until eye opening 37"v3. After eye opening (P12-13), the number of synaptic contacts dramatically increases ~7" 73. It has been suggested that visual activity may stimulate the formation of intrinsic networks, including GABAergic connections 3v. Yet it was not known at which time GABAergic synapses are formed in vivo. The present results indicate that the first GABAergic contacts are formed before day P9. Since GABAergic neurons were found among the cells dissociated from the embryonic SGS it is quite possible that these neurons form functional connections even before birth. The morphogenesis of GAD-labeled synapses has been described for several brain structures 15"~smJ'3, but no G A D / E M investigation has, to our knowledge, been carried out using the preand early postnatal rat SC. Since SGS-derived cultures were prepared at the end of the rat's gestational period and maintained for 4-6 weeks, the observation almost covers the whole time course of synapse formation in this brain structure. The formation of GAD-positive synaptic connections between DIV 14 and 21 corresponds to the period of massive synaptogenesis in the postnatal SGS. Our experiments showed that SGS-derived neurons already display spontaneous GABA-mediated activity by D I V 2. G A B A release may thus play a role in the regulation of neuronal growth and initiation of synaptogenesis 375. Our demonstration of spontaneous ICI(GABA)SYNin very young cultures is in line with previous observations indicating that G A B A might be released from nonsynaptic terminal structures such as growth c o n e s 17'~s'41"77 or dendrites 2643. The discrete character of evoked and

109 spontaneous ICI(GABA)SYNand the reduced frequency of these currents in the presence of Ca 2÷ channel blockers suggests that the G A B A release is due to vesicle exocytosis, rather than inverse G A B A uptake from an extravesicular G A B A pool. It is rather puzzling that discrete G A B A release can be demonstrated with patch clamp recording as early as DIV 2, although no vesiclecontaining terminals were found before DIV 6 or 10. This contradiction can only be resolved with more specific experiments to clarify the nature of G A B A release by immature terminals of GABAergic neurons. The fact that GAD-staining of axon terminals was virtually absent until DIV 10-12 may reflect the relatively low sensitivity of our GAD-immunostaining procedure, as compared to the sensitivity of the patch clamp technique as a probe for terminal maturation. A low, as compared to G A B A levels, activity of G A D at birth, and a slow increase of G A D activity during the first postnatal weeks was shown not only in culture, but also in vivo 1°. GABAergic synapses in vitro have previously been demonstrated in other culture systems by labeling the terminals and subsequent EM investigation 7, by recording spontaneous or evoked synaptic activity m6°, or by measuring G A B A release 1"13'23'28'32'57'62'78. Generally, both G A D activity and G A B A release were found to remain at low levels until DIV 7-1023'29'32, and to increase roughly in parallel between the second and third week in culture. The possibility that long-term cultures provide conditions which are preferential for the survival and development of GABAergic neurons has been repeatedly muted 24"59"67'78. This should certainly be the case if the dissociated brain structures contain a large number of target-dependent neurons which cannot be maintained under the given culture conditions. However, the projection neurons of the rat SGS issue numerous intrinsic collaterals 21 and may therefore find targets in culture. Moreover, the fraction of neurons with [3H]GABA uptake or perikaryal GAD-labeling in SGS-derived cultures was similar to that in the normally grown (although feline) SGS 44. It can therefore be assumed that the composition of SGS-derived cultures roughly matches the in vivo situation. The more crucial difference between the SGS in vivo and SGS-derived cultures would seem to be the presence or absence of extrinsic (visual) afferents and the spatial REFERENCES 1 Aloisi, E, Ciotti, M.T. and Levi, G., Characterization of GABAergic neurons in cerebellar primary cultures and selective neurotoxic effects of a serum fraction, J. Neurosci., 5 (1985) 2001-2008. 2 Altman, J. and Bayer, S.A., Time of origin of neurons of the rat superior colliculus in relation to other components of the visual and visuomotor pathways, Exp. Brain Res., 42 (1981) 424-434.

order imposed by the latter. The results of Houser et al. 26 predict that in culture a noticeably higher fraction of GAD-positive terminals would form asymmetrical contacts. Our results are consistent with this hypothesis in that about 30% of G A D - I R terminals participate in junctions which have postsynaptic densities of more than 25 nm. The reduced quality of the EM material after the GAD-immunohistochemistry often makes it difficult to qualify synaptic contacts as belonging to either the symmetrical or the asymmetrical type. Since the responses to exogenous G A B A and spontaneous G A B A ergic activity did not display any unusual features, it remains unclear which receptor- or other properties correspond to the diversity in the appearance of G A B A ergic synapses. As all synaptic complexes were formed de novo, it can be concluded that GAD-positive terminals do not occupy vacant sites liberated from other inputs, but rather induce an ultrastructural pattern which is considered as characteristic for non-GABAergic synapses. The observed termination of G A D - I R boutons on both GAD-positive and GAD-negative perikarya and dendrites is in contrast with previous notions suggesting that some SGS neurons, in particular the presumed GABAergic neurons, are free of GABAergic inhibition 3°'6°. The extracellular recording techniques in the experiments on intact animals were probably not always adequate for testing neuronal responsiveness to GABA. Our experiments on the isolated superfused SGS of postnatal rats revealed large IPSPs in all intrasomatically impaled neurons from day P14 onwards and, thus underline the significance of GABAergic inhibition for the visual functions carried out in this structure. Acknowledgements. The authors wish to express their gratitude to Dr. W.H. Oertel for providing the GAD-antiserum and for critical reading of the manuscript, Prof. H. Holl/inder for providing electron microscopic facilities, Dr. A. Rodriguez-T6bar for his help during the experiments with [3H]GABA uptake, Mrs. A.K.E. Horn for advice with the GAD-immunostaining, Dr. C. Parsons for useful comments on the manuscript, and Mrs. Ch. Pfitzner for excellent secretarial assistance. Dr. R. Schfimann reconstructed the HRPfilled neurons shown in Figs. 1A and 3E S.S.W. was a stipendiate of the Max Planck Society. The study was supported by the National Health and Medical Research Council of Australia and by Sigma Tau, Rome. The experiments on the isolated superfused tectum were carried out by R.G. while being at the Carl Ludwig Institute of Physiologyin Leipzig (G.D.R.)

3 Balcar, V.E, Dammasch, I. and Wolff, J.R., Is there a non-synaptic component in the K+-stimulated release of GABA in the developing rat cortex?, Dev. Brain Res., 10 (1983) 309-311. 4 Blue, M.E. and Parnavelas, R.G., The formation and maturation of synapses in the visual cortex of the rat, J. Neurocytol., 12 (1983) 599-616. 5 Bormann, J., Electrophysiology of GABAA and GABAB receptor subtypes, Trends Neurosci., 11 (1988) 112-116.

110 6 Bunt, S.M., Lund, R.D. and Land, P.W., Prenatal development of the optic projection in albino and hooded rats, Dev. Brain Res., 6 (1983) 149-168. 7 Burry, R.W. and Lasher, R.S., Electron microscopic autoradiography of the uptake of [3H]GABA in dispersed cell cultures of rat cerebellums. I. The morpholgy of the GABAergic synapse, Brain Res., 151 (1978a) 1-17. 8 Burry, R.W. and Lasher, R.S., Electron microscopic autoradiography of the uptake of [3H]GABA in dispersed cell cultures of rat cerebellums. II. The development of GABAergic synapses, Brain Res., 151 (1978b) 19-29. 9 Carbone, E. and Lux, H.D., o)-Conotoxin blockade distinguishes Ca from Na permeable states in neuronal calcium channels, Pflfiger's Arch., 413 (1988) 14-22. 10 Coyle, ,I.T. and Enna, S.J., Neurochemical aspects of the ontogenesis of GABAergic neurons in the rat brain, Brain Res., 111 (1976) 119-133. 11 Devon, R.M. and Jones, D.G., Synaptic terminal parameters in unanaesthetized rat cerebral cortex, Cell Tissue Res., 203 (1979) 189-200. 12 Dichter, M.A., Physiological identification of GABA as the inhibitory transmitter for mammalian cortical neurons in cell culture, Brain Res., 190 (1980) 111-121. 13 Farb, D.H., Berg, D.K. and Fischbach, G.D., Uptake and release of [3H]-aminobutyric acid by embryonic spinal cord neurons in dissociated cell culture, J. Cell Biol., 80 (1979) 651-661. 14 Fisher, R.S., Levine, M.S., Adinolfi, A.M., Hull, C.D. and Buchwald, N.A., The morphogenesis of glutamic acid decarboxylase in the neostriatum of the cat: neuronal and ultrastructural localization, Dev. Brain Res., 33 (1987) 215-234. 15 Fonnum, F., The distribution of glutamate decarboxylase and aspartate transaminase in subcellular fraction of rat and guinea pig brain, Biochem. J., 106 (1968) 401-412. 16 Fosse, V.M., Heggelund, P., Iversen, E. and Fonnum, E, Effects of area 17 ablation on neurotransmitter parameters in efferents to area 18, the lateral geniculate body, pulvinar and superior colliculus in the cat, Neurosci. Lett., 52 (1984) 323-328. 17 Gordon-Weeks, P.R., Lockerbie, R.O. and Pearce, B.R., Uptake and release of [3H]GABA by growth cones isolated from neonatal rat brain, Neurosci. Lett., 52 (1984) 205-210. 18 Gotow, T. and Sotelo, C., Postnatal development of the inferior olivary complex in the rat. IV. Synaptogenesis of GABAergic efferents, analyzed by glutamic acid decarboxylase immunocytochemistry, J. Comp. Neurol., 263 (1987) 526-552. 19 Grantyn, R., Grantyn, A. and Schierwagen, A., Passive membrane properties, afterpotentials and repetitive firing of superior colliculus neurons studied in the anaesthetized cat, Exp. Brain Res., 50 (1983) 377-391. 20 Grantyn, R., Ludwig, R. and Eberhardt, W., Neurons of the superficial tectal gray. An intracellular HRP-study on the kitten superior colliculus in vitro, Exp. Brain Res., 55 (1984) 172-176. 21 Grantyn, R., Gaze control through superior colliculus: structure and function. In J. Buettner-Ennever (Ed.), Anatomy of the Oculomotor System, Reviews of Oculomotor Research, Vol. 2, Elsevier, Amsterdam, 1988, pp. 273-331, 22 Grantyn, R. and Lux, H.D., Similarity and mutual exclusion of NMDA- and proton-activated transient Na ÷ currents in rat tectal neurons, Neurosci. Lett., 89 (1988) 198-203. 23 Hauser, K., Balcar, V.J. and Bernasconi, R., Development of G A B A neurons in dissociated cell culture of rat cerebral cortex, Brain Res, Bull., 5 (1980) 37-41. 24 Hoch, D.B. and Dingledine, R., GABAergic neurons in rat hippocampal culture, Dev. Brain Res., 25 (1986) 53-64. 25 Hodgson, A.J., Penke, B., Erdei, A., Chubb, I.W. and Somogyi, P., Antisera to y-aminobutyric acid. I. Production and characterization using a new model system, J. Histochem. Cytochem., 33 (1985) 229-239. 26 Houser, C.R., Lee, M. and Vaughn, J.E., Immunocytochemical localization of glutamic acid decarboxylase in normal and

27

28

29

30

31

32

33

34

35

36

37 38

39

40

41

42

43 44

45

46

47

deafferented superior colliculus: evidencc for ~eorganizauon oi ),-aminobutyric acid synapses, ]. Neurosci., 3 (!083) 2030-2042. Huerta, M.E and Harting, J.K.. The mammalian superior colliculus: studies of its morphology and connections. In I! Vanegas (Ed.), Comparative Neurology o£ the Omic Tectam Plenum, New York, 1984, pp. 687-773. Hyndman, A.G. and Adler, R., G A B A uptake and release in purified neuronal and non-neuronal cultures from Chick embryo retina, Dev. Brain Res., 3 (1982) 167-180. Jong, Y.-J., Thampy, K.G. and Barnes, E . M , Ontogeny of GABAergic neurons in chick brain: studies in vivo and in vitro, Dev. Brain Res., 25 (1986) 83-90. Kayama, Y., Fukuda, Y. and Iwama, K., GABA sensitivity of neurons of the visual layer in the rat superior colliculus, Brain Res., 192 (1980) 121-131. Kettenmann, H. Backus, K.H. and Schachner, M., /-Aminobutyric acid opens CI channels in cultured astrocytes, Brain Res., 404 (1987) 1-9. Kuriyama, K.. Tomono, S., Kishi, M., Mukainaka, T. and Ohkuma, S., Development of y-aminobutyric acid (GABA)ergic neurons in cerebral cortical neurons in primary culture, Brain Res., 416 (1987) 7-21. Langer, T.P. and Lund, R.D., The upper layers of the superior colliculus of the rat: a Golgi study, J. Comp. Neurol.i 158 (1974) 405-436. Lasher, R.S., The uptake of [3H]GABA and differentiation of stellate neurons in cultures of dissociated postnatal rat cerebellum, Brain Res., 69 (1974) 235-254. Lauder, J.M., Han, V.K.M., Henderson, P., Verdoorn, 1'. and Towle, A.C., Prenatal ontogeny of the GABAergic system in the rat brain: an immunocytochemical study, Neuroscience, 19 (1986) 465-493. Lund, R.D., Synaptic patterns of the superficial layers of the superior colliculus of the rat, J. Comp. Neurol., 135 (1969) 179-208. Lund, R.D. and Lund, J.S., Development of synaptic patterns in the superior colliculus of the rat, Brain Res., 42 (1.972) 1-20. Lund, R.D. and Bunt, A.H., Prenatal development of central optic pathways in albino rats, J. Comp. Neurol., 165 (1976) 247-264. Lund Karlsen, R. and Fonnum, F., Evidence for glutamate as a neurotransmitter in the corticofugal fibres to the dorsal lateral genieulate body and the superior coUiculus in rats, Brain Res., 151 (1978) 457-467. McLaughlin, B.J., Wood, J.G., Saito, K., Barber, R., Vaughn, .I.E., Roberts, E. and Wu, ,I.-Y., The fine structural localizatiOn of glutamate decarboxylase in synaptic terminals of rodent cerebellum, Brain Res., 76 (1974) 377-391. McLaughlin, B.J., Wood, J.G., Saito, K., Roberts, E. and Wu, J.-Y., The fine structural localization of glutamate decarboxylase in developing axonal processes and presynaptic terminals of rodent cerebellum, Brain Res., 85 (1975) 355"371. Mize, R.R., Spencer, R.F. and Sterling, P., Neurons and glia in cat superior colliculus accumulate [JH]gamma-aminobutyric acid (GABA), J. Comp. Neurol., 202 (1981) 385-396. Mize, R.R., Spencer, R.F. and Sterling, P,. Two types of GABA-accumulating neurons in the superficial gray layer of the cat superior colliculus, J. Comp. Neurol., 206 (1982) 180-192. Mize, R.R., Immunoeytochemical localization of gamma-aminobutyric acid (GABA) in the cat superior colliculus, J: Comp, Neurol., 275 (1988) 168-187. Mugnaini, E. and Oertel, W.H., An atlas of the distribution of GABAergic neurons and terminals in the rat CNS as revealed by GAD immunohistochemistry. In A. Bj6rklund and T. H6kfelt (Eds.), Handbook of Chemical Neuroanatomy, Vol. 4, Part I, Elsevier, Amsterdam, 1985, pp. 436-622. Mustari, M.J., Lund, R.D. and Graubard, K., Histogenesis of the superior colliculus of the albino rat: a tritiated thymidine study, Brain Res., 164 (1979) 39-52. Nagai, T., McGeer, P.L. and McGeer, E.G.. Distribution of

111

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

GABA-T-intensive neurons in the rat forebrain and midbrain, J. Comp. Neurol., 218 (1983) 220-238. Neale, E.A., Oertel, W.H., Bowers, L.M. and Weise, V.K., Glutamate decarboxylase immunoreactivity and [3H]aminobutyric acid accumulation within the same neurons in dissociated cell cultures of cerebral cortex, J. Neurosci., 3 (1983) 376-382. 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 injection of an antigen-antibody complex, Neuroscience, 6 (1981) 2689-2700. Okada, Y., Distribution of y-aminobutyric acid (GABA) in the layers of superior colliculus of the rabbit, Brain Res., 75 (1974) 362-365. Okada, Y., Distribution of G A B A and GAD activity in the layers of superior colliculus of the rabbit. In E. Roberts, T.N. Chase and D.B. Tower (Eds.), GABA in Nervous System Function, Raven, New York, 1976, pp. 229-233. Ottersen, O.P. and Storm-Mathisen, J., Neurons containing or accumulating transmitter amino acids. In A. Bj6rklund and T. H6kfelt (Eds.), Handbook of Chemical Neuroanatomy, VoL 3, Part II, Elsevier, Amsterdam, 1984, pp. 141-246. Panula, P., Wu, J.-Y. and Emson, P., Ultrastructure of GABAneurons in cultures of rat neostriatum, Brain Res., 219 (1981) 202-207. Perouansky, M. and Grantyn, R., Separation of quisqualateand kainate-selective glutamate-receptors in cultured neurons from the rat superior colliculus, J. Neurosci., 9 (1989) 70-80. Raft, M.C., Fields, K.L., Hakomori, S.I., Mirsky, R., Pruss, R.M. and Winter, J., Cell-type-specific markers for distinguishing and studying neurons and the major classes of glial cells in culture, Brain Res., 174 (1979) 283-308. Ramon y Cajal, S., Beitrag zum Studium der Medulla Oblongata des Kleinhirns und des Ursprungs der Hirnn'erven, Johann Abrosius Barth, Leipzig, 1896, pp. 26-32. Reisert, I., Jirikowski, G., Pilgrim, C. and Tappaz, M.L., GABAergic neurons in dissociated cultures of rat hypothalamus, septum, and midbrain, Cell Tissue Res., 229 (1983) 685-694. Ribak, C.E., Vaughn, J.E. and Barber, R.P., Immunocytochemical localization of GABAergic neurones at the electron microscopical level, Histochem. J., 13 (1981) 555-582. Robain, O., Barbin, G., Ben-Ari, Y., Rozenberg, E and Prochiantz, A., GABAergic neurons of the hippocampus: development in homotopic grafts and in dissociated cell cultures, Neurosci., 23 (1987) 73-86. Segal, M., Rat hippocampal neurons in culture: responses to electrical and chemical stimuli, J. Neurophysiol., 50 (1983) 1249-1264. Segal, M. and Barker, J.L., Rat hippocampal neurons in culture: voltage clamp analysis of inhibitory synaptic connections, J. Neurophysiol., 52 (1984) 469-487. Shalaby, I.A., Won, L. and Wainer, B., Biochemical and morphological studies on GABA neurons in reaggregate culture, Brain Res., 402 (1987) 68-77. Shotweli, S.L., Shatz, C.J. and Luskin, M.B., Development of glutamic acid decarboxylase immunoreactivity in the cat's lateral geniculate nucleus, J. Neurosci., 6 (1986) 1410-1423. Sotelo, C., Gotow, T. and Wassef, M., Localization of glutamic-

65

66

67

68

69

70

71

72

73

74

75

76

77

78

acid-decarboxylase-immunoreactive axon terminals in the inferior olive of the rat, with special emphasis on anatomical relations between GABAergic synapses and dendrodendritic gap junctions, J. Comp. Neurol., 252 (1986) 32-50. Storm-Mathisen, J., Leknes, A.K., Bore, A.T., Vaaland, J.L., Edminson, P., Haug, E-M.S. and Ottersen, O.P., First visualization of glutamate and G A B A in neurones by immunocytochemistry, Nature (Lond.), 301 (1983) 517-520. Straschill, M. and Perwein, J., Effect of iontophoretically applied biogenic amines and of cholinomimetic substances upon the activity of neurons in the superior colliculus and mesencephalic reticular formation of the cat, Pfliiger's Arch., 324 (1971) 43-55. Thampy, K.G., Sauls, C.D., Brinkley, B.R. and Barnes, E.M., Neurons from chick embryo cerebrum: ultrastructural and biochemical development in vitro, Dev. Brain Res., 8 (1983) 101-110. Thong, I.G. and Dreher, B., The development of the corticotectal pathway in the albino rat, Dev. Brain Res., 25 (1986) 227-238. Tigges, M. and Tigges, J., Presynaptic dendrite cells and two other classes of neurons in the superficial layers of the superior colliculus of the chimpanzee, Cell Tissue Res., 162 (1975) 279-295. Valverde, E, The neuropil in superficial layers of the superior colliculus of the mouse. A correlated Golgi and electron microscopic study, Z. Anat. Entwickl.-Gesch., 142 (1973) 117147. Walker, C.R. and Peacock, J.H., Development of GABAergic function of dissociated hippocampal cultures from fetal mice, Dev. Brain Res., 2 (1982) 541-555. Warton, S.S. and Jones, D.G., Postnatal development of the superficial layers in the rat superior colliculus: a study with Golgi-Cox and Klfiver-Barrera techniques, Exp. Brain Res., 58 (1985) 490-502. Warton, S.S. and McCart, R., Synaptogenesis in the stratum griseum superficiale of the superior colliculus, Synapse, 3 (1989) 136-148. Weiss, D.S., Barnes, E.M. and Hablitz, J.J., Whole-cell and single channel recordings of GABA-gated currents in cultured chick cerebral neurons, J. Neurophysiol., 59 (1988) 495-513. Wolff, J.R., Evidence for a dual role of G A B A as a synaptic transmitter and a promoter of synaptogenesis. In EV. DeFeudis and P. Mandel (Eds.), Amino Acid Neurotransmitters, Raven, New York, 1981, pp. 459-466. Wu, J.-Y., Matsuda, T. and Roberts, E., Purification and characterization of glutamate decarboxylase from mouse brain, J. Biol. Chem., 248 (1973) 3029-3034. Young, S.H. and Poo, M., Spontaneous release of transmitter from growth cones of embryonic neurones, Nature (Lond.), 305 (1983) 634-637. Yu, A.C.H., Hertz, E. and Hertz, L., Alterations in uptake and release rates for GABA, glutamate, and glutamine during biochemical maturation of highly purified cultures of cerebral cortical neurons, a GABAergic preparation, J. Neurochem., 42 (1984) 951-960.