Int. J. Devl. Neuroscience. Vol. 5. No. 3, pp. 263-269. 1987. Printed in Great Britain.

(1736-5748/87 $03.(111+11.(10 Pergamon Journals Ltd.

~) 1987 ISDN

TEMPORAL DEVELOPMENT OF GABA AGONIST INDUCED ALTERATIONS IN ULTRASTRUCTURE AND GABA RECEPTOR EXPRESSION IN CULTURED CEREBELLAR GRANULE CELLS

GERT H. HANSEN, BO BELHAGE, ARNE SCHOUSBOE and EDDI MEIER*

Department of Biochemistry A, Panum Institute, University of Copenhagen, DK-2200 Copenhagen, Denmark

(Received 13 November 1986; in revised form 16 March 1987; accepted 20 March 1987)

Abstract--The temporal development of the effect of THIP (4,5,6,7-tetrahydroisoxazolo[5,4-c]pyridin- 3-ol) on the ultrastructure composition and GABA receptor expression in cerebellar granule cells was investigated by quantitative electron microscopy (morphometric analysis) and GABA binding assays. It was found that the cytoplasmic density of smooth endoplasmic reticulum was decreased, while the cytoplasmic density of rough endoplasmic reticulum, Golgi apparatus, vesicles and coated vesicles was greatly enhanced after exposure of the cells to THIP (150 p.M) for only 1 hr. In cerebellar granule cells exposed to THIP (150 p.M) for 3 hr low affinity GABA receptors were induced. These findings show that the effect of THIP on the ultrastructure composition and GABA receptor expression in cultured cerebellar granule cells may be interrelated and moreover it is likely that the turn-over of GABA receptors is extremely fast.

Key words: Cerebellar granule cells, Temporal development, Cultured neurons, THIP, Trophic action, Ultrastructure, GABA receptors.

T h e inh ib i to ry n e u r o t r a n s m i t t e r G A B A has been d e m o n s t r a t e d to exhib i t a t r o p h i c effect on neu rona l d e v e l o p m e n t in the b ra in in vivo, 34"35 in re t ina t3-nS"23 and in d i f fe ren t types of cu l tu red neu rons such as n e u r o b l a s t o m a cells 7"3° and c e r e be l l a r g ranu le cells. 3"n°'n8"27'28 The effect o f

G A B A on the m o r p h o l o g y of c e r ebe l l a r g ranu le cells "~ could be in i t i a ted by e i the r an ext ra- or in t race l lu la r ac t ion , s ince the ce r ebe l l a r g ranu le cells a re p e r m e a b l e to G A B A . 36 Us ing the specific G A B A r e c e p t o r agonis t T H I P 26 (4 ,5 ,6 ,7 - te t rahydro isoxazolo[5 ,4-c ]pyr id in-3-01) and the an tagon i s t b icucul l ine m e t h o b r o m i d e 5 it has found tha t the effect o f G A B A on cell m o r p h o l o g y was m i m i c k e d by T H I P and b locked by bicucul l ine . 2°'27"2s This s t rongly ind ica t ed tha t the neuro - t roph ic effect o f G A B A is m e d i a t e d by G A B A recep to rs .

In the b ra in two d i f fe ren t b ind ing sites with d i f fe ren t affinit ies for G A B A have been d e m o n s t r a t e d . Is C e r e b e l l a r g ranu le cells in cu l tu re exhib i t , however , on ly the high affini ty G A B A r e c e p t o r s unless they are cu l tu red in the p re sence of G A B A Z9"27'28 o r T H I P , 3'27"2s which

induce low affini ty G A B A recep to rs . A l so this effect of G A B A has been shown to be m e d i a t e d by G A B A recep to r s . 3

V e r y l i t t le is, h o w e v e r , known a b o u t the t e m p o r a l d e v e l o p m e n t of these t roph ic ac t ions of G A B A and G A B A agonis ts . T o e luc ida te the t ime course of the T H I P induced changes in u l t ra- s t ruc tu re c o m p o s i t i o n as well as in G A B A r e c e p t o r expres s ion , the p re sen t s tudy was ca r r i ed out in which 4 -day-o ld ce r ebe l l a r g ranu le cell cu l tu res were e x p o s e d to T H I P (150 IxM) for , respec- t ive ly , 1, 3, 6 and 24 hr. Subsequen t ly , the u l t r a s t ruc tu re c o m p o s i t i o n as well as G A B A r e c e p t o r express ion were inves t iga ted by , respec t ive ly , quan t i t a t ive e l ec t ron mic roscopy ( m o r p h o m e t r i c analysis) and G A B A b ind ing assays.

A p r e l i m i n a r y accoun t of this work has been given prev ious ly . I t

E X P E R I M E N T A L P R O C E D U R E S

M a t e r i a l s

Seven -day -o ld rats (Wis t a r ) were o b t a i n e d f rom the an imal q u a r t e r in the P a n u m Ins t i tu te . Plast ic t issue cu l tu re d ishes and cu l ture f lasks were pu rchased f rom N U N C A/S , D e n m a r k and

*Present address: Department of Pharmacology, H. Lundbeck & Co., DK-2500 Valby, Denmark. Address correspondence to: Dr Gert H. Hansen, Department of Biochemistry A, Panum Institute, University of

Copenhagen, Blegdamsvej 3, DK-2200 Copenhagen N, Denmark.

263

264 G.H. Hansen et al.

fetal calf serum from Gibco/Biocult Lab. Ltd., Scotland. Poly-L-lysine (molecular weight greater than 300,000), trypsin, trypsin inhibitor, DNAse and amino acids were obtained from Sigma Chemicals (St. Louis, MO, U.S.A.) , insulin from NOVO, Denmark and penicillin from Leo, Denmark. [2,3-~H]GABA (specific radioactivity 33 Ci/mmol) was purchased from New England Nuclear Company, Boston, MA, U.S.A. Methaacrylic acid-2-hydroxy-propylester was from Merck, Darmstadt, F.R.G. and EPON 812 resin kit from TAAB, Reading, Berks, England. THIP was synthesized by Dr P. Krogsgaard-Larsen, Department of Chemistry, B C , Royal Danish School of Pharmacy, Copenhagen, Denmark.

Cell cultures

Granule cells were cultured from cerebella of 7-day-old rats essentially as described by Messer 2j and Hansen et al. "~ Cells were isolated according to Wiikin et al. ~'- by mild trypsinization (0.025% w/v) followed by trituration in a DNAse solution (0.004% w/v) containing a trypsin in- hibitor from soybeans (0.03% w/v). The cells were suspended in Dulbecco's minimum essential medium containing 10% (v/v) fetal calf serum at a concentration of 3 z 1 0 7 cells/mi. The cell sus- pension was subsequently inoculated into poly-L-lysine coated dishes 29 or culture flasks. The medium contained 24.5 mM KCI, 30 mM glucose, 7 ~M p-aminobenzoic acid and 100 mU/l insulin. After 48 hr in culture 20 IxM cytosine arabinoside was added to the culture medium and the cells were cultured in this medium for the remaining part of the culture period. After 96 hr in culture 150 IxM THIP was added to half of the cultures. The cells were subsequently fixed for ultrastructure analysis or used for G A B A binding assays after, respectively, 1, 3, 6 and 24 hr exposure to THIP.

M o r p h o m e t r i c analysis

For ultrastructure analysis cultures were fixed at 4°C by draining off the culture media and adding 0.1 M sodium phosphate buffered glutaraidehyde (2.5% v/v) for 10 min. The cells were postfixed in 1% (w/v) osmium tetroxide in 0.1 M sodium phosphate buffer (pH 7.2) for i hr at 4°C. After careful washing in distilled water the cells were blockstained in 1% (w/v) uranyl acetate for 1 hr at 20°C in complete darkness.

Dehydration was carried out at room temperature in graded series of ethanol 35% (v/v) for 5 min, 50% (v/v) for 5 min, 70% (v/v) for 5 min and 90% (v/v) for 15 min and graded methaacrylic acid-2-hydroxy-propylester (HPMA)/ethanol series [90% (v/v) for 3 x 15 min, 95% (v/v) for 15 min and 97% (v/v) for 15 mini. The cells were subsequently exposed to graded series of HPMA/Epon (2:1 for 15 min, 1:1 for 15 min and 1:2 for 30 min) and then to pure Epon (3 x 10 min). Finally, the cultures were embedded in Epon and polymerized at 60°C for 24 hr.

Ultrathin sections (approximately 75 nm) were cut on a LKB Ultrotome 8800 III using glass knives prepared on a LKB knifemaker 7800. Subsequently, sections were collected on 400 mesh nickel grids (previously rinsed in 20% acetic acid, 99% ethanol and distilled water) and stained for 30-45 sec in lead citrate. 24 Electron micrographs were taken using a Philips electron micro- scope 201c operated at 60 kV. Morphometric analyses were carried out by the point method. 33 A transparent screen of 400 points was applied at random to the micrographs and the cytoplasmic density of the investigated organelles was calculated. Only cells having characteristic granule cell morphology, i.e. prominent nuclei "~'27 were analysed. Before sections entered the electron microscope grids were randomized and assigned a code, the meaning of which was unknown to the person carrying out the analysis. When all data from the morphometric analyses had been collected, the code was broken and data subjected to statistical analysis.

Statistical evaluat ion

The data from the morphometric analyses were subjected to two-way analysis of variance as described in standard textbooks of statistical methods.t'9 It was, however, first tested if the pre- requisite for this type of analysis was fulfilled, i.e. to test if the estimates of variance calculated from each group of data to be compared were estimates of the same population value. 2 If signifi- cant differences between the calculated estimates of variance were observed, the data were trans- formed to fulfill the prerequisite for variance analysis. 6 In cases where the two-way analysis of variance showed significant interaction between the two criteria, drug-treatment and time in

Time course of the neurotrophic activity of GABA 265

vitro, the variation between the two criteria was split to test which group of data gave rise to the significant interaction.

Preparation of membranes The granule cells were harvested in 0.32 M sucrose (0-5°C) and sedimented for 10 min at 900 g.

The pellet was resuspended in 0.2 ml 0.32 M sucrose (0-5°C) (0.2 ml per culture flask) and homogenized with a glass homogenizer equipped with a teflon piston. The homogenate was cen- trifuged at 35,000 g, 20 min, 2°C. The pellet was dispersed in water (0°C) by ultrasonication and centrifuged at 35,000 g, 20 min, 2°C. The ultrasonication and last centrifugation were done twice, and the final pellets were frozen. After thawing, the pellets were resuspended in 50 mM Tris- citrate buffer (pH 7.4) containing 0.04% (v/v) Triton X-100 (0.35 ml per culture flask) and incubated for 30 min at 37°C. The suspension was subsequently centrifuged (35,000 g, 20 min, 2°C), resuspended (0.35 ml/culture flask) in the Tris-citrate buffer at 0°C (twice), and the pellet from the last centrifugation was resuspended in 4 ml ice-cold 50 mM Tris-citrate buffer.

Receptor binding GABA-recep to r binding assays were performed as described by Enna and Snyder 8 and Meier

and Schousboe. 17 Ice-cold membrane suspension (100 Ixl) was added to 100 txl ice-cold 3H-GABA in different concentrations. The two suspensions were carefully mixed and incubated for exactly 20 min at 0°C before centrifugation at 35,000 g, 20 min, 2°C. After removal of the supernatant the pellet was cautiously rinsed with 1 ml ice-cold distilled water, and subsequently the tubes were dried with a piece of Kleenex and the pellets dissolved in 100 Ixl 0.1 M NaOH by incubation for 30 min at 37°C. Samples were taken for measurement of radioactivity, using a Packard Tricarb liquid scintillation spectrometer and for protein determination using the Lowry method 12 with serum albumin as the standard. The following concentrations of 3H-GABA were used for the binding assays: 0.5, 1.0, 3.0, 6.0, 12.0, 25, 50, 100,250,600, and 1460 nM. Non-specific binding which was linear as a function of the G A B A concentration was estimated at 25 nM G A B A by addition of 1 mM non-radioactive GABA. The kinetic constants were derived from the Scatchard plots using computer analysis based upon a program by McPherson.16"22

RESULTS

GABA receptor expression Table 1 shows the kinetic constants (Kd and Bmax) for G A B A binding to membranes prepared

from 4-day-old cultures of cerebellar granule cells exposed to TH IP (150 I~M) for, respectively, 1, 3, 6 and 24 hr. It can be seen that low affinity G A B A receptors in addition to high affinity recep- tors were observed after exposure of the cells to THIP for 3 hr. After 1 hr exposure low affinity receptors could not be detected but the Ko for the high affinity receptor was somewhat higher

Table 1. Kinetic constants (K,~ and B,,,,x) for GABA binding to membranes prepared from 4-day-old cultures of cerebellar granule cells exposed to THIP

for, respectively, I, 3, 6 and 24 hr

Time of treatment Kd (nM) B ...... (pmol/mg) with 15111J.M THIP

(hr) High Low High Low

1 22.3 +- 5.8 1.00 +- 0.06 3 6.5 +-2.4 537+376 0.81 +-0.02 5.31 +-11.39 6 5 .7+0.9 5[X1+- 410 1.79+-0.09" 5.42_+0.41

24 7 . 6 - + 1.8 502+-265 1.35+-11.18 4.69+-0.27

Cells were cultured for 4 days in plain culture media and subsequently for 1, 3, 6 and 24 hr in the presence of 150 p,M THIP. Values for Ko and B ..... are derived from the binding curves by computer analysis "-22 and represent the statistically significantly best fits to the curves assuming either one or two bind- ing sites. Values are given as means -+ S.E.M. The asterisk indicates that the B .... is statistically significantly higher (P<0.001) than the B ...... at 1 hr exposure.

266 G . H . Hansen e t al.

than the corresponding Ka values calculated for longer exposure times. Membranes from cells cultured in plain culture media for 4 days exhibited, as expected from previous experience, 3JT-t9 only high affinity receptors (results not shown). No further increase in the formation of low affinity GABA receptors was observed after prolonged exposure to THIP for, respectively, 6 and 24 hr but from 1 to 6 hr there was a statistically significant (P<0.001) increase in the number of high affinity GABA receptors (Table 1).

Ultrastructure composition Morphometric analyses of 4-day-old cerebellar granule cell cultures grown in the presence or

absence of THIP (150 ixM) for, respectively, 1, 3, 6 and 24 hr were carried out. Figure 1A shows the relative cytoplasmic density of rough endoplasmic reticulum (RER) in these cells. It is seen that already after exposure to THIP for 1 hr the density of RER was significantly increased com- pared to the untreated cultures. A similar increase in the cytoplasmic density of RER in cere- bellar granule cells grown in the presence of THIP for 3, 6 and 24 hr was observed compared to the control cultures. The cytoplasmic density of smooth endoplasmic reticulum (SER) was, in contrast, significantly decreased in 4-day-old cerebellar granule cell cultures exposed to THIP for only 1 hr compared to the control culture (Fig. 1B). The same effect was observed in cells grown in the presence of THIP for, respectively, 3, 6 and 24 hr (Fig. 1B).

The cytoplasmic density of the Goigi apparatus was significantly increased in 4-day-old cere- bellar granule cell cultures after exposure to THIP for 1 hr compared to the control cultures (Fig. 2). The same trend was observed after exposure of the cells to THIP for, respectively, 3, 6 and 24 hr.

IE _ J

Z m ~ - r

o

o , , y

z

0

f f l

- - i i I i

A

I

B

4

I I I 214 1 3 6

TIME OF EXPOSURE TO THIP ( L O G H O U R S )

Fig. 1. Four-day-old cultures of cerebellar granule cells exposed to 150 IxM THIP for 1,3, 6 or 24 hr (©) and corresponding control cultures (n). Density of rough endoplasmic reticulum (A) and smooth endoplasmic reticulum (B) expressed as percent of total cytoplasmic volume as a function of the culture period. Each point represents the average of countings of 27 different electron micrographs and S.E.M. are indicated by vertical bars. THIP-treated cells differed significantly from control cultures

(A, P<0.0001; B, P<0.0018).

Time course of the neurot rophic activity of G A B A 267

o

N t 9 4

o... v

I I I I 1 3 6 2 4

TIME OF EXPOSURE TO TH[P ( LOG HOURS I

Fig. 2. Four-day-old cultures of cerebellar granule cells exposed to 150 IJ.M THIP for I, 3, 6 or 24 hr (O) and corresponding control cultures (o). Density of the Golgi apparatus expressed as percent of total cytoplasmic volume as a function of the culture period. Each point represents the average of countings of 27 different electron micrographs and S.E.M. are indicated by vertical bars. THIP-treated cells

differed significantly from control cultures (P<0.0005).

12

> lo

g_ ,~o

> 6

~ 4

3

8 ~

I I I i A

I B

[ I i I I 3 6 2 / *

TIME OF EXPOSURE TO THIP ( LOG HOURS)

Fig. 3. Four-day-old cultures of cerebellar granule cells exposed to 150 IzM THIP for 1,3, 6 or 24 hr (O) and corresponding control cultures (a). Density of vesicles (A) and coated vesicles (B) expressed as percent of total cytoplasmic volume as a function of the culture period. Each point represents the average of countings of 27 different electron micrographs and S.E.M. are indicated by vertical bars.

THIP-treated cells in both (A) and (B) differed significantly from control cultures (P<0.0001).

268 G .H. Hanscn et al.

Figure 3A shows the cytoplasmic density of vesicles in 4-day-old cerebellar granule cell cultures grown in the presence and absence of THIP. It is seen that already after exposure to THIP for 1 hr the cytoplasmic density of vesicles is significantly increased compared to the control cultures. The same effect was observed in cells grown in the presence of THIP for, respectively, 3, 6 and 24 hr. Also the cytoplasmic density of coated vesicles (Fig. 3B) was significantly increased in 4-day- old cerebellar granule cell cultures after exposure to THIP for 1 hr compared with the control cul- tures. Again the same trend was observed after exposure to THIP for prolonged periods of time (3, 6 and 24 hr).

DISCUSSION

It has previously been shown that low affinity G A B A receptors are induced in cerebellar granule cell cultures after exposure to G A B A for 8 days ~8't9 or to THIP for 12 days, 3'27'28 and that this effect is blocked by the GABA-receptor antagonist 5 bicuculline. 3 However, the present study demonstrates that this induction of low affinity G A B A receptors can be achieved after exposure of the cells to THIP for only 3 hr. This extremely fast appearance of low affinity G A B A receptors may indicate that THIP either activates pre-existing but functionally inactive low affinity G A B A receptors in the plasma membrane or modifies the high affinity G A B A receptors which are always present on the cells, t7 It has, however, recently been demonstrated that inhibitors of protein synthesis (actinomycin D and cycloheximide) prevent the induction by THIP of low affinity G A B A receptors in 4-day-old cerebellar granule cell cultures. 4 These findings together with the present observation of a parallel increase in numbers of both high and low affinity G A B A receptors strongly indicate that the induction of low affinity G A B A receptors is a result of a d e n o v o synthesis. This in turn, suggests that the turn-over of G A B A receptors is extremely rapid.

The conclusion that the induction requires d e n o v o protein synthesis is in line with the results obtained from the morphometric analyses on corresponding 4-day-old cerebellar granule cell cul- tures grown in the presence and absence of THIP for, respectively, 1, 3, 6 and 24 hr, It was demonstrated that the THIP induced increase in the cytoplasmic density of organelles such as RER, Golgi apparatus, vesicles and coated vesicles followed a slightly accelerated temporal development compared to that seen for the low affinity G A B A receptors. Since RER, Golgi apparatus, vesicles and coated vesicles are known to participate in synthesis and transport of membrane proteins, 25 e.g. low affinity G A B A receptors, it might be anticipated that alterations in these structures would precede the appearance of low affinity G A BA receptors on the cells. The cytoplasmic density of SER was significa, ntly decreased in cerebellar granule cells after exposure to THIP. This may indicate that free ribosomes are attached to the endoplasmic reticulum in order to synthesize membrane proteins.

Although the present study has revealed an extremely rapid onset of the trophic actions of THIP, only little is known about the molecular mechanism by which THIP mediates the changes in ultrastructure composition and the induction of low affinity G A B A receptors. THIP itself might be internalized through receptor mediated endocytosis and mediate its neurotrophic effect intracellularly. It seems, however, more likely that the association between the ligand THIP and the high affinity G A B A receptor triggers a secondary messenger which is responsible for the changes in the ultrastructure composition and the induction of the low affinity G A B A receptor. Another possibility may be that THIP mediates these effects through the opening of ion channels leading to hyperpolarization of the cells. This would be in line with the observation that bromide has a t rophic action on neuroblastoma cells 7 and that the effect of G A B A or THIP on cerebellar granule cells can be blocked by picrotoxin (B. Belhage and A. Schousboe, unpublished results) which specifically acts on G A B A receptor coupled chloride channels. 3~ Studies aimed at elucidat- ing these different mechanisms of action of G A B A agonists are currently in progress in our laboratory.

Acknowledgements- -This work has been supported financially by the following grants: Danish State Natural Science (511-20817) and Medical (12-4492; 12-5878) Research Councils. Dr P. Krogsgaard-Larsen, Department of Chemistry BC, The Royal Danish School of Pharmacy, Copenhagen, Denmark, is cordially thanked for the supply of THIP.

Time course of the neuro t rophic activity of G A B A 269

R E F E R E N C E S

1. Armitage P. (1971) Statistical Methods in Medical Research. Blackwell, London, 504 pp. 2. Bartlett M. S. (1937) Properties of sufficiency and statistical tests. Proc. R. Soc. 160, 268-282. 3. Belhage B., Meier E. and Schousboe A. (1986) GABA-agonists induce the formation of low affinity GABA-receptors

on cultured cerebellar granule cells via preexisting high affinity GABA receptors. Neurochem. Res. I I, 599-606. 4. Belhage B., Hansen G. H., Meier E. and Schousboe A. (1986) Effect of inhibitors of protein synthesis and axonal

transport on THIP-induced development of GABA receptors in cultured cerebellar granule cells. In Molecular Basis o f Neural Fanction (eds Tucek S., Stipek S., Stastny F. and Krivanek J.), p. 207. Eur. Soc. Neurochem., Prague.

5. Curtis D. A., Duggan A. W., Felix O., Johnston G. A. R. and McLennan H. (1971) Antagonism between bicuculline and GABA in the rat brain. Brain Res. 33, 57-73.

6. Curtiis J. H. (1943) On transformations used in the analysis of variance. Ann. Math. Star. 14, 107-122. 7. Eins S., Spoerri P. E. and Heyder E. (1983) GABA or sodium bromide-induced plasticity of neurites of mouse

neuroblastoma cells in culture. A quantitative study. Cell Tiss. Res. 229, 457-461). 8. Enna S. J. and Snyder S. H. (1975) Properties of "y-aminobutyric acid (GABA) receptor binding in rat brain synaptic

membrane fractions. Brain Res. 100, 81-97. 9. Hald A. (1948) Statistiske Metoder. Akademisk Forlag, Copenhagen, 654 pp.

10. Hansen G. H., Meier E. and Schousboe A. (1984) GABA influences the ultrastructure composition of cerebellar granule cells during development in culture. Int. J. devl Neurosci. 2, 247-257.

11. Hansen G. H., Belhage B., Meier E. and Schousboe A. (1986) Developmental aspects of the trophic activity of GABA agonist on cerebellar granule cells in culture. Int. J. devl Neurosci. 4, suppl. 1, $36.

12. Lowry O. H., Rosebrough N. J., Farr A. L. and Randall R. J. (1951) Protein measurement with the folin phenol reagent. J. biol. Chem. 193, 265-275.

13. Madtes P. C. and Redburu D. A. (1983) Synaptic interactions in the GABA system during postnatal development in retina. Brain Res. Bull. 10, 741-745.

14. Madtes P. C. and Redburu D. A. (1983) GABA as atrophic factor during development. Life Sci. 33, 979-984. 15. Madtes P. C. and Bashir-Elahi R. (1986) GABA receptor binding site "induction" in rabbit retina after nipecotic acid

treatment. Changes during development. Neurochem. Res. !1, 55-61. 16. McPberson G. A. (1983) A practical computer-based approach to the analysis of radioligand binding experiments.

Comp. Prog. Biomed. 17, 107-114. 17. Meier E. and Schousboe A. (1982) Difference between GABA receptor binding to membranes from cerebellum

during postnatal development and from cultured cerebellar granule cells. Devl Neurosci. 5, 546-553. 18. Meier E., Drejer J. and Schousboe A. (1983) Trophic actions of GABA on the development of physiologically active

GABA receptors. In CNS-Receptors f rom Molecular Pharmacology to Behavior (eds Mandel P. and DeFeudis F. V.), Adv. Biochem. Psychopharmac. 37, 47-58. Raven Press, New York.

19. Meier E., Drejer J. and Schousboe A. (1984) GABA induces functionally active low affinity GABA-receptors on cul- tured cerebellar granule cells. J. Neurochem. 43, 1737-1744.

20. Meier E., Hansen G. H. and Schousboe A. (1985) The trophic effect of GABA on cerebellar granule cells is mediated by OABA-receptors. Int. J. devl Neurosci. 3, 401-407.

21. Messer A. (1977) The maintance and identification of mouse cerebellar granule cells in monolayer cultures. Brain Res. 130, 1-12.

22. Munson P. J. and Rodbard D. (1980) A versatile computerized approach for characterization of ligand binding systems. Analyt. Biochem. 107, 220-237.

23. Redburn D. A. and Madtes P. C. (1986) Postnatal development of ~H-GABA-accumulating cells in rabbit retina. J. comp. Neurol. 243, 41-57.

24. Reynolds E. A. (1963) The use of lead citrate at high pH as an electron-opaque stain in electron microscopy. J. Cell Biol. 17, 208-212.

25. Sabatini D. D., Kreibich G., Morimoto T. and Adesnik A. (1982) Mechanisms for the incorporation of proteins in membranes and organelles. J. Cell Biol. 92, 1-22.

26. Schousboe A., Larsson O. M. and Krogsgaard-Larsen (1985) Lack of a high affinity uptake system for the GABA agonists THIP and isoguvacine in neurons and asterocytes cultured from mouse brain. Neurochem. Int. 7, 505--508.

27. Schousboe A., Drejer J., Hansen G. H. and Meier E. (1985) Cultured neurons as model systems for biochemical and pharmacological studies on receptors for neurotransmitter amino acids. Devl Neurosci. 7, 252-262.

28. Schousboe A., Hansen G. H., Belhage B. and Meier E. (1986) Role of the inhibitory neurotransmitter GABA as a signal for neuronal growth and differentiation. In Metabolism and Development o f the Nervous System (ed. S. Tucek), in press. John Wiley, New York.

29. Sensenbrenner M., Maderspach K., Latzkovitz L. and Jaros G. G. (1978) Neuronal cells from chick embryo cerebral hemispheres cultivated on polylysine-coated surfaces. Devl Neurosci. 1, 90-101.

30. Spoerri P. E. and Wolff J. R. (1981) Effect of GABA-administration on murine neuroblastoma cells in culture. Cell Tiss. Res. 218, 567-579.

31. Ticku M. K., VanNess P. C., Haycock J. W., Levy W. B. and Olsen R. W. (1978) Dihydropicrotoxinin binding sites in rat brain: Comparison to GABA receptors. Brain Res. 150, 642-647.

32. Wilkin G. P., Balazs R., Wilson J. E., Cohen J. and Dutton G. R. (1976) Preparation of cell bodies from the de- veloping cerebellum: Structural and metabolic integrity of the isolated cells. Brain Res. 115, 181-199.

33. Williams M. A. (1977) Quantitative methods in biology. In Practical Methods in Electron Microscopy, Vol. 6 (ed. Glavert A. M.), pp. 1-84. North Holland/Elsevier, Amsterdam.

34. Wolff J. R., Joo F. and Dames W. (1978) Plasticity in dendrites shown by continuous GABA administration in superior cervical ganglion of adult rat. Nature 274, 72-74.

35. Wolff J. R., Rickmann M. and Chronwall B. M. (1979) Axoglial synapses and GABA-accumulating glial cells in the embryonic neocortex of the rat. Cell Tiss. Res. 201,239-248.

36. Yu A. C. H. and Hertz L. (1982) Uptake of glutamate, GABA, and glutamine into a predominantly glutamatergic nerve cell population in culture. J. Neurosci. Res. 7, 23-35.