GABA Neurotransmission in the Hypothalamus: Developmental Reversal from Ca*+ Elevating to Depressing

The Journal of Neuroscience, July 1995, 75(7): 5065-5077

Karl Obrietan’ and Anthony N. van den POISED ‘Department Biological Science, Stanford University, Stanford, California 94305 and 2Section of Neurosurgery, Yale University, School of Medicine, New Haven Connecticut 06520

GABA is the primary inhibitory transmitter of the adult hy- responses in cultured olfactory bulb, cortex, medulla, stria- pothalamus, synthesized by many neurons and found in turn, thalamus, hippocampus, and colliculus. The majority 50% of the presynaptic boutons. GABA causes a decrease (75%) of developing neurons from each region showed a in Ca*+ in mature hypothalamic neurons in vitro by de- Ca2+ rise in response to GABA. pressing cellular activity through opening Cl- channels. Together these data suggest that GABA elevates Ca*+ in Despite the early expression of GABA, receptors in the em- developing, but not mature, neurons from the hypothala- bryonic hypothalamus (E15), the cellular function of GABA mus and all other brain regions examined. As Ca*+ plays a in the developing hypothalamus has received little atten- crucial role in modulating gene expression, enzymatic tion. In the present study the role of GABA in modulating function, neurite outgrowth, and transmitter release, GABA intracellular Ca*+ in developing hypothalamic neurons was may serve as an excitatory intercellular messenger in- studied with fura- digital imaging. volved in developmental signalling prior to the time when GABA (0.5-500 PM) applied to embryonic hypothalamic its primary function is to inhibit neuronal activity. neurons elicited a dramatic and rapid increase in intracel- [Key words: neuroendocrine, bicuculline, GABA, recep- lular Ca2+ This Ca2+ rise could be completely blocked by for, chloride, glutamate] the GABA, antagonist bicuculline (20 PM) and persisted in the presence of tetrodotoxin (1 PM). The Ca*+ elevation in- During brain development, the amino acid transmitter GABA duced by GABA was greater than that of equimolar con- may modulate neuronal function and growth. GABA increases centrations of the excitatory transmitter glutamate in early neurite outgrowth of brain and retinal neurites (Spoerri, 1988; development. The number of El5 neurons that responded Michler, 1990; Barbin et al., 1993) and modulatessynapse for- to GABA with a Ca*+ rise increased during the first few mation of cultured cells (Meier et al., 1984; Hansenet al., 1987). days of culture, reaching 78% after 4 d in vitro. The Ca2+ Although GABA directly openschloride channels, GABA can rise was 87% blocked by cadmium (100 PM) and 85% influence the levels of intracellular Ca2+by influencing voltage blocked by nimodipine (1 PM), indicating that the mecha- activated Ca2+channels. Ca*+ exerts a wide variety of effects nism of Ca*+ increase was primarily via L-type voltage op- on the developing brain, where it can modulate the rate and erated Ca*+ channels. direction of neuritic growth (Mattson and Kater, 1987) and in- Addition of bicuculline to synaptically coupled cultures fluence gene expression(Vaccarino et al., 1992; Bading et al., caused a significant decrease in Ca2+ 4-10 d after cultur- 1993). ing, indicating hypothalamic neurons were secreting GABA GABA appearsto be the primary inhibitory transmitter in the at an early age of development, and that sufficient GABA adult hypothalamus. Immunocytochemical studieswith the an- was released to elicit an increase in Ca*+. This effect was tisera againstGABA or againstthe GABA synthesizingenzyme seen even after blocking all glutamatergic activity with glu- glutamate decarboxylase (GAD) show strong staining in cell tamate receptor antagonists. In contrast, GABA elicited no bodies and axons throughout the rat hypothalamus(Tappaz et Ca*+ rise in older neurons (>I8 d in vitro), and the action al., 1982; van den Pol, 1985, 1986; van den Pol and Tsujimoto, of bicuculline reversed and caused a large increase in Ca2+ 1985). Nearly 50% of all presynaptic boutons in the hypothal- in spontaneously active neurons. Similar findings were ob- amusare immunoreactivewith GABA, as studiedquantitatively tained in cultures enriched in GABAergic neurons from the with postembeddingimmunogold ultrastructural immunocyto- suprachiasmatic nucleus. To determine if the Caz+ stimu- chemistry (Decavel and van den Pol, 1990); of the hundredsof lating role of GABA on developing neurons was restricted neuronsstudied with electron microscopy, all had a GABAergic to the hypothalamus and a few other regions, or whether input (Decavel and van den Pol, 1990, 1992). Physiologically, it might exist throughout the brain, we examined the Ca2+ application of GABA in the adult hypothalamusleads to neuronal inhibition mediated by opening Cl- channels, and application of GABA, receptor antagonistsgreatly reduce or elimi- Received Nov. 2, 1994; revised Jan. 23, 1995; accepted Feb. 20, 1995. nate IPSPs (Randle et al., 1986; Kim and Dudek, 1990; Nissen We thank Drs. Gong Chen, Paul Trombley, Jeff Kocsis, and Craig Heller for suggestions and helpful discussions. Research and facilities were supported by and Renaud, 1994). NIH NS16296. NSl0174. and the AFOSR to A.N.v. In the courseof studying the development of glutamate me- Correspondence should be addressed to Anthony N. van den Pol, Section of diated Ca*+ excitability in hypothalamic neurons(Obrietan and Neurosurgery, Yale University, School of Medicine, 333 Cedar Street, New Haven, CT 06520. van den Pol, 1995; van den Pol et al., 1995), we found some Copyright 0 1995 Society for Neuroscience 0270.6474/95/155065-13$05.00/O unusualneurons in young cultures that rather than showing the 5066 Obrietan and van den Pol * GABA: Excitatory Transmitter during Development

treated for 30 min. The tissue was then pelleted by centrifugation, pa- GABA AND MUSClMOL INDUCED Ca*+ RISE pain solution removed by aspiration, and tissue mechanically triturated and pelleted three times to generate a single cell suspension. The cell suspension was then plated onto 22 mm square coverslips that had been A. 2 DIV 3 2 4 4 washed in a mild soap solution, rinsed and autoclaved and coated with high molecular weight poly-L-lysine (540,000 Da; Collaborative Re- -d -d T~RODOTO~N search). To ensure high local neuronal density, cells were plated within a 7 200- mm diameter glass ring placed on top of the coverslip; 45 min after z plating the ring was removed. Most of the cells adhered to the coverslip 5+- loo-+#$J surface within 10 min of plating. Cultures were maintained at 37°C and 5% CO, in a Napco 5410 incubator in glutamate- and glutamine-free MEM (GIBCO) supplemented with 10% fetal bovine serum, 100 units/ O\ I I I I I ml penicillin/streptomycin, and 6 gm/liter glucose. 1 To limit the proliferationof non-neuronalcell types,cytosine arabi- 0 2 4 6 6 10 12 __------nofuranoside (1 FM) was added to the tissue culture medium 4 d after plating and maintained until 14 d in vitro (14 DIV). To maintain high B. 6 DIV density neuronal cultures the glutamate receptor antagonists CNQX (10 FM) and AP5 (100 FM) were added to the tissue culture medium at 4 DIV. CNQX and AP5 enhance long term survival of hypothalamic cultures through the inhibition of glutamate mediated excitotoxicity (Choi - - BICUCULLINE I- I- 1987; Furshpan and Potter, 1989; Obrietan and van den Pol, 1995). Medium was changed twice a week. The region of the suprachiasmatic nuclei (SCN) was dissected from El8 brains. Due to the small size of the SCN at El 8, these cultures contained neurons not only from the SCN, but also from the immediate surrounding anterior hypothalamus. These cells were treated to an identical digestion, plating, and feeding protocol as that described for hypothalamic cultures. 0 5 10 15 20 25 Calcium digital imaging. Prior to the beginning of the experiment, MIN cells were incubated in standard HEPES perfusion solution (137 mM NaCl, 25 mM glucose, 5 mM KCl, 1 mM MgCl,, 3 mM CaCl,, pH 7.4) containing 5 FM fura- acetoxymethyl ester (Molecular Probes) for 20 min at 37°C. To inhibit endogenous synaptic activity during fura- load- C. 5 DIV GABA DOSE-RESPONSE SO I I ing, AP5 and CNQX were added to the incubation solution of cells cultured longer than 18 d. The cells were then washed and allowed to recover for 15 min prior to the start of the experiment. Coverslips were then loaded into an eight port 180 pl microscope perfusion chamber (Forscher et al., 1987) and perfused at a constant rate of 1 ml/min. Solutions moved as a straight wave across the chamber allowing for rapid application and removal of receptor agonists and antagonists. Cells were imaged using a 40X Olympus objective with high 340/380 nm transmittance on a Nikon Diaphot 300 inverted microscope. Neu- “’ .05 .5 5 50 500 rons were identified by their responsiveness to the application of NMDA GABA (PM) and by their phase-bright appearance. After 2 DIV neurons had extend- ed long dendritic and axonal processes and were usually found growing Figure 1. A, After 2 d in vitro (2 DIV) embryonic day 18 (E18) on top of an astrocyte monolayer. Ca2+ digital recordings were made hypothalamic neurons were exposed to 20 FM GABA. Ca2+ levels in- from the cell soma. All experiments were performed at room temper- creased and decreased immediately with the repeatedaddition and re- ature. To examine NMDA responses, glycine (2 PM) was added to per- moval of GABA. Even in the presence of 1 FM tetrodotoxin (TTX), fusion solutions that contained 0 Mg2+ (Nowak et al., 1984) to enhance GABA was still capable of increasing Ca 2+ levels. The response of this NMDA receptor activity (Johnson and Ascher, 1987). neuron is characteristic of all neurons assayed. B, A typical Ca*+ rise Data were recorded and all peripheral devices were controlled by a in response to 5 PM muscimol is shown; 6 DIV. C, The Ca*+ responses Universal Imaging 486 computer with FLUOR software. Calibrated Ca*+ of neurons to different concentrations of GABA are plotted. 5 DIV. values from up to 64 cells could be recorded simultaneously. Calibra- tions of Ca*+ were performed as described by Grynkiewicz et al. (1985) with Ca*+ standards and fura- free acid from Molecular Probes. Ex- expecteddecrease in Ca2+, citation light from a 150 W xenon lamp powered by an Optiquip trans- showedan increasewhen GABA was former was filtered through a Sutter filter wheel driven by a Lambda- added. Previous work with nonhypothalamic neuronshad sug- 10 microprocessor. Attenuation of excitation light by 90% was accom- gested that GABA may exert a transient depolarizing influence plished with neutral density filters allowing for long recording periods on developing neurons (Connor et al., 1987; Ben-Ari et al., without any sign of phototoxicity; 16 video frames of data were re- 1989; Yuste and Katz, 1991) suggestingthat if GABA had a corded from both wavelengths every two seconds. CaZ+ data from single cells were transferred to an Apple 840AV computer and analyzed with similar effect on developing hypothalamic neurons,a rise in in- IGOR PRO software (WaveMetrics). Responses to the addition of excit- tracellular Ca2+ might result. The present study examines the atory amino acids are reported as the mean Caz+ rise from basal Ca*+ role of GABA in elevating Ca2+m young hypothalamicneurons, levels + the standard error of the mean. but decreasingCa2+ m older hypothalamic neurons. Cytosine arabinofuranoside, GABA, glycine, and glutamate were purchased from Sigma, AP5, CNQX, bicuculline, and TTX were purchased Materials and Methods from Research Biochemicals International. Papain was purchased from Worthington Biochem. Tissue culture. The hypothalamus was removed from embryonic Sprague-Dawley rats, stripped of meninges, and washed three times in Results standard tissue culture medium, and then incubated in an enzymatic GABA induced rises in intracellular Caz+ solution (10 U/ml papain, 500 FM EDTA, 1500 PM CaCl,, 0.2 mg/ml L-cysteine in Earl’s balanced salt solution). Embryonic day 15 (E15) The addition of 20 PM GABA to El8 hypothalamic neurons tissue was incubated for 20 min in the protease solution, El8 tissue was cultured for 2 d causedan immediate and reproducible rise in The Journal of Neuroscience, July 1995, 75(7) 5067

INHIBITION OF GABA INDUCED Ca*+ RISE

5 DIV 3 DIV 4 DIV aaaaaa itic!Lss 2 aaaa $$92$$ QQUQ s UQ(3(3QU ------cd+&?+ A. BICUCULLINE B. NIMODIPINE C. - - 1607

150

100

50 0-I 1601

150 150 120

100 100 80

50 50 40 ,il

0 4 8 0 5 10 15 20 MIN

Figure 2. Inhibition of GABA induced Ca*+ rises. A, The addition of 20 PM bicuculline to the perfusion solution inhibited GABA induced Ca2+ increases in a reversible manner. 5 DIV. B, 1 FM nimodipine reversibly suppressed the Caz+ rise, indicating the involvement of L-type voltage activated Ca*+ channels. 3 DIV. C, The addition of 100 p,~ Cd2+ also inhibited Ca2+ Increases. Due to its toxic effects if used for long durations, Cd2+ was applied just prior to and after the removal of 20 p,~ GABA. 4 DIV. intracellular Ca2+levels (Fig. 1A). GABA washout quickly re- creasesignificantly with higher GABA concentrationsup to 500 turned the neuronsto resting Ca2+levels. GABA also increased PM (Fig. 1C). Ca2+levels in the presenceof 1 pM tetrodotoxin (Fig. lA), sug- gesting that secondary synaptic activity is not the causeof the Modulation of Ca2+levels by endogenousrelease of GABA Ca2+rise. The Ca2+rise elicited by the addition of GABA could Baseline Ca2+levels in cultured hypothalamic neuronscan be be completely inhibited by the GABA, receptor antagonistbi- reduced by the addition of the glutamate receptor antagonists cuculline (Fig. 2A). GABA (5 pM) causeda meanCa2+ rise of AP5 and CNQX (van den Pol and Trombley, 1993). Figure 3, 89.8 + 8.6 nM (SEM), whereas5 pM GABA in the presence AI-A3, are examplesof this from El8 culturesafter 6 DIV, the of 20 FM bicuculline induced a negligible meanrise (1.3 ? 1.3 removal of APSKNQX from the perfusion solution increased nM) in 42 neurons.This suggeststhat the GABA induced Ca*+ basal Ca2+levels. The addition of bicuculline to the neuron in rise is initiated by the activation of the GABA, and not the Figure 3A3 had virtually no effect on basalCa2+ levels, whereas GABA, receptor. Caz+ rises could also be induced by the ad- the Ca2+levels of the neuronsshown in Figure 3, A2 and A3, dition of the GABA, receptor agonistmuscimol (Fig. IB). Stim- were depressedby the addition of bicuculline. To determine ulation with 5 p,M muscimolincreased Ca2+ in 77% of 128 El8 whether GABA was being releasedby hypothalamic cells, bi- neuronsin vitro for 6 d, with a mean Ca2+rise of 65 & 5 nM. cuculline was addedto neuronsconstantly perfused with AP5/ GABA-induced Ca2+rises could be inhibited by the addition of CNQX to block glutamatereceptor activity. The addition of bi- the L-type Ca2+channel blocker nimodipine (Fig. 2B). Nimo- cuculline reversibly depressedbasal Ca2+ levels (Fig. 3B) in dipine (1 FM) reduced the Ca*+ increaseby an average of 86% 37% of 128 neuronstested. These findings suggestthat in early in 50 neurons.Similarly, the addition of 100 pM Cd2+(Fig. 2C) hypothalamic developmentGABA is actively secretedand that inhibited GABA induced Ca2+rise by 87% in 35 neuronstested. it increasesintracellular Ca2+. Ca2+ rises could be evoked using as little as 500 nM GABA. In the CNS GABA is primarily thought of as an inhibitory The effects of GABA were maximal at 5 p,M and did not in- neurotransmitter.Contrary to this, early in hypothalamic devel- 5068 Obrietan and van den Pol - GABA: Excitatory Transmitter during Development

ENDOGENOUS GABA SECRETION

B. APS/CNQX

1501 Al I 6 DIV 5 DIV 300- 100 200- 50 1 oo-

0 \ I I I I 200

100

00 0-I 1561 A3 100

50 0-I 5 10 15 20

Figure 3. A, Endogenous glutamatergic activity was visible with the removal of the glutamate receptor antagonists AP5 (100 p,M) and CNQX (10 FM) from the 0 Mg*+ perfusion solution. A3 shows a neuron that increases Ca 2+ levels with the removal of glutamate receptor agonist. The elevated Ca2+ level was unaffected by the addition of the GABA, receptor antagonist bicuculline (20 FM). Unlike A3, the addition of bicuculline depressed elevated Ca2+ levels in the neurons Al and A2 (arrows). All cells are cultures of El8 hypothalami. 6 DIV. B, Neurons were perfused in the constant presence of AP5 (100 JLM) and CNQX (10 FM). Basal Ca z+ levels of all three neurons were reversibly depressed by the addition of bicuculline (20 /AM). 5 DIV.

opment GABA may function as a Ca2+elevating neurotrans- vitro, reaching a maximum of 97% at 25 DIV. The changing mitter. We examinedthe developmentaltime point at which GA- role of the GABA, receptor is clearly visible at 10 DIV. At this BA, receptorsswitch rolesfrom Ca*+-elevatingto Ca2+-depress- time point bicuculline increasesand decreasesCa2+ levels in ing. Figure 4 shows neurons from three developmental ages nearly equal numbersof neuronson the same coverslip. Data treated to the identical experimental protocol. At 8 DIV (Fig. collected prior to 10 DIV show the GABA, receptor playing a 4A) the addition of 20 pM bicuculline depressedCa2+ levels and predominately Ca2+elevating role. After 10 DIV activation of Ca2+fluctuations. At 13 DIV (Fig. 4B), the role of the GABA, the GABA, receptor depressedCa2+. receptor had changed. The application of bicuculline dramatically increasedCa 2+ in all three neurons. Similar results were Transition of GABA from excitatory to inhibitory activity seenin neuronscultured for 33 d (Fig. 4C). Thesedata indicate The excitatory effects of GABA diminish by 13 DIV (Fig. 6). that during the period between 8 and 13 DIV the GABA, re- In all three neurons,excitatory synaptic activity initiated by the sponsechanged from Caz+-elevatingto Ca*+-depressing. removal of APS/CNQX from the perfusion solution was inhib- Figure 5 showsthe effects of bicuculline on unblocked,basal ited by the addition of GABA. Unlike Figure 1 where in the Ca*+ levels. From 2 DIV through 13 DIV, bicuculline depressed presence of TTX, GABA induced a Ca2+ rise in the majority of Ca*+ levels. The percentageof neurons showing reductions in 2 DIV neurons(71% of 64 neurons)GABA induceda Ca2+rise Ca*+ with bicuculline increaseddramatically from 2 DIV (2%) in only 17% of 110 neurons in vitro for 13 d. to 4 DIV (2.5%) reaching a maximum at 8 DIV (nearly 40%). GABA’s transition from Ca2+-elevating to Ca2+-depressing After 8 DIV the percentageof neuronsdepressed by bicuculline neurotransmittertemporally correspondswith the development dropped dramatically. By 13 DIV only 4% of the neuronsre- of complex glutamatergicneurotransmission. Of interest wasthe sponded,and by 2.5 d the effect wascompletely lost. In contrast, finding that bicuculline potentiated endogenousglutamatergic bicuculline’sability to increaseneuronal Ca*+ levels startedvery activity primarily in neurons displaying synchronized Ca*+ low (2% at 8 DIV) and increasedsteadily throughout time in spikes.Neurons (10 DIV) in Figure 7, A and B, exhibited syn- The Journal of Neuroscience, July 1995, 15(7) 5069

DEVELOPMENTAL REVERSAL OF Ca 2+ MODULATION WITH BICUCULLINE

A. 8 DIV B. 13 DIV C. 33 DIV

800 1 Cl 600-

400-

200-

O-l

O-1

6oo B3 1 I

MIN Figure 4. The effects of bicuculline on Ca*+ levels reversed for El8 hypothalamic neurons between 8 and 13 DIV. A, 8 DIV neurons were removed from glutamate receptor blockade [AP5 (100 FM) and CNQX (10 FM)] in a 0 Mg2+ perfusion solution. The application of 20 FM bicuculline to unblocked neurons depressed Ca*+ levels. CaZ+ levels in Al were not affected by the removal of glutamate receptor blockers. B, 13 DIV neurons were treated to the identical protocol used for 8 DIV neurons. Of interest is the reversed effect of bicuculline on Ca2+levels; bicuculline reproducibly increased Ca2+levels in all three neurons. The effects of bicuculline on 33 DIV neurons (C) were identical to those found for 13 DIV neurons. chronized Ca*+ spikes with the removal of APSKNQX from the with 20 pM GABA, 20 pM glutamate, or 20 FM glycine. Figure perfusion solution. Endogenous activity was potentiated (larger 8 shows examples of characteristic cells from three different spikes and increased Ca*+ level) by the application of bicucul- durations in vitro. In Figure 8A neurons were plated 4 hr prior line. The Caz+ level in neuron C was elevated by the removal to imaging. The neuron in Figure 8A1 had nearly equivalent, of APSKNQX but, unlike neurons A and B, the Ca2+ spikes reproducible Ca 2+ increases to the application of GABA and were not synchronized. When bicuculline was applied to neuron glutamate and a slightly less robust response to glycine. In Fig- C, the basal Ca*+ level was depressed. The neuron in Figure 70 ure 8A2 glutamate induced a much larger Ca*+ rise than GABA, was unaffected by the removal of APYCNQX but the addition whereas in Figure 8A3, the converse was true. A similar hetero- of bicuculline reduced basal CaZ+ levels. Of 27 neurons (studied geneity of reproducible responses was seen in Figure 8B. Neu- simultaneously with digital imaging) with complex, glutamate- rons from 12 DIV (Fig. SC) had a notable absence of strong mediated synchronized activity, 8 1% had Ca*+ levels potentiated GABA induced Ca*+ increases. All cells that showed a Ca2+ rise and 19% had Ca2+ levels depressed by bicuculline. Of 35 neu- in response to glycine application also responded to GABA. A rons not exhibiting glutamate-mediated synchronized activity, subpopulation of neurons were found that responded only to 14% showed potentiation and 86% showed depression of Ca2+ GABA and not to glycine. levels. Several interesting trends appeared when peak Ca2+ responses from each agonist were combined into mean responses for each Comparison of GABA, glycine, and glutamate induced Ca2+ time point (Fig. 9A). As was the case with bicuculline’s ability responses during development to depress basal CaZ+ levels [percentage of neurons increased To compare the Ca*+ responses to GABA with other amino acids over the first 8 DIV, then decreased slowly (Fig. 5)] GABA’s at different developmental ages, El5 neurons were stimulated ability to increase Ca*+ levels started at low amplitudes (12.1 ? 5070 Obrietan and van den Pole GABA: Excitatory Transmitter during Development

MODULATION OF Ca * + BY BICUCULLINE was the finding that the mean Ca2+response to the combined addition of agonists(79 ? 3.6 nM) was intermediate between 100 the responsesto GABA and glutamate (N = 119). Of the 119 n BIC DECREASE neuronstested, 91% had a greater responseto either GABA or BIC INCREASE glutamatethan to the responseelicited by the combinedaddition of the agonists.The responseof the remaining 9% was slightly greater than the largest responseto either glutamate or GABA, but in only 4% of the neuronswas the combinedCa2+ rise sta- tistically greater (>2 X the SEM). GABA in the suprachiasmaticnucleus Whereasthe hypothalamusin generalhas a high density of GA- BAergic neurons,the suprachiasmaticnucleus (SCN), the circadian clock of the mammalianbrain, is composedalmost en- tirely of GABA immunoreactive cells (Card and Moore, 1984; van den Pol and Tsujimoto, 1985; Okamura et al., 1989). The

2 4 8 10 13 25 region of the SCN was cultured on El8 and treated to similar protocols as hypothalamic cultures. Histological analysis con- DAY IN VITRO firmed that the SCN and the immediately surroundinghypotha- Figure 5. The ability of bicuculline (20 p,M) to either decrease or lamic area were included in these cultures. Figure 1OA shows increase Ca*+ levels of El8 hypothalamic neurons was plotted from 2 to 2.5 DIV. From 4 to 8 DIV the predominant effect of bicuculline was three neurons from these cultures that respondedto the appli- to decrease Ca2+ levels. By 10 DIV, Ca 2+ levels in roughly equal num- cation of GABA with a rise in basal Ca2+levels, both in the bers of neurons were either increased or decreased by bicuculline. At absenceand presenceof TTX. Figure 10B shows4 DIV neurons 13 DIV the effect was dominated by potentiation and by 25 DIV the responding to multiple applications of GABA, glutamate and sole response to bicuculline was an increase in Caz+. glycine. Both the meanresponse and the number of responsive neurons were compared at 4 DIV and 18 DIV; 128 neurons were tested at each time point. Over time in culture, an increased 1.5 nM at 0 DIV), gradually increasedto a meanrise of 77.5 + meanCa2+ rise wasfound with the applicationsof glutamate(69 4.5 nM by 4 DIV, and subsequentlydiminished. In contrast, glu- -t- 4nMat4DIVand283 + 8nMat18DIV)andK+55mM tamate-inducedresponses start out low (13.9 2 2.4 nM at 0 (180 ? 6 nM at 4 DIV and 283 & 8 nM at 18 DIV). In contrast DIV) and continue to rise throughout the time tested. GABA a decreasedmean response was found with GABA (39 ? 3 nM induced a larger rise after 1 and 2 DIV than did glutamate. At at 4 DIV and 13 k 2 nM at 18 DIV) and glycine (14 k 3 nM 4 DIV meanGABA responselevels peaked,whereas glutamate at 4 DIV and 5 + 1 nM at 18 DIV). Similarly, the total number responsevalues continued to rise. At 16 DIV the responsegen- of neuronswith a Ca*+ rise greater than 30 nM changedwith eratedby the application of glutamatewas over 14 times greater time in culture. At 4 DIV 46% of the cells respondedto GABA than that induced by GABA. Similar to glutamate, 55 mM K+ with a Ca2+rise, 81% respondedto glutamate, 4% to glycine, induced Ca2+rises increaseddevelopmentally. Glycine was the and 96% to K+. By 18 DIV 13% of the cells respondedto least effective agonist, only slightly increasingCa*+ levels dur- GABA with a Ca2+rise, 100% respondedto glutamate, 1% to ing early development. glycine, and 100% to K+. Inhibition of the GABA, receptor by The percentageof neuronsthat respondedto each agonist is bicuculline after 6 DIV (in the constantpresence of APS/CNQX) depicted in Figure 9B. A similar pattern to that seenin Figure depressedbasal Ca*+ levels in 50% of the 128 neuronstested. 9A developed. The percentageof cells responsiveto the appli- This 50% was the largest number of neuronsthat respondedto cation of GABA rose steadily from 18% at 0 DIV up to 78% bicuculline with a Ca2+depression at any time during develop- by 4 DIV and thereafter fell. The percentageof glutamate re- ment, either in whole hypothalamusor SCN-enrichedcultures. sponsivecells increasedthroughout the time in culture, reaching This indicatedthat in theseSCN enriched cultures with blocked a plateauat 98% at 12 DIV. The percent of glutamateresponsive glutamate receptors, GABA secretedfrom cells in the culture cells lagged behind GABA responsive cells (the number of elevated cytoplasmic Ca2+of many neurons. GABA responderswas nearly twice that of glutamateresponders at 1 DIV) up to 3 DIV. Cells that responded to 55 mM K+ Other brain regions increasedthrough time in culture, reaching 100% by 12 DIV. To determineif GABA (30 PM) would induce Ca2+rises only Glycine responsive cells followed the same general trend as in a few regions of the brain, or whether its effects were more GABA responsivecells; a slow rise to a peak at 4 DIV and universal on developing neurons,we comparedthe responsesof declining thereafter. 3 DIV hypothalamic culturesto culturesfrom sevenother brain To study the combined effect of GABA and glutamate in regionsfrom El8 rats. The majority (75%) of neuronsfrom each young hypothalamic neurons, El8 4 DIV cultures were stimu- region showedGABA-elicited Ca2+rises. This was basedon a lated with 5 pM GABA, 5 PM glutamate, or a combination of total 508 neurons,with 64 from each region except hypothala- 5 p,M GABA and 5 FM glutamate.In someneurons the response mus which had 60 neurons.The percent of neuronsshowing a to glutamatewas greater, and in others the responseto GABA rise of at least30 nM in each areawas as follows: hypothalamus was greater. No neuronswere found that showeda decreasein (61%), cortex (89%), hippocampus(83%), thalamus(64%), me- Ca2+in responseto either amino acid. The mean Ca2+rise in dulla (76%), colliculus (83%), striatum (72%), and olfactory responseto the application of GABA was 55 + 3.2 nM. The bulb (61%). The mean Ca*+ rise (mean + SEM [nM]) of all meanCa2+ response to glutamatewas 101 2 5.3 nM. Of interest cells studiedin each area was as follows: hypothalamus(66 & The Journal of Neuroscience, July 1995, f5(7) 5071

GABA INHIBITION OF GLUTAMATE MEDIATED Ca*+ INCREASES

13 DIV

Figure 6. Spontaneousactivity in El8 neurons was completely sup- pressedby the addition of 20 pM GABA. Bv 13DIV. 20 LM GABA was unableto induceCa2+ rises in the ma- 1’5 jority of neurons.1 p.M TTX wasused to suppressendogenous synaptic activ- MIN ity.

9) cortex (151 + 13), hippocampus(93 2 lo), thalamus (51 blocked by bicuculline, indicating the specific action of the GA- ? 7), medulla (80 2 S), colliculus (79 & S), striatum (69 ? BA, receptor; the role of the GABA, receptor was further sup- S), and olfactory bulb (41 f 5). When neurons(n = 739) from ported by the activity of muscimol, a GABA, receptor agonist the hippocampus,cortex, hypothalamus, spinal cord, striatum, that also produceda Ca*+ rise in young neurons. and olfactory bulb were studied after 3 weeks in culture, the The blockade of the Ca*+ rise by cadmium indicatesthat the GABA antagonistbicuculline (20 p,M) elicited either no change Ca2+entered the cell from the extracellular milieu. The blockade or an increasein intracellular Ca2+.Less than 1% of the 3 week of most of the Ca2+rise by nimodipine indicated the activation old cells showed a bicuculline-induced Ca*+ decrease.These of L-type voltage-gated Ca*+channels to be the primary mech- data indicate that GABA had reversed its role and become a anism of CaZ+entry into the cell. Together these data suggest Ca2+-depressingtransmitter in thesemore mature neurons,con- that GABA opensCll channelsthat depolarize the neuron; the sistent with our previously describedexperiments focusing on depolarization activates voltage gated Ca*+ channelsthat allow hypothalamic neurons. the entry of extracellular Ca2+.This phenomenonhas apparently not been studied previously in hypothalamic neurons,but work Discussion in other parts of the nervous systemincluding the retina, hip- Mechanism of GABA-induced Ca2+rise pocampusand spinal cord suggesta parallel function for GABA In young hypothalamic neuronsGABA increasedCaZ+ even dur- in early development.In studiesof the developing avian retina, ing blockade of action potentials with TTX. This indicates that spinal cord, rat neocortex, and neonatal pituitary cells in vitro, the Ca*+ rise was not due to secondaryrelease of some other activation of the GABA, receptor has been reported to elicit a transmitter initiated by GABA. The effect was completely depolarizing current that in turn activates voltage dependent 5072 Obrietan and van den Pol * GABA: Excitatory Transmitter during Development

DIFFERENTIAL EFFECTS OF BICUCULLINE ON NEURONAL Ca*+

INCREASE

0 I I I I I I DECREASE

Figure 7. The reversal of the response to bicuculline was most strik- ingly seen at 10 DIV. Bicuculline reproducibly potentiates Ca2+ levels in A and B, whereas Ca*+ levels in C and D were suppressed. Complex synchronized CaZ+ transients are seen in neurons A and B in an unblocked state, both in the presence and absence of bicuculline. To verify that the cells were neurons, 100 FM NMDA was used to - 0I 5 1’0 1’5 2’0 2’5 3’0 stimulate a Ca*+ rise (not shown). MIN

Ca*+channels (Yuste and Katz, 1991; Horvath et al., 1993; Ya- study showedan immediate(2-4 set) Ca2+rise in responseto mash&aand Fukuda, 1993). Similarly, in a cell line developed GABA, and a rapid (seconds)recovery upon GABA washout. from LHRH secreting neurons,GABA elicited a CaZ+rise, and A depolarizing responseto GABA can be found in restricted dependingon the chloride equilibrium potential, induced action regions of adult hippocampalneurons. GABA applied to den- potentials (Hales et al., 1994). Application of GABA in early drites causeda depolarization;in contrast, applicationto the per- but not late embryogenesiselicited action potentials in the de- ikaryon elicited a hyperpolarizing response(Anderson et al., veloping chick spinal cord (Obata et al., 1978). Embryonic hip- 1980; Alger and Nicoll, 1982). Axons of developing retinal gan- pocampal cells showed depolarizing responsesto nanomolar glion cells were depolarized by GABA agonists;this sensitivity concentrationsof GABA agonists(Fiszman et al., 1990). Hip- was lost or reduced in the adult optic nerve (Sakatani et al., pocampalcells showedgiant depolarizing potentials during the 1992; Lim et al., 1993). An extreme example of depolarization first week of postnatal developmentthat appearedto be due to through the opening of Cl- channelsis found outsidethe brain, GABA release;these disappeared by P12 (Ben-Ari et al., 1989). and even outsidethe animal kingdom. Algae cells that normally Additional mechanismsof GABA-induced Ca2+rises may exist have a negative membranepotential showa slow action potential in other cells. For instance, GABA stimulation of cerebellar that, rather than basedon Na+ influx, is basedon efflux of high granule cells evoked a Ca2+rise that took 2 min to develop, and levels of internal Cl- (Gaffey and Mullins, 1958; Kishimoto, did not recover even 15 min after washout of the GABA (Con- 1965). nor et al, 1987).In contrast,hypothalamic neuronsin the present In parallel experiments, glycine also evoked a modestCa*+ The Journal of Neuroscience, July 1995, 15(7) 5073

RESPONSE TO 20 pM GABA, GLUTAMATE AND GLYCINE

A. 0 DIV B. 4 DIV C. 12 DIV

200- B2 150- ,

loo-

0 5 10 15 20 0 5 10 15 20

MIN Figure 8. Representative cells plated at El5 are shown at three time points (0, 4, 12 DIV) during development. Cells were treated to the repeated application of 20 FM GABA, 20 FM glutamate, and 20 FM glycine. Of note is the heterogeneous, yet reproducible response to the application of the neurotransmitter receptor agonists. Perfusion solutions contained 1 FM TTX. rise in our developing hypothalamic neurons,probably due to recording show GABA responses(Randle et al., 1986; van den opening Cll channelsby strychnine-sensitiveglycine receptors, Pol et al., 1995). Thesedata suggestthat all hypothalamic neu- with subsequentactivation of voltage regulated Ca2+channels. rons probably have GABA receptors,further supportingthe im- The ability of glycine to evoke a Ca2+rise was lost in older portance of early GABA neurotransmission. hypothalamic neurons.Strychnine-sensitive glycine receptor activation evoked depolarizations in hippocampal neurons from Developmentalchanges in GABA-induced Caz+ response rats lessthan 4 d old (Ito andCherubini, 1991)and in developing In early hypothalamic development,GABA increasesCaZ+, and spinal cord motoneurons(Wu et al., 1992). in more mature neurons,GABA either has no effect on Ca2+or In the spinalcord, all cells that showeda Ca2+rise in response decreasesit. What is the substrateunderlying this change?One to GABA also showed one to glycine, and the magnitude of possibility is that the membranepotential changeswith devel- their rise washigh correlated(Reichling et al., 1994). In contrast, opment. Changesin membranepotential would lead to differ- not all hypothalamic cells that responded to GABA also re- encesin Cl- movement. spondedto glycine, althoughall cells that did respondto glycine It is unlikely the developmental changesin the GABA-in- also respondedto GABA. In cells that did respondto both, in duced Ca*+ responseare due to changesin extracellular Cl- some the responsewas larger to GABA, and in others the re- becausein our experimentsthe external Cl- levels were held at sponsewas larger to glycine. This difference between the hy- a constantlevel. Changesin intracellular Cll, on the other hand, pothalamusand spinal cord Ca*+ responsemay be due to the may play an important role in the developmental alteration in absence of glycine receptors from some hypothalamic neurons GABA activity by changing the equilibrium potential of Cl-, (van den Pol and Gores, 1988). Other neuronsthat do express thereby influencing the voltage changeinduced by GABA elic- the glycine receptorin the hypothalamusmay have more glycine ited opening of Cl- channels.One cellular mechanismfor this than GABA receptors. All cells we have examined with ultra- may be through developmentalchanges in the operation of Cll structural immunocytochemistryin the adult hypothalamuswere transporters.Different cells may have different Cll transporter postsynaptic to GABA immunoreactive boutons (Decavel and mechanisms.For instance,hippocampal granule cells can be de- van den Pol, 1990) and all neurons tested with patch clamp polarized by GABA, while at the same stageof development, 5074 Obrietan and van den Pole GABA: Excitatory Transmitter during Development

A. Ca*+ INCREASE WITH 20 PM GLYCINE, GABA, Opening the Cl- channel could result in HCO,- passingout of GLUTAMATE the cell, depolarizing the cell. An electrophysiological study in 200 1 the hippocampusargued that HCO,- appearedunlikely to play a significant role in depolarizing GABA responsesin hippocam- s pal pyramidal cells (Grover et al., 1993). 5 0 GLUTAMATE The effect of GABA in reducing intracellular Ca*+ is partic- N= 1455 ularly dramatic in older neuronsthat show an elevated Ca2+due ii to glutamateexcitation. GABA acting as an inhibitory transmit- c + loo- ter reducesthe glutamate-mediatedactivity either by hyperpo- N larization or by a slight depolarization and current shunt,result- 8 ing in a large decreasein Ca2+.Parallel decreasesin Ca2+can z be producedby blocking action potentialswith TTX, or blocking 4 glutamate receptorswith APYCNQX. I Changesin the molecularexpression of GABA, receptor sub- i units occur during development(Fritschy et al., 1994; Mathews o- *. et al., 1994). Physiologically, GABA receptorsshow lessdesen- 012348 12 18 sitization in young neuronsthan in adult cells (Oh and Dichter, DAY IN VITRO 1992). A changein the density of GABA receptors(Hansen et al., 1987, 1991) or in the subunit configuration (Fritschy et al., 1994; Mathews et al., 1994) could alter the Cl- movement through the channel. B. % CELLS WITH Ca *+ RISE GREATER THAN 30 nM

100 Developmental role of GABA-induced Ca2+ rises 1 q GLYCINE In early hypothalamic development GABA appearsto play a q GABA 80 greater role than the excitatory transmitter glutamatein evoking tl GLUTAMATE Ca2+rises. This is consistent with experiments near the beginning of hypothalamic neurogenesis(E1.5) when all cells tested 80 with patch-clamp recording showeda large responseto GABA, 8 but only someshowed a responseto glutamate (van den Pal et 40 al., 1995). Studies in the spinal cord showedthat GABA receptors are expressedin early embryogenesisand may be found even earlier than glutamate receptors (Walton et al., 1993). GABA and its synthetic enzyme are found in E6 neuronsin the 1 chick retina (De Mello et al., 1976; Frederick, 1987); this is prior 0 1 2 3 8 1: 18 to significant synaptogenesisdescribed at El4 (Sheffield and DAY IN VITROI!./ I Fischman, 1970). GABA may also be releasedduring development, perhapseven before synaptogenesis(Gordon-Weeks et al., Figure 9. A, The mean Ca?+ rise to the application of 20 PM glycine, glutamate and GABA is plotted over time in culture (O-16 DIV). The 1984; Balcar et al., 1986; Hicks et al., 1986); GABA release mean response to GABA peaks at 4 DIV, whereas the response to glu- could be induced from a cell fraction containing axon endings tamate continues to rise. Bars represent SEM. B, At each time point the prior to the time when the synaptic proteins P65 and synapto- total number of cells that responded with a minimum rise of 30 nM to physin could be detected immunocytochemically (Taylor et al., the application of agonist is plotted. A greater percentage of cells responded to the addition of GABA than glutamate over the first 3 d in 1990). Spontaneoushyperpolarizing potentials associatedwith culture. All data were from the same initial plating of cells. GABA transmissionhave beenfound developmentally later than depolarizing potentials attributed to glutamate in the retina (Ro- rig and Grantyn, 1993). NMDA receptor currents were found in pyramidal cells are hyperpolarized; this differential effect has the somatosensorycortex before inhibitory currents (Agmon and been attributed to Cl- transportersof different polarities and is O’Dowd, 1992). It is possiblethat someof the depolarizing po- dependent on the different reversal and resting membranepo- tentials throughout the brain in early stages of development tentials (RMP) of the two cell types (Misgeld et al., 1986). Re- could be elicited by spontaneousGABA release. duced outward transport of Cl- was reported in younger hip- A number of studieshave shown that intracellular Ca*+ may pocampalneurons (Luhmann and Prince, 1991), indicating that play important roles in neurite outgrowth in vitro (Mattson and the full activity of outward Cl- transport may develop late. The Kater, 1987). GABA can stimulate process outgrowth; hippo- reversal potential for IPSPschanges during developmentrelative campal neuronsraised in the presenceof GABA receptor antag- to the RMP For instance in hippocampalCA1 neurons,the Cl- onists had fewer neuritesand a smallertotal neuritic length than dependent IPSP reversal potential in neurons2-5 d after birth controls (Barbin et al., 1993). GABA also influenced neuritic was near the RMP, but by 15-20 d after birth the reversal po- branching and synaptogenesisin avian retina and brain cultures tential was 25 mV more negative than the RMP (Zhang et al., (Spoerri, 1988; Michler, 1990). GABA agonistshave been re- 1991). ported to alter the expressionof GABA receptorsand to influ- Intracellular changes in Cl- appear the most likely ionic ence the ultrastructure of developing cerebellar granule cells mechanismto explain the shift from Ca2+elevating to depress- (Hansenet al., 1987, 1988, 1991; Belhage et al., 1988). ing. However, bicarbonatemay also passthrough Cl- channels, Another role for GABA, in collaboration with glutamate, may and may play a role in GABA responses(Kaila et al., 1993). be to regulate and stabilize the Ca2+setpoint of developing neu- The Journal of Neuroscience, July 1995, 15(7) 5075

SUPRACHIASMATIC NUCLEUS-GABA INCREASES Ca2+

4 DIV -lTX 6 DIV --- P 24 P 2501 200

150 80 100

0 0-I 2501 160

120 120

80 80

40 i 0 o-

120

80- 80

0-I 0 I I I I o- 0 4 8 12 MIN 0 5 10 15 20 0 5 10 15 20

Figure 10. Characteristic responses of neurons from the suprachiasmatic region of the hypothalamus are shown. A, Both in the absence and presence of TTX (1 PM), Ca’+ levels increased in response to 20 FM GABA. 4 DIV. B, Equimolar concentrations of GAB A, glutamate, and glycine (20 FM) evoked reproducible increases in Ca*+. 4 DIV. C, The addition of 20 PM bicuculline depressed basal Ca 2+ levels in neurons constantly perfused with AP.5 (100 FM) and CNQX (10 FM) after 6 DIV, indicating that GABA was being released by cells in these cultures. rons. In neurons in vitro for 4 d, examined in the present study, cellular events via the increase of intracellular Ca2+. Experi- the addition of either GABA or glutamate evoked an increase mentswith 2-deoxyglucosesuggest that cells in the SCN in vivo in Ca*+. However, the combined addition of GABA and gluta- (Reppert and Schwartz, 1984) or in vitro (Shibata and Moore, mate resulted in a Ca*+ rise of an intermediate value between 1988) show an orchestratedrhythm of uptake in early stagesof the two, regardless of which amino acid initially produced the development.A potential explanation for this, basedon the data greater Ca*+rise. In neurons in which GABA induced a greater in the presentstudy, may be that GABA releasedby young SCN Ca*+ rise than glutamate did, the addition of glutamate to GABA cells could serve to orchestratethe juvenile neuronsin the SCN causeda decreasein Ca2+, suggestinga paradoxical reversed that are involved in circadian timekeeping. Ca2+inhibitory role for glutamatein somedeveloping hypotha- The studies describedhere focus on hypothalamic neurons. lamic neurons.The intermediate,rather than additive rise in Ca*+ Our experimentssurveying seven other regionsof the brain in- in the presenceof both Ca2+-stimulatingGABA and glutamate dicated that GABA elicited a CaZ+rise in the majority of all suggeststhat the combinedeffect of both would not lead to Ca*+ developing neuronstested. Together with work from other labs excitotoxicity, but rather would tend to stabilize with a more describedabove, our data support the probability that the early modest Ca2+rise. Many cellular processesfunction optimally role of GABA as an excitatory transmitter is not restricted to a within a narrow rangeof cytoplasmic CaZ+concentrations. Neu- few brain regions,but rather is widespreadthroughout the CNS. rons may die both as a result of Ca*+ levels that are too high, for instancecaused by glutamateexcitotoxicity (Olney and Shar- References pe, 1969; Choi, 1988), but neuronsmay die also if Ca*+ levels Agmon A, O’Dowd D (1992) NMDA receptor-mediated currents are are too low (Franklin and Johnson, 1992). prominent in the thalamocortical synaptic response before maturation of inhibition. J Neurophysiol 68:345-349. In older neurons from the suprachiasmaticnucleus, GABA Alger B, Nicoll R (1982) Pharmacological evidence for two kinds of can inhibit ultradian Ca*+oscillations (van den Pol et al., 1992). GABA receptor on rat hippocampal pyramidal cells studied in vitro. GABA releasefrom immatureneurons may serve to synchronize J Physiol (Lond) 328:125-141. 5076 Obrietan and van den Pol * GABA: Excitatory Transmitter during Development

Anderson P, Dingledine R, Gjerstad L, Langmoen I, Mosfeldt Laursen Hansen G, Belhage B, Schousboe A, Meier E (1988) Gamma-amino- A (1980) Two different responses of hippocampal pyramidal cells butyric acid agonist-induced alterations in the ultrastructure of cul- to applications of gamma-aminobdtyric acid. J Physiol (Lond) 305: tured cerebellar granule cells is restricted to early development. J 279-296. Neurochem 5 1:243-245. Bading H, Ginty D, Greenberg M (1993) Regulation of gene expres- Hansen GH, Belhage B, Schousboe A (1991) Effect of a GABA ag- sion in hippocampal neurons by distinct calcium signaling pathways. onist on the expression and distribution of GABA, receptors in the Science 260:181-186. plasma membrane of cultured cerebellar granule cells: an immuno- Balcar V, Joo F, Kasa P, Dammasch I, Wolff J (1986) GABA receptor cytochemical study. Neurosci Lett 124: 162-l 65. binding in rat cerebral cortex and superior cervical ganglion in the Hicks T, Ruwe W, Veale W (1986) Release of gamma-aminobutyric absence of GABAergic synapses. Neurosci Lett 66:269-274. acid from the visual cortex of young kittens. Dev Brain Res 24:299- Barbin G, Pollard H, Gaiarsa J, Ben-Ari Y (1993) Involvement of 304. GABA, receptors in the outgrowth of cultured hippocampal neurons. Horvath G, Acs Z, Mergl Z, Nagy I, Makara G (1993) Gamma-ami- Neurosci Lett 152:150-152. nobutyric acid-induced elevations of intracellular calcium concentra- Belhage B, Hansen G, Schousboe A, Meier E (1988) GABA agonist tions in pituitary cells of neonatal rats. Neuroendocrinology 57: 1028- promoted formation of low affinity GABA receptors on cerebellar 1034. granule cells is restricted to early development. Int J Dev Neurosci Ito S, Cherubini E (1991) Strychnine-sensitive glycine responses of 6:125-128. neonatal rat hippocampal neurones. J Physiol (Lond) 440:67-83. Ben-Ari Y, Cherubini E, Corradetti R, Gaiarsa J (1989) Giant synaptic Johnson J, Ascher P (1987) Glycine potentiates the NMDA response potentials in immature rat CA3 hippocampal neurones. J Physiol in cultured mouse brain neurons. Nature 325:529-53 1. (Lond) 416:303-325. Kaila K, Voipoi J, Paalasmaa P, Pasternak M, Deisz RA (1993) The Card J, Moore R (1984) The suprachiasmatic nucleus of the golden role of bicarbonate in GABA, receptor-mediated IPSPs of rat neo- hamster: immunohistochemical analysis of cell and fiber distribution. cortical neurones. J Physiol (Lond) 464:273-289. Neuroscience 13:415-43 1. Kim YI, Dudek FE (1992) Intracellular electrophysiological study of Cherubini E, Gaiarsa J, Ben-Ari Y (1991) GABA:an excitatory trans- suprachiasmatic nucleus neurones in rodents: inhibitory synaptic mitter in early postnatal life. Trends Neurosci 14515-519. mechanisms. J Physiol (Lond) 458:247-260. Choi DW (1988) Glutamate neurotoxicity and diseases of the nervous Kishimoto U (1965) Voltage clamp and internal perfusion studies on system. Neuron 1:623-634. Nitella internodes. J Cell Comp Physiol 66:43-54. Choi DW, Maulucci-Gedde M, Kriegstein AR (1987) Glutamate neu- Lim JY, Utzschneider DA, Sakatani K, Kocsis JD (1993) The attenu- rotoxicity in cortical cell culture. J Neurosci 7:357-368. ation of GABA sensitivity in the maturing myelin-deficient rat optic Connor J, Tseng H, Hockberger P (1987) Depolarization-and trans- nerve. Dev Brain Res 72:15-20. mitter-induced changes in intracellular Ca2+ of rat cerebellar granule Luhmann H, Prince D (1991) Postnatal maturation of the GABAergic cells in explant cultures. J Neurosci 7: 1384-1400. system in rat neocortex. J Neurophysiol 65:247-263. Decavel C, van den Pol AN (1990) GABA: a dominant transmitter in Mathews G. Bolos-Sv A. Holland K. Isenberg K. Covev D. Ferrendelli the hypothalamus. J Comp Neurol 302:1019-1037. J, Rothman S (1954) ‘Developmental alte&tidn in GABA, receptor Decavel C, van den Pol AN (1992) Converging GABA-and glutamate- structure and physiological properties in cultured cerebellar granule immunoreactive axons make synaptic contact with identified hypo- neurons. Neuron- 13:14%158. - thalamic neurosecretory neurons. J Comp Neural 3 16: 104-l 16. Mattson MI? Kater SB (1987) Calcium regulation of neurite elongation De Mello F, Bachrach U, Nirenberg M (1976) Ornithine and glutamic and growih cone mohlity. j Neurosci 72034-4043. acid decarboxylase activities in the developing chick. J Neurochem Meier E, Drejer J, Schousboe A (1984) GABA induces functionally 27~847-851. active low-affinity GABA receptors on cultured cerebellar granule Fiszman M, Novotny E, Lange G, Barker J (1990) Embryonic and cells. J Neurochem 43:1737-1744. early postnatal hippocampal cells respond to nanomolar concentra- Michler A (1990) Involvement of GABA receptors in the regulation tions of muscimol. Dev Brain Res 53:186-193. of neurite growth in cultured embryonic chick tectum. Int J Dev Forscher P, Kaczmarek L, Buchanan J, Smith S (1987) Cyclic AMP Neurosci 8:463472. induces changes in distribution and transport of organelles within Misgeld U, Deisz R, Dodt H, Lux H (1986) The role of chloride trans- growth cones of Aplysia bag cell neurons. J Neurosci 7:3600-3611. port in postsynaptic inhibition of hippocampal neurons. Science 232: Franklin J, Johnson E (1992) Suppression of programmed neuronal 1413-1415. death by sustained elevations of cytoplasmic calcium. Trends Neu- Nissen R, Renaud L (1994) GABA receptor mediation of median rosci 15:501-508. preoptic nucleus-evoked inhibition of supraoptic neurosecretory neu- Frederick J (1987) The emergence of GABA-accumulating neurons rones in rat. J Physiol (Lond) 479:207-216. during retinal histogenesis in the embryonic chick. Exp Eye Res 45: Nowak L, Bregestovski F’, Ascher P, Herbert A, Prochiantz A (1984) 933-945. Magnesium gates glutamate-activated channels in mouse central neu- Fritschy J-M, Paysan J, Enna A, Mohler H (1994) Switch in expression rones. Nature 307:462-465. of rat GABA,-receptor subtypes during postnatal development: an Obata K, Oide M, Tanaka H (1978) Excitatory and inhibitory actions immunohistochemical study. J Neurosci 14:5302-5324. of GABA and glycine on embryonic chick spinal neurons in culture. Furshpan EJ, Potter DD (1989) Seizure-like activity and cellular dam- Brain Res 144: 179-184. age in rat hippocampal neurons in cell culture. Neuron 3:199-207. Obrietan K, van den Pal AN (1995) Calcium hyperexcitability in neu- Gaffey C, Mull&s L (1958) Ion fluxes during the action potential in rons cultured with glutamate receptor blockade. J Neurophysiol 73: Chara. J Phvsiol (Land) 144:505-524. 1524-1536. Gordon-Weeks-P, Lockerbie R, Pearce B (1984) Uptake and release of Oh D, Dichter M (1992) Desensitization of GABA-induced currents in ZH GABA by growth cones isolated from neonatal rat brain. Neurosci cultured rat hippocampal neurons. Neuroscience 49:571-576. Lett 52:202-212. Olney JW, Sharpe LG (1969) Brain lesions in an infant rhesus monkey Grover L, Lambert N, Schwartzkroin P, Teyler T (1993) Role of HCO, treated with monosodium glutamate. Science 166:386-388. - in depolarizing GABA, receptor-mediated responses in pyramidal Randle J. Bouraue C. Renaud L (1986)j I Characterization of suontane-I cells of rat hippocampus. J Neurophysiol 69:1541-1555. ous and evoked postsynaptic potentials in rat supraoptic neurosecre- Grynkiewicz G, Poenie M, Tsien R (1985) A new generation of cal- tory neurons in vitro. J Neurophysiol 56:1703-1717. cium indicators with greatly improved fluorescence properties. J Biol Reichling DB, Kyeozis A, Wang J, MacDermott AB (1994) Mecha- Chem 260: 1440-1447. nisms of GABA and glycine depolarization-induced calcium tran- Hales T, Sanderson M, Charles A (1994) GABA has excitatory actions sients in rat dorsal horn neurons. J Physiol (Lond) 476:411421. on GnRH-secreting immortalized hypothalamic (GTl-7) neurons. Reppert SM, Schwartz WJ (1984) The suprachiasmatic nuclei of the Neuroendocrinology 59:297-308 fetal rat: characterization of a functional circadian clock using 14C- Hansen G, Belhage B, Schousboe A, Meier E (1987) Temporal devel- labeled deoxyglucose. J Neurosci 4:1677-1682. opment of GABA agonist induced alterations in ultrastructure and Rorig B, Grantyn R (1993) Glutamatergic and GABAergic synaptic GABA receptor expression in cultured cerebellar granule cells. Int J currents in ganglion cells from isolated retinae of pigmented rats Dev Neurosci 5:263-269. during postnatal development. Dev Brain Res 74:98-l 10. The Journal of Neuroscience, July 1995, 75(7) 5077

Sakatani K, Black JA, Kocsis JD (1992) Transient presence and func- nocytochemical localization in presynaptic axons in the suprachias- tional interaction of endogenous GABA and GABA, receptors in mat& nucleus. Neuroscience 17-643-659. developing rat optic nerve. Proc R Sot Lond [Biol] Sci 247:155- van den Pol AN. Tromblev PO (1993) Glutamate neurons in hvuo- 161. thalamus regulate excitatory transmission. J Neurosci 13:2829-2g36. Sheffield J, Fischman D (1990) Calcium hotspots caused by L-channel van den Pol AN, Tsujimoto K (1985) Neurotransmitters of the hypo- clustering promote morphological changes in neuronal growth cones. thalamic suorachiasmatic nucleus: immunocvtochemical analvsis of Nature 343:751-754. 25 neuronaiantigens. Neuroscience 15:104911086. Shibata S, Moore RY (1988) Development of the fetal circadian van den Pol AN. Finkbeiner S. Cornell-Bell A (1992) Calcium excit- rhythm after disruption of the maternal circadian system. Dev Brain abilitv and oscillations in suprachiasmatic nucleus neurons and glia Res 41:313-317. in vi&. J Neurosci 12:2648-2664. Spoerri P (1988) Neurotrophic effects of GABA in cultures of embry- van den Pol AN, Obrietan KO, Cao V, Trombley P (1995) Embryonic onic chick brain and retina. Synapse 2: 1 l-22. hypothalamic expression of functional glutamate receptors. Neuro- Tappaz M, Oertel W, Wassef M, Mugnaini E (1982) Central GA- science, in press. BAergic neuroendocrine regulation: pharmacological and morpholog- Walton M, Schaffner A, Barker J (1993) Sodium channels, GABAa ical evidence. Prog Brain Res 55:77-96. receptors, and glutamate receptors develop sequentially on embryonic Taylor J, Docherty M, Gordon-Weeks P (1990) GABAergic growth rat spinal cord cells. J Neurosci 13:2068-2084. cones: release of endogenous gamma-aminobutyric acid precedes the Wu W, Ziskind-Conhaim L, Sweet M (1992) Early development of expression of synaptic versicle antigens. J Neurochem 54:1689-1699. glycine- and GABA-mediated synapses in rat spinal cord. J Neurosci Vaccarino F, Haywark M, Nestler E, Duman R, Tallman J (1992) Dif- 12:3935-3945. ferential induction of immediate early genes by excitatory amino acid Yamashita M, Fukuda Y (1993) Calcium channels and GABA receptors in the early embryonic chick retina. J Neurobiol 24:1600-1614. receptor types in primary cultures of cortical and striatal neurons. Yamashita M, Yoshimoto Y, Fukuda Y (1994) Muscarinic acetvlcho- Mol Brain Res 12:233-241. line responses in the early embryonic chick retina. J Neurobcol 25: van den Pol AN (1985) Dual ultrastructural localization of two neu- 114-1153. rotransmitter-related antigens: colloidal gold labeled neurophysin im- Yuste R, Katz L (1991) Control of postsynaptic Ca2+ influx in devel- munoreactive supraoptic neurons receive peroxidase labeled gluta- oping neocortex by excitatory and inhibitory neurotransmitters. Neu- mate decarboxylase or gold labeled GABA immunoreactive synapses. ron 6:333-344. J Neurosci 5:2940-2954. Zhang L, Spigelman I, Carlen P (1991) Development of GABA-me- van den Pol AN (1986) Gamma-aminobutyrate, gastrin releasing pep- diated, chloride-dependent inhibition in CA1 pyramidal neurones of tide, serotonin, somatostatin, and vasopressin: ultrastructural immu- immature rat hippocampal slices. J Physiol (Lond) 444:25-49.