CALCIUM-~N~EP~N~E~ GABA RELEASE STRIATAL SLICES: THE RULE OF CALCIUM CHANNELS
~ttsburgh,
FROM
PA 15260, U.S.A.
have investigated the role of Ca’+ and Ca2+ channels in the modulation of GABA release. Brain slices prepared from rat striatum were ~~in~~~t~ with [%fGABA, superfused with Krebs bicarbonate buffer, and exposed to electrical fiefd stimulation (2 Hz for 3 min). Tritium efllux was measured as an index ofGABA release. Both resting and evoked efflux were greatly accelerated by deleting Ca’+ from tke medium and adding EGTA (I m&4). However, when the concentration of iI&*+ in tke buffer was &vat& to 10 mM. no effect of the Ca2+-de%5ency was observed on resting release and its impact on evoked overflow was diminisked. Moreover, ~~~~on of verapamil (IO ,aM). a Ca2” channel blocking agent, reduced evoked overflow even in the. absence of external Caz*, while 4-aminopyridine (10 PM), a K” channel inhibitor, enhanced GABA efflux in normal buffer but had no effect in the absence of Ca*+. Finally, we have shown previously that nipecotic acid, an inhibitor of high affinity GABA transport, increases GABA overflow in normal buffer, but blocks it in Ca *+-free buffer. Collectively, these results suggest that Caz+ channeis may play two roles in the regulation of d~ola~at~on-jnd~ GABA release. Firstly, t&se channels permit a dE~la~zation_~~d~c~ infiux ofCazi which then promotes GABA release. In addition, these channels influence GABA release thraugh a rn~ka~isrn that does not involve external CaZ+. Although the precise nature of this latter involvement is unclear, we propose that the Cazs channels serve to permit an ir&ux of Na+, wkick in turn promotes Cal+-independent release through an Influence on the kigh affinity GABA transport system. Abstr&--We
EXPERIMENTALPROCEDURES
assumed that depolarization-induced release of nemotransmitter involves Ca*+-dependent exc;pcytosis.27.33 However, recently we have reported that GABA can be &eased from rat striatal dices by It is generally
electrical. depolarizatian in the absence of external Ca*+ 7.8 and similar observatians have been made by many others.S,1s,32~3P-33 Thus, it would appear that secretian-coupling of GABA can occur through a mechanism that is distinct from the usual Ca*+-
( + )~e~aparn~l k~dr~kIo~~, TlX3 ~arnj~opy~d~~e (4-AP), and ~m~ooxya~tic acid h~~kydr~~oride were obtained from Sigma Chemical Co, (St Louis, MO), (2,3-3H(N)]GABA (25 Ci/mmoi. NET-191X), from New England Nuclear (Boston, MA), and ‘IS-1 tissue solubilizer from Research Products Inte~a~onal Corp. (Mt Prospect, IL). All other chemicals and reagents were obtained from Fisher Scientific Co. (Pittsburgh, PA).
dependezx:ntprocess. We have previously ducidated several features of this rne~han~~~?*~ Firstly, it cas be blocked by nipecotic acid, an inhibitor of the high affinity GABA transport system present in GABAergic neurons. Secondly, although it is more prominent when examined by preIa~ili~g with f%jGABA, it also can be detected by measuring the overflow of ~~dogenous GABA. Thirdly, it is inhibited by tetrodotoxin (TTX). In this report we have further characterized the Ca’+kdependent release of GABA lati&ing several ~~arna~~o~~al tools that influence the transport of Gaze through the plasma membrane.
Male, Spra~e-~w~~~ rats from ~~~-M~ller Labomtories (Allison Park, PA), weighing 2&l-3OOg, were used throughout these experiments. They were housed individually for at least one week in a temperature-controlled room (2 l-23%) in which lights were on from 8~00a.m. to 8~00p.m. Animals were killed by decapitation, the brain was @eklg removed, and the striatum was dissected and placed in ice-cold Krebs bi~r~~te buffer (see below). Coronal slices (350 pm) were made beginning at the anterior pole of the striatum using a McIlwain-B~nkmnnn tissue chopper (Brinkmann Inst~m~~, Westbury, NY). The first two slices were discarded and the next four slices were used in the assessment of GABA release.
*To whom correspondence should be addressed. Abbrmiutions : 4-AP, 4-aminopyridine; EGTA,
GABA r&we was assessed as ~~vj~~~~~ ~~~~ Briefly. slices were ~rein~~~ at &C for 25 min in 1 ml of Krebs bicarbonate buffer (see below) containing 0.1 FM [‘HJGABA (2.5pCi). Individual slices were then rinsed with fresh buffer, transferred to a superfusion chamber, and superfused at 37°C at a rate of 0.24ml/min for 135&n.
eneglycol-bi@-amiaoetkylether)tetra-acetic tetrodotoxin.
add:
ethylTTX,
647
678
S.
BERNATH and
M. J.
ZIGMOND
Effluents were discarded during the first 60 min and therecompared with the response of slices prepared from the after S-nun fractions were collected and analysed for same animai and exposed to excess K’ in the absence of total tritium content by liquid scintillation spectroscopy. At drug. Means+ S.E.M. are given throughout this paper. the end of the experiments each slice was solubihzed with Treatment effects were examined using an analysis of variTS-1 and afso anafysed to determine the tritium content. ance followed by pofr-hoc comparisons using H Neumanfn comparable experiments employing a Krebs buffer conKeuls test. Differences were regarded to be significant if tainingaminooxy&tic acid, virtu&yall tritium collected in P < 0.05. the sunerfusion fluid was found to be I”H1GABA.23~30~3’~37 Unless otherwise noted, the medium for mcubation and superfusion contained (in mM): NaCI 117,KC1 4.7, MgCLz t.2, CaCI, 1.25, NaH,PO, 1.2, NaHCU, 2.5, glucose It.5 and aminooxyacetic acid 0.1. When external Ca2+ was eliminated, the buffer also contained the Ca’” chelating After an initial equilibration period, spontaneous agent, EGTA (1 .OmM). The buffer was gassed with 5% GABA efflux reached a low rate which gradually CO,-95% 0, and the OH was maintained at 7.4, adiustina _ decreased throughout the course of the subsequent witt; 0. I M NaGH as necessary. superfusion period. Thus, for example, the rate of Two methods were used to depolarize the tissue slices. electrical field stimulation and high K+. spontaneous efflux observed prior to and immediately Electrical $eld slimulation. Slices were stimulated twice after the first stimulation period (B,) was 18% higher with biphasic pulses for 3-n&t periods coinciding with the than that observed at the time of the second such beginning of the third (S,) and 12th (S,) 5-min fractions. Constant current stimulation was dehvered by a Grass S88 period (I&). Three minutes of electrical stimulation significantly increased GABA overfIow, which Stimulator (Grass Medical Instruments, Quincy, MA) coupled through two Grass SIU5 Stimulus Isolation Units reached a peak during the first 5 min. As in the case to two Grass CCUI Constant Current Units. A 5DOn of spontaneous efflux, the e~~tr~cail~evoked resistor was connected in series with the slice chambers to overflow of GABA declined during the experimental permit monitoring of the stimuhts with 8n oscilloscope. period, such that overtlow in reponse to the first Each stimulus consisted of a positive 20-mA, 3-ms rectangular pulse followed 0.1 ms later by a negative pulse of comstimulation (S,) was 22% higher than that occurring parable amplitude and duration. These pulse pairs were in response to the second (&). delivered at a frequency of 2 Hz. Drugs were added 20 min prior to Sr. Stimulation with excess K+. K” concentration was eleSlices were superfused from the outset of the vated to 25 mM and the NaCl concentration was reduced to maintain constant osmolarity. Tissues were exposed to the experiment with a Ca ‘+-free buffer containing EGTA depola~~jng buffer for the 10 min comprising the third and (I mM) and the results compared with those obtained the fourth samples. Drugs were added to the depolarizing with normal buffer. Cal+-free buffer increased basat buffer. c
The amount ofi’H]GABA present in the slice at any given time was estimated retrospectively by summing a% tritium collected in the superfusate in subsequent fractions with the amount present in the slice at the end of the experiment. Efflux of [)H]GABA was calculated by computing the amount of tritium present in each 5-min fraction as a percentage of the tritium estimated to be present in the slice at the beginning of that fraction. Baseline efflux (B, , B,) was estimated by determining the mean efflux far the 5-min fraction immediately before and after the fractions used to determine overgow. Overflow (S, , SJ was caiculated by first determining e&x from the onset of stimulation until the prestimulation baseline was restored and then subtracting estimated basal efflux for this period. To quantify the effect of alterations in the constituents of the medium, basal efflux and spon~neous overtTow in the presence of the drug were divided by the comparable vafues obtained in the absence of the drug. When electricat field stimulation was used, this involved a determination of the ratios E2/8, and S,jS, . For studies employing high K’, the response to this stimuhzs in the presence of drug was
efflux J&fold and increased electrically-evoked overflow 13-fold. in addition, the gradual decline in resting and e~~trica~Iy-evoke efflux observed during an experiment carried out under normal conditions did not occur in the absence of Ca*+ (Table 1).
In sufficiently high concentration, Mg*’ is known to block Ca’+ channels? To examine the impact of excess Mg*+ (10 mM) the #n~ntration of this ion was elevated 20min before the second stimulation period. In the presence of a normal concentration of Ca2+. excess Mg”” reduced electrically-evoked GABA overflow by 83%. A simiiar effect was observed when Mg’” was increased in the absence of CaZ*. Elevated Mg*+ had no effect on basal GABA efflux either in the presence or absence of Ca: + (Table 2).
Table 1. Effect of Ca2+-deficiency and excess Mg*+ on spontaneous and electrically-evoked [3H]GABA release Buffer
n
Normal Ca’+sfree
8 8
%, 0.46 + 0.02 2.03 40.11*
B* 0.39 & 0.02 2.26 k 0.14*
%t& 0.82 * 0.02 1. I I + 0.05*
S, 0.45 f 0.08 5.86 * 0.96*
S, 0.35 i_ 0.06 5.54 & 0.91’
S,!S, 0.78 & 0.02 0.97 ri: 0.08*
3, and B, refer to the estimated value of basal efflux; S, and S, are overflow due to electrical stimulation. Both basal and evoked et&x are expressed as a percentage of tissue tritium per 5 min. Shown are the mean + S.E.M. @‘-free buffer contained t mM EGTA and was apphed throughout the superfusion. *P < 0.01 when compared with standard buEer (1.25mM Ca2*, 1.2mM Ma”+).
Cakium-i~de~ndent
GABA release from striatal slices
679
Effect of verapamil on CABA release in the presence and absence of Ca2’ In order to further examine the role of Cal+ in GABA release, studies were carried out in the
presence of verapamil, a Ca2+ channel antagon~st.2~ Verapamif (to PM) enhanced basal GABA e@ux by 28%, while decreasing evoked overflow by 45%. In the absence of Ca”, the drug reduced basal efflux by 45% and decreased electrically-evoked overflow by 85% (Table 3). EJfct qf ~-~~~o~~ridi~e on GABA release in the presence and absence of Ca” Next, the CaZ+ channel was manipulated indirectly by means of the K* channe1 antagonist, 4-AP.4’ By prolonging the action potential, this drug causes the voltage-de~ndent Ca2+ channel to remain open for an extended period. We observed that in the presence of CaZi, 4-AP (10mM) enhanced both the spontaneous efflux (+ 52%) and evoked overflow ( + 78%) of GABA. However, in the absence of Caa+, the drug had no significant impact on either spontaneous or evoked efflux (Table 3). Efect of tetrodotoxin on GABA overtop evoked by K+ Fin&y, we examined the impact of selectively blocking voltage-dependent Na+ channels while opening
voltage-dependent
Ca2+ channels.
This was
Table 2, Effect of excess M$+ on sponianeous electrically-evoked [31-I]GABAefflux
and
Buffer
n
4/B,
sz is,
Normalt
8
Excess Mg:* Excess Mgl’, Ca”-free
4 4
0.82 + 0.02 0.88 & 0.04 0.71 + 0.05
0.78 + 0.02 0.13 & o.oz* O.12 f o&2*
B, and & refer to the estimated value of basal efflux; S, and S, are overflow due to electrical stimulation. Both basal and evoked efflux are expressed as a percentage of tissue titer per 5 min. Shown are the mean & S.E.M. The addition of excess Mg2+ (10 mM) with or without CaZ+ occurred 20 min before the S2. *P < 0.01 when compared with standard buffer (1.25 mM Ca”, 1.2mM Mg’+). Waluea from Table I provided for comparison. Table 3. Effect of verapamil and 4-aminopyridine efflux of [3H]GABA
on the
Buffer
n
WB,
&is,
Standard Control Verapamil 4-AP
8 4 4
0.82 I: 0.02 I .os& 0.03* i .25 & 0.02*
0.78 & 0.02 0.43 _t o.os* 1.36&0.10*
Ca’+-free Control Verapamit 4-AP
8 4 4
1.11 rfro.05 0.61 + 0.10* 1.14i0.23
0.97 _+0.08 0.1.5*0.06* 0.76iO.11
Ca’+ -free conditions (i 1 mM EGTA added) were applied throughout the superfusion and verapamil (1Opm) and 4-AP (10 mM) were administered 20 min before Sz. *P < 0.05 when compared with the corresponding control.
K+ (25 mM) 3=(5&M
Fig. 1. Effect of TTX on [)H]GABA overflow evoked by high K+ in the presence or absence of Ca2+. Excess K’ (25 mM) was added to the buffer for 10 min as indicated (bar). This occurred in the presence (a) or absena: (of of TTX @PM). The superfusion buffer contained a normal Ca2+ concentration (1.25 mM; top panel) or was Ca2+-free and contained EGTA (1 mM; bottom panel). Results are expressed as a percentage of tissue tritium per 5 min and each point represents the mean 2 S.E.M. of six experiments. Note change in scale between the top and bottom panels. *P < 0.05 compared with control. 0-Q control; *--*, Tl-X.
~~ornpl~shed by elevating extra~~Iu~a~ K+ to produce a local depolarization and then examining overflow in the presence of TTX. Elevating K+ to 2SmM elicited a large overflow of GABA. TTX (5 FM) had no signi~cant effect on basal efRux, but reduced GABA overflow by 25% (S, control, 17.2 + 1.7, n = 6; TTX, 12.1 f 1.2, n = 6) (Fig. 1). The response to excess K+ in the absence of external Ca’” was increased over that observed with standard buffer. However, under these conditions, the inhibition caused by TTX rose to 70% (S, control, 26.0 + 8.3, n = 6; TTX. 6.7 f 2.4, n = 6) (Fig. 1). DISCUSSION
We have previously observed that depo~a~zationinduced GABA overf?ow is not blocked by removing external Ca*+; indeed, it is greatly facilitated.?” Moreover, spontaneous GABA efflux is also increased in the absence of external Ca’+. This study was designed to further explore these phenomena.
680
SBERNATH
and
Again, we observed that the overflow of GABA from striatal slices was facilitated by withdrawing Ca” from the medium. However, Ca’+-deficiency had no effect on evoked GABA efflux when Mg2+ concentration was elevated. Mg’+ can block Ca2+ channels,’ presumably explaining its well known ability to antagonize the effect of Ca’+ on transmitter release.‘7.25.40Thus, our results suggest that Ca’+ channels continue to play a role in GABA release even in the absence of external Ca’+. Further support for a role of Ca’.+ channels in Ca”-independent GABA release derives from our observations with verapamil. This drug, which can inhibit Ca2+ flux across excitable membranes,“,2y also depressed the electrically-evoked GABA overhow in Ca’+-free conditions. Similarly, verapamil has been shown to inhibit the overtIow of GABA evoked by excess K + in the absence of external Ca”. I4 Ca’+ channel antagonists, such as verapamil, may interact with sodium channels as well, especially when they are used at higher concentrations.‘~,‘~.4~ However, we have reported that TTX has no effect on basal [3H]GA%A efflux in the presence of either normal or Ca?+-free buffer.’ Therefore, it is more likely that the significant effect of verapamil on GABA release is via Ca’+ channels rather than on voltage-dependent Na+ channels. How might Ca?’ channels play a role in GABA release in the absence of Cal+ itself? One possibility involves the reported ability of Na’ to move through voltage-sensitive Ca2+ channels.3~4~‘2~22 The absence of external Ca’+ should reduce competition for Na+ entry through these channels and, thus, increase the inward flux of Na”. Mgz+ and verapamil, on the other hand, would act to offset this increased Nat flux by blocking Ca’+ channels. The increase in intracellular Nat promoted by removing Ca’+ from the medium could affect GABA release in one of two ways. Firstly. the resulting entry of Na + could release Cal+ from intraneuronal pools through a Na+ /Ca?+ exchange system.‘8.34~4’ This, in turn, would be expected to enhance spon~neous transmitter release.‘.6,39However, although this model provides an explanation for an increased GABA overflow in the absence of external Ca’.‘. it does not explain the reported failure of a Ca’+-free buffer to increase the release of other neurotransmitters.27.33 A second explanation involves the apparent relation between GABA release and the high affinity GABA transport system. In the presence of a normal concentration of Ca”, the high affinity uptake of GABA plays a major role in removing GABA from the synaptic cleft. 7s3’.MIndeed, this may explain the ability of verapamil to enhance the resting outfiow of GABA. since it has been reported that verapamil can inhibit Na+-coupled GABA uptake.19~‘”The driving force for GABA transport, as for a variety of other carrier-mediated transport processes. is the electroInceasing the intrachemical potential of Na + ,‘1.20,3h cellular concentration of Nat by removing external
M.J.
ZIGMOND
Ca” would be expected to reduce or even reverse the action of such a transport system.42 Elsewhere we have noted that nipecotic acid, an inhibitor of this transport, blocks the release of GABA under Ca’+free conditions. Thus, we have suggested that Ca’+-independent GABA release may be mediated by the operation of the GABA transport system in an outward direction.7s8 Removal of external Ca’+ increased s~ntaneous GABA efflux as well as GABA overlIow. The efflux of GABA during depolarization differs from resting efflux in that it is the consequence of explosive ion movements through excitable membranes. Nonetheless, the antagonizing effect of Mgz+ and verapamif on both spontaneous efflux and evoked overflow in the absence of external Ca2+ suggests that both phenomena share a common m~hanism. Our findings may provide some insight into the basis for the discrepancies that exist in the literature regarding the role of Ca2+ in GABA release. In examining this issue, ~nv~tigators have employed a wide variety of approaches to alter Ca’+ concentration. For example, some have lowered the Ca2+ concentration to 0.1 mM, others have withdrawn external Ca2+ entirely (with or without added EGTA), and still others have substituted Mg2+ for Ca2 +. In our experience, although these procedures are superficially comparable with each other, they can produce quite different results. For exampie, we have observed in the present series of experiments that the electrically-evoked overflow of GABA is greatly enhanced during superfusion with a Ca2+-free solution containing EGTA, but is diminished when the Mg2+ concentration is elevated. Moreover, we have previously reported that although removing Ca2 + greatly increases GABA overflow, reducing external Ca’+ to 0.1 mM actually depresses GABA overflow.’ Such apparently contradictory findings regarding the relation between Ca2 + and GABA release may be explicable in terms of the competition between Na+ and other ions for the Ca2+ channel. Inward Nat flux through Ca2+ channels should be greatest in the presence of a Cal+-free solution containing EGTA when competition between Nat and Ca2+ has been eliminated, somewhat lower in the presence of 0.1 mM Ca” + when some competition between Na+ and Ca2+ remains, and nearly abolished in the presence of l0mM Mg*+ when the presence of ions to compete with Na+ has been increased several fold. The predicted changes in Na + flux in response to these manipulations parallels the observed responses of GABA efflux, suggesting a causal relation between these two variables. In the foregoing discussion, we have proposed that GABA can be released through a process that does not require external Ca 2+. On the other hand, our results suggest that GABA can also be released by classical Ca?+-dependent exocytosis and that this latter mechanism predominates in the presence of a normal concentration of external Ca’+. Firstly, Mg’+
~aleium-inde~nden~ and
verapamil
both
decreased
GABA release from striatal slices
electrically-evoked
GAEIA o~erfIow in the presence of CaZ’. Since both Mgz+ and verapamil nels, these results when
present
woufd act to block
can be interpreted
in the external
Ca2’
to suggest
medium,
Cazc
chanthat influx
plays an ~m~ortaut
roie in GABA release. Secondly, although 4AI? had no elect on GABA efRux in the absence of Ca*+, it greatiy increased GABA avertlow in the presence of Cazt . By blocking K” channels, 4-AP enhames the time-course of the action potential, thereby enhancing Ca’+ uptake?.2a4s Though it appears likely that Na’ entry can directly contribute to an increased GABA e&ix under Ca2+ deficient conditions, the question still remains as to whether Cafe-~~de~ndent, ~~~~de~nd~nt GABA release exists under physiological conditions, enters mainly through when Na+ presumably voltage-de~ndent Naf channels. In order to determine how Ca’* and Na+ ion movements contribute to GABA efflux, we examined the impact of an elevated concentration of external K' . Electrical depolarization appears to release transmitter primariiy via an indirect route: ~~oItage-d~~ndent Ha’ channels are opened, initiating an action potentiaf. which ultimately opens Ca2+ channels in the nerve terminals. In contrast, excess K+ seems to act by local depolarization and, as a result, opens voltagedependent Na+ and Ca2* channels directly.io~24Thus, while TTX comoletelv abolished the effects of electrical stimulation on ~a2-~-de~ndent GABA release,’ 1
681
this inhibitor
of Na’ channels only depressed the K+-evoked overffow of GABA by about 25%. since the potassium concentration appiied was high enough to totally depolarize tissues,24 this suggests that Na+ mov~ent through voltage-de~ndent Na+ channels can directly contribute to GABA release3 although this contribution is relatively small, The higher impact of TTX on the Kc-evoked overflow of GABA in Ca2+-free conditions may be related to the enhanced axonal excitability produced by withdrawing Ca2 + 28and this may underline the jrn~ortan~ of voltage-de~ndent Na* channels.
Our ~x~riments suggest that depolarizationinduced GABA reIease can occur by two processes, one which requires an inward movement of Ca2” via voltage-dependent Ca2+ channels. The second process, which can occur even in the absence of external Ca*+, may involve an inward Na+ flux which could act to alter the ~~i~b~urn betwen the inward and outward transport of GABA via a high a~nity GABA pump. ~c~5~le~~~~~~~-W~ thank Drs Ray Ana Wallace and David C. Wood for befpful discussions, and MS ‘Terri L. Komar for assistance in preparing this rrqort. The research was supported in part by grants from the U.S.P.H.S. (NS- 19608, MN-43947 and MH-~~8~.
REFERENCES
f. Adams D, J., Takeda K. and Umbach f. A. (1985) Inhibitors of calcium buffering depress evoked transmitter release at squid giant synapse. J. f&s&L,
Land. 369, 145-159.
2. Agoston D,, Har&tai P. and Magy A. (1983) Elects of a ~a~inopyridine
on calcium ~ove~~~~ and caulk of membrane potential in pinched-off nerve terminals from rat cerebral cortex. J. Neur~c~e~. 41, 745-751. 3. Alrners W., McCluskey E. W. and Palade P. T. (1984) A nonselective caation conductance in frog muscle membrane blocked by micromolar external calcium ions. J. Physiol., Ltxstf. 353, 565-533. 4. Alsners W. and MCeleskq E. W. (t984) Non-selective conductance in cal&m cbm~~ls of frog m~~!e: c&cium selectivity in a single&e pore. 3. Phjlsiaf., Land. 353, 585-608. 5. Arias C. and Tapia R. (1986) Differential calcium dependence of y-aminobutyric acid and acetylcholine release in mouse brain synaptoso~~s. J. ~~~r~~~~e~~47, 396-404. 6. Bemarh S. and Vizi E. S. (1987) Inhibitory e&c$ af ionized free in~~~llular calcium enhanced by Ru~he~~u~ Red and nr-chfanr-farbonykyanidt: phenyl hydrazon on the evoked release of a~ty~cbol~ne. B&&w. ~~~~ac. 36, 3683-3687. 7. Bernath S. and Zigmond M. .I. (1988) Characterization of [3H]GABA release from striatal slices: evidence for a calcium-independent process via the GABA uptake system, Neuroscience 27, 563-570. 8. Bemath S., Keller: R, W. Jr and Zigmond M. J. (1989) Release of endogenous GABA can occur through Ca”+-dependent and Caz*-independent processes. ivewochenr. Rt. I& 439-445. 9. Bixby 1. L. and Spitzer 8. C. ((984) Early d~~e~n~atio~ of vertebrate spinaX neurons in the absence of voltage dependent Cal+ and Na+ influx. I>ati &al. X06,89-96. 10. Blaustein M. P. (1975) Effect of potassium, veratridine and scorpion venam on calcium accumulation and transmittet release by nerve terminals in vim. J. Physiol., Land. 241, 617-655. Ii. Blmstein M. P. and King A. C. (1976) Influence of membrane potential on the ~i~~e~~~en~ uptake of ~~~~a-a~jnobuty~c acid by presynaptic nsrve terminals: experimental observations and theoretical considerations. 3. .~ernbrune Biol. 30, 153-173. 12. Bossu S.-L. and F&z A. (1984) Patch ciamp study of the tetrodotoxin resistant sodium current in group C sensory newones. Neurosci. tetl. 51, 241-246. 13. Bus~a~an~e 3. 0. (1985) Block of Na+-currents by the Ca2’ -antagonist D-600 in human beart ceil semen&, Q&#gers A&. g&Y.P&&J!. 4@3, 225-227. 14. Carvalho C. M., Santos S. V. and Carvaiho A. P. (1986) y-Aminobutyric acid release from synaptosomes as influenced by Ca?’ and Ca*+ channel blockers. Eur. J. Pharmac. 131, l-12. 1.5. Caudill W. L., Bigelow J. C. and Wightman R. M. (19gS) Comparison of release of endogenous dopamine and y-aminobutyric acid from rat caudate synaptosomes. 3eurochem. Res. 16, 3I9-331.
I”7
S. BERNATH and M. J. ZIGMOND
on:
16. Clay J. R. and Shrier A. (1984) Effects of D-600 on sodium current in squid axons. J. Membrune Biol. 79, 211-214. 17. Cotman C. W., Haycock J. W. and White W. F. (1976) Stimulus-secretion coupling processes in brain: analysis of noradrenaline and GABA release. J. Physiol., Land. 254, 475-505. 18. Cunningham J. and Neal M. J. (1981) On the mechanism by which veratridine causes a calcium-independent release of ~-amjnobutyric acid from brain slices. Br. J. Phurmac. 73, 655-667. 19. Erdreich A., Spanier R. and Raham~moff H. (1983) The inhibition of Na-dependent Ca uptake by verapamil in synaptic plasma membrane vesicles. Eur. J. Pharmac. 90, 193-202. 20. Erecinska M. (1987) The neurotransmitter amino acid transport systems. Biochem. Pharmac. 34, 3547-3555. 21. Fleckenstein A. (197 1) Specific inhibitors and promoters of calcium action in the excitation--contraction coupling of heart muscle and their role in the prevention or production of myocardial lesions. In Calcium and the Heurt (eds Harris P. and Opie L.), pp. 1355188. Academic Press, London. 22. Fukuda J. and Kameyama M. (1980) Tetrodotoxin sensitive and tetrodotoxin resistant sodium channels in tissue cultured spinal ganglion neurones from adult mammals, Brain Res. 182, 191-197. 23. Haycock J. W., Levy W. B., Dibber L. A. and Cotman C. W. (1978) Effects of elevated (K’ )0 on the release of nenrotransmitters from cortical synaptosomes: efflux or secretion? 1. Meurochem. 30, it 13-l 125. 24. Hillman H. H. and Mcllwain H. (1961) Membrane potentials in mammalian cerebral tissues in r&o: dependence on ionic environment. J. PhyGol., Lond. 157, 263-278. ‘5. Hubbard J. I., Jones S. F. and Landau E. M. (1968) On the mechanism by which calcium and magnesium affect the release of transmitter by nerve impulses. J. Physiol., Lond. 196, 75-86. 26. Jones R. S. G. and Heinemann U. (1987) Pre- and post-synaptic K+ and Ca’+ fluxes in area CA1 of the rat hippocampus in vitro: effects of Ni*+, TEA and 4AP. &pi Brain Res. 68, 205-209. 27. Katz B. and Miledi R. (1970) Further study of the role of calcium in synaptic transmission. J. Physiol., Lond. 207, 789 -80 I. 28. Keesey J. C.. Wallgren H. and McIlwain H. (1965) The sodium. potassium and chloride ofcerebral tissues: maintenance. change on stimulation and subsequent recovery. Biochem. J. 95, 289-300. 29. Kohlradt M. B.. Baner B.. Krausse M. and Fleckenstein A. (1972) New selective inhibitors of transmembrane calcium conductivity in mammalian myocardial fibers. Studies with voltage clamp technique. Experientia 28, 288-289. 30. Levy W. B. (1973) Stimulus-coupled secretion of y-aminobutyric acid from rat brain synaptosomes. Science 181. 676 -678. 31. Limberger N., Spath L. and Starke K. (1986) Release of previously incorporated g-[‘H]aminobutyric acid in rabbit caudate nucleus slices. J. .~euroehem. 46, 1102-I 108. 32. Liron Z., Roberts E. and Wong E. (1985) Verapamil is a competitive inhibitor of p-aminobutyric acid and calcium uptakes by mouse brain subcellular particles. Life Sci. 36, 321-327. 33. Llinas R., Steinberg I. Z. and Walton K. (1976) Presynaptic calcium currents and their relation to synaptic transmission: voltage clamp study in squid giant synapse and theoretical model for the calcium gate. Proc. natn. Acad. Ski. U.S.A. 73, 29 18-2922.
34. Lockerbie R. 0. and Gordon-Weeks P. R. (1986) Further characterization of [3H]~-aminobutyric acid release from isolated neuronal growth cones: role of intracellular Car+ stores. Neuroscience 17, 1257-1266. 35. Maurin Y., Arbilla S. and Langer S. Z. (1985) Inhibition by yohimbine of the calcium-dependent evoked release of [‘H]GABA in rat and mouse brain shces irt rifro. Eur. J. Pharmac. 111, 3748. 36. Nelson M. T. and Blaustein M. P. (1982) GABA efflux from synaptosomes: effects of membrane potential, and external GABA and cations. J. Membrane Biol. 69, 213-223. 31. Okada K. and Hassler R. (I 973) Uptake and release of 7 -aminobutyricacid (GABA) in slices of substantia nigra of rat. Brain Res. 49, 214-217. 38. Peris. J.. Harris R. A. and Zahniser N. R. (1987) Modulation of y-PHjaminobutyric acid release from rat cortical slices by cc?-adrenoceptors. Neurosis. Lett. 80, 309..3 14. 39. Rahamimoff R. and Atnaes E. (1973) Inhibitory action of Ruthenium Red on neuromuscular transmission. Proc. nom. Acad. SC;. U.S.A. 70, 3613-3616. 40. Richards C. D. and Sercombe R. (1970) Calcium, magnesium and the electrical activity of guinea-pig olfactory cortex iti vitro. 1. Phwiol., Lund. 211, 571 -584. 41. Sandoval M. E. (1980) Sodium-de~ndent efBux of rZH]GABA from synaptosomes probably related to mitochondrial calcium mobilization. J. Neurochem. 35, 915-92 1. 42. Schwartz E. A. (I 987) Depolarization without calcium can release y-aminobutyric acid from a retinal neuron. Science 238, 350-355. 43. Szerb .J. C. (1979) Relationship between Ca-” -dependent and independent release of [‘HIGABA evoked by high Kt . veratridine or electrical stimulation from rat cortical slices. J. Neurochem. 32, 1565-I 573. 44. Szerb J. C. f 1982) Effect of nipecotic acid. a y-aminobutyric acid transport inhibitor, on the turnover and release of y-aminobutyric acid in rat cortical slices. J. Neurochem. 39, 850-858. 45. Thesleff S. (1980) Aminopyridines and synaptic transmission. Neuroscience 5, 1413-1419. 46. Yatani A. and Brown A. M. (1985) The Ca 2+ channel blocker nitrendipine blocks sodium channels in neonatal rat cardiac myocytes. Cir~z~~ati[~n Rex 56, 868-875. (Accepted 27 NotTemher 1989)
~ttsburgh,
FROM
PA 15260, U.S.A.
have investigated the role of Ca’+ and Ca2+ channels in the modulation of GABA release. Brain slices prepared from rat striatum were ~~in~~~t~ with [%fGABA, superfused with Krebs bicarbonate buffer, and exposed to electrical fiefd stimulation (2 Hz for 3 min). Tritium efllux was measured as an index ofGABA release. Both resting and evoked efflux were greatly accelerated by deleting Ca’+ from tke medium and adding EGTA (I m&4). However, when the concentration of iI&*+ in tke buffer was &vat& to 10 mM. no effect of the Ca2+-de%5ency was observed on resting release and its impact on evoked overflow was diminisked. Moreover, ~~~~on of verapamil (IO ,aM). a Ca2” channel blocking agent, reduced evoked overflow even in the. absence of external Caz*, while 4-aminopyridine (10 PM), a K” channel inhibitor, enhanced GABA efflux in normal buffer but had no effect in the absence of Ca*+. Finally, we have shown previously that nipecotic acid, an inhibitor of high affinity GABA transport, increases GABA overflow in normal buffer, but blocks it in Ca *+-free buffer. Collectively, these results suggest that Caz+ channeis may play two roles in the regulation of d~ola~at~on-jnd~ GABA release. Firstly, t&se channels permit a dE~la~zation_~~d~c~ infiux ofCazi which then promotes GABA release. In addition, these channels influence GABA release thraugh a rn~ka~isrn that does not involve external CaZ+. Although the precise nature of this latter involvement is unclear, we propose that the Cazs channels serve to permit an ir&ux of Na+, wkick in turn promotes Cal+-independent release through an Influence on the kigh affinity GABA transport system. Abstr&--We
EXPERIMENTALPROCEDURES
assumed that depolarization-induced release of nemotransmitter involves Ca*+-dependent exc;pcytosis.27.33 However, recently we have reported that GABA can be &eased from rat striatal dices by It is generally
electrical. depolarizatian in the absence of external Ca*+ 7.8 and similar observatians have been made by many others.S,1s,32~3P-33 Thus, it would appear that secretian-coupling of GABA can occur through a mechanism that is distinct from the usual Ca*+-
( + )~e~aparn~l k~dr~kIo~~, TlX3 ~arnj~opy~d~~e (4-AP), and ~m~ooxya~tic acid h~~kydr~~oride were obtained from Sigma Chemical Co, (St Louis, MO), (2,3-3H(N)]GABA (25 Ci/mmoi. NET-191X), from New England Nuclear (Boston, MA), and ‘IS-1 tissue solubilizer from Research Products Inte~a~onal Corp. (Mt Prospect, IL). All other chemicals and reagents were obtained from Fisher Scientific Co. (Pittsburgh, PA).
dependezx:ntprocess. We have previously ducidated several features of this rne~han~~~?*~ Firstly, it cas be blocked by nipecotic acid, an inhibitor of the high affinity GABA transport system present in GABAergic neurons. Secondly, although it is more prominent when examined by preIa~ili~g with f%jGABA, it also can be detected by measuring the overflow of ~~dogenous GABA. Thirdly, it is inhibited by tetrodotoxin (TTX). In this report we have further characterized the Ca’+kdependent release of GABA lati&ing several ~~arna~~o~~al tools that influence the transport of Gaze through the plasma membrane.
Male, Spra~e-~w~~~ rats from ~~~-M~ller Labomtories (Allison Park, PA), weighing 2&l-3OOg, were used throughout these experiments. They were housed individually for at least one week in a temperature-controlled room (2 l-23%) in which lights were on from 8~00a.m. to 8~00p.m. Animals were killed by decapitation, the brain was @eklg removed, and the striatum was dissected and placed in ice-cold Krebs bi~r~~te buffer (see below). Coronal slices (350 pm) were made beginning at the anterior pole of the striatum using a McIlwain-B~nkmnnn tissue chopper (Brinkmann Inst~m~~, Westbury, NY). The first two slices were discarded and the next four slices were used in the assessment of GABA release.
*To whom correspondence should be addressed. Abbrmiutions : 4-AP, 4-aminopyridine; EGTA,
GABA r&we was assessed as ~~vj~~~~~ ~~~~ Briefly. slices were ~rein~~~ at &C for 25 min in 1 ml of Krebs bicarbonate buffer (see below) containing 0.1 FM [‘HJGABA (2.5pCi). Individual slices were then rinsed with fresh buffer, transferred to a superfusion chamber, and superfused at 37°C at a rate of 0.24ml/min for 135&n.
eneglycol-bi@-amiaoetkylether)tetra-acetic tetrodotoxin.
add:
ethylTTX,
647
678
S.
BERNATH and
M. J.
ZIGMOND
Effluents were discarded during the first 60 min and therecompared with the response of slices prepared from the after S-nun fractions were collected and analysed for same animai and exposed to excess K’ in the absence of total tritium content by liquid scintillation spectroscopy. At drug. Means+ S.E.M. are given throughout this paper. the end of the experiments each slice was solubihzed with Treatment effects were examined using an analysis of variTS-1 and afso anafysed to determine the tritium content. ance followed by pofr-hoc comparisons using H Neumanfn comparable experiments employing a Krebs buffer conKeuls test. Differences were regarded to be significant if tainingaminooxy&tic acid, virtu&yall tritium collected in P < 0.05. the sunerfusion fluid was found to be I”H1GABA.23~30~3’~37 Unless otherwise noted, the medium for mcubation and superfusion contained (in mM): NaCI 117,KC1 4.7, MgCLz t.2, CaCI, 1.25, NaH,PO, 1.2, NaHCU, 2.5, glucose It.5 and aminooxyacetic acid 0.1. When external Ca2+ was eliminated, the buffer also contained the Ca’” chelating After an initial equilibration period, spontaneous agent, EGTA (1 .OmM). The buffer was gassed with 5% GABA efflux reached a low rate which gradually CO,-95% 0, and the OH was maintained at 7.4, adiustina _ decreased throughout the course of the subsequent witt; 0. I M NaGH as necessary. superfusion period. Thus, for example, the rate of Two methods were used to depolarize the tissue slices. electrical field stimulation and high K+. spontaneous efflux observed prior to and immediately Electrical $eld slimulation. Slices were stimulated twice after the first stimulation period (B,) was 18% higher with biphasic pulses for 3-n&t periods coinciding with the than that observed at the time of the second such beginning of the third (S,) and 12th (S,) 5-min fractions. Constant current stimulation was dehvered by a Grass S88 period (I&). Three minutes of electrical stimulation significantly increased GABA overfIow, which Stimulator (Grass Medical Instruments, Quincy, MA) coupled through two Grass SIU5 Stimulus Isolation Units reached a peak during the first 5 min. As in the case to two Grass CCUI Constant Current Units. A 5DOn of spontaneous efflux, the e~~tr~cail~evoked resistor was connected in series with the slice chambers to overflow of GABA declined during the experimental permit monitoring of the stimuhts with 8n oscilloscope. period, such that overtlow in reponse to the first Each stimulus consisted of a positive 20-mA, 3-ms rectangular pulse followed 0.1 ms later by a negative pulse of comstimulation (S,) was 22% higher than that occurring parable amplitude and duration. These pulse pairs were in response to the second (&). delivered at a frequency of 2 Hz. Drugs were added 20 min prior to Sr. Stimulation with excess K+. K” concentration was eleSlices were superfused from the outset of the vated to 25 mM and the NaCl concentration was reduced to maintain constant osmolarity. Tissues were exposed to the experiment with a Ca ‘+-free buffer containing EGTA depola~~jng buffer for the 10 min comprising the third and (I mM) and the results compared with those obtained the fourth samples. Drugs were added to the depolarizing with normal buffer. Cal+-free buffer increased basat buffer. c
The amount ofi’H]GABA present in the slice at any given time was estimated retrospectively by summing a% tritium collected in the superfusate in subsequent fractions with the amount present in the slice at the end of the experiment. Efflux of [)H]GABA was calculated by computing the amount of tritium present in each 5-min fraction as a percentage of the tritium estimated to be present in the slice at the beginning of that fraction. Baseline efflux (B, , B,) was estimated by determining the mean efflux far the 5-min fraction immediately before and after the fractions used to determine overgow. Overflow (S, , SJ was caiculated by first determining e&x from the onset of stimulation until the prestimulation baseline was restored and then subtracting estimated basal efflux for this period. To quantify the effect of alterations in the constituents of the medium, basal efflux and spon~neous overtTow in the presence of the drug were divided by the comparable vafues obtained in the absence of the drug. When electricat field stimulation was used, this involved a determination of the ratios E2/8, and S,jS, . For studies employing high K’, the response to this stimuhzs in the presence of drug was
efflux J&fold and increased electrically-evoked overflow 13-fold. in addition, the gradual decline in resting and e~~trica~Iy-evoke efflux observed during an experiment carried out under normal conditions did not occur in the absence of Ca*+ (Table 1).
In sufficiently high concentration, Mg*’ is known to block Ca’+ channels? To examine the impact of excess Mg*+ (10 mM) the #n~ntration of this ion was elevated 20min before the second stimulation period. In the presence of a normal concentration of Ca2+. excess Mg”” reduced electrically-evoked GABA overflow by 83%. A simiiar effect was observed when Mg’” was increased in the absence of CaZ*. Elevated Mg*+ had no effect on basal GABA efflux either in the presence or absence of Ca: + (Table 2).
Table 1. Effect of Ca2+-deficiency and excess Mg*+ on spontaneous and electrically-evoked [3H]GABA release Buffer
n
Normal Ca’+sfree
8 8
%, 0.46 + 0.02 2.03 40.11*
B* 0.39 & 0.02 2.26 k 0.14*
%t& 0.82 * 0.02 1. I I + 0.05*
S, 0.45 f 0.08 5.86 * 0.96*
S, 0.35 i_ 0.06 5.54 & 0.91’
S,!S, 0.78 & 0.02 0.97 ri: 0.08*
3, and B, refer to the estimated value of basal efflux; S, and S, are overflow due to electrical stimulation. Both basal and evoked et&x are expressed as a percentage of tissue tritium per 5 min. Shown are the mean + S.E.M. @‘-free buffer contained t mM EGTA and was apphed throughout the superfusion. *P < 0.01 when compared with standard buEer (1.25mM Ca2*, 1.2mM Ma”+).
Cakium-i~de~ndent
GABA release from striatal slices
679
Effect of verapamil on CABA release in the presence and absence of Ca2’ In order to further examine the role of Cal+ in GABA release, studies were carried out in the
presence of verapamil, a Ca2+ channel antagon~st.2~ Verapamif (to PM) enhanced basal GABA e@ux by 28%, while decreasing evoked overflow by 45%. In the absence of Ca”, the drug reduced basal efflux by 45% and decreased electrically-evoked overflow by 85% (Table 3). EJfct qf ~-~~~o~~ridi~e on GABA release in the presence and absence of Ca” Next, the CaZ+ channel was manipulated indirectly by means of the K* channe1 antagonist, 4-AP.4’ By prolonging the action potential, this drug causes the voltage-de~ndent Ca2+ channel to remain open for an extended period. We observed that in the presence of CaZi, 4-AP (10mM) enhanced both the spontaneous efflux (+ 52%) and evoked overflow ( + 78%) of GABA. However, in the absence of Caa+, the drug had no significant impact on either spontaneous or evoked efflux (Table 3). Efect of tetrodotoxin on GABA overtop evoked by K+ Fin&y, we examined the impact of selectively blocking voltage-dependent Na+ channels while opening
voltage-dependent
Ca2+ channels.
This was
Table 2, Effect of excess M$+ on sponianeous electrically-evoked [31-I]GABAefflux
and
Buffer
n
4/B,
sz is,
Normalt
8
Excess Mg:* Excess Mgl’, Ca”-free
4 4
0.82 + 0.02 0.88 & 0.04 0.71 + 0.05
0.78 + 0.02 0.13 & o.oz* O.12 f o&2*
B, and & refer to the estimated value of basal efflux; S, and S, are overflow due to electrical stimulation. Both basal and evoked efflux are expressed as a percentage of tissue titer per 5 min. Shown are the mean & S.E.M. The addition of excess Mg2+ (10 mM) with or without CaZ+ occurred 20 min before the S2. *P < 0.01 when compared with standard buffer (1.25 mM Ca”, 1.2mM Mg’+). Waluea from Table I provided for comparison. Table 3. Effect of verapamil and 4-aminopyridine efflux of [3H]GABA
on the
Buffer
n
WB,
&is,
Standard Control Verapamil 4-AP
8 4 4
0.82 I: 0.02 I .os& 0.03* i .25 & 0.02*
0.78 & 0.02 0.43 _t o.os* 1.36&0.10*
Ca’+-free Control Verapamit 4-AP
8 4 4
1.11 rfro.05 0.61 + 0.10* 1.14i0.23
0.97 _+0.08 0.1.5*0.06* 0.76iO.11
Ca’+ -free conditions (i 1 mM EGTA added) were applied throughout the superfusion and verapamil (1Opm) and 4-AP (10 mM) were administered 20 min before Sz. *P < 0.05 when compared with the corresponding control.
K+ (25 mM) 3=(5&M
Fig. 1. Effect of TTX on [)H]GABA overflow evoked by high K+ in the presence or absence of Ca2+. Excess K’ (25 mM) was added to the buffer for 10 min as indicated (bar). This occurred in the presence (a) or absena: (of of TTX @PM). The superfusion buffer contained a normal Ca2+ concentration (1.25 mM; top panel) or was Ca2+-free and contained EGTA (1 mM; bottom panel). Results are expressed as a percentage of tissue tritium per 5 min and each point represents the mean 2 S.E.M. of six experiments. Note change in scale between the top and bottom panels. *P < 0.05 compared with control. 0-Q control; *--*, Tl-X.
~~ornpl~shed by elevating extra~~Iu~a~ K+ to produce a local depolarization and then examining overflow in the presence of TTX. Elevating K+ to 2SmM elicited a large overflow of GABA. TTX (5 FM) had no signi~cant effect on basal efRux, but reduced GABA overflow by 25% (S, control, 17.2 + 1.7, n = 6; TTX, 12.1 f 1.2, n = 6) (Fig. 1). The response to excess K+ in the absence of external Ca’” was increased over that observed with standard buffer. However, under these conditions, the inhibition caused by TTX rose to 70% (S, control, 26.0 + 8.3, n = 6; TTX. 6.7 f 2.4, n = 6) (Fig. 1). DISCUSSION
We have previously observed that depo~a~zationinduced GABA overf?ow is not blocked by removing external Ca*+; indeed, it is greatly facilitated.?” Moreover, spontaneous GABA efflux is also increased in the absence of external Ca’+. This study was designed to further explore these phenomena.
680
SBERNATH
and
Again, we observed that the overflow of GABA from striatal slices was facilitated by withdrawing Ca” from the medium. However, Ca’+-deficiency had no effect on evoked GABA efflux when Mg2+ concentration was elevated. Mg’+ can block Ca2+ channels,’ presumably explaining its well known ability to antagonize the effect of Ca’+ on transmitter release.‘7.25.40Thus, our results suggest that Ca’+ channels continue to play a role in GABA release even in the absence of external Ca’+. Further support for a role of Ca’.+ channels in Ca”-independent GABA release derives from our observations with verapamil. This drug, which can inhibit Ca2+ flux across excitable membranes,“,2y also depressed the electrically-evoked GABA overhow in Ca’+-free conditions. Similarly, verapamil has been shown to inhibit the overtIow of GABA evoked by excess K + in the absence of external Ca”. I4 Ca’+ channel antagonists, such as verapamil, may interact with sodium channels as well, especially when they are used at higher concentrations.‘~,‘~.4~ However, we have reported that TTX has no effect on basal [3H]GA%A efflux in the presence of either normal or Ca?+-free buffer.’ Therefore, it is more likely that the significant effect of verapamil on GABA release is via Ca’+ channels rather than on voltage-dependent Na+ channels. How might Ca?’ channels play a role in GABA release in the absence of Cal+ itself? One possibility involves the reported ability of Na’ to move through voltage-sensitive Ca2+ channels.3~4~‘2~22 The absence of external Ca’+ should reduce competition for Na+ entry through these channels and, thus, increase the inward flux of Na”. Mgz+ and verapamil, on the other hand, would act to offset this increased Nat flux by blocking Ca’+ channels. The increase in intracellular Nat promoted by removing Ca’+ from the medium could affect GABA release in one of two ways. Firstly. the resulting entry of Na + could release Cal+ from intraneuronal pools through a Na+ /Ca?+ exchange system.‘8.34~4’ This, in turn, would be expected to enhance spon~neous transmitter release.‘.6,39However, although this model provides an explanation for an increased GABA overflow in the absence of external Ca’.‘. it does not explain the reported failure of a Ca’+-free buffer to increase the release of other neurotransmitters.27.33 A second explanation involves the apparent relation between GABA release and the high affinity GABA transport system. In the presence of a normal concentration of Ca”, the high affinity uptake of GABA plays a major role in removing GABA from the synaptic cleft. 7s3’.MIndeed, this may explain the ability of verapamil to enhance the resting outfiow of GABA. since it has been reported that verapamil can inhibit Na+-coupled GABA uptake.19~‘”The driving force for GABA transport, as for a variety of other carrier-mediated transport processes. is the electroInceasing the intrachemical potential of Na + ,‘1.20,3h cellular concentration of Nat by removing external
M.J.
ZIGMOND
Ca” would be expected to reduce or even reverse the action of such a transport system.42 Elsewhere we have noted that nipecotic acid, an inhibitor of this transport, blocks the release of GABA under Ca’+free conditions. Thus, we have suggested that Ca’+-independent GABA release may be mediated by the operation of the GABA transport system in an outward direction.7s8 Removal of external Ca’+ increased s~ntaneous GABA efflux as well as GABA overlIow. The efflux of GABA during depolarization differs from resting efflux in that it is the consequence of explosive ion movements through excitable membranes. Nonetheless, the antagonizing effect of Mgz+ and verapamif on both spontaneous efflux and evoked overflow in the absence of external Ca2+ suggests that both phenomena share a common m~hanism. Our findings may provide some insight into the basis for the discrepancies that exist in the literature regarding the role of Ca2+ in GABA release. In examining this issue, ~nv~tigators have employed a wide variety of approaches to alter Ca’+ concentration. For example, some have lowered the Ca2+ concentration to 0.1 mM, others have withdrawn external Ca2+ entirely (with or without added EGTA), and still others have substituted Mg2+ for Ca2 +. In our experience, although these procedures are superficially comparable with each other, they can produce quite different results. For exampie, we have observed in the present series of experiments that the electrically-evoked overflow of GABA is greatly enhanced during superfusion with a Ca2+-free solution containing EGTA, but is diminished when the Mg2+ concentration is elevated. Moreover, we have previously reported that although removing Ca2 + greatly increases GABA overflow, reducing external Ca’+ to 0.1 mM actually depresses GABA overflow.’ Such apparently contradictory findings regarding the relation between Ca2 + and GABA release may be explicable in terms of the competition between Na+ and other ions for the Ca2+ channel. Inward Nat flux through Ca2+ channels should be greatest in the presence of a Cal+-free solution containing EGTA when competition between Nat and Ca2+ has been eliminated, somewhat lower in the presence of 0.1 mM Ca” + when some competition between Na+ and Ca2+ remains, and nearly abolished in the presence of l0mM Mg*+ when the presence of ions to compete with Na+ has been increased several fold. The predicted changes in Na + flux in response to these manipulations parallels the observed responses of GABA efflux, suggesting a causal relation between these two variables. In the foregoing discussion, we have proposed that GABA can be released through a process that does not require external Ca 2+. On the other hand, our results suggest that GABA can also be released by classical Ca?+-dependent exocytosis and that this latter mechanism predominates in the presence of a normal concentration of external Ca’+. Firstly, Mg’+
~aleium-inde~nden~ and
verapamil
both
decreased
GABA release from striatal slices
electrically-evoked
GAEIA o~erfIow in the presence of CaZ’. Since both Mgz+ and verapamil nels, these results when
present
woufd act to block
can be interpreted
in the external
Ca2’
to suggest
medium,
Cazc
chanthat influx
plays an ~m~ortaut
roie in GABA release. Secondly, although 4AI? had no elect on GABA efRux in the absence of Ca*+, it greatiy increased GABA avertlow in the presence of Cazt . By blocking K” channels, 4-AP enhames the time-course of the action potential, thereby enhancing Ca’+ uptake?.2a4s Though it appears likely that Na’ entry can directly contribute to an increased GABA e&ix under Ca2+ deficient conditions, the question still remains as to whether Cafe-~~de~ndent, ~~~~de~nd~nt GABA release exists under physiological conditions, enters mainly through when Na+ presumably voltage-de~ndent Naf channels. In order to determine how Ca’* and Na+ ion movements contribute to GABA efflux, we examined the impact of an elevated concentration of external K' . Electrical depolarization appears to release transmitter primariiy via an indirect route: ~~oItage-d~~ndent Ha’ channels are opened, initiating an action potentiaf. which ultimately opens Ca2+ channels in the nerve terminals. In contrast, excess K+ seems to act by local depolarization and, as a result, opens voltagedependent Na+ and Ca2* channels directly.io~24Thus, while TTX comoletelv abolished the effects of electrical stimulation on ~a2-~-de~ndent GABA release,’ 1
681
this inhibitor
of Na’ channels only depressed the K+-evoked overffow of GABA by about 25%. since the potassium concentration appiied was high enough to totally depolarize tissues,24 this suggests that Na+ mov~ent through voltage-de~ndent Na+ channels can directly contribute to GABA release3 although this contribution is relatively small, The higher impact of TTX on the Kc-evoked overflow of GABA in Ca2+-free conditions may be related to the enhanced axonal excitability produced by withdrawing Ca2 + 28and this may underline the jrn~ortan~ of voltage-de~ndent Na* channels.
Our ~x~riments suggest that depolarizationinduced GABA reIease can occur by two processes, one which requires an inward movement of Ca2” via voltage-dependent Ca2+ channels. The second process, which can occur even in the absence of external Ca*+, may involve an inward Na+ flux which could act to alter the ~~i~b~urn betwen the inward and outward transport of GABA via a high a~nity GABA pump. ~c~5~le~~~~~~~-W~ thank Drs Ray Ana Wallace and David C. Wood for befpful discussions, and MS ‘Terri L. Komar for assistance in preparing this rrqort. The research was supported in part by grants from the U.S.P.H.S. (NS- 19608, MN-43947 and MH-~~8~.
REFERENCES
f. Adams D, J., Takeda K. and Umbach f. A. (1985) Inhibitors of calcium buffering depress evoked transmitter release at squid giant synapse. J. f&s&L,
Land. 369, 145-159.
2. Agoston D,, Har&tai P. and Magy A. (1983) Elects of a ~a~inopyridine
on calcium ~ove~~~~ and caulk of membrane potential in pinched-off nerve terminals from rat cerebral cortex. J. Neur~c~e~. 41, 745-751. 3. Alrners W., McCluskey E. W. and Palade P. T. (1984) A nonselective caation conductance in frog muscle membrane blocked by micromolar external calcium ions. J. Physiol., Ltxstf. 353, 565-533. 4. Alsners W. and MCeleskq E. W. (t984) Non-selective conductance in cal&m cbm~~ls of frog m~~!e: c&cium selectivity in a single&e pore. 3. Phjlsiaf., Land. 353, 585-608. 5. Arias C. and Tapia R. (1986) Differential calcium dependence of y-aminobutyric acid and acetylcholine release in mouse brain synaptoso~~s. J. ~~~r~~~~e~~47, 396-404. 6. Bemarh S. and Vizi E. S. (1987) Inhibitory e&c$ af ionized free in~~~llular calcium enhanced by Ru~he~~u~ Red and nr-chfanr-farbonykyanidt: phenyl hydrazon on the evoked release of a~ty~cbol~ne. B&&w. ~~~~ac. 36, 3683-3687. 7. Bernath S. and Zigmond M. .I. (1988) Characterization of [3H]GABA release from striatal slices: evidence for a calcium-independent process via the GABA uptake system, Neuroscience 27, 563-570. 8. Bemath S., Keller: R, W. Jr and Zigmond M. J. (1989) Release of endogenous GABA can occur through Ca”+-dependent and Caz*-independent processes. ivewochenr. Rt. I& 439-445. 9. Bixby 1. L. and Spitzer 8. C. ((984) Early d~~e~n~atio~ of vertebrate spinaX neurons in the absence of voltage dependent Cal+ and Na+ influx. I>ati &al. X06,89-96. 10. Blaustein M. P. (1975) Effect of potassium, veratridine and scorpion venam on calcium accumulation and transmittet release by nerve terminals in vim. J. Physiol., Land. 241, 617-655. Ii. Blmstein M. P. and King A. C. (1976) Influence of membrane potential on the ~i~~e~~~en~ uptake of ~~~~a-a~jnobuty~c acid by presynaptic nsrve terminals: experimental observations and theoretical considerations. 3. .~ernbrune Biol. 30, 153-173. 12. Bossu S.-L. and F&z A. (1984) Patch ciamp study of the tetrodotoxin resistant sodium current in group C sensory newones. Neurosci. tetl. 51, 241-246. 13. Bus~a~an~e 3. 0. (1985) Block of Na+-currents by the Ca2’ -antagonist D-600 in human beart ceil semen&, Q&#gers A&. g&Y.P&&J!. 4@3, 225-227. 14. Carvalho C. M., Santos S. V. and Carvaiho A. P. (1986) y-Aminobutyric acid release from synaptosomes as influenced by Ca?’ and Ca*+ channel blockers. Eur. J. Pharmac. 131, l-12. 1.5. Caudill W. L., Bigelow J. C. and Wightman R. M. (19gS) Comparison of release of endogenous dopamine and y-aminobutyric acid from rat caudate synaptosomes. 3eurochem. Res. 16, 3I9-331.
I”7
S. BERNATH and M. J. ZIGMOND
on:
16. Clay J. R. and Shrier A. (1984) Effects of D-600 on sodium current in squid axons. J. Membrune Biol. 79, 211-214. 17. Cotman C. W., Haycock J. W. and White W. F. (1976) Stimulus-secretion coupling processes in brain: analysis of noradrenaline and GABA release. J. Physiol., Land. 254, 475-505. 18. Cunningham J. and Neal M. J. (1981) On the mechanism by which veratridine causes a calcium-independent release of ~-amjnobutyric acid from brain slices. Br. J. Phurmac. 73, 655-667. 19. Erdreich A., Spanier R. and Raham~moff H. (1983) The inhibition of Na-dependent Ca uptake by verapamil in synaptic plasma membrane vesicles. Eur. J. Pharmac. 90, 193-202. 20. Erecinska M. (1987) The neurotransmitter amino acid transport systems. Biochem. Pharmac. 34, 3547-3555. 21. Fleckenstein A. (197 1) Specific inhibitors and promoters of calcium action in the excitation--contraction coupling of heart muscle and their role in the prevention or production of myocardial lesions. In Calcium and the Heurt (eds Harris P. and Opie L.), pp. 1355188. Academic Press, London. 22. Fukuda J. and Kameyama M. (1980) Tetrodotoxin sensitive and tetrodotoxin resistant sodium channels in tissue cultured spinal ganglion neurones from adult mammals, Brain Res. 182, 191-197. 23. Haycock J. W., Levy W. B., Dibber L. A. and Cotman C. W. (1978) Effects of elevated (K’ )0 on the release of nenrotransmitters from cortical synaptosomes: efflux or secretion? 1. Meurochem. 30, it 13-l 125. 24. Hillman H. H. and Mcllwain H. (1961) Membrane potentials in mammalian cerebral tissues in r&o: dependence on ionic environment. J. PhyGol., Lond. 157, 263-278. ‘5. Hubbard J. I., Jones S. F. and Landau E. M. (1968) On the mechanism by which calcium and magnesium affect the release of transmitter by nerve impulses. J. Physiol., Lond. 196, 75-86. 26. Jones R. S. G. and Heinemann U. (1987) Pre- and post-synaptic K+ and Ca’+ fluxes in area CA1 of the rat hippocampus in vitro: effects of Ni*+, TEA and 4AP. &pi Brain Res. 68, 205-209. 27. Katz B. and Miledi R. (1970) Further study of the role of calcium in synaptic transmission. J. Physiol., Lond. 207, 789 -80 I. 28. Keesey J. C.. Wallgren H. and McIlwain H. (1965) The sodium. potassium and chloride ofcerebral tissues: maintenance. change on stimulation and subsequent recovery. Biochem. J. 95, 289-300. 29. Kohlradt M. B.. Baner B.. Krausse M. and Fleckenstein A. (1972) New selective inhibitors of transmembrane calcium conductivity in mammalian myocardial fibers. Studies with voltage clamp technique. Experientia 28, 288-289. 30. Levy W. B. (1973) Stimulus-coupled secretion of y-aminobutyric acid from rat brain synaptosomes. Science 181. 676 -678. 31. Limberger N., Spath L. and Starke K. (1986) Release of previously incorporated g-[‘H]aminobutyric acid in rabbit caudate nucleus slices. J. .~euroehem. 46, 1102-I 108. 32. Liron Z., Roberts E. and Wong E. (1985) Verapamil is a competitive inhibitor of p-aminobutyric acid and calcium uptakes by mouse brain subcellular particles. Life Sci. 36, 321-327. 33. Llinas R., Steinberg I. Z. and Walton K. (1976) Presynaptic calcium currents and their relation to synaptic transmission: voltage clamp study in squid giant synapse and theoretical model for the calcium gate. Proc. natn. Acad. Ski. U.S.A. 73, 29 18-2922.
34. Lockerbie R. 0. and Gordon-Weeks P. R. (1986) Further characterization of [3H]~-aminobutyric acid release from isolated neuronal growth cones: role of intracellular Car+ stores. Neuroscience 17, 1257-1266. 35. Maurin Y., Arbilla S. and Langer S. Z. (1985) Inhibition by yohimbine of the calcium-dependent evoked release of [‘H]GABA in rat and mouse brain shces irt rifro. Eur. J. Pharmac. 111, 3748. 36. Nelson M. T. and Blaustein M. P. (1982) GABA efflux from synaptosomes: effects of membrane potential, and external GABA and cations. J. Membrane Biol. 69, 213-223. 31. Okada K. and Hassler R. (I 973) Uptake and release of 7 -aminobutyricacid (GABA) in slices of substantia nigra of rat. Brain Res. 49, 214-217. 38. Peris. J.. Harris R. A. and Zahniser N. R. (1987) Modulation of y-PHjaminobutyric acid release from rat cortical slices by cc?-adrenoceptors. Neurosis. Lett. 80, 309..3 14. 39. Rahamimoff R. and Atnaes E. (1973) Inhibitory action of Ruthenium Red on neuromuscular transmission. Proc. nom. Acad. SC;. U.S.A. 70, 3613-3616. 40. Richards C. D. and Sercombe R. (1970) Calcium, magnesium and the electrical activity of guinea-pig olfactory cortex iti vitro. 1. Phwiol., Lund. 211, 571 -584. 41. Sandoval M. E. (1980) Sodium-de~ndent efBux of rZH]GABA from synaptosomes probably related to mitochondrial calcium mobilization. J. Neurochem. 35, 915-92 1. 42. Schwartz E. A. (I 987) Depolarization without calcium can release y-aminobutyric acid from a retinal neuron. Science 238, 350-355. 43. Szerb .J. C. (1979) Relationship between Ca-” -dependent and independent release of [‘HIGABA evoked by high Kt . veratridine or electrical stimulation from rat cortical slices. J. Neurochem. 32, 1565-I 573. 44. Szerb J. C. f 1982) Effect of nipecotic acid. a y-aminobutyric acid transport inhibitor, on the turnover and release of y-aminobutyric acid in rat cortical slices. J. Neurochem. 39, 850-858. 45. Thesleff S. (1980) Aminopyridines and synaptic transmission. Neuroscience 5, 1413-1419. 46. Yatani A. and Brown A. M. (1985) The Ca 2+ channel blocker nitrendipine blocks sodium channels in neonatal rat cardiac myocytes. Cir~z~~ati[~n Rex 56, 868-875. (Accepted 27 NotTemher 1989)
