Modulation of calcium channels by metabotropic glutamate receptors in cerebellar granule cells

Neuropharmacology Vol. 34, No. 8, pp. 929-931, 1995 Copyright 0 1995 Elsevier Science Ltd

002th3908(95)00082-8

Pergamon

Modulation Glutamate

Printed in Great Britain. All rights reserved 0028-3908/95$9.50+ 0.00

of Calcium Channels by Metabotropic Receptors in Cerebellar Granule Cells

P. CHAVIS,‘.2*

L. FAGNI,’

J. BOCKAERT’

and J. B. LANSMAN*

‘CNRS UPR 9023, CCIPE, rue de la Cardonille, 34094 Montpellier, Cedex OS, France and 2Department of Pharmacology, School of Medicine, University of California, San Francisco, CA 94143-0450, U.S.A. (Accepted 27 April 1995) Summary-We investigated the mechanisms by which metabotropic glutamate receptors (mGluRs) modulate specific Ca2+ channels in cerebellar granule cells. A large fraction of the current in granule cells is carried by L- and Q-type Ca’ ’ channels (about 26% each), whereas N- and P-type contribute proportionally less to the global current (9 and 15%, respectively). 1-Aminocyclopentane-dicarboxylate (t-ACPD), (2S,3S,4S)-a-(carboxycyclopropyl)-glycine (L-CCGI) and (S)-4-carboxy-3-hydroxyphenylglycine [(S)-4C3HPG], but not L( +)-2-amino-4-phosphonobutyrate (L-AP4) reduced the Ca*+ current amplitude. The t-ACPD-induced inhibition was fully antagonized by (+)-methyl-4-carboxyphenylglycine [( i-)-MCPG] and blocked by pertussis toxin (PTX). These results are consistent with inhibitory response mediated by mGluR2/R3. The use of specific Cal+ channel blockers provided evidence that mGluR2/R3 inhibited both L- and N-type Ca2+ currents. In PTX-treated cells, Glu or t-ACPD, but not L-CCGI or L-AP4, increased the Ca2+ current. Consistent with the activation of mGluR1, the antagonists (+)-MCPG and (S)-4C3HPG prevented the facilitation of Ca2+ current produced by t-ACIPD. The mGluRI-activated facilitation was completely blocked by nimodipine, indicating that L-type Ca*+ currents were selectively potentiated. Keywords-Cultured mouse-patch-clamp.

cerebellar

granule

cells,

metabotropic

Metabotropic glutamate receptors (mGluRs) have been cloned and subclassified in three groups, according to their coupling mechanisms and pharmacological profiles (Nakanishi, 1992; Pin et af., 1993): group-I (mGluR1 and mGluRS), group-II (mGluR2 and mGluR3) and group-III (mGluR4, mGluR6 and mGluR7). 1-Aminocyclopentane-dicarboxylate (t-ACPD) was first characterized as a selective mGluR. agonist (Manzoni et al., 1990; Palmer et al., 1989), but it does not discriminate between the different mGluR groups. On the other hand, (2S,3S,4S)-ol-(carboxycyclopropyl)-glycine (L-CCGI) at micromolar concentration (Hayashi et al., 1992; Ishida et al., 1990; Nakagawa el’ al., 1990; Pin et al., 1994) and L( +)-2-amino-4-phosphonobutyrate (L-AP4) (Nakanishi, 1992; Shigemoto, 1!)93; Thomsen et al., 1992) are specific agonists of group-II and group-III, respectively. Recently, several phenylglycine derivatives have been described as new pharmacological tools for investigating the role of mGluRs. For example, (+)-methyl-Ccarboxyphenylglycine, (+)-MCPG, the active form of MCPG (Jane et al., 1993), competitively antagonizes various actions of t-ACPD [for a review, see Watkins and *To whom all correspondence

should

be addressed.

glutamate

receptors,

calcium

currents,

Collingridge (1994)] and L-APCmediated responses (Manzoni et al., 1995), although it has been found inactive on cloned mGluR4 (Hayashi et al., 1994). Another phenylglycine derivative, (S)-4-carboxy-3-hydroxyphenylglycine [(S)-4C3HPG], displays a potent and a partial agonistic activity on cloned mGluR2 and mGluR5, respectively (Hayashi et al., 1994; Joly et al., 1995), but also acts as a potent antagonist on cloned mGluR1 (Hayashi et al., 1994). Previous studies have shown that activation of mGluR leads to the subsequent inhibition of voltage-sensitive Ca*+ channels in various neuronal preparations [see Pin et al. (1993) for a review]. Group-I mGluRs selectively inhibit L-type Ca2+ currents in neocortical neurons (Sayer et al., 1992) and N-type Ca2+ currents in acutely dissociated CA3 hippocampal pyramidal cells (Swartz and Bean, 1992). t-ACPD inhibits both L- and N-type Ca2+ currents in cultured hippocampal neurons (Sahara and Westbrook, 1993). Group-III mGluRs have been found to block high-threshold Ca2+ currents in olfactory bulb (Trombley and Westbrook, 1992). Although, these studies have shown that mGluRs inhibit Ca2+ channels, in many cases, the precise signalling pathways remain to be elucidated. Cultured cerebellar granule cells express several types 929

930

P. Chavis et al.

of mGluRs. mGluR1 and mGluR4 mRNA are highly expressed in these cells, whereas the mGluR2 mRNA is present at a lower level. The mGluR5 mRNA has not been detected (Prezeau et al., 1994). Cultured cerebellar granule cells also express multiple components of high-threshold Ca2+ currents that can be distinguished by their sensitivity to the specific Ca*+ channel blockers. These blockers are dihydropyridines, o-Conotoxin GVIA, o-Agatoxin IVA and w-Conotoxin MVIIC, which selectively block L-, N-, P- and Q-type Ca*+ channels, respectively. In addition, a component of Ca*+ current, called R-type, is insensitive to all of these blockers (Bossu et al., 1994; Pearson and Dolphin, 1993; Pietrobon et al., 1993; Randall et al., 1993; Zhang et al., 1993). The existence of multiple mGluRs and Ca2+ channel subtypes raises the question of whether different mGluRs regulate the same or different types of Ca2+ channels. We have previously shown that the activity of single L-type Ca2+ channels, recorded from cell-attached patches on cerebellar granule cells, is selectively inhibited by mGluR2/3 (Chavis et al., 1994) and activated by mGluRl/S (Chavis et al., 1995). The aim of the present study was to examine whether mGluRs regulate the activity of other types of Ca*+ currents in these cells. For that purpose, we recorded whole-cell Ca2+ currents and studied the effects of mGluR agonists on specific pharmacological component of the current. We found that cerebellar neurons express L- and Q-type Ca*+ currents in higher proportions than P- and N-type. Only N- and L-type are inhibited by mGluRs. The mGluRs involved in this effect were found to belong to group-II. Treatment of granule cells with PTX abolished the inhibitory response and revealed a selective facilitatory effect of the group-I mGluRs on L-type Ca2+ currents.

METHODS

Tissue culture Cultures of mouse cerebellar granule cells were prepared following the procedure described previously (Van-Vliet et al., 1989). Seven-day-old male mice (C57BL/6: Simonsen) were decapitated, the brain removed and the cerebellum dissected out. After mechanical dissociation of the cerebella, cells were plated at a density of 0.5 x lo6 cells/ml on glass cover-slips precoated with 25 mg/ml poly-L-lysine (Sigma). Cultures were kept in a humidified atmosphere of 5% CQ2-95% air, at 37°C in a medium containing defined minimal essential medium with Earle’s basal salts and 2 mM glutamine (UCSF Tissue Culture Facility). The medium was supplemented with 5% horse serum, 5% foetal bovine serum, 25 mM KC1 and 0.2% glucose (Sigma). Under these conditions, granule neurons, which represented 90% of the cells, were identified by their small (5-10 pm) phase-bright round or oval body and thin long neurites.

Solutions The bathing medium contained (in mM): BaC12 (20 or 5 as indicated in the figure legends), HEPES (lo), tetraethylammonium chloride (20) TTX (3 x 10 - 4), glucose (10) and sodium gluconate (120) adjusted to pH 7.4 with TEA-OH. The osmolarity of the solution was adjusted to 3 15 mOsm with glucose. The mGluR agonists and antagonists were prepared in the bathing medium and the pH of the solution adjusted to 7.4. The intracellular solution contained (in mM): Cs-aspartate (120), MgCl* (3) HEPES (lo), glucose (15) and NaCl(16). EGTA (10 mM), Na2ATP (3 mM) and CAMP (1 mM) were included in the pipette solution in order to minimize Ca2+ current run-down. Similar results were obtained in the absence of CAMP. The pH of this solution was adjusted to 7.4 by adding CsOH. The osmolarity was adjusted to 300 mOsm with CsOH. Electrophysiology Recordings were made from cerebellar granule cells after 8 f 1 days in culture. Glass cover-slips with attached cells were placed in a recording chamber mounted on a Nikon phase-contrast microscope. Currents were recorded under the whole-cell configuration at room temperature. Patch pipettes were made from borosilicate glass with filament (Sutter Inst. Co) coated with Sylgard (Dow Corning Corp.) and the tips fired polished. Pipettes had resistances of 4-6 MR when filled with internal solution and immersed in the bath. Membrane capacitance and 70-90% of the series resistance were compensated electronically. A recording was continued only when the series resistance after breaking into a cell was < 10 Mn. We analysed the results from cells in which the voltage error from current flowing through the series resistance was <3 mV. Currents were evoked by step voltage-clamp pulses of 450 msec duration from a holding potential of - 80 mV to a test potential of - 10 mV. Voltage steps were applied at a rate of once every 7 sec. Current signals were recorded with a LIST EPC-7 amplifier, filtered at 1 kHz with an g-pole Bessel filter and sampled at 3 kHz on a 80386 PC computer. Data were analysed with the pClamp software (version 5.5) of Axon Instruments (Foster City, CA). All records were corrected for linear leak and capacity currents by the P/N subtraction method. Control and drug solutions were applied using a fast perfusion system previously described (Fagni et al., 199 1). Perfusion was achieved by gravity using an g-barrel pipette system. The delay for reaching a drug concentration clamp around the cell was ~30 msec (Fagni et al., 1991). Materials mGluR agonists and antagonists were purchased from Tocris Cockson (U.K.). Nimodipine, w-Conotoxin MVIIC and PTX were purchased from RBI (U.S.A.). o-Agatoxin IVA was purchased from Peptide Institute

Glutamate Inc. (Japan). w-Conotoxin Sigma (U.S.A.).

GIVA

modulation

and TTX were from

R:ESULTS

Pharmacological components of high- threshold currents in cerebellar granule celEs

Ca”

This series of experiments determined the types and the relative proportions of Ca2+ channels present in our cultured cerebellar granule cells. Figure l(A) shows an example of a typical experiment where a granule cell was exposed sequentially to: nimodipine (1 PM; trace 2), cu-Conotoxin GVIA (5 pIti; trace 3), co-Agatoxin VIA (30 nM; trace 4) and w-Conotoxin MVIIC (10 PM; trace 5). The records shown were obtained when the inhibition reached steady-state during the application of each blocker. In this cell, all of the toxins plus nimodipine blocked 72% of the total Ca2+ current. Consequently, 28% of the total current was R-type. As shown in Fig. l(B), the inhibition of the Ca*+ current produced by either cu-Conotoxin MVIIC and w-Agatoxin IVA was complete within 3 and 1 min, respectively. Nimodipine and w-Conotoxin GVl.4, by contrast, inhibited the current much more quickly (3&40 set). Figure l(C) shows that nimodipine and w-Conotoxin MVIIC inhibited roughly similar proportions of the current [26.5 + 1.8% (n = 21)anld25.S f 3.4% (n - 15),respectively]. On the other hand, o-Agatoxin IVA and cu-Conotoxin GVIA depressed the current by 14.8 f 2.5% (n = 15) and 9.2 + 1.3% (n = 15), respectively. This shows that cerebellar granule cells express roughly the same number of L-, Q- and R-type Ca*+ channels, which carry most of the high-threshold Ca*+ current. N- and P-type channels carry a smaller fraction of the current.

of currents

931

by 35 f 2% (n = 5) and 31 .O f 1.7% (n = 20), respectively. The t-ACPD-induced inhibition was fully antagonized by 1 mM of (+)-MCPG. After wash-out of (k)-MCPG, t-ACPD (200 PM) was able to inhibit the Ca2+ current (not shown), indicating a reversible effect of the antagonist. (-i-J-MCPG, by itself, did not alter the Ca2+ current amplitude. L-AP4 (1 mM) had no effect on

120PA 60

msec

B

L

o-CgTx MVIIC o-Agatx IVA w-CgTx GVIA NLMODIPINE

-1501 0

200

Time

400

600

(set)

C w-CgTx MVIIC w-AgaTx

Pharmacological characterization of the mGluRs involved in the inhibition of high-threshold Ca” currents

o-CgTx

IVA GVIA

NIMODIPINE

We found that L-CCGI (1 PM), inhibited highthreshold Ca2+ curren1.s in all the neurons tested [Fig. 2(A, B)]. As shown in Fig, 2(B), the time course of the Ca’+ current inhibition induced by L-CCGI was slow. The inhibition progressed after a delay of approximately 2 min after applying the agonist and reached a plateau after 5 min. The L-CCGI-induced inhibition was irreversible over the lifztime of the patch. Since the amount of inhibition remained constant after washing out the agonist, the possibility of a time-dependent run-down of the Ca2+ current during and after the application of L-CCGI was unlikely. L-CCGI inhibited the Ca2+ current by 41.0 k 3.0% (~1= 15) [Fig. 2(C)]. Other mGluR agonists also irreversibly and slowly inhibited the Ca’+ current by a similar amount. Figure 2(C) shows that Glu (100 PM), in the presence of 50 PM CNQX and 10 ,uM CPP, inhibited the Ca2+ current by 32.2 + 3.0% (n = 7). 50 PM (S)-4C3HPG and 200 FM t-ACPD inhibited the current

%

Of

Inhibition

Fig. 1. Pharmacological characterization of Ca?+ currents in granule cells. (A) The records show Ca*+ currents recorded in control conditions (trace 11, in the presence of 1 PM nimodipine (trace 2), nimodipine + 5 PM o-Conotoxin GVIA (o-CgTx GVIA; trace 3), 30 nM u-Agatoxin IVA (w-AgaTx WA) + nimodipine + w-CgTx GVIA (trace 4) and 10 PM w-Conotoxin MVIIC (co-CgTx MVIIC) + all the preceding blockers (trace 5). The bathing solution contained 5 mM Ba’+. (B) Time-course of the block ofthe CaZ+ current by the indicated channel blockers. Drug applications are indicated by the horizontal bars. Numbers indicate the times at which the representative records in (A) were taken. (C) The graph shows the mean ( * SEM) fractional reduction of the total Ca’+ current after exposure to each of the channel blockers according to the sequence described in (A): nimodipine (1 mM, n = 211, w-Conotoxin GVIA (5 PM, n = 15), w-Agatoxin IVA (30 nM, n = 15) and w-Conotoxin MVIIC (10 PM, n = 15).

P. Chavis et al.

932

A

t-ACPD and, subsequently, the Ca2+ channel blockers as shown in Fig. 3(A). The t-ACPD-insensitive Ca2+ current [Fig. 3(A), trace 21 was not affected by further applications of either nimodipine (1 PM) or o-Conotoxin GVIA (5 PM) [Fig. 3(A), traces 3 and 41, but was depressed by a-Agatoxin IVA [30 nM; Fig. 3(A), trace 51 or o-Conotoxin MVIIC [lo PM; Fig. 3(A), trace 61

*ri

L-

40pA

50msec

A

B -6O-

L-CCGI

z 3 -120-

E

9"

E ; 0

" -160-

: Y P

6 c wd 1

-2001 0

, 100

200

300

Time

400

600

, 600

B w-CgTx w-Agafx

(set)

WA

o-CgTx GVIA NIMODIPINE t-ACPD

C t-ACPD+MCPG L-AP4 (S)-4C3HPG L-CCGI Glutamate t-ACPD 0

10 %

20

Of

30

40

200

Inhibition

the current, indicating that group-III mGluRs were not involved in this effect. Thus, inhibition of the high-threshold Ca2+ currents is triggered by group-II mGluRs. characterization

of the mGluR2/3-sensi-

five Ca-” currents To examine the types of Ca2+ channels that are the target for inhibition by group-II mGluRs, we first applied

600

400

Time

Fig. 2. Inhibition ofwhole-cell Ca2+ current by different mGluR agonists. (A) Ca’+ currents recorded in the absence (I) and presence (2) of 1 PM L-CCGI. The bathing solution contained 20 mM Ba’+. (B) Time-course of the inhibitory effect of L-CCGI on the Ca*+ current. The cell was exposed to 1 PM L-CCGI during the time indicated by the horizontal bar. The records shown in (A) were taken before adding L-CCGI and after the plateau was reached in the presence of the agonist, as indicated by numbers. (C) Mean (_+SEM) fractional reductions of the current after exposure to various mGluR agonists: t-ACPD (200 PM; n = 20) Glu (100 PM) in the presence of CNQX (50 PM) and CPP (10 PM) to block the current through ionotropic channels (n = 7) L-CCGI (1 PM; n = I S),(S)-4C3HPG (50 PM; n = S), L-AP4 (I mM; n = 10) and (+)-MCPG (I mM) + tACPD (200 PM) (n = 5).

Pharmacological

0

50

600

(set)

C o-CgTx o-AgaTx o-CgTx

MVIIC IVA GVIA

NIMODIPINE t-ACPD 0

10 %

2.0

Of

30

40

Inhibition

Fig. 3. t-ACPD inhibits both w-Conotoxin GVIA-sensitive and nimodipine-sensitive Ca’+ currents. (A) Ca2+ currents recorded before (I) and after adding t-ACPD (200 PM) in the absence (2) and presence of nimodipine (1 PM) (3) u-Conotoxin GVIA (5 /*M) + nimodipine (4), w-Agatoxin IVA (30 nM) + the former inhibitors (5), u-Conotoxin MVIIC (10 PM) + the other Ca?+ channels blockers (6). The bathing solution contained 5 mM Ba’+. (B) Time-course of the inhibitory effect of the indicated Ca’+ channels blockers in the presence of t-ACPD. The cell was exposed to the drugs for the time indicated by the horizontal bars. The numbers indicate the time at which the traces in (A) were taken. (C) Mean (_+ SEM; n = 1I) fractional reductions of the total Ca’+ current after exposure to t-ACPD (200 PM) and the Ca*+ channel blockers: nimodipine (1 /cM), u-Conotoxin GVIA (5 PM), w-Agatoxin IVA (30 nM) and w-Conotoxin MVIIC (10 PM); according to the sequence described in (A).

Glutamate

modulation

within 1 and 3 min, respectively [Fig. 3(B)]. The kinetics of inhibition by these toxins were not altered by a prior exposure to t-AGPD [cf. Fig. 3(B) and l(B)]. Figure 3(C) summarizes the results obtained from all the experiments performed, as shown in Fig. 3(A). The mean fractional reductions of the total current induced by w-Agatoxin IVA and o-Conotoxin MVIIC were 15.1 + 2.1 and 26.2 f 2.5%, respectively, which is comparable to that obtained in the absence of t-ACPD [Fig. l(C)]. Thus, t-ACPD appears to have no effect on P- and Q-currents in these cells. Moreover, the total percentage of inhibition [35%, Fig. l(C)] produced by L- and N-type Ca2+ channel blockers was similar to the percentage of inhibition induced by t-ACPD alone [33.2 f 1.5%, Fig. 3(C)]. All these observations indicated that t-ACPD completely blocked both L- and N-type Ca2+ currents, but was ineffective on P- and Q-type Ca2+ currents.

A

!-

40pA

60

Pharmacological

characterization of mGluRs cells Ca’+ current in PTX-treated

-601

Pharmacological

characterization

Ca2+ current

in PTX-treated

of

the

t-ACPD

100

200

Time

which

300

400

(set)

C

We asked whether inhibition of L- and N-type Ca2+ currents by group-II mGluR involved a G-proteincoupled pathway. Overnight preincubation of the cells with PTX (200 ng/ml) completely abolished the inhibition of the Ca*+ current by t-ACPD (400 PM). Treatment with PTX also revealed a reversible facilitatory effect of this agonist on the current in 96% of the cells tested [Fig. 4(A)]. Figure 4(B) shows that the t-ACPD-induced increase in the amplitude of the Ca2+ current was very rapid (within 5 set) and reversible within 3 min of washing-out the agonist. t-ACPD increased the current by 93.0 f 5.3% [n = 30; Fig. 4(C)]. This effect was fully prevented by the competitive antagonists (+)-MCPG (500 PM) and (S)-4C3HPG (350 ,uM). Glu (1 mM) also increased the Ca2+ current amplitude (84.0 + 8.0%, n = 50) [Fig. 4(C)]. Neither L-AP4 (1 mM) (data not shown) nor L-CCGI (1 PM) [Fig. 4(C)] affected the Ca2+ current in these PTX-treated cells, ruling out the involvement of group-II and group-III mGluRs.

facilitated

msec

B

0

facilitates

933

of currents

mGluRs-

cells

Control experiments shLowed that treatment of the cells with PTX did not significantly alter the relative proportion of the different Ca2+ currents. L-, N-, P-, and Q-types accounted for 2~6.3 f 2.2, 9.0 + 1.0, 15.0 f 3.0 and 25.0 + 2.0% of the whole-cell Ca2+ current, respectively (n = 6). Figure 5(A) shows a PTX-treated cell in which nimodipine (‘I ,uM) inhibited 22% of the high-threshold current (trace 2). Subsequent application of t-ACPD failed to increase the Ca2+ current in 6/6 cells tested [Fig. 5(A), trace 3; Fig. 5(B)]. In other cells, neither the N-type Ca*+ channel blocker, w-Conotoxin GVIA (5

MCPG+t-ACPD (S)-4C3HPG+bACPD L-CCGI Glutamate t-ACPD 0

25 %

50

Of

75

100

increase

Fig. 4. Facilitation of the whole-cell Ca*+ current by mGluR1 /R5 agonists in PTX-treated cells. (A) Records of Ca’+ currents before (I), after (2) adding t-ACPD (400 PM) and following wash-out of the agonist (3). Currents were recorded in a 20 mM Ba2+-containing bathing solution. The pronounced inactivation displayed by the records shown in Fig. 4(A), compare to Figs. 1, 2, 3, is not a typical feature of PTX-treated cells. (B) Time-course of the facilitating effect of t-ACPD. The numbers indicate the time at with the records in (A) were taken. The cell was exposed to the agonist for the time indicated by the horizontal bar. (C) Mean (+ SEM) increase of the Ca2+ current after exposure to the indicated drugs: t-ACPD (400 PM; n = 30), Glu 1 mM + CNQX (50 PM) and CPP (10 PM) (n = 50), L-CCGI (1 PM; n = IO), (+)-MCPG (500 PM) + tACPD (400 PM) (n = 7) and (S)-4C3HPG (350 PM) + tACPD (400 PM) (n = 6).

FM; n = 5), the P-type Ca2+ channel blocker, w-Agatoxin IVA (30 nM; n = 5) nor the Q-type channel blocker w-Conotoxin MVIIC (10 PM; n = 5), were effective in preventing the response to t-ACPD. In these cells, the t-ACPD-mediated facilitation following the blockade of N-, P-, and Q-types Ca2+ channels was completely abolished by nimodipine (data not shown). This provided evidence that, in PTX-treated cells, the mGluR-enhanced

934

P. Chavis et al.

Pharmacological characteristics of the mGluRs which modulate Ca2+ currents Our experiments were designed to elucidate which subtypes of mGluRs were involved in the modulation of Ca2+ channels. The following pharmacological evidence strongly suggests that mGluR2/3 mediated the inhibition of Ca2+ current:

L

(1) The very potent and selective agonist of cloned mGluR2, (S)-4C3HPG (Hayashi et al., 1994), inhibited Ca2+ currents at low concentration. (2) L-CCGI was active at 1 ,uM, a concentration that has been shown to selectively block cloned mGluR2 (Hayashi et al., 1992, 1994; Nakanishi, 1992) and mGluR3 (Pin et al., 1994). (3) (f)-MCPG, a competitive antagonist of cloned mGluR2 (Hayashi et al., 1994), fully antagonized the t-ACPD effect. (4) The lack of efficacy of L-AP4 ruled out the involvement of the group-III mGluRs.

40 pA

50

msec

6

E

P

I

NIMODIPINE

-2001 0

100

200

Time

300

400

(set)

Fig. 5. The t-ACPD-facilitated current is carried exclusively by L-type Ca2+ channels in PTX-treated cells. (A) Currents were recorded in a bathing solution containing 20 mM Ba2+ in the absence of drug(l), in the presence of nimodipine (1 PM) (2) and nimodipine + t-ACPD (400 PM) (3). (B) Time-course of the blocking effect of nimodipine on the response to t-ACPD. Horizontal bars indicate the period of drug applications and the numbers the time of the records in (A).

Ca2+ current channels.

was carried

exclusively

by L-type

Ca2+

DISCUSSION

Our results showed that mGluRs activated two distincts signalling pathways in cerebellar granule cells. One pathway is activated by group-II mGluRs and leads to inhibition of Ca2+ current through L- and N-type channels. This pathway is slow, irreversible and blocked by PTX. The second pathway was revealed after treatment of the cells with PTX. This pathway is activated by mGluR1 and causes a large facilitation of the Ca2+ current through L-type channels. The present results were consistent with our previous observations showing facilitation and inhibition of single L-type Ca*+ channels by mGluRs of group-I and group-II, in cell-attached patches (Chavis ef al., 1994, 1995). We now extend these observations by showing that facilitation was restricted to L-type Ca2+ channels while inhibition involved both L- and N-type Ca2+ channels.

The inhibitory pathway mediated by group-II mGluRs may operate under physiological conditions since Prezeau and colleagues (Prezeau et al., 1994) showed that the mRNA for mGluR2/3 is expressed in cultured cerebellar granule cells from mouse. Since there are no ligands with selectivity for mGluR2 over mGluR3, we were unable to examine the relative contribution of these two receptor subtypes to the inhibitory effect of mGluR agonists. Our results also show that facilitation of the Ca2+ currents was mediated solely by mGluR1. This is supported by the following observations obtained after blockade of the inhibitory pathway by PTX: (1) The selective group-II agonist, L-CCGI did not affect Ca2+ currents (2) The group-III ~G~uRs agonist L-AP4 had no effect on the Ca2+ current. Interestingly, although Prezeau and colleagues (Prezeau et al., 1994) detected mGluR4 mRNA in cultured cerebellar granule cells, they were unable to show inhibition of adenylate cyclase by L-AP4. Therefore, the coupling mechanisms and regulatory pathways of this receptor subtype remain to be identified in these cells. (3) (+)-MCPG, a competitive antagonist of cloned mGluR1 (Hayashi et al., 1994), prevented the t-ACPDinduced Ca2+ current facilitation. (4) Facilitation of the Ca*+ current by t-ACPD was antagonized by (S)-4C3HPG, which has been described as a competitive antagonist at mGluR1 (Hayashi et al., 1994; Thomsen and Suzdack, 1993) and as a partial agonist at mGluR5 (Joly et al., 1995). As shown by Prezeau and colleagues the mGluR1 mRNA is the only representative member of group-I mGluRs found in lo-day-old cultured cerebellar granule cells (Prezeau et al., 1994). This finding reinforces the hypothesis that this effect was mediated by mGluR 1.

Glutamate

modulation

Pharmacological characterization of Ca’+ channels present in cerebellar granule cells

In previous studies focusing on the characterization of Ca*+ currents in cerebellar granule cells from rat, L- and N-type currents were found to represent 39% of the total Ca*+ current (Randall et al., 1993). Pietrobon et al. (1993) and Amico et al. (1994) showed that 25% of the current was of the L-type (Pietrobon et al., 1993), while only lO-15% was of the N-type (Amico et al., 1994). These findings are consistent with the results presented here (26 and 9% of L- and N-type Ca*+ currents, respectively). On the other hand, we found that 15% of the current is carried by P-type channels, somewhat more than the 5.5% described by Randa. and colleagues (Randall et al., 1993). Furthermore, we found that only 26% of the current is carried by Q-type Ca*+ channels, which is substantially less than the amount (43%) found by Randall et al. (1993). These discrepancies may represent differences in the age of the cells studied, since it has been shown that Ca*+ channel expression is developmentally regulated. Indeed, Haws et al. (1993) showed dihydropyridine-and o-Conotoxin GVIA-sensitive currents represent the same fraction (50%) of the total current in 2-3 DIV granule cells, while these comprised 26 and 9% of the global current, respectively in our 7-9 DIV cultures. Moreover, differences in the amount of P-type Ca*+ current could arise from differencies in the ionic composition of the pipette solution, which has been shown to influence the P-type Ca*+ current (Pearson and Dolphin, 1993).

of currents

935

effective in blocking the t-ACPD-insensitive Ca*+ current. By contrast, t-ACPD completely occluded the response to the L- and N-type channel blockers. This is consistent with the fact that nimodipine plus w-Conotoxin GVIA inhibited the same fraction of the total current as t-ACPD alone. This indicated that mGluR2/3 inhibited both Land N-type Ca*+ currents. The inhibition of L-type Ca*+ currents by group-II mGluRs is consistent with our previous results on isolated Ca*+ channels (Chavis et al., 1994). It has been reported that L-type Ca*+ channels are also regulated by mGluRs with a pharmacology close to that of group-I, in pyramidal cortical neurons (Sayer et al., 1992). N-type Ca*+ currents are also inhibited by phospholipase C-coupled mGluR in accutely dissociated CA3 hippocampal neurons (Swartz and Bean, 1992) cloned mGluR2 expressed in rat sympathetic neurons (Ikeda et al., 1994) and mGluR4 in mouse retinal ganglion neurons (Rothe et al., 1994). This leaves open the possibility that there are cell-specific pathways through which Glu can depress one type of Ca*+ current via different receptor subtypes. There are few examples of activation of Ca*+ channels by G-protein-coupled receptors. Facilitation of N-type Ca*+ channels is triggered by group-III mGluRs in mouse retinal ganglion neurons (Rothe et al., 1994). The present results corroborate our previous findings (Chavis et al., 1995) and further show that in PTX-treated cells facilitation is triggered by mGluR1 and restricted to L-type Ca*+ currents. These results suggest that different Ca*+ channel types could be up-regulated by distinct pathways, depending on the cell type.

IdentiJication of the Ca-‘+ channel types modulated by mGluRs

Possible mechanisms of coupling between mGluRs and Ca2 + channels

The concentrations of Ca*+ channel blockers used in this study were sufficient to block completely each component of high-threshold Ca*+ current. Thus, micromolar concentrations of dihydropyridine and o-Conotoxin GVIA and low nanomolar concentrations of o-Agatoxin IVA would entirely and selectively block L-, N- and P-type Ca*+ currents (Randall et al., 1993). One potential problem in assigning the contribution of Q-type Ca*+ channels, is that w-Agatoxin IVA at 100 nM partially inhibits these channels (Randall et al., 1993). However, partial block of Q-type Ca2+ current is achieved only with higher concentrations of w-Agatoxin IVA than those used here (30 nM). A second potential problem is the lack of selectivity of to-Conotoxin MVIIC for N- and Q-type Ca*+ channels [5 PM blocks both N- and Q-type Ca*+ currents (Wheehler et al., 1994)]. With our protocol [Figs l(B) and 3(B)], o-Conotoxin GVIA would have inhibited all the N-type current, so that subsequent addition of o-Conotoxin MVIIC would inhibit only Q-type Ca*+ currents. Consequently, it was possible to use these blockers to identify the nature of both the t-ACPD-sensitive and -insensitive Ca*+ currents. Indeed, both o-Agatoxin IVA and o-Conotoxin MVIIC were

It is becoming increasingly clear that Ca*+ channels are modulated by neurotransmitters and hormones via two types of independent pathways. These include a fast (CO.5 set) membrane-delimited G-protein-coupled mechanism and a slower process involving diffusible cytoplasmic messengers. These two mechanisms can be distinguished experimentally by the effect of bath-applied agonist on channel activity recorded from a cell-attached patch [for a review, see Hille (1994)]. There is considerable evidence showing that various neurotransmitter receptors are coupled to N-type Ca*+ channels in the CNS via an inhibitory G-protein which acts in a membrane-delimited manner. On the other hand, evidences also suggest that neurotransmitter receptors are coupled to L-type Ca*+ channels by a Ca*+-dependent pathway involving second messengers [for a review, see Hille (1994)]. Our results with mGluR modulation of L- and N-type Ca*+ currents, in granule cells, are consistent with both membrane-delimited and second messenger-involving pathways, but with clear differences to those previously described. One characteristic of a membrane-delimited pathway is that it is relatively rapid, involving close-range

P. Chavis et al.

936

molecular interactions within the plane of the membrane. We found however that inhibiton of N- and/or L-type Ca2+ channels, recorded in the whole-cell (present study) and cell-attached (Chavis et al., 1994) configurations, is quite slow and inconsistent with the simplest membranedelimited mechanism. The kinetics of inhibition are more consistent with the mobilization of a second messenger. The nature of such a putative messenger, however, remains to be identified. The slow inhibitory pathway is not likely to involve a decrease in intracellular CAMP level resulting from activation of group-11 mGluRs, since the recording electrode contained saturating concentrations of CAMP. Furthermore, the effect was not blocked by inhibitors of either protein kinase A, protein kinase C or protein phosphatases (unpublished data). Inhibition of Ca2+ channels persisted even in the presence of high concentration of EGTA or BAPTA in the pipette, suggesting that global changes in intracellular Ca2+ were not required. A number of studies have previously pointed out an unidentified second messenger which is involved in the inhibition of Ca2+ channels triggered by neurotransmitters and hormones (Hille, 1994). On the other hand, activation of L-type Ca2+ channels in PTX-treated cells was very rapid (5 set). This suggests that the coupling between L-type Ca2+ channels and mGluRs is likely to occur in a membrane-delimited pathway. The finding that L-type Ca2+ channels are also activated in the cell-attached configuration (Chavis et al., 1995) is however inconsistent with a direct facilitatory interaction between Ca2+ channels and a G-protein. One hypothesis that the rapid generation of a long-range second messenger, such as IP3 [for a review, see Kasai and Petersen (1994)], causes an increase in the L-type current. Such a mechanism would be consistent with the results of DeWaard et al. (1992) who showed that IP3 potentiates Ca2+ currents in cerebellar granule cells. Specific experiments are needed to elucidate the precise signalling pathways and the irreversibility of the inhibitory process. Acknowledgements-We thank Drs 0. Manzoni and J.-P. Pin for their helpful comments. We thank Angie Turner-Madeuf for her helpful technical assistance. This work was funded by grants from Bayer (France)/Troponwerke (Germany), DRET (Grant 91/161) and the Human Frontier ScienceProgram (RG 5792B).

REFERENCES Amico C., Marchetti C. and Usai C. (1994) DHP-and agatoxin-sensitive calcium currents in rat cerebellar granule cells. Neurosci. Abstr. 20: 35.16. Bossu J.-L., DeWaard M., Fagni L., Tanzi F. and Feltz A. (1994) Characteristics of calcium channels responsible for voltage-activated calcium entry in rat cerebellar granule cells. Eur. J. Neurosci. 6: 335-344. Chavis P., Shinozaki H., Bockaert J. and Fagni L. (1994) The metabotropic glutamate receptor types 2/3 inhibit L-type

calcium channels via a pertussis toxin-sensitive G-protein in cultured cerebellar granule cells. J. Neurosci. 14: 7067-7076. Chavis P., Nooney J., Bockaert J., Fagni L., Feltz A. and Bossu J. L. (1995) Facilitatory coupling between a glutamate metabotropic receptor and dihydropyridine-sensitive calcium channels in cultured cerebellar granule cells. J. Neurosci. 15: 135-143. DeWaard M., Seagar M., Feltz A. and Couraud F. (1992) Inositol phosphate regulation of voltage-dependent calcium channels in cerebellar granule neurons. Neuron 9: 4977503. Fagni L., Bossu J. L. and Bockaert J. (1991) Activation of a large-conductance Ca’+-dependent K+ channel by stimulation of glutamate phosphoinositide-coupled receptors in cultured cerebellar granule cells. Eur. J. Neurosci. 3: 778-789. Haws C. M., Slesinger P. A. and Lansman J. 8. (1993) Dihydropyridine-and w-conotoxin-sensitive Ca*+ currents in cerebellar neurons: persistent block of L-type channels by a pertussis toxin-sensitive G-protein. J. Neurosci. 13: 11481158. Hayashi Y., Tanabe Y., Aramori I., Masu M., Shimamoto K., Ohfune Y. and Nakanishi S. (1992) Agonist analysis of 2-(carboxycycloproyl)glycine isomers for cloned metabotropic glutamate receptor subtypes expressed in Chinese hamster ovary cells. Br. J. Pharmac. 107: 539-543. Hayashi Y., Sekiyama N., Nakanishi S., Jane D. E., Sunter D. C., Birse E. F., Udvarhelyi P. M. and Watkins J. C. (1994) Analysis of agonist and antagonist activities of phenylglycine derivatives for different cloned metabotropic glutamate receptor subtypes. J. Neurosci. 14: 3370-3377. Hille B. (1994) Modulation of ion-channel function by G-protein-coupled receptors. Trends Neurosci. 17: 53 l-536. Ikeda S. R., Lovinger D. M., McCool B. A. and Lewis D. L. (I 994) Heterologous expression and functional coupling of a metabotropic glutamate receptor (mGluR2) to N-type Ca’+ channels in adult rat sympathetic neurons. Neurosci. Abstr. 20: 375.17. Ishida M., Akagi N., Shimamoto K., Ohfune Y. and Shinozaki H. (1990) A potent metabotropic glutamate receptor agonist: Electrophysiological actions of a conformationally restricted glutamate analogue in rat spinal cord and Xenopus oocytes. Brain Res. 537: 31 l-314. Jane D. E., St Jones P. L., Pook P. C.-K., Salt T. E., Sunter D. C. and and Watkins J. C. (1993) Stereospecific antagonism by (+)-cc-methyl-4-carboxyphenylglycine (MCPG) of (1 S,3R)ACPD-induced effects in neonatal rats motoneurons and rat thalamic neurons. Neuropharmacology 32: 725-727. Joly C., Gomeza J., Brabet I., Curry K., Bockaert J. and Pin J.-P. (1995) Molecular, functional and pharmacological characterization of the metabotropic glutamate receptor type 5 splice variants: comparison with mGluR1. J. Neurosci. In press. Kasai H. and Petersen 0. H. (1994) Spatial dynamics of second messengers: IP3 and CAMP as long-range and associative messengers. Trends Neurosci. 17: 95-101. Manzoni O., Fagni L., Pin J.-P., Rassendren F., Poulat F., Sladeczek F. and Bockaert J. (1990) (trans)-l-Amino-cyclopentyl-1,3-dicarboxylate stimulates quisqualate phosphoinositide-coupled receptors but not ionotropic glutamate receptors in striatal neurons and Xenopus oocytes. Molec. Pharmac. 38: I-6. Manzoni 0. J., Castillo P. E. and Nicoll R. A. (1995) Pharmacology of metabotropic glutamate receptors at the

Glutamate mossy

fiber

synapses

Neuropharmacology

of

the

34: 9(j5-97

ginea

pig

modulation

hippocampus.

1.

Nakagawa Y., Saitoh K., Ishida M. and Shinozaki H. (1990) (2S,3S,4S)a-(carboxycyclopropyl)glycine is a novel agonist of metabotropic glutamate receptors. Eur. J. Pharmac. 184: 205-206.

Nakanishi S. (1992) Molecular diversity of glutamate receptors and implications for brain functions. Science 258: 597-603. Palmer E., Monaghan D. T. and Cotman C. W. (1989) trans-ACPD, a selective agonist of the phosphoinositide coupled excitatory amincm acid receptor. Eur. J. Pharmac. 166: 585-587.

Pearson H. A. and Dolphin A. C. (1993) Inhibition of w-conotoxin-sensitive Ca’+ channel currents by internal Mg2 + in cultured rat cerebellar granule neurones. f’jh’igers Arch.

Ges. Physiol.

425: 518-527.

Pietrobon D., Forti L., Tottene A. and Moretti A. (1993) Functional diversity of dihydropyridine and w-conotoxinGVIA-insensitive calcium channels in rat cerebellar granules. Neurosci. Abstr. 19: 607.7. Pin J.-P., Fagni L. and Bockaert J. (1993) Metabotropic glutamate receptors: targets for new neuropharmacologically active drugs. Curr. Drugs: Neurodegen. Disord. 1: 11I-137. Pin J.-P., Joly C., Heinemann S. F. and Bockaert J. (1994) Domains involved in the specificity of G protein activation in phospholipase C-coupled metabotropic glutamate receptors. EMBO Jll3: 342-348. Prezeau L., Carette J., Helpap B., Curry K., Pin J.-P. and Bockaert J. (1994) Pharmacological characterization of metabotropic glutamate receptors in several types of brain cells in primary cultures. Molec. Pharmac. 45: 570-577. Randall A. D. and Tsien R. W. (1995) Pharmacological dissection of multiple types of Ca’+ channel currents in rat cerebellar granule neurons. J. Neurosci. 15: 299553012. Rothe T., Bigl V. and Grantyn R. (1994) Potentiating and depressant effects of metabotropic glutamate receptor agonists on high-voltage-activated calcium currents in cultured retinal ganglion neurons from postnatal mice. Pfliigers Arch. Ges. PhySloI. 426: 161-170. Sahara Y. and Westbrook G. L. (1993) Modulation of calcium

of currents

937

currents by a metabotropic glutamate receptor invloves fast and slow kinetic components in cultured hippocampal neurons. J. Neurosci. 13: 3041-3050. Sayer R. J., Schwindt P. C. and Crill W. E. (1992) Metabotropic glutamate receptor-mediated suppression of L-type calcium current in acutely isolated neocortical neurons. J. Neurophysiol. 68: 833-843.

Shigemoto R. (1993) Structure, function and localization of the metabotropic glutamate receptor subtypes. J. Neurochem. 61: Sl23C. Swartz K. J. and Bean B. P. (1992) Inhibition of calcium channels in rat CA3 pyramidal neurons by a metabotropic glutamate receptor. J. Neurosci. 12: 43584371. Thomsen C. and Suzdack P. D. (1993) 4-Carboxy-3-hydroxyphenylglycine, an antagonist at type I metabotropic glutamate receptor. Eur. J. Pharmac. (Molec. Pharmac. Sec.) 245: 299-30

I.

Thomsen C., Kristensen P., Mulvihill E., Haldeman B. and Suzdack P. D. (1992) L-2-amino-4-phosphonobutyrate (L-AP4) is an agonist at the type IV metabotropic glutamate receptor which is negatively coupled to adenylate cyclase. Eur. J. Pharmac.

227: 361-362.

Trombley P. Q. and Westbrook G. L. (1992) L-AP4 inhibits calcium currents and synaptic transmission via a G-proteincoupled glutamate receptor. J. Neurosci. 12: 2043-2050. Van-Vliet B. J., Sebben M., Dumuis A., Gabrion J., Bockaert J. and Pin J. P. (1989) Endogenous amino acid release from cultured cerebellar neuronal cells: effect of tetanus toxin on glutamate release. J. Neurochem. 52: 1229-l 239. Watkins J. and Collingridge G. L. (1994) Phenylglycine derivatives as antagonists of metabotropic glutamate receptors. Trends Pharmac. Sci. 15: 333-342. Wheehler D. B., Randall A. and Tsien R. W. (1994) Role of N and Q type Ca channels in supporting hippocampal synaptic transmission. Science 264: 107-l 11. Zhang J. F., Randall A. D., Ellinor P. T., Horne W. A., Sather W. A., Tanabe T., Schwarz T. L. and Tsien R. W. (1993) Distinctive pharmacology and kinetics of cloned neuronal Ca’+ channels and their possible counterparts in mammalian CNS neurons. Neuropharmacology 32: 1075-1088.