European Journal of Pharmacology, 211 (1992) 195-202 © 1992 Elsevier Science Pubhshers B.V. All rights reserved 0014-2999/92/$05.00

195

EJP 52243

2-(Carboxycyclopropyl) glycines: binding, neurotoxicity and induction of intracellular free Ca 2+ increase M a s a n o r i Kawai

a,

Y o s h i k o H o r i k a w a a, T a k a f u m i I s h i h a r a a, K e i k o S h i m a m o t o b and Yasufumi Ohfune b

Suntory Institute for Btornedtcal Research and b Suntory Instttute for Btoorgantc Research, Shtmamoto-cho, Osaka 618, Japan Received 25 July 1991, revised MS recewed 22 October 1991, accepted 29 October 1991

The excitatory actions of the eight stereoisomers of 2-(carboxycyclopropyl)glycine (CCG), conformationally rigid glutamate analogues, were analyzed for the glutamate receptor subtypes by means of binding assays with rat brain membranes. All CCG isomers inhibited the binding of [3H]3-(2-carboxypiperazine-4-yl)propyl-l-phosphonlc acid ([3H]CPP) to N-methyl-D-aspartate (NMDA) receptors. The (2S,3R,4S) isomer (L-CCG-IV) was the most potent agonlst for the NMDA receptor and its binding potency was 17- and 790-fold higher than that of L-glutamate and NMDA, respectively. The (2S,3S,4R) isomer (L-CCG-III) showed a potent inhibitory activity for [3H]D-aspartate uptake. Further, L-CCG-IV caused a marked increase of lntracellular free Ca 2+ concentration ([Ca2+],) and potent neurotoxiclty In the single rat cerebral cortical neurons in vitro, and both were blocked effectively by the NMDA antagonists. Significant correlations were observed between neurotoxicity and the increase of [CaZ+], and [3H]CPP binding affinity to the NMDA receptor. L-Glutamate; 2-(Carboxycyclopropyl)glycme, NMDA receptors; D-Aspartate uptake; [Ca2+], increase; Neurotoxicity

1. Introduction

Electrophysiological and biochemical studies have revealed the presence of two major types of glutamate receptors in the mammalian CNS (Monaghan et al., 1989), i.e., the ionotropic receptor, which is further divided into NMDA, kainate and a-amino-3-hydroxy5-methyl-4-isoxazolepropionic acid (AMPA) receptor subtypes (Monaghan et al., 1989; Watkins et al., 1990), and the metabotropic receptor (Sladeczek et al., 1985; Sugiyama et al., 1987). The molecular characteristics of the glutamate receptors are not yet well understood. How is L-glutamate recognized by these distinct glutamate receptor subtypes? It is reasonable to assume that a specific conformer (extended or folded form) of L-glutamate is responsible for activating individual glutamate receptor subtypes (Watkins et al., 1990). Glutamate is a conformationally flexible molecule which is capable of binding to all receptor subtypes. In order to gain insight into the conformation-activity relationship of glutamate, we synthesized eight stereoisomers of 2-

Correspondence to: Y. Ohfune, Suntory Institute for Bioorganlc Research, Shlmamoto-cho, Osaka 618, Japan. Tel 81 75.962 1660, fax 81 75.962 2115

(carboxycyclopropyl)glycine (CCG) (Kurokawa and Ohfune 1985; Yamanoi et al., 1988). CCG is a conformationally restricted glutamate analogue in which the cyclopropyl group fixes the glutamate chain in an extended or a folded form; the a-amino acid moiety of CCG may rotate. Electrophysiological studies by the authors and others (Shinozaki et al., 1989a,b; Ishida et al., 1990), using newborn rat spinal cords, have shown that (1) the (2S,3R,4S) isomer (L-CCG-IV) and the (2R,3S,4S) isomer (D-CCG-II) are potent and selective agonists of the N M D A receptor, (2) the (2S,3S,4R) isomer (L-CCG-III) potentiated the response to Lglutamate, and (3) the (2S,3S,4S) isomer (L-CCG-I) is a selective agonist of the m e t a b o t r o p i c receptor (Nakagawa et al., 1990; Ishida et al., 1990). In addition, the (2R,3S,4R) isomer (D-CCG-IV) has been shown to be a potent N M D A agonist by binding studies using [3H]L-glutamate (Pellicciari et al., 1990). These results, suggesting that all CCG isomers are useful as probes for elucidating the conformational requirements of the glutamate molecule for activating its receptor subtypes and as a tool in neuropharmacology, make it important to know the binding characteristics of the series of CCG isomers. Described in this report are the radioligand binding studies on the eight diastereomers of CCG in the rat brain membranes. Effects of the increase in intracellular free Ca 2+ con-

196 centration ([Ca2+],) and neurotoxicity of the CCG isomers in single rat cerebral cortical neurons in vitro are described.

2. Materials and methods

2.1. Materials T h e eight s t e r e o i s o m e r s of 2-(carboxycyclopropyl)glycine (CCG) (the (2S,3S,4S) isomer (LCCG-I), the (2S,3R,4R) isomer (L-CCG-II), the (2S,3S,4R) isomer (L-CCG-III), the (2S,3R,4S) isomer (L-CCG-IV), the (2R,3R,4R) isomer (D-CCG-I), the (2R,3S,4S) isomer (D-CCG-II), the (2R,3R,4S) isomer (D-CCG-III) and the (2R,3S,4R) isomer (D-CCG-IV)) (fig. 1) were synthesized according to the previously reported procedure (Yamanoi et al., 1988). Enantiomeric purity of these compounds (> 99%) was ascertained by high performance liquid chromatography (HPLC) using an optically active column (Daicel Chemical Industries Ltd., Crownpack CR(+), elution with aqueous HCIO 4 (pH 2.0)). [3H]3-(2-Carboxypiperazine-4-yl)propyl-l-phosphonic acid ([3H]CPP) (999 GBq/mmol), [ 3H]a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid ([3H]AMPA)(1021 GBq/mmol), [3H]kainate (2220 G B q / m m o l ) , [3H]N-[1-(2thienyl)cyclohexyl]-piperidine ([3H]TCP) (2220 GBq/mmol) and [3H]D-aspartate (500 GBq/mmol) were purchased from New England Nuclear Research Products (Boston, MA, U.S.A.). Other chemicals were all of the highest purity commercially available from Nacalai Tesque (Japan) and Sigma (U.S.A.). 2.2. Membrane preparation for the binding assay Male Wistar rats were decapitated and the brain tissues were rapidly dissected. The cortex or the whole

brain except for the cerebellum was homogenized in ice-cold buffer for total particulate membrane preparation, or in 0.32 M sucrose for removal of the nuclear fraction, followed by the preparation of a synaptic membrane fraction. The membrane suspension for each binding assay was prepared according to published methods (Murphy et al., 1987; Loo et al., 1987; Honor6 and Nielsen, 1985; Simon et al., 1976). For the study of [3H]CPP binding to the NMDA receptor, the membrane suspension was incubated with 0.04% Triton X-100 at 37°C for 30 min (Murphy et al., 1987). For the study of [3H]TCP binding to phencyclidine recognition sites, the membrane suspension was incubated with 10 mM EDTA at 37°C for 30 min (Loo et al., 1987). Triton X-100 or EDTA was removed by successive cycles of washing and centrifugation and the membrane pellet was stored at - 80°C until use. On the day of the assay, the frozen pellet was thawed to room temperature and was washed and centrifuged three times before resuspension in a buffer suitable for the binding assay. 2.3. Radioligand binding assays Binding assays were performed according to published methods. The [3H]CPP binding assay was carried out at a ligand concentration of 4 nM at 25°C for 20 min (Murphy et al., 1987). The conditions for the other binding assays were as follows: [3H]AMPA, 5 nM, 0°C, 30 min (Honor6 and Nielsen, 1985), [3H] kainate, 5 nM, 0°C, 60 min (Simon et al., 1976) and [3H]TCP, 3 nM, 25°C, 120 min (Loo et al., 1987). Specific binding was determined from the difference in radioactivity in the absence and in the presence of 1 mM L-glutamate (for [3H]CPP, [3H]AMPA, [3H]kainate binding) or 10 /xg 5-methyl-10,11-dihydro-5H-dibenzo[a,d]cyclohepten-5,10-imine (MK-801) (for [3H]TCP binding). IC50 values were determined by logic-log analysis.

HO2C H NH2

H NH2

I

~CO2H L-Glu (folded form)

H02C~ ' ~ ' C 0 2 H L-Glu (extended form)

.~

H

H ~'~CO2H H O 2 C ~ 0 0 2

HO2C" " H~ Z~NH2 L-CCG-I

V

D-CCG-I

NH2

L-CCG-II

Ho20-.dNH2 V

H H ~'CO2H"H CO2H H ~ C O 2

002..o20-

H

H

T

'°02.

D-CCG-II

L-CCG-III

L-COG-IV

D-CCG-III

D-CCG-IV

Fig 1. Chemical structures of eight CCG isomers and L-glutamate.

H

197

2.4. Uptake inhibition Cerebral cortex was homogenized in 10 volumes of 0.32 M sucrose in a glass-Teflon homogenizer. The homogenate was centrifuged at 1000 × g for 10 min, and the synaptosomes were isolated from the supernatant by centrifugation at 12000 × g for 20 min. The synaptosome pellet was resuspended in a physiological salt solution of the following composition (mM): NaC1 140, KC1 3.5, MgC12 1.2, CaC12 1.3, N a z H P O 4 8.5, N a H z P O 4 2.1, glucose 10 (gassing with 95% 0 2 and 5% CO2), pH 7.4. After 15 min preincubation at 37°C, the synaptosome was incubated for 5 min with 0.1 ~ M [3H]D-aspartate. [3H]D-Aspartate uptake was terminated by rapid vacuum filtration through Whatman G F / C glass fiber filters. The filters were washed three times with 3 ml of ice-cold 0.9% NaC1 solution. Specific uptake was determined from the difference in radioactivity in the absence and in the presence of 1 mM D-aspartate. Counting was performed with a scintillation spectrometer.

2.5. Preparation of cultured cerebral corttcal neuron Cerebral cortical neurons from fetal Wistar rat brains (18-20 d gestation) were isolated and cultured as a monolayer according to a reported procedure (Banker and Cowan, 1977). Cerebral neurons, cultured for 2-3 weeks in vitro before use, were placed on a poly-L-lysine coated glass coverslip with a silicon rubber wall (Heraus, Flexiperm) for measurement of [Ca2+],, or on a 35-mm dish (Falcon) for assessment of neurotoxicity. About the 5th day, the culture ceils were treated with cytosine arabinofuranoside (10 /xM) for 24-48 h in order to suppress any growth of glial cells.

2.6. Detectton of increase m [Ca 2 +]1 [Ca2+], was monitored by the ratio of the fluorescence intensities of fura-2 at 340 nm and 380 nm excitations (Grynkiewicz et al., 1985). The cultured cells were loaded with fura-2 by incubation with 5/~M f u r a - 2 / A M , a membrane permeable dye, in a basal salt solution (BSS composition: 130 mM NaC1, 5.5 mM glucose, 5.4 mM KC1, 1.8 mM CaCI 2 and 20 mM H E P E S - N a O H , pH 7.3) for 60 min at 37°C. After the incubation, the coverslips containing the dye-loaded cells were mounted on an inverted fluorescence microscope (Nikon T M O ) and superfused at a constant flow rate of 5 m l / m i n with BSS for at least 10 min before starting fluorescence measurements. [Ca2+] 1 was measured by the use of a calcium measurement system (Argus-100/CA, Hamamatsu Photonics, Hamamatsu, Japan). All experiments were performed at room temperature and excitatory amino acid (EAA) agonists and

antagonists dissolved in BSS at an appropriate concentration were applied for the perfusion medium.

2. 7. Neurotoxic effect Neurotoxicity in cerebral cortical cells was quantified by the procedure reported by Choi et al. (1988) and Koh et al. (1990). The cultured cells were exposed to L-glutamate, NMDA, or CCGs in BSS substituted for culture medium by triple exchange. The day after the start of the exposure of the ceils to the test compounds, neurotoxicity was assayed by morphological examination and by measurement of cytosolic lactate dehydrogenase (LDH) activity in the bathing medium. The L D H activity measured was corrected for background activity and was expressed as a percentage of the activity of L D H in matched control cell cultures exposed to 1 mM L-glutamate.

3. Results

3.1. Binding of CCG tsomers to glutamate receptor subtypes The characteristics of the binding of the eight CCG isomers to the glutamate receptor subtypes were determined by radioligand binding techniques, using rat brain membrane preparations. NMDA, A M P A and kainate receptors were labeled with [3H]CPP, [3H] A M P A and [3H]kainate, respectively. The [3H]CPP binding inhibition curves for L-glutamate, N M D A and CCG isomers are shown in fig. 2A, and their ICs0 values are in table 1. Of the four L-CCG isomers, L-CCG-IV most potently inhibited [3H]CPP binding and its ICs0 value is 19 nM, which corresponds to 17-fold and 790-fold higher affinity for N M D A receptor than L-glutamate (ICs0 = 0.327 ~ M ) and N M D A (ICs0 = 15.0 /zM), respectively. Of the four D-CCG isomers, the affinity of D-CCG-II and D-CCG-IV (ICs0 = 0.412/zM and 0.215 ~M, respectively) was similar to that of L-glutamate. The other three L- and the two D-CCG isomers showed less inhibition of [3H]CPP binding (ICs0 > 10 ~M). [3H]TCP was used as a functional probe to determine whether the CCG isomers functioned as N M D A receptor agonists or antagonists. As shown in fig. 3A, L-glutamate and the CCG isomers enhanced [3H]TCP binding in a concentration-dependent manner. L-CCGIV caused the most marked enhancement of [3H]TCP binding. The concentration-response curve for Lglutamate (Foster and Wong, 1987) and L-CCG-IV reached a plateau at 1 ~ M and 0.1 ~M, respectively. Ranking of the isomers with regard to [3H]TCP binding enhancement was as follows: L-CCG-IV > D-CCGIV, L-glutamate, D-CCG-II >> other CCG isomers. The

198 (~) NMDA L-glutamate O L-CCG-I A L-CCG-II V L-CCG-III [ ] L-CCG-IV • D-CCG-I • D-CCG-II • D-CCG-III • D-CCG-IV

100 ca ~" '10 f-

8O

'R

6o

13. ~.

4O

t.J

o E

soo

"6

600

(~) O A V [] • • • •

O O

v ca t-

4O0

'15 t." •R

200

L-glutamate L-CCG-I L-CCG-II L-CCG-III L-CCG-IV D-CCG-I D-CCG-II D-CCG-III D-CCG-IV

O.

l.-

(~) Qulsqualate L-glutamate O L-CCG-I A L-CCG-II V L-CCG-III [] L-CCG-IV • D-CCG-I • D-CCG-II • D-CCG-Ill • D-CCG-IV

100

'N 60

g

o

3

ca ° •7 4

9

8 7 6 5 Compound -log (M)

4

B

6

ca ........

"O

ca

~

60 40

0

9

8 7 6 5 Compound -log (M)

4

Fig 2. Effect of L-glutamate, NMDA, kainate, qulsqualate and CCG isomers on [3H]CPP binding to rat cortical membranes (A), [3H]AMPA binding to rat cortical membranes (B) and [3H]kainate binding (C) to the rat whole brain membranes The values represent the means_+S.EM of at least three separate experiments, each performed in duplicate

enhancement of [3H]TCP binding was completely reversed by AP5 (data not shown). Glycine and D-serine also enhanced [3H]TCP binding (data not shown). None

5

C

~) Kamate L-glutamate O L-CCG-I A L-CCG-II V L-CCG-III [] L-CCG-IV • D-CCG-I • D-CCG-II • D-CCG-III • D-CCG-IV

100

¢"

.E

.......

4

"D Q.

6

3 ~

1: L-CCG-IV 2" D-CCG-IV

~ ~ 2

7 8

3. L - g l u t a m a t e

j

4 D-CCG-II 5: D-CCG-, 6: L-CCG-I 7: D-CCG-III

O'~-1

UJ

ICs0 for [3H]CPP binding

-log (M)

Fig. 3. Effect of L-glutamate and CCG isomers on [3H]TCP binding to the rat cortical membranes (A) and correlation between IC50 values for [3H]CPP binding and ECso values for [3H]TCP binding (B). The values represent the means _+S.E.M. of at least three separate experiments, each performed in duplicate.

of the CCG isomers inhibited [3H]glycine binding at concentrations up to 100/zM (data not shown). Figure 3B shows the linear correlation between IC50 for inhibition of [3H]CPP binding to the NMDA receptor and ECs0 for enhancement of [3H]TCP binding by the CCG isomers. These results indicate that the CCG isomers are classified as NMDA receptor agonists.

TABLE 1 Effects of eight 2-(carboxycyclopropyl)glyclne

isomers on excitatory amino acid binding sites and [3H]D-aspartate uptake.

Data represent the means_+ S.E.M from at least three separate experiments, each performed in duplicate IC50 and Hill coefficient values were determined from logic-log analysis. Compound

[3H]CPP binding 1C50 ( / z M )

L-Glutamate NMDA Qulsqualate Kainate L-CCG-I L-CCG-II L-CCG-III L-CCG-IV D-CCG-I D-CCG-II D-CCG-III D-CCG-IV

[3H]AMPA binding Hill

IC50 ( / z M )

[3H]Kalnate binding Hill

IC50 ( / x M )

[3H]D-Aspartate uptake Hill

IC50 ( / z M )

0'.327 + 0.045

0 67

0 426 + 0.053

0 99

0 362 + 0.045

0.70

33.7 + 2.5

15.0 + 2.5 21.7 __ 8 2 327 _+57 53 3 _+ 8.6 > 100 > 100 0.019_+ 0 004 14 2 _+ 0.8 0 412_+ 0 . 0 5 0 76.9 _+ 7.4 0 215_+ 0.021

0.56 052 0.94 0.82 0.72 0.74 0.65 0.68 0.72

. . 0020_+0.003 8 8 9 _+0.35 100 100 100 2 37 _+0.11 48 1 _+9 9 100 100 29.4 _+ 6 0

. 1 11 0.74 0.95 0.93 0 96

. . 0.131 _+ 0.015 0 0 0 9 6 _ + 0.0001 100 100 26.1 _+ 2 8 1.67 _+ 0.19 36.1 _+ 10.8 100 100 42.1 _+ 19 8

0.90 0.93 0.88 0.66 0 61 0.52

100 100 24.5_+ 4 2 100 101 _+7 100 100 100

> > >

> >

> >

> >

> > > > > >

199 -3

All CCG isomers exhibited less inhibitory activity than did L-glutamate on [3H]AMPA and [3H]kainate binding to the AMPA and kainate receptor subtypes (figs. 2B and 2C). Of the eight CCG isomers, L-CCG-IV showed the highest affinity (IC50 = 2.37/zM and 1.67 /zM, respectively), but this was 5- to 6-fold less than that of L-glutamate (IC~0 = 0.426 /zM and 0.362 /zM, respectively) (table 1). D-CCG-II had no effect (IC5o > 100/zM) on either [3H]AMPA or [3H]kainate binding.

A Qutsqualale o

O

-5

O

NMDA

o "E

F ~-

-6

m

~- ~

-7

y--064329+11735x R^2=0988

3.2. The effects of CCG tsomers on [3H]D-aspartate uptake L-Glutamate released from the nerve terminals has been shown to be rapidly taken up into neuronal or glial cells (Logan and Snyder, 1971). D-Aspartate is also rapidly taken up into these cells in a manner similar to that for L-glutamate (Davies and Johnston, 1976). Since D-aspartate is a metabolically stable substrate of the glutamate uptake systems, we examined the effect of CCG isomers on [3H]D-aspartate uptake into rat cortical synaptosomes. Of these, only L-CCGIII potently inhibited [3H]D-aspartate uptake (table 1). Its inhibitory activity (ICs0 = 24.5 /zM) was a little greater than that of L-glutamate (IC50 = 33.7/zM) and D-aspartate (IC50 = 40.5/xM). The other CCG isomers were ineffective (ICs0s > 1 mM).

Kamate O

-4

=

-9 -8

g o

=

I

~

-5

O Kamate

o

N

I

-6

B

-s

o ~ -6

~d

I

-7

L-CCG-IV

o

a5 _

-7

~ 7

-8

o

-9

L-glutamate

o Qulsqualate

-7

8

-6

-5

g g

-s

C o

3.3. The increase in [Ca 2 +], induced by CCG tsomers

L-CCG-IV

o O

Most of the CCG isomers induced a sustained and monophasic increase in [Ca2+], (fig. 4). L-CCG-IV proved to be the most potent to increase [Ca2+],. The threshold concentration of L-CCG-IV was about 0.01 /zM. The ECs0 for L-CCG-IV (the response to 100/zM NMDA was regarded as Emax) was 0.06 p.M, approximately 20- and 300-fold lower than that for L-gluta15

A O

~0 IL 0

"~05

• • • •

L glutamate Qu=squalale NMDA Kamate

O A V []

L-CCG I L CCG-II L CCG III L-CCG-IV

® /'~

D D D D

[]

00

-9

-8

-7

-6

-5

CCG CCG CCG CCG

i II IH IV

-4

log [Agonist Concentratmn (M)]

Fig. 4 Response of [CaZ+], to L-glutamate, N M D A , kamate, qmsqualate and C C G Isomers in cultured rat cerebral cortical neurons: [Ca2+], was estimated from the ratio of the fluorescence intensities of fura-2 at 340 and 380 n m exeltatmns (F340/F380). For this and subsequent figures, the values of (F340/F380)EA A after additmn of E A A s were corrected for (F340/F380)r~stlng m resting cells, i e , A (F340/F380) = (F340/F380)EA A - ( F 3 4 0 / F 3 8 0 ) r ~ t m g (mean + S.E.M, n greater than 7).

L-glutamate

Qu~squalate

g_ -6

o Kamale

"-~

-9

I

-7

=

I

-6

I

I

-5

ECs0for [Ca2+]1 Increase m Cortical Neurons log [ Agomst Concentrat=on (M) ]

Fig. 5. The IC50 for Inhibition by L-CCG-IV and other E A A s of [3H]CPP binding (A), [3H]AMPA binding (B) and [3H]kalnate binding (C) in the rat cerebral cortical m e m b r a n e s were plotted versus ECso for stimulation of [Ca 2+], increase in rat cerebral cortical neurons. For calculation of ECso, the response to 100 /xM N M D A was taken as Emax.

mate (ECs0 = 1.2 /xM) and NMDA (ECs0 = 19 /xM), respectively. The increase in [Ca2+]t by D-CCG-II and D-CCG-IV was more than that with NMDA but less than with L-glutamate. The other CCG isomers showed less [Ca2+], increase even at higher concentrations. Figure 5A shows the linear correlation (R = 0.99) between ICs0 for inhibition of [3H]CPP binding to the NMDA receptor and ECso for [Ca2+], increase by L-CCG-IV, L-glutamate and NMDA. On the other hand, the ECso for [Ca2+], increase by L-CCG-IV did not correlate with the IC5o for [3H]AMPA or [3H]kainate binding (figs. 5B and 5C).

200 In order to determine whether the mobilization of [CaZ+], by L-CCG-IV resulted from activation of the N M D A receptor, we examined the effects of N M D A antagonists such as AP5 (10/zM), MK801 (1 /xM) and Mg 2÷ (3 mM) on L-CCG-IV-induced increase in [Ca2÷],. The sustained increase of [Ca2÷], induced by L-CCG-IV (0.1 /xM) or N M D A (10 /xM) was almost completely reversed by a subsequent addition of any of the N M D A antagonists (fig. 6A). The responses induced by quisqualate (1 /zM) or kainate (10/xM) were not blocked by the N M D A antagonists (data not shown). La 3+ (10 /zM), a voltage-dependent Ca 2+ channel blocker, had little effect on the L-CCG-IV-induced (0.1 /zM) increase in [Ca2+], (data not shown). In the presence of AP5 (100/zM) and Mg 2+ (3 mM), the concentration-dependent curve for L-CCG-IV on [Ca2+], was shifted to the right, so that the L-CCG-IV induced [Ca2+], increase was observed at 10/xM. Under these conditions, the [CaZ+], response to N M D A (10 and 100 /.~M) was almost undetectable (fig. 6B). The [Ca2+]1 response-concentration curves for kainate and quisqualate were not affected by N M D A antagonists such as AP5 and Mg z+ (data not shown). Therefore, L-CCG-IV is a potent N M D A agonist. All the D-isomers of CCG were classified as N M D A agonists in the same manner as above.

A 05

L-CCG-IV 10-7 M • ~ ~

/"

O:CuVVe -r S

50

-7

B

-6 -5 -4 -3 log [ Agomst Concer~tratlon(M) ]

so -~ -------3

40 30 "r

o, 20 10 h,-.--J . . . . . . . . . . .

•. . . . . . . . .

10-6M L-CCG-IV 10"5M MK801 10-5M MK801 + 10-6M

L - C C G IV

Fig. 7. Neurotoxlclty of L-CCG-IV and L-glutamate m cultured cerebral cortical neurons Neurotoxlclty as a function of log [agomst concentration (M)] Matched cultures were exposed to L-CCG-IV or L-glutamate for 24 h (A) Protection from the neurotoxlc effects of L-CCG-IV by MK801. Matched cultures were exposed to L-CCG-IV (1 /xM) alone, MKS01 (10 /xM) followed by L-CCG-IV (1 /xM) or MK801 ( 1 0 / z M ) alone (B) In both (A) and (B), neuronal injury was quantified by the appearance of lactate dehydrogenase (LDH) in the bathing m e d m m L D H release (mean + S.E M., n greater than 3) is expressed as a percentage of that produced by exposure of s]ster cultures to L-glutamate (1 mM) for 24 h. The asterisk indicates a statistically slgmflcant difference (P less than 0.05, S t u d e n t - N e w m a n - K e u l s test)

3.4. Neurotoxic effects AP5 10"5M MK801 10-5M

Mg2+3xl0-3M

NMDA 10-5 M

AP5 10-5 M

MK801 10-5 M

Mg2+3xl0"3M - - 10sec

g

10 C) L-CCG IV I L-CCG'IV+ 3 x103 M Mg2*+10-4M AP5 / / / ~ /~ NMDA /

oo

loo

•

~ o

O

A

,

0~,

-9

,

,

,

(~

~

~

-8 -7 -6 -5 log [ Agontst Concentration (M) ]

[ ~

'

Among the CCG isomers, L-CCG-IV exhibited the most potent neurotoxicity, both morphologically and biochemically. The ECs0 values of the CCG isomers for L D H efflux were as follows: L-CCG-IV (2.6/zM) < D-CCG-II (3.4/xM) < D-CCG-IV (5.1 /xM) < D-CCGI (32 / z M ) < L-CCG-II, L-CCG-III, L-glutamate, N M D A (130 / x M ) < L-CCG-I (260 /xM). Figure 7A shows that the potency of L-CCG-IV for inducing L D H effiux was about 50 times greater than that of L-glutamate. Prior exposure of the cells to MK801 (10 /zM) prevented the morphological damage (data not shown) and L D H efflux caused by L-CCG-IV (1 /zM) (fig. 7B).

-4 4.

Fig 6 Reversal of the [Ca 2+ ], increasing effect of L-CCG-IV by subsequent apphcatlon of N M D A antagonists (A). [Ca2+]~ vs log[agomst concentration (M)] for L-CCG-IV and N M D A as agomsts m the absence and presence of AP5 and Mg 2+ (B) Antagonists were added to the perfusion m e d m m at least 1 m m prior to agonist application and were present untd the end of the agomst action

Discussion

It was found that all CCG isomers had a greater affinity for binding to the N M D A receptor than to the AMPA and kainate receptors in the rat brain membranes. Of the eight CCG isomers, L-CCG-IV, a folded

201 isomer, proved to be the most potent agonist of the N M D A receptor. Two of the D-isomers (D-CCG-II and D-CCG-IV) showed almost the same affinity as L-glutamate for binding to the N M D A receptor. DCCG-II was the most selective N M D A agonist since it had little affinity for the A M P A and kainate receptors. The other isomers and the known N M D A agonists (Watkins and Olverman, 1987; Madsen et al., 1990) had less affinity ( K l > 10/xM) for the N M D A receptor. Electrophysiological experiments using newborn rat spinal cord showed D-CCG-II to have a greater depolarizing activity than L-CCG-IV (Shinozaki et al., 1989b). D-CCG-IV has been shown to be the most potent N M D A agonist as estimated from [3H]Lglutamate binding (Pellicciari et al., 1990). In the present experiments using [3H]CPP binding, however, the binding affinities of D-CCG-II and D-CCG-IV for the N M D A receptor were found to be much less than that of L-CCG-IV. The marked enhancement of [Ca2+], in cultured rat cerebral cortical neurons by L-CCG-IV also shows that it is a potent N M D A agonist. Compared to L-glutamate, the other CCG isomers and EAAs, L-CCG-IV was the most potent to induce a [Ca2+], increase. The mobilization of [Ca2+], was prevented by prior treatment with N M D A antagonists such as MK801, AP5 and Mg 2+. These results suggest that the [Ca2+], increase caused by L-CCG-IV was mainly due to the activation of the N M D A receptor. Both D-CCG-II and D-CCG-IV were less potent than L-glutamate, and the other CCG isomers showed little effect on [Ca:+], increase. Furthermore, L-CCG-IV was the most potent to induce neurotoxicity, which was largely prevented by MK801. This suggests that neurotoxicity is induced by the activation of the N M D A receptor. It is interesting to note that there are significant correlations between neurotoxicity and the increase of [Ca2+], and the N M D A receptor affinity. It has been reported that activation of the N M D A receptor leads to neurotoxicity as a consequence of a toxic influx of extracellular Ca 2+ (Abele et al., 1990; Choi et al., 1988; Michael and Rothman 1990; Madsen et al., 1990). The present results add further evidence to the above conclusion. L-CCG-IV led to greater neurotoxicity than did Lglutamate compared to the increase of [Ca2+], or binding affinity to N M D A receptor. It is tentatively proposed that these differences are mainly due to the uptake process for L-glutamate (Logan and Snyder, 1971); L-CCG-IV, an exogenous amino acid, is taken up slowly into the synaptic environments as compared to L-glutamate. Only L-CCG-III inhibited [3H]D-aspartate uptake into the rat cerebral cortical synaptosomes. Electrophysiological experiments showed L-CCG-III to cause a marked potentiation of the depolarizing response to

L-glutamate and D- and L-aspartate in the isolated spinal cord of newborn rats (Shinozaki et al., 1989b). The present experiments suggest that the potentiation of the L-glutamate response by L-CCG-III results from inhibition of L-glutamate uptake in the synaptic environments. Of the four L-CCG isomers, L-CCG-I and II closely mimic the extended conformation of L-glutamate, while L-CCG-III and IV mimic the folded one, except the rotamers around the a-amino acid moiety. Since the affinity of L-CCG-IV for the N M D A receptor was much greater than that of the extended isomers, it is suggested strongly that L-glutamate activates the N M D A receptor in its folded conformation. Of the folded isomers, L-CCG-III had far less binding affinity to the N M D A receptor than did L-CCG-IV. The cyclopropane ring of L-CCG-III probably sterically hinders its binding to the receptor surface. No conformationactivity relationships could be detected among the four D isomers, because both the extended and the folded isomers of D-CCG showed similar affinities for the N M D A receptor. This may be taken as additional support for the suggestion that the N M D A receptor tolerates acidic amino acids with the D-configuration better than do other receptor subtypes (Watkins et al., 1990). The extended isomers (L-CCG-I and II) showed much less affinity for binding to the N M D A receptor and less activity with regard to mobilization of [Ca2+]~ increase and neurotoxicity. It has recently been found that these isomers show preferential affinity for the metabotropic receptor (Nakagawa et al., 1990; Ishida et al., 1990). The metabotropic receptor itself may not contribute very much to glutamate neurotoxicity, although it might contribute somewhat through associative action with other receptor subtypes (Dumuis et al., 1990). Thus, each CCG isomer has been shown to possess characteristic affinity for the different glutamate receptors and is expected to be a useful pharmacological tool for analyzing molecular mechanisms underlying glutamate receptors.

Acknowledgements We thank Drs. H. Shmozakl and M Ishida, The Tokyo Metropohtan Institute of Medical Science, for valuable suggestions and discussions This work was supported m part by a Grant m Aid for Scientific Research from the Ministry of Education, Science and Culture, Japan

References Abele, A E, K.P. Scholz, W K Scholz and R.J. Mdler, 1990, Excitotoxtclty induced by enhanced excitatory neurotransmlsslon m cultured hlppocampal pyramidal neurons, Neuron 2, 413.

202 Banker, G.A. and W.M. Cowan, 1977, Rat hlppocampal neurons in dispersed cell culture, Brain Res 126 397. Choi, D W 1987, Ionic dependence of glutamate neurotoxlclty, J. Neuroscl. 7, 369 Chol, D.W., J Y Koh and S. Peters, 1988, Pharmacology of glutamate neurotoxlclty In cortical cell culture: Attenuation by NMDA antagonists, J. Neuroscl. 8, 185 Davies, L.P. and G A.R Johnston, 1976, Uptake and release of Dand L-aspartate by rat brain slices, J Neurochem 26, 1007 Dumuis, A , J P Pin, K. Oomagari, M. Sebben and J Bockaert, 1990, Arachldonic acid released from stnatal neurons by joint stimulation of lonotroplc and metabotroplc qmsqualate receptors, Nature 347, 182 Foster, A.C and E H F Wong, 1987, The novel anticonvulsant MK-801 brads to the activated state of the N-methyl-D-aspartate receptor in rat brain, Br. J Pharmacol. 91,403 Grynkiewlcz, G., M. Poenle and R.Y. Tslen, 1985, A new generation of Ca z+ indicators with greatly improved fluorescence properties, J. Biol. Chem 260, 3440. Honor6, T. and M Nielsen, 1985, Complex structure of quisqualatesensitive glutamate receptors m rat cortex, Neurosci Lett 54, 27. Ishida, M., H. Akagl, K. Shlmamoto, Y. Ohfune and H. Shmozakl, 1990, A potent metabotroplc glutamate receptor agomst: electrophysiological actions of a conformatlonally restricted glutamate analogue in the rat spinal cord and Xenopus oocytes, Brain Res. 537, 311. Koh, J.Y, M.P. Goldberg, D.H Hartley and D.W. Choi, 1990, Non-NMDA receptor-medmted neurotoxlclty in cortical culture, J. Neuroscl. 10, 693 Kurokawa, N. and Y Ohfune, 1985, The palladium (II)-asslsted synthesis of a-(carboxycyclopropyl)glycIne, two bloactwe amino acids, Tetrahedron Lett 26, 83. Logan, W.J and S H Snyder, 1971, Unique high affinity uptake systems for glycine, glutamlc and aspartlc acids in central nervous tissue of the rat, Nature 234, 297. Loo, P.S., A.F. Braunwarder, J. Lehmann, M Williams and M A Sills, 1987, Interaction of L-glutamate and magnesmm with phencychdine recognition sites in rat brain evidence for multiple affimty states of the phencyclidme/N-methyl-D-aspartate receptor complex, Mol. Pharmacol. 32, 820 Madsen, U., J.W. Ferkany, B.E. Jones, B Ebert, T N. Johansen, T Holm, P. Krogsgaard-Larsen, 1990, NMDA receptor agomsts derived from ibotemc acid. Preparation, neuroexcitation and neurotoxIcity, Eur J Pharmacol Mol. Pharmacol 189, 381 Michael, R.L. and S.M Rothman, 1990, Glutamate neurotoxiclty m vitro' Antagomst pharmacology and intracellular calcmm concentrations, J Neuroscl. 10, 283.

Monaghan, D T., R.J. Bridges and C.W Cotman, 1989, The excitatory amino acid receptors their classes, pharmacology, and distinct properties m the function of the central nervous system, Ann. Rev Pharmacol. Toxicol. 29, 365 Murphy, D.E, J Schneider, C. Boehm, J Lehman and M. Williams, 1987, Binding of [3H]3-(2-carboxyplperazm-4-yl)propyl-l-phosphonic acid to rat brain membranes, a selectwe, high-affinity hgand for N-methyl-D-aspartate receptors, J Pharmacol Exp Ther. 240, 778. Nakagawa, Y , K Saltoh, T Ishihara, M Ishlda and H. Shlnozakl, 1990, (2S,3S,4S)-a-(carboxycyclopropyl)glycine is a novel agontst of metabotropic glutamate receptors, Eur. J Pharmacol. 184, 205 Pelhcclan, R , B Natahnl, M Marlnozzl, J B Monahan and J.P. Snyder, 1990, D-3,4-'Cyclopropylglutamate' isomers as NMDA receptor hgands. Synthesis and enantioselectwe actwlty, Tetrahedron Lett 31, 139 Shmozaki, H., M. Ishlda, K. Shlmamoto and Y Ohfune, 1989a, A conformatlonally restricted analogue of L-glutamate, the (2S,3R, 4S) isomer of L-a-(carboxycyclopropyl)glyclne, actwates the NMDA-type receptor more markedly than NMDA m the isolated rat spinal cord, Brain Res. 480, 355 Shlnozaki, H., M Ishida, K. Shlmamoto and Y Ohfune, 1989b, Potent NMDA-hke actions and potentiation of glutamate responses by conformatlonal variants of a glutamate analogue m the rat spinal cord, Br J Pharmacol. 98, 1213. Simon, J R., J F. Contrera and M J Kuhar, 1976, Binding of [3H]kalmc acid, an analogue of L-glutamate, to brain membranes, J Neurochem 26, 141 Sladeczek, F., J.P. Pin, M. Recasens, J Bockaert and S Weiss, 1985, Glutamate stimulates mositol phosphate formation m striatal neurones, Nature 317, 717 Suglyama, H., I Ito and C Hlrono, 1987, A new type of glutamate receptor linked to mositol phospholipld metabolism, Nature 325, 531 Watklns, J.C and H J Olverman, 1987, Agonlsts and antagonists for excitatory amino acid receptors, Trends Neurosci. 10, 265. Watkms, J C., P Krogsgaard-Larsen and T Honor6, 1990, Structureactivity relationships in the development of excitatory amino acid receptor agonlsts and competitive antagonists, Trends Pharmacol. Scl. 11, 25. Yamanoi, K., Y. Ohfune, K Watanabe, P.-N L1 and H. Takeuchl, 1988, Synthesis of trans- and cis-a-(carboxycyclopropyl)glycme, novel neuroinhibltory amino acids as L-glutamate analogue, Tetrahedron Lett. 29, 1181.