Neuroscience Research, 17 (1993) 241-248 241 © 1993 Elsevier Scientific Publishers Ireland, Ltd. All rights reserved 0168-0102/93/$06.00

NSR 00663

Billionfold difference in the toxic potencies of two excitatory plant amino acids, L-BOAA and L-BMAA:

biochemical and morphological studies using mouse brain slices

K a r n i r e S. Pa i a Susar la K. S h a n k a r b and Vi j aya l akshmi R a v i n d r a n a t h a

Departments of "Neurochemistry and b Neuropathology, National Institute of Mental Health and Neuro Sciences, Bangalore, India

(Received 23 March 1993; revised 28 May 1993; accepted 7 June 1993)

Key words: L-BMAA; L-BOAA; Neurotoxicity; In vitro; Brain slice; Excitatory amino acid; Quinoxalinedione; NMDA receptor; Glutamate receptor

Summary

Plant amino acids /3-N-oxalylamino-L-alanine (L-BOAA, present in Lathyrus sativus) and /3-N-methylamino-L- alanine (L-BMAA, present in Cycas circinalis) have been implicated in the pathogenesis of human neurological disorders lathyrism and amyotrophic lateral sclerosis-Parkinson's dementia complex of Guam (ALS-PD), respec- tively. In view of the conflicting reports that have emerged on the role of L-BMAA in ALS-PD, we reinvestigated the comparative toxicity of L-BMAA and L-BOAA. We report here the potent toxicity of L-BOAA as examined in an in vitro model consisting of sagittal slices of mouse brain. Incubation of sagittal slices of mouse brain with L-BOAA (1 pM) resulted in significant leakage of lactate dehydrogenase (LDH) and potassium from the slices into the medium. Under similar conditions, L-BMAA-induced LDH leakage from the slices into the medium was observed only at very high concentration of the toxin, namely 1 mM. N-Methyl-D-aspartate (NMDA) receptor antagonists ameliorated the toxic effects of L-BMAA, while non-NMDA receptor antagonists (quinoxalinediones) protected against the toxicity of L-BOAA. Incubation of slices with L-BOAA for 1 h resulted in extensive vacuolation and degeneration of neurons in the thalamus and brain stem, and to a lesser extent in the hippocampus and cerebellar nuclei. The large sized neurons appeared to be affected to a greater extent than the smaller ones. The neurons in other areas of the brain also revealed variable degree of degeneration with swelling of axons and dendrites. Thus, the present study demonstrates the potent toxicity of L-BOAA and elucidates for the first time, the billion-fold difference in the concentration of L-BOAA and L-BMAA required to elicit similar toxic response in vitro, using mouse brain slices. The study also demonstrates the selective vulnerability of certain regions of the brain to toxic insult by L-BOAA.

Introduction

Consumption of L-BOAA and L-BMAA, present in Lathyrus sativus and Cycas circinalis, respectively, has been implicated in the aetiopathogenesis of central nervous system disorders, namely lathyrism and amy- otrophic lateral sclerosis-Parkinson's dementia corn-

Correspondence to: Vijayalakshmi Ravindranath, Department of Neurochemistry, NIMHANS, Hosur Road, Bangalore 560 029, India.

plex in Guam (ALS-PD) (Spencer et al., 1987). Al- though conclusive evidence exists for the neurotoxic effects of L-BOAA, both in vitro (Nunn et al., 1987; Ross et al., 1987; Weiss et al., 1989) and in vivo (Spencer et al., 1986) the role of L-BMAA in the pathogenesis of ALS-PD is still controversial (Duncan et al., 1988; Gajdusek, 1990). Morphological changes indicative of degeneration of neurons have been ob- served in humans who consumed the kernels of Cycas circinalis (Spencer et al., 1987) and in mouse cortical cultures exposed to L-BMAA (Nunn et al., 1987; Ross

242

et al., 1987). However, pathological changes similar to ALS have also been reported in primates fed on low calcium, high aluminium and manganese diet with or without cycad supplementation (Garruto et al., 1985, 1989).

In view of the above controversial reports on the involvement of L-BMAA in the pathogenesis of ALS- PD, the present study was carried out to examine comparative toxicity of L-BMAA and L-BOAA, which share structural similarities. Mouse brain sagittal slices were used as an in vitro model to evaluate the toxic effect of the plant amino acids, since our earlier stud- ies have demonstrated the efficacy of this in vitro model for examination of the neurotoxic effects of excitatory amino acids (Pai and Ravindranath, 1991, 1992). Histological and histochemical examination was also carried out to determine the morphological alter- ations in brain slices following exposure to L-BOAA.

Materials and methods

L-BOAA, L-BMAA, CNQX, DNQX and MK-801 were procured from Research Biochemicals Inc., USA, while NMDA and GDEE were obtained from Sigma Chemical Co., USA. NBQX was gift from Novo Nordisk, Denmark. All other chemicals and reagents were of analytical grade.

Swiss Albino mice (3-4 months old, weighing 25-30 g) from NIMHANS Central Animal Research Facility were used in the study. Sagittal slices of uniform thick- ness (500/xm thick) were prepared using an indigenous slicer (Pai et al., 1991). Slices were incubated in artifi- cial cerebrospinal fluid (ACSF, pH 7.4) containing (in mM): NaCI 122, KC! 3.1, CaCI 2 1.3, MgSO 4 1.2, glu- cose 10, NaHCO 3 25, and KH2PO 4 0.4 under an oxygen atmosphere, as described (Pai and Ravin- dranath, 1991, 1992; Pai et al., 1991). Brain slices were incubated in ACSF (1.5 ml ACSF/slice) containing various concentrations of L-BMAA or L-BOAA for 1 hr at 37°C. Following incubation, the leakage of the cytosolic enzyme, lactate dehydrogenase (LDH) and potassium from the slice into the medium was moni- tored as a measure of cell damage (Pai and Ravin- dranath, 1991, 1992). Control incubations without the toxins were also run simultaneously. LDH activity was measured in the ACSF by monitoring the change in absorbance at 340 nm using pyruvate as substrate, in the presence of NADH as described earlier (Pai and Ravindranath, 1991, 1992). A flame photometer was used to measure the concentration of potassium in the

medium after incubation of the slices (Pai and Ravin- dranath, 1991, t992).

Brain slices were preincubated with glutamate re- ceptor antagonists, namely 5-methyl-10,11-dehydro- 5H-benzo(a, 6)-cyclohepten-5,10-imine maleate (MK- 801, 1-100 nM), glutamate diethyl ester (GDEE, 10- 300 /zM) or 6,7-dinitroquinoxaline-2,3-dione (DNQX, 1 ~zM) for 30 min at 37°C. L-BOAA (1 or 10 pM) or L-BMAA (1 mM was then added and the incubations were continued for a further period of 1 h. Slices incubated with antagonists alone served as controls. Quinoxalinediones, namely, 2,3,-dihydroxy-6-nitro-7- sulfamoylbenzo(F)quinoxaline (NBQX) and 6-cyano- 7-nitroquinoxaline-2,3-dione (CNQX) were adminis- tered subcutaneously (30 mg/kg body wt.) to mice and the animals killed 4 h later. Brain slices were prepared from control (vehicle-treated) and quinoxalinedione- treated mice, and incubated with or without L-BOAA (10 pM) for 1 h at 37°C. Statistical analyses were carried out using Student's t-test or one-way analyses of variance (ANOVA) with Duncan's multiple range test, where appropriate.

For morphological studies, frozen sections of 15-/~m thickness were prepared from mouse brain slices (500 /zm thick) incubated with and without L-BOAA (1 pM to 1 /xM). Sections were stained histochemically for LDH. Another set of mouse brain slices incubated under identical conditions were fixed in paraformal-de- hyde-lysine-periodate at 4°C for 24 h and processed for histology. Serial paraffin sections 8-10/~m thick were stained with haematoxylin-eosin, cresyl violet for Nissl bodies in neurons and Luxol fast blue for myelin. Sections were also immunostained with the peroxidase anti-peroxidase technique, using a monoclonal anti- body against phosphorylated, high molecular weight neurofilament protein (200 kDa) isolated from rat spinal cord (SMI-31, Sternberger Mayer Immunochem- icals Inc., Jerrsetteville, MD, USA). The primary anti- body, the secondary antibody (rabbit antimouse IgG) and mouse peroxidase-antiperoxidase complex were used at 1 : 2000, 1 : 40 and 1 : I00 dilution, respectively. The incubations were carried out for 2 h at room temperature and the colour was developed using hy- drogen peroxide and diaminobenzidine. Coded sec- tions were examined for the effect of L-BOAA on neurons, myelin and axons, and compared with control sections incubated in artificial CSF alone for the same duration of time. The histochemical staining intensity and the topographic distribution were evaluated in slices incubated with L-BOAA and compared with the control by studying the sections grossly with naked eye and under low magnification microscopy.

Results

Dose-dependent leakage of LDH from the slice into the medium was noted following incubation of mouse brain slices with various concentrations of L-BOAA (Fig. 1A). Significant leakage was observed when the slices were incubated with L-BOAA (1 pM). LDH activity in the incubation medium increased with in- creasing concentrations of L-BOAA. In contrast, signif- icant leakage of LDH from the slice into the medium was noted only when the slices were incubated with a very high concentration of L-BMAA (1 mM). L- BOAA-induced leakage of potassium from the slice into the medium is shown in Fig. lB. Similar to LDH, the amount of potassium in the ACSF increased signif- icantly as compared to controls when slices were incu- bated with L-BOAA (1 pM). In contrast, significant leakage of potassium was observed only when the slices were incubated with a very high concentration of L- BMAA namely, 5 mM.

Mouse brain slices preincubated with MK-801 (1 nM) for 30 min at 37°C were incubated for a further period of 1 hr after the addition of L-BMAA (1 mM). MK-801 completely protected the slices against L- BMAA-induced LDH leakage (Fig. 2). However, prein- cubation of mouse brain slices with the non-NMDA receptor antagonist, GDEE (300/xM) had no effect on L-BMAA-induced toxicity (Fig. 2). L-BOAA-induced leakage of both LDH and potassium was completely blocked by GDEE (100 nM to 100/zM), while MK-801 (1-100 nM, NMDA receptor antagonist) had no effect on L-BOAA induced toxicity (Table 1).

Brain slices prepared from mice treated with NBQX or CNQX were incubated with and without L-BOAA

120

100

8O

I--- z 60 S

40

20

. . , T _

I

-L

243

BMAA(ImM) + + +

MWS01(InM ) +

GDEE (300pM) - 4-

Fig. 2. Effect of MK-80I and G D E E on L-BMAA induced LDH leakage from the slice into medium. Mouse brain slices were prein- cubated with MK-801 (1 nM) or G D E E (300 tzM) for 30 min at 37°C. L-BMAA (1 mM) was added to one set of slices and the incubations were continued for a further h. Leakage of LDH from the slice into the medium was monitored at the end of incubation period. Values are m e a n + S D (n = 4-6). Asterisks represent values significantly different from respective controls (P < 0.05). LDH activity in medium

containing control slices was 0.4/zmol N A D H ox id i zed /min /ml .

(10 pM) for 1 h at 37°C. Both the quinoxalinediones protected the slices from the toxic action of L-BOAA. L-BOAA-induced LDH and potassium leakage was completely abolished by treatment with NBQX and CNQX. Preincubation of mouse brain slices with DNQX, in vitro, also offered protection against the toxic effects of L-BOAA (Table 2).

The histochemical and histological differences be-

05

~- O45 < E , ' ? ~6~ o.~

o~3 E'-- :x ~ 0~5

A

8.0

7.0

.J

o- 6"0 L,LI E

5.0 ;i I I i i i .i i i i i i

16~ ~69 163 ~1 oi0~ ~69 ~63 ic~I CONCENTRATION (M) CONCENTRATION (M)

Fig. 1. Dose-dependent release of (A) L DH and (B) potassium from the slice into medium following incubation with L -BOAA or L-BMAA. Mouse brain slices were incubated with various concentrations of L-BOAA (e . ) or L -BMAA (r, zx) in ACSF for 1 h at 37°C. Control incubations did not contain any L-BOAA or L-BMAA and were run simultaneously. Activity of LDH (A) and concentration of potassium (B) were measured in ACSF following incubation. Values are expressed as mean + SD (n = 4-6). * values significantly different from

controls (P < 0.05).

244

tween the control and L-BOAA treated slices could be best appreciated when slices were incubated with 1 # M L-BOAA. At lower concentrations, the changes were less apparent. In the frozen sections from the untreated mouse brain slices, the staining intensity and topographic distribution of the histochemical staining for LDH and NADH-ni t roblue tetrazolium reductase were relatively uniform and reproducible, with greater concentration along the grey matter and nuclear areas. Histochemical staining for L D H activity in slices incu- bated with L-BOAA (1 /.tM) revealed a reduction in the intensity of the staining. This was particularly ap- parent to the naked eye in the thalamus, striatum, the colliculi, brain stem and hippocampus (Fig. 3A and B). In the cerebral cortex and cerebellum it was variable. Following incubation with L-BOAA, the histochemical staining for N A D H nitroblue tetrazolium reductase

revealed mild reduction in staining intensity, especially in the thalamus, striatum and brain stem, in compari- son to the control (data not shown).

Closer examination of the paraffin-processed and cresyl violet-and LFB-stained sections revealed striking features of neuronal degeneration in the thalamus (Fig. 4), brain stem, colliculi, striatum and cortex in descend- ing order of severity, following treatment with L-BOAA (1 nM and 1 p,M). The neurons, especially the large sized cells in the brain stem (Fig. 5A and B), thalamus, colliculi and cerebellar nuclei revealed swelling, rari- faction, vacuolation, depletion of Nissl substance and focal disruption of cell membrane. These changes were seen to extend along the dendrites into the neuropil (Fig. 5A). In the brainstem, the neurons in superior olivary complex, the nuclei of" spinal tract of trigeminal, facial, vagus cranial nerves and the magnacellular retic-

A

Fig. 3. Whole mount of 15-p.M thick frozen sections of mouse brain sagittal slice incubated with (A) ACSF alone (control) and (B) L-BOAA ( 1 gM) and stained histocbemically for LDH. Note the reduction in staining intensity in diencephelic nuclei, brain stem, and to a lesser extent in

cerebellum and cerebral cortex; x 10.

245

Fig. 4. Neuronal vacuolation in thalamus in response to L-BOAA toxicity (slices treated with 1 p.M L-BOAA); x 60.

u lar fo rma t ion showed the degene ra t ive changes (Fig. 5 A and B). S imi lar f ea tu res were no ted in the ce rebe l - lar nucle i as well. The rar i fac t ion and vacuola t ion , though less, was also no ted in the py ramida l cells of S o m m e r sec tor and subiculum, and the g ranu le cells of d e n t a t e gyrus (Fig. 6) in the h i p p o c a m p u s and neurons

of the ce reb ra l cortex. The Purkinje cells of ce r ebe l lum were not affected. T h e r e was n o app rec i ab l e myel in loss. However , immunos ta in ing for the neu ro f i l amen t p ro te in as a m a r k e r for axonal changes revea led swelling and d is tens ion of the axons in the b ra in s tem f iber t racts (Fig. 7 A and B).

Fig. 5. Higher magnification showing the large neurons of the (A) magnocellular reticular formation in brainstem and (B) large neurons of the spinal tract of trigeminal nucleus in the brainstem after after treatment with L-BOAA. Note marked swelling of the neuron, distension of the

dendrites and indistinct cell membranes (arrow heads) and relatively well preserved glial elements; x 300.

246

Discussion

The present study demonstrates the potent toxicity of L-BOAA in contrast to L-BMAA, as examined using sagittal slices of mouse brain, in vitro; L-BOAA was toxic at a concentration of 1 pM and the toxicity increased in a dose-dependent manner (Fig. 1).

L-BMAA-induced leakage of LDH and potassium was observed following exposure of slices to very high concentrations, i.e., 1 mM and 5 mM, respectively. Leakage of LDH was observed on incubation with a lower dose of L-BMAA (1 mM) while potassium leak- age was noted at a higher dose of L-BMAA (5 mM). The very low concentration of L-BOAA (1 pM) suffi- cient to bring about toxicity under similar conditions, indicates the high potency of L-BOAA for causing cerebral damage. Under similar experimental condi- tions, the minimum dose of other glutamate agonists (i.e., quisqualate, N-acetylaspartylglutamate, kainate, NMDA and AMPA) required to bring about signifi- cant leakage of LDH from the slices into the medium varied from 0.1 pM to 1 nM (data not shown). L-BMAA was the only glutamate agonist tested that was toxic to brain slices at the very high concentration of 1 raM. L-BMAA is thus the least potent excitotoxin as com- pared to other glutamate agonists.

The higher concentration of L-BMAA required to bring about toxic insult both in vitro and in vivo has been attributed to the lack of a dicarboxylic acid struc-

ture or possible formation of a toxic metabolite prior to acting on glutamate receptor (Spencer et al., 1987). Presence of bicarbonate has been shown to be essential for L-BMAA to exert its toxic action (Weiss and Choi, 1988). Bicarbonate probably forms a complex with L- BMAA which in turn binds to the NMDA receptor (Nunn et al., 1991). In the present study, the ACSF medium contained bicarbonate (25 mM) and hence the high concentration of L-BMAA necessary to cause toxic insult was not due to the lack of bicarbonate.

Plant excitotoxins L-BOAA and L-BMAA exert their toxic effects via specific glutamate receptors. L-BMAA probably mediates its toxic effect by acting through NMDA receptor, while, L-BOAA is thought to exert its effect via the non-NMDA class of glutamate recep- tor (Ross et al., 1987; Weiss et al., 1989). To determine if the toxic action of L-BMAA and L-BOAA observed in the present study was indeed mediated through glutamate receptors, the effect of glutamate receptor antagonists on L-BMAA and L-BOAA- induced toxic- ity was also examined. L-BOAA-induced LDH and potassium leakage was selectively blocked by non- NMDA receptor antagonist, GDEE, while MK-801, the NMDA receptor antagonist had no effect (Table 1). MK-801 selectively protected the slices from L- BMAA induced LDH leakage (Fig. 2), substantiating earlier findings (Ross et al., 1987; Weiss e ta! . , 1989). The effect of the more recently discovered antagonists to non-NMDA receptors, namely, quinoxalinediones

Fig. 6. Granule cells of dentate gyrus in the hippocampus showing swelling and vacuolation of the soma in L-BOAA (luM) treated mouse brain slice; x 240.

247

Fig. 7. Swollen axons (arrows) of the brain stem fibre tracts visualised immunocytochemically using monoclonal antibody to the 200 kDa neurofi lament protein, control (A) and L-BOAA (1 ~zM) (B) treated slice; X 360.

(Garthwaite and Garthwaite, 1989, Sheardown et al., 1990) on L-BOAA induced toxicity has not been hith- erto to reported. Prior addition of DNQX to mouse brain slices, in vitro, ameliorated the toxic action of L-BOAA. However, when CNQX and NBQX were added to mouse brain slices, in vitro, no protective effect was observed (data not shown). Hence, these quinoxalinediones (CNQX and NBQX) were injected

TABLE 1

EFFECT OF G L U T A M A T E R E C E P T O R ANT AGONIS T S ON L-BOAA I N D U C E D L E A K A G E OF L DH F R O M SLICES INTO M E D I U M

Additions LDH leakage (% control)

L -BOAA (1 pM) L-BOAA (10 pM)

1. NIL 112.0+_5.1 * 120.1+8.2 * 2. G D E E (100 nM) nd 105.0_+ 1.1 3. G D E E (10 p.M) 96.5 _+ 6.0 nd 4. G D E E (100 ~z M) 98.5 +_ 6.3 nd 5. MK-801 (1 nM) 115.6+8.3 * 116.8_+8.1 * 6. MK-801 (100 nM) 114.6-+4.9 * 118.1 -+3.3 *

Slices were preincubated with various concentrations (in parenthesis) of glutamate receptor antagonists, namely, G D E E or MK-801 for 30 min at 37°C prior to the addition of L -BOAA (1 pM or 10 pM). Incubations were continued for a further h after which L DH activity was measured in the medium. Data are mean_+ SD (n = 4 -6 slices) and are expressed as % control activity. * Values significantly different from respective controls, nd, not determined.

to mice and brain slices were prepared from treated animals to study the effect of these antagonists on BOAA toxicity. Both NBQX and CNQX completely attenuated the toxic action of L-BOAA, when they were administered in vivo prior to L-BOAA exposure. The protective effect of NBQX during reperfusion injury following ischemia was also observed after in

TABLE 2

EFFECT OF Q U I N O X A L I N E D I O N E S ON L - B O A A - I N D U C E D LDH AND POTASSIUM L E A K A G E F R O M M O U S E BRAIN SLICES

Con- L -BOAA N B Q X + C N Q X + D N Q X + trol L -BOAA L-BOAA L-BOAA

LDH (% control) 100

Potassium (% control) 100

1 1 2 . 7 + 1 . 9 " 98.4+2.7 95.1+5.7 102.4+2.4

111.7_+1.7 * 103.3_+2.1 101.0_+1.9 99.5_+2.1

Mice were injected with NBQX (30 m g / k g body wt.) or CNQ X (30 m g / k g body wt.) subcutaneously and killed 4 h later. Brain slices were prepared from the treated animals and L-BOAA (10 pM) was added to one set of slices and incubated for 1 h at 37°C. Studies with D N Q X were carried out by preincubating mouse brain slices from untreated animals with D N Q X (1 # M ) for 30 min at 37°C. Following preincubation, L -BOAA (1 pM) was added to one set of slices and the incubation was continued for a further period of 1 h. L D H activity and potassium concentration were measured in the medium following incubation. Values are mean_+ SD (n = 4); * values signifi- cantly different from corresponding controls (P < 0.05).

248

vivo administration of NBQX (Sheardown et al., 1990). Thus, both CNQX and NBQX are effective antagonists when administered in vivo, while DNQX is effective when added in vitro. Nevertheless, the present study demonstrates that quinoxalinediones could serve as effective glutamate antagonists and protect against L- BOAA induced brain damage (Table 2). NBQX has been described as a specific antagonist to DL-o~-amino- 3-hydroxy-5-methylisoxazole-4-propionic acid (AMPA) receptor (Sheardown et al., 1990). The present findings thus demonstrate that L-BOAA acts via the AMPA class of glutamate receptors.

The capability of L-BOAA and L-BMAA to dis- place binding of [3H]glutamate to brain membrane preparations has been examined earlier. The IC50 con- centrations for L-BOAA (Ross et al., 1989) and L- BMAA (Richter and Mena, 1989) required to displace [3H]glutamate were 100 /xM and 1 mM respectively. Thus, a 10-fold difference was observed in the IC50 concentration of L-BOAA and L-BMAA. However, in the present study the concentrations of the least toxic doses of these amino acids differed by a billion-fold. This shows that the affinity for glutamate receptors may not always be indicative of excitotoxic potency of the agonist.

The cytological changes observed, especially the swelling of the neuronal perikarya and dendritic ar- borisation are indicative of postsynaptic pathology typi- cal of those associated with glutamate like neuroexci- tant amino acids (Olney et al., 1976; Ross et al., 1987). The axonal swelling observed could be a secondary change. The topographic distribution of cytopathology to thalamus, striatum, brain stem and to a lesser extent the hippocampus and cerebral cortex, suggest differen- tial vulnerability of the mouse brain neurons to the excitotoxin.

Acknowledgements The authors thank NOVO Nor- disk, Denmark, for the generous gift of NBQX. K.S. Pai thanks the CSIR, Government of India, for the award of a Senior Research Fellowship. The authors also acknowledge the encouragement given by Dr. S.M. Channabasavanna, Director, NIMHANS, and Prof. B.S. Sridhara Rama Rao.

References

Duncan, M.W., Kopin, l.J., Garruto, R.M., Lavine, L. and Markey, S.P. (1988) 2-Amino-3 (methylamino)-propionic acid in cycad-de- rived foods is an unlikely cause of amyotrophic lateral sclerosis/ Parkinsonism. Lancet., 2: 631-632.

Gajdusek, D.C. (1990) Cycad toxicity not the cause of high incidence amyotrophic lateral sclerosis/Parkinsonism-dementia on Guam. Kii Peninsula of Japan, or in West New Guinea. In: A.J. Hudson (Ed.), Amyotrophic lateral Sclerosis: Concepts in Pathogenesis and Etiology, University of Toronto Press, T(~ronto, pp. 317-325.

Garruto, R.M., Swyt, C., Fiori, C,E., Yanagihara, R. and Gajdusek. D.C. (1985) lntraneuronal deposition of calcium and aluminium in amyotrophic lateral sclerosis of Guam. l_ancet. 2: 1353,

Garruto, R.M., Shankar, S.K., Yanagihara, R., Salazar. A.M. Amyx, H.L. and Gajdusek, D.C. (19891 Low-calcium, high aluminium diet-induced motor neuron pathology in cynomolgus monkeys. Acta Neuropathol., 78: 210-219.

Garthwaite, G. and Garthwaite, J. (19891 Quisqualate neurotoxicity: A delayed, CNQX-sensitive process triggered by a CNQX-insen- sitive mechanism in young rat hippocampal slices. Neurosci. Lett., 99: 113-118.

Nunn, P.B., Seelig, M., Zagoren, J.C. and Spencer. P.S. (19871 Stereospecific acute neurotoxicity of 'uncommon' plant amino acids linked to human motor-system diseases. Brain Res., 410: 375-379.

Nunn, P.B., Davis, A.J., O'Brien, P., Olney, J.W., Zorumski, C., Price, M.T. and Labruyere, J. (1991) Carbamate formation and neurotoxicity of L-amino acids. Science., 251: 1619-162(I.

Olney, J.W., Misra, C.H. and Rhee, V. (19761 Brain and retinal damage from Lathyrus excitotoxin, [3-N-oxalyl-i.-a,13-diamino- propionic acid. Nature, 264: 659-661.

Pal, K.S. and Ravindranath, V. (1991) Toxicity of N-acctyl-aspar- tylglutamate and its protection by NMDA and non-NMDA re- ceptor antagonists. Neurosci Lett., 126: 49-51.

Pai, K.S. and Ravindranath, V. (19921 Quisqualic acid-induced neu- rotoxicity is protected by NMDA and non-NMDA receptor an- tagonists. Neurosci Lett., 143: 177-180.

Pai, K.S., Shankar, S.K. and Ravindranath, V. (1991) A simple and inexpensive slicer for preparation of brain slices. J. Neurosci. Methods, 37: 209-214.

Richter, K.E. and Mena, E.E. (19891 L-/3-Methylaminoalanine in- hibits [3H]glutamate binding in the presence of bicarbonate ions. Brain Res., 492: 385-388.

Ross, S.M., Seelig, M. and Spencer, P.S. (1987) Specific antagonism of excitotoxic action of "uncommon" amino acid assayed in organotypic mouse cortical cultures. Brain Res., 425: 120-127.

Ross, S.M., Roy, D.N. and Spencer, P.S. (19891/3-N-Oxalylamino-L- alanine action of glutamate receptors. J. Neurochem., 53: 710- 715.

Sheardown, M.J,, Nielsen, E.O,, Hansen, A.J,, Jacobsen, P. and Honore, T. (1990) 2,3-Dihydroxy-6-nitro-7-sulfamoyl-benzo(F)- quinoxaline: a neuroprotectant for cerebral ischemia. Science, 247: 571-577.

Spencer, P.S., Ludolph, A., Dwivedi, M.P., Roy, D.N., Hugon, J. and Schaumburg, H.H. (1986) Lathyrism: evidence for role of the neuroexcitatory amino acid, BOAA. Lancet, 2 (1986) 1066-1067.

Spencer, P.S., Nunn, P.B., Hugon, J., Ludolph, A.C., Ross, S.M., Roy, D.N. and Robertson, R.C. (1987) Guam amyotrophic lateral sclerosis-Parkinsonism-dementia linked to a plant excitant neuro- toxin. Science, 237: 517-522.

Weiss, J.H. and Choi, D.W. (1988)/3-N-methylamino-t:alanine neu- rotoxicity: requirement for bicarbonate as a cofactor. Science., 241: 973-975.

Weiss, J.H., Koh, J. and Choi, D,W. (1989) Neurotoxicity of /3-N- methylamino-L-alanine (BMAA) and /~-N-oxalylamino-L-alanine (BOAA) on cultured cortical neurons. Brain Res., 497: 64-71.