Nitric oxide modulates NMDA-induced increases in intracellular Ca2+ in cultured rat forebrain neurons

310

Brain Research, 592 (1992) 310-316 © 1992 Elsevier Science Publishers B.V. All rights reserved 0006-8993/92/$05.00

BRES 18168

Nitric oxide modulates NMDA-induced increases in intracellular in cultured rat forebrain neurons Kari R. Hoyt a, L i a n g - H o n g T a n g b, Elias A i z e n m a n b and Ian J. R e y n o l d s

Ca 2+

a

Departments of '~Pharmacology and b Physiology, University of Pittsburgh School of Medicine, Pittsburgh, PA 15261 (USA) (Accepted 19 May 1992)

Key words: Nitric oxide; NMDA receptor; Intracellular calcium

We studied the effects of nitric oxide (NO) and the NO-releasing agents sodium nitroprusside (SNP), S-nitroso-N-acetylpenicillamine(SNAP) and isosorbide dinitrate (ISDN) on N-methyl-o-aspartate (NMDA)-induced increases in intracellular Ca 2+ ([Ca 2+ ]i), whole-cell patch-clamp currents and on glutamate.stimulated ['~H]dizocilpine binding. NO and agents that release NO partially inhibit increases in [CaZ÷]i at concentrations between 1 /zM and ! raM. These agents also decrease [Ca 2÷ ]i changes produced by kainate and potassium, but to a smaller extent. As the effects of NO are still present following alkylation of the redox modulatory site on the NMDA receptor this action of NO is probably not :t consequence of oxidation of the redox site. in contrast to SNP, ISDN does not inhibit NMDA-induced whole cell patch-clamp currents suggesting that NO modulates [Ca 2+ )~ via perturbation of a Ca 2+ homeostatic process, Furthermore, SNP may have a direct action on the NMDA receptor complex in addition to the generation of NO. 8.Bromo.cGMP does not mimic the inhibitory effect of NO suggesting that this effect is not the result of NO stimulation of neuronal cGMP production. As the production of NO in neurons is dependent on increases in [Ca-'+ Ji associated with NMDA receptor activation, these data suggest thai NO-mediated decreases in [Ca 2+ ]t may represent a novel feedback inhibitory mechanism for NO production in the brain.

INTRODUCTION Nitric oxide (NO) is a short lived, highly reactive messenger molecule that is synthesized in a number of tissues including the brain :~, It is produced by the action of nitric oxide synthase (NOS) on arginine in a Ca2+-dependent fashion Is. The localization of NOS exclusively in neurons in the CNS 4 supports the role of NO as an important intercellular messenger in the brain. For example, Ca2+.dependent NO production is believed to be an essential intermediate in glutamate. induced cyclic GMP production ,~a3. This phenomenon appears to result largely from the activation of the N-methyl-D-aspartate (NMDA) sub-type of glutamate receptor, which is permeable to Ca 2+ (refs. 20,23,28). Studies in our laboratories have characterized the regulation of the NMDA receptor by a wide variety of ligands, including a number of agents that chemically modify a site that is sensitive to sulfhydryl redox reagents3. These studies have shown that oxidation of the NMDA receptor-associated redox site decreases

NMDA responses, while reduction of this site potentiates the effects of NMDA2,x'~:. As NO is an effective sulfhydryl oxidizing agent "~qt we speculated that NO might act as a feedback inhibitor of its own production by oxidizing the redox site on the NMDA receptor and thereby decreasing the Ca2+-dependent synthesis of NO. In this study we demonstrate that NO as well as the NO-generating agents sodium nitroprusside (SNP), isosorbide dinitrate (ISDN) and S-nitroso,N.acetylpenicillamine (SNAP) can decrease NMDA-mediated changes in intracellular free Ca 2+ ([Ca2+]i) in neurons. However, these studies also demonstrate that this effect appears to be independent of the NMDA-receptor associated redox site and independent of cGMP production, MATERIALS AND METHODS PrJ.':~arycell cidture Neurons for [Ca2+]i experiments were cultured from 17-day embryonic rat forebrain isolated as previously described 32. Cells

Correspondence: l.J, Reynolds, Department of Pharmacology, University of Pittsburgh, E1354 Biomedical Science Tower, Pittsburgh, PA 15261, USA, Fax: (I) (412) 648-1945.

311 were plated onto poly-o-lysine coated glass coverslips at a concentration of 3 × l0 s cells/ml in Dulbecco's modified Eagle's medium (DMEM) containing 10% v/~" fetal bovine serum, 100 U / m l streptomycin and 5 p.g/ml penicillin. This medium was replaced 24 h after plating to DMEM with 10% horse serum and antibiotics, and the coverslips were inverted. Cells were kept in a humidified incubator at 370(2 in 95% air/5% CO2 for 12-25 days until [Ca 2+ ]i recordings were made. For electrophysiology studies cells were plated on 12 mm coverslips coated with collagen and poly-L-lysine in DMEM with 10% 1-'-12, I0% heat inactivated iron supplemented calf serum and antibiotics as indicated above. The medium was changed on a Monday-Wednesday-Friday schedule. Cells were treated with 2 / z M cytosine arabinoside after two weeks and used one week later.

Intracelhdar Ca 2 + recordings Microspectrofluorimetric [Ca2+]i recordings of single neurons were made essentially as described 32. On the day the recordings were made, the coverslips were rinsed with HEPES buffered salt solution (HBSS), which contained (raM): NaCI 137, KCI 5, MgSO4 0.9, CaCI 2 1.4, NaHCO 3 3, Na2HPO4 0.6, KH2PO 4 0.4, glucose 5.6, and HEPES 20, pH-adjusted to 7.4 with NaOH. The cells were then incubated in 5 p.M fura-2 AM (Molecular Probes, Eugene, OR) in HBSS containing 5 mg/ml bovine serum albumin for approximately I hour at 37*C. After incubation the cells were rinsed with HBSS and then mounted in a recording chamber. Pretreatment of cells was accomplished by perfusing the indicated solution through the recording chamber to exchange the chamber volume (about 0.5 ml) at least 15 times and allowing the solution to contact the neurons for the indicated amount of time. All pretreatments were washed thoroughly out of the chamber (about 15 volume exchanges) before agonists were applied. Agonist application was accomplished by perfusing the recording chamber with the agonist solution for approximately 30 s followed by pertusion with agonist-free HBSS. Nitric oxide solutions were prepared by bubbling NO gas (Matheson Gas Products, Twinsbur8, OH) through deoxygenated HBSS for 15-311 rain '4. We assumed a saturated NO concentration of approximately 3 mM after this time and made our dilutions based on this concentration. All NO solutions were prepared intmediately before application to the neurons, while other agents were freshly prepared the day of the experiment. Whole.c,,fl patch.clamp Patch-clamp recordings were made as previously described". Cells were continuously superfused at a rate of 0,5 ml/min with a physiological solution based on Hanks salts (composition in raM: NaCI, 137; NaHCO3 I; NaHPO+, 0.34; KCI, 5,36; KH2PO 4, 0,44; CaCI~, 2.5; HEPES, 5; dextrose, 22.2, Phenol red, 0,011 g/I; adjusted to pH 7.2 with NaOH). Five hundred nanomolar TTX and 10/zM bicuculline methiodide were added to decrease synaptic activity, and 1 p,M glycine was included to mostly saturate the glycine modulatory site on the NMDA receptor Is. The intracellular pipette solution contained (in raM): CsCI, 140; MgCI 2, 1; CaCI 2, 1; EGTA, 2.25;

HEPES 10, adjusted to pH 7.2 with concentrated CsOH. NO-releasing drugs were added either by superfusion or by puffer application. Similar results were obtained with either method. NMDA 30 /zM together with 1 ttM glycine were applied by puffer.

Receptor binding assays [3H]Dizocilpine biading assays we performed using well washed rat brain membranes as previously described 31. [3H]Dizocilpine was incubated with membranes, together with glutamate, glycine and drugs at appropriate concentrations for 2 h at room temperature in a volume of 1 ml HEPES-NaOH, pH 7.4. Assays were terminated by vacuum filtration over glass-fiber filters and radioactivity determined by liquid scintillation counting. Materials NMDA, glycine, kainate, sodium nitroprusside, isosorbide dinitrate, N-ethylmaleimide, sodium nitrate, sodium nitrite and 8bromo-cGMP were obtained from Sigma (St. Louis, Me). S-NitrosoN-acetylpenicillamine was kindly provided by Burroughs Welleome (Beckenham, Kent). DTNB was obtained from Calbiochem (La Jolla, CA) and DT'I" was from Boehringer-Mannheim Biochemicals (Indianapolis, IN). [3H]Dizoeilpine (22.5 Ci/mmol) was obtained from Du Pont/NEN (Boston, MA).

RESULTS

Pretreatment of neurons with NO generated by bubbling gaseous NO into hypoxic buffer solutions resulted in inhibition of NMDA and glycine induced increases in [Ca2+]i (Fig. 1). Similar levels of inhibition of the NMDA response were observed using NO concentrations between 1-100 /~M (Table I), and the level of inhibition was not increased by including NO with the agonist. The effect of NO was rapidly reversible by superfusion of cells with drug-free solutions (Fig. 1). A similar level of inhibition was produced by ! m M ISDN (Fig. 2A), while 1 mM SNP usually resulted in a greater diminution of the NMDA response (Fig. 2B). SNAP also inhibited NMDA responses and this inhibition became progressively greater with time (Fig. 3). These results are summarized in Table I. As NO is very unstable, and degenerates into nitrite and nitrate upon exposure to oxygen, we also tested these st~ecies for their effects on NMDA-induced

TABLE ! Comparison of the inhibitory effects of NO, NSP, ISDN and SNAP pretreatment on changes in [Ca 2 +]t elicited b), 30 ~ M NMDA / I lz M glycine in single cuhured embryonic rat forebrain neurons Neurons were exposed to the indicated agents for 200 s (except SNAP = 400 s), washed with agent-free HBSS, and then exposed to NMDA/glycine. Data are expressed as percent of untreated control NMDA response and are the mean + S.E.M. of the number of experiments indicated in parentheses. NT indicates not tested. Mean control NMDA responses ranged from 95 to 651 nM.

Concentration 1/.tM 10 p,M 100 p.M 500 ttM 1 mM

hzcreae in [Ca 2 +]i (% centred NO SNP

ISDN

SNAP

69.5 + 9.6 (6) * 66.3+4.4 (6) * 63.8 + 5.6 (l 6) * NT NT

NT 88.7+5.8 (4) 81.3 + 6.3 (4) * NT 71.9-1-4.7(4) *

NT NT 85.2 + 7.4 (9) 80.0 + 7.7 (9) * 66.0_+6.6(9) *

NT NT 76.1 + 5.6 (11) * NT 33.4+13.2 (5) *

* Significantly different to control (P < 0.05, Student's t-test).

....

312

400

300

3OO c

200

+

% 2OO u

m

too

tOO

o-.

tl~NO

t O l ~ NO

lOO~No

,;oo

o

0

0

Fig. I. Nitric oxide (NO) reversibly inhibits NMDA/glycine-induced increases in [CaZ ÷ ]i. In this trace from a single neuron. N M D A (30 p M ) and glycine (1 p M ) were added at the arrowheads. The first NMDA-induced response is a control response prior to drug exposure. subsequent responses were obtained after a 200 s pretreatment with the indicated NO concentration. The inhibition produced by NO was rapidly reversible by perfusion with drug-free HBSS. Similar effects of NO were obtained in 6-16 additional cells• Mean control NMDA responses ranged from 95-651 nM.

[Ca2*]i increases (Table I!). Nitrate (0.01-10 mM)was effective at decreasing NMDA-induced [Ca2+]i changes, while nitrite had a smaller effect, in addition NO-bubbled solutions that were exposed to air for at least 30 rain ('oxidized' NO) were able to decrease responses to NMDA to levels similar to those produced by nitrate. Nevertheless, the level of inhibition produced by nitrate, nitrite or 'oxidized' NO was lower than that resulting from NO itself (Table II). This suggests that NO breakdown products may also contribute to the overall actions of NO. We also monitored the ability of NO, SNP and ISDN to inhibit [Ca'+]l changes produced by other

A

%

i¸

1

50 f

!

fi00

I mM $NP

0( 800

i

i

i

1500

2000

2500

Time, sec Fig. 3. S-Nitroso-N-acetylpenicillamine (SNAP) inhibits NMDA-induced increases in [Ca 2+ ]i in a time-dependent manner. NMDA (30 gM) and glycine (I g M ) were added at the arrowheads. The response in this cell was measured both before and after treatment (400 s) with 100 g M SNAP. followed by washing with drug-free HBSS. Unlike NO-induced inhibition, this SNAP-induced inhibition does not reverse with washing, in fact the inhibition becomes progressively greater with time. Similar responses were observed in 3 additional cells•

stimuli. Depolarization by KCI and by kainate (50 pM) resulted in large increases in [Ca 2+]i that were probably mediated by voltage sensitive Ca z+ channels. Increases in [Ca2+]i produced by these agents were less sensitive to inhibition by NO and NO-releasing drugs. However ISDN and SNP significantly diminished KC! responses, while NO was more effective against kainate (Table liD, We originally hypothesized that NO would decrease NMDA responses by oxidizing the redox site on the NMDA receptor complex, and thus performed two series of experiments to test this idea. Firstly, we pretreated cells with 5,5'-dithio-bis-2-nitrobenzoic acid (DTNB) to oxidize the NMDA receptor. Under this condition NO still caused a significant inhibition of the

Neurons were exposed to the indicated agents for 200 s, washed with agent-free HBSS, and then exposed to NMDA/glycine. 'Oxidized' NO was generated by exposing a 100 g M NO solution to air for at least 30 rain, Data are expressed us percent of untreated control NMDA response (control -- 100%) and are the mean ± S.E.M. of the number of experiments indicated in parentheses,

,

~0

400

i

1000

Sodsurn nitrate, sodium ttitritt, attd '~tffdized' NO inhibit 30 IzM NMDA / ! I~M glycine induced chanl~es in [C'az +]~ in single etdtured etnbtyotffc rat forebrain neurons

tOO

200

500

TABLE II

~'~ I00

0

A

B

150

150

I mM tSDN

1 O0 pM SNAP

l OO

Time, I¢¢

2oo

A

0

Concemration

Increase in [Ca 2 +1~(% control) Sodium nitrate

Sodium nitrite

10ttM 100 g M 1 mM 10raM

87.4+3.2(7)* 78.14-6,5 (8) * 76.1 +4.8 (8) * 82.44-5.8(4)

92.84.2.6(5) 98.3 4- 6.3 (5) 87.6±5.6 (5) 78.24-4•8(5) *

100

Fig. 2. Isosorbide dinitrate (ISDN) and sodium nitroprussid¢ (SNP) inhibit NMDA/glycine-induced increases in [Ca 2+ ]i. NMDA (30 p M ) and glycine (I ~M)were added at the arrowheads. Responses in single neurons were recorded before and after 200 s treatment with I mM (A) ISDN or (B) SNP, Similar results were found in 4-11 additional cells.

"Oxidized'NO 84,7 + 5.8 (9) *

* Significantly different from control ( P < 0.05, Student's t-test).

313

A

TABLE !Ii

Comparison of the effects of NO. SNP, or ISDN on changes in [Ca z +]~ elicited by 30 p M NMDA / 1 p M glycine, 50 IzM kainate, or 50 mM KCI in single cultured embryonic rate forebrain neurons Neurons were treated with the indicated agents for 200 s, washed with HBSS, and then exposed to the indicated agonist. Data are expressed as percent of untreated control agonist response (control = 100%) and are the mean+S.E.M, of the number of experiments indicated in parentheses. Mean control kainate responses ranged from 149 to 741 nM; mean KCI responses ranged from 165 to 670 nM.

Agonist

100 I~M NO

I mM SNP

B

I mM ISDN

Control

0.5 mM SNP

Control

I m M ISDN

NMDA/glycine 63.8+5.6 (16) * 3.4+ 13.2 (5) * 71.9+4.7 (4) * Kainate 84.3+4.4 (6)* 90.8+ 3.9(6) 91.9+3.7(3) KCI 94.9+1.7 (4) 91.9+3,0(6)* 63.1+3.5(4)* • Significantly different from control (P < 0.05, Student's t-test)

NMDA response (Fig. 4A). Following oxidation by DTNB, NO decreased NMDA responses to 73.8 + 4.7% of control (mean + S.E.M. compared to untreated control NMDA response, n = 7). Secondly, we tested NO on cells that were initially reduced with dithiothreitol (DTT) after which we exposed the reduced receptor to the alkylating agent N-ethylmaleimide (HEM). Under these conditions DTNB can no longer reverse the enhancement of the NMDA response produced by DTT 34. In 4 cells (out of 17 cells tested) in which NMDA responses were enhanced by DTT and not subsequently injured by NEM treatment, DTNB failed to substantially decrease the NMDA response. In these cells, DTNB diminished the response to 93.1 + 1.5% ( n - 4) of the DTT.enhanced levels, compared to 57.6 + 5.1% (n = 12) of DTT-enhanced levels in cells not treated with NEM. These

A

300

Fig. 4. SNP but not ISDN inhibits NMDA- and glycine-induced whole-cell patch-clamp currents. (A) SNP applied by bath perfusion for 1 min substantially reduced response to puffer-applied NMDA (30 p,M) and glycine (1 /zM), while (B) ISDN applied in the same way has not effect. Cells were held at - 6 0 inV. Calibration bars represent 100 pA and 2 s.

results indicate that the redox site had been successfully and irreversibly alkylated. In these four cells 100 pM NO was still able to inhibit NMDA responses to 74.0 + 11.5% of the D T T / N E M pretreated levels. We were unable to use the remaining 13 cells used in this series of experiments, as the D T T / N E M sequence of treatments was either not successful in alkylating the redox site or not always well tolerated by the cells. In 8 of these 13 cells the D T T / N E M treatment failed to permanently enhance or, in fact, produced run-down of the NMDA response, and in 5 cells the DTT/NEM treatment injured or killed the cells.

B

400

300

200

too

~

wrNn

DTT/NEM

NO

i

i

i

i

i

i

0

500

1000

1500

2000

0

DTNB

NO

i

i

i

1000

2000

3000

Time, s~ Time, s~ Fig. 5, A: NO further inhibits NMDA-induced increases in [Ca :+ ]i in single neurons after oxidation with DTNB. NMDA (30 p.M) and glycine (I /zM) were added at the arrowheads, After a 200 s treatment with 5 mM dithiothreitol (DTT) followed by a 150 s incubation with 0.5 mM DTNB, NO (100/zM) applied for 200 s further decreased NMDA-induced [Ca 2+ ]i fluxes. Similar results were obtained in 6 additional cells. B: NO inhibits NMDA-elicited increases in [Ca 2+ ]i in cells locked in a reduced state. Cells were pretreated with 4 mM DTF for 200 s, washed with DTT-free HBSS and then exposed to 300 p.M N.ethylmaleimide (NEM) for 300 s, resulting in an enhanced NMDA response. This enhancement was not reversed by oxidation for 150 s with 0.5 mM DTNB, but was substantially inhibited by 200 s exposure to I00 p~M NO. Similar results were obtained in 3 additional cells.

314 To determine whether an intracellular component contributed to the action of NO we monitored NMDA responses using the whole-cell patch-clamp configuration. In this paradigm SNP (0.5 mM) applied by bath perfusion significantly decreased NMDA responses to 57.6 + 10.4% of control levels (mean ± S.E.M., n -- 7; P < 0.05, paired t-test) after 30 s. Responses were further diminished to 28 ± 8.1% (n -- 7) of control after 1 min (Fig. 5) and to 10.4 + 6.2% (n -- 5) of control values after 2 min. NMDA responses inhibited by SNP recovered very slowly, requiring > 20 rain of drug-free perfusion for complete reversal of the blockade. SNP was also fully effective following receptor alkylation with NEM (data not shown). In contrast ISDN (1 mM) had no significant inhibitory effect after 30 s or 1 rain applications. After these treatments, responses were 92.2 ± 3.4 and 91.5 ± 5.2% of control levels (mean ± S.E.M., n = 6, P > 0.05), respectively. The effects of SNAP could not be reliably tested in whole-cell recordings as this drug produced a prottounced but reversible increase in the access resistanc~ which confounded the results. The observation that SNP was more effective in inhibiting NMDA response than NO and ISDN led us to question whether all of these compounds shared common mechanism of action. To determine whether SNP was acting directly at the NMDA receptor complex we monitored the binding of [3H]dizocilpine in well washed rat brain membranes. SNP proved to be an effective inhibitor of [~H]dizocilpine binding, in the presence of saturating concentrations of glycine SNP inhibited [~H]dizocilpine binding with an IC50 of 225 ± 76 ttM (mean + S,E.M., n m 3, Fig. 6), Interestingly, this inhibition was sensitive to changes in glutamate concentrations, as elevating glutamate from 0.1 to 100 ttM increased the ICs0 above 10 mM. ISDN and SNAP

tO0

0:t}hMOlu30ttMOly

"~L~,,.

0 "~"

"6

-5

~ o g [SNpi3M

-2

-I

Fig. 6, SNP inhibits [3H]dizocilpine binding to rat brain membranes, SNP was more effective in the presence of low concentrations of glutamate, but was less effective when glutamate was raised to 100 ttM. Data represents the mean of 3 experiments performed in duplicate.

had minimal effects on [3HI dizocilpine binding at the concentrations used in the [Ca 2 + ]i experiments. One prominent action of NO on neuronal cell function is its ability to activate guanylate cyclase with resulting production of cGMP. Accordingly, we tested 8-bromo-cGMP for its effects on NMDA-induced increases in [Ca2+]i as a possible mechanism of action for NO. Concentrations of 8-bromo-cGMP up to 0.5 mM had no significant inhibitory effect on NMDA-induced increases in [Ca2+]i (data not shown) suggesting that the inhibitory effect of NO is not mediated by cGMP. DISCUSSION In this study we have demonstrated that NO and NO-generating compounds decrease NMDA- and glycine-mediated increases in [Ca2+] i in cultured neurons. Similar but smaller effects of NO were seen on the action of other agents that increase [Ca2+]i, including kainate and KCl. As the production of NO in neurons is dependent on an increase in [Ca2+]~3, these findings suggest that NO may act as a feedback inhibitor of its own synthesis by limiting increases in [Ca2+]i. While the precise significance of this observation remains to be determined, it is interesting to note that neurons that contain NOS are less sensitive to NMDA-receptor mediated neurotoxicity 14'1~, while nitric oxide produc. tion may mediate glutamate neurotoxicity H. As this form of cell death require~ extracellular Ca 2+ (ref. 6), and is associated with large increases in [Ca~+]l I,~,22,~.~, the ability of NO to limit increases in [Ca2÷]~ may be responsible for the reduced sensitivity of NOS-containing cells to neurotoxic stimuli, We originally hypothesized that NO might modify the site on the NMDA receptor that is sensitive to sulfhydryl redox reagents as a consequence of its known action as an oxidizing agent. However, our findings do not support this hypothesis. We found that NO still effectively diminished NMDA responses in ceils that had been fully oxidized by DTNB. While it is possible that NO is a more effective oxidizing agent than DTNB, the spontaneous reversal of NO-inhibited responses is in contrast to the effects of the oxidizing agent DTNB 3. We also tested NO in cells that had been reduced by DTT and subsequently 'locked' in the reduced state by subsequent application of NEM 34. The insensitivity of these cells to DTNB (Fig. 3B) clearly demonstrates successful alkylation of the redox site. Nevertheless, NO still significantly inhibited NMDA-induced [Ca2+]i increases in these cells, again suggesting that the actions of NO are not mediated by the redox site. Finally, as NO decreased responses produced by kainate and

315 KCI which are not sensitive to redox reagents in the same way as NMDA receptors 3'32, a simple redox mechanism for the action of NO seems unlikely. From our studies, the precise mechanism by which NO produces its effects on NMDA-mediated [Ca2+]i changes is not clear. In our experiments we used NO and NO-generating drugs by pretreating cells, and then washing the drugs out prior to agonist addition. The effectiveness of these agents in this paradigm is suggestive of an effect that is not simply the result ~f a drug-receptor interaction on a cell surface receptor. The time course observed may be more consistent with the production of an intracellular second messenger. As each of these agents was effective, it seems unlikely that the difference in the mechanisms by which they release NO 33 is of critical importance to their overall action, although the difference in the time course of action might be altered by this variable. It also seen:~ unlikely that NO acts directly on NMDA receptors because ISDN had no effect on whole-cell patch-clamp currents, As the cytoplasm becomes dialyzed with this procedure, this technique usually measures drug interaction directly with the receptor, is less sensitive to effects that require second messenger production, and is usually insensitive to downstream events. The absence of an effect with ISDN on whole-cell currents while effective in [Ca2+]+ assays implies that the principal effect of NO may be mediated by one of the Ca 2+ homeostatic mechanisms operating in these cells. This would also account for the ability of NO and related agents to modify [Ca2÷]l changes produced by the other agonists. Inhibition of [Ca2+]~ increases by agents that release NO has been observed in a number of non-neuronal cell types including vascular smooth muscle cells, platelets and fibroblasts. Specifically, SNP has been shown to increase Na+-independent Ca 2+ effiux tt, and reversibly inhibit Ca 2+ influx through voltage sensitive Ca 2+ channels 7 in vascular smooth muscle cells. In platelets, SNP inhibits intracellular Ca 2+ translocation induced by platelet activating factor and arachidonic acid 29, as well as ADP-mediated Ca 2+ influx 2~. SNP also inhibits thrombin-induced Ca 2+ mobilization in plateiets t~. Finally, in Balb/c 3T3 fibroblasts, NO, ISDN, and SNAP decrease [Ca2+]i (ref. 12). Interestingly, not all these effects on Ca 2+ homeostasis are mimicked by cGMP. These observations, together with the results from this study, support the possibility that NO exerts its inhibitory effect on NMDA responses via modification of Ca 2+ homeostasis similar to the effects of NO on other cell types, although the actual mechanism by which it has its effect may vary with respect to cell type. The precise effect of NO on specific neuronal Ca 2+ homeostatic mechanisms re-

mains to be determined. As 8-bromo-cGMP did not mimic the effect of NO in this study, we propose that the effect of NO on neuronal Ca 2+ homeostasis is not cGMP mediated. This statement is supported by the recent observation that 8-bromo cGMP does not inhibit NMDA-induced whole cell currents nor NMDAinduced increases in [Ca2+]~I. The precise molecular species responsible for the effects of NO also remains to be determined. Aware of the instability of NO in solution, we also tested the effects of the major oxidative metabolites of NO. The sodium salts of nitrate and nitrite also decreased [Ca2+]i, with the former being more effective. We canno~ exclude the possibility that all of the effects of NO occur subsequent to its conversion to nitrate. However, as this conversion would also occur in vivo, this would not preclude the possible physiological significance of our findings. Finally, we were surprised to see the differences in the magnitude of the effects of SNP compared to NO and ISDN and its ability to inhibit NMDA-induced whole cell currents whereas ISDN does not. Using assays that more directly monitor the effects of drugs on the NMDA receptor we observed that SNP directly inhibited [3H]dizocilpine binding and NMDA-mediated whole cell currents. These results suggest that SNP exerts an inhibitory effect on the NMDA receptor directly in addition to actions arising from the release of NO. This is consistent with other studies that have suggested that the ferrocyanate moiety of SNP directly interacts with the NMDA receptor I°'tT. In conclusion, these studies have shown that NMDA-induced changes in [Ca2+]l are diminished by NO and by NO generating agents. Our findings suggest that this effect of NO is probably not mediated directly by the NMDA receptor-associated sulfhydryl redox site as originally hypothesized. Instead some down-stream event appears to be the target for NO. Finally, SNP appears to have limited usefulness for investigating the interaction of NO with the NMDA receptor due to a direct inhibitory action on the receptor complex. Acknowledgements. We gratefullyacknowledgethe expert technical assistanceof Kristi Rothermundand Karen Hartnett. This studywas supported in part by NIH Grant NS 29365 (EA).

ABBREVIATIONS DMEM DTNB DTT EGTA cGMP

HBSS

Dulbccco's modified Eagle's medium 5,5'.dithio.bis.(2-nitrobenzoic acid) dith~othreitol ethyleneglycol-bis-(,8-aminoethyl ether)N,N,N',N' tetraacetic acid guanosine 3' : 5'-cyclic monophosphate HEPES-buffered saline solution

316 HEPES [Ca2. ], ISDN NEM NO NMDA SNAP SNP Trx

N.(2,hydroxyethyl)piperazine-N '.(2,ethanesulfonic acid) intracellular free calcium isosorbide dinitrate N-etbylmaleimide nitric oxide N-methyI-D-aspartate S-nitroso-N-acetyipenicillamine sodium nitroprusside tetrodotoxin

REFERENCES ! Abele, A.E., Scholz. K.P., Scholz, W.K. and Miller, RJ., Excitetoxicity induced by enhanced excitatory neurotransmission in cultured hippocampal primary neurons. Nearon, 4 (1990) 413-419. 2 Aizenman, E., Hartnett, K.A. and Reynolds. l.J., Oxygen free radicals regulate N-methyI-D-aspartate receptor function via a redox modulatory site, Neuron, 5 (1990) 841-846. 3 Aizenman, E,, Lipton, S.A. and Loring, R.H., Selective modulation of NMDA responses by reduction and oxidation, Neuron, 2 (1989) 1257-1263. 4 Bredt, D.S., Hwang, P.M. and Snyder. S.H., Localization of nitric oxide synthase indicating a neural role for nitric oxide, Nature, 347 (1990) 768-770. 5 IBredt, D.S. and Snyder, S.H., Nitric oxide mediates glutamatelinked enhancement of cGMP levels in the cerebellum, Prec. Natl. Acad. Sci. USA, 86 (1989) 9030-9033. 6 Choi, D.W., Ionic dependence of glutamate neurotoxicity, J. Neurosci., 7 (1987) 369-379. 7 Clapp, L.H. and Gurney, A.M., Modulation of calcium movements by nitroprusside in isolated vascular smooth muscle cells. Pflugers Arch., 418 (1991) 462-470. 8 Dawson, V.L., Dawson, T.M., London. E.D., Bredt. D.S. and Snyder, S.H., Nitric oxide mediates glutamate neurotoxicity in primary cortical cultures, Pmc. Not/. Acad. Set. USA, 88 (1991) 6368-637 I. 9 De Erausquin. G.A.. Manev. H., Guidotti, A., Costa, E, and Brooker. O.. Gangliosides normal[,.~ distorted single-cell intracel. lular free Ca =* dynamics after toxic doses of glutamate in cure. helhtr granule cells, Prec. Nat/. Acad. Sci. USA, 87 (1990) 81ll78I)21. II) Fujimori, II. and Pan-Hou, H., Effect of nitric oxide on I.. ['~H]81utamate binding to rat brain synaptic membranes, Brain Res., 554 (1991) 355-357, !1 Furukawa, K.-I,, Tawada, Y. and Shil+ekawa, M,, Regulation of the plasma membrane Ca2+ pump by cyclic nucleolides in cal. tured vascular smooth muscle cells, J. Biol. Chem., 263 (1988) 8058-8065, 12 Gar8, U.C and Hassid, A,, Nitric oxide decreases cytosolic free calcium in Balb/c 3T3 fibroblasts by a cyclic GMP-independent mechanism, J, Biol. Chum,, 266 (1991) 9-12, 13 Garthwaite, J,, Charles, S,L, and Chess-Williams, R., Endothe. lium deri~ed relaxing factor release on activation of NMDA receptors ~uggests role as intercellular messenger in brain, Nature, 336 (1988) 385-388. 14 Hope, B.T., Michael, G,J,, Knigge, K,M, and Vincent, S,R,, Neuronal NADPH diaphorase is a nitric oxide synthase, Proc, NaIL Acad. Sci. USA, 88 (1991) 2811-2814, 15 Johnson, J.W. and Ascher, P., Glycine potentiates the NMDA response in cultured mouse brain neurons, Natron,, 325 (1987) 529-531. 16 Kawahara, Y,, Yamanishi, J, and Fukuzaki, It., Inhibitory aclion of guanosine 3',5'-monophosphate on thrombin-induced calcium mobilization in human platelets, Thromb. Res., 33 (1984)203-209.

17 Kiedrowski, L., Costa, E. and Wroblewski, J.T., Inhibition of N-methyI-D-aspartate-induced calcium influx in cerebellar neurons by sodium nitroprusside is not mediated by nitric oxide, Sac. NeuroscL, 17 (1991) 349. (Abstract) 18 Knowles, R.G., Palacios, M., Palmer, R.M.J. and Moncada, S., Formation of nitric oxide from I-arginine in the central nervous system: a transduction mechanism for the stimulation of soluble guanylate cyclase, Prec. Natl. Acad. Sci. USA, 86 (1989) 51595162. 19 Koh, J.Y. and Choi, D.W., Vulnerability of cultured cortical neurons to damage by excitotoxins: differential susceptibility of neurons containing NADPH-diaphorase, jr Neurosci., 8 (1988) 2153-2163. 20 MacDermott, A.B., Mayer, M.L., Westbrook, G.L., Smith, S.J. and Barker, J.L., NMDA-receptor activation increases cytoplasmic calcium concentration in cultured spinal cord neurones, Nature. 321 (1986) 519-522. 21 Manzoni, O.JJ., Fagni, L., Marin, P., Prezeau, L., Sahuquet, A., Sladeczeck, F. and Bockaert, J., Nitric oxide (NO) selectively inhibits NMDA receptor activation via a cGMP-independent pathway, Sac. Neurosci.. 17 (1991) 72. (Abstract) 22 Mattson. M.P., Rychlik, B., Chu, C. and Christakos, S., Evidence for calcium-reducing and excite-protective roles for the calciumbinding protein calbindin-D2s k in cultured hippocampal neurons, Neuron, 6 (1991) 41-51. 23 Mayer, M.L., MacDermott, A.B.. Westbrook, G.L., Smith, SJ. and Barker, J.L., Agonist and voltage gated calcium entry in cultured mouse spinal cord neurons under voltage clamp measured using arsenazo 111, J. Neurosci,, 7 (1987) 3230-3244. 24 Menon. N.K., Wolf, A., Zehetgruber, M, and Bing, R.J., An improved chemiluminesence assay suggests non nitric oxide mediated action of lysophosphatidylcholine and acetylcholine, Prec. Sac. Exp. Biol. Mud., 191 (1989)316-319. 25 Michaels, R.L. and Rothman, S.M., Glutamate neurotoxicity in vitro: antagonist pharmacology and intracellular calc~,.~mconcentrations, J. Neurosci., 10 (1990) 283-292. 26 Moncada, S., Palmer, R.M.J, and Higgs, E.A., Nitric oxide: physiology, pathophysiology and pharmacology, Pha~macol, Ret',, 43 (1991) 11)0-142, 27 Morgan. R.O. and Newby, A.C,. Nitroprusside differentially inhibits ADP-stimulated calcium influx and mobilization in Imman platelets, Bkwhem. J., 2~8 (1989)447-454. 28 Murphy, S,N., Thayer, S.A. and Miller, R,J., The effects of excitatory amino acids on intracellular calcium in single mouse striatal neurons in vitro, 3. Neumsci., 7 (1987) 4145-4158. 29 Pasqui, A,L., Capecchi, P,L,, Ceccatelli, L,, Mazza, S,, Gistrl, A,, Laghi Pasini, F. and Di Perri, T., Nitroprusside in vitro inhibits platelel aggregation and intraceUular calcium translocation, Effect of haemoglobin, Thromb. Res., 61 (1991) 113-122. 30 Pryor, W.A., Church, D,F,, Govindan, C,K, and Crank, G,, Oxidation of thiols by nitric oxide and nitrogen dioxide: synthetic utility and toxicological implications, J. Org. Chem,, 47 (1982) 156-159, 31 Reynolds, l,J,, Arcaine uncovers dual interaction of polyamines with the N-methyl-D-aspartate receptor, J, PharmacoL Exp. Ther., 255 (1990) 1001-1007, 32 Reynolds, l.J., Rush, E,A. and Aizenman, E., Reduction of NMDA receptors with dithiothreitol increases [3H]MK801 binding and NMDA-induced Ca~+ fluxes, Br. J, Pharmacol,, 101 (1990) 178-182, 33 Southam, E, and Garthwaite, J,, Comparative effects of some nitric oxide donors on cyclic GMP levels in rat cercbellar slices, Neura~vci, Left,, 130 (1991) 107-111. 34 Tang, L,-H, and Aizenman, E., NMDA receptor blockade by H +, ~+ M8* , and Zn 2. following alkylation of the redox modulatory site, Sac. Neurosci,, 17 (1991) 1167. (Abstract)