J~URNALOF NEUROPHYSIOLOGY Vol. 68, No. 2. August 1992. Printed
RAPID
in U.S.A.
PUBLICATION
Evidence for Nitric Oxide Synthase Inhibitor-Sensitive and Insensitive Hippocampal Synaptic Potentiation VALENTIN
K. GRIBKOFF
AND
Department Wallin&rd,
uf’Neuvopharmacology, Connecticut 06492
SUMMARY
AND
JANET
T. LUM-RAGAN
Bristol-Myers
Squibb Pharmaceutical
CONCLUSIONS
1. Nitric oxide (NO) has been proposed as a retrograde messenger, mediating the postsynaptic to presynaptic transfer of the effects of conditioning stimulation, responsible for the initiation of hippocampal long-term potentiation (LTP). To further test this hypothesis, we inhibited nitric oxide synthase (NOS) to determine whether synaptic potentiation produced by different conditioning stimulus patterns and intensities was differentially affected by reduction of stimulation-dependent NO production. 2. Synaptic potentiation was produced in hippocampal slices from young F-344 rats by two different conditioning stimulation protocols. Conditioning stimuli were delivered to the Schaffer-collateral commissural system, and moderate levels of potentiation of the population excitatory postsynaptic potential (EPSP) in area CA 1 were produced by a single 100 Hz, l-s conditioning train delivered at half-maximal stimulus intensity. Higher levels of potentiation of the population EPSP were obtained by delivering two 100 Hz, l-s conditioning stimulus trains, with a 60-s intertrain interval, at high stimulus currents. 3. Application of the nitric oxide synthase inhibitors NG-nitroL-arginine ( NOARG; 0.1-200 PM ) and N G-monomethyl-L-arginine (NMMA; 100 ,uM) produced no significant direct effects on synaptic responses. 4. In slices that received a single conditioning stimulus train, both NOARG and NMMA were ineffective in blocking or reducing potentiation at concentrations between 0.1 and 200 PM. In slices receiving the more intense pair of conditioning stimulus trains, levels of potentiation in control slices were higher, and there was a very significant reduction by both NOARG (50 and 100 PM) and NMMA (100 PM). 5. Application of the specific N-methyl-D-aspartate (NMDA) excitatory amino acid antagonist DL-Zamino-$phosphonopentanoic acid (AP5; 50 FM) significantly reduced potentiation in the 2-train conditioning model, and also reduced potentiation in the 1-train model. 6. These results provide further support for the hypothesis that NO release after conditioning stimulation of the Schaffer-collatera1 commissural system in hippocampal slices can contribute to synaptic potentiation. Our results suggest, however, that NO plays a modulatory rather than an obligatory role in the expression of synaptic potentiation in vitro.
INTRODUCTION
Research Institute,
synaptic plasticity in both the cerebellum, where it may be involved in long-term depression of purkinje cell output (Crepe1 and Jaillard 1990; Shibuki and Okada 199 1; see also Linden and Connor 1992) and in the hippocampus. Hippocampal long-term potentiation (LTP) is dependent on the postsynaptic activation of N-methyl-D-aspartate (NMDA) excitatory amino acid receptors and the subsequent entry of calcium, leading to a cascade of sensitizing intracellular events in area CA 1 (Collingridge 1985; Madison et al. 199 1; Malenka 199 1). In addition, there is evidence linking LTP to presynaptic changes in release probability in both CA 1 and the dentate gyrus (Bekkers and Stevens 1990; Bliss et al. 1986; Malinow and Tsien 1990). This combination of pre- and postsynaptic involvement requires a “retrograde diffusable factor” to initiate presynaptic changes leading to enhanced probability of transmitter release (Bekkers and Stevens 1990; Malinow and Tsien 1990). A prime candidate for this factor is NO (Gally et al. 1990). In support of the NO-LTP hypothesis, several studies have recently reported that inhibition of nitric oxide synthase (NOS), the synthetic enzyme responsible for the production of NO by the conversion of L-arginine to L-citrulline, can virtually eliminate LTP in vitro (Bohme et al. 199 1; Haley et al. 1992; O’Dell et al. 199 1; Schuman and Madison 199 1). Not all laboratories, however, have reported significant reduction of LTP in area CA 1 in the presence of NOS inhibitors (Errington et al. 199 1). This suggests that the role of NO in area CA 1 may be more complex than previously indicated. The present experiments result from our initial inability to reduce synaptic potentiation by NOS inhibition in area CA1 of rat hippocampal slices, when a single high-frequency stimulus train was used as the conditioning stimulus. We have examined the NOS inhibitor-sensitivity of synaptic potentiation produced by two different conditioning paradigms and report that NOS inhibitor sensitivity of hippocampal synaptic potentiation may be conditioning dependent. METHODS
The diffusable free radical gas nitric oxide (NO) plays an important role in hemodynamic regulation as an endothelium-derived relaxation factor and as the “transmitter” at certain nonadrenergic, noncholinergic synapses in the periphery (Moncada et al. 199 1). The functions of NO in brain, however, are not well understood (Bredt and Snyder 1992). Recent evidence indicates that NO participates in
Hippocampal slices were obtained from young ( l-3 mo) F-344 male rats ( y1 = 46) using standard techniques. Slices were - 500 PM coronal sections, and after preparation were either maintained in control medium (composition below) in a “prechamber” at room temperature, or were directly transferred to a recording chamber (Haas et al. 1979) and maintained at 32°C. In the recording chamber, slices were submerged in a continuous flow of me-
0022-3077192 $2.00 Copyright 0 1992 The American Physiological Society
639
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (192.236.036.029) on December 19, 2018.
640
V. K. GRIBKOFF
AND J. T. LUM-RAGAN
dium ( 1.7 ml/ min) provided by a peristaltic pump. All solutions were addedin the bath. Rapid switching wasaccomplishedwith miniature flow valves. The control bathing medium had the following composition(in mM): 124NaCl, 3.0 KCl, 2.5 NaH,PO,, 1.3 MgS04, 26 NaHCO,, 2.4 CaCl,, 10 glucose.Drug solutions weremadefrom freshstocksolutionspreparedat the beginningof eachday, and diluted to their final concentration in control medium just before use. Extracellular population EPSPswererecordedfrom stratum radiatum. All data were storedon a computer for analysis.Pre- and postconditioning test stimuli (0.3-ms constant-current square pulses)weredeliveredin all casesat 1/min to the Schaffer-collatera1commissuralfibers in stratum radiatum at the CA2-CA3 border, via a bipolar insulatedtungstenelectrode.Beforeand after conditioning, all test stimuli in both conditioning protocolswere delivered at a current that, before conditioning, produced a halfmaximal population EPSP.In all slices,baselineresponsestability wasmonitored for 230 min beforethe addition of drug solutions or, in the caseof experimentsperformed in normal medium, the conditioning stimulation. Slicesused in theseexperiments had < 10%responsevariation during this period. The preconditioning control value for eachslicein eachexperiment wasthe averageof six teststimuli collectedimmediatelybeforethe conditioning stimulation. Conditioning stimuli werealsodelivered to the Schaffercollateral fibers. In the l-train model, the conditioning stimulus wasa single 100 Hz l-s train delivered at half-maximal current, whereasin the 2-train model the conditioning stimulus was 2 - 100 Hz 1-s trains, separatedby 60 s, delivered at maximal current. Postconditioningtestresponses werecollectedfor 60 min, and the slopesof the population EPSPswere comparedwith control values.Direct effectsof drugson synaptic transmissionwere assessed by applying the drug before the conditioning stimulation for a minimum of 1 h for NOS inhibitors, and a minimum of 30 min for NMDA receptorantagonism.Drug solutionswerepresent during the entire experiment, including the 60 min postconditioning. Drugs used in theseexperiments were the NOS inhibitors NG-monomethyl-L-arginine(NMMA; Calbiochem) (Reeset al. 1990) and NG-nitro-L-arginine (NOARG; Sigma Chemical) (Dwyer et al. 1991; East and Garthwaite 1990), and the specific competitive NMDA receptor antagonist DL-2-amino-5-phosphonopentanoicacid (AP5; ResearchBiochemicals)(Davies et al. 1981). Light sensitivesolutions,suchasNMMA in solution, were protected from light for the entire experiment. Data were combined from each experimental group in each conditioning paradigm. Thesevalues were expressedastime-responseprofilesof potentiation levelsfrom l-60 min after the conditioning train, assessed at 10-min intervals. Statistical analyses consistedof simpleone-way analysesof variance (ANOVAs) for determination of the direct effectsof drug treatmentson synaptic transmission.RepeatedmeasuresANOVAs wereusedto compare the time-responseprofiles of the decay of potentiation for each experimentalgroup in the two experiments.If significantF-values for the main effect of treatment werefound, preplannedpost hoc comparisons(Fisher’s Protected LSD) were made betweenthe curves for the various treatment groups,with time asthe dependent variable and treatment asthe effect. A significantdifference betweengroupswasdefined asP 5 0.05. Statisticalanalyseswere carried out with “SuperANOVA” General Linear Modeling software (Abacus Concepts,Berkeley, CA). RESULTS
A total of 107 slices were used in this study. In the 1-train model, slices were conditioned in control medium (n = 19), 0.1 pM NOARG (n = 5), 50 pM NOARG (n = 9), 200 UM NOARG (n = 11). 100 UM NMMA (n = 8). and
50 pM AP5 (n = 11). Similarly, in the 2-train model, slices were conditioned in control medium (n = lo), 50 pM NOARG (n = 8), 100 pM NOARG (n = 9), 100 pM NMMA (n = 9), and 50 pM AP5 (n = 8). No significant effect of any of these drug solutions was observed on the
A co4ltrol NoARGsouM NOARG 100 UN NMMA
100
uM
APSSOUUY
10
0
20
30
40
50
Time After Conditioning
60
Stimuli
Contrd NOARG NMMA
loo 100
YY UY
APsso&J)rY
: d2
50 +
10I
20I
301
40,
601
50.
0
Time After Conditioning
Stimuli
2 Train
1 Train
NOARG 200uM
* 2 mV
_J G 5 m3 FIG. 1. A : slices conditioned in the 1-train model did not display sensitivity to NOS inhibition by either NOARG or NMMA and were only partially sensitive to the NMDA antagonist AP5. Potentiation levels were actually higher in the presence of NOARG (and were significantly increased at 200 PM). B: slices conditioned in the 2-train model were very sensitive to both NOARG and NMMA; potentiation levels in slices incubated in the NOS inhibitors were reduced to about the same level observed in APS-treated slices. C: representative examples of population EPSPs recorded from slices conditioned in the 1-train or 2-train models and incubated in control medium or the NOS inhibitor NOARG. In all cases the larger potential is the postconditioning response. Values represent means SE.: see text for groun n values.
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (192.236.036.029) on December 19, 2018.
NITRIC
OXIDE
AND
HIPPOCAMPAL
population EPSP slope during the preconditioning incubation period (Fs < 1). In the l-train conditioning model, slices incubated in drug-free medium exhibited moderate levels of synaptic potentiation that were still significantly enhanced at 60 min, relative to preconditioning control values (Fig. 1A). Application of NOARG at 0.1 (not shown in Fig. 1A), 50 and 200 PM produced no reduction of potentiation; rather, the group incubated in 200 PM NOARG had higher levels of potentiation throughout the 60-min test period. NMMA ( 100 FM) was likewise ineffective in reducing these levels. Incubation in AP5 (50 ,uM) reduced potentiation. Statistical analyses demonstrated a significant main effect for treatment (F = 3.498; P = 0.008) and a significant effect of time (F = 3.549; P = 0.002), but no significant interaction between time and treatment (F = 1.326; P = 0.122). The post hoc comparisons revealed significant differences from the drug-free control group only for the 200-PM NOARG group, and potentiation levels in this group were significantly enhanced by the NOS inhibitor (Fig. 1, A and c>. Although potentiation was depressed in the AP5 group, this did not reach statistical significance. The 2-train model produced a greater level of synaptic potentiation in control slices. The difference between control potentiation levels produced by the two different conditioning models was highly significant (F = 20.423; P < 0.000 1). In this experiment, incubation in either NOARG (50 or 100 ,uM; data for 50 ,uM not shown in Fig. 1L3) or NMMA produced a profound reduction of potentiation (Fig. 1, B and C). The NMDA antagonist AP5 reduced potentiation to a level almost identical to that seen in the NOS inhibitors. Statistical analyses revealed that synaptic potentiation in groups of slices incubated in NOS inhibitors or AP5 were significantly different from drug-free control values (significant main effects for treatment [F = 10.159; Y
RAPID
in U.S.A.
PUBLICATION
Evidence for Nitric Oxide Synthase Inhibitor-Sensitive and Insensitive Hippocampal Synaptic Potentiation VALENTIN
K. GRIBKOFF
AND
Department Wallin&rd,
uf’Neuvopharmacology, Connecticut 06492
SUMMARY
AND
JANET
T. LUM-RAGAN
Bristol-Myers
Squibb Pharmaceutical
CONCLUSIONS
1. Nitric oxide (NO) has been proposed as a retrograde messenger, mediating the postsynaptic to presynaptic transfer of the effects of conditioning stimulation, responsible for the initiation of hippocampal long-term potentiation (LTP). To further test this hypothesis, we inhibited nitric oxide synthase (NOS) to determine whether synaptic potentiation produced by different conditioning stimulus patterns and intensities was differentially affected by reduction of stimulation-dependent NO production. 2. Synaptic potentiation was produced in hippocampal slices from young F-344 rats by two different conditioning stimulation protocols. Conditioning stimuli were delivered to the Schaffer-collateral commissural system, and moderate levels of potentiation of the population excitatory postsynaptic potential (EPSP) in area CA 1 were produced by a single 100 Hz, l-s conditioning train delivered at half-maximal stimulus intensity. Higher levels of potentiation of the population EPSP were obtained by delivering two 100 Hz, l-s conditioning stimulus trains, with a 60-s intertrain interval, at high stimulus currents. 3. Application of the nitric oxide synthase inhibitors NG-nitroL-arginine ( NOARG; 0.1-200 PM ) and N G-monomethyl-L-arginine (NMMA; 100 ,uM) produced no significant direct effects on synaptic responses. 4. In slices that received a single conditioning stimulus train, both NOARG and NMMA were ineffective in blocking or reducing potentiation at concentrations between 0.1 and 200 PM. In slices receiving the more intense pair of conditioning stimulus trains, levels of potentiation in control slices were higher, and there was a very significant reduction by both NOARG (50 and 100 PM) and NMMA (100 PM). 5. Application of the specific N-methyl-D-aspartate (NMDA) excitatory amino acid antagonist DL-Zamino-$phosphonopentanoic acid (AP5; 50 FM) significantly reduced potentiation in the 2-train conditioning model, and also reduced potentiation in the 1-train model. 6. These results provide further support for the hypothesis that NO release after conditioning stimulation of the Schaffer-collatera1 commissural system in hippocampal slices can contribute to synaptic potentiation. Our results suggest, however, that NO plays a modulatory rather than an obligatory role in the expression of synaptic potentiation in vitro.
INTRODUCTION
Research Institute,
synaptic plasticity in both the cerebellum, where it may be involved in long-term depression of purkinje cell output (Crepe1 and Jaillard 1990; Shibuki and Okada 199 1; see also Linden and Connor 1992) and in the hippocampus. Hippocampal long-term potentiation (LTP) is dependent on the postsynaptic activation of N-methyl-D-aspartate (NMDA) excitatory amino acid receptors and the subsequent entry of calcium, leading to a cascade of sensitizing intracellular events in area CA 1 (Collingridge 1985; Madison et al. 199 1; Malenka 199 1). In addition, there is evidence linking LTP to presynaptic changes in release probability in both CA 1 and the dentate gyrus (Bekkers and Stevens 1990; Bliss et al. 1986; Malinow and Tsien 1990). This combination of pre- and postsynaptic involvement requires a “retrograde diffusable factor” to initiate presynaptic changes leading to enhanced probability of transmitter release (Bekkers and Stevens 1990; Malinow and Tsien 1990). A prime candidate for this factor is NO (Gally et al. 1990). In support of the NO-LTP hypothesis, several studies have recently reported that inhibition of nitric oxide synthase (NOS), the synthetic enzyme responsible for the production of NO by the conversion of L-arginine to L-citrulline, can virtually eliminate LTP in vitro (Bohme et al. 199 1; Haley et al. 1992; O’Dell et al. 199 1; Schuman and Madison 199 1). Not all laboratories, however, have reported significant reduction of LTP in area CA 1 in the presence of NOS inhibitors (Errington et al. 199 1). This suggests that the role of NO in area CA 1 may be more complex than previously indicated. The present experiments result from our initial inability to reduce synaptic potentiation by NOS inhibition in area CA1 of rat hippocampal slices, when a single high-frequency stimulus train was used as the conditioning stimulus. We have examined the NOS inhibitor-sensitivity of synaptic potentiation produced by two different conditioning paradigms and report that NOS inhibitor sensitivity of hippocampal synaptic potentiation may be conditioning dependent. METHODS
The diffusable free radical gas nitric oxide (NO) plays an important role in hemodynamic regulation as an endothelium-derived relaxation factor and as the “transmitter” at certain nonadrenergic, noncholinergic synapses in the periphery (Moncada et al. 199 1). The functions of NO in brain, however, are not well understood (Bredt and Snyder 1992). Recent evidence indicates that NO participates in
Hippocampal slices were obtained from young ( l-3 mo) F-344 male rats ( y1 = 46) using standard techniques. Slices were - 500 PM coronal sections, and after preparation were either maintained in control medium (composition below) in a “prechamber” at room temperature, or were directly transferred to a recording chamber (Haas et al. 1979) and maintained at 32°C. In the recording chamber, slices were submerged in a continuous flow of me-
0022-3077192 $2.00 Copyright 0 1992 The American Physiological Society
639
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (192.236.036.029) on December 19, 2018.
640
V. K. GRIBKOFF
AND J. T. LUM-RAGAN
dium ( 1.7 ml/ min) provided by a peristaltic pump. All solutions were addedin the bath. Rapid switching wasaccomplishedwith miniature flow valves. The control bathing medium had the following composition(in mM): 124NaCl, 3.0 KCl, 2.5 NaH,PO,, 1.3 MgS04, 26 NaHCO,, 2.4 CaCl,, 10 glucose.Drug solutions weremadefrom freshstocksolutionspreparedat the beginningof eachday, and diluted to their final concentration in control medium just before use. Extracellular population EPSPswererecordedfrom stratum radiatum. All data were storedon a computer for analysis.Pre- and postconditioning test stimuli (0.3-ms constant-current square pulses)weredeliveredin all casesat 1/min to the Schaffer-collatera1commissuralfibers in stratum radiatum at the CA2-CA3 border, via a bipolar insulatedtungstenelectrode.Beforeand after conditioning, all test stimuli in both conditioning protocolswere delivered at a current that, before conditioning, produced a halfmaximal population EPSP.In all slices,baselineresponsestability wasmonitored for 230 min beforethe addition of drug solutions or, in the caseof experimentsperformed in normal medium, the conditioning stimulation. Slicesused in theseexperiments had < 10%responsevariation during this period. The preconditioning control value for eachslicein eachexperiment wasthe averageof six teststimuli collectedimmediatelybeforethe conditioning stimulation. Conditioning stimuli werealsodelivered to the Schaffercollateral fibers. In the l-train model, the conditioning stimulus wasa single 100 Hz l-s train delivered at half-maximal current, whereasin the 2-train model the conditioning stimulus was 2 - 100 Hz 1-s trains, separatedby 60 s, delivered at maximal current. Postconditioningtestresponses werecollectedfor 60 min, and the slopesof the population EPSPswere comparedwith control values.Direct effectsof drugson synaptic transmissionwere assessed by applying the drug before the conditioning stimulation for a minimum of 1 h for NOS inhibitors, and a minimum of 30 min for NMDA receptorantagonism.Drug solutionswerepresent during the entire experiment, including the 60 min postconditioning. Drugs used in theseexperiments were the NOS inhibitors NG-monomethyl-L-arginine(NMMA; Calbiochem) (Reeset al. 1990) and NG-nitro-L-arginine (NOARG; Sigma Chemical) (Dwyer et al. 1991; East and Garthwaite 1990), and the specific competitive NMDA receptor antagonist DL-2-amino-5-phosphonopentanoicacid (AP5; ResearchBiochemicals)(Davies et al. 1981). Light sensitivesolutions,suchasNMMA in solution, were protected from light for the entire experiment. Data were combined from each experimental group in each conditioning paradigm. Thesevalues were expressedastime-responseprofilesof potentiation levelsfrom l-60 min after the conditioning train, assessed at 10-min intervals. Statistical analyses consistedof simpleone-way analysesof variance (ANOVAs) for determination of the direct effectsof drug treatmentson synaptic transmission.RepeatedmeasuresANOVAs wereusedto compare the time-responseprofiles of the decay of potentiation for each experimentalgroup in the two experiments.If significantF-values for the main effect of treatment werefound, preplannedpost hoc comparisons(Fisher’s Protected LSD) were made betweenthe curves for the various treatment groups,with time asthe dependent variable and treatment asthe effect. A significantdifference betweengroupswasdefined asP 5 0.05. Statisticalanalyseswere carried out with “SuperANOVA” General Linear Modeling software (Abacus Concepts,Berkeley, CA). RESULTS
A total of 107 slices were used in this study. In the 1-train model, slices were conditioned in control medium (n = 19), 0.1 pM NOARG (n = 5), 50 pM NOARG (n = 9), 200 UM NOARG (n = 11). 100 UM NMMA (n = 8). and
50 pM AP5 (n = 11). Similarly, in the 2-train model, slices were conditioned in control medium (n = lo), 50 pM NOARG (n = 8), 100 pM NOARG (n = 9), 100 pM NMMA (n = 9), and 50 pM AP5 (n = 8). No significant effect of any of these drug solutions was observed on the
A co4ltrol NoARGsouM NOARG 100 UN NMMA
100
uM
APSSOUUY
10
0
20
30
40
50
Time After Conditioning
60
Stimuli
Contrd NOARG NMMA
loo 100
YY UY
APsso&J)rY
: d2
50 +
10I
20I
301
40,
601
50.
0
Time After Conditioning
Stimuli
2 Train
1 Train
NOARG 200uM
* 2 mV
_J G 5 m3 FIG. 1. A : slices conditioned in the 1-train model did not display sensitivity to NOS inhibition by either NOARG or NMMA and were only partially sensitive to the NMDA antagonist AP5. Potentiation levels were actually higher in the presence of NOARG (and were significantly increased at 200 PM). B: slices conditioned in the 2-train model were very sensitive to both NOARG and NMMA; potentiation levels in slices incubated in the NOS inhibitors were reduced to about the same level observed in APS-treated slices. C: representative examples of population EPSPs recorded from slices conditioned in the 1-train or 2-train models and incubated in control medium or the NOS inhibitor NOARG. In all cases the larger potential is the postconditioning response. Values represent means SE.: see text for groun n values.
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (192.236.036.029) on December 19, 2018.
NITRIC
OXIDE
AND
HIPPOCAMPAL
population EPSP slope during the preconditioning incubation period (Fs < 1). In the l-train conditioning model, slices incubated in drug-free medium exhibited moderate levels of synaptic potentiation that were still significantly enhanced at 60 min, relative to preconditioning control values (Fig. 1A). Application of NOARG at 0.1 (not shown in Fig. 1A), 50 and 200 PM produced no reduction of potentiation; rather, the group incubated in 200 PM NOARG had higher levels of potentiation throughout the 60-min test period. NMMA ( 100 FM) was likewise ineffective in reducing these levels. Incubation in AP5 (50 ,uM) reduced potentiation. Statistical analyses demonstrated a significant main effect for treatment (F = 3.498; P = 0.008) and a significant effect of time (F = 3.549; P = 0.002), but no significant interaction between time and treatment (F = 1.326; P = 0.122). The post hoc comparisons revealed significant differences from the drug-free control group only for the 200-PM NOARG group, and potentiation levels in this group were significantly enhanced by the NOS inhibitor (Fig. 1, A and c>. Although potentiation was depressed in the AP5 group, this did not reach statistical significance. The 2-train model produced a greater level of synaptic potentiation in control slices. The difference between control potentiation levels produced by the two different conditioning models was highly significant (F = 20.423; P < 0.000 1). In this experiment, incubation in either NOARG (50 or 100 ,uM; data for 50 ,uM not shown in Fig. 1L3) or NMMA produced a profound reduction of potentiation (Fig. 1, B and C). The NMDA antagonist AP5 reduced potentiation to a level almost identical to that seen in the NOS inhibitors. Statistical analyses revealed that synaptic potentiation in groups of slices incubated in NOS inhibitors or AP5 were significantly different from drug-free control values (significant main effects for treatment [F = 10.159; Y
