Thiopental Inhibits Increases in [Ca²⁺]ⁱInduced by Membrane Depolarization, NMDA Receptor Activation, and Ischemia in Rat Hippocampal and Cortical Slices 

This study was approved by the Committee for the Guidelines on Animal Experimentation of Niigata University. Experiments were done on brain slices obtained from male Wistar rats that weighed 100 - 130 g and were 30 - 35 days old. Rats were anesthetized with isoflurane and decapitated, and their brains were quickly removed and placed into ice-cold oxygenated Krebs solution, which contained 117 mM NaCl, 3.6 mM KCl, 1.2 mM NaH²PO⁴, 2.5 mM CaCl², 1.2 mM MgCl², 25 mM NaHCO³, 11.5 mM glucose, and had a pH of 7.35 (at 23 [degree sign]C equilibrated with 95% oxygen and 5% carbon dioxide. Brains were coronally cut into 350 [micro sign]m-thick sections using a vibrating tissue slicer (DTK-1000, Dosaka, Kyoto, Japan). Hippocampi and somatosensory cortices were isolated.

To load fura-2, slices were incubated in oxygenated Krebs solution containing 10 [micro sign]M acetoxymethyl ester of fura-2 (fura-2/AM; Dojindo, Kumamoto, Japan) and 0.1% Pluronic F-127 (Boehringer Diagnostics, La Jolla, CA) for 90 min at room temperature using an interface chamber. These slices were rinsed twice and incubated again in oxygenated Krebs solution for an additional 20 - 30 min to allow hydrolysis of the acetoxymethyl ester.

Fura-2 - loaded slices were transferred individually into a cylindrical chamber glued onto a microscope glass slide that was mounted on the stage of an inverted fluorescence microscope (Diaphot, Nikon, Tokyo, Japan). The chamber had an effective volume of 0.3 ml, and the slices were superfused with Krebs solution, equilibrated with 95% oxygen and 5% carbon dioxide. The flow rate was 5 ml/min, and the temperature was maintained within 36.5 to 37.5 [degree sign]C using a heating bath. The levels of [Ca (²⁺)]ⁱin the CA1 pyramidal cell layer of hippocampal slices or layers II to III of cortical slices were monitored as the fluorescence of fura-2 using a microspectrofluorimeter (JASCO CAM-500, Japan Spectroscopic Co., Tokyo, Japan), as previously described. [8]Excitation lights were obtained from a xenon lamp (150 W) equipped with a rotating wheel filter and two monochromators. The intensity of fura-2 fluorescence at 510 nm emission was obtained at 340 and 380 nm excitation wavelengths. Changes in the intensities of fluorescence at 340 (I³⁴⁰) and 380 nm (I³⁸⁰) collected from a 100 [micro sign]m x 25 [micro sign]m area of slices with an adjustable diaphragm inserted in the light path were detected using a photo-multiplier at 1 Hz. The I³⁴⁰and I³⁸⁰values and their ratio (I^340/I380, referred to as R340/380) were recorded simultaneously using a chart writer. Estimates of [Ca²⁺]ⁱwere represented by R340/380.

Aglycemic anoxia in slices was made by switching the 95% oxygen and 5% carbon dioxide-gassed Krebs solution to glucose-free Krebs solution (supplemented with equimolar sucrose) that was preequilibrated with 95% nitrogen and 5% carbon dioxide (this decreased the oxygen tension of the perfusate in the recording chamber from 484 +/- 20 mmHg to 30 +/- 5 mmHg). Changes in [Ca²⁺]ⁱcaused by aglycemic anoxia were evaluated based on (1) the latency from the onset of aglycemic anoxia to the appearance of the [Ca²⁺]ⁱplateau, and (2) the changes in [Ca²⁺]ⁱ8, 10, and 15 min after the onset of aglycemic anoxia, which are expressed as the normalized changes in R340/380 calculated by dividing R340/380 at those points to just before the onset of aglycemic anoxia. The [Ca²⁺]ⁱplateau indicates the period from the finish of rapid [Ca²⁺]ⁱelevated to the point at which 15 min of ischemia was stopped.

Membrane depolarization and NMDA receptor activation in slices were produced by exposure to 60 mM K⁺(60 s) and 100 [micro sign]M NMDA (30 s), respectively. To block membrane depolarization-generated sodium influx, 0.5 [micro sign]M tetrodotoxin was added to the bathing media before, during, and after exposure to high K⁺or NMDA. Because the [Ca²⁺](ⁱ) response caused by 60 mM K⁺was completely repeatable in the same slice, the change in [Ca²⁺]ⁱcaused by first exposure to 60 mM K (⁺) was used as a control in each slice. The high K^+-mediatedincreases in [Ca²⁺]ⁱin the presence of thiopental were normalized to the control levels by the net increase in R340/380 in each slice. The changes in [Ca²⁺]ⁱevoked by 100 [micro sign]M NMDA were represented by the net increase in R340/380. Interventions were started 5 min before the onset of aglycemic anoxia, 60 mM K⁺or 100 [micro sign]M NMDA exposure, and were continued at the same concentration during aglycemic anoxia.

The effect of thiopental on the NMDA-mediated change in [Ca²⁺]ⁱwas also examined in cultured cortical neurons. Cortical neurons were prepared from 18- to 19-day-old embryos of Wistar rats according to the procedure of Mizuno et al. [9]Whole neocortices were mechanically dissociated and plated onto glass coverslips (1 cm in diameter and 0.17 mm thick) that were precoated with poly-L-lysine and laminin at a density of 1 [tilde operator] 2 x 10³cells/mm². Dissociated cells were first grown for 24 h with Dulbecco's modified Eagle's medium containing 2 mM purified glutamine, 2% fetal bovine serum, a nutrient mixture of N2 (100 [micro sign]g/ml transferrin, 5 [micro sign]g/ml bovine insulin, 100 nM putrescine, 20 nM progesterone, and 30 nM sodium selenite), 10 mM HEPES (pH 7.3), and 100 [micro sign]g/100 ml bovine serum albumin (all from Sigma Chemical Co., St. Louis, MO). The next day, the original medium was replaced by the medium lacking serum (serum-free N2 medium), which topically contained 20 ng/ml purified human recombinant neurotrophines (i.e., NGF, BDNF, NT-3, or NT-5). Cultures were maintained at 37 [degree sign]C in humidified 6% carbon dioxide-94% air atmosphere, and the media were replaced every 2 or 3 days. After 7 - 9 days in culture, cells were washed with Krebs solution and then incubated in Krebs solution containing 5 [micro sign]M fura-2-AM and 0.1% Pluronic F-127 for 45 min at 37 [degree sign]C to load fura-2. After the loading period, the cells were incubated in fura-2 - free Krebs solution for an additional 30 min at 21 - 23 [degree sign]C. The coverslips were then attached to a chamber (effective volume, 0.4 ml) and placed on the stage of a Nikon Diaphot inverted microscope equipped with a 100 x oil immersion objective (numeric aperture, 1.3). Cells were superfused continuously with oxygenated Krebs solution at a rate of 4 ml/min, and the temperature was maintained at 37 [degree sign]C. Cells were illuminated alternately with 340 and 380 nm light from a xenon lamp (150 W) and were recorded using an SIT video camera (model C2400–8; Hamamatsu Photonics, Hamamatsu, Japan) controlled by software (ARGUS 200/CA, Hamamatsu Photonics) running on a personal computer (model A2008; Canon Inc., Tokyo, Japan). Hardware gain and black settings were optimized so that the somatic region of the neurons could be seen clearly. Images were sampled and collected at 1-min intervals for 44 min. Three or four neurons were monitored simultaneously in each coverslip. The protocol for eliciting the effect of thiopental on the NMDA-mediated increase in [Ca²⁺]ⁱwas similar to that in the slice preparation, except the concentration of NMDA was 200 [micro sign]M and the duration of exposure was increased to 2 min.

Drugs were applied by superfusion. DL-2-amino-5-phosphonovaleric acid, NMDA, tetrodotoxin, bicuculline, and [small gamma, Greek]-amino-butyric acid (GABA) were purchased from Sigma Chemical Company. Lanthanum chloride was obtained from WAKO Chemical Industries (Osaka, Japan). Sodium thiopental was a gift of Tanabe Pharmaceutical Co. (Osaka, Japan).

Data are expressed as mean +/- SD and were analyzed, as appropriate, by Student's t test, one- or two-way analysis of variance, with statistical significance determined by Dunnett's post hoc test. A P value <or= to 0.05 was considered significant.

Control experiments were performed in the fura-2 - unloaded hippocampal slices to determine how autofluorescence would affect the R340/380 after the onset of aglycemic anoxia, or application of 60 mM K⁺and 100 [micro sign]M NMDA. Figure 1A shows that exposing slices to aglycemic anoxia caused an increase in R340/380 with concomitant increases in both I³⁴⁰and I³⁸⁰. The increases in R340/380 began within 45 s after the onset of aglycemic anoxia and reached a plateau within 1 min. In seven slices examined, the increases in R340/380 caused by aglycemic anoxia are 110 - 117% at 2 min and 115 - 125% at 15 min after the onset of aglycemic anoxia (R340/380 at the two time points compared with those before the onset of aglycemic anoxia). Based on this, it is possible to subtract autofluorescence-related changes in R340/380 in the fura-2 - loaded slices. Exposure of slices to 60 mM K⁺(n = 5) and 100 [micro sign]M NMDA (n = 5) did not affect R340/380 significantly (Figure 1B and Figure 1C).

In the fura-2 - loaded slices, the intensities of fluorescence at 340 nm and 380 nm were 8 - 10 times higher than those in the fura-2 - unloaded slices. The use of 0.1% Pluronic F-127 did not harm slices within 6 h, as evaluated by population spikes (data not shown). As shown in Figure 2A, in hippocampal or cortical slices responding to aglycemic anoxia for 15 min, a characteristic increase in R340/380 was observed: an initial slow phase was followed by a rapid elevation and then reached a plateau phase. The initial increases in R340/380 may not reflect changes in [Ca²⁺]ⁱbecause they were observed with increases in both I³⁴⁰and I³⁸⁰. The rapid increases of [Ca²⁺]ⁱin the pyramidal cell layer of the CA1 region began 4.5 to 6.5 min after the onset of aglycemic anoxia and reached the plateau phase within 1 min. In layers II to III of cortical slices, latencies to the rapid increase in [Ca²⁺]ⁱwere shorter than those in hippocampal slices, ranging from 3.5 to 5.5 min. The latencies to the occurrence of the [Ca²⁺]ⁱplateau after the onset of aglycemic anoxia are 416.0 +/- 23.2 s in hippocampal slices (n = 9) and 348 +/- 60.8 s in cortical slices (n = 6; P < 0.05 by unpaired t test), respectively. Table 1lists the normalized increases in R340/380 at 8 (R (⁸)), 10 (R¹⁰), and 15 min (R¹⁵) after the onset of aglycemic anoxia in both hippocampal and cortical slices. The R⁸, R¹⁰, and R (¹⁵) values were calculated by dividing R340/380 at 8, 10, and 15 min after the onset of aglycemic anoxia by R340/380 at 2 min after the onset of aglycemic anoxia but not that before the onset of aglycemic anoxia in each slice to offset autofluorescence-related changes in R340/380.

Pretreatment with thiopental (50 - 400 [micro sign]M) did not affect basal [Ca²⁺]ⁱin either hippocampal or cortical slices. The pattern of increase in [Ca²⁺]ⁱcaused by aglycemic anoxia, however, was dose dependently altered by thiopental, which slowed the rapid increase phase, prolonged the [Ca²⁺]ⁱplateau, and reduced R⁸, R¹⁰, and R¹⁵(Figure 2B). Table 1summarizes the [Ca²⁺]ⁱplateau and R⁸, R¹⁰, and R¹⁵in the hippocampal and cortical slices responding to aglycemic anoxia in the presence of thiopental. The calculations of R⁸, R¹⁰, and R¹⁵were identical to those in thiopental-untreated slices, as mentioned before.

In both hippocampal and cortical slices, the [Ca²⁺]ⁱresponse caused by membrane depolarization with 60 mM K⁺(60 s) could be completely repeated in the same slice at least three times, as evaluated by the net increase in R340/380 (Figure 3A and Figure 3A'). Thus changes in [Ca²⁺]ⁱinduced by 60 mM K⁺in the presence of thiopental were normalized with respect to the first exposure (referred to as control) by the net change in R340/380 in each slice. Pretreatment with thiopental for 5 min dose dependently reduced the 60 mM K^+-mediatedincreases in [Ca²⁺](ⁱ) in either hippocampal or cortical slices (Figure 3B, Figure 3C, Figure 3D and Figure 3B', Figure 3C', Figure 3D'). At the same concentration, the inhibition of thiopental on the 60 mM K^+-inducedincrease in [Ca²⁺]ⁱwas not significantly different for hippocampal and cortical slices (Figure 4). In addition, the depressant effect of thiopental on the 60 mM K^+-inducedincrease in [Ca²⁺]ⁱwas not affected by bicuculline (20 [micro sign]M), a selective GABA^Areceptor antagonist (Figure 5). The peak increase in [Ca²⁺]ⁱ(% control) caused by 60 mM K⁺is 40.5 +/- 8.9% after pretreatment with 400 [micro sign]M thiopental alone and 40.2 +/- 9.1% after pretreatment with a combination of 400 [micro sign]M thiopental and 20 [micro sign]M bicuculline (P > 0.05 by paired t test; n = 4).

In the presence of magnesium (1.2 mM), exposure to 100 [micro sign]M NMDA (for 30 s) caused a transient increase in [Ca²⁺]ⁱin both hippocampal and cortical slices, and the responses were not completely repeatable in the same slice, as evaluated by the net change in R340/380 (Figure 6A and Figure 6A'). The ratio of second to first responses is 0.63 +/- 0.09 in hippocampal slices (n = 8) and 0.55 +/- 0.10 in cortical slices (n = 6). The NMDA-mediated increases in [Ca²⁺]ⁱwere reversibly abolished by DL-2-amino-5-phosphonovaleric acid (50 [micro sign]M), a competitive NMDA receptor antagonist, but rarely were affected by lanthanum chloride (30 [micro sign]M), a nonspecific voltage-gated calcium channel (VGCC) blocker, [10,11]in either hippocampal or cortical slices (Figure 6B and Figure 6C and Figure 6B' and Figure 6C'). Thiopental also dose dependently reduced the NMDA-induced increase in [Ca²⁺]ⁱin both the pyramidal cell layer of the CA1 region and layers II to III of cortical slices, as shown in Figure 7and Figure 8. The [Ca²⁺]ⁱresponses caused by a second exposure to NMDA were significantly enhanced by increasing the concentration of thiopental in either hippocampal or cortical slices. The net increases in R340/380 produced by 100 [micro sign]M NMDA are 0.239 +/- 0.028 in hippocampal slices (n = 8) and 0.237 +/- 0.029 in cortical slices (n = 6; P > 0.05 by unpaired t test). The inhibition of thiopental on the NMDA-induced increases in [Ca²⁺]ⁱwere pronounced in layers II to III of cortical slices (Figure 7and Figure 8). In the presence of 400 [micro sign]M thiopental, the net increase in R340/380 caused by 100 [micro sign]M NMDA was 0.134 +/- 0.011 in the CA1 pyramidal cell layer (n = 6), compared with 0.018 +/- 0.014 in layers II to III of cortical slices (n = 6; P < 0.0001 by unpaired t test).

The effect of thiopental on the NMDA-mediated [Ca²⁺]ⁱresponse was also examined in 15 cultured cortical neurons of four sister coverslips. In the presence of 400 [micro sign]M thiopental, 7 of 15 neurons did not respond to NMDA (200 [micro sign]M), whereas 14 of 15 neurons responded to NMDA after thiopental was washed out. Figure 9shows the effect of thiopental on the NMDA-mediated increases in [Ca²⁺]ⁱin the somatic region of four neurons.

We found that thiopental could attenuate increases in [Ca²⁺](ⁱ) induced by membrane depolarization, NMDA receptor activation, and ischemia in both the hippocampus and cortex in vitro, and its attenuation of the NMDA-mediated increases in [Ca²⁺]ⁱwas more profound in layers II to III of the cortex.

The excessive increases in [Ca²⁺]ⁱhave been hypothesized to be responsible for neuronal injury under excitotoxic, anoxic, and ischemic insults. [1,2]Several studies [12,13]have revealed that net intracellular Ca²⁺accumulation measured by the radioactive⁴⁵Ca²⁺isotope correlates with neuronal injury, whereas most attempts have failed to show a correlation between neuronal injury and the levels of [Ca (²⁺)]ⁱ, measured using Ca^2+-sensitivedyes. [14,15]Given that a large amount of Ca²⁺can be sequestrated by internal organelles (such as mitochondria) and the peak levels of [Ca²⁺]ⁱcould not be estimated precisely using some Ca^2+-sensitivedyes, [16]the results obtained by the radioactive⁴⁵Ca²⁺isotope and by Ca^2+-sensitivedyes actually may not be in conflict. On the other hand, intramitochondrial Ca²⁺accumulation would cause the permeability transition pore to open, [17,18]further leading to mitochondrial dysfunction, which may be a key step preceding apoptotic cell death. [19] 

Using the fura-2 fluorescent technique for brain slice preparations enabled us to determine an average change in [Ca²⁺]ⁱof a heterogeneous population, including neuronal somata, dendrites, axons, and glial cells. A true elevation of [Ca²⁺]ⁱcan be identified as an increase in R340/380, with mirror changes of I³⁴⁰and I³⁸⁰. In both hippocampal and cortical slices, ischemia caused a typical increase in [Ca²⁺]ⁱ. The ischemia-induced [Ca²⁺]ⁱoverload was shown to result from several routes, including the opening of both voltage-gated and NMDA-gated calcium channels, reverse of Na^+-Ca2+exchange, [20]release of Ca²⁺from internal stores, and activation of nonselective ion conductance. [1,21]Thiopental delayed the occurrence of the [Ca²⁺]ⁱplateau and reduced the magnitude of increase in [Ca (²⁺)]ⁱduring ischemia. This result is consistent with the report that thiopental was effective in reducing anoxia-caused calcium influx in the CA1 region of hippocampal slices. [6]Because the occurrence of the [Ca²⁺]ⁱplateau was associated with the irreversible loss of synaptic activity, [22]attenuation of ischemia-induced increases in [Ca²⁺](ⁱ) may be an important mechanism for thiopental to protect neurons against ischemia.

It is conceivable that attenuation of ischemia-induced increases in [Ca²⁺]ⁱby thiopental may be related to its ability to reduce the cellular metabolic rate. Nevertheless, studies in vitro [6]and in vivo [23]have not shown that thiopental decreases energy consumption during anoxia or ischemia. Even in a slice model, thiopental accelerated the decrease in adenosine triphosphate level during anoxia. [6]Thus mechanisms other than decreases in energy consumption appear to be involved in its attenuation of ischemia-induced increases in [Ca²⁺]ⁱ. In the current study, we have provided evidence that attenuation of ischemia-induced increases in [Ca²⁺]ⁱby thiopental is parallel to its inhibitory effects of both VGCCs and NMDA receptors.

A potential explanation for the inhibition of thiopental on high K (^+-evoked) increases in [Ca²⁺]ⁱmay be its potentiation of GABA (^A) receptor activity. However, direct exposure of GABA (100 [micro sign]M) had no effect on the high K^+-evokedincrease in [Ca²⁺]ⁱ(data not shown), and bicuculline, a GABA^Areceptor antagonist, did not reduce thiopental-mediated inhibition, suggesting that activation of GABA^Areceptors is not the cause. It is well known that increases in [Ca²⁺](ⁱ) induced by high K⁺are the result of membrane depolarization, which opens VGCCs. In non-neuronal cell lines, thiopental has been known as a VGCC blocker. [24,25]In rat striatal synaptosomes, thiopental was found to reduce high K^+-inducedGABA release. [26]When these observations are combined with our findings, we would expect that thiopental also works as a VGCC blocker to neurons, and its inhibition of VGCCs should contribute in part to its attenuation of ischemia-induced increases in [Ca²⁺]ⁱ.

Barbiturates at higher concentrations can block the NMDA receptor. [27]Because NMDA receptor channels are highly permeable to Ca²⁺, recording NMDA-mediated [Ca²⁺]ⁱresponses can be used as an indicator to assess the activity of NMDA receptors. In the presence of 1.2 mM Mg²⁺, NMDA caused a transient increase in [Ca²⁺]ⁱin either pyramidal cell layer of the CA1 region or layers II to III of cortical slices, and the response could be abolished by DL-2-amino-5-phosphonovaleric acid, indicating that the [Ca²⁺]ⁱresponses caused by NMDA initially resulted from NMDA receptor activation. Given that glial cells lack NMDA receptors [28]and their [Ca²⁺]ⁱlevels are insensitive to NMDA, [29]the NMDA-induced increases in [Ca²⁺]ⁱin both hippocampal and cortical slices appear to occur in neurons but not in glial cells. Indeed, the NMDA-mediated increase in [Ca²⁺]ⁱin a pattern that is comparable to the NMDA-induced inward current recorded from neostriatal neurons. [30]Although it is possible that the increases in [Ca²⁺]ⁱafter NMDA receptor activation may result in part from the opening of VGCCs secondary to membrane depolarization, we found that lanthanum chloride, a nonspecific VGCC blocker, [10,11]did not reduce the NMDA-mediated increase in [Ca²⁺]ⁱ.

In both hippocampal and cortical slices, the [Ca²⁺]ⁱresponse caused by a second exposure to NMDA was smaller than that by the first exposure in each slice, which was reversed by thiopental at higher concentrations. Two explanations are possible: First, thiopental might actually protect slices from NMDA-induced damage [31]; and second, a larger increase in [Ca²⁺]ⁱcaused by NMDA receptor activation may lead to long-term inactivation of NMDA receptors. [32,33]Interestingly, the effect of thiopental on the NMDA-induced increases in [Ca²⁺]ⁱvaries regionally; in the presence of 200 or 400 [micro sign]M thiopental, the NMDA-induced increases in [Ca²⁺]ⁱwere reduced to a greater extent in layers II to III of cortical slices than were those in the pyramidal cell layer of the CA1 region. The potent inhibition of thiopental on the NMDA-induced increases in [Ca²⁺]ⁱwas confirmed in cultured cortical neurons. In addition, recording the NMDA-induced currents in CA1 pyramidal neurons by blind whole-cell patch revealed that the inhibitory effect of thiopental (400 [micro sign]M) is only partial (data not shown), according to its effect on the NMDA-mediated increases in [Ca²⁺]ⁱ.

It is now clear that multiple NMDA receptor subtypes exist in the brain and differ in their anatomic distribution, pharmacologic profiles, and regulatory and physiologic properties. [34,35]Barbiturates differentially affecting NMDA receptors in the cortex and hippocampus have been suggested indirectly by two investigations, [36,37]in which pentobarbital and barbital increased [(³) H] dizocilpine binding sites in the cortex but not in the hippocampus. Because overstimulation of NMDA receptors is considered an important factor in causing ischemic neuronal injury, [1,2]it is reasonable to speculate that thiopental protecting neurons against ischemia would be region specific. Indeed, Sano et al. [7]showed that thiopental reduces ischemic neuronal injury in the cortex but not in the hippocampus.

The significance of our results might be as follows. First, a combined inhibition of both VGCCs and NMDA receptors may be one of the mechanisms by which thiopental protects neurons against ischemia. In an earlier study, Raley-Susman and Lipton [38]found that neither Ca²⁺removal nor ketamine alone, but combining Ca²⁺removal with ketamine, could prevent the inhibition of protein synthesis and the morphological changes caused by ischemia in the CA1 pyramidal neurons in vitro. Similarly, an in vivo investigation [39]also found that neither VGCC blocker nor NMDA receptor antagonist, but rather a combination of VGCC blocker and NMDA receptor antagonist could reduce ischemic neuronal injury. Furthermore, the protective effect of sigma ligands such as dextromethorphan in neurons against ischemia in vivo [40,41]also has been related to the inhibition of both VGCCs and NMDA receptors. [41]Second, although thiopental attenuated the increases in [Ca²⁺]ⁱcaused by membrane depolarization, NMDA receptor activation, and ischemia dose dependently, its inhibitions at lower concentrations were not significant, indicating that high doses of thiopental would be needed to achieve appreciable protection. Third, because calcium influx through NMDA receptor channels is especially neurotoxic, [42]a regionally differential inhibition of NMDA receptors by thiopental may result in a region-specific action against ischemia.

In conclusion, thiopental attenuates increases in [Ca²⁺]ⁱcaused by ischemia in both CA1 pyramidal cell layer of the hippocampus and layers II and III of the cortex in vitro, probably by its inhibition of VGCCs and NMDA receptors. The inhibition of thiopental on NMDA receptors is more profound in the cortex than in the hippocampus. These results indicate that stabilizing Ca²⁺homeostasis by inhibiting VGCCs and NMDA receptors may be one of the mechanisms through which thiopental protects neurons against ischemia, and its protection would be region specific.

The authors thank Professor Hiroyuki Nawa and Dr. Huabao Xiong, Department of Molecular Neurobiology, Brain Institute, Niigata University, for technical assistance in preparing the cultured cortical neurons; and Dr. Manabu Okamoto, Department of Anesthesiology, for performing the blind whole-cell patch in hippocampal slices.