γ-Aminobutyric Acid-A Receptors Contribute to Isoflurane Neuroprotection in Organotypic Hippocampal Cultures

Abstract

The mechanisms by which anesthetics such as isoflurane reduce cell death in rodent models of cerebral ischemia remain incompletely defined. Reduction in glutamate excitotoxicity explains some but not all of isoflurane’s neuroprotection. Because isoflurane potentiates γ-aminobutyric acid (GABA) receptor-mediated ion fluxes and GABA_A receptor agonists have neuroprotective effects, we hypothesized that GABA_A receptors contribute to isoflurane neuroprotection. As a model of cerebral ischemia and recovery, we used rat hippocampal slice cultures. Survival of CA1, CA3, and dentate neurons was examined 2 and 3 days after 1-h combined oxygen-glucose deprivation (OGD) at 37°C. To define the role of GABA_A receptors in mediating protection, the effect of 1% isoflurane on cell survival was examined in the presence of the GABA_A antagonist bicuculline during OGD. Cell death was measured with propidium iodide fluorescence. Isoflurane and the selective GABA_A agonist muscimol (25 μM) reduced cell death after OGD to values similar to slices not exposed to OGD, with the exception that muscimol did not reduce cell death in CA3 neurons 2 days after OGD. The GABA_A antagonist bicuculline reduced the neuroprotective effects of isoflurane on hippocampal neurons 2 and 3 days after OGD. We conclude that GABA_A receptors contribute to neuroprotection against OGD produced by isoflurane in the hippocampal slice model. Based on this and other studies, it is likely that neuroprotection produced by isoflurane is multifactorial and includes actions at both GABA_A and glutamate receptors and possibly other mechanisms.

In Brief

IMPLICATIONS: Isoflurane is neuroprotective in rodent brain ischemia models, but the mechanisms for this effect remain incompletely defined. In organotypic cultures of rat hippocampus, we show that protection of CA1, CA3, and dentate neurons by 1% isoflurane from death caused by oxygen and glucose deprivation involves GABA_A receptors.

Isoflurane reduces brain damage resulting from both focal and global experimental cerebral ischemia ^(1–4). Despite numerous investigations in both intact animals and in in vitro models of brain ischemia, the mechanisms responsible for these neuroprotective actions remain unclear. A proposed mechanism for neuroprotection by isoflurane is amelioration of glutamatergic neurotoxicity either by reducing ischemia-induced release and accumulation of glutamate in cerebral tissue ⁽⁵⁾ or inhibition of glutamate receptors ^(6,7).

The neuroprotective qualities of isoflurane have also been tested in in vitro models of cerebral ischemia. Isoflurane protects brain slices from acute injury and death, an effect possibly related to direct inhibition of N-methyl-d-aspartate (NMDA) or glutamate toxicity ^(8,9). However, isoflurane is not a potent antagonist of NMDA receptors compared with compounds such as MK-801. For example, isoflurane is relatively weak in reducing NMDA-induced death in cultured neurons ⁽⁷⁾ and it does not completely block glutamate toxicity in hippocampal slice cultures ⁽¹⁰⁾. Previous in vitro studies investigating anesthetic neuroprotection suggested that interactions between volatile anesthetics and the γ-aminobutyric acid receptor, type A (GABA_A), may also contribute to neuroprotection ⁽⁷⁾. However, these interactions have not been specifically tested.

An important pharmacologic effect of isoflurane is the enhancement of GABA_A-receptor activity in the hippocampus ⁽¹¹⁾. Because GABA_A receptor agonists, such as benzodiazepines or barbiturates, are neuroprotective in the hippocampus ⁽¹²⁾, we hypothesized that part of isoflurane’s neuroprotection is because of enhancement of GABA_A receptors. Accordingly, the purpose of this study was to determine if GABA_A receptors play a role in the reduction of neuron death in hippocampal neurons subjected to oxygen and glucose deprivation (OGD) in the presence of isoflurane. We used the same organotypic slice culture model of the hippocampus that was used to show that isoflurane reduces, but does not eliminate, cell death in CA1, CA3, and dentate neurons observed for 2 wk after a 1-h period of OGD ⁽¹⁰⁾. The presence of intact glutamatergic and GABAergic synapses in the cultured slice model is an important similarity with the hippocampus of intact animals and different from other in vitro models (dissociated cultures) used to study neuroprotective effects.

Methods

All studies were approved by the University of California San Francisco (UCSF) Committee on Animal Research and conform to relevant National Institutes of Health guidelines.

Organotypic cultures of the hippocampus were made as described by Stoppini et al. ⁽¹³⁾ and Laake et al. ⁽¹⁴⁾ and modified by Sullivan et al ⁽¹⁰⁾. Briefly, 8–14-day-old Sprague-Dawley rats were anesthetized with halothane and given an intraperitoneal injection of ketamine (10 mg/kg) and diazepam (0.2 mg/kg). The rats were decapitated, and the hippocampi were removed and placed in 4°C Gey’s Balanced Salt Solution. Next, the hippocampi were transversely sliced (400-μm thick) with a MacIlwain tissue chopper or a wire grid-type tissue slicer (Siskiyou Design Instruments, Grants Pass, OR) and stored in Gey’s Balanced Salt Solution containing 0.038 mg/mL of ketamine and 50 μM of adenosine at 4°C for 1 h ⁽¹⁵⁾. No ketamine was present in the culture media beyond this point. The slices were then transferred onto 30-mm-diameter membrane inserts (Millicell-CM, Millipore, Bedford, MA) and put into 6-well culture trays with 1.5 mL of slice culture medium per well. The slice culture medium consisted of 50% Minimal Essential Medium (Eagles with Earle’s balanced salt solution, UCSF Cell Culture Facility), 25% Earle’s balanced salt solution (UCSF Cell Culture Facility), 25% heat inactivated horse serum (Hyclone Laboratories, South San Francisco, CA) with 6.5 mg/mL of glucose, and 5 mM of KCl. The culture media was changed every 2–3 days. Slices were kept in culture for 7–14 days before study.

In vitro OGD was achieved by anoxia combined with glucose-free media. The slices were washed three times with glucose-free Hank’s balanced salt solution (HBSS). The cultures were then placed into a 2-L airtight Billups-Rothenberg Modular Incubator chamber (Del Mar, CA) through which 95% N₂-5% CO₂ gas, preheated to 37°C, was passed at 5–10 L/min. The temperature of the chamber was kept at 37°C by both passing preheated gas through the chamber and by placing a heat lamp over the chamber. The temperature inside the chamber was monitored with a thermocouple thermometer. After 10 min of gas flow, the chamber was sealed and placed in a 37°C incubator. The partial pressure of oxygen was <0.2 mm Hg, measured with a Clark-type oxygen electrode. For studies involving isoflurane, a calibrated isoflurane vaporizer was used to deliver 1% isoflurane in the anoxic gas. The isoflurane remained in the chamber during the insult. A Datex OSCAR multigas analyzer (Datex, Helsinki, Finland) was used to measure isoflurane concentration in the inflow gas. In slices treated with the GABA_A antagonist bicuculline, slices were rinsed with glucose-free HBSS containing 100 μM of bicuculline; the bicuculline remained for the duration of the 1-h insult. The 100-μM concentration of bicuculline was chosen based on studies involving different bicuculline concentrations in brain slices; 100 μM is the minimum concentration demonstrated to produce a maximal antagonism of GABA_A-mediated neurotransmission ⁽¹⁶⁾. After the insult, the culture tray was removed from the chamber, the anoxic-glucose-free HBSS was aspirated from the wells, and standard (oxygenated) slice culture media was added. The GABA_A agonist muscimol was used similarly; a 25-μM concentration was used because this results in minimal peak augmentation of GABA_A receptor-mediated effects in hippocampal slices ⁽¹⁷⁾.

Cell viability was assessed with propidium iodide (PI) fluorescence (Molecular Probes, Eugene OR). PI, a highly polar fluorescent dye, penetrates damaged plasma membranes and binds to DNA. Before imaging, slice culture media containing 2.3 μM of PI was added to the wells of the culture trays. After 1 h, the slices were examined with a Nikon Diaphot 200 inverted microscope (Nikon, Melville, NY), and fluorescent digital images were taken using a SPOT Jr. Digital Camera (Kodak). Excitation light wavelength was 520 nm, and emission was 640 nm. The sensitivity of the camera and intensity of the excitation light was standardized to be identical from day to day. PI fluorescence was measured in the dentate gyrus, CA1, and CA3 regions of the hippocampal slices. Slices were discarded if they showed more than slight PI fluorescence in these regions after 7–10 days in culture. Slices were imaged before OGD (signal equated to 0% cell death) and after 2 and 3 days after OGD. In preliminary studies, we observed that maximum post-OGD death occurred by Day 3 and did not increase significantly by 7–14 days ⁽¹⁰⁾. Serial measurements of PI fluorescence intensity were made in predefined areas (manually outlining CA1, CA3, and dentate separately) for each slice using NIH Image software (developed at the United States National Institutes of Health and available on the Internet at http://rsb.info.nih.gov/nih-image/). Thus, cell death was followed in the same regions of each slice after OGD. After the measurement of PI fluorescence on the 3rd post-OGD day, all the neurons in the slice were killed to produce a fluorescence signal equal to 100% neuron death. This was completed by adding 100 μM of KCN and 2 μM of iodoacetate to the cultures for at least 20 min. One day later, final images of PI fluorescence (equated to 100% cell death) were acquired. Percentage of dead cells at 0, 2, and 3 days after OGD were then calculated based on these values. The relationship between cell death and PI fluorescence intensity is linear ^(14,15).

Only slices showing negligible baseline PI fluorescence after 7–14 days in culture were selected for study. The study was designed to determine if selectively blocking GABA_A receptors with bicuculline altered the capacity of isoflurane to reduce cell death after OGD and to determine if a GABA_A agonist mimicked the capacity of isoflurane to protect CA1, CA3, and dentate neurons. In the studies involving bicuculline treatment, groups included controls (no OGD) with or without bicuculline, OGD, OGD plus isoflurane, and OGD plus isoflurane and bicuculline. In the studies with the GABA agonist muscimol, treatments included controls (no OGD) with or without muscimol, OGD, and OGD with muscimol. The numbers of slices in each treatment group ranged from 7 to 17 and were prepared from 2 to 3 animals in each group.

Initial analysis indicated that cell death was not normally distributed. Therefore, the Kruskal-Wallis test followed by the Mann-Whitney U-test (Statistica, Statsoft, Inc, Tulsa, OK) was used to compare different treatment groups to each other for the same slice regions at 2 and 3 days after OGD. Differences were considered significant for P < 0.05. Data are presented as mean ± se.

Results

During a 3-day period after OGD, death was greatest in CA1 neurons (23% ± 3%), intermediate in CA3 neurons (15% ± 3%), and least in dentate neurons (8% ± 2%). In previous studies with this same model, cell death in CA1, CA3, and dentate peak near Day 3 after OGD ⁽¹⁰⁾. Isoflurane reduced cell death after OGD in each of these regions (Fig 1, A–C; P < 0.01). The addition of the GABA_A antagonist bicuculline to the cultures reduced the protective effect of isoflurane in CA1, CA3, and dentate at 2 and 3 days after OGD. That is, there was a significant increase in cell death in the slices treated with bicuculline combined with isoflurane and OGD compared with isoflurane and OGD alone (indicated by # in Fig. 1). At 3 days after OGD, bicuculline treatment eliminated the difference between cell death in slices treated with OGD alone and those treated with isoflurane and OGD, whereas at Day 2, these groups were distinct (indicated by * in Fig. 1), suggesting that GABA receptors play a major role in isoflurane neuroprotection. Bicuculline did not increase cell death when present alone (no OGD) and only slightly when present during OGD in CA1 and CA3 (P = 0.001–0.02). It is thus unlikely that the effect of bicuculline in reducing the neuroprotective action of isoflurane was simply because of an increase in cell death caused by GABA_A antagonism. Examples of slices from each treatment group are shown in Figure 2.

Figure 1: (A) Effects of isoflurane (1%) and the GABA_A antagonist bicuculline (100 μM) on cell death in CA1 neurons in cultured hippocampal slices after oxygen and glucose deprivation (OGD; 1 h at 37°C). Data are mean and se, with the number of slices in each group indicated in the legend. *Significant difference between OGD group and bicuculline/isoflurane/OGD group; P = 0.007. #Significant difference between isoflurane/OGD group and bicuculline/isoflurane/OGD group; P = 0.01–0.02. (B) Effects of isoflurane and the GABA_A antagonist bicuculline on cell death in CA3 after OGD. *P = 0.016 between OGD and bicuculline/isoflurane/OGD; #P = 0.01 between isoflurane/OGD and bicuculline/isoflurane/OGD. (C) Effects of isoflurane and the GABA_A antagonist bicuculline on cell death in dentate after OGD. *P = 0.017 between OGD and bicuculline/isoflurane/OGD; #P = 0.02–0.04 between isoflurane/OGD and bicuculline/isoflurane/OGD.

Figure 2: Example of propidium iodide (PI) fluorescence images of hippocampal slices 3 days after oxygen and glucose deprivation (OGD). Dark areas indicate cells that have taken up PI, i.e., they are dead. Bicuc, Iso, OGD = 100 μM bicuculline and 1% isoflurane present during OGD; Bicuc. = bicuculline 100 μM, no OGD; Bicuc., OGD = bicuculline present during OGD.

Slices in these studies were prepared with a minor difference in technique compared with those in the studies with bicuculline. The wire-grid tissue slicer used here prepared all slices from one hippocampus simultaneously, reducing early damage from handling and producing more uniform slices with larger numbers of OGD-vulnerable neurons in the various strata of the slice. The effect of this was to increase the percentage of dead neurons after the standard OGD insult (compare cell death in control OGD slices in CA1 in Figs. 1 and 3). In slices prepared this way, cell death 3 days after OGD was greatest in CA3 (45% ± 4%), somewhat less in CA1 (39% ± 5%), and least in dentate (14% ± 4%). At 2 and 3 days after OGD, the GABA_A agonist muscimol decreased cell death in CA1 (P = 0.001–0.029;Fig. 3) compared with the group treated only with vehicle (OGD). This protective effect of muscimol was evident only 3 days after OGD in CA3 and dentate (P = 0.036). Further, cell death in the muscimol OGD group was larger than in the control group for both the 2- and 3-day time points in the CA3 cells. This lesser protective effect of muscimol on CA3 neurons is clearly seen in the example in Figure 4. Muscimol, when present by itself, did not alter the percentage of living cells in any of the slice regions compared with untreated (vehicle only) controls.

Figure 3: (A) Effects of the γ-aminobutyric acid (GABA)_A agonist muscimol (25 μM) on cell death in CA1 neurons in cultured hippocampal slices after oxygen and glucose deprivation (OGD; 1 h at 37°C). *P = <0.001–0.003 between muscimol/OGD and OGD. (B) Effects of muscimol on cell death in CA3 neurons after OGD. *P = 0.036 between OGD and muscimol/OGD; #P = 0.001 = 0.011 between control (no OGD) and muscimol/OGD. (C) Effects of muscimol on cell death in dentate neurons after OGD. *P = 0.029 between OGD and muscimol/OGD.

Figure 4: Example of propidium iodide (PI) fluorescence images of hippocampal slices 3 days after oxygen and glucose deprivation (OGD). Muscimol OGD = muscimol (25 μM) present during OGD. Note the lack of muscimol protection of neurons in the CA3 region in the muscimol OGD slice.

Discussion

We found that the volatile anesthetic isoflurane, in an in vitro model of cerebral ischemia and recovery, reduces delayed cell death by a mechanism dependent on GABA_A receptors. This conclusion is based on the following two observations: First, antagonizing GABA_A receptors eliminated the protection afforded by isoflurane against cell death 3 days after the OGD insult, when cell death in this model peaks. Second, a selective GABA_A agonist reduced neuron death in this model of OGD. Because antagonizing GABA_A receptors did not completely reverse the protection of isoflurane at all time points and cell regions and a GABA_A agonist was not a potent neuroprotectant in CA3, the possibility remains that other mechanisms contribute to isoflurane neuroprotection.

Previous studies have found that isoflurane and other volatile anesthetics protect against selective neuronal necrosis resulting from near-complete forebrain ischemia in vivo^(2,3,18), against focal cerebral ischemic insults ⁽¹⁾, and against excitotoxic (glutamate or NMDA) or anoxic injury in cultured neurons ^(7,19), acute brain slices ^(8,9), and brain slice cultures ⁽¹⁰⁾. In addition, isoflurane administered before ischemia can induce ischemic tolerance in the brain ^(20,21). Because glutamate receptors play a central role in ischemic neuron death both in intact animals and in in vitro models ^(9,15,22) and isoflurane inhibits glutamatergic processes, it has been suggested that isoflurane neuroprotection involves glutamate receptors. Reduction of glutamate excitotoxicity is consistent with data showing that isoflurane decreases NMDA-receptor activity ^(6,23). Additionally, volatile anesthetics such as isoflurane may increase glutamate uptake ⁽²⁴⁾ and suppress glutamate release ^(5,25), contributing to reduced excitotoxicity. However, unlike the case with isoflurane, selective glutamate receptor antagonism is effective against only focal ^(26–28) but not global ischemic insults in vivo^(29–32). NMDA receptor antagonists such as MK-801 protect against cerebral injury in focal but not global ischemia.

The cumulative data suggest that isoflurane’s neuroprotective action, at least in the case of focal ischemia, may involve glutamate receptor antagonism as an important mechanism of action. In the case of global ischemia, this seems less likely. Indeed some evidence suggests that the antiglutamate effect of isoflurane is not potent, particularly in dissociated neuron cultures ^(7,19). Further, Sullivan et al. ⁽¹⁰⁾ found that isoflurane did not completely block glutamate toxicity (500 μM of glutamate applied to organotypic hippocampal slices) via a NMDA-receptor-dependent mechanism but that the specific NMDA antagonist MK-801 was a potent protectant. Isoflurane is a relatively weak antagonist of the NMDA receptor, both in comparison to antagonists such as MK-801 in reducing cell death from NMDA exposure in cultured neurons ⁽⁷⁾ and in terms of the percentage inhibition of NMDA-receptor activity at clinical concentrations. The 50% effective concentration of isoflurane for inhibition of recombinant NMDA receptors expressed in Xenopus oocytes is 1.3 minimum alveolar anesthetic concentration ⁽³³⁾, suggesting it is not an exceptionally sensitive site for anesthetic or neuroprotective action.

Abundant evidence suggests that GABA receptors can play a role as a site of neuroprotective action ⁽¹²⁾. Importantly, GABA receptors have significant influence on NMDA-receptor activity. This is because NMDA-receptor currents depend on the membrane potential. In the presence of physiological concentrations of magnesium, the NMDA receptor is inhibited by membrane hyperpolarization such as that produced by GABA_A-mediated chloride currents ⁽³⁴⁾. Thus, GABA- and NMDA-receptor modulation are complexly intertwined.

Thus, available evidence suggests that isoflurane both inhibits glutamate toxicity and augments GABAergic processes. This is significant for two reasons. First, isolated antagonism of glutamate receptors (particularly NMDA receptors) is neurotoxic to neurons in the cingulate cortex ⁽³⁵⁾. The reduction of NMDA antagonist neurotoxicity can be achieved with GABA agonists ⁽³⁵⁾. Importantly, halothane reduces neurotoxicity from the NMDA antagonist MK-801 ⁽³⁶⁾. Second, the simultaneous inhibition of glutamate receptors and activation of GABA receptors may produce synergistic neuroprotective effects. The combination of GABAergic and antiglutamatergic strategies has produced powerful synergistic neuroprotection in intact animal models of cerebral ischemia ⁽³⁷⁾. Therefore, volatile anesthetics such as isoflurane, which act via these two mechanisms simultaneously, offer an important model for neuroprotection that may ultimately be clinically significant, particularly in the case of global ischemia, which is resistant to many forms of experimental pharmacologic intervention that otherwise are effective against focal ischemic insults.

The neuroprotective effect of the GABA_A agonist muscimol against cell death in CA1 but not in CA3 neurons was interesting. This may indicate that GABA_A-based neuroprotection may be regionally restricted, possibly depending on the distribution of particular GABA receptor subtypes or other variables. The GABA_A receptor subunits α 5 and α 6 are confined to CA3 pyramidal neurons in the hippocampus ⁽³⁸⁾, and muscimol binding to GABA_A receptors varies with the type of GABA_A receptor subunits present, although relatively little effect of α subunit on muscimol binding was noted by Ebert et al ⁽³⁹⁾. Murine knockouts of GABA receptor subunits or antisense oligonucleotide knock downs of the various receptor subunits may provide useful tools to better define the site of isoflurane neuroprotective action.

The combination of GABAergic and antiglutamate effects of isoflurane may be a feature of the intact brain or of slice cultures but less so in dissociated cultures of neurons. As shown by Kudo et al. ⁽⁷⁾, isoflurane has comparatively little potency to reduce glutamate receptor toxicity in dissociated neuron cultures, possibly because GABA receptor-based processes are present at a low level of activity. Isoflurane does not directly activate GABA receptors; rather, it potentiates currents flowing through the receptor ion channel when GABA or another agonist is present ⁽⁴⁰⁾. This may explain the relatively limited neuroprotection of isoflurane in dissociated neuron cultures; there is probably little GABA release and accumulation during simulated ischemia in these cultures and thus limited opportunity for isoflurane to enhance GABA currents and initiate protective actions based on the GABA receptor.

Only one concentration of muscimol and one concentration of bicuculline were examined in this study because we were interested in the mechanisms, not the potency, of isoflurane neuroprotection. Although there were sound pharmacologic reasons for choosing these doses (in each case, the concentration was the minimal concentration required to maximally block or augment GABA_A receptors in hippocampal slices ^(16,17) (see Methods), our conclusions may be limited. A further limitation was that the study involved only a single concentration of isoflurane. The conclusion that neuroprotective effects of 1% isoflurane involve a combination of GABA and glutamatergic mechanisms may not be valid at other isoflurane concentrations.

In summary, our studies show that preventing the action of GABA_A receptors in a hippocampal slice model of OGD reduces isoflurane neuroprotection. In addition, we found that augmenting GABA_A activity with an agonist (muscimol) produces neuroprotection. We conclude that the GABA_A receptor contributes to neuroprotection produced by 1% isoflurane in the organotypic hippocampal slice culture model.