Xenon Attenuates Hippocampal Long-term Potentiation by Diminishing Synaptic and Extrasynaptic N -methyl-D-aspartate Receptor Currents

• Amnesia is essential for successful anesthesia, but whether xenon, which antagonizes N -methyl-D-aspartate (NMDA) receptors, affects hippocampal circuit responses associated with memory formation and storage has not been examined

• In hippocampal slices from mice, xenon blocked long-term potentiation, a synaptic and circuit function linked to memory formation by a reduction of NMDA receptor-mediated responses

THE inhibition of both implicit and explicit memory formation is a fundamental component of general anesthesia. A failure of this component may contribute to intraoperative awareness, an adverse event during anesthesia that occurs in a stunningly high number of patients.1Despite the clinical importance of the memory-ablating properties of general anesthetics, their neuronal basis remains not entirely clarified.

Memory formation occurs at various sites in the brain. Among other regions,2the hippocampus has been suggested to be one of the key structures for anesthetics mediating their amnestic properties.3Long-term potentiation (LTP), a long-lasting enhancement of synaptic efficacy upon repetitive and/or simultaneous stimulation of different afferents, represents (especially in the hippocampus) an important and widely studied neuronal correlate for memory formation.4A major mechanism of LTP induction is an increased calcium influx through both N -methyl-D-aspartate (NMDA) receptors and voltage-gated calcium channels (VGCCs).5 

An impairment of synaptic plasticity has been reported for several anesthetics. The volatile anesthetics isoflurane6and sevoflurane7and the intravenous anesthetics propofol8and etomidate9attenuate hippocampal synaptic plasticity through γ-aminobutyric acid (GABA)-ergic mechanisms. It has been suggested that this effect during the application of the anesthetics may account for the desirable intraoperative amnesia.6,7,9 

During the last decade, the inhalational anesthetic xenon increasingly came into the focus of clinical and basic research because of its profound anesthetic and analgesic properties and low side effect profile. The safety and efficiency of xenon anesthesia has been evaluated in large clinical trials.10,11Pharmacologically, xenon antagonizes NMDA receptors12and thus also may affect the induction of LTP.

In the current study, we investigated in the CA1 region of the hippocampus the impact of xenon on LTP, NMDA receptor-mediated neurotransmission, and calcium channel currents in a murine brain slice preparation.

Experimental protocols were approved by the Ethical Committee on Animal Care and Use of the Government of Bavaria (Munich, Germany). After male C57BL/6 mice (p20–30) were anesthetized with isoflurane, brains were quickly removed and sagittal slices (350 μm thick), including the hippocampus, were prepared in ice-cold, carbogen-saturated (95% O²/5% CO²) artificial cerebrospinal fluid (aCSF; pH 7.4) using a vibratome (HM 650 V; Microm International, Walldorf, Germany). Slices were allowed to recover for at least 1 h at 34°C before being transferred to the recording chamber and continuously perfused with aCSF at a rate of 5 ml/min. The aCSF contained (in mM): NaCl 125, KCl 2.5, NaHCO³25, CaCl²2, MgCl²1, D-glucose 25, NaH²PO⁴1.25. All experiments were performed at room temperature (23–25°C).

Recordings of field excitatory postsynaptic potentials (fEPSPs) were performed in the CA1 stratum radiatum of the hippocampus using borosilicate glass micropipettes (Clark Electromedical Instruments, Pangbourne Reading, United Kingdom) filled with aCSF, resulting in an open tip resistance of 1–2 MΩ. Data were recorded with an Axopatch 200B patch-clamp amplifier, a Digidata 1200 interface (both from Axon Instruments, Foster City, CA), and the WinLTP Program version 1.11b (WinLTP Ltd., Centre for Synaptic Plasticity, School of Physiology & Pharmacology, University of Bristol, Bristol, United Kingdom).13fEPSPs were evoked by alternately delivering a 50-μs voltage pulse via  one of two bipolar tungsten electrodes (insulated to the tip; 50 μm diameter), placed at either side of the recording pipette, thus stimulating nonoverlapping populations of fibers of the Schaffer collateral-associational commissural pathway (fig. 1). Stimuli of each electrode were delivered alternatively at 15-s intervals, and two consecutive fEPSPs were averaged for noise minimization. The slope of the rising phase of the fEPSP (taken between 20 and 80% of the peak amplitude) was used as measure of the strength of synaptic transmission. For baseline recordings, stimulation intensity was adjusted to a value that evoked a response of approximately 25–30% of the maximal obtainable response. For the induction of LTP, a high-frequency stimulation (HFS) train (100 pulses delivered at 100 Hz) was applied to the Schaffer collateral-associational commissural pathway, and subsequent recordings were made for an additional 60 min without changing the stimulation rate and intensity used for baseline recordings. fEPSP slopes were normalized with respect to the responses recorded during the last 5 min before HFS.

Infrared videomicroscopy14(Zeiss, Oberkochen, Germany) was used to visualize CA1 pyramidal neurons, from which excitatory postsynaptic currents (EPSCs) were recorded using standard whole cell patch-clamp technique. Pipettes had an open tip resistance of 4–6 MΩ when filled with a solution containing (in mM): K-D-gluconate 130, KCl 5, EGTA 0.5, MgCl²2, HEPES 10, D-glucose 5, and Na²-phosphocreatine 20. Currents were recorded with a switched voltage-clamp amplifier (SEC 10 l; NPI Electronic, Tamm, Germany) with switching frequencies of 60–80 kHz (25% duty cycle). Series resistance was continuously monitored. EPSCs were elicited by square pulse stimuli (6–30 V, 50–500 μs; interstimulus interval 15 s) delivered via  a bipolar tungsten electrode placed in the Schaffer collateral-associational commissural pathway. Before each tissue stimulation pulse, neuronal input resistance was determined by a hyperpolarizing voltage step (−10 mV for 200 ms). NMDA receptor-mediated EPSCs (NMDA-EPSCs) were recorded in the presence of 1,2,3,4-tetrahydro-6-nitro-2,3-dioxo-benzo[f]quinoxaline-7-sulfonamide (NBQX; 5 μM), 3-amino-propyl(diethoxymethyl)phosphinic acid (CGP35348; 200 μM), and bicuculline methiodide (20 μM) at a holding potential of −40 mV. At the end of the experiments, we also applied d(-)-2-amino-5-phosphonopentanoic acid (AP5; 50 μM), resulting in a complete blockade of current responses (data not shown). For measurement of α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor-mediated currents, NBQX was omitted from the aCSF, and AP5 (50 μM) was added instead.

For measuring the effect of xenon on L-type calcium channels in CA1 pyramidal cells, we substituted calcium in the extracellular solution with barium and thus recorded barium currents through calcium channels. The extracellular solution contained (in mM): NaCl 125, KCl 2.5, NaHCO³25, CaCl²2, BaCl²1, D-glucose 25, NaH²PO⁴1.25. Pipettes were filled with an intracellular solution containing (in mM): K-D-gluconate 130, KCl 5, EGTA 0.5, MgCl²2, HEPES 10, D-glucose 5, and Na²-phosphocreatine 20. L-type calcium channel currents were recorded by adding specific antagonists to the aCSF: ω-Conotoxin GVIA (1 μM) for N-type channels, ω-Agatoxin IVA (100 nM) for P/Q-type channels and NiCl²(50 μM) for T-type channels. Barium currents were evoked by a 100-ms depolarizing step from −70 mV to 0 mV.

All current responses were amplified, low-pass filtered (3 kHz), digitized (ITC-16 Computer Interface; Instrutech Corp., Port Washington, NY) with a sampling frequency of 9 kHz, and stored on a hard drive (Power Macintosh G3 computer [Apple, Cupertino, CA], data acquisition software Pulse version 8.5 [HEKA Electronic GmbH, Lambrecht, Germany]).

In a subset of experiments, we measured the intracellular calcium influx during HFS by fluorometric calcium imaging. For this purpose, slices were loaded with the membrane-soluble acetoxymethyl ester of Fluo-4.15The dye was dissolved in dimethyl sulfoxide + 20% Pluronic F-127 at a concentration of 5 mM. For slice loading, the solution was diluted 1:10 in a pipette solution containing HEPES-aCSF (in mM: NaCl 125, KCl 3, HEPES 25, MgSO⁴2, CaCl²2, NaH²PO⁴1.25). This resulted in a final dye concentration of 500 μM Fluo-4 acetoxymethyl ester. The dye was applied to the dendrites of hippocampal CA1 neurons by ejecting it from a micropipette (open tip resistance 0.5 MΩ) with a pressure pulse. Cells were allowed to take up the dye for 1 h before the beginning of recordings. Imaging was performed with a ×60 water immersion objective (numerical aperture 0.9; Olympus, Hamburg, Germany) and a filter block containing a 470/40 filter (band pass) and 495 dichroic mirror. A monochromator (Polychrome V; T.I.L.L. Photonics, Martinsried, Germany) was used for excitation with a wavelength of 488 nm, and emission was collected at a maximum of 516 nm. For measurements of calcium-dependent changes in fluorescence, images were sampled during HFS to the Schaffer collateral-associational commissural pathway at 10 Hz for 5 s and recorded with a high-speed, charge-coupled device camera (Retiga-2000RV; QImaging, Surrey, Canada). Data acquisition and offline data analysis were performed with TillVision (T.I.L.L. Photonics). Baseline fluorescence before HFS (F0) and relative fluorescence changes (ΔF/F0) in defined regions of interest were measured throughout the image sequence. Changes in intracellular calcium are expressed as ΔF/F0 × 1,000 using the average of 10 images before HFS as a baseline. The time integral (area under the curve) was used to assess changes in total calcium.

Under control conditions, the aCSF was bubbled with carbogen, whereas for xenon application it was also bubbled with a mixture of 65% xenon, 30% O², and 5% CO². Concentration measurements of dissolved xenon were accomplished using headspace gas chromatography (Harlan Laboratories Ltd., Itingen, Switzerland). The final concentration of xenon in the aCSF was 1.9 ± 0.5 mM (n = 6). Concentrations of dissolved oxygen and carbon dioxide as well as pH were determined using a calibrated blood gas analyzer (Rapidlab 860 blood gas analyzer; Diamond Diagnostics, Holliston, MA) and did not change during application of the xenon gas mixture. All gas mixtures were purchased from Linde AG (Unterschleissheim, Germany) and applied at a flow rate of 0.3–0.5 l/min to the aCSF reservoir. Tubing was made of polytetrafluoroethylene (Teflon; VWR International, Darmstadt, Germany) to minimize loss of xenon.

All salts and chemicals were purchased from Sigma–Aldrich (Munich, Germany), except CGP35348 (Novartis Laboratories, Basel, Switzerland), and Fluo-4 acetoxymethyl ester (Invitrogen GmbH, Darmstadt, Germany).

For all statistical evaluations, SPSS Statistics version 16 (SPSS GmbH Software, Munich, Germany) was used. The recording times of fEPSPs, NMDA-EPSCs, and calcium channel currents were first portioned in equidistant subintervals of 5-min length. The averaged relative charge, slope, and amplitude were calculated for each of them. Two factorial multivariate analyses of variance with repeated measures design were applied on the averaged relative charge, slopes, and amplitudes with interval as a within-subjects factor and set of experiments as a between-subjects factor. Differences between the various intervals were studied by tests with contrasts, whereas differences between the various experiments by Bonferroni post hoc  tests. Multivariate analysis of variance was also performed for testing differences in the area under the curve of relative changes in fluorescence during HFS between the control and xenon recordings. In contrast to some other multivariate analyses, multivariate analysis of variance is robust against violations of the normality and homogeneity conditions. Multivariate analysis of variance is based on F-tests, which are, unlike some other parametric tests, very robust against the aforementioned conditions. We accepted α = 0.05 as the nominal level of significance. It was corrected (according to Bonferroni procedure) for all a posteriori  tests (tests with contrasts and post hoc  tests). Numerical data are presented as means ± SEM with the number of experiments indicated, if not stated otherwise. When error bars were smaller than the size of the symbols, they are not shown in the graphs.

In a first set of experiments, we determined the impact of xenon on basal synaptic transmission at CA3-CA1 synapses. After 20 min of stable baseline recordings, application of 1.9 mM xenon resulted in a reduction of fEPSPs, which reached a stable plateau after approximately 15 min. Averaged slopes of fEPSPs recorded during 16–20 min after start of xenon application were 70.4 ± 3.5% compared with the normalized fEPSP slopes recorded during the last 5 min before start of xenon application (n = 7; P = 0.001; fig. 2A). The xenon-induced decrease of fEPSP slopes was completely reversed after removal of xenon (fig. 2A).

In a next step, we determined the action of xenon on LTP. Under control conditions, HFS potentiated fEPSP slopes to 127.2 ± 5.8% (n = 25; P < 0.001; fig. 3). In the presence of xenon, HFS applied via  the second stimulation electrode produced only a short-term potentiation, and fEPSPs returned almost to baseline levels (105.9 ± 2.9%; n = 25; P = 0.090; fig. 3).

The effect of xenon on NMDA-EPSCs and α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid-EPSCs was evaluated using patch-clamp recordings from CA1 stratum radiatum neurons. Xenon (1.9 mM) reduced NMDA-EPSCs to 65.9 ± 9.4% (n = 7; P = 0.015; fig. 4) and α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid-EPSCs to 75.2 ± 7.6% (n = 6; P = 0.013; fig. 2B) of control responses. Upon termination of xenon application, the current responses recovered to the control level.

Because activation of extrasynaptic NMDA receptors is involved in the formation of LTP,16the influence of xenon on such receptors was investigated. The application of the glutamate reuptake inhibitor DL-threo-β-Benzyloxyaspartic acid (DL-TBOA; 30 μM) induces a pronounced spillover of glutamate, thereby also activating extrasynaptic NMDA receptors,17which preferentially comprise NR2B subunits.18To test whether the DL-TBOA–induced spillover of glutamate resulted in activation of extrasynaptic NMDA receptors, we investigated the influence of the NR2B-specific antagonist Ro-25-6981 (1 μM) on NMDA-EPSCs in the absence and presence of DL-TBOA. In the absence of DL-TBOA, Ro-25-6981 decreased the NMDA-EPSC amplitudes to 53.3 ± 4.9% of control (n = 5; P < 0.001; fig. 4A), whereas in the presence of DL-TBOA, Ro-25-6981 attenuated NMDA-EPSCs to a significantly higher degree (32.9 ± 5.7%; n = 5; P < 0.001; fig. 4A), pointing to an additional recruitment of extrasynaptic, NR2B-containing receptors. The xenon-induced decrease of NMDA-EPSCs was not significantly different in the presence (58.2 ± 5.8%; n = 5; P = 0.004; fig. 4B) or in the absence (65.9 ± 9.4%; n = 7; P = 0.022; fig. 4B) of DL-TBOA. These findings suggest that xenon antagonizes both synaptic and extrasynaptic NMDA receptors with the same potency.

Long-term potentiation requires an increase in intracellular calcium.19,–,21Using fluorometric calcium imaging, we tested the influence of xenon on calcium influx during HFS. Xenon reduced area-under-the-curve values to 30.2 ± 6.9% (n = 8; P < 0.001; fig. 5) of control. We also evaluated the effect of the specific NMDA receptor antagonist AP5 on calcium transients during HFS. AP5 nearly completely abolished fluorometric calcium signals (area under the curve 7.4 ± 2.6% of control; n = 8; P < 0.001; fig. 5), proving an essential role of NMDA receptors in the increase in intracellular calcium concentrations during HFS in our brain slice preparation.

In addition to NMDA receptors, an increase in intracellular calcium during HFS can be mediated via  VGCCs, especially L-type calcium channels. Thus, we investigated the effect of xenon on calcium channel currents evoked by a depolarizing step from −70 to 0 mV in CA1 pyramidal neurons. For these experiments, the fraction of L-type calcium channels mediating the barium currents was increased pharmacologically by specific antagonists against N-, P/Q-, and T-type channels. Xenon did not affect currents recorded under these conditions (95.4 ± 7.4% of control; n = 6; P = 0.650; fig. 6). Coapplication of the L-type calcium channel antagonist nifedipine (5 μM) reduced calcium channel currents to 47.0 ± 13.1% and revealed a fraction of approximately 50% L-type calcium channels contributing to the entire barium current. Application of the unspecific calcium channel antagonist CdCl (100 μM) completely abolished the currents (data not shown).

In the current study, we found that the inhalational anesthetic xenon blocks LTP in the murine hippocampus most likely by antagonizing calcium influx through NMDA receptors during HFS. Xenon antagonized synaptically and extrasynaptically located NMDA receptors with similar potency and had no effect on L-type calcium channels.

All these effects occurred under a clinically relevant concentration of xenon. Published human minimum alveolar concentrations of xenon for immobilization (MAC^immobility) are 63 vol%22or 71 vol%23and for hypnosis (MAC^awake) 33 vol%.24Using a solubility coefficient of 0.0887 at 37°C,25the calculated minimum alveolar concentration equivalents of dissolved xenon are 2.2 or 2.5 mM for MAC^immobility, and 1.2 mM for MAC^awake, respectively. The applied xenon concentration of 1.9 ± 0.5 mM dissolved in aCSF at room temperature most probably resembles a concentration at which amnestic properties are likely to arise in humans. MAC^immobilityfor rodents has been shown to be hyperbaric (1.61 atm26), so because of methodological limitations, experiments in the current study were performed at concentrations below MAC^immobility. Thus, the effects observed may even underestimate the effects of xenon in humans under anesthesia.

Long-term potentiation, a long-lasting change in synaptic efficacy, is considered to be the major in vitro  correlate for memory formation and learning.4LTP is defined as an increase in the strength of synaptic transmission measured by an increase of fEPSP slopes for at least 1 h. In accordance with this definition, xenon inhibited the induction of LTP in the CA1 region of the hippocampus.

An inhibition of hippocampal LTP by volatile anesthetics,6,27,28propofol,29and etomidate9has been reported. Dependent on the mechanisms of action, isoflurane,6sevoflurane,7and etomidate9reduce LTP, at least partly, because of their potentiating effect on the GABAergic system. In contrast, numerous reports from different laboratories suggest that xenon exerts little or no potentiating action on GABA receptor type A receptors.30,–,32Recently, it has been shown that xenon affects neurotransmission by antagonizing NMDA and α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptors.32,–,34In line with these results, we provide evidence that xenon impairs LTP because of its antagonistic properties against the NMDA receptor.

Induction of LTP is NMDA receptor dependent, and an impairment of LTP has been shown for a variety of NMDA receptor antagonists.35,36Antagonistic properties of xenon against the NMDA receptor have been shown in other rodent central nervous system areas and in studies using heterologously expressed human NMDA receptors.37,38NMDA receptors are highly permeable for calcium ions, and in addition to an activation of VGCCs, the recruitment of NMDA receptors during HFS leads to an increase in the intracellular calcium concentration. It has been shown that induction of LTP can be inhibited by intracellular injection of calcium chelators.39,–,41By fluorometric calcium imaging, we could demonstrate that xenon decreases intracellular calcium influx during HFS. Two important calcium sources are involved crucially in the induction of LTP: NMDA receptors and calcium channels of the L-type.4,5,42,–,47A previous study showed no influence of xenon on L-type VGCCs in human atrial cardiomyocytes.48In the current study, we show that xenon did not reduce Ba²⁺-carried currents through VGCCs when P/Q-, N-, and T-type calcium channels were antagonized pharmacologically. Nifedipine antagonized these currents to approximately 50%. (The nifedipine-resistant current probably was mediated by R-type VGCCs.44Unfortunately, no specific antagonists are available for this type of calcium channel current.) These results indicate that a significant fraction of this current is mediated by L-type VGCCs. Thus, an effect of xenon on L-type VGCCs in CA1 pyramidal neurons is highly unlikely.

These results suggest that the xenon-induced inhibition of LTP induction is mediated primarily by an antagonism against NMDA receptors and the resulting decrease in calcium influx during HFS. Confirmative results were obtained when NMDA receptors were blocked specifically with AP5.

Subunit composition and subunit localization of NMDA receptors changes during neuronal development. In early stages, NR2B subunits are predominantly expressed, and NR2A/NR2B ratio increases during development.45NR2A incorporation into the synapses replaces NR2B receptors so that in mature neurons,46NR2B-containing NMDA receptors are predominantly,47but not exclusively,48found in extrasynaptic receptors. At least in cortical synapses, the peak of developmental NMDA receptor remodeling can be found between postnatal days 4 and 9, and subunit expression reaches a stable state until P15.49Therefore, in hippocampal slices from mice used in our study (P20–30), developmental changes of subunit expression are most likely accomplished.

Synaptic NMDA receptors respond to synaptically released glutamate from the presynaptic membrane. In contrast, extrasynaptic NMDA receptors are by definition located outside synapses and are not activated by synaptic release of glutamate. They are expressed in a low surface density but represent approximately one half of the total NMDA receptor surface population.50Ambient glutamate has been shown to be present in a nanomolar range51and leads to a small tonic current mediated by extrasynaptic NMDA receptors. Extrasynaptic NMDA receptors play an important role in modulation of protein synthesis and slow network oscillations.18In addition, they are capable of initiating neuronal degeneration because strong activation of extrasynaptic NMDA receptors in vivo  can contribute to calcium loading and depolarization of mitochondria that is likely to precede necrotic cell death.52Synaptic release of glutamate does not increase ambient glutamate concentrations, but HFS has been shown to overwhelm glutamate reuptake transporters,53leading to a glutamate spillover54and thus increasing ambient glutamate concentrations. The maximal activation of such receptors is found at stimulation frequencies between 100 and 200 Hz55and resembles the HFS used in our experiments. In addition to their neurotoxic action, extrasynaptic NR2B-containing receptors are critical for the induction of synaptic plasticity because a low NR2A/NR2B ratio lowers the threshold for LTP induction, making it more likely that a modest response can increase calcium and induce LTP.45A reduction of extrasynaptic NMDA receptor activity has been shown to reduce neuronal degeneration56,57and synaptic plasticity.58,59It is still under debate whether synaptic NR2A-containing receptors or extrasynaptic NR2B-containing receptors are more important for the induction of LTP. A xenon-mediated reduction of extrasynaptic NMDA receptors might be an important mechanism of how xenon affects LTP in the hippocampus and how it mediates its neuroprotective properties.

Spillover of glutamate predominantly activates NR2B-containing NMDA receptors.16,60,61In fact, when glutamate transporters were blocked, the NR2B subunit-selective antagonist Ro-25-6981 reduced NMDA-EPSCs more extensively, demonstrating that extrasynaptic NR2B-containing NMDA receptors also had been activated. The allosteric NMDA receptor antagonist Ro-25-6981 has a 3,000-fold selectivity of NR1/NR2B NMDA receptors over NR1/NR2A receptors.62An activity against NR2A-containing receptors has been reported for high concentrations with an IC⁵⁰greater than 30 μM.62In contrast, IC⁵⁰for NR2B-containing NMDA receptors is reported to be 0.009 μM. The concentration used in our experiments (1 μM) is far below the IC⁵⁰for NR2A-containing NMDA receptors but in the range of selective concentrations (0.5–3 μM)63for NR2B-containing NMDA receptors. It has been shown recently that synaptically located NMDA receptors were antagonized by xenon nonselectively for receptors containing either the NR2A or NR2B subunit.12Although we were not able to record currents from isolated extrasynaptic NMDA receptors, we could demonstrate that DL-TBOA increased the fraction of NR2B-containing extrasynaptic NMDA receptors governing NMDA-EPSCs. Under these experimental conditions, the application of xenon antagonized NMDA-EPSCs with similar potency as in the absence of DL-TBOA. This shows that xenon exerts no specificity to synaptically or extrasynaptically located NMDA receptors. These results are in line with previous studies showing that xenon affects NMDA receptors regardless of subunit composition (NR2A or NR2B)12and inhibits NMDA receptors, presumably by binding to the glycine site at the NR1 subunit.64 

Extrasynaptic GABA receptor type A receptors are also expressed in the hippocampus and are critically involved in learning and memory processes.65However, there is increasing evidence that xenon does not affect GABA receptor type A receptors, so that possible amnestic actions of xenon are highly unlikely to be mediated by a modulation of extrasynaptic GABA receptor type A receptors.

It has been suggested that attenuation of synaptic plasticity in the hippocampus may contribute to the amnestic properties of anesthetics.6,7,9Although LTP in the hippocampus is accepted to underlie certain forms of memory (e.g. , spatial learning42), other mechanisms and brain regions also contribute to memory formation and recall.66,67However, recent work3could demonstrate a participation of the hippocampus in the amnestic action produced by propofol. Impairment of memory formation during surgery is one of the key properties of anesthesia. Meanwhile, xenon anesthesia in humans has been evaluated in a number of clinical trials10,11,68and studies using healthy volunteers.69,70In none of these studies has the occurrence of awareness during xenon anesthesia been systematically investigated, and to our knowledge no case of intraoperative awareness has been reported. Given the lack of a systemic investigation and the small numbers of patients included in these trials, it is too early to draw a definite conclusion as to whether xenon has less or more amnestic potency than do other inhalative anesthetics. However, apparently a prevention of intraoperative memory formation occurs during xenon anesthesia. The demonstrated xenon-induced attenuation of hippocampal LTP most probably is attributable to an antagonism against NMDA receptors and represents one mechanism of how the inhalative anesthetic xenon exerts its amnestic properties and prevents patients from undesirable intraoperative memory formation.

The authors thank Alexander Yassouridis, Ph.D. (Biostatistician, Head of the Research Group Biostatistics, Max Planck Institute of Psychiatry, Munich, Germany), for statistical advice.