| ||||||||||||||
|
|
|||||||||||||
-Aminobutyric Acid Type A Receptor Binding in Humans





From the *Turku PET Centre and the #Departments of Pharmacology and Clinical Pharmacology, University of Turku, Turku, Finland; the
Department of Psychology, Åbo Akademi University, Turku, Finland; the
Department of Anesthesiology and Intensive Care, Turku University Hospital, Turku, Finland, the
Department of Otorhinolaryngology – Head and Neck Surgery, Turku University Hospital, Turku, Finland; the ||Department of Child Neurology, Helsinki University Hospital, Helsinki, Finland and the ¶Institute of Biomedicine, Pharmacology, University of Helsinki, Helsinki, Finland.
Address correspondence to Elina Salmi, MD, Turku PET Centre, PO Box 52, FIN-20521 Turku, Finland. Address e-mail to anelsa{at}utu.fi.
| Abstract |
|---|
|
|
|---|
-aminobutyric acid type A (GABAA) receptors also have been reported. The mechanism of anesthetic action of xenon in the living human brain still remains to be determined. METHODS: We used the specific GABAA receptor benzodiazepine-site ligand 11C-flumazenil and positron emission tomography to study the GABAergic effects of xenon in eight healthy male volunteers. Each subject underwent two dynamic 60-min positron emission tomography studies awake and during approximately one minimum alveolar concentration of xenon (65%). Bispectral index was recorded. Cortical and subcortical gray matter regions were analyzed using both automated regions-of-interest analysis and voxel-based analysis.
RESULTS: During anesthesia, the mean ± sd bispectral index was 23 ± 7, and there were no significant changes in heart rate or mean arterial blood pressure. Xenon did not significantly affect 11C-flumazenil binding in any brain region.
CONCLUSIONS: Xenon did not affect 11C-flumazenil binding in the living human brain, indicating that the anesthetic effect of xenon is not mediated via the GABAA receptor system.
| Introduction |
|---|
|
|
|---|
-aminobutyric acid type A (GABAA) receptors.7 Nevertheless, the GABAergic effects of xenon have also been reported.8,9 The effects of anesthetics on the GABAergic system in the living human brain can be studied using positron emission tomography (PET) and carbon-11 (11C)-labeled flumazenil. The volatile anesthetics isoflurane and sevoflurane as well as the IV anesthetic propofol enhance 11C-flumazenil binding in both cortical and subcortical gray matter regions implying the involvement of GABAA receptors in the mechanism of anesthetic action of these agents.10,11 Ketamine, which is also believed to primarily act via NMDA receptors, has been shown not to affect 11C-flumazenil binding in subanesthetic doses,12 despite causing significant increases in cerebral blood flow and glucose metabolism.13,14 Recently, the xenon-induced decrease in cerebral glucose metabolism was interpreted to reflect other mechanisms of anesthetic action than NMDA receptor antagonism.15 Furthermore, in rodent experiments, xenon has counteracted the paradoxical neuronal excitation induced by ketamine16 and the dopamine-releasing effect of ketamine,17 which suggests that xenon affects brain mechanisms differently from ketamine.
The aim of the present study was to evaluate the effects of xenon on the GABAergic system in humans in vivo using 11C-flumazenil binding and PET.
| METHODS |
|---|
|
|
|---|
Each subject underwent two dynamic 60-min 11C-flumazenil PET scans on the same day at least 2 h apart. The first scan was done before and the second scan during 1 minimum alveolar concentration (MAC) xenon anesthesia. The subjects were breathing room air during the baseline scan. In the same study protocol, we also assessed the effects of xenon on regional cerebral blood flow using H215O and PET.18
Anesthesia and Monitoring
A peripheral vein was cannulated for 11C-flumazenil administration and fluid infusion (0.9% saline, 100 mL/h). A radial artery cannula was placed under local anesthesia for blood sampling and invasive arterial blood pressure monitoring. Vital signs and bispectral index were monitored with Datex-Ohmeda S/5 monitor (GE Healthcare, Helsinki, Finland) and recorded into a portable computer. The partial pressure for arterial carbon dioxide was maintained at baseline (awake) level by adjusting ventilation.
No premedication was given. To enable 63% xenon (the estimated MAC for xenon)19 in the closed-system ventilation, a 1-h denitrogenation with subjects breathing spontaneously 100% oxygen through 5 cm H2O continuous positive airway pressure mask was needed. After the denitrogenation, the subjects were anesthetized with xenon (Xenon Pro Anesthesia, Air Liquide Deutschland GmbH, Krefeld, Germany) delivered with a PhysioFlexTM closed-system ventilator (Dräger, Lübeck, Germany). The induction, maintenance and monitoring of xenon anesthesia have been reported in more detail by Laitio et al.18
After the induction, a 40-min minimum stabilization period was allowed before the second 11C-flumazenil PET scan. After the scan, xenon was discontinued and muscle relaxation was reversed with a neostigmine-glycopyrrolate combination (Robinul Neostigmin; Wyeth Lederle, Vantaa, Finland). Subjects were tracheally extubated as they recovered spontaneous breathing and regained consciousness. They were monitored for stable vital signs for a minimum of 1 h. The subjects were discharged following hospital criteria for ambulatory surgery patients.
Data Acquisition
11C-Flumazenil binding was assessed at baseline (awake) and during 1 MAC xenon anesthesia with a GE Advance PET Scanner (GE Medical System, Milwaukee, WI, USA). Positioning of the head was done using anatomical landmarks and laser alignment lights. The preparation of 11C-flumazenil was done according to a previously published procedure,20 and an IV dose of 11C-flumazenil was administered manually in 60 s. The injected mass was 1.738 ± 0.295 µg and the dose 256.3 ± 13.6 MBq. The image acquisition was started at the same time with dosing and was performed in 3D-mode (septa retracted) for 60 min. During the study, 21 arterial blood samples were taken to determine arterial radioactivity and seven arterial plasma samples were taken for 11C-flumazenil metabolite analysis. The image reconstruction and the timing and processing of arterial plasma samples were performed in a similar way as in our previous sevoflurane and propofol study.11 A Hill type function was fitted to the fraction curve of unchanged 11C-flumazenil to improve the quality of arterial data.
|
|
Image Analysis Methodology
A ligand-specific 11C-flumazenil template was created for spatial normalization of images according to a previously presented procedure.21 Parametric images were calculated representing the three-compartmental model of total tissue distribution volume (VT) of 11C-flumazenil at voxel-level. Parametric images were spatially normalized to Montreal Neurological Institute space, using a standard estimation procedure implemented in SPM9922,23 and the 11C-flumazenil template described above. Normalization parameters for parametric images were estimated from the mean summated images created with within-subject realigning procedure, and both realigned images (baseline and intervention scan) of each individual were written using the same parameters. The normalized images were written using bilinear interpolation. In addition, for a voxel-based image analysis, parametric images were smoothed using a 12-mm Gaussian kernel.
To test the effects of xenon on VT of 11C-flumazenil, a voxel-based statistical analysis was made using SPM99 and MatLab 6.1 for Windows (Math Works, Natick, MA) as described earlier.12 The analysis was performed as an exploratory analysis covering the whole brain, i.e., without any a priori hypothesis or spatial constrictions concerning the location of possible differences. A P value below 0.05 (corrected for multiple comparisons) was considered significant.
To achieve quantitative regional estimates of VT values an automated region-of-interest (ROI) analysis24,25 was performed. ROIs for automated ROI analysis were defined using a magnetic resonance image fitted onto common stereotactic space in accordance with coordinates of the Montreal Neurological Institute database. The ROIs were drawn using Imadeus (version 1.50, Forima Inc., Turku, Finland) on the dorsolateral prefrontal and medial part of the superior frontal gyruses; the medial and lateral part of the temporal cortex; and the inferior parietal and occipital cortices; the anterior and posterior cingulate cortices, the thalamus, caudate, putamen, cerebellum, white matter, and the pons. The average VT values were calculated from spatially normalized parametric 11C-flumazenil images. The average cortical gray matter value was calculated using individual cortical gray matter ROIs weighted with the size of individual ROIs. As this method is based on common stereotactic space, i.e., spatially normalized parametric images, it is not hampered by operator-induced error in drawing ROIs individually for each subject in a sample having differences in brain anatomy. Using regional values provided by automated ROI analysis, a further confirmation and estimation of the results of voxel-based statistical image analysis can be achieved.24,25
Statistical Analysis
All enrolled subjects were included in the statistical analyses. Quantitative VT values were analyzed using repeated-measures analysis of variance having the condition (i.e., awake and anesthesia) as a within factor. The pairwise comparisons were performed using the linear contrasts of the same model. The mean of the left and right VT values were used in the analyses because there were no statistically significant side-by-condition interactions in any of the ROIs studied. A two-sided P value of <0.05 was considered statistically significant. Ninety-five percent confidence intervals were calculated for percentual VT changes. Statistical analysis was conducted with SAS (version 8.2; SAS Institute Inc., Cary, NC) by 4Pharma Ltd., Turku, Finland.
| RESULTS |
|---|
|
|
|---|
|
The VT values are shown in Table 2. There were no statistically significant changes in any of the brain areas studied. Furthermore, the voxel-based analysis did not reveal any significant changes in 11C-flumazenil binding.
|
| DISCUSSION |
|---|
|
|
|---|
Based on in vitro studies on various recombinant GABAA receptor subtypes expressed in human embryonic kidney cells and Xenopus oocytes, xenon has been shown to enhance GABAergic transmission.8,9 In the study of Hapfelmeier et al.,8 the concentration of xenon (3.9 mM) was slightly over 1 MAC (2.9 mM) and it produced about 40% increase in peak amplitudes of 10-µM GABA-evoked responses. Yamakura and Harris9 also detected a minor 15% enhancement of GABAA receptors at 46% of xenon (2 mM), i.e., at lower concentrations than in our study. In their study, GABAA receptors were much less enhanced by xenon than by isoflurane and the effects of xenon on, e.g., NMDA receptors were more prominent. In cultured hippocampal neurons, Franks et al.6 found no effects of xenon on GABAA receptor currents, but about 60% inhibition of NMDA receptor currents. Furthermore, de Sousa et al.26 studied autaptic synapses in individually cultured hippocampal neurons and found that 3.4 mM xenon strongly inhibited excitatory, mainly NMDA receptor-mediated currents, but hardly affected the GABAA receptor-mediated inhibitory currents. Thus, the GABAA receptor augmentation of xenon appears only in recombinant receptors, while hardly any enhancement has been observed in native receptor populations, which might be due, for example, to different posttranslational modifications and/or receptor subunit combination/stoichiometry of artificial versus native receptors.
Recently the effects of 1 MAC xenon on cerebral glucose metabolism in humans were investigated.15 Xenon reduced the whole-brain metabolic rate of glucose resembling the effects of volatile anesthetics. As other drugs with NMDA receptor antagonistic properties, such as ketamine, increase cerebral glucose metabolism, it was concluded that xenon more likely has other mechanisms of anesthetic action than NMDA receptor antagonism. This conclusion, however, rests on the assumption that metabolic changes during anesthesia are mediated via specific receptor systems that are also engaged with the mechanism of anesthetic action itself.27 Even if this is the case, the mechanism of anesthesia in the living human brain has not been verified for any general anesthetics, and it is plausible that most general anesthetics affect numerous different neurotransmission systems of the brain.7 For instance, in addition to NMDA receptor antagonism, ketamine interacts with, e.g., the cholinergic,28 the opioid,29 and the serotonergic system,30 and therefore it is not certain that NMDA receptor antagonism induces the increase in cerebral glucose metabolism. Thus, there are several caveats in extrapolating the mechanism of action of xenon from its effects on cerebral glucose metabolism.
11C-Flumazenil has become the PET tracer of choice in studying the GABAergic system of the living human brain.10,11 In PET tracer doses, it has no pharmacological effects of its own and it binds highly selectively to GABAA receptors.31,32 In addition, the binding is independent from alterations in cerebral blood flow.33 On the basis of the present results, we could exclude more than approximately 7% average increase of binding of 11C-flumazenil in the cortical gray matter.
The interaction of different drugs with the GABAA receptor depends on the subunit composition of receptor complex and certain receptor subtype selectivity has been shown with various drugs.34 It is possible that xenon affects only the GABAA receptor subtypes lacking the benzodiazepine binding site, such as the flumazenil-insensitive
subunit-containing receptors.34 Xenon-induced enhancement of recombinant GABAA receptor function is not dependent on the
2 subunit8 that is obligatory for the binding of flumazenil.
Several studies suggest that xenon inhibits excitatory glutamatergic signaling, and although it is not absolutely definitive which subtype of glutamatergic receptors is responsible for the effects of xenon, NMDA receptor antagonism is currently regarded as the prime mechanism.6,35 There is also evidence that xenon affects the inhibitory neurotransmission system via glycine receptors36 and activates two-pore potassium leak background channels.37 Other possible targets include nicotinic-acetylcholine receptors, second-messenger signaling systems, and neurotransmitter release of the brain.38 It would be highly intriguing to investigate the effects of xenon on the glutamatergic system and the NMDA receptor, but unfortunately there is no PET ligand available for this purpose.
We have previously shown that both sevoflurane and propofol increase 11C-flumazenil binding during anesthesia,11 but based on these studies, it could not be confirmed if the change was a specific effect of these drugs or a nonspecific effect of anesthetic state itself. Current results with 1 MAC xenon suggest that enhanced 11C-flumazenil binding during sevoflurane and propofol anesthesia is a drug-specific phenomenon, thus conforming to the current understanding of the mechanism of action of these drugs.
In conclusion, xenon did not affect 11C-flumazenil binding suggesting that other mechanisms than the enhancement of the GABAergic system exert the anesthetic effects of xenon in humans.
| ACKNOWLEDGMENTS |
|---|
| Footnotes |
|---|
Supported by Turku University Hospital EVO-grant No. 13323, Turku, Finland; Alfred Kordelin Foundation, Helsinki, Finland; Finnish Norwegian Medical Foundation, Helsinki, Finland; and the National Graduate School of Clinical Investigations, Helsinki, Finland.
Reprints will not be available from the author.
| REFERENCES |
|---|
|
|
|---|
, and
opioid receptors expressed in Chinese hamster ovary cells. Anesthesiology 1999;90:174–82[Web of Science][Medline]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|