| JOURNAL HOME | CME HOME | THIS MONTH | PAST ISSUES | ETOC | COLLECTIONS |
| AUTHORS | REVIEWERS | EDITORIAL BOARD | FEEDBACK | RSS | HELP |
| ||||||||||||||
| ||||||||||||||
|
|
|
|
||||||||||||
|
||||||||||||||
Turku PET Centre, University of Turku and the Department of Anesthesiology and Intensive Care, Turku University Hospital, Turku, Finland.
Address correspondence to Elina Salmi, MD, Turku PET Centre, PO Box 52, FIN-20521 Turku, Finland. Address email to elina.salmi{at}utu.fi
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
-aminobutyric acid type A (GABAA) receptors. However, this fundamental characteristic has not been extensively investigated in humans. We studied 11C-flumazenil binding to GABAA receptors during sevoflurane and propofol anesthesia in the living human brain using positron emission tomography (PET). Fourteen healthy male subjects underwent 2 60-min dynamic PET studies with 11C-labeled flumazenil, awake and during anesthesia. Anesthesia was maintained with 2% end-tidal sevoflurane (n = 7) or propofol at a target plasma concentration of 9.0 ± 3.0 (mean ± SD) µg/mL (n = 7). The depth of anesthesia was measured with bispectral index (BIS). Values of regional distribution volumes (DV) of 11C-flumazenil were calculated in several brain areas using metabolite-corrected arterial plasma curves and a two-compartment model. Separate voxel-based statistical analysis using parametric DV images was performed for detailed visualization. The average BIS index was 35 ± 6 in the sevoflurane group and 28 ± 8 in the propofol group (P = 0.02). Sevoflurane increased the DV of 11C-flumazenil significantly (P < 0.05) in all brain areas studied except the pons and the white matter. In the propofol group the increases were significant (P < 0.05) in the caudatus, putamen, cerebellum, thalamus and the frontal, temporal, and parietal cortices. Furthermore, the DV increases in the frontal, occipital, parietal, and temporal cortical areas and in the putamen were statistically significantly larger in the sevoflurane than in the propofol group. Our findings support the involvement of GABAA receptors in the mechanism of action of both anesthetics in humans. IMPLICATIONS: Both sevoflurane and propofol enhanced gamma-aminobutyric acid (GABA)A receptor binding in the living human brain as assessed with 11C-labeled flumazenil and positron emission tomography, thus supporting the involvement of GABAA receptors in the mechanism of action of both volatile anesthetics and propofol.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
-aminobutyric acid type A (GABAA) receptors (1,2). At clinically relevant concentrations they enhance the affinity of GABA to GABAA receptors (3), increasing GABA-induced chloride currents (4,5). The anesthetic potency of isoflurane and enflurane correlates well with their abilities to strengthen GABA-mediated chloride currents (5,6). Also, the IV anesthetic propofol enhances GABA-induced chloride currents and increases both the duration and intensity of GABA-mediated inhibition (5,7). Positron emission tomography (PET) allows the effects of anesthetics on neurotransmission to be studied in the living human brain. By using PET with 11C-labeled flumazenil and an appropriate kinetic model, the amount of available GABAA receptors can be quantified. The effects of isoflurane on GABAA receptor binding were investigated in healthy volunteers with PET and 11C-flumazenil (8). In that study, the distribution volume ratio (DVRATIO) increased concentration-dependently during isoflurane anesthesia in comparison to awake state in all brain areas, indicating that the 11C-flumazenil PET method is sensitive to conformational changes in GABAA receptor in humans.
We have now extended this finding by studying the effects of two different types of anesthetics, a volatile anesthetic (sevoflurane) and an IV anesthetic (propofol), on regional cerebral GABAA receptor binding in humans with PET. Based on the study of Gyulai et al. (8), our hypothesis was that both drugs would increase 11C-flumazenil binding in humans, as the anesthetics are supposed to enhance the affinity of GABA to GABAA receptors.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Eight of the 16 subjects were anesthetized with sevoflurane and 8 with propofol. Each subject underwent 2 11C-flumazenil PET scans on the same day. The first scan was done awake (no drug) and the second during anesthesia either at 1 minimum alveolar concentration (MAC) of sevoflurane or 50% effective plasma concentration (EC50) of propofol. In the same study protocol, we also assessed the effects on sevoflurane and propofol on regional cerebral blood flow using H215O and PET at different depths of anesthesia. These results and details of the experimental setting have been published earlier (9).
The left radial artery was cannulated for arterial blood sampling and two large veins in the right forearm were cannulated for administration of propofol, 11C-flumazenil boluses, and 2.5% dextrose + 0.45% NaCl infusion (100 mL/h). The subjects received no premedication. In the sevoflurane group, anesthesia was induced via face mask with 8% sevoflurane (Sevorane, Abbot Oy, Espoo, Finland) in 100% oxygen. The inhaled anesthetic concentration was thereafter immediately reduced to reach and maintain 2% end-tidal level (1 MAC). In the propofol group, the induction was accomplished with IV target controlled infusion of propofol (Diprivan 20 mg/mL, AstraZeneca Oy, Masala, Finland). The target venous plasma concentration of propofol was set at 6 µg/mL (10). The infusion was delivered with a Harvard 22 syringe pump (Harvard Apparatus, South Natick, MA) connected to a portable computer running STANPUMP software with pharmacokinetic variables by Marsh et al. (11).
After the loss of eyelid reflex, subjects were administered 0.6 mg/kg rocuronium IV and a laryngeal mask was inserted. A semiclosed ventilator system (Ventilator 710; Siemens-Elema Ab, Solna, Sweden) was used with 2.5 L/min of fresh gas flow of 30% oxygen-air mixture. Breathing frequency was set to 15 breaths/min. The ETCO2 was kept strictly at 4.5% (33.75 mm Hg) throughout the anesthesia by adjusting the tidal volume. Muscle relaxation was maintained with bolus doses of 10 mg rocuronium.
The depth of anesthesia was measured with the bispectral index (BIS®) monitor (Aspect Medical Systems, Natick, WA). BIS values were manually recorded at 1-min intervals throughout the study (12). ECG, peripheral oxygen saturation, breathing gases, noninvasive arterial blood pressure, and heart rate were monitored with Datex AS3 equipment (Datex-Ohmeda Division, Instrumentarium Corp, Helsinki, Finland). After completing the PET assessments at 1 MAC/EC50, residual neuromuscular block was reversed with neostigmine/glycopyrrolate combination and the subjects were tracheally extubated.
11C-flumazenil was used as the PET tracer. As a benzodiazepine antagonist flumazenil binds selectively to GABAA receptors (13,14). 11C-flumazenil was prepared by the reaction of 11C-methyl triflate and Ro 15–5528, followed by high-performance liquid chromatography (HPLC) purification and formulation of the product according to a published procedure (15). The radiochemical purity (>99.5%) and the specific radioactivity of the product were determined using HPLC and UV-detection at 246 nm.
All PET studies were performed in a dimly lit room with no sudden loud noises. The subjects wore ear plugs and kept their eyes shut during the scanning. A plastic head holder was used to minimize head movement. Descriptions of the PET scanner and image reconstruction are given in our previous article (9). For anatomical reference, individual magnetic resonance image (MRI) scans were acquired with a 1.5 Tesla scanner (Siemens Magnetom SP63, Erlagen, Germany) on a separate occasion.
An IV dose of 370 MBq of 11C-flumazenil was administered manually in 60 s. The mean ± SD injected amount of 11C-flumazenil was 2.40 ± 0.85 µg. The image acquisition was started at the same time with dosing and was performed in two-dimensional mode (septa in place) for 60 min. Twenty-one arterial blood samples were taken during the study (40 s, 50 s, and 1, 1.5, 2, 2.5, 3, 4, 5, 6, 7, 8, 9, 10, 15, 20, 25, 30, 40, 50, and 60 min after the dosing of 11C- flumazenil) to determine arterial radioactivity with an automatic counter (Wizard 1480; Wallac, Turku, Finland). Furthermore, 7 arterial blood samples (at 4, 10, 20, 30, 40, 50, and 60 min) were taken for flumazenil metabolite analysis. Plasma proteins were precipitated with acetonitrile containing flumazenil. The solution obtained after centrifugation was separated with reversed-phase HPLC and a gradient of acetonitrile in phosphoric acid (16). The results were calculated as the percentage of decay corrected radioactivity in all peaks.
During the second PET scan in the propofol group, a 5 mL arterial blood sample was collected for determination of plasma propofol concentration. Plasma was immediately separated and kept frozen at –70°C until analyzed with HPLC (17). The intra-assay coefficient of variation of the assay was 7%.
The dynamic 11C-flumazenil images were realigned (within-subject) by using Statistical Parametric Mapping (SPM) version 99 (Wellcome Department of Cognitive Neurology, University College London, England) (18) to estimate the realign parameters for the 11C-flumazenil summation images. The MRIs were coregistered using mutual information procedure to the mean 11C-flumazenil images. Regions of interest (ROIs) were drawn using the Imadeus software (version 1.10; Forima Inc, Turku Finland) on the coregistered individual MRIs in the caudate, putamen, cerebellum, thalamus, pons, the frontal, temporal, parietal, and occipital cortices, the posterior and anterior cingulate cortices, and in the white matter (frontal periventricular area). Mean tissue time activity data of each ROI were calculated from the realigned PET images.
A two-compartment two-parameter model was used to describe 11C-flumazenil kinetics of regional and pixel tissue time activity curves. Distribution volumes (DV) were estimated directly without division (19). This method is optimal for studying the effects of anesthetics because the probable changes in perfusion and peripheral clearance have no effect on DV. However, the effects of plasma protein binding and nonspecific binding in tissue are inherent in the DV, in addition to the receptor density and the affinity. Either of these binding effects can be corrected by dividing or subtracting the DV values of ROIs by DV from a receptor-free reference region (20) if data from such a brain region are available. Pons is often used as a reference region in 11C-flumazenil studies, even though it contains some GABAA receptors (21). The present data show that the anesthetics tend to increase the DV in the pons, although not statistically significantly, indicating that it cannot be used as a reference tissue in anesthesia studies. Pontial change could not be the result of changes in plasma protein or nonspecific binding because a similar change was not seen in the white matter (see Table 2). White matter is unfortunately equally unsuitable as a reference region because the nonspecific binding in white and gray matter cannot be assumed to be similar. In addition, in the present study the variance of both pontial and white matter DV values was large (see Table 2).
|
Quantitative DV values and hemodynamic variables were analyzed as changes from baseline (awake) to anesthesia with analysis of variance with the drug (sevoflurane versus propofol) as a between factor. The mean of the left and right DV values were used in the analyses because there were no statistically significant side-by-treatment interactions in any of the ROIs studied. BIS and ETCO2 were analyzed with the two-sample Student’s t-test with equal variances. A two-sided value of P < 0.05 was considered statistically significant. Data are presented as mean ± SD unless otherwise stated.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Both drugs induced a significant reduction in mean arterial blood pressure, the decrease being larger in the sevoflurane (24 ± 6 mm Hg) than the propofol group (8.2 ± 7.9 mm Hg, P = 0.002) (Table 1). In the propofol group, heart rate increased significantly during anesthesia. There were no significant differences between groups in ETCO2 or oxygen saturation (Table 1). In both groups, BIS and all physiologic variables remained stable during the second PET scan. The average BIS was 35 ± 6 in the sevoflurane group and 28 ± 8 in the propofol group; the difference between the groups was statistically significant (P = 0.02).
|
In the SPM-analysis, a significant increase in all cortical and subcortical gray matter structures was shown during sevoflurane anesthesia. There also occurred significant but more restricted cortical changes during propofol anesthesia. Instead, in the cerebellum and striatum the increase of DV during propofol anesthesia did not reach statistical significance at cluster- level. Figure 1 visualizes the results of the voxel-based analysis showing the regions where propofol induced significant cluster-level increases in 11C-flumazenil binding, and Figure 2 visualizes the regions where sevoflurane induced significantly larger increases than propofol. In the sevoflurane group, the increase in the SPM analysis was so global that visualization was not considered informative.
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The effect of isoflurane, another volatile anesthetic, on GABAA receptor ligand binding in healthy human subjects has been studied by Gyulai et al. (8). They demonstrated, using PET and 11C-flumazenil, that isoflurane enhances GABAA receptor binding in brain areas of high, medium, and low GABAA receptor densities. They also demonstrated a several-fold increase in receptor-specific 11C-flumazenil binding when anesthetic exposure was increased from 1 MAC to 1.5 MAC. Our results with sevoflurane are in keeping with their findings with 1 MAC isoflurane.
We now demonstrate that propofol also significantly enhances GABAA receptor binding at a moderate depth of anesthesia. The targeted propofol concentration (EC50) was clearly exceeded, and according to BIS data, the depth of anesthesia was also greater in the propofol group. Therefore, our finding that propofol clearly affected GABAA binding less than sevoflurane is of special interest. This could be a result of differences in their interactions with the GABAA receptor complex, albeit both volatile anesthetics and propofol seem to enhance GABA binding to the GABAA receptor (3,22). The GABAA receptor is a pentameric transmembrane protein with numerous allosteric effector sites and integral anion channel. GABAA receptors can be assembled from combinations of almost 20 subunits at poorly known stoichiometry, which creates considerable heterogeneity of the receptors in various brain cells and brain regions in terms of pharmacological actions (23). Anesthetics may affect the GABAA receptor function by potentiating the action of the transmitter and/or by directly causing the channel activation, which phenomena seem to depend on the anesthetic and on the receptor subunit combination (23). Structural requirements of the receptor for the actions of propofol and volatile anesthetics are most likely different. Propofol has shown little receptor subunit dependence in its actions (24), but certain receptor subunit residues in the transmembrane domains have been found to selectively affect propofol’s actions (25,26). Critical receptor domains for the action of volatile anesthetics also reside in the transmembrane regions, although they do not include the residues needed for the action of propofol (27,28). Therefore, the different effects of sevoflurane and propofol on the GABAA receptor binding in vivo could be explained by their different allosteric effects on the multisubunit GABAA receptor protein affecting the extracellular flumazenil-binding site.
In the present study, changes in regional DV values of 11C-flumazenil indicated an effect of the anesthetics on GABAA receptors. Unfortunately, we cannot exclude a nonspecific effect of the anesthetic state on GABAA receptors and/or 11C-flumazenil binding because we did not have a proper (negative) control group. It would be necessary to compare the present anesthetics with compounds such as ketamine and xenon that are supposed to act via other mechanisms (N-methyl-D-aspartic acid antagonism).
Furthermore, some methodological issues should be noted. Flumazenil, the prototype benzodiazepine antagonist, binds highly selectively to GABAA receptors (13,14), and it has become the PET ligand of choice for in vivo quantification of GABAA receptors in the human brain. Its advantages and basic characteristics have been thoroughly discussed (8). As a receptor antagonist, the relevance of altered flumazenil binding as a marker of activity of the GABAergic transmission is, however, somewhat open to interpretation. Results from in vitro studies have suggested that only the binding of benzodiazepine agonists is affected by GABA (13). Ex vivo experiments with mice have demonstrated, however, that compounds that increase the amount of GABA in the brain also enhance the binding of flumazenil to the GABAA receptors, thus suggesting that in vivo flumazenil has a similar (positive) "GABA shift" as the benzodiazepine site full agonists (29). As discussed in more detail (8), the lack of agreement between in vitro and in vivo experiments could be the result of masking of GABA shift in vitro by residual GABA in the membrane preparations.
Hansen et al. (30) studied the effects of halothane and isoflurane on GABAA receptors in vivo and in vitro in rats using [3H]-flumazenil and autoradiography. They found that cortical binding of flumazenil was enhanced in vivo by both halothane and isoflurane, but interpreted it to be at least partly attributable to diminished clearance of flumazenil during anesthesia. They calculated the clearance as a change in radioactivity, which is not as reliable as determining clearance with chemical measurements. Furthermore, the change in flumazenil clearance was much smaller (20%) than the change in [3H]-flumazenil binding (40%). Flumazenil clearance was also reduced by the anesthetics (by 30%–44%) in our human subjects. It should be noted that arterial blood sampling and metabolite analysis applied in the present study effectively corrected the tissue data for errors resulting from changes in tracer clearance.
The 11C-flumazenil DV changes are attributable to changes in specific binding to GABAA receptors because the unchanged DV in the white matter eliminates any changes in plasma protein binding and nonspecific binding. In earlier studies neither halothane nor isoflurane induced any effects on nonspecific binding (8,30). In addition, nonspecific binding of 11C-flumazenil (as compared with specific binding) can be assumed negligible, unless specific binding is decreased in the study conditions (31). In the present study, the binding was increased, further decreasing the effect of nonspecific binding in DV.
Our study only revealed widespread overall effects of anesthetics on the GABAergic system, and it is not possible to make conclusions concerning the importance of the different brain regions in the anesthetic action. The regions of the most intensive receptor density changes reflect the effects of the anesthetic, not the mechanism of anesthesia per se. One of the interesting brain regions is the thalamus and especially its reticular nucleus, in which the GABA-mediated inhibitory circuit plays an important role (32). It has been suggested that a hyperpolarization block of the thalamocortical neurons could be the uniform neurophysiologic mechanism for unconsciousness induced by the anesthetics (33). In our study, the changes in the thalamus were 16.4% ± 6.5% and 12.1% ± 7.4% in the sevoflurane and propofol groups, respectively. Both of these increases were highly significant (P < 0.001).
Our findings indicate that both sevoflurane and propofol affect the GABAergic system. However, the effect might not be the only, and not even the most important, mechanism of action of these anesthetics (34). For example, both volatile anesthetics and propofol also act on other neurotransmitter systems, such as neuronal nicotinic acetylcholine and inhibitory glycine receptors (35). It is thus possible that anesthesia results from molecular effects on multiple neurotransmitter systems.
In conclusion, both sevoflurane and propofol significantly affected GABAA receptor ligand binding in humans, thus supporting at least a partial involvement of GABAA receptors in the mechanism of action of volatile anesthetics and propofol.
|
Acknowledgments |
|---|
We thank Berner Oy (Helsinki, Finland) for providing the BIS® monitor for this study. We would also like to thank Steven L. Shafer, MD (Stanford University, Department of Anesthesia) for the free use of his STANPUMP computer program and Professor W. Hunkeler (Roche, Basel, Switzerland) for kindly supplying the precursor des-methyl flumazenil, Ro 15–5528 and flumazenil, Ro 15–1788.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
-aminobutyric acidA receptors expressed by Xenopus oocytes: lack of involvement of intracellular calcium. J Pharmacol Exp Ther 1992; 263: 569–78.
2ß32 -aminobutyric acidA receptors. J Neurochem 2002; 80: 140–8.[ISI][Medline]
-aminobutyric acid and glycine receptors change alcohol cutoff: evidence for an alcohol receptor? Proc Natl Acad Sci U S A 1998; 95: 6504–9.This article has been cited by other articles:
|
|
 |
|
  E. Salmi, R. M. Laitio, S. Aalto, A. T. Maksimow, J. W. Langsjo, K. K. Kaisti, R. Aantaa, V. Oikonen, L. Metsahonkala, K. Nagren, et al. Xenon Does Not Affect {gamma}-Aminobutyric Acid Type A Receptor Binding in Humans Anesth. Analg., January 1, 2008; 106(1): 129 - 134. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. R. Atack, P. Scott-Stevens, J. S. Beech, T. D. Fryer, J. L. Hughes, M. C. Cleij, J.-C. Baron, J. C. Clark, R. J. Hargreaves, and F. I. Aigbirhio Comparison of Lorazepam [7-Chloro-5-(2-chlorophenyl)-1,3-dihydro-3-hydroxy-2H-1,4-benzodiazepin-2-one] Occupancy of Rat Brain {gamma}-Aminobutyric AcidA Receptors Measured Using in Vivo [3H]Flumazenil (8-Fluoro 5,6-dihydro-5-methyl-6-oxo-4H-imidazo[1,5-a][1,4]benzodiazepine-3-carboxylic Acid Ethyl Ester) Binding and [11C]Flumazenil Micro-Positron Emission Tomography J. Pharmacol. Exp. Ther., March 1, 2007; 320(3): 1030 - 1037. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Salmi, J. W. Langsjo, S. Aalto, K. Nagren, L. Metsahonkala, K. K. Kaisti, E. R. Korpi, J. Hietala, and H. Scheinin Subanesthetic Ketamine Does Not Affect 11C-Flumazenil Binding in Humans Anesth. Analg., September 1, 2005; 101(3): 722 - 725. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
