The present leading theory of the mechanism of action of inhaled anesthetics proposes that anesthetics act by enhancing inhibitory synaptic neurotransmission, decreasing excitatory neurotransmission, or both. Consistent with this theory, inhaled anesthetics modulate the function of many excitatory and inhibitory neurotransmitter receptors. Glutamate, the major excitatory neurotransmitter in the mammalian central nervous system (CNS), acts on receptors from three families: N-methyl-d-aspartate, α–amino-3-hydroxy-5-methyl-4-isoxazole propionic acid, and kainate. Subunits derived from five different genes (GluR5–7 and KA1–2) combine to form kainate receptors (1). Kainate receptors expressed presynaptically in the spinal cord (2) regulate gamma-aminobutyric acid (GABA) and glycine release (3) and those expressed in the hippocampus regulate GABA release presynaptically (4) and effects postsynaptically (5). They can modulate synaptic transmission and plasticity in the hippocampus (6,7) and influence the generation of kainate-induced seizures (8). Presynaptic kainate receptors in the hippocampus enhance both glutamate and GABA release (9).
Inhaled anesthetics enhance the peak currents mediated by GluR6-containing kainate receptors expressed in oocytes (10). Because kainate receptors appear to play a modulatory role in the CNS, they might also modulate the actions of anesthetics in vivo. As there are no specific antagonists for the GluR6 subunits of kainate receptors, we assessed the importance of GluR6 receptors to inhaled anesthetic actions by examining mice bearing mutations in GluR6 receptors. We reasoned that if GluR6 receptors modulate inhaled anesthetic actions, then wild-type animals should differ in their anesthetic requirement from animals bearing mutations that alter GluR6 receptor function.
We defined anesthesia by two end-points: immobility in response to a noxious stimulus and amnesia. Immobility was quantified as the minimum alveolar concentration (MAC) (11) of inhaled anesthetic that prevented movement in 50% of animals subjected to a noxious stimulus. Immobility depends largely on actions of anesthetics on the spinal cord (12–14). Amnesia was quantified using Pavlovian fear conditioning (15). Amnesia probably depends, at least in part, on anesthetic actions on higher brain structures, such as the amygdala (16).
These measures of anesthesia were applied to two types of genetically modified animals. The first were GluR6 knockout animals (8). GluR6 knockout animals behave normally on various sensorimotor tests and on the Morris water maze. However, they are less susceptible to kainate-induced seizures than wild-type littermate animals having functional GluR6 genes (8). Even in the absence of GluR6, some kainate receptors in these animals may still be assembled from the remaining kainate receptor subunits, GluR5, GluR7, KA1, and KA2.
The second genetically modified mouse had a mutation that blocks the editing of GluR6 subunits (17). In the normal adult, GluR6 pre-RNA undergoes post-transcriptional editing before removal of introns. In this process, editing converts a codon for glutamine (Q) to a codon for arginine (R) in the region encoding the second membrane domain of the subunit. In wild-type mice, approximately 85% of CNS GluR6 transcripts are edited from a Q to an R at the Q/R site. The presence of the edited subunit GluR6R in wild-type mice makes kainate receptors containing GluR6 less calcium permeable. Mice bearing the editing mutation differ from wild-type mice in that they are engineered to lack an intronic RNA editing site. This intronic sequence enables the folding of the RNA in such a way that the editing enzyme can work. The mutation engineered into this sequence disables the creation of that secondary structure. As a result, mice harboring this mutation have only unedited GluR6 transcripts; that is, the mutant mice contain only a Q at the Q/R site in the GluR6 subunit. The presence of GluR6Q confers greater calcium permeability on kainate receptors containing the GluR6Q subunit compared with receptors containing only normally edited GluR6R subunits. The GluR6Q mutation, unlike the knockout, is a dominant negative mutation that can affect receptors containing the GluR6Q subunit combined with any other subunit.
All studies in animals were approved by institutional animal care and use committees. GluR6 knockout mice were made via homologous recombination using a targeting construct that deleted the second membrane domain of the gene (8). Heterozygous mice containing the knockout were intercrossed to obtain 19 knockout (−/−), 21 wild-type (+/+), and 64 heterozygous (±) littermates. The founder 129SvImJ strain was purchased from the Jackson Laboratories (Bar Harbor, ME) while the C57BL/6Nhsd mice were from Harlan Sprague Dawley (Indianapolis, IN).
GluR6 editing mutants were created using homologous recombination to eliminate the intronic Q/R editing site (17). The homozygous editing mutants and wild-type animals were maintained and bred separately. Forty editing mutants and 40 wild-type mice were studied.
MAC for desflurane, halothane, and isoflurane was measured using the average of the largest concentration of anesthetic at which an animal responded to a tail clamp and the smallest concentration of anesthetic at which the animal would not respond (18). For a given genotype, MAC was the average of the measurements on all animals of that genotype.
Fear conditioning followed 1 wk of behavioral accommodation, during which time each animal was brought to our laboratory daily (including weekends) and handled for approximately 1 min. On the day of study, they were brought to the laboratory and exposed in their home cages to anesthetic or oxygen for 30 min to allow tissue equilibration with anesthetic. Based on results from pilot studies, the cage bottoms were heated to maintain normothermia. After 30 min of equilibration, each animal was quickly transferred to an individual gas tight conditioning (training) chamber, which contained the same concentration of anesthetic as the home cage. Each chamber included a footshock grid, heater, and speaker. Four chambers, connected in parallel to a fresh gas source, vaporizer, carbon dioxide absorbent, and fan, permitted concurrent study of 4 animals. After 3 min to allow each animal to explore its chamber, a 90 dB 2000-Hz tone sounded, coterminating with a 2-s footshock. This was repeated 3 times, separated by 1 min between tone and footshock. A 1-mA footshock was used for animals inhaling only oxygen, a 2 mA footshock was used for animals receiving 0.25% isoflurane, and a 3 mA footshock was used for animals breathing 0.5% or 0.75% isoflurane. All animals were observed continuously by closed circuit television during this time, and all sessions were videotaped. Immediately after the training session, the animals were returned to their home cage.
The next day, we determined whether the animals feared either the tone that preceded the footshock (fear to tone) or the box they had been conditioned in (fear to context). To test the effect of tone alone, we placed each mouse in an environment different from that in which they had been conditioned. This differently shaped chamber lacked a footshock grid, had been cleaned with a different cleaning solution (different odor) than that used in the original conditioning chamber, and was in a different room lit with only a dim red incandescent light (the room in which training occurred was brightly lit with fluorescent lights). After allowing 3 min of exploration in this chamber, the animals were exposed to a 90 dB, 2000-Hz tone for 8 min.
To assess fear to context, approximately 4 h after testing fear to tone the animals were placed in the training chamber (i.e., the chamber in which they had originally been conditioned) for 8 min. No tone was imposed.
For both the test for fear conditioning to tone and for fear conditioning to context, animals were observed by closed circuit television. No personnel were in the room during testing. Fear was assessed as immobility (i.e., freezing except for respiration). Immobility was measured in each animal every 8 s for 8 min for each animal, for a total of 60 measurements per animal. The number of freezes of the possible 60, expressed as a percentage, gave the probability of freezing (the “freeze score”) for each animal.
For each anesthetic, MAC values for different genotypes were compared using either analysis of variance or Student’s t-test with a Bonferroni correction for multiple comparisons. To establish whether the dose-response curves for fear conditioning differed among genotypes, a two-way analysis of variance was performed.
MAC did not differ among the wild-type littermate mice, GluR6 knockout mice, heterozygous mice, or the parental strains that contributed to the genetic background for these animals (129SvIMJ and C57Bl/6N hsd) for desflurane (P = 0.38), isoflurane (P = 0.07) or halothane (P = 0.78) (Table 1.)
For the GluR6 knockout animals and their littermates, increasing isoflurane concentration significantly decreased conditioning to tone (P = 0.001), but neither genotype (P = 0.358) nor the interaction of genotype with isoflurane concentration (P = 0.786) influenced conditioning to tone (Fig. 1). Fear conditioning to context showed similar relationships, where isoflurane concentration significantly affected conditioning (P = 0.012), but genotype (P = 0.58) or the interactions of genotype with isoflurane (P = 0.13) did not (Fig. 2).
MAC for desflurane in GluR6 editing mutants was smaller than in wild-type animals (P < 0.001), larger for isoflurane (P = 0.010), and not different for halothane (P = 0.252) (Table 2).
Fear conditioning to tone did not differ between GluR6 editing mutants and wild-type animals (P = 0.852) (Fig. 3). Isoflurane concentration dependently significantly influenced fear conditioning to tone (P = 0.008) but the interaction of concentration with genotype did not (P = 0.416). Likewise, isoflurane concentration significantly influenced fear conditioning to context (P = 0.003) but neither genotype alone (P = 0.82) nor the interaction of genotype with concentration influenced conditioning (P = 0.28) (Fig. 4).
Because kainate receptors play a modulatory role in synaptic transmission and plasticity and because alteration of synaptic transmission may mechanistically underlie anesthesia, we hypothesized that alterations in kainate receptors might, directly or indirectly, change MAC or modify the production of amnesia.
A loss-of-function mutation (knockout) in the GluR6 subunit from kainate receptors did not affect the capacity of desflurane, halothane, or isoflurane to produce immobility (MAC) or the capacity of isoflurane to impair conditioning to context or conditioning to tone. Such data, however, do not exclude the possibility that kainate receptors assembled from the remaining kainate receptor subunits (GluR5, GluR7, KA1, and KA2) may contribute to MAC or amnesia. They also do not exclude the possibility that these effects may have been unchanged owing to compensation for the gene knockout.
We reasoned that a gain-of-function mutation would not be expected to have the same behavioral effect as a loss-of-function mutation. Isoflurane MAC for gain-of-function (editing) mutants exceeded that in wild-type mice by 12%, but desflurane MAC for wild types exceeded that in editing mutants by 18%, and halothane MAC did not differ between the two strains. The observation of any change in MAC is intriguing given the in vitro finding that isoflurane modifies currents through kainate receptors comprised solely of GluR6 subunits. The different effects observed for the three different anesthetics in the editing mutant mice might be explained by the three anesthetics acting differently on GluR6 or differential modulation of the primary site of actions of the three anesthetics by the signaling mechanisms activated by kainate receptors containing GluR6. A mechanistically meaningful effect would require that the in vitro action of isoflurane would have to be opposite to that of desflurane whereas the effect of halothane would have to lie between.
This possibility remains to be tested. The finding of differential modulation of the effect of two volatile ether anesthetics (isoflurane and desflurane) provides behavioral evidence in mice that chemically related anesthetics may have differences in the way they are modulated. If this modulatory effect were direct on GluR6Q containing receptors, then these receptors should show the agent-specific effects we observed in animals. The effect may, however, be an indirect one, which would require an intact neuronal or synaptic system for investigation. The data from the studies reported here are consistent with this latter possibility for MAC. Knockout of GluR6 containing receptors would not be expected to have downstream effects (which are indirect), and indeed no effect of the knockout is observed. By contrast, the dominant negative GluR6Q mutation should have downstream effects because it increases calcium flux into neurons through the kainate subtype of glutamate receptors, which affects many second messenger pathways in neurons. These second messengers may, in turn, modulate other mechanisms of actions of inhaled anesthetics.
Neither GluR6 mutation in mice affected the capacity of isoflurane to interfere with fear conditioning to context or to tone. Based on the variability in fear conditioning scores and the number of animals studied, all of the dose response studies involving fear conditioning were adequately powered to detect a difference had there been one: there was more than an 80% probability (power) of detecting a difference of 10%–15% in the freeze score between genetically modified animals and control animals (where α = 0.05 defines the type I error rate) in all studies.
Potential limitations apply to the interpretation of behavioral results (MAC or amnesia) in genetically modified animals. Variations in the genetic background of animals may, for example, confound results. In our studies, the same cohort of mice was used for all our studies with all anesthetics. In addition, when mice received halothane they showed no change in MAC. This provides a control demonstrating that the effect of genetic background on MAC is negligible. Thus the difference in anesthetic actions observed are not attributable to a variability between different groups of animals, as might occur with variations in genetic background.
The similarity of the results in the GluR6 knockouts and editing mutants for fear conditioning supports the conclusion that these GluR6 subunits have little role in the amnesic effects of isoflurane as measured by fear conditioning. Because fear conditioning depends on higher brain structures, we conclude that GluR6 in the brain has little role in the amnestic effects of isoflurane as assessed by fear conditioning. By contrast, MAC depends largely on the spinal cord and our results obtained in the GluR6 editing mutants provides an intriguing result leaving open the possibility that kainate receptors containing GluR6 might either directly or indirectly modulate MAC by a small amount. In vitro studies that document opposing effects of isoflurane versus desflurane would support this interpretation.
1. Chittajallu R, Braithwaite S, Clarke V, Henley J. Kainate receptors: subunits, synaptic localization and function. Trends Pharmacol Sci 1999;20:26–34.
2. Hwang S, Pagliardini S, Rustioni A, Valtschanoff J. Presynaptic kainate receptors in primary afferents to the superficial laminae of the rat spinal cord. J Comp Neurol 2001;436:275–89.
3. Kerchner G, Wang G, Qiu C, et al. Direct presynaptic regulation of GABA/glycine release by kainate receptors in the dorsal horn: an ionotropic mechanism. Neuron 2001;32:477–88.
4. Kamiya H, Ozawa S. Kainate receptor-mediated inhibition of presynaptic Ca2+ influs and EPSP in area CA1 of the rat hippocampus. J Physiol 1998;509(Pt 3):833–45.
5. Ruano D, Lambolez B, Rossier J, et al. Kainate receptor subunits expressed in single cultured hippocampal neurons: molecular and functional variants by RNA editing. Neuron 1995;14:1009–17.
6. Bortolotto Z, Clarke V, Delany C, et al. Kainate receptors are involved in synaptic plasticity. Nature 1999;402:297–301.
7. Contractor A, Swanson G, Sailer A, et al. Identification of the kainate receptor subunits underlying modulation of excitatory synaptic transmission in the CA3 region of the hippocampus. J Neurosci 2000;20:8269–78.
8. Mulle C, Sailer A, Pérez-Otaño I, et al. Altered synaptic physiology and reduced susceptibility to kainate-induced seizures in GluR6-deficient mice. Nature 1998;392:601–5.
9. Kullman D. Presynaptic kainate receptors in the hippocampus: slowly emerging from obscurity. Neuron 2001;32:561–4.
10. Dildy-Mayfield J, Eger EI II, Harris R. Anesthetics produce subunit-selective actions on glutamate receptors. J Pharmacol Exp Ther 1996;276:1058–65.
11. Sonner J. Issues in the design and interpretation of minimum alveolar concentration (MAC) studies. Anesth Analg 2002;95:609–14.
12. Antognini JF, Schwartz K. Exaggerated anesthetic requirements in the preferentially anesthetized brain. Anesthesiology 1993;79:1244–9.
13. Borges M, Antognini J. Does the brain influence somatic responses to noxious stimuli during isoflurane anesthesia? Anesthesiology 1994;81:1511–5.
14. Rampil IJ. Anesthetic potency is not altered after hypothermic spinal cord transection in rats. Anesthesiology 1994;80:606–10.
15. Dutton R, Maurer A, Sonner J. The concentration of isoflurane required to suppress learning depends on the type of learning. Anesthesiology 2001;94:514–9.
16. Maren S, Fanselow M. The amygdala and fear conditioning: has the nut been cracked? Neuron 1996;16:237–40.
17. Vissel B, Royle G, Christie B, et al. The role of RNA editing of kainate receptors in synaptic plasticity and seizures. Neuron 2001;29:217–27.
© 2005 International Anesthesia Research Society
18. Sonner J, Gong D, Eger EI II. Naturally occurring variability in anesthetic potency among inbred mouse strains. Anesth Analg 2000;91:720–6.