Effect of Isoflurane and Other Potent Inhaled Anesthetics on Minimum Alveolar Concentration, Learning, and the Righting Reflex in Mice Engineered to Express α¹γ-Aminobutyric Acid Type A Receptors Unresponsive to Isoflurane

Institutional animal care and use committees (University of California, San Francisco, California, and University of Pittsburgh, Pittsburgh, Pennsylvania) approved our studies of male and female mice bred from mice heterozygous for the S270H and L277A mutations, producing wild-type (SL/SL), heterozygous knock-in (SL/HA), and homozygous knock-in (HA/HA) mice. Animals were housed under a 12-h light-and-dark cycle and had continuous access to standard mouse chow and tap water.

Brain slices were prepared from adult mice (age range, 28–52 days), and miniature inhibitory postsynaptic currents (mIPSCs) were recorded and analyzed as previously described.10Interneurons and pyramidal cells in the CA1 stratum radiatum and stratum lacunosum-moleculare layer of the hippocampus were identified using differential interference contrast microscopy.10The identity of pyramidal cells was then verified by pronounced accommodation of action potential firing, to distinguish them from interneurons also present in the pyramidal cell layer.11As in a previous study,10fewer than 10% of the recorded neurons in the CA1 pyramidal cell layer were interneurons. Whole cell patch clamp recordings were obtained with borosilicate glass pipettes that had a resistance of 2–4 MΩ when filled with pipette solution containing 130 mm cesium methanesulfonate (CH³SO³Cs), 8.3 mm sodium methanesulfonate (CH³SO³Na), 1.7 mm NaCl, 1 mm CaCl², 10 mm EGTA (intracellular free calcium ion concentration: approximately 0.01 μm), 2 mm Mg-ATP, 0.3 mm Na-GTP, and 10 mm HEPES (pH 7.2, osmolarity 295 mOsm). In the recording chamber, brain slices were constantly superfused at a rate of 3.5 ml/min with artificial cerebrospinal fluid containing 117 mm NaCl, 3.6 mm KCl, 25 mm NaHCO³, 1.2 mm NaH²PO⁴, 1.2 mm MgCl², 2.5 mm CaCl², and 11 mm glucose (pH 7.4, osmolarity 305 mOsm). To eliminate action potential–evoked synaptic events, 0.5 μm tetrodotoxin was added to the artificial cerebrospinal fluid. Drugs were applied via  a gravitational glass-syringe/polytetrafluoroethylene-tube system to reduce the loss of volatile anesthetics. A 10-min preapplication of each drug was performed before the 5-min data recording epoch that was used for analysis; this ensured that the anesthetic drugs were at equilibrium throughout the data collection period.12The mIPSCs were recorded at room temperature for 5 min with an acquisition frequency of 10 kHz and a filter frequency of 2 kHz. The holding potential was −60 mV. Analysis of the mIPSC data traces was performed as described previously.10Bath application of 20 μm bicuculline blocked all mIPSC activity (n = 3, data not shown). The weighted decay time constant (τ^decay) was calculated as τ^decay= (A^f× t^f+ A^s× t^s)/(A^f+ A^s).10To ensure an adequate recording configuration throughout the experiments, we determined the access resistance before and after each recording using the membrane properties feature of the acquisition software (pClamp 9.0; Axon Instruments, Foster City, CA). Recordings were excluded from the analysis when the difference between membrane and access resistance decreased below a 10-fold value.

Volatile anesthetic solutions were prepared in airtight surgical bags. Each bag contained 100 ml artificial cerebrospinal fluid including 0.5 μm tetrodotoxin plus volatile anesthetic, as described previously. The values obtained for amplitudes and τ^decaywere compared to the respective control values in volatile anesthetic–free solution to calculate the volatile anesthetic effect on mIPSCs as percentage change for each experiment, and comparisons between genotypes were made using analysis of variance.10 

Studies of learning and memory applied techniques described previously.13–15Mice (n = 254) were exposed to a target concentration of isoflurane (confirmed with gas chromatography) for 30 min before training. Each animal was then rapidly transferred to a training chamber containing the target concentration of isoflurane and was allowed to explore the chamber for 3 min before training began. Mice then received three tone-shock pairs (tone training) consisting of a 30-s tone (90 dB, A-scale, 2,000 Hz) coterminating with a 2-s electric shock (11-Hz bipolar square waves). Ninety seconds separated tone-shock pairs. Animals were returned to their home cages within 60 s after the last shock. The shock currents were 2 mA at 0, 0.1%, 0.2%, 0.3%, and 0.4% isoflurane. At 0.5% and 0.75% isoflurane, we used 3-mA currents. For each test, the anesthetic concentration was calculated as the mean of the concentrations measured in the training chambers before and after training of that set of four mice.

Both context and tone testing took place the day after training. For tone testing, each animal was placed in a special test chamber and, after 3 min of exploration, a tone (90 dB, A-scale, 2,000 Hz) was continuously sounded for 8 min; no shock was administered. Context testing was conducted 1–2 h later. For context testing, the mice were returned to the chambers used to supply the electric shocks, whereas tone testing took place in chambers providing an entirely different environment from that provided by the training chambers. For both context and tone testing, four mice were observed simultaneously, one in each of four separate test chambers, via  a video camera. No personnel were in the tone or context training or testing rooms during training or testing. To score freezing to either tone or context, an observation of one of the four animals was made every 2 s. Therefore, each animal was scored once every 8 s. Behavior was judged as freezing if there was no visible movement except for breathing.16The observation periods were video-recorded for scoring by a blinded observer.

The percentage of time an animal froze during the 8-min observation periods was calculated as the number of observations judged to be freezing divided by the total number of observations in 8 min, i.e. , 60 observations.16For each group score at a given isoflurane concentration, the mean freeze score and standard error of the mean (SE) were calculated.

A least-squares linear regression was applied to the raw data for fear to context for 0–0.5% isoflurane. For fear to tone data, a least-squares linear regression was applied to the data for 0–0.75% isoflurane. The concentration (EC⁵⁰) producing a 50% decrease in freezing scores from control (no isoflurane) and the SE of this concentration were calculated and used to compare freezing to context and freezing to tone for the three genetic groups. In addition, two-way analysis of variance using genotype and anesthetic concentration as factors was performed to determine whether there was a difference in the effect of genotype on freeze scores. A value of P < 0.05 was regarded as significant for all comparisons.

We measured MAC for desflurane, isoflurane, and halothane in 88 mice (28 SL/SL, 36 SL/HA, and 24 HA/HA) as described previously.17These mice were a random subset of mice after measurements of fear conditioning had been made. Each mouse was used as a subject for one, two, or three of the test anesthetics but was used only once for a test of a given anesthetic. For each mouse, MAC was calculated as the mean of the greatest inspired concentration that permitted movement in response to tail clamp and the smallest concentration that prevented movement. Genotyping of each test mouse was accomplished after the determinations of learning and memory, and MAC. Differences in MAC between genotypes were determined using analysis of variance.

Adult (8- to 12-week-old) male and female mice (n = 15–18 per genotype) were tested for sensitivity to inhaled anesthetics using the LORR assay as described.18,19Briefly, mice were placed in individual wire mesh cages in a rotating carousel in a sealed acrylic chamber and anesthetized. Within the chamber, carbon dioxide was maintained at less than 1% atm, and temperature was maintained at 35°± 0.2°C. Chamber atmosphere and anesthetic concentrations were monitored continuously with a Datex Capnomac Ultima device (Datex-Ohmeda, Helsinki, Finland), and fresh oxygen was delivered at a rate of 1.5 l/min. Mice were equilibrated with the desired concentration (% atm) of halothane (Halocarbon Laboratories, River Edge, NJ), isoflurane (Halocarbon Laboratories), or enflurane (Anaquest, Madison, WI) for 15 min, after which they were scored by an observer blind to the genotypes. Scores were quantal; a positive response for LORR occurred when mice were not able to right themselves two times during five revolutions of the carousel at 4 rpm. Mice were allowed to recuperate in oxygen for at least 20 min before being equilibrated with the next anesthetic concentration. The dose–response relation for each anesthetic was analyzed using the Z statistic.20 

Miniature inhibitory postsynaptic currents were recorded from interneurons (n = 36) and pyramidal cells (n = 23) in the CA1 subfield of the hippocampus. The cell properties (input resistance, cell capacitance) were not significantly different between genotypes in interneurons and pyramidal cells, and there was also no significant difference between the groups of interneurons and pyramidal cells. In interneurons, the decay phase of mIPSCs recorded in neurons from HA/HA animals was significantly faster compared with mIPSCs recorded in SL/SL and SL/HA animals (figs. 1A and B). In pyramidal cells, we observed a significant faster decay time constant in HA/HA animals compared with SL/SL animals (fig. 1B). However, mIPSC amplitude and frequency were not significantly different between the genotypes (data not shown).

Bath application of isoflurane (0.16 and 0.31 mm; 1 MAC corresponds to 0.31 mm) prolonged the decay phase of mIPSCs in SL/SL animals, and this effect was concentration dependent. The effect of isoflurane was significantly reduced in HA/HA animals compared with wild-type animals as shown in example current traces in figure 1C. Halothane application (0.27 mm, the aqueous concentration at 1 MAC) significantly prolonged the decay time of mIPSCs in SL/SL animals and in the mutant genotypes, with no significant differences (fig. 1D). To determine the average effect of the volatiles on the decay time constants, we calculated the percentage change between control conditions and after 10 min of volatile application. The average data are shown in two bar diagrams in figures 2A and B. We found that the effect of isoflurane on mIPSC decay times was reduced in interneurons and pyramidal cells (figs. 2A and B). The amplitudes and frequencies of mIPSCs were not significantly affected by volatile anesthetic application in any of the three genotypes (data not shown).

Mice were tested for the amnestic effects of inhaled anesthetics using a fear conditioning assay. Control values for fear to context and fear to tone (i.e. , in the absence of anesthetic) did not differ among the three genotypes of mice. Slopes and intercepts did not differ by genotype among mice conditioned to context or among those conditioned to tone (figs. 3 and 4). Mice of different genotypes did not differ in conditioning to tone (F^2,248= 0.11, P = 0.89) or context (F^2,226= 0.16, P = 0.85). The EC⁵⁰± SE (in % atm isoflurane) for freezing to tone was 0.34 ± 0.05 for SL/SL mice (n = 72), 0.34 ± 0.04 for SL/HA mice (n = 113 mice), and 0.31 ± 0.06 for HA/HA mice (n = 66). For freezing to context, these values were 0.28 ± 0.05 for SL/SL (n = 64), 0.25 ± 0.04 for SL/HA (n = 101), and 0.24 ± 0.06 for HA/HA (n = 64).

Mice were tested for the immobilizing effects of inhaled anesthetics in response to a noxious stimulus using the standard tail clamp/withdrawal assay. MAC values did not differ between genotypes for any anesthetic tested (table 1). Using one-way analysis of variance, for isoflurane, F^2,48= 0.05 (P = 0.95); for desflurane, F^2,32= 1.32 (P = 0.29); and for halothane, F^2,38= 0.07 (P = 0.94).

Mice were also tested for the motor ataxic effects of inhaled anesthetics using the standard LORR assay. In this assay, halothane EC⁵⁰values did not differ between SL/SL (0.68 ± 0.02) and HA/HA (0.69 ± 0.02) mice (fig. 5A). However, differences in the EC⁵⁰values were observed for isoflurane (0.63 ± 0.02 for SL/SL vs.  0.72 ± 0.02 for HA/HA; P < 0.01; fig. 5B) and enflurane (1.09 ± 0.03 for SL/SL vs.  1.21 ± 0.03 for HA/HA; P < 0.001; fig. 5C).

The knock-in mice harboring the HA/HA mutation were designed to enable us to evaluate the significance of the α¹subunit to the anesthetic effects of isoflurane. The electrophysiologic recordings from hippocampal neurons show that the decay time constants of mIPSCs in interneurons from HA/HA mice decay significantly faster compared with the mIPSCs of wild-type mice (SL/SL). In addition, whereas isoflurane and halothane prolonged mIPSCs in the wild-type animals, the effects of isoflurane (but not halothane) was significantly reduced in the HA/HA animals. These findings are all consistent with the expression in the hippocampus of mutated α¹subunits, and with a role for GABA^ARs containing α¹subunits in generating inhibitory synaptic currents in pyramidal cells and interneurons. The smaller residual effect of isoflurane in the HA/HA animals presumably reflects the activation of a distinct population of isoflurane-sensitive GABA^ARs at hippocampal synapses, perhaps containing α²and/or α³subunits. At GABAergic synapses onto cerebellar Purkinje cells,21for example, or thalamocortical relay neurons,22the IPSC is generated by activation of a more homogeneous population of GABA^ARs containing α¹subunits, and IPSC kinetics are fast (τ= 5–10 ms). The kinetics of hippocampal IPSCs10are slower and presumably reflect a mixed population of receptors.

The mice harboring the α¹subunit HA/HA mutation displayed normal sensitivity to the amnestic effects of isoflurane. This suggests that GABA^Areceptors containing the α¹subunit do not mediate the capacity of isoflurane to produce amnesia. This is an interesting result, because GABA^Areceptors containing the α¹subunit make up approximately half of the total receptor population in the rodent brain.7In addition, our hippocampal electrophysiology shows a decrease in the pharmacologic effect of the anesthetic in vivo  in a structure known to be important for learning, the hippocampus. Clearly, such receptors are not critical to the amnestic effect of isoflurane, as they have been shown to be for benzodiazepines.23The amygdala is another structure known to be important for the acquisition of fear conditioning, and its circuitry is thought to use GABA^ARs that contain other subunits, such as α²and α⁵, that have been implicated in the anxiolytic and amnestic effects of the benzodiazepines.23Therefore, several factors limit the generalization of our results from the GABA^AR-containing α¹subunits and isoflurane to other GABA^AR subtypes and anesthetics. In addition, the faster decay of IPSCs in the HA/HA mice compared with wild-type mice may also lead to compensation for the HA/HA mutation. Although GABA^ARs containing the α¹subunit may not mediate the capacity of isoflurane to produce amnesia, this may not apply to other inhaled anesthetics.

We also conclude that GABA^ARs containing the α¹subunit do not mediate the capacity of desflurane, isoflurane, or halothane to produce immobility in the face of noxious stimulation (“MAC”). This result is less surprising in the sense that inhaled anesthetics are believed to act at the level of the spinal cord, where the GABA^AR α¹subunit is not highly expressed, although immobility induced by the intravenous anesthetics propofol and etomidate is reduced by mutations in the GABA^AR β³subunit.3The current results for MAC are also consistent with results from pharmacologic studies that suggest no mediation by GABA^AR of the immobility produced by inhaled anesthetics.6,24,25 

The HA/HA mice showed an increase in the EC⁵⁰for isoflurane and enflurane, but not halothane in the LORR test. Because the GABA^AR containing the mutated α¹subunit does not respond to isoflurane or enflurane, but retains sensitivity to halothane, this result provides evidence that the circuits involved in the LORR response involve GABA^ARs containing α¹subunits. The LORR response produced by ethanol was not altered by this mutation26—this difference may reflect the much larger potentiation of GABA^AR function by anesthetic concentrations of isoflurane as compared with ethanol. The α¹HA/HA mutation produces a 14% change in LORR for isoflurane, suggesting that other GABA^AR subtypes or additional anesthetic targets are important in generating LORR.

Taken together, these studies suggest the surprising conclusion that several behavioral actions of isoflurane do not require activation of GABA^ARs containing α¹subunits, despite the ubiquity of this receptor subtype at subsynaptic locations in a variety of cortical and subcortical regions, whereas LORR does indeed require GABA^ARs containing α¹subunits. Recent studies suggest that extrasynaptic GABA^ARs, which can contain α⁴, α⁵, α⁶, and/or Δ subunits, may be especially sensitive to ethanol and volatile anesthetics,27–30and these may also prove to be significant for the behavioral actions of these drugs. Mice bearing mutations containing α⁴, α⁵, α⁶, and/or Δ subunits in these receptors that parallel those engineered for the current report for α¹subunits will allow an evaluation of this hypothesis.