γ-Aminobutyric Acid Type A Receptor β2 Subunit Mediates the Hypothermic Effect of Etomidate in Mice

All animal experiments were performed in accordance with the United Kingdom Animals (Scientific Procedures) Act 1986 and associated guidelines, including Local Ethical Review Process (Merck Sharp & Dohme, Harlow, United Kingdom) approval.

Generation and characterization of GABA^Areceptor β2N265S mutant mice has been previously described. 7Neuroscreening evaluation revealed that β2N265S mice did not differ from wild-type (WT) mice in body weight, rectal temperature, grip strength, beam balancing, or swimming ability. Mice used for these experiments were from the F4 generation and were aged between 2 and 5 months.

During isoflurane anesthesia (2–3% in O²), eight mice (four GABA^Aβ 2N265S and four WT controls) were implanted with intraperitoneal radiotelemetry transmitters (Data Sciences International, St. Paul, MN) for measuring CBT (model No. TA-F20). Adequate anesthetic depth was assessed by the tail withdrawal reflex. Strict asepsis was maintained, and mice received preoperative and postoperative analgesia (5 mg/kg subcutaneous carprofen) and antibiotics (long-acting preparation of 150 mg/kg subcutaneous amoxicillin). Mice were allowed to recover for 2 weeks before the first experiment.

Mice were singly housed in standard laboratory conditions with a 12 h light:dark lighting schedule and free access to food and water. For drug administration experiments, all dosing was conducted at circadian time 5 (CT5; 5 h after lights on). Room temperature was maintained at 21 ± 2°C.

Body temperature was recorded in the absence of drugs to ensure that there was no genotype difference in these parameters. CBT and locomotor activity were recorded for 72 h with 8-s sampling every 5 min.

Etomidate (Hypnomidate; Janssen-Cilag, Berchem, Belgium; 2 mg/ml) or vehicle (35% propylene glycol in water for injection) was administered via  either the intraperitoneal (5 and 20 mg/kg) or the intravenous route (5-mg/kg single bolus via  lateral tail vein). These same doses and routes have previously been examined for their effect on spontaneous locomotor activity and rotarod performance in GABA^Aβ2N265S mice and WT controls. 4Data were sampled for 8 s each minute and recorded for 6 h after return of righting reflex for animals dosed with anesthetic. Propofol (Rapinovet; Genusxpress, York, United Kingdom; 10-mg/ml solution in an aqueous isotonic emulsion) was administered as a single intravenous bolus at a dose of 24 mg/kg. This dose of propofol was selected to produce a similar duration of LORR 7to 5 mg/kg intravenous etomidate to compare the hypothermic responses to these two agents.

Pretreatment CBT was the mean ± SEM of the previous 10-min data. Normothermia  was defined as return to pretreatment mean CBT.

Loss of righting reflex occurred at 20 mg/kg intraperitoneal and 5 mg/kg intravenous etomidate and 24 mg/kg intravenous propofol. After intravenous administration, mice immediately lost consciousness. For intraperitoneal administration, mice lost consciousness within 60 s of drug administration. Mice were placed on their backs, and the time until mice were able to recover normal righting posture was measured.

Effect of genotype on differences in mean CBT over time were tested for statistical significance using a repeated-measures analysis of variance (Statistica; Statsoft, Tulsa, OK). Significant genotype differences between minimum CBT (CBT^min), maximum CBT (CBT^max), and LORR duration were determined using the Student t  test.

Baseline measurement of CBT in the absence of any drug treatment was performed in both genotypes. Both GABA^Areceptor β2N265S mice and their WT controls display a well-defined circadian variation in CBT (fig. 1). There was no difference in the mean active (dark) phase CBT (WT, 36.89°± 0.3°C; β2N265S, 36.89°± 0.3°C). During the rest (light) phase, there was also no genotype difference in CBT (WT, 36.13°± 0.3°C; β2N265S, 36.18°± 0.3°C). These data were used to select time of drug administration in further experiments in which CBT did not significantly vary (CT5). This reduced any compounding effect of circadian CBT change during anesthetic challenge.

With an intraperitoneal dose that produced an LORR (20 mg/kg; duration: WT, 66 ± 5.1 min; β2N265S, 68.5 ± 6.2 min), the mean CBT^minin WT mice was significantly lower (28.46°± 0.72°C) compared with GABA^Aβ2N265S mice (CBT^min, 32.26°± 0.50°C; P < 0.05) (fig. 2). Predosing temperatures were not significantly different (WT, 36.51°± 0.23°C; β2N265S, 35.89°± 0.35°C). After recovery of righting reflex, return to normothermia was faster in β2N265S mice. GABA^Areceptor β2N265S mice attained predose CBT 2 h 31 min after LORR; however, WT controls did not regain normothermia during the 5-h postdosing recording period.

After dosing of vehicle (35% wt/vol propylene glycol in water for injection), both genotypes showed an increase in CBT, which then diminished over time. The profile of this change was not significantly affected by the genotype of the mice (F < 0.05) (fig. 3A). As before, there was no genotype difference in pretreatment CBT (WT, 36.32°± 0.4°C; β2N265S, 36.49°± 0.45°C).

An intraperitoneal dose of 5 mg/kg etomidate does not produce LORR in either mouse genotype. 4After dosing, a significant hypothermic response was observed in WT mice (CBT^min, 34.28°± 1.1°C; P < 0.05). In GABA^Areceptor β2N265S mice, a significant increase in CBT was observed (CBT^max, 37.28°± 0.46°C; P < 0.05); this is a “stress-induced hyperthermic response” resulting from dosing and restraint stress and is well described in the literature. 20As before, pretreatment CBT did not differ between genotypes (WT, 36.06°± 0.25°C; β2N265S, 36.37°± 0.46°C) (fig. 3B).

To examine the hypothermic response during a brief period of anesthesia, etomidate was administered intravenously (5-mg/kg single intravenous bolus). There was no significant effect of genotype on duration of LORR (WT, 359 ± 7 s; β2N265S, 381.5 ± 9 s).

As with the intraperitoneal dosing experiment, CBT in WT mice was significantly lower compared with GABA^Aβ2N265S mice (WT CBT^min, 32.21°± 0.77°C; β2N265S CBT^min, 33.88°± 0.48°C; P < 0.05) (fig. 4). Temperature began to recover shortly after recovery of righting reflex in GABA^Areceptor β2N265S mice; however, in WT mice, CBT continued to decrease after recovery of righting reflex. WT mice took significantly longer to recover normothermia (mean = 188 min) compared with GABA^Areceptor β2N265S mice (mean = 37 min; P < 0.05).

The effect of the intravenous anesthetic propofol on CBT was examined. For comparison with the hypothermic response of etomidate, doses were chosen that result in a similar duration of LORR in WT mice (24 mg/kg intravenous propofol: LORR, 349.6 ± 9.1 s; 5 mg/kg intravenous etomidate: LORR, 327 ± 7 s). Propofol produced a similar degree of hypothermia in both WT and GABA^Areceptor β2N265S mice (WT CBT^min, 35.1°± 0.16°C; β2N265S CBT^min, 34.5°± 0.4°C), whereas etomidate produced an enhanced hypothermic effect in WT mice (CBT^min, 32.2°± 0.24°C) compared with GABA^Areceptor β2N265S mice (CBT^min, 33.9°± 0.09°C) (figs. 5A and B). These conclusions are supported by a two-way repeated-measures analysis of variance that indicated significant effects of both genotype (F^1,12= 50.0, P < 0.0001) and anesthetic (F^1,12= 110.9, P < 0.0001), and there was a significant interaction between the two (F^1,12= 45.1, P < 0.0001).

We have demonstrated that the β2 subunit of the GABA^Areceptor mediates a significant proportion of the hypothermic response to etomidate. Previously, we have shown that the sedative/ataxic effects of etomidate can be separated from the anesthetic activity of etomidate. Anesthetic effects (as measured by LORR and the presence of burst-suppression activity in the electroencephalogram seem to be mediated by the β3 GABA^Areceptor subunit, whereas the sedative and ataxic effects are mediated by the β2 subunit. After etomidate anesthesia, β2N265S mice recover rotarod performance and attain normal levels of locomotor activity faster compared with WT controls. 7It is highly likely that the reduced hypothermia and ability to recover normothermia at a faster rate are major contributing factors in the improved functional recovery observed in β2N265S mice.

Although β2 subunit–containing GABA^Areceptors constitute approximately 50% of all GABA^Areceptors in the central nervous system, remarkably, loss of the β2 subunit is not lethal. 21In β2^−/−mice, etomidate produces a LORR although the period of LORR is reduced. 22This confirms the role of β3 in anesthesia as measured by LORR. The reduction in duration of LORR in β2^−/−mice may result from a loss of sedative and ataxic effect mediated by the β2 subunit. Etomidate produces some LORR in mice containing the corresponding β3 mutation (β3N265M), 23and the duration of the LORR period is reduced, but these animals show no anesthetic effect of etomidate as measured by loss of pedal withdrawal reflex. Etomidate still produces a LORR in β3^−/−mice, which is reduced compared with WT 24; however, β3^−/−mice present with severe developmental abnormalities, 25and it is possible that developmental compensation, perhaps by other GABA subunits, in brain regions critical for anesthesia may underlie the reduced effect of etomidate in these mice.

Baseline recordings of CBT in the absence of drug treatment were made to ensure that the β2N265S mutation did not result in a change in CBT. Recordings made over 3 days revealed that both genotypes expressed a well-defined circadian variation in CBT, with an increase during the active (dark) period compared to the rest (light) phase. Data from this study were also used to select a time of day for drug administration when CBT was relatively stable (i.e. , during the nadir), and results could be interpreted without confounding effects of circadian variation. All drug administration experiments were conducted at the same circadian time point (CT5) to allow direct comparison between experiments. In all experiments, there was no effect of genotype on pre-treatment CBT. At 20 mg/kg (intraperitoneal), we observed a very pronounced and extended period of hypothermia in the WT mice, which was much reduced in the β2N265S mice. We have previously shown that this dose significantly impairs postrecovery rotarod performance in WT mice. 7The hypothermic response in both genotypes was of a longer duration that the period of LORR.

After intraperitoneal administration of a subanesthetic dose of etomidate (5 mg/kg), a hypothermic response was present in the WT mice that was absent in the β2N265S mice. This indicates that etomidate, acting via β2-containing receptors, mediates hypothermic effects that are independent of anesthesia (as defined by LORR). In β2N265S mice, a modest “stress hyperthermia” response was observed with the subanesthetic etomidate dose. This is due to the stress of restraint and injection and was also observed with administration of vehicle solution in both genotypes.

Although etomidate has a favorable cardiovascular profile, it is used clinically only for the induction of general anesthesia (rather than maintenance) because it suppresses cortisol production. 26Conversely, propofol is widely used for both induction and maintenance of anesthesia. Because of their different uses, it is difficult to compare their hypothermic effects directly in the clinical population; however, intraoperative hypothermia occurs with both agents. 27In a study in which the effects of both agents were compared in hypothermic patients for cerebral oxygen supply and demand during cardiopulmonary bypass, propofol seemed to have better recovery characteristics after rewarming recommenced, perhaps suggesting that propofol resulted in less hypothermia. 28 

It was observed in both genotypes that after recovery of righting reflex, the mice engaged in a brief period of intense locomotor activity. Although β2N265S mice have higher levels of locomotor activity over a few hours after recovery (data not shown), it is unlikely that higher levels of locomotor activity can account wholly for the faster return to normothermia in β2N265S mice because there is no sudden increase in CBT after recovery of righting reflex in either genotype. During a brief anesthetic period induced intravenously with 5 mg/kg etomidate, CBT continued to decrease after recovery in the WT group and did not increase during the brief recovery period of locomotor activity, again suggesting that activity is not a significant factor in improved normothermic recovery in β2N265S mice.

Although β2N265S mice have a much reduced hypothermic response to etomidate anesthesia compared with WT mice, a significant hypothermic effect is still evident when LORR occurs. The hypothermic effect of propofol, which has in vitro  activity at β1-, 2-, and 3-containing receptors, 29is similar to that of etomidate in β2N265S mice. Although our study strongly implicates the GABA^Aβ2 receptor subunit in the hypothermic response to etomidate, during anesthesia, many factors are involved in the generation of anesthetic hypothermia. These include abolition of behavioral responses, interference with cutaneous vascular smooth muscle tone, and reduction of nonshivering thermogenesis. Although a significant proportion of anesthetic hypothermia could be overcome with an anesthetic agent lacking in vivo  activity at β2-containing receptors, it is unlikely that anesthetic hypothermia can be completely abolished because of hypothermic mechanisms that are GABA independent. Indeed, hypothermia produced by etomidate in β2N265S mice was only evident when loss of consciousness occurred, with subanesthetic doses not producing hypothermia. This indicates that mechanisms other than theGABA^Areceptor β2 subunit may also contribute to anesthetic hypothermia. Given that the hypothermic effect of propofol is much less pronounced than that produced by etomidate in WT mice, the reasons for this remain to be investigated. Consistent with the reduced hypothermia observed in this study, both WT and β2N265S mice demonstrated rapid recovery in rotarod performance after propofol-induced anesthesia. 7We have previously reported that etomidate is without activity at other receptors and ion channels at clinically relevant concentrations, 7and based on previously measured effective plasma concentrations at the doses used in the current study, it is unlikely that direct activation of the GABA^Areceptor occurs. It is therefore feasible to assume that the differential hypothermic effect observed in β2N265S mice is dependent on specific GABA^Aβ2 subunit properties. Although we suggest the major effects observed here are centrally mediated, GABA^Areceptors in the spinal cord also express β2 and β3 subunits, and these may also play a role.

In mammals, the principle brain region controlling body temperate is the POAH and contains both warm-and cold-sensitive neurons. Warm-sensitive neurons are activated at higher temperatures and stimulate processes that reduce body temperature. Cold-sensitive neurons are activated by relatively cooler brain temperatures and generate hypothermic responses. GABA, acting via  GABA^Areceptors, directly inhibits cold-sensitive neurons, whereas modulation of warm-sensitive neurons by GABA occurs at the site of afferent input. 30Central or systemic administration of GABA and GABA agonists results in hypothermia, 18whereas GABA antagonists produce hyperthermia. 30It is interesting to note that in the POAH, there is a high ratio of β2-containing GABA^Areceptors compared with β3-containing receptors. An immunohistochemical study 31found intense labeling of processes within the POAH with a specific β2 antibody along with moderate “diffuse” immunoreactivity. In contrast, no processes were labeled with the β3 antibody, although some diffuse labeling was reported. Studies examining the distribution of GABA^Aβ2 subunit expression reveal that expression of β2 messenger RNA is weak in the POAH. 32Consistent with the results showing strong β2 immunoreactivity in processes but no cell body staining, 31it is likely that the source of β2 immunoreactivity in the POAH is from afferent sources. It is highly possible that the hypothermic effect of etomidate observed in WT mice results from activation of β2-containing GABA^Areceptors in the POAH and that this effect is much reduced in β2N265S mice because of the relative lack of effect of etomidate at the mutant β2-containing receptors.

Previously, we have demonstrated that an equivalent depth of anesthesia is attained in GABA^Aβ2N265S mice and WT controls as determined by analysis of isoelectric activity in the electroencephalogram. 7Hypothermia can affect the Bispectral Index value, an integrated measure of the human intraoperative electroencephalogram. 33–35Because β2N265S mice experience much less hypothermia during etomidate anesthesia compared with WT controls, it is curious that no detectable differences were measured in the electroencephalograms of these mice. However, it is possible that a much greater hypothermic response than that produced in the WT mice in this study is required before differences in the electroencephalogram are apparent. 34 

Previously, we have observed that when a heat lamp was used to maintain the surrounding air temperature at 27°± 2°C, that the duration of LORR was significantly longer in WT compared with knock-in mice at higher doses (10 and 15 mg/kg) of intravenous etomidate. 7Because body temperature of the mice was not measured in these experiments, it is not possible to determine what effect active warming had on CBT. However, it is highly likely that a more pronounced hypothermic effect would have occurred in the WT mice compared with the knock-in mice even when actively warmed. Hypothermia has been shown to augment anesthetic potency and to significantly delay the metabolism of anesthetics and concomitantly administered agents resulting in a longer recovery period and increased hypnotic hangover. 36,37It would be interesting to study plasma concentrations of etomidate on recovery to determine any difference in etomidate metabolism. Faster metabolism of etomidate in the knock-in mice could be a contributing factor to the improved functional recovery and reduced sedation observed in the knock-in β2N265S mice. Anesthetic hypothermia can also result in a depression of cognitive functioning on awakening, which could also contribute to the improved functional performance observed in the β2N265S mice.

The data presented here support the role of GABA^Aβ2-containing receptors being involved in the hypothermic response to etomidate. We have previously shown that the β2 subunit mediates the ataxic and sedative effects of etomidate. Evidence that the β2 subunit also produces hypothermia indicates that several undesirable properties of etomidate anesthesia are mediated by this subunit. Because the anesthetic qualities (i.e. , loss of consciousness, lack of purposeful movement in response to noxious stimuli, and presence of burst-suppression activity on the electroencephalogram) of etomidate seem to be mediated by the β3 subunit, this further indicates that anesthetics that selectively activate specific subunits may produce surgical anesthesia with improved recovery characteristics.