Halothane and desflurane requirements in alcohol-tolerant and -nontolerant rats{dagger}

L. L. Firestone1,*, E. R. Korpi2, L. Niemi3, P. H. Rosenberg3, G. E. Homanics1 and J. J. Quinlan1

1Anesthesiology Research Laboratories, University of Pittsburgh, Pittsburgh, PA, USA. 2Department of Pharmacology and Clinical Pharmacology, University of Turku, Turku, Finland. 3Department of Anaesthesia, Helsinki University Central Hospital, Helsinki, Finland, and Biomedical Research Center, Alko Group, Ltd, Helsinki, Finland

{dagger}Presented in part to the 1995 annual meeting of the Society for Neuroscience (November 12 – 16, 1995, San Diego, CA).

Accepted May 30, 2000


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
On the basis of data implicating GABAA receptors in the effects of volatile general anaesthetics, we hypothesized that alcohol-, barbiturate-, and benzodiazepine-sensitive alcohol-nontolerant (ANT) rats would also be more sensitive than alcohol-tolerant (AT) rats to two clinical general anaesthetics with differing potencies, halothane and desflurane. The obtunding effect of halothane and desflurane on mature ANT (n=17) and AT (n=16) rats was assessed by the loss-of-righting reflex endpoint. ANT rats were significantly (P<0.0001) more sensitive to the obtunding effects of both halothane and desflurane (ED50=0.45±0.03% atm for ANT vs 0.95±0.04% atm for AT and 2.16±0.17 vs 3.69±0.13% atm, respectively). The immobilization effect of halothane and desflurane was assessed with the tail clamp/withdrawal endpoint. ANT rats were more sensitive to the effects of halothane (ED50=1.10±0.08% atm for ANT vs 1.72±0.09% atm for AT; P<0.0001) but not desflurane (ED50=6.25±0.25% atm for ANT vs 5.85±0.21% atm for AT). The data presented support the hypothesis that volatile anaesthetics interact with specific neuronal proteins (possibly GABAA receptors) and agree with recent hypotheses that different elements of the anaesthetic state are produced by separate sites or mechanisms.

Br J Anaesth 85; 2000: 757–62

Keywords: anaesthetics volatile, halothane; anaesthetics volatile, desflurane; brain, GABA; rat


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
{gamma}-Aminobutyric acid type A (GABAA) receptors mediate the postsynaptic actions of the major inhibitory neurotransmitter in mammalian brain. GABAA receptors are believed to be pentameric combinations of various subunits, of which 16 ({alpha}1–6, ß1–3, {gamma}1–3, {rho}1–2, {delta}, and {epsilon}) have been so far identified.1 Subunit combinations determine diverse pharmacological GABAA receptor subtypes in a brain cell- and region-specific manner.2 Thus, it is possible that various drug actions and behaviours are dependent on certain GABAA receptor subtypes.

Abundant in vitro data demonstrating that structurally heterogeneous general anaesthetics enhance GABA-gated hyperpolarizing currents (reviewed by Tanelian et al.3) have led to the hypothesis that general anaesthetics act primarily via the GABAA receptor. One approach to testing this hypothesis is to identify experimental animals with a specific alteration in the genes encoding brain GABAA receptors and assay for changes in their anaesthetic requirement. Previous investigations have utilized GABAA receptor mutants arising spontaneously,4 due to ionizing radiation,5 or as a result of targeted gene disruption.68 In the case of a spontaneous mutation, a selectively outbred rat line highly susceptible to the motor ataxia produced by alcohol,9 benzodiazepines and barbiturates is available.10 The putative underlying mutation involves a single nucleotide substitution in the GABAA receptor {alpha}6 subunit resulting in an Arg100->Gln substitution.11 Given the susceptibility of this rat line, termed alcohol-nontolerant (ANT), to central depressants known to act via the GABAA receptor, we hypothesized that ANT rats would also be cross-sensitive to the actions of volatile general anaesthetics. To test this, anaesthetic requirements were determined in ANT and control, alcohol tolerant (AT) rats, using two volatile agents with widely differing potencies, as well as two separate endpoints reflecting obtundation and immobilization.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
The protocol for these experiments was approved by the Animal Ethics Committee of Helsinki University Central Hospital.

ANT and AT rats are outbred lines derived from crossbred Wistar, Sprague-Dawley, and Long-Evans rats,9 selected and bred for sensitivity (ANT) or resistance (AT) to ethanol. The selection procedure has been based on motor performance under the influence of a moderate, nonhypnotic dose of ethanol (2 g kg–1, i.p.). The phenotype is rechecked every second generation. Stocks of AT and ANT rats are kept in stainless steel wire mesh cages on a 12-h light/dark cycle (light on at 6 a.m.) at the ambient temperature (20±2°C) and relative humidity (50±5%). For the present study, male rats of the F50 generation were housed four per cage and had continuous access to RMI(E)SQC pellet feed (SDS Ltd, Witham, Essex, UK) and tap water. The age of rats used in this study was 3.5 months (350–430 g).

Six groups (three ANT, three AT) of four to six rats of either genotype (unknown to the observer) were studied during halothane and desflurane anaesthesia, in random order. Each group was placed within a Plexiglas chamber warmed by heating pads from below and an infrared lamp from above. The composition of the chamber’s atmosphere was monitored continuously on-line with Datex Capnomac UltimaTM and NormacTM devices (for oxygen/carbon dioxide, and volatile anaesthetic concentrations, respectively). Fresh oxygen flow was delivered via a Dameca oxygen vernitrol at rates of 1.5–8 litres min–1. Volatile agents were added to the fresh gas stream by means of either a Fluotec 3TM (Cyprane Ltd, Keighley, UK) or an Ohmeda Tec 4TM vaporizer (Ohmeda, Inc., Steeton, West Yorkshire, UK). Carbon dioxide concentration was maintained at <1.0% atm by means of soda lime scattered on the floor of the chamber.

After a 20-min equilibration period at a given anaesthetic concentration, rats were gently placed supine and observed for the ability to roll over spontaneously to the fully prone position within 30 s. Responses were scored in a quantal fashion. Immediately thereafter, the base of the tail of each rat in the group was clamped with a full-sized Kelly haemostat to the second ratchet position, and gross withdrawal within 5 s scored quantally. Withdrawal was defined as jerking of the head, twisting of the neck, or movement of an extremity; coughing, swallowing, chewing, and grimacing motions were not scored as positive responses, nor were increases in the respiratory rate.12 Rectal temperatures were then measured using a digital thermometer (Exacon MC8700, Denmark) with an accuracy of ±0.1°C. The rats were allowed to recover in an oxygen-rich atmosphere for at least 20 min before they were equilibrated with the next higher anaesthetic concentration and retested for both the obtundation and immobilization endpoints. For both agents, 9–13 different concentrations were tested in each group of rats.

After all experiments were completed, halothane and desflurane response data from the three ANT groups (n=17 rats) or three AT groups (n=16 rats) were pooled. The dose-response relationships were analysed by fitting to a logistic equation,13 which yielded estimates of the ED50 and slopes for each agent, as well as their respective standard errors. The ANT and AT group responses were then compared, using a method making the fewest assumptions about the distribution of errors.14 Briefly, the estimated variances for the AT and ANT groups were calculated from the standard errors; the sum of these then yielded the estimated variance of the difference in ED50 between groups. The ratio of this difference to its standard error was referred to a standard normal distribution.


    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Halothane- and desflurane-induced obtundation and immobilization were reversible in all rats from both the AT and ANT groups. Rectal temperatures were maintained throughout the experiments to within 1.5°C of the mean value obtained after each dose was tested in a given group.

Data obtained during halothane anaesthesia are summarized in Table 1 and Figures 1 and 2. These six experiments included three with AT rats (n=16) and three with ANT rats (n=17). For each genotype/endpoint combination, the arithmetic mean of the ED50 values obtained separately closely agreed with the ED50 obtained by a free fit of pooled data (for example, 0.98 vs 0.95% atm, respectively, for AT rats tested by loss-of-righting-reflex).


View this table:
[in this window]
[in a new window]
 
Table 1 Halothane requirements* in AT and ANT rats as measured by loss of righting reflex and withdrawal following tail clamp stimulus. *Median effective doses (ED50, expressed in % atm), dose-response curve slopes, and their respective standard errors (SE) are listed for individual experiments. Results of refitting combined data from triplicate experiments are also shown. {dagger}P<0.0001 vs AT ED50. {ddagger}P>0.49 vs AT slope. §§P<0.0001 vs AT ED50. ¶¶P=0.10 vs AT slope
 


View larger version (14K):
[in this window]
[in a new window]
 
Fig 1 Halothane requirements in ANT vs AT rats defined by the endpoint of loss of righting reflex. Data from six separate experiments are plotted: each data point is the fractional response of the group (four to six rats) to a given halothane concentration. Curves are derived as in Materials and methods.

 


View larger version (14K):
[in this window]
[in a new window]
 
Fig 2 Halothane requirements in ANT vs AT rats defined by withdrawal following a tail clamp stimulus. Data and curve fitting as in Figure 1.

 
Results from all individual experiments with desflurane are summarized in Table 2 and Figures 3 and 4, including three experiments each with AT (n=15) and ANT (n=15) rats. Again, means of the separate ED50 values agreed with the values from pooled data.


View this table:
[in this window]
[in a new window]
 
Table 2 Desflurane requirements* in AT and ANT rats as measured by loss of righting reflex and withdrawal following tail clamp stimulus. *Median effective doses (ED50, expressed in % atm), dose-response curve slopes, and their respective standard errors (SE) are listed for individual experiments. Results of refitting combined data from triplicate experiments are also shown. {dagger}P<0.0001 vs AT ED50. {ddagger}P>0.25 vs AT slope. §§P>0.50 vs AT ED50. ¶¶P>0.15 vs AT slope
 


View larger version (15K):
[in this window]
[in a new window]
 
Fig 3 Desflurane requirements in ANT vs AT rats defined by the endpoint of loss of righting reflex. Data and curve fitting as in Figure 1.

 


View larger version (14K):
[in this window]
[in a new window]
 
Fig 4 Desflurane requirements in ANT vs AT rats defined by withdrawal following a tail clamp stimulus. Data and curve fitting as in Figure 1.

 
Since the slopes (Tables 1 and 2) of the pooled loss-of-righting-reflex data for the AT vs ANT groups were not significantly different (P>0.49 for halothane; P>0.25 for desflurane), it is valid to compare the ED50s for those groups. Halothane sensitivity in the AT line (0.95±0.04% atm) was less than previously reported values in wild-type mice (generally 0.6–0.7% atm),6 8 15 but ANT rats (0.45±0.03% atm) were more sensitive than previous reports. No previous reports of desflurane sensitivity in the loss-of-righting reflex assay in rodents are available for comparison. For both agents, the ANT rats were significantly more sensitive with respect to obtundation than AT rats (P<0.0001 for halothane; P<0.0001 for desflurane) (Tables 1 and 2). The magnitude of the difference in sensitivity was slightly greater for halothane (ANT ED50 equal to 47% of the AT value) than for desflurane (ANT ED50 equal to 58% of the AT value).

Similarly, the slopes of the pooled tail clamp/withdrawal data for the AT vs ANT groups were not significantly different (P>0.10 for halothane; P>0.15 for desflurane). Halothane sensitivity in the AT line (1.72±0.09% atm) was again less than that previously reported for MAC in wild-type rats (approximately 1–1.2% atm), while the ANT value (1.1±0.08% atm) was similar to previous reports. Desflurane ED50 values for tail clamp for both AT (5.85±0.21% atm) and ANT (6.25±0.25% atm) lines were within the range of previous reports in rats (5.7–6.8% atm).16 17 Comparison of the ED50 values for these groups showed that for halothane, the ANT rats were significantly more sensitive than were AT rats with respect to the immobilization endpoint (P<0.0001), but ANT and AT rats had similar sensitivity to the immobilizing effects of desflurane (P>0.50).


    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
ANT rats were originally bred for their sensitivity to alcohol9 and subsequently were found to be cross-sensitive to benzodiazepines and barbiturates.10 In the present study, we hypothesized that ANT rats would also exhibit cross-sensitivity with volatile anaesthetics.

We found that ANT rats were significantly more sensitive than AT rats to the obtunding effects of halothane and desflurane. ANT rats were also more sensitive to the immobilizing effects of halothane, but not desflurane. This inconsistent response to desflurane is unlikely to be anomalous, since our measurement of tail-clamp ED50 for desflurane in AT rats (5.85±0.21% atm) agreed closely with the minimum alveolar concentration (MAC) of desflurane in oxygen reported in normothermic Sprague-Dawley rats (5.72±0.40% atm).14 It is also unlikely that the immobilization assay result with desflurane was an artefact of the group sizes studied: a 20% difference between ED50 values would have been detectable in these groups by the statistical methods employed. Instead, this nonuniformity of response to the volatile agents more likely indicates that the immobilizing effect has different neuronal and neurochemical mechanisms than does the obtunding effect. Therefore, these data agree with the hypothesis that separate elements of the anaesthetic state are produced by independent mechanisms.18 The oil/gas partition coefficient of halothane is about 12 times greater than that of desflurane.19 The relatively greater lipid solubility of halothane allows for the possibility that halothane interacts with more hydrophobic regions of the membrane surrounding the GABAA or another receptor.20

ANT rats were first noted to display unusual cerebellar GABAA receptor binding pharmacology,21 which were subsequently proven to have a point mutation in the N-terminal extracellular portion of the {alpha}6 subunit.11 This {alpha}6 subtype is expressed almost exclusively in cerebellar granule cells,22 which are intrinsic to the cerebellar cortex and are the most ubiquitous neuron found in the mature mammalian nervous system. The role of cerebellum in motor coordination, posture maintenance, balance, and muscle tone is clear,23 but its role in conscious behaviours, including motor learning, is currently an area of intense research (reviewed by Glickstein).24 Classically, granule cells relay information between cerebellar input (mossy fibre) and output (Purkinje) neurons. The {alpha}6 subunit-containing GABAA receptor is among the most sensitive subtype to inhibition by GABA;25 thus, potentiation of GABA actions at granule cells by an anaesthetic could sharply reduce Purkinje cell output from the cerebellar cortex to the deep cerebellar nuclei, and subsequently to the brain. Indeed, attenuation of Purkinje output could explain, at least in part, the motor ataxia of intoxicated ANT rats, but might not be critical for deep anaesthesia.

In vitro, some anaesthetics (propofol and trichloroethanol) have significantly lower efficacy, but not apparent affinity, at {alpha}6 subunit-containing GABAA recombinant receptors expressed in cell lines, as compared with those containing the {alpha}1 subunit.26 However, pentobarbital potentiates GABA-gated currents more significantly in {alpha}6-containing cells than in those containing {alpha}1.27 The significance of these findings in vivo remains to be determined. Studies of mice with {alpha}5 and {gamma}3 GABAA receptor gene deletions suggest that these subtypes are not essential for the halothane response.5 In contrast, targeted gene disruption of the ß3 subunit of the GABAA receptor increases halothane and enflurane MAC but not the ED50 for loss of righting.8

Despite extensive characterization of the ANT {alpha}6 GABAA receptor genotype, there is still a possibility that other mutations coexist in ANT rats that affect the central nervous system.28 Additional but unidentified mutations could, in principle, have become enriched in the selected population along with the defective {alpha}6 gene and be the actual basis for the drug-response phenotype reported here. The role of possible unidentified mutations could be assessed in crossbreeding experiments in which AT rats are mated with ANT, siblings of the first generation are mated again, and anaesthetic sensitivity is determined in this resulting generation of rats. The degree of dominance of the {alpha}6 mutation should determine the distribution of anaesthetic sensitivity in rats of the second generation. However, such experiments are costly and require large numbers of animals. Additional unidentified mutations may be excluded by using targeted disruption, or site-specific alteration, of the gene in question. Such an approach was recently used to establish that the GABAA {gamma}2 receptor subunit is essential to benzodiazepine agonist action in vivo; targeted disruption of the subtype gene by homologous recombination resulted in mice for whom diazepam was behaviourally inactive.7 Mice completely lacking the {alpha}6 subunit, created using similar techniques, do not differ from wild type mice in their sensitivity to either halothane or enflurane in the loss-of-righting reflex assay or the tail clamp assay.6 These mice also exhibit normal responses to ethanol and pentobarbital6 29 but their response to the motor ataxic effects of diazepam was enhanced.29 The dramatic difference in the pharmacological profile of rats with an intact but altered {alpha}6 subunit and null allele mice completely lacking any {alpha}6 subunits suggests that other undocumented mutations in the ANT line may be responsible for at least some of their abnormal behavioural responses, but the difference could also be explained by substitution of other subunits for the missing {alpha}6 subunits in the null allele mice.6 An alternative explanation is that the {alpha}6 null allele is pharmacologically and behaviorally very different from the gain of function mutation that is present in the ANT rat line.

In conclusion, we have documented that the obtunding effects of halothane and desflurane are markedly enhanced in ANT rats compared to AT controls. The immobilizing effects of halothane, but not desflurane, were also enhanced in the ANT rat line. Since ANT rats are known to contain a point mutation in the {alpha}6 subunit of the GABAA receptor that segregates with the ANT phenotype, it is tempting to speculate that the enhanced behavioral sensitivity to the volatile agents is also due to this mutation. However, it is equally plausible that an additional mutation(s) is the basis for the observed phenotype.


    Acknowledgements
 
The authors thank Mr Markku Hamalainen of Pharmacia, Finland, for the generous donation of desflurane for these experiments, Drs Tarja Randell and Markku Paloheimo for valuable technical assistance; and Ulla Rankamo, of Datex-Engstrom/Instrumentarium Ltd, Helsinki, for generously providing the Capnomac UltimaTM, the NormacTM, and factory calibration standards. Special appreciation is expressed to Lisa Cohn for editorial advice and to Dr Peter Winter and the University Anesthesiology and Critical Care Medicine Foundation for long-term support. This study is supported by grants from the National Institutes of Health (GM35900 and GM52035 to LLF); the Foundation for Anesthesia Education and Research (Starter Award to JJQ).


    Footnotes
 
* Corresponding author: Department of Anesthesiology and Critical Care Medicine, University of Pittsburgh School of Medicine, A1305 Scaife Hall, Pittsburgh, PA 15261, USA Back


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
1 Macdonald RL, Olsen RW. GABAA receptor channels. Annu Rev Neurosci 1994; 17: 569–602[ISI][Medline]

2 Lüddens H, Korpi ER. Biological function of GABAA/benzodiazepine receptor heterogeneity. J Psychiatr Res 1995; 29: 77–94[ISI][Medline]

3 Tanelian DL, Kosek P, Mody I, MacIver MB. The role of the GABAA receptor/chloride channel complex in anesthesia. Anesthesiology 1993; 78: 757–76[ISI][Medline]

4 Firestone L, Korpi E, Niemi L, Rosenberg P, Homanics G, Quinlan J. Volatile general anesthetic responses in alcohol-nontolerant (ANT) rats. Soc Neurosci Abstracts 1995; 21: A499

5 Firestone L, Quinlan J, Homanics G, Firestone S, Winter P, Russell L, Rinchik E. Halothane responses in mice with deletion of the genes encoding for GABAA receptor subunits {alpha}5 and {gamma}3. Anesthesiology 1994; 81: A861[ISI]

6 Homanics GE, Ferguson C, Quinlan JJ, Daggett J, Snyder K, Lagenaur C, Mi ZP, Wang XH, Grayson DR, Firestone LL. Gene knockout of the alpha-6 subunit of the gamma-aminobutyric acid type A receptor: Lack of effect on responses to ethanol, pentobarbital, and general anesthetics. Mol Pharmacol 1997; 51: 588–96[Abstract/Free Full Text]

7 Gunther U, Benson J, Benke D, Fritschy J, Reyes G, Knoflach F, Crestani F, Aguzzi A, Arigoni M, Lang Y, Bluethmann H, Mohler H, Luscher B. Benzodiazepine-insensitive mice generated by targeted disruption of the gamma 2 subunit gene of gamma-aminobutryic acid type A receptors. Proc Natl Acad Sci USA 1995; 92: 7749–53[Abstract]

8 Quinlan JJ, Homanics GE, Firestone LL. Anesthesia sensitivity in mice lacking the ß3 subunit of the GABAA receptor. Anesthesiology 1998; 88: 775–80[ISI][Medline]

9 Eriksson K, Rusi M. Finnish selection studies on alcohol-related behaviors: general outline. In: McCLearn GE, Deitrich GE, Erwin G, eds. Development of animal models as pharmacogenetic tools. Washington DC: US Government Printing Office, 1981; 87–117

10 Hellevuo K, Kiianmaa K, Juhakoski A, Kim C. Intoxicating effects of lorazepam and barbital in rat lines selected for differential sensitivity to ethanol. Psychopharmacology 1987; 91: 263–7[ISI][Medline]

11 Korpi ER, Kleingoor C, Kettenmann H, Seeburg PH. Benzodiazepine-induced motor impairment linked to point mutation in cerebellar GABA-A receptor. Nature 1993; 361: 356–9[ISI][Medline]

12 Quasha AL, Eger EI, Tinker JH. Determination and applications of MAC. Anesthesiology 1980; 53: 315–34.

13 Waud D. On biologic assays involving quantal responses. J Pharmacol Exp Ther 1972; 183: 577–697[ISI][Medline]

14 Alifimoff J, Firestone L, Miller K. Anesthetic potencies of secondary alcohol enantiomers. Anesthesiology 1987; 66: 55–9[ISI][Medline]

15 Deady JE, Koblin DD, Eger EI, Heavner JE, D’Aoust B. Anesthetic potencies and the unitary theory of narcosis. Anesth Analg 1981; 60: 380–4[Abstract]

16 Laster MJ, Liu J, Eger EI, Taheri S. Electrical stimulation as a substitute for the tail clamp in the determination of minimum alveolar concentration. Anesth Analg 1993; 76: 1310–12[Abstract]

17 Eger EI, Johnson BH. MAC of I-653 in rats, including a test of the effect of body temperature and anesthetic duration. Anesth Analg 1987; 66: 974–6[Abstract]

18 Eger EI, Koblin DD, Harris RA, Kendig JJ, Pohorille A, Halsey MJ, Trudell JR. Hypothesis – Inhaled anesthetics produce immobility and amnesia by different mechanisms at different sites. Anesth Analg 1997; 84: 915–18[ISI][Medline]

19 Koblin DD. Mechanisms of action. In: Miller RD, ed. Anesthesia. New York: Churchill Livingstone, 1994; 80

20 Hales TG, Olsen RW. Basic pharmacology of intravenous induction agents. In: Bowdle TA, Horita A, Kharasch ED, eds. Pharmacological basis of anesthesiology. New York: Churchill Livingstone, 1994; 295–306

21 Uusi-Oukari M, Korpi ER. Diazepam sensitivity of the binding of an imidazobenzodiazepine, [3H]Ro15–4513, in cerebellar membranes from two rat lines developed for high and low alcohol sensitivity. J Neurochem 1990; 54: 1980–7[ISI][Medline]

22 Lüddens H, Pritchett DB, Kohler M, Killisch I, Keinanen K, Monyer H, Sprengel R, Seeburg PH. Cerebellar GABA-A receptor selective for a behavioural alcohol antagonist. Nature 1990; 346: 648–51[ISI][Medline]

23 Ito M. The cerebellum and neural control, Raven Press, 1984

24 Glickstein M. The cerebellum and motor learning. Curr Opinion Neurobiol 1992; 2: 802–6[Medline]

25 Korpi ER, Lüddens H. Regional {gamma}-aminobutyric acid sensitivity of t-butylbicyclophosphoro[35S]thionate binding depends on {gamma}-aminobutyric acid A receptor {alpha} subunit. Mol Pharmacol 1993; 44: 87–92[Abstract]

26 Krasowski M, O’Shea S, Rick C, Hadingham K, Whiting P, Czajkowski C, Harrison N. A subunit isoform influences GABAA receptor modulation by propofol. Neuropharmacology 1997; 36: 941–9[ISI][Medline]

27 Thompson S, Whiting P, Wafford K. Barbiturate interactions at the human GABAA receptor: dependence on receptor subunit combination. Br J Pharmacol 1996; 117: 521–7[Abstract]

28 Harris R. Alcohol sensitivity. Nature 1990; 348: 589[ISI][Medline]

29 Korpi ER, Koikkalainen P, Vekovischeva OY, Makela R, Kleinz R, Uusi-Oukari M, Wisden W. Cerebellar granule-cell-specific GABAA receptors attenuate benzodiazepine-induced ataxia: evidence from alpha 6-subunit-deficient mice. Eur J Neurosci 1999; 11: 233–40[ISI][Medline]