Flumazenil

There has been continuing debate as to whether conditioned suppression (Estes & Skinner, 1941) or discriminated punishment (Geller & Seifter, 1960) are appropriate animal models for the assessment of clinical anxiolytics and related compounds. Arguably, conditioned suppression has higher face validity as it involves the signaled delivery of an unavoidable aversive event, and the conditioned suppression of positively reinforced operant behavior that occurs is not instrumental in reducing the frequency of aversive events. This procedure, however, has been criticized as not being selectively responsive to anxiolytics (e.g., Cook & Sepinwall, 1975; Dantzer, 1977; but see Millenson & Leslie, 1974, for a review that reached a different conclusion).

We examined its utility in a recent series of experiments in which conditioned suppression (of lever pressing by rats) was modified by acute or chronic treatment with an anxiolytic (chlordiazepoxide) or an anticonvulsant that acts at the GABA/benzodiazepine receptor complex (valproate), or by one of these drugs in combination with a possible antagonist. In various experiments these included bicuculline, picrotoxin, Ro15-1788 (flumazenil) and delta-amino-n-valeric acid (Leslie, Shephard, & Toal, 1994; Shephard, Toal, & Leslie, 1990; Toal, Leslie, & Shephard, 1991). The most striking findings were as follows: Chlordiazepoxide and valproate both reduced conditioned suppression; both these effects were antagonized by picrotoxin; and the effect of chlordiazepoxide was also antagonized by flumazenil. These findings are consistent with an account of the GABA/benzodiazepine receptor complex in which both valproate and picrotoxin act at the chloride ion channel, while chlordiazepoxide acts at the benzodiazepine receptor and this is antagonized by flumazenil (in earlier publications this drug was referred to as Ro15-1788). However, we have failed on several occasions to find effects on conditioned suppression of compounds believed to act at GABAa or GABAb sensitive sites, respectively, muscimol and bicuculline, and baclofen and delta-amino-n-valeric acid (Shephard et al., 1990; Toal et al., 1991).

The present experiments extended our previous findings by using a barbiturate, pentobarbital, as the anxiolytic along with some of the putative antagonists that had been effective in our previous experiments. In general, barbiturates have been found to have the same effects as benzodiazepines in paradigms used to assess the effects of anxiolytics (DeLorey, Kissin, Brown, & Brown, 1993; Leidenheimer, Whiting, & Harris, 1993; and see Iversen & Iversen, 1981; Polc, 1988, for reviews). We sought to establish the similarity of action for a barbiturate to that of benzodiazepines for our conditioned suppression paradigm and to identify the pattern of interaction with putative antagonists.

In the first experiment, pentobarbital was given at 10 mg/kg, a dose found to be effective in several behavioral paradigms, and the putative antagonists used were flumazenil and picrotoxin at 1.5 mg/kg, a dose found to be effective in our earlier experiments. Because picrotoxin had no systematic effect, a second experiment was conducted with the same dose of pentobarbital, a higher dose of picrotoxin (2.0 mg/kg), and Ro5-3663 (5.0 mg/kg), a compound believed to act in the same fashion as picrotoxin at the chloride ion channel associated with the GABA/benzodiazepine receptor complex.

Method

Subjects

Eight experimentally naive male Sprague-Dawley rats, approximately 100 days old at the start of the experiment, were used. Rats were housed two to a cage with water freely available and were maintained at close to 85% of their free-feeding weight.

Apparatus

Four two-lever Campden Instruments rat test chambers (Model CI 410) were used. Only the left lever was operative. A 2.8-W stimulus light was situated 4 cm above each lever and a third was 15 cm above the floor and midway between the two levers. During sessions the chamber was lit by a 2.8-W houselight. The reinforcer was 5-s access to 5% sucrose solution (by weight) that was delivered by a motor-driven dipper into an aperture in the floor of a recessed tray situated between the two levers at floor level. The tray was covered by a lightly hinged clear plastic flap and was illuminated by a 2.8-W bulb during reinforcer delivery. Scrambled shock could be delivered to the grid floor, made of stainless steel rods 1.3 cm apart, from a constant current shock source (Campden Instruments Model 521C). Each test chamber was encased in a sound-attenuating housing that was fitted with a ventilating fan. A Data General NOVA 2/10 minicomputer programmed in ACT (Millenson, 1975) controlled the experiment and collected data.

Procedure

Experiment 1 (Rats L102, L106, L110, L114). Rats were feeder trained and lever press responses were shaped. After this training period, lever pressing was reinforced according to a variable-interval (VI) 48-s schedule. Each daily session (there were 5 test days per week) lasted 30 minutes. After 15 sessions of the VI 48-s schedule, a stimulus was presented for 60 s at irregular intervals on three occasions during each session. The stimulus consisted of the three stimulus lights flashing at 2.5 Hz (ontime = offtime). After five sessions of stimulus presentations, the rats were exposed to an additional five sessions during which a 0.5-s duration electric shock, unconditioned stimulus (US), was delivered contiguous with termination of the flashing lights, which became the conditioned stimulus (CS). The intensity of the US was initially 0.1 mA but was increased to 0.25 mA over the five sessions of CS-US pairings. The drug phase followed immediately. Drugs used were sodium pentobarbital (10 mg/kg), picrotoxin (Sigma: 1.5 mg/kg), and flumazenil (Roche: 10 mg/kg) and all were dissolved or suspended in the vehicle of 0.9% saline/1% Tween 80. All drugs were given by intraperitoneal injection 20 minutes before experimental sessions in a volume of 1 ml/kg. Each drug session was followed by one session on which a vehicle injection was given. The drug treatments were given in a semirandom sequence to each of the subjects. Following two initial vehicle sessions and one pentobarbital session, pentobarbital, pentobarbital plus picrotoxin, and pentobarbital plus flumazenil were each given three times. All doses are expressed as salt and were chosen on the basis of pilot and published work. Combined drug administrations were given as two injections.

Experiment 2 (Rats L13, L14, L15). The apparatus and procedure were the same as in the first experiment. After 10 sessions of the VI 48-s schedule, there were 4 sessions in which the CS was presented alone, and then 10 sessions of CS-US pairings. The drug phase followed immediately. Drugs used were pentobarbital (10 mg/kg), picrotoxin (2.0 mg/kg), and Ro5-3663 (Roche: 5 mg/kg). In the drug phase two vehicle days were followed by three pentobarbital sessions (each followed by a vehicle session) and there were then five pentobarbital, five pentobarbital plus picrotoxin, and five pentobarbital plus Ro5-3663 sessions given in a semirandom sequence, with drug sessions alternating with vehicle sessions. Again, combined drugs were given as two injections. Four rats completed the experiment. However, persistent apparatus failure occurred with one rat, and only data from the remaining three subjects will be presented.

Results

The data presented are rates of responding during the CS (CS rates) and rates of responding during the 60 s preceding each 60-s CS period (pre-CS rates). In the first experiment, conditioned suppression was maintained on CS-US pairing sessions and sessions where injections of the saline vehicle were given throughout the drug phase (although one rat, L114, showed some temporary loss of conditioned suppression in the middle of the drug phase). Figure 1 gives average pre-CS and CS rates for each animal for saline vehicle administration sessions, and the various types of sessions on which drugs were administered. (Figure 1 omitted)

On drug sessions, which alternated with vehicle sessions, pentobarbital markedly increased the CS rate without systematically affecting the pre-CS rate (the exception to this was Rat L110 where the pre-CS rate did increase on four successive drug sessions but not systematically thereafter). On drug sessions where pentobarbital was given with picrotoxin, there was no additional or antagonistic effect of the picrotoxin. However, when pentobarbital was given with flumazenil, there was on most occasions an additional effect of flumazenil with less conditioned suppression being observed than under any other condition. On several such sessions, the CS rate was very similar to the pre-CS rate.

Sign tests were conducted to make the following comparisons: pre-CS with CS rates on all the vehicle sessions that preceded drug sessions (X= 36, N = 36, p

The results of Experiment 2 are shown in Figure 2. (Figure 2 omitted) As in Experiment 1, all three rats showed sustained conditioned suppression on CS-US pairing sessions and vehicle sessions in the drug phase, and pronounced increases in CS rates occurred on pentobarbital days with some evidence of increases in pre-CS rates. On drug sessions when picrotoxin (2.0 mg/kg) was given with pentobarbital, CS rates were lower than in the nearest pentobarbital sessions although higher than on vehicle sessions, whereas pre-CS rates did not change systematically. On drug sessions when Ro5-3663 was given with pentobarbital a similar pattern was seen, with CS rates lower than on the nearest pentobarbital sessions, but no systematic changes in pre-CS rates.

For Experiment 2, sign tests were conducted for the following comparisons: pre-CS with CS rates on all vehicle sessions that preceded drug sessions (X = 51, N = 51, p

Discussion

Conditioned suppression was maintained across the two experiments. As no systematic effects were found of any drug combination on pre-CS rate, the results can be discussed solely in terms of the changes in CS rates. However, it is worth noting that an analysis of a suppression ratio measure (for example, CS rate/CS rate + pre-CS rate) would yield qualitatively similar results.

Previous behavioral (Tenen, 1967) and pharmacological (Leeb-Lundberg & Olsen, 1982) findings suggested that pentobarbital should reduce conditioned suppression, and this occurred within both of the present experiments. This adds to the extensive literature showing comparable effects of barbiturates to those of benzodiazepines in anxiety-related behavioral procedures. Similar actions have been found with the Geller-Seifter conflict procedure (Geller & Seifter, 1960), neophobia (Poschel, 1971), punished drinking (Petersen, Paschelke, Kehr, Neilsen, & Braestrup, 1982), and pentylenetetrazole discrimination (Gherezghiher & Lal, 1982). In relation to our previous studies of conditioned suppression, pentobarbital was as efficacious as chlordiazepoxide and valproate. The dose of pentobarbital used here (10 mg/kg) was selective in that it did not systematically affect pre-CS rates, and it is in the range of doses generally found to be behaviorally active.

Although picrotoxin at 1.5 mg/kg had previously been found to antagonize the effects on conditioned suppression of valproate (Shephard et al., 1990) and chlordiazepoxide (Leslie et al., 1994; Toal et al., 1991), it was only effective at the higher dose of 2.0 mg/kg in the present study. The antagonism of the effects of pentobarbital was replicated with Ro5-3663. These findings are consistent with a model of the GABA/benzodiazepine receptor complex in which barbiturates facilitate opening of the chloride-ion channel, and picrotoxin and Ro5-3663 act at the same site within the receptor complex but have the opposite effect (Polc, 1988). This interpretation implies, for example, that higher doses of barbiturates would overcome the antagonistic effects of the dose levels of picrotoxin and Ro5-3663 used in the present study.

The most striking result in the present study was the further increase in CS rates produced by adding flumazenil to pentobarbital. This combination of drugs produced less conditioned suppression than any other drug or combination used. This contrasts with our previous findings that flumazenil antagonized the effects of chlordiazepoxide administered acutely (Toal et al., 1991) or chronically (Leslie et al., 1994) on conditioned suppression. The effect of valproate was not modified by the addition of flumazenil (Shephard et al., 1990). Some previous behavioral studies (Gherezghiher & Lal, 1982; Petersen et al., 1982) and pharmacological studies (Polc, 1981; Scholfield, 1983) have found different interactions of flumazenil with barbiturates and benzodiazepines. The present observation that flumazenil potentiates the effect of pentobarbital may be explained by one or both of two considerations. Firstly, flumazenil shows anxiolytic effects in its own right in some circumstances (File & Pellow, 1986). However, the dose used to induce this effect had to be well in excess of 10 mg/kg and this dosage produces anxiogenic actions in a majority of studies where intrinsic activity of flumazenil is detected (see File and Pellow, 1986, for review). Moreover, we have failed to detect intrinsic actions of flumazenil (10 mg/kg) on conditioned suppression (Toal et al., 1991), although that particular study employed a low shock intensity and was largely designed to detect anxiogenic effects. A second possibility is that blocking the benzodiazepine site of the receptor complex sensitizes it to drug actions at the chloride-ion channel, perhaps by preventing the action of endogenous ligands (see File & Pellow, 1986; Polc, 1988, for a further discussion of this possibility). This hypothesis would also explain why flumazenil potentiates the anxiogenic effects of picrotoxin and pentylenetetrazole (File & Pellow, 1985).

The present findings are valuable in three ways: They add to the data available on the effects of anxiolytics and related compounds on conditioned suppression, they provide further information on the possible nature of the GABA/benzodiazepine receptor complex, and they suggest that a more extensive study of the interaction between flumazenil and pentobarbital would be illuminating. In particular, the effect of flumazenil in combination with a range of doses of pentobarbital would indicate whether the interaction observed in the present experiments reflects a shift in the dose-response curve or whether addition of flumazenil produces a greater anxiolytic effect than any dose of the barbiturate.

References

COOK, L., & SEPINWALL, J. (1975). Reinforcement schedules and extrapolations to humans from animals in behavioral pharmacology. Federation Proceedings, 34, 1889-1897.

DANTZER, R. (1977). Behavioral effects of benzodiazepines: A review. Biobehavioral Reviews, 1, 71-86.

DELOREY, T. M., KISSIN, I., BROWN, P., & BROWN, G. B. (1993). Barbiturate-benzodiazepine interactions at the gamma aminobutyric acid A receptor in rat cerebral cortical synaptoneurosomes. Anesthesia and Analgesia, 77, 598-605.

ESTES, W. K., & SKINNER, B. F. (1941). Some quantitative properties of anxiety. Journal of Experimental Psychology, 29, 390-400.

FILE, S. E., & PELLOW, S. (1985). Does the benzodiazepine antagonist Ro15-1788 reverse the actions of picrotoxin and pentylenetetrazole on social and exploratory behaviour? Archives Internaionales de Pharmacodynamie et de Therapie, 277, 272-279.

FILE, S. E., & PELLOW, S. (1986). Intrinsic actions of the benzodiazepine receptor antagonist Ro15-1788. Psychopharmacology, 88, 1-11.

GELLER, I., & SEIFTER, J. (1960). The effects of meprobamate, barbiturates, d-amphetamine and promazine on experimentally induced conflict in rats. Psychopharmacologia, 1, 482-492.

GHEREZGHIHER, T., & LAL, H. (1982). Ro15-1788 selectively reverses antagonism of pentylenetetrazole-induced discriminative stimuli by benzodiazepines but not by barbiturates. Life Sciences, 31, 2455-2460.

IVERSEN, S. D., & IVERSEN, L. L. (1981). Behavioral pharmacology (2nd ed.). New York: Oxford University Press.

LEEB-LUNDBERG, L. M. F., & OLSEN, R. W. (1982). Heterogeneity of benzodiazepine receptor interactions with gamma-aminobutyric acid and barbiturate receptor sites. Molecular Pharmacology, 23, 315-325.

LEIDENHEIMER, N. J., WHITING, P. J., & HARRIS, R. A. (1993). Activation of calcium-phospholipid-dependent protein-kinase enhances benzodiazepine and barbiturate potentiation of the GABA-A receptor. Journal of Neurochemistry, 60, 1972-1975.

LESLIE, J. C., SHEPHARD, R. A., & TOAL, L. (1994). Chronic effects of chlordiazepoxide, valproate and antagonists on conditioned suppression of operant behaviour in rats. Manuscript submitted for publication.

MILLENSON, J. R., & LESLIE, J. C. (1974). The conditioned emotional response (CER) as a baseline for the study of antianxiety drugs. Neuropharmacology, 13, 1-9.

PETERSEN, E. N., PASCHELKE, G., KEHR, W., NIELSEN, M., & BRAESTRUP, C. (1982). Does the reversal of the anticonflict effect of phenobarbital by beta-carboline and FG7142 indicate benzodiazepine receptor-mediated anxiogenic properties? European Journal of Pharmacology, 82, 217-221.

POLC, P. (1981). Effects of the selective benzodiazepine antagonist Ro15-1788 on the cat spinal cord. Experientia, 37, 674.

POLC, P. (1988). Electrophysiology of benzodiazepine receptor ligands: Multiple mechanisms and sites of action. Progress in Neurobiology, 31, 349-423.

POSCHEL, B. P. N. (1971). A simple and specific screen for benzodiazepine-like drugs. Psychopharmacologia, 19, 193-198.

SCHOLFIELD, C. N. (1983). Ro15-1788 is a potent antagonist of benzodiazepines in the olfactory cortex slice in vitro. Pfluegers Archives, 396,292-296.

SHEPHARD, R. A., TOAL, L., & LESLIE, J. C. (1990). Effects of agonists and antagonists at the GABA/benzodiazepine receptor on conditioned suppression in rats. Pharmacology, Biochemistry and Behavior, 36, 39-43.

TENEN, S. (1967). Recovery time as a measure of CER strength: effect of benzodiazepines, amobarbital, chlorpromazine and amphetamine. Psychopharmacologia, 12, 1-7.

TOAL, L., LESLIE, J. C., & SHEPHARD, R. A. (1991). Effects of chlordiazepoxide and putative anxiogenics on conditioned suppression in rats. Physiology and Behavior, 49, 1085-1090.

JULIAN C. LESLIE, LIAM TOAL, AND ROBERT A. SHEPARD

University of Ulster

We thank Roche Products Ltd. for their generous gifts of flumazenil and Ro5-3663. Correspondence and reprint requests should be sent to J.C. Leslie, Department of Psychology, University of Ulster, Jordanstown, Northern Ireland BT37 0QB, U.K.

Copyright Psychological Record Spring 1995
Provided by ProQuest Information and Learning Company. All rights Reserved