Tolerance To The Benzodiazepine Diazepam In An Animal Model Of Anxiolytic Activity

Psychopharmacology

Psychopharmacology (1985) 87: 322-327

9 Springer-Verlag 1985

Tolerance to the benzodiazepine diazepam in an animal model of anxiolytic activity D.N. Stephens and H.H. Schneider Research Laboratories of Schering AG, t000 Berlin 65 and Bergkamen, FRG Abstract. The antipunishment properties of diazepam (DZP) were investigated in mice treated acutely, or following nine daily treatments with either DZP (5 mg/kg, PO) or its vehicle. Acutely, or following chronic vehicle treatment, DZP produced a dose-related increase in activity punished by footshock. Following chronic DZP, test doses of DZP given 24 or 48 h following the last chronic treatment were no longer, or less effective in enhancing punished activity. Effects on unpunished activity were unaffected. In a study of the time course of tolerance development, tolerance was not seen after one or three daily treatments but was present after 6 days. Following establishment of tolerance by 9 days' treatment, the antipunishment activity of DZP reappeared after 8 days' withdrawal and was restored to acute levels after 16 days. Tolerance was not associated with changes in benzodiazepine (BZ) receptor affinity or numbers, but the ability of GABA to enhance BZ binding was increased. There was no change in the ability of DZP or the convulsant fl-carboline D M C M to modulate 35STBPS binding. The mechanism of tolerance to the antipunishment properties of DZP therefore remains unknown. Key words: Anxiety - Tolerance - Benzodiazepine - Diazepare - Passive avoidance - Punishment - Benzodiazepine binding - Mice

Although the development of tolerance to the anticonvulsant and sedative effects of the benzodiazepines (BZ) is well documented both in the clinic (e.g. Greenblatt et al. 1979; Browne and Penry 1973; Lader 1980) and in animal studies (e.g. File 1981 ; Sepinwall et al. 1978; Margules and Stein 1968; Frey et at. 1984), it is widely held that the anxiolytic activity of the BZs is persistent (Sepinwall and Cook 1980; Greenblatt and Shader 1978; Overstreet and Yamamura 1979). Although this view has sometimes been questioned (Lader 1980; File 1984, 1985), great weight has been attached to findings that in animal conflict models an antipunishment action of BZs is revealed or enhanced as their sedative properties undergo tolerance (Sepinwall et al. 1978; Margules and Stein 1968; Sepinwall and Cook 1980). As File (1985) has pointed out, however, the empirical evidence to support the belief that tolerance does not develop is rather slight, since the duration of the reported experiOffprint requests to: D.N. Stephens, D Nettropsychopharmacology, Schering AG, Postfach 650311, 1000 Berlin 65

ments was only 5-7 days and it is possible that with longer treatment, tolerance to the antipunishment effects of BZs would also have been observed. Perhaps of equal importance, it is impossible to interpret these conflict experiments in terms of anxiolytic tolerance, since the measure of anxiolytic activity, lever pressing, was influenced not only by the antipunishrnent properties of BZs but also by their sedative properties. Since following the first doses the sedative properties overwhelmingly predominate, it is not possible to obtain a true measure of the antipunishment effect. When after several doses the sedative effects have themselves undergone tolerance it is possible to observe the antipunishment activity of BZs revealed as an increase in the rate of lever pressing, unconfounded by sedation. However, it is totally inappropriate to compare this rate with the rate of lever pressing during the first application, since it cannot be known what the antipunishment effect would have been in the absence of sedation. Using a less conventional animal model of anxiolytic action, File and her colleagues (Vellucci and File 1979) have demonstrated tolerance to the ability of BZs to enhance social interaction between rats placed in a mildly stressful novel test situation. In the present experiments we have investigated the development of tolerance to the antipunishment properties of the BZ diazepam in a model of anxiety utilising a passive avoidance test in mice (Boissier et al. 1978; Stephens and Kehr 1985), which allows estimation of a drug's antipunishment activity on its first administration. We have also carried out studies on the duration and extent of BZ receptor occupation, as well as some other measures of BZ receptor function, i.e. GABA-stimulated BZ binding and the properties of the functionally related TBPS binding sites after tolerance development.

Methods Animals

N M R I mice (either male or female) weighing about 20 g were obtained from the Department of Tierzucht und -haltung, Schering AG, Berlin. Treatment

Diazepam (DZP) 5 mg (a gift from Hoffmann La-Roche, Basle, Switzerland) was given orally each day for 9 consecu-

323 rive days (defined here as "chronic" treatment). Control groups received an equivalent volume (10 ml/kg) of vehicle (10% Cremophor EL in saline). Behavioural experiments were carried out at either 24 or 48 h or at the time points specified following the last chronic treatment. Parallel groups of mice were treated identically and concurrently for estimation of biochemical parameters. For the assessment of the antipunishment action of DZP, the mice were given DZP or vehicle PO 1 h before testing. In all cases, independent groups of mice (all n = 8) were tested at each dose and time point.

Behavioural experiments Assessment of the antipunishment activity of DZP was carried out using the four-plate test (Boissier et al. 1968) with minor modifications (Stephens and Kehr 1985). Briefly, individual mice, which were naive to the test, were allowed to explore an apparatus consisting of a perspex box, 23 x 18 x 30 cm high, the floor of which consisted of four metal plates. Following 20 s habituation, mice in the punished groups (n = 8) received a mild (1 mA), brief (60 ms) footshock each time they crossed from one plate to another. The total number of crossings in a 1-min period was counted. Mice in independent unpunished groups were similarly treated, except that they received no footshock.

Biochemical experiments In vitro binding." Stimulation of 3H-lormetazepam (3HLMZ) binding by GABA. Individual forebrains were homogenized 1:20 (weight:volume) in 5 0 m M tris citrate buffer, pH 7.1 using an Ultra Turrax homogeniser. Following centrifugation for 15 rain at 30,000 g the pellets were washed twice by the same procedure. Duplicate samples (200 gl) of the 1 : 100 membrane suspension in Krebs phosphate buffer, pH 7.1, were added to 250 gl buffer or GABA solution and 50 ~tl of tracer (3H-lormetazepam, 56.6 Ci/ mmol, Schering AG; final concentration 0.88 nM). Nonspecific binding was measured in the presence of 1 gM clonazepam. Following 30 rain incubation at 23 ~ C the samples were diluted with 2 ml cold buffer and bound radioactivity was separated from free by filtration through Whatman

~

Punishedcrossings

GF/B filters, followed by three washes with 2 ml cold buffer (Filter prep, Ismatec, Zfirich). Radioactivity on the filters was determined by liquid scintillation counting.

3ss-t-butylbicyclophosphorothionate

(3SS-TBPS) binding.

This was estimated after Nielsen et al. (1985), using repeatedly frozen and washed membrane preparations. Following three freezing and washing cycles of an aliquot of the first pellet described above with 50 mM tris citrate, pH 7.1, the final pellet was resuspended 1 : 100 in 50 m M tris citrate, 1 M NaC1, pH 7.1. The membrane suspension (500 111)was incubated in duplicate with 25 gl ethanol, or compound solution in ethanol for 60 min at 23 ~ C. Nonspecific binding was measured in the presence of 10 gM picrotoxinin. Tracer (3ss-t-butylbicyclophosphothionate, 84.6 Ci/mmol, NEN, Boston; 75 gl; final concentration 0.17 nM) was added, and after a further 20 min incubation, membrane bound radioactivity was determined as above.

3H-LMZ binding in vivo. Twenty minutes before sacrifice 200 gCi 3H-LMZ/5 ml/kg was administered IV. Following decapitation, forebrain was removed as fast as possible ( < 1 min), weighed, and homogenized in 20 ml ice-cold 25 mM sodium phosphate buffer, pH 7.4. Three 2 ml aliquots were immediately pipetted onto Whatman GF/B filters and washed three times with 3 ml cold buffer. Filters were transferred into counting vials and shaken with 10 ml scintillation fluid for at least 30 min before determination of radioactivity. Nonspecific binding in vivo was determined in an additional group receiving 50 mg lorazepam/ kg, PO, 30 min prior to sacrifice. Results

Antipunishment activity. Mice which had received Cremophor EL (CEL) vehicle for 9 days did not differ from mice which had received no pretratment in their sensitivity to the antipunishment effects of DZP (Fig. 1) and in these groups DZP exhibited a dose-dependent ability to increase punished activity [F(3,28)=16.24; P<0.001]. Analysis of variance also showed significant effects of chronic DZP on the antipunishment activity of DZP test doses [F(3,91)=

Unpunished crossings

V Acute

9 48 h followingchronicvehicle % 48 h following chronic DZP (5 mg/kg p.o. for 9 days)

]f.

20

;.--;/%.

1816-

/ / ,/ ./ / ~ ./ / /

1412-

\\'I

10 8 6 4-

i,-/

S

}i,) 7

2

0.63 2.'5 10 40 Diazepamtest dose (mg/kg, p.o.)

0.63 215 1'0 413 Diazepamtest dose (mg/kg, p.o.)

Fig. 1. The effect of test doses of DZP on unpunished activity and activity suppressed by footshock (1 mA, 60 ms) in mice treated acutely or 48 h following 9 days of chronic treatment with DZP (5mg/kg/day, PO) or its vehicle. Chronic vehicle and acute groups did not differ from each other, but the effects of test doses of DZP were significantly less than in either control group and did not differ from the effect of vehicle (see text)

324 The effect of 10 mg/kg Diazepam (p.o.) or vehicle on punished and unDunished activity when given acutely or following 9 days chronic treatment (5 mg/kg p.o.)

Punished crossings test treatment J vehicle

2V Number of crossings . . . . . unpunished, DZP . . . . . unpunished, vehicle punished, DZP -~ punished, vehicle

i 10 m g / k g diazepam

* different from vehicle / p < 05 t different from acute ] " I

I

i

}

- .........

....................

.....................

!ii i;i i~

iii;iiiii Ac;te ; +

;

4

t6

24 Acu~te

Days following last chronic DZP treatment .x+:,:!

iiiiiiiiiii

Fig. 4, The effects of DZP (10 mg/kg PO) or vehicle on punished and unpunished activity in female mice when given acutely, and when given t, 2, 4, 8, 16 or 24 days following 9 days of chronic DZP treatment. The "Acute" groups were mice age matched for the 1 and 24 day withdrawal periods, respectively + different from corresponding "Acute" group, P < 0.05 (Dunnett's test) * different from corresponding vehicle group, P < 0.05

ilili i i!i i ":W:2 ":.Y;2

24 h

48 h

24 h

48

Time since last chronic application [ ] chronic vehicle

[] chronic diazepam

Fig. 2. The effect of a single test dose (10 mg/kg) of DZP or of vehicle on locomotor activity suppressed by punishment at two time periods (24 and 48 h) following the last chronic DZP or vehicle treatment. * significantly different from chronic vehicle group, P < 0.05 (Dunnett's test)

[•

9 chronic DZP o chronic vehicle

Punished crossings

Acute

t Day

3 Days

6 Days

9

3 H - L M Z b o u n d ( % of control)

200-

9 Days

2018 16 1412

t~

150-

64 ( 2 0.6 25 10

0

0.6 2.5 10

0

0.6 2,5 10

0

0.6 2,5 10

0

0.6 2,5 10

Diazepam Test Dose (mg/kg, p.o.)

Fig. 3. The effect of t, 3, 6 or 9 daily DZP (5 mg/kg, PO) doses on the ability of test doses of DZP given 48 h after the last "chronic" dose of the schedule to reinstate locomotor activity suppressed by punishment. * DZP-pretreated group differs from vehicle-pretreated group, P < 0.05 (Scheffe's test)

100-

/l --t ~-.-.-------- T . 3.10 -7

10-6

. T

.

3.10 `6

r.

I

10 -5

3.10 -5

1 - 10 -4

[M]

GABA

6.5; P < 0.001]. In particular, the ability of D Z P to increase punished activity was no longer present following 9 days of chronic treatment ( P > 0 . 0 5 ; Scheffe's test). The ability of low doses of D Z P to increase unpunished activity was slightly but non-significantly diminished by chronic treatment (Fig. 1). Similar, though less marked tolerance was found after 9 days of D Z P using a 24-h withdrawal period between chronic dosing and testing, but at this time period, 10 and 40 m g / k g D Z P PO exhibited some residual antipunishment activity (data not shown). The apparent increase in tolerance 48 h following withdrawal relative to that at 24 h was confirmed in a separate experiment in which the antipunishment of 10 mg/kg D Z P PO was assessed, in separate groups,

Fig. 5. GABA stimulation of BZ binding to mouse forebrain membranes 48 h following nine daily treatments with DZP, 5 mg/kg PO, ( e - - e ) or vehicle ( o - - o ) . Shown are the means (n= 5) • expressed as percentages of control levels, which were 678 + 32 and 6584-36 fmol SH-LMZ/mg protein for the chronic vehicle and chronic DZP groups, respectively * significantly different from chronic vehicle group, P < 0.05 (t-test) both 24 and 48 h following 9 days chronic treatment (Fig. 2). In an investigation of the time course of tolerance development, the antipunishment activity of D Z P was assessed 48 h following 1,3, 6 or 9 daily administrations of D Z P (5 mg/kg PO). N o tolerance was evident following one or

325 Table 1. Characteristic parameters (4-SEM) of 35S-TBPS binding

to mouse forebrain membranes 48 h following nine daily treatments with DZP or vehicle (n = 5)

35S-TBPS binding Specifically bound at 0.17 nM (fmol/mg protein) In the presence ofDZP, t gM (% of control binding) In the presence of DMCM, I gM (% of control binding)

Chronic vehicle

Chronic DZP

20.3 + t .2

21.7 _+3.7

119_+3

125 _+6

67_+5

63 _+7

three doses but was significant after six daily treatments (Fig. 3). Figure 4 shows the time course of recovery of the antipunishment activity of 10 mg/kg DZP following 9 days oral treatment. In this experiment, no difference was apparent between the 24 h and 48 h withdrawal periods and DZP was no more effective than vehicle in reinstating activity suppressed by punishment until 4 days following withdrawal, when a weak antipunishment effect became evident. Full antipunishment potency returned within 16 days.

Biochemistry. In vivo receptor binding studies at 5, 24 and 48 h following a single dose of 5 mg/kg DZP PO gave a half-life for BZ receptor occupation of approximately 11 h in forebrain. Binding of 3H-lormetazepam was decreased to 78_+8% (mean_+ SEM) at 24 h and to 96___6% of control at 48 h following DZP (all n = 5), indicating significant remaining receptor occupation by DZP 24 [t(8)=3.26; P < 0.01] but not 48 h following a single dose. Assays carried out following 9 days of treatment revealed binding o f 3 H LMZ to be 98_+3% of control values at 48 h following the last treatment, indicating no residual receptor occupation by DZP, and 92_+4% of control at 24 h (all n = 6). These values in the chronically-treated animals do not appear different from those in the acutely-treated mice, though because the estimations were carried out in different experiments, a direct comparison is not strictly valid. To control for induction of increased DZP metabolism, we investigated 3H-LMZ binding in vivo 1 h following a 5 mg/kg dose of DZP PO to parallel groups of chronicallytreated mice. There was no change in receptor occupation at that time (9 days vehicle: 25_+4% of control; 9 days DZP: 20_+6%, all n=6). Specific BZ binding was not changed 48 h following 9 days of DZP treatment, as revealed by a single 3H-LMZ concentration (0.88 nM); see legend to Fig. 5. In contrast, the stimulation of BZ binding in the presence of GABA was enhanced significantly (Fig. 5), but with an unchanged ECso (controls: 1.0 gM; DZP group: 0.8 gM). 3SS-TBPS binding was unaffected, both with regard to basal and to stimulated (DZP, I ~tM) and inhibited (DMCM, 6,7-dimethoxy-fl-carboline-3-carboxylic acid methyl ester, 1 [aM) binding (Table 1). Discussion

The present findings indicate clearly that the antipunishment potency of the benzodiazepine DZP is markedly re-

duced following a relatively short period of repeated daily administration. While a single dose had no effect, even six daily administrations of a moderate dose led to a marked attenuation of the antipunishment effect of BZs. This observation, which was repeated in three separate experiments, stands in contradiction to the widely held view that in animal studies tolerance does not develop to the anxiolytic effects of the BZ. The present findings, however, are in close agreement with the observations of File and her colleagues (Vellucci and File 1979; File 1985) using a nonpunishment paradigm (the social interaction test in rats). Taken together, these observations provide animal experimental support for the view of the Committee for the Review of Medicines (CRM; 1980) that there is little evidence for continued efficacy of BZs as anxiolytics after 4 months of administration. Indeed, the present experiments suggest that DZP may retain its anxiolytic efficacy for a much shorter period. Whether the more recently developed BZs with shorter half-lives or other classes of BZ receptor ligand also lose their anxiolytic efficacy cannot be predicted from these experiments. The reason for tolerance to the antipunishment effects of benzodiazepines having lain hidden for more than 20 years is not entirely clear, but one possible reason is the undue weight placed upon observations in the Geller type of conflict test. In this test it is well known that clinicallyeffective anxiolytic benzodiazepines produce significant anticonflict effects only after several doses (Margules and Stein 1968). During the first doses the anticonflict action is masked by a sedative effect. Tolerance develops rapidly to the sedation, unmasking the anxiolytic action. The inappropriate conclusion has been that tolerance has developed to the sedative but not to the anxiolytic activity of BZs. In the four-plate test, "anxiolytic" doses of DZP are below those which decrease unpunished exploratory activity, so the problem of masking of the anxiolytic effect does not arise. For this reason, testing can be carried out in drugand test-naive animals, allowing a comparison of the effects of repeated treatments with those of first-time administration. In this case a clear tolerance effect is seen. A rather curious phenomenon observed in the present experiment was the reduction in the efficacy of D Z P 48 compared with 24 h after the last chronic treatment. This effect was observed in two out of three independent experiments. One possible explanation could be the residual receptor occupation we saw in some animals 24 h following the last DZP dose, though it would be surprising if such residual receptor occupation contributed meaningfully to total receptor occupation following a 40 mg/kg test dose. Although we were able to demonstrate tolerance to the antipunishment properties of DZP, its basis is not yet clear. As with other BZ effects, the antipunishment potency of BZ ligands in the four-plate test is highly correlated with their in vivo affinity for central BZ receptors (Stephens et al. 1984), and the reduced potency observed here could relate to reduced receptor occupation either as a result of reduced affinity or as a result of the more rapid metabolism of DZP. However, in common with most (e.g. Braestrup and Nielsen 1983; Gallager et al. 1984a) but not all (e.g. Rosenberg et al. 1982) previous authors, we found no change in the ability of DZP to displace a labelled BZ (3H-LMZ) from its receptors. This result would seem to rule out both dispositionaI tolerance and changes in receptor affinity or numbers.

326 Recently, the phenomenon of learned tolerance has received growing attention (Demellweek and Goudie 1984; Siegel and MacRae 1984). Important as this mechanism may be in other circumstances, it is difficult to see how a learning mechanism might affect the efficacy of DZP in the four-plate procedure, in which the animals were exposed to the test apparatus only once and for a short time. Learned tolerance can therefore probably be excluded as an explanation of the present observations. BZs are thought to achieve their pharmacological effects by facilitating the action of GABA (Costa and Guidotti 1979; Haefely etal. 1975). The aUosteric interaction between BZ and GABA receptors is reflected in the ability of GABA to enhance the affinity of BZs for their receptor (Braestrup et al. 1982; Martin and Candy 1978; Tallman et al. 1978). A plausible mechanism for tolerance could be a reduced influence of BZs on GABAergic transmission, perhaps through the uncoupling of the BZ receptor from the GABA receptor and its associated chloride channel. In support of this possibility, Gallager et al. (1984a) have reported that following chronic BZ administration the ability of GABA to enhance BZ binding is diminished. We were unable to confirm this observation and indeed, in our tolerant mice the ability of GABA to increase BZ binding was increased. These results were confirmed in a repeat experiment (data not shown). Apart from the species difference, our experiments differed from those of Gallager et al. in the time between the last BZ treatment and sacrificing the animals for biochemical analysis. However, this seems an unlikely explanation for the difference in our respective observations. Our own observations strongly suggest that tolerance to the anxiolytic activity of DZP is not due to an uncoupling of the BZ/GABA allosteric complex, and this interpretation is also favoured by our observations on the binding of 35S-TBPS. This ligand is thought to bind to the chloride channel protein associated with the GABA/ BZ receptor complex (Squires et al. 1983) and its binding is enhanced in the presence of BZ receptor agonists and decreased in the presence of inverse agonists (Supavilai and Karobath 1983). Like the GABA ratio, the ability of BZ receptor ligands to enhance TBPS binding is a good predictor of their activity in animal models of anxiety (Stephens et al. 1984; Stephens and Kehr 1985). If the mechanism of tolerance is a change in the ability of BZs to alter chloride channel function, then this should be reflected in a change in the ability of BZ ligands to modulate TBPS binding. In the present experiment in mice, the ability of DZP to enhance and of the inverse agonist/?-carboline DMCM to decrease TBPS binding was unaffected by chronic treatment. Perhaps more important than the GABA receptor/BZ binding site interaction may be the state of the GABA receptor/chloride channel coupling following chronic BZ. Acute BZ treatment decreases GABA turnover (Bernasconi et al. 1982), and it might be supposed that chronically-reduced GABA release would induce an increase in GABA receptor function. The increased number of low affinity GABA binding sites (perhaps the functional form) observed following chronic BZ treatment (Gallager etal. 1984b; Rago et al. 1983) would be consistent with such an interpretation. Whether such an increased number of GABA binding sites would result in an increased ability of GABA to modulate BZ binding as we observed is a matter for further investigation.

The previous paucity o f evidence on the development of tolerance to the anxiolytic properties of BZs, in contrast to the wealth of information on tolerance to some of their other actions, has been a weakness in the argument that the various properties of the BZs are all attributable to a common mechanism, the facilitation of GABAergic transmission. The present study indicates that the anxiolytic properties of BZs, like their sedative, hypnotic and anticonvulsant properties, do undergo tolerance, and that the development of such tolerance is associated with a change in GABAergic transmission, albeit not that predicted by the simplest model of BZ/GABA receptor/chloride channel interaction. Other possible mechanisms of tolerance, including changes in monoamine turnover (Wise et al. 1972; Lister and File 1983), remain to be investigated in our model.

Acknowledgements. We thank Christel Schneider, Ronald Weid~ mann, Monika Mfiller-Seewald, Simone Fritz and J6rg Seidler for skilled assistance and Tage Honor6 for kind advice on the TBPS binding method.

References Bernasconi R, Maitre L, Martin P, Raschdorf F (1982) The use of inhibitors of GABA-transaminase for the determination of GABA turnover in mouse brain regions: an evaluation of aminooxyacetic acid and gabaculine. J Neurochem 38 : 57-66 Boissier JR, Simon P, Aron C (1968) A new method for the rapid screening of minor tranquillisers in mice. Eur J Pharmacol 4:145-150 Braestrup C, Nielsen M (t 983) Benzodiazepine receptors. In: Iversen LL, Iversen SD, Snyder SH (eds) Handbook of Psychopharmacology Vol. 17 Plenum, New York, pp 285-384 Braestrup C, Schmiechen R, Neef G, Nielsen M, Petersen EN (1982) Interaction of convulsive ligands with benzodiazepine receptors. Science 216:1241-1243 Browne TR, Penry JK (1973) Benzodiazepines in the treatment of epilepsy: a review. Epilepsia 14:277-310 Committee for the Review of Medicines (1980) Benzodiazepines. Drug Ther Bull 18:97-98 Costa E, Guidotti A (1979) Molecular mechanisms in the receptor action of benzodiazepines. Ann Rev Pharmacol Toxicol 19:531-535 Demellweek C, Goudie AJ (1983) Behavioural tolerance to amphetamine and other psychostimulants : the case for considering behavioural mechanism. Psychopharmacology 80:287-307 File SE (1981) Rapid development of tolerance to the sedative effects of lorazepam and triazolam in rats. Psychopharmacology 73 :240-245 File SE (1984) Behavioural pharmacology of benzodiazepines. Prog. Neuropsychopharmacol. Biol Psychiatry 8:19-31 File SE (1985) Tolerance to the behavioural actions of benzodiazepines. Neurosci Biobehav Rev 9 : (in press) Frey HH, Philippin HP, Scheuler W (1984) Development of tolerance to the anticonvulsant effect of diazepam in dogs. Eur J Pharmacol 104:27-38 Gallager DW, Lakoski JM, Gonsalves SG, Rauch SL (1984a) Chronic benzodiazepine treatment decreases postsynaptic GABA sensitivity. Nature 308:74-77 Gallager DW, Rauch SL, Malcolm LK (t984b) Alterations in a low affinity GABA recognition site following chronic benzodiazepines. Eur J Pharmacol 98:159-160 Greenblatt DJ, Shader RI (1978) Dependence, tolerance, and addiction to benzodiazepines: clinical and pharmacokinetic considerations. Drug Metab Rev 8:13-28 Greenblatt D J, Shader RI, Harmatz JS, Georgotas A (1979) Selfrated sedation and plasma concentrations of desmethyldiazepam following single doses of chlorazepate. Psychopharmacology 66 : 289~90

327 Haefely W, Kulcsar A, M6hler H, Pieri L, Polc P, Schaffner R (1975) Possible involvement of GABA in the central actions of benzodiazepines. In: E. Costa, P. Greengard (eds) Mechanism of action of benzodiazepines Raven Press, New York pp 131-151 Lader MH (1980) The present status of benzodiazepines in psychiatry and medicine. Arzneimittelforsch 30:851-916 Lister RG, File SE (1983) Changes in regional concentrations of 5-HT and 5-HIAA during development of tolerance to the sedative action of chlordiazepoxide. J Pharm Phamacol 35 : 601-603 Margules DL, Stein L (1968) Increase of antianxiety activity and tolerance to behavioral depression during chronic administration of oxazepam. Pschopharmacology 13:74-80 Martin IL, Candy JM (1979) Facilitation of benzodiazepine binding by sodium chloride and GABA. Neuropharmacology 17 : 993-998 Nielsen M, Honore T, Braestrup C (1985) Radiation inactivation of brain 35S-TBPS binding sites reveals complicated molecular arrangements of the GABA/benzodiazepine receptor chloride channel complex. Biochem Pharmacol (in press) Overstreet DH, Yamamura HI (1979) Receptor alterations and drug tolerance. Life Sci 25:1865-1878 Rago LK, Sarv KA, Allikmets LK (1983) Effect of a ten-day course of fenibut and diazepam on GABA and benzodiazepine receptors in mouse brain. Bull Exp Biol Med 96:1708-1709 Rosenberg HC, Smith S, Chiu TH (1982) Benzodiazepine-specific and nonspecific tolerance following chronic flurazepam treatment Life Sci 32:279-285 Sepinwall J, Cook L (1980) Mechanism of action of the benzodiazepines: behavioral aspects. Fed Proc 39 : 3024-3031

Sepinwall J, Grodsky FS, Cook L (1978) Conflict behavior in the squirrel monkey: effects of chlordiazepoxide, diazepam and Ndesmethyldiazepam. J Pharmacol Exp Ther 204:88-102 Siegel S, MacRae J (1984) Environmental specificity of tolerance. Trends Neurosci 7:140-144 Squires RF, Casida JE, Richardson M, Saederup E (1983) [35S]-& butylbicyclophosphorothionate binds with high affinity to brain-specific sites coupled to 7-aminobutyric acid A and ion recognition sites. Mol Pharmacol 23 : 326-336 Stephens DN, Kehr W (1985) fl-carbolines can enhance or antagonize the effect of punishment in mice. Psychopharmacology 85:143-147 Stephens DN, Kehr W, Schneider HH, Braestrup C (1984) Bidirectional effects on anxiety of fl-carbolines acting as benzodiazepine receptor ligands. Neuropharmacology 23 : 879-880 Supavilai P, Karobath M (1983) Differential modulation of [35S]TBPS binding by the occupation of benzodiazepine receptors with its ligands. Eur J Pharmacol 91 : 145-146 Tallman JF, Thomas JW, Gallager DW (1978) GABAergic modulation of benzodiazepine binding site sensitivity. Nature 274:383-385 Vellucci SV, File SE (1979) Chlordiazepoxide loses its anxiolytic action with long-term treatment. Pschopharmacology 62:61 65 Wise CD, Berger BD, Stein L (1972) Benzodiazepines: anxiety reducing activity by reduction of serotonin turnover in the brain. Science 177:180-183

Received March 23, 1985; Final version June 12, 1985