Previous experience of withdrawal from chronic diazepam ameliorates the aversiveness of precipitated withdrawal and reduces withdrawal-induced c-fos expression in nucleus accumbens Sarah J. Dunworth, Andy N. Mead* and David N. Stephens Laboratory of Experimental Psychology, University of Sussex, Brighton BN1 9QG, UK Keywords: conditioned place aversion, conditioned taste aversion, repeated withdrawal, seizure sensitivity Abstract Flumazenil (20mg/kg, i.p.)-precipitated withdrawal from chronic treatment with diazepam (DZP, 15mg/kg, s.c. in sesame oil for 21days)resultedinadecreasedseizurethresholdtotheconvulsant,pentylenetetrazole(PTZ),infusedintothetailvein;withdrawal from 21-day chronic diazepam treatment, interspersed with two periods of drug withdrawal, resulted in a greater decrease in convulsant threshold. A separate experiment showed that consumption of a sucrose solution immediately prior to precipitated withdrawal resulted in a decreased subsequent consumption of the sucrose solution; no such evidence of a conditioned taste aversion(CTA)wasseeninmicegivenpriorexperienceofwithdrawal.Thus,priorexperienceofwithdrawalenhancedtheeffectsofa subsequent precipitated withdrawal in increasing seizure sensitivity, but weakened the ability of this withdrawal to serve as an aversiveunconditionedstimulus(US).Theweakeningoftheaversivepropertiesofprecipitatedwithdrawalmayre¯ecthabituationto the withdrawal stimulus, and was accompanied by a loss of the ability of withdrawal to induce c-fos expression in the shell of the nucleus accumbens, an area sensitive to both novel, and stressful, as well as rewarding stimuli. Introduction Chronic treatment with benzodiazepines may result in the develop- ment of dependence, revealed as withdrawal signs when drug treatment is terminated (Woods etal., 1992). If animals are given repeated episodes of chronic benzodiazepine treatment, with inter- mittent drug withdrawal, the severity of the withdrawal syndrome alters. In common with repeated withdrawal from chronic ethanol (Kokka etal., 1993; Becker, 1994), repeated withdrawal from diazepam (DZP) induces an increased sensitivity to seizures (Ward & Stephens, 1998; Dunworth & Stephens, 1998). In contrast, certain behavioural signs of withdrawal show reduced intensity following previous withdrawal experience. Thus, mice which had undergone repeated withdrawal showed reduced anxiogenic effects of with- drawal compared with mice undergoing their ®rst withdrawal experience (Ward & Stephens, 1998). A reduction in the intensity of behavioural signs of benzodiazepine withdrawal has been observed previously in several species, following repeated precipitated with- drawal induced by administration of the antagonist ¯umazenil (Lamb & Grif®ths, 1985; Gallager etal., 1986; File & Baldwin, 1987; Baldwin & File, 1988; Sannerud etal., 1989). Why should repeated experience of withdrawal lead to increased severity of certain withdrawal signs (seizures), but reduced intensity of others? One obvious possibility is that the neural systems underlying convulsant sensitivity and anxiety differ, and undergo different adaptations as a result of withdrawal. An alternative possibility is that experience of withdrawal results in a behavioural adaptation to withdrawal-induced anxiety, which might even mask any underlying change in neuronal sensitization. The ability of aversive events to protect against the behavioural consequences of future aversive experience is well established (Jackson etal., 1978; Solomon & Stanford, 1992). In the context of anxiety, exposure of animals to operant partial punishment schedules results in increased resistance to punishment (Miller, 1960; Brown & Wagner, 1964), while previous exposure of rats and mice to the plus-maze results in a decrease in anxious behaviour in the apparatus (e.g. Takulis & McCay, 1992). The reduction in the severity of anxiety-related behaviour, following repeated experience of withdrawal, may re¯ect this attenuation of the aversiveness of the withdrawal experience. We tested the idea that the aversiveness of withdrawal is lessened following prior experience using a conditioned taste aversion (CTA, Riley & Tuck, 1985) paradigm to assess the strength of conditioning to the aversiveness of withdrawal. In our previous work (Ward & Stephens, 1998; Dunworth & Stephens, 1998), we have tested the mice 72 h following withdrawal of diazepam treatment, at a time when brain levels of drug and metabolites are low (Steppuhn etal., 1993). In such experiments, the aversive withdrawal state is likely to be drawn out over a lengthy, but indeterminate time period. In order to facilitate the formation of an association of the withdrawal state with a novel taste, the ®nal withdrawal was achieved using the benzodiazepine receptor antagonist ¯umazenil to precipitate with- drawal; the intermittent withdrawal periods during chronic treatment were obtained by interrupting drug administration, as previously. In order to con®rm that this regimen resulted in an increase of Correspondence: Dr D. N. Stephens, as above. E-mail: dns@biols.susx.ac.uk *Present address: Internal Research Program, NIDA, Baltimore, MD 21222, USA. Received 3 November 1999, revised 18 January 2000, accepted 19 January 2000 European Journal of Neuroscience, Vol. 12, pp. 1501±1508, 2000 ã European Neuroscience Association