Effects of chronic nicotine, nicotine withdrawal and ...

Psychopharmacology (2012) 219:453–468 DOI 10.1007/s00213-011-2558-z

ORIGINAL INVESTIGATION

Effects of chronic nicotine, nicotine withdrawal and subsequent nicotine challenges on behavioural inhibition in rats K. Z. Kolokotroni & R. J. Rodgers & A. A. Harrison

Received: 28 April 2011 / Accepted: 24 October 2011 / Published online: 29 November 2011 # Springer-Verlag 2011

Abstract Rationale Drug addiction is a chronically relapsing disorder characterised by compulsive drug use and loss of control over drug intake. Although several theories propose impulsivity as a key component of addiction, the precise nature of this relationship remains unclear. Objectives This study aims to investigate the short- and longer-term effects of chronic nicotine administration on behavioural inhibition. Methods Rats were trained on a symmetrically reinforced go/no-go task, following which they were subcutaneously prepared with osmotic minipumps delivering either nicotine (3.16 mg kg−1 day−1 (freebase)) or saline for 7 days. Performance was assessed daily during chronic treatment, in early and late abstinence, and in response to acute nicotine challenges following prolonged abstinence. Results Chronic nicotine resulted in a transient reduction in inhibitory control. Spontaneous withdrawal was associated with a nicotine abstinence syndrome, the early stages of which were characterised by a significant increase in inhibitory control. This was, however, short-lived with a decrease in inhibition observed in the second week of Electronic supplementary material The online version of this article (doi:10.1007/s00213-011-2558-z) contains supplementary material, which is available to authorized users. K. Z. Kolokotroni : R. J. Rodgers : A. A. Harrison Behavioural Neuroscience Laboratory, Institute of Psychological Sciences, University of Leeds, Leeds LS2 9JT, UK K. Z. Kolokotroni (*) Faculty of Health and Social Sciences, Leeds Metropolitan University, D420 Civic Quarter, Calverley Street, Leeds LS1 3HE, UK e-mail: [email protected]

abstinence. Whilst performance returned to baseline by the end of the third week, acute challenges (0.125, 0.25, 0.5 mg/kg, SC) revealed that nicotine exposure had sensitised animals to the disinhibitory effects of the compound. Conclusions Drug-induced loss of inhibitory control may be critically involved both in the initial and later stages of addiction. Neuroadaptations occurring during chronic exposure to and/or withdrawal from nicotine render animals more sensitive to the disinhibitory effects of the drug. Longer-term changes in behaviour may play an important role in the increased susceptibility to relapse in those with a history of nicotine abuse. Keywords Impulsivity . Drug abuse . Behavioural inhibition . Go/no-go task . Nicotine . Withdrawal . Sensitisation . Rats

Introduction Impulsivity is a multidimensional concept including a failure of inhibitory control and a preference for immediate over delayed gratification (Evenden 1999; Reynolds et al. 2006). The DSM IV-TR defines substance abuse disorder in terms of dysfunctional impulse control, comprising an inability to refrain from inappropriate drug-seeking/taking behaviour, as well as a preference for small immediate rewards over larger delayed rewards. Consistent with this perspective, a growing literature indicates that addicts exhibit heightened impulsivity on measures of behavioural disinhibition (as assessed by Go/No-go and stop tasks) and sensitivity to delayed reward (as assessed by delay discounting paradigms). Such effects have been reported for chronic cocaine users (e.g. Kirby and Petry 2004), opiate addicts (e.g. Madden et al. 1997) methamphetamine users (Monterosso et

454

al. 2005), alcoholics (e.g. Vuchinich and Simpson 1998) and cigarette smokers (e.g. Bickel et al. 1999; Spinella 2002). Despite such robust differences, however, the precise role/s of varying subcomponents of impulsivity in the initiation, maintenance and relapse of addiction has yet to be fully determined. Recent theories of addiction argue that impulsivity may be a stable trait contributing to the establishment and maintenance of drug addiction and/or that drug use itself may directly impair inhibitory control/impulsive decision making leading to continued drug use. Support for impulsivity as a risk factor includes reports that childhood impairments in inhibitory control significantly predict substance abuse during late adolescence and early 20s (Aytaclar et al. 1999; Tarter et al. 2007; see Verdejo-Garcia et al. 2008 for review) and that baseline sensitivity to delayed reward and behavioural disinhibition in rodents are predictive of later self-administration of cocaine, ethanol and nicotine (e.g. Dalley et al. 2007; Diergaarde et al. 2008; Oberlin and Grahame 2009; Poulos and Parker 1995). The view that heightened impulsivity may be a consequence of drug use is supported by a wealth of studies showing reductions in inhibitory control in Go/No-go, stop and five-choice serial reaction time tasks (5CSRTT) following acute treatment with cocaine, D-amphetamine, morphine, alcohol and nicotine (e.g. Blondel et al. 2000; Fillmore and Vogel-Sprott 1999; Fillmore et al. 2002; Kolokotroni et al. 2011; Pattij et al. 2009; Van Gaalen et al. 2006). These findings however appear to be dependent on a number of variables, including dose level, behavioural task and baseline impulsivity. Comparably, the same compounds administered acutely have also been shown to increase impulsive choice (e.g. Cardinal et al. 2000; Evenden and Ryan 1996; Kolokotroni et al. 2011; Reynolds and Schiffbauer 2004; see Perry and Carroll 2008 for review). Against this background, comparatively fewer animal studies have investigated the effects of chronic administration of drugs of abuse on performance in behavioural tasks of impulsivity. This is surprising in view of the fact that chronic treatment mirrors more precisely the pattern of drug use in drug-dependent individuals and helps to ascertain the level of drug exposure necessary to induce the neuroadaptations associated with heightened impulsivity. In one such study, Richards et al. (1999) demonstrated an increase in impulsive choice in rats following 14 days of methamphetamine treatment (however, see Stanis et al. 2008). Similarly, chronic administration of cocaine has been found to decrease tolerance to delayed rewards (e.g. Anker et al. 2009; Paine et al. 2003), an effect that has been shown to persist for up to 3 months following treatment (Simon et al. 2007). Less consistent are the effects of cocaine on tasks of behavioural disinhibition, with both increases in impulsive responding (Jentsch et al. 2002; Winstanley et al. 2009) and

Psychopharmacology (2012) 219:453–468

no changes from baseline being reported (Paine et al. 2003). Less well studied are the chronic effects of nicotine on impulsive behaviour. Although without initial effect, 7-day nicotine administration was found to increase anticipatory responding in a 5CSRTT (Blondel et al. 1999). However, caution is warranted in that a concomitant decrease in response latency renders it difficult to dissociate nicotine’s effects on inhibitory control from its effects on locomotor sensitisation (e.g. Clarke and Kumar 1983). On measures of impulsive choice, animals treated with nicotine for 65 days displayed an increase in impulsivity that continued for a month following treatment (Dallery and Locey 2005). In order to better understand the role played by impulsivity in drug maintenance and relapse, studies have also assessed such behaviour during drug withdrawal. In animals, termination of cocaine, amphetamine and heroin self-administration had no apparent effect on anticipatory responding in the 5CSRTT (Dalley et al. 2005a, b; however; see Winstanley et al. 2009), whilst repeated withdrawal from cocaine in high trait impulsive animals was associated with a normalisation of premature impulsive responding (Dalley et al. 2007). These data are inconsistent with the continued dysfunction in inhibitory control observed following longer-term abstinence in humans. Thus, 2 weeks following termination of drug use, both abstinent alcoholics and cocaine users display greater deficits in performance of the go/no-go task in comparison to control subjects (e.g. Noël et al. 2007; Verdejo-Garcia et al. 2007). In contrast, increases in impulsive choice have been reported in the early stages of drug withdrawal in both smokers and opioid-dependent patients (Field et al. 2006; Giordano et al. 2002; Mitchell 2004). The less consistent research findings for later phases of abstinence (e.g. Bickel et al. 1999; Heil et al. 2006) may indicate that the heightened impulsive choice observed both during drug use and drug withdrawal is transient in nature. Much of the human research in this area has employed cross-sectional designs; therefore, it is not clear whether heightened impulsivity is a reversible effect of chronic drug use, or whether low impulsive individuals find it easier to remain abstinent. The latter possibility is suggested by reports that more impulsive individuals displaying greater levels of delay discounting are at greater risk of future relapse (e.g. Krishnan-Sarin et al. 2007; Yoon et al. 2007). Prospective within-subject research designs that assess impulsivity prior to the initiation of drug treatment, during chronic drug exposure, and during initial and long-term withdrawal can be used to address these important issues. In view of these considerations and given the limited research exploring the chronic effects of nicotine on impulsivity, the present series of experiments employed a symmetrically reinforced Go/No-go conditional visual discrimination task to assess the effects on inhibitory


control in the same animals of: (1) chronic nicotine administration, (2) early and late nicotine abstinence and (3) residual sensitivity to acute nicotine challenge following sustained abstinence in attempt to understand the relationship between chronic drug use and impulsive behaviour.

Methods Subjects Subjects were 19 adult male Lister hooded rats (Charles River, UK). On arrival in the laboratory, animals were housed in pairs (46×26.5×26 cm) and maintained under a 12-h light/dark cycle (lights on at 0700 hour) in a temperature (21±4°C) and humidity (50±10%) controlled environment. At the start of testing, animals weighed 360– 380 g. A food deprivation schedule of 18.6 g/day (inclusive of food consumed during testing) kept animals at 85% of their free-feeding bodyweight throughout the experiment. Water was available ad libitum in home cages and feeding occurred at the end of each day. All procedures were conducted under Home Office licence in accordance with the UK Animals (Scientific Procedures) Act 1986. Drugs (−)–Nicotine hydrogen tartrate salt ((−)-1-methyl-2-(3-pyridyl) pyrrolidine (+)-bitartrate salt), obtained from Sigma-Aldrich (Poole, UK), was dissolved in 0.9% saline and pH adjusted to approximately seven using 0.1 M sodium hydroxide. Nicotine was chronically administered through subcutaneously implanted osmotic minipumps (Model 2ML1; Alzet, Charles River, UK) with drug concentration calculated such that animals received 3.16 mg kg−1 day−1 (rate of release, 10 μl/h) for a 7-day period. Chronic administration of nicotine via osmotic minipumps (rather than by injection) mimics more effectively the nicotine intake of smokers (e.g. McMorrow and Foxx 1983) by ensuring the maintenance of constant plasma levels throughout the study. This daily dose maintains plasma levels of nicotine at approximately ∼44 ng ml−1—a level which is similar to smokers who consume 30 cigarettes/day (Benowitz 1988; Murrin et al. 1987). Acute nicotine challenges were administered subcutaneously (SC) in a volume of 1 ml/kg at challenge doses of 0, 0.125, 0.25 and 0.5 mg/kg, 10 min prior to testing. All doses were calculated as free base and freshly prepared on each test day. Apparatus Subjects performed the behavioural task in four operant test chambers (dimensions 30.5×24.1×21 cm; Med Associates

455

Inc., USA). Each aluminium chamber was enclosed within a soundproof wooden box fitted with a ventilation fan. Chambers were illuminated by a 2.8-W stimulus house light and equipped with two retractable levers on the front wall. Above each lever was a 2.8 W stimulus light and between the two levers there was a food magazine to which 45-mg sucrose pellets (Noyes pellets, Sandown Scientific, UK) were delivered from a pellet dispenser. The magazine was illuminated by a white LED and head entries were detected by an infrared photobeam. The apparatus was controlled by Med-PC software running on a Pentium 3 Processor. During the observation of somatic withdrawal signs, animals were placed in a glass observation tank (dimensions 40.5×37×11 cm) the floor of which was lightly covered with wood shavings. Behaviours were recorded by a video camera positioned horizontally in front of the observation tank. The video signal was relayed to a monitor and VCR in an adjacent laboratory. Procedure This symmetrically reinforced Go/No-go conditional visual discrimination task was based on that described by Harrison et al. (1999). Unlike asymmetrically reinforced go/no-go tasks, both discriminative stimuli in the present paradigm represent the availability of reward, and animals are trained to learn when to, and when not to respond in order to gain a reward. Symmetrical reinforcement thus enables the impact of drug manipulations on behavioural disinhibition and reduced aversiveness to stimuli associated with non-reward to be disentangled. Initially all animals were magazinetrained by permitting access to several sucrose pellets placed in an illuminated magazine. This phase was followed by continuous reinforcement training. Only one lever was presented during these training sessions, with the position (right or left) counterbalanced across rats. A lever press resulted in the illumination of the magazine light which remained on until the animal entered the magazine to receive a single sucrose pellet. This training continued until rats earned more than 50 pellets in 30 min on two consecutive sessions: this usually required no more than five training sessions. The go/no-go task consisted of 40 Go trials and 40 Nogo trials presented in random order. Animals were required to discriminate between two visual stimuli comprising fast (0.1 s pulses presented at 5 Hz) and slow (0.4 s pulses presented at 0.83 Hz) synchronised flashings of the stimulus lights. The stimulus-response contingencies were counterbalanced such that, for half the animals, fast flashing lights indicated a Go trial while slow flashing lights indicated a No-go trial while, for the remainder, the converse applied. Each trial began with illumination of the house light and the initiation of a 5-s inter-trial interval

456

(ITI), following which the discriminative visual stimuli were presented for 10 s. During the first 1.2 s of stimulus presentation (the pre-discrimination period), any lever press was recorded as an early response but had no consequence as this time period allowed the presentation of one complete cycle of the slow stimulus and therefore identification of the stimulus frequency prior to this time point was unlikely. As go trials require an active lever response, early lever responses were expected to be considerably greater during these trials in comparison to No-go trials. Presentation of the stimuli continued for a maximum period of 10 s or until either a response on the lever occurred, or animals entered the food magazine. If animals entered the magazine during the 10-s stimulus presentation, a 5-s time-out period followed and an inappropriate magazine entry was recorded. During the time-out period, the house light was extinguished and any further responses had no consequences. After the time-out period, the same trial was restarted. If a lever response occurred during the 10-s stimulus presentation period, one of two outcomes followed. During a go trial, a lever press resulted in the stimulus lights being turned off and the magazine being illuminated to signal the availability of a food reward. This response was recorded as a correct go trial. The animal then had a period of 5 s to enter the food magazine which would in turn switch off the magazine light and deliver a single sucrose pellet. However, during no-go trials, a lever press resulted in the termination of the flashing stimulus lights, followed by a 5-s time-out period of darkness. In this case, an incorrect no-go trial was recorded. If no lever response occurred during the 10-s stimulus presentation then, again, then one of two outcomes followed. If no response occurred during a No-go trial, at the end of the stimulus presentation the magazine was illuminated and animals had a 5-s period in which to enter the magazine and receive a food reward. A correct no-go trial was recorded. However, if animals failed to make a lever response during a go trial, a 5-s time-out period of darkness followed and the trial was recorded as an incorrect go trial. Response latencies were recorded for both trials and were measured from the end of the pre-discrimination period of the stimulus presentation until the lever was pressed. Response latencies were recorded as correct and incorrect during Go and No-go trials, respectively. Both correct and incorrect response latencies were used to determine the ITI prior to the following trial. This was achieved by subtracting the response latency from the total possible duration of the stimulus had no lever response occurred (i.e. 10 s). This duration was then added to the 5 s ITI period. If no response occurred, the ITI duration was simply 5 s. Therefore, irrespective of the type of trial that had preceded and how the animals had responded, the duration between trials always remained constant. This


technique was implemented to minimise the likelihood of a Go response bias developing due to an increase in the rate of reinforcement delivery. Magazine entry latencies were calculated from the time of the correct response (i.e. a lever press during Go trials or following the stimulus presentation during No-go trials) until the animal’s entry into the food magazine. If animals failed to enter the magazine within 5 s, this was recorded as a magazine omission and no rewards were delivered. Training continued until animals reached a criterion of 85% total correct trials on two consecutive sessions. Animals typically reached this level of performance following 8 weeks of training. Once this level of accuracy had been achieved, sessions continued for one week to ensure that performance accuracy had stabilised prior to testing. Once trained each session took approximately 40 min to complete. All dependent variables recorded in the go/no-go task are summarised below. The main index of disinhibition in the paradigm was the inability to withhold responding during No-go trials. Impulsive responding was therefore indicated by reductions in the percentage of correctly completed No-go trials. Enhancements of anticipatory responses (both early responses and inappropriate magazine entries) were also indicative of increased impulsive responding, and these measures were utilised to provide further support of an impulsive behavioural profile. Accuracy of responding & & &

Total percentage correct trials (no. of correct Go trials+ correct No-go trials/80*100) Percentage correct Go trials (no. of correct go trials/ 40*100) Percentage correct No-go trials (no. of correct no-go trials/40*100) Anticipatory responding

& & & &

No. No. No. No.

of Go trials with early responses of No-go trials with early responses of Go trials with inappropriate magazine entries of No-go trials with inappropriate magazine entries

Speed of responding & & & &

Correct response latency during Go trials (s) Incorrect response latency during No-go trials (s) Magazine latency following correct Go trials (s) Magazine latency following correct No-go trials (s) Omissions

& &

No. of magazine omissions following correct Go trials No. of magazine omissions following correct No-go trials

A series of three experiments was conducted which sequentially examined the performance of the same animals


in the go/no-go task during chronic nicotine treatment (Experiment 1), nicotine withdrawal (Experiment 2) and in response to acute nicotine challenges following a period of sustained abstinence (Experiment 3). Experiment 1: effect of chronic nicotine treatment on performance of the symmetrically reinforced go/no-go conditional visual discrimination task A mixed design was employed with treatment group (chronic saline (n=9) or chronic nicotine (n=10)) as the between-subject factor and test day as the within-subject factor. Once trained, animals were assigned to a treatment group according to their baseline accuracy (total percentage of correct trials). Osmotic minipumps were surgically implanted SC under isoflurane/oxygen anaesthesia and the incision closed with 9-mm steel wound clips (Vet-Tech, UK). Behavioural testing commenced 18 h following implantation and continued daily for a 7-day period. All operant testing took place during the light phase of the LD cycle (0800–1830 hours). Experiment 2: effect of short- and long-term spontaneous nicotine withdrawal on performance of the symmetrically reinforced go/no-go conditional visual discrimination task Spontaneous nicotine withdrawal was initiated on the seventh day of drug treatment. The minipumps were surgically removed under isoflurane/oxygen anaesthesia, and the incision closed with Vetbond surgical adhesive (Vet-Tech, UK). Testing on the go/no-go task began 12 h later and continued for a 3-week period. In order to identify any early behavioural effects of drug withdrawal, animals were tested twice daily for the first two days following pump removal with 6 hours between each test session i.e. at 12-, 18-, 36- and 42-h post-removal. Data from previous validation studies demonstrated that running animals twice per day on the task had little or no effect on performance (Kolokotroni, unpublished findings). During the remainder of the 3-week period, operant testing was conducted once daily. To assess the severity and time course of spontaneous nicotine withdrawal, the frequency of somatic withdrawal signs was recorded over 7 days following pump removal. In animals, spontaneous nicotine deprivation induces observable physical symptoms (gasps, writhes, body shakes, head shakes, chews, teeth chattering, cheek tremors, paw tremors, genital grooming, foot licks, yawns, ptosis and scratches) that can be quantified to determine withdrawal intensity (e.g. Malin et al. 1992). Prior to any assessment, animals were habituated both to the test laboratory and observation chamber on two consecutive days. Somatic signs were observed the day before pump implantation (baseline assessment), the final day of chronic drug infusion, and during the first week following the termination of drug treatment. On the basis of previous findings (Malin et al. 1992; Epping-Jordan et al. 1998; Harrison et al. 2001),

457

observations were made at 6, 12.5, 18.5, 24, 36.5, 42.5, 60.5, 84.5, 108.5, 132.5, 156.5 and 180.5 h following pump removal. These observations were normally made after operant testing and under normal laboratory illumination during the light phase of the LD cycle. The only exception was during the early onset of withdrawal, when observations were also made under low intensity red light during the dark phase of the LD cycle. For testing, animals were individually placed in a glass observation chamber for 10 min and the frequency of behaviours recorded by an observer seated at a distance of approximately 1.5 m; sessions were also video recorded. Between animals, the observation tank was thoroughly cleaned and fresh wood shavings provided. Experiment 3: effect of acute nicotine challenge on the performance of the symmetrically reinforced go/no-go conditional visual discrimination task following a period of nicotine abstinence Three weeks after pump removal, all animals had returned to pre-drug baseline levels of performance accuracy with additional checks confirming the equivalence of the two original treatment groups. Acute nicotine challenges were assessed using a mixed design, with chronic drug treatment group (saline or nicotine) as the between-subject factor and dose (0, 0.125, 0.25 and 0.5 mg/kg, SC) as the within-subject factor. Treatments were administered according to a Latin square design with a minimum of 72 h separating consecutive treatments. The performance accuracy of all animals had returned to baseline levels prior to subsequent drug challenges. All operant testing took place during the light phase of the LD cycle (0830–1730 hour). Statistical analysis All data were initially evaluated for normality (Shapiro– Wilks test) and homogeneity of variance (between-subject designs, Levene’s test; within-subject designs, Mauchly’s test of sphericity). Data that violated these assumptions were subjected to appropriate transformations: arcsine transformations were used for proportional data (e.g. percentage correct trials) while all other datasets were subjected to square root, log10 or inverse transformations. Comparisons of performance on the task by each treatment group were analysed across each stage of the study, i.e. baseline, chronic drug administration, and withdrawal weeks 1, 2 and 3. Two-way mixed analysis of variance (ANOVA; treatment group×test session) were used to assess group differences in performance (accuracy, anticipatory responses and response latencies) during baseline. All significant main effects were further analysed by Bonferroni tests while significant interactions were followed by Bonferroni-adjusted simple effects analyses. As there was both within and between-group variability in

458

baseline performance on all test parameters, and to control for the possibility that drug effects may have been influenced by differences in pre-drug baseline, all data from chronic drug treatment and drug withdrawal stages were subjected to a two-way mixed ANCOVA (e.g. Tabachnick and Fidell 2007). At each of the treatment stages, behavioural parameters were assessed with the average baseline performance over seven days acting as the covariate within the model. Homogeneity of regression was initially examined through interactions of the covariate with both the within and between-subject variables. On the very rare occasions when homogeneity of regression was violated (i.e. variables interacted significantly with the covariate), data were re-expressed as a percentage change from baseline and analysed by a two-way mixed ANOVA (Tabachnick and Fidell 2007). Significant main effects of treatment and test day were further explored by Bonferroni tests, while significant interactions were examined by Bonferroni-adjusted simple effects analysis where individual means were adjusted to control for the effects of the covariate. All somatic signs from the nicotine abstinence syndrome were aggregated and analysed by a two-way mixed ANOVA (treatment group×observation session). Significant main effects were explored by Bonferroni tests while Bonferroni-adjusted simple effects analysis was applied to further examine significant interactions. The effects of acute nicotine challenges were analysed by two-way mixed ANOVAs (chronic drug treatment group×acute challenge dose). All significant main effects were further explored by Bonferroni tests. In the case of significant interactions, Bonferroni-adjusted simple effects analysis examined the response of each chronic treatment group across acute drug challenges, and also compared performance across groups at each challenge dose. In view of their rarity throughout the study, data for omissions were not analysed across all experiments. In all cases, α values of p