sign in sign up
<<
Home , Dopamine

Psychopharmacology DOI 10.1007/s00213-007-0995-5

ORIGINAL INVESTIGATION

Effects of acute / depletion on the selective processing of -related cues and the relative value of in smokers

Brian Hitsman & James MacKillop & Anne Lingford-Hughes & Tim M. Williams & Faheem Ahmad & Sally Adams & David J. Nutt & Marcus R. Munafò

Received: 16 May 2007 /Accepted: 18 October 2007 # Springer-Verlag 2007

Abstract tive mood, and expired (CO) levels were Rationale Acute tyrosine/phenylalanine depletion (ATPD) is measured at various timepoints through 300 min. Participants a validated neurobiological challenge that results in reduced smoked at hourly intervals to prevent acute with- , allowing examination of drawal during testing. the effects of a hypodopaminergic state on craving-related Results The TYR/PHE-free mixture, as compared to the processes. BAL mixture, was associated with a greater increase in CO Objectives We studied 16 nonabstaining smokers (>10 levels from baseline ( p=0.01). Adjusting for the potential cigarettes/day; 9 males; age 20–33 years) to whom was confounding influence of between-condition differences in administered a tyrosine/phenylalanine-free mixture (TYR/ CO levels across , TYR/PHE-free mixture was asso- PHE-free) and a balanced amino mixture (BAL) in a ciated with increased demand for cigarettes ( p=0.01) and double-blind, counterbalanced, crossover design. decreased attentional bias toward smoking-related words Methods Subjective craving, attentional bias to ( p=0.003). There were no significant differences between smoking-related word cues, relative value of cigarettes, nega- conditions in either subjective craving or depressed or anxious mood ( p values>0.05). Conclusion Among nonabstaining daily smokers, acute dopaminergic depletion via ATPD may influence smoking B. Hitsman (*) behavior and indices of smoking-related , such Centers for Behavioral and Preventive Medicine, as attentional bias to smoking cues and relative cigarette The Miriam Hospital and Brown Medical School, Coro Building, Suite 500, 1 Hoppin Street, value, which are not readily captured by subjective craving. Providence, RI 02903, USA e-mail: [email protected] Keywords Cigarette smoking . Attentional bias . . . J. MacKillop Relative value Cigarette craving Center for and Studies, Brown University, Providence, RI, USA Introduction A. Lingford-Hughes : T. M. Williams : F. Ahmad : D. J. Nutt Academic Unit of Psychiatry, University of Bristol, Bristol, UK Converging evidence from animal and studies : indicates that cigarette smoking behavior is under partial S. Adams M. R. Munafò (*) influence of dopamine-dependent mesolimbic and meso- Department of Experimental Psychology, University of Bristol, cortical systems. Nicotine stimulates burst firing of 12a Priory Road, Bristol BS8 1TU, UK dopaminergic in animals when binding to nicotinic e-mail: [email protected] receptors in the tegmentum (Di

Chiara 2000), which in turn enhances release of dopamine degree of attentional bias and level of subjective cigarette in the outer shell of the and prefrontal craving. Consequently, cognitive constructs such as atten- cortex (Benwell and Balfour 1992; Corrigall et al. 1994). tional bias to smoking cues may index facets of reward and The cascade of events, including changes in postsynaptic motivational processes not necessarily entirely captured by signaling, is theorized to be positively reinforcing with self-report measures of craving (Perkins et al. 1997; Tiffany strong hedonic and appetitive motivational properties and Conklin 2000). A similar distinction has been observed (Epping-Jordan et al. 1998; Koob and Le Moal 1997). in studies of the relative value of addictive substances vs The synthesis of brain dopamine is dependent on the money. Relative value has been measured by eliciting self- availability of its precursor, tyrosine, in plasma. report of the desired level of consumption of a substance Administration of an amino acid mixture that lacks tyro- under multiple (and increasing) levels of price. In general, sine, as well as its precursor phenylalanine, decreases avail- the more a person is willing to spend to obtain a substance, ability of tyrosine and phenylalanine in plasma through such as a cigarette, the greater its relative value. These processes of increased protein synthesis. This results in studies have found moderate correlations between relative increased competition for transport across the blood–brain value and subjective craving, but far from collinear rela- barrier (Oldendorf and Szabo 1976; Pardridge 1977). In tionships (e.g., Mackillop et al. 2007). More objective animals, administration of the tyrosine/phenylalanine- cognitive and behavioral indicators of motivational drive to free (TYR/PHE-free) mixture decreases the availability smoke, such as attentional bias to smoking-related cues and of tyrosine in the brain, resulting in direct action on relative value of cigarettes, may be especially sensitive to dopamine (McTavish et al. 1999b). Noradrenergic acute changes in dopaminergic neurotransmission. response to either or is unaffected In the current study, we investigated the influence of (McTavish et al. 1999a), suggesting a selective effect on ATPD via administration of a TYR/PHE-free mixture, as dopamine. Administration of the TYR/PHE-free mixture compared with a BAL mixture, on motivation to smoke, in increases levels (Harmer et al. 2001) as indexed by attentional bias for smoking-related cues, and decreases extracellular dopamine levels, as indexed by relative value of cigarettes, and subjective cigarette changes in [11C] binding (Leyton et al. 2004; craving, and mood among non-nicotine deprived male Montgomery et al. 2003). and female smokers. Based on the role of dopamine in The acute tyrosine/phenylalanine depletion (ATPD) mediating nicotine’s rewarding and motivational effects, technique represents a valuable paradigm for examining we hypothesized that administration of the TYR/PHE-free the effects of reduced dopamine availability on addictive mixture would be associated with increased attentional behaviors. Despite the promise of this approach for iden- bias for smoking-related cues and increased value of tifying core biobehavioral processes underlying chronic cigarettes. Based in part on Casey and colleagues’ (2006) addiction, to our knowledge only one study has applied it to results, we did not hypothesize an effect on subjective . A recent study of 15 male smokers craving. We also did not predict effects on negative mood (Casey et al. 2006) found no effect of ATPD on self- as ATPD has been found to influence mood only in healthy reported craving or smoking topography. Subjective crav- female social drinkers and only after exposure to a ing may be distinct from other craving-related mechanisms psychological stress challenge (Leyton et al. 2000). The that contribute to relapse following acute smoking absti- TYR/PHE-free and BAL challenges were administered nence (Munafò and Hitsman 2006; Perkins et al. 1997). The 1 week apart using a double-blind, counterbalanced, incentive- model (Robinson and Berridge crossover design. 1993, 2000) asserts that repeated use of substances that enhance mesolimbic dopaminergic neurotransmission results in permanent or semipermanent changes in the Methods system. This process results in stimuli associated with substance use being assigned higher levels of incentive Participants (i.e., subjective “wanting” of the substance). One hypothesized consequence is an attentional bias towards Participants were 16 healthy volunteers who reported associated stimuli, which is not necessarily closely coupled smoking 10 to 25 cigarettes per day. Exclusion criteria with subjective craving. were the following: current intention to quit smoking, Numerous studies have employed the modified Stroop to defined by a score >5 on a 10-point scale of readiness to investigate attentional biases towards smoking-related cues quit (with 10=strong ); any current illicit drug use; in daily smokers (Munafò and Hitsman 2006). In most current or past psychiatric disorder (DSM-based), including cases, only a weak correlation has been observed between substance abuse or dependence other than nicotine depen- Psychopharmacology dence; history of neurological illness; significant current and sedation. The composition for males was iso- medical illness; pregnancy or lactation; current or recent leucine 15 g, leucine 22.5 g, lysine 17.5 g, methionine 5 g, (past 3 months) use of psychotropic (antide- valine 17.5 g, threonine 10 g, and 2.5 g. The pressants, anxiolytics, or neuroleptics) or herbal supple- BAL mixture for males additionally contained tyrosine ments known to affect mood (e.g., St. John’s Wort); current 12.5 g and phenylalanine 12.5 g. Females received 20% use of smoking cessation pharmacotherapy; or uncorrected less by weight of each amino acid than males. The amino impaired vision or color blindness. All participants gave were suspended in tap water, which was flavored with their written informed consent prior to screening. lemon and lime or cherry vanilla (participant’s preference) to disguise the unpalatable taste of the mixture. Eligibility screening To minimize any possibility of during testing, participants took monitored cigarette breaks Candidates who responded to recruitment flyers placed on at 60-, 120-, 180-, 240-, 300-, and 360-min intervals. An the University of Bristol precinct attended an initial session exhaled CO breath sample was taken 10 min after each for eligibility screening, which included a medical and cigarette. Subjective craving and mood ratings, along with psychiatric history and physical examination, electrocardio- a second blood sample for amino acid analysis, were gram, and routine blood testing. A urine specimen was taken again at 300 min. The Minnesota Withdrawal Scale- collected for illicit drug use screening and a breath test to Revised (MWS-R) (Hughes and Hatsukami 1986) was screen for recent alcohol consumption. Finally, assessments completed to verify that participants were not in withdraw- of nicotine dependence [Fagerstrom test for nicotine al. Finally, participants completed the reassessment of dependence (FTND)] (Heatherton et al. 1991) and depres- depressive symptoms and the assessments of relative sive symptoms [21-item Hamilton depression rating scale cigarette value and attentional bias to smoking-related cues. (HAM-D); Hamilton 1967] were completed. Eligible Upon completion of the 300-min assessments, they were participants then completed the two 7-h test sessions offered a balanced meal and then allowed to leave when conducted at the Bristol Royal Infirmary. The study was subjectively and objectively stable. The second session approved by the Bath Research Ethics Committee (UK) and was identical to the first, except that the alternate mixture the Institutional Review Board of The Miriam Hospital was given. Eight participants completed the BAL condi- (USA). Participants were reimbursed £60 plus travel tion first and eight completed the TYR/PHE-free condition expenses. first.

Experimental sessions Measures

Participants came to the laboratory at 08:00 hour on the The primary outcomes were attentional bias to smoking- morning of each test session after having followed a low- related cues and relative value of cigarettes in response to protein diet (total protein content less than 20 g) for the ATPD vs balanced amino acid challenge. The modified preceding 24 h and after having fasted from midnight. They Stroop task was identical to that used in prior research were instructed to smoke normally between awakening and (Mogg and Bradley 2002; Munafò et al. 2003)and arriving at the laboratory. To minimize the possibility of consisted of 12 practice trials followed by 12 blocks, each withdrawal during testing, participants who were consisting of 2 buffer and 12 experimental trials (i.e., 144 daily coffee or tea drinkers were instructed to have their experimental trials in total). The stimulus words were (1) 12 usual morning amount. Female participants were tested smoking-related words (cigarette, fags, lighter, matches, during the follicular phase of their menstrual cycle. For the inhale, ashtray, smoke, nicotine, cigar, , puff, filter) baseline assessment, participants provided a blood sample and (2) 12 categorized neutral words (blanket, garage, for amino acid analysis and an exhaled CO breath sample to shampoo, handle, sofa, curtain, switch, bathroom, vase, quantify recent smoking rate. They also completed assess- hallway, duster, lounge). Each block presented one set of ments of mood, craving, and depressive symptoms. stimulus words (smoking or neutral) under either masked or Immediately following the baseline assessment, partici- unmasked exposure conditions. The order of the 12 blocks pants consumed either the TYR/PHE-free or BAL amino was randomized across participants. Sets were matched for acid mixture in a randomized, double-blind, counter- length and frequency using published norms (Caroll et al. balanced manner. We used the 90-g TYR/PHE-free mixture 1971). Uncategorized neutral words (e.g., lucky, donkey, developed by McTavish and colleagues (McTavish et al. cultural, calliper) were used for the practice and buffer 1999b) to sufficiently challenge the dopaminergic system trials. An awareness check was given to participants yet minimize potential physical side effects, especially immediately after completing the Stroop to confirm the Psychopharmacology masked exposure condition. This was a lexical decision was used as an internal standard. The detection limit was task to identify the presence of a word or nonword and 1.3 pg/ml and inter- and intra-assay coefficients of varia- consisted of 12 practice and 24 test trials as described by tion were 13 and 8%, respectively. Munafo and colleagues (2003). Relative value of cigarettes vs money was assessed Statistical analysis using a cigarette purchase task (CPT) (Jacobs and Bickel 1999). The CPT is a self-report measure that is an analog We first examined the changes in plasma levels of tyrosine of a progressive-ratio operant task (Hodos 1961) using and phenylalanine, as well as in the ratio of tyrosine + self-reported consumption under various levels of price. phenylalanine/LNAAs, from baseline to 300 min during the Relative value is assessed by generating a cigarette de- TYR/PHE-free condition compared with the BAL condi- mand curve or a quantitative representation of economic tion. Data were analyzed using repeated measures 2×2×2 demand for cigarettes at multiple levels of price. In turn, multivariate analysis of variance (MANOVAs), with mix- the demand curve is used to characterize several facets of ture (TYR/PHE-free, BAL) and time (baseline, 300 min) as relative value. Prior studies have used the CPT as a trait- within-subjects factors and condition order (TYR/PHE-free type measure (i.e., typical smoking during a typical day) first, BAL first) as a between-subjects factor. Next, we (Jacobs and Bickel 1999; Mackillop and Murphy 2007; examined change in exhaled CO values across the two Murphy and MacKillop 2006), but we used it as a state conditions, as well as nicotine withdrawal scores at 300 min measure, assessing demand for cigarettes over a 3-h period. between conditions because any difference in smoking or Participants were assessed at the following 16 costs: zero withdrawal would confound the analyses of attentional bias (free), £0.01, £0.05, £0.13, £0.25, £0.50, £1, £3, £6, £11, and cigarette reward value. CO data were analyzed using £35, £70, £140, £280, £560, and £1,120 (Jacobs and Bickel repeated measures 3×2×2 MANOVA, with time (baseline, 1999; Murphy and MacKillop 2006). 180 min, 300 min) and mixture (TYR/PHE-free, BAL) as Secondary outcomes were subjective craving, mood, within-subjects factors and condition order (TYR/PHE-free and depressive symptoms as measured using the brief first, BAL first) as a between-subjects factor. The inter- questionnaire of smoking urges (QSU-Brief) (Cox et al. mediate (180 min) timepoint was chosen as the midpoint 2001), state-trait anxiety inventory—state subscale (STAI) between baseline and likely peak depletion (300 min). (Spielberger et al. 1983), profile of mood states (POMS) MWS-R data were analyzed using repeated measures 2×2 (McNair 1971), and visual analog scale (VAS) ratings of MANOVAwith mixture (TYR/PHE-free, BAL) as a within- the following symptoms: nausea, of fingers, diffi- subjects factor and condition order (TYR/PHE-free first, culty concentrating, restlessness, depression, anxiety, irrita- BAL first) as a between-subjects factor. bility, drowsiness, headache, nervousness, fatigue, , After these important experimental checks, we analyzed excess energy, and . the primary outcomes variables. Modified Stroop data were Depressive symptoms were measured using a modified analyzed using a repeated measures 2×2×2×2 MANOVA version of the Hamilton depression rating scale (M- of color-naming response with word type (smoking- HAMD). The M-HAMD included only those nine items related, control), exposure (masked, unmasked), and mix- that could change meaningfully over the course of an ture (TYR/PHE-free, BAL) as within-subjects factors and experimental session: low energy, depressed mood, feelings condition order (TYR/PHE-free first, BAL first) as a of guilt, decreased concentration and motor activity, between-subjects factor. CPT data were analyzed using anxious mood, decreased appetite, loss of , feelings methods similar to those of prior studies involving pur- of worthlessness, and loss of interest. chase tasks to generate demand curves (Jacobs and Bickel Plasma levels of tyrosine and phenylalanine were 1999; Murphy and MacKillop 2006), including the appli- measured at baseline and 300 min (approximate time of cation of nonlinear demand curve equation (Hursh et al. peak depletion) to confirm that the ATPD challenge pro- 1988). Five indices of relative value were derived: (1) cedure produced the predicted decline in levels. The ratio of intensity of demand (i.e., cigarette consumption at the tyrosine + phenylalanine to other large neutral amino acids lowest level of cost); (2) maximum cigarette expenditure

(LNAAs; valine, tryptophan, isoleucine, leucine) was also (Omax; i.e., maximum amount of money a participant was measured. Plasma was separated by centrifugation and willing to spend across various increments of price); (3) stored at −30°C. Total concentrations of tyrosine, phenyl- price at maximum consumption (Pmax; the price corre- , valine, tryptophan, isoleucine, and leucine were sponding to Omax, reflecting the point at which demand measured using automated high-performance liquid chro- becomes elastic); (4) elasticity of demand (sensitivity to matography with fluorescence end-point detection and escalating cost); and (5) breakpoint (i.e., first price at which precolumn sample derivatization adapted from methods consumption was fully suppressed). These indices reflect described by Furst and colleagues (1990). Norvaline different topographical facets of the demand curve. Data Psychopharmacology were analyzed using 2×2 repeated-measures MANOVA, in levels from baseline to 300 min in the TYR/PHE-free with mixture (TYR/PHE-free, BAL) as a within-subjects condition (tyrosine F [1, 10]=33.3, p<0.001; phenylalanine factor and condition order (TYR/PHE-free first, BAL first) F [1, 10]=30.8, p<0.001) and a significant increase in as a between-subjects factor. levels during the same time period in the BAL condition The secondary outcomes of mood (anxiety, total mood (tyrosine F [1, 10]=7.1, p=0.02; phenylalanine F [1, 10]= disturbance), craving, and depressive symptoms were 13.3, p=0.005). analyzed using repeated-measures 2×2×2 MANOVAs, The TYR/PHE-free mixture caused a 36.0% (SD=22.9%) with mixture (TYR/PHE-free, BAL) and time (baseline, reduction in tyrosine from baseline (M=34.2 nmol/ml, SD= 300 min) as within-subjects factors and condition order 9.2) to 300 min (M=22.6 nmol/ml, SD=11.3) and a 65.9% (TYR/PHE-free first, BAL first) as a between-subjects fac- (SD=18.0) reduction in phenylalanine from baseline (M= tor. We conducted correlational analyses of the plasma amino 43.6 nmol/ml, SD=22.3) to 300 min (M=14.7 nmol/ml, acid data to evaluate whether change in tyrosine or the ratio SD=10.2). In contrast, the BAL mixture caused a 118% of tyrosine + phenylalanine to LNAAs were associated with (SD=181.2%) increase in tyrosine from baseline (M= change on any dependent measure demonstrating a signifi- 28.5 nmol/ml, SD=8.5) to 300 min (M=54.2 nmol/ml, cant effect of ATPD. SD=29.4) and a 259.4% (SD=257.6%) increase in phenyl- All statistically significant interaction effects among alanine from baseline (M=43.0 nmol/ml, SD=19.6) to mixture condition and time were decomposed with simple 300 min (M=140.1 nmol/ml, SD=96.7). main effects analyses. Where no significant condition order Analysis of the tyrosine + phenylalanine/LNAAs ratio effect was evident, the order factor was removed from the revealed a significant mixture × time interaction (F [1, 10]= analysis. Where an order effect was present, it was retained 12.9, p=0.005). There was significant decline from baseline and the direction of the effect was probed using simple to 300 min in the TYR/PHE-free condition (F [1, 10]= interaction effects analyses stratifying on order. An alpha of 132.5, p<0.001), but not in the BAL condition (F [1, 10]= 0.05 was used for all outcomes. 1.6, p=0.23). The TYR/PHE-free mixture caused a 82.7% (SD=5.3) decrease from baseline (0.15, SD=0.04) to 300 min (0.02, SD=0.01). In the BAL condition, the ratio Results was 0.16 (SD=0.03) at baseline and 0.13 (SD=0.06) at 300 min, and represented only a 11.3% decline. Characteristics of participants Exhaled CO and acute nicotine withdrawal Two participants withdrew from the study due to nausea and following ingestion of the drink, so that the A paired samples t test revealed no differences in baseline final sample for analysis consisted of 14 participants. On CO between conditions (TYR/PHE-free, M=16.2, SD=6.5; average, participants (56% male) were 25 years old (SD= BAL, M=17.3, SD=7; p=0.35), indicating comparable 4 years; range 20–33 years) and had a FTND score of 4 rates of smoking prior to each session. The MANOVA (SD=2; range 2–7), indicating moderate nicotine depen- of CO data indicated a significant main effect of time dence. Their Hamilton depression score at baseline was 2 (F [2,12]=6.67, p=0.01), with increasing CO levels over (SD=2), reflecting euthymic mood. A series of independent time. This was qualified by a significant time × mixture samples t tests and χ2 tests indicated no difference between interaction (F [2,12]=4.63, p=0.03). The main effect of participants randomized to either condition order (n=7 mixture was not significant ( p=0.71). Post hoc tests TYR/PHE-free first, n=7 BAL first) on any measures indicated that the change in CO from baseline was greater ( p values>0.45). Two participants did not complete the in the TYR/PHE-free condition compared to the BAL CPT adequately, either providing nonnumeric data or condition at 180 min (p=0.01), although this difference omitting items. Plasma samples from three participants was no longer significant at 300 min ( p=0.24). These data for the amino acid analyses could not be assayed because are presented graphically in Fig. 1. The MANOVA of of blood processing errors. One participant had incomplete MWS-R scores at 300 min revealed no main effect of data on the QSU-Brief. mixture (F [1,12]=0.21, p=0.66). Scores were comparable between conditions and low overall (TYR/PHE-free, M= Plasma amino acid levels 7.4, SD=4.3; BAL, M=8.1, SD=6.0). Given the aforementioned difference between condi- Analysis of amino acid levels revealed significant mixture × tions in CO values across time, analyses of primary and time interactions for both tyrosine (F [1, 10]=12.1, p=0.01) secondary outcomes, except for plasma amino acids, and phenylalanine (F [1, 10]=19.1, p=0.001). Post hoc were adjusted for the between-condition difference in CO analyses of both amino acids revealed a significant decline change from baseline to 300 min using multivariate anal- Psychopharmacology

iate, indicated a significant main effect of exposure (F [1, 11]=13.7, p=0.003), with slower RTs for stimuli presented in the unmasked condition compared to the masked condition. The word type × mixture × order interaction was significant (F [1, 11]=14.3, p=0.003). No other main effects or interactions were significant (p values>0.11). Stratified analyses were conducted to explore the signifi- cant interaction involving the condition order term. Among participants who received the BAL mixture first (n=7), there was a trend for an effect of word type (F [1, 5]=4.2, p=0.10) and a significant effect of exposure (F [1, 56]= 13.7, p=0.02). The effect of word type was qualified by a significant word type × mixture interaction (F [1, 5]=18.3, p=0.01), which reflected greater slowing of RTs for smoking stimuli compared to control stimuli in the BAL conditioncomparedtotheTYR/PHE-freecondition.No Fig. 1 Mean change in exhaled CO levels from baseline by challenge other main effects or interactions were significant ( p condition. Errors bars represent SEM values>0.40). Among participants who received the TYR/ PHE-free mixture first, there were no significant main effects or interactions ( p values>0.13). These data are ysis of covariance (MANCOVA) instead of the planned presented graphically in Fig. 2. MANOVAs. A comparison of RT data from the first session only, with mixture (TYR/PHE-free/BAL) as a between-subjects Modified stroop task factor, indicated a significant main effect of word type (F [1, 11]=11.6, p=0.006). This was qualified by a Reaction time (RT) data were excluded from error trials. In marginal word type × mixture interaction (F [1, 5]=4.4, addition, to remove outliers, RTs less than 200 ms, more p=0.06). A similar analysis of data from the second than 2,000 ms, or more than two standard deviations above test session indicated no main effects or interactions the mean were excluded (Ratcliff 1993). The mean color- ( p values>0.65). During the TYR/PHE-free condition, the naming time data are presented in Table 1. difference in RTs for smoking stimuli compared to control The repeated-measures MANCOVA of color-naming stimuli was not significantly correlated with reduction in times, with the difference between conditions in change in either plasma tyrosine or the ratio of tyrosine + phenyl- exhaled CO from baseline to 300 min included as a covar- alanine to other LNAAs (p values>0.26).

Table 1 Means and standard deviations (ms) for reaction times to unmasked and masked smoking-related and control stimuli by challenge condition and condition order

Order Smoking-related Control

Unmasked Masked Combined Unmasked Masked Combined

TYR/PHE-free condition BAL first 577.89 572.80 575.35 591.17 556.65 573.91 (186.59) (103.87) (139.02) (160.43) (146.39) (152.30) TYR/PHE-free first 585.00 560.88 572.94 562.75 551.48 557.11 (90.53) (75.40) (82.59) (87.34) (77.06) (81.48) Combined 581.44 566.84 574.14 576.96 554.06 565.51 (140.95) (87.42) (109.86) (124.97) (112.42) (117.67) BAL condition BAL first 573.18 538.83 556.00 531.45 512.86 522.15 (120.74) (106.42) (112.03) (115.64) (113.70) (111.54) TYR/PHE-free first 551.66 538.68 545.17 567.66 534.01 550.84 (77.28) (65.25) (69.01) (88.04) (83.50) (84.20) Combined 562.42 538.75 550.59 549.56 523.43 536.50 (98.03) (84.81) (89.57) (100.51) (96.46) (96.11) Psychopharmacology

Fig. 2 Mean color-naming interference for smoking-related stimuli by challenge condition and condition order. Interference scores were calculated for each participant by subtracting the mean RTs for control stimuli from those for smoking-related stimuli. This index reflects the word type effect in the RT data, with positive values indicating an interference effect of smoking-related stimuli. Errors bars represent SEM

The percentage of awareness check trials with correct responses was calculated for all participants. The overall mean percentage correct was 52% in the BAL condition and 51% in the TYR/PHE-free condition, neither of which was significantly above chance performance of 50% (BAL, t[13]=1.32, p=0.21; TYR/PHE-free, t[13]=0.46, p=0.65).

Relative value of cigarettes

Performance on the CPT topographically conformed to expectations: demand for cigarettes exhibited a decelerating Fig. 3 Mean performance on the CPT by challenge condition. a curve in response to escalating price and expenditure Demand for cigarettes (self-reported consumption) over the session – exhibited the characteristic inverted U-shaped curve (see and in proportionate log log coordinates. Demand curves after both ’ challenges exhibit the prototypical deceleration in response to Fig. 3). Hursh and colleagues (1988) demand equation escalating cost, initially exhibiting inelastic demand and then provided a very good fit to the aggregated data for both the becoming elastic and decreasing to zero consumption. Zero values BAL and TYR/PHE-free conditions (R2=0.94 and R2= are recoded as trivial nonzero values to permit logarithmic transfor- 0.93, respectively). In addition, the equation provided a mation, but values are only shown to two decimal points. b Output for cigarettes (associated monetary expenditure) over the experimental very good fit to the individual participant data under period. The prototypical inverted U-shaped curve is evident, reflecting 2 both conditions (BAL, median R =0.94 [IQR=0.93– escalating expenditure to a peak followed by subsequent decreases, 0.99]; TYR/PHE-free: median R2 =0.93 [IQR=0.91– eventually to zero, is evident, but no effect on maximum expenditure 0.96]). Initial CPT analyses revealed no significant order (i.e., Omax). Zero values are recoded as trivial nonzero values to permit logarithmic transformation, but values are only shown to two decimal effects ( p values>0.25), so, as in all other analyses, the points. c The significant effect of the TYR/PHE-free challenge on order term was removed from the analytic model. Adjust- intensity of demand for cigarettes (i.e., initial demand at zero cost), ing for the difference between conditions in change in which is obscured by the logarithmic coordinates in a;**p<0.01. This exhaled CO from baseline to 300 min, the TYR/PHE-free analysis included depletion-related differences in exhaled CO as a covariate challenge was associated with a significant increase in intensity of demand for cigarettes (F [1, 10]=8.96, p= 0.01). There were no differences in breakpoint (F [1, 10]=

0.29, p=0.60), Omax (F [1, 10]=0.48, p=0.50), Pmax (F [1, Psychopharmacology

10]=1.18, p=0.30), or elasticity (F [1, 10]=1.46, p=0.26). increase from baseline to 300 min, with this effect greater Examination of expenditure on cigarettes suggested that, among participants who received the TYR/PHE-free mix- while the peak expenditure was not affected, the TYR/ ture second. No other symptoms showed changes, includ- PHE-free challenge resulted in a more persistent willing- ing those for depression and anxiety ( p values>0.06). ness to purchase cigarettes. Both session cigarette demand curves, including the difference in intensity of demand, are depicted in Fig. 3. Discussion During the TYR/PHE-free condition, there were no significant associations between the indices of relative We found that ATPD among nonabstaining male and value and reduction in plasma levels of either tyrosine or female smokers was associated with increased intensity of the tyrosine + phenylalanine/LNAA ratio ( p values>0.18), demand for cigarettes relative to money and reduced with the exception of a high-magnitude negative correlation attentional bias to smoking-related cues, although the latter between Pmax and percent change in tyrosine, r=−0.69, p< effect depended upon the order in which the challenge 0.05. This relationship indicated an inverse relationship conditions were administered. Despite standardizing the such that the greater the reduction in plasma tyrosine, the number of cigarettes smoked and achieving comparably higher the level of Pmax (i.e., price at which demand very low levels of nicotine withdrawal between conditions, becomes elastic), reflecting greater desire for cigarettes. we also found that ATPD was associated with increased intensity of smoking behavior as measured by exhaled CO, Subjective cigarette craving although this effect only constituted a trend when all inter- mediate timepoints between baseline and approximate time The repeated-measures MANCOVA of QSU-Brief scores of peak depletion were included in the analysis. The effects revealed that mixture × time interaction was statistically on smoking behavior, attentional bias to smoking-related nonsignificant (F [1, 11]=0.01, p=0.92). A significant cues, and cigarette reward value appeared in the absence of main effect of time (F [1, 11]=30.7, p<0.001), however, subjective craving. They also occurred independently of indicated that scores decreased from baseline to 300 min any changes in anxious or depressed mood or symptoms, as comparably across both conditions. Mean scores in the all were unaffected by ATPD. BAL condition were 31.0 (SD=11.0) at baseline and 17.1 Our results are consistent with preclinical and smoking (SD=7.8) at 300 min. Those in the TYR/PHE-free con- treatment studies evaluating acute and longer-duration dition were 34.2 (SD=11.2) at baseline and 19.9 (SD=10.1) manipulations of dopamine. Dopaminergic have at 300 min. The main effect of mixture was also non- been found to reduce smoking behavior, whereas antago- significant (F [1, 11]=1.6, p=0.23). Results for the two nists have the opposite effect (Caskey et al. 1999, 2002; subscales were identical. Dawe et al. 1995; George et al. 2003; Houtsmuller et al. 2002; Jarvik et al. 2000). Caskey and colleagues (2002) Depressive symptoms, subjective mood, and physical found that the dopaminergic reduced symptoms smoking behavior, as indicated by alterations in smoking topography, whereas the dopaminergic antagonist halo- No statistically significant main effects or interactions were peridol increased smoking. , which enhances observed for depressive symptoms (M-HAMD), anxiety dopaminergic functioning through inhibition of mono- (STAI-State), or total mood disturbance (POMS) (all p oxidase B, has been shown to exert a modest effect values>0.08). For the VAS ratings of psychological and on short-term smoking abstinence (George et al. 2003; physical symptoms, there were significant main effects of Houtsmuller et al. 2002). Finally, treatment involving time for difficulty concentrating (F [1,12]=11.0, p=0.006), , which acts as both a dopaminergic and norad- restlessness (F [1,12]=19.0, p=0.001), and headache (F renergic inhibitor, doubles long-term quit rates as [1,12]=7.1, p=0.02), indicating in all cases an increase compared with placebo (Mooney and Sofuoglu 2006). from baseline to 300 min in both conditions. For irritability, Unlike in the current study, the aforementioned pharma- there was a significant main effect of time (F [1,12]=5.3, cological studies typically have observed effects on sub- p=0.04), which was qualified by a time × mixture inter- jective craving (Brody et al. 2004; Caskey et al. 2002; action (F [1,12]=10.72, p=0.007), indicating an increase Houtsmuller et al. 2002). As in Casey and colleagues’ from baseline to 300 min in the TYR/PHE-free condition (2006) study of male smokers after overnight abstinence, only. For hunger, there was a significant main effect of time ATPD did not influence craving in our sample of non- (F [1,11]=5.9, p=0.03), which was qualified by time × abstaining smokers. This discrepancy with the literature mixture (F [1,11]=12.1, p=0.005) and time × mixture × may reflect differing mechanisms of action between ATPD order (F [1,11]=11.4, p=0.006) interactions, indicating an and agents used in the general pharmacological challenge Psychopharmacology and treatment studies. Whereas ATPD influences dopami- repeated administration from session 1 to session 2), which nergic function by decreasing the amount of presynaptic has been observed in a prior study (Munafò et al. 2003), dopamine available to postsynaptic receptors, most phar- plus attenuation caused by ATPD. This would serve to macological challenges studied either stimulate or block maximize the difference between challenge conditions in postsynaptic sites leading to more extreme changes the BAL-first group and decrease it in the TYR/PHE-free in dopaminergic function. They also have substantial sec- group. This interpretation is supported by our observation ondary influences on other systems. The that, at the first test session, there was a marginal effect of dissociation of smoking-related reward and motivational ATPD (reflecting a depletion effect), while at the second processes from conscious awareness is consistent with for- test session, there was no evidence of a modified Stroop mulations of incentive motivation (Berridge and Winkielman effect in either group (reflecting a practice effect). 2003; Uslaner et al. 2006) and compulsive drug use (Tiffany Despite observing no effects of ATPD on subjective and Carter 1998; Tiffany and Conklin 2000). craving, it was notable that ATPD also increased the rela- That ATPD-reduced attentional bias was also unexpected tive value of cigarettes. Five hours after ATPD, participants given the theory that attentional bias is a cognitive correlate reported wanting to smoke on average two more cigarettes of craving (Sayette et al. 2000). If accepted, a reduction in over the next 3 h than at the same time-point during the central dopamine levels, which may serve to model some of balance amino acid challenge. The pattern of expenditure the effects of acute nicotine abstinence, should increase reflected more persistent motivational drive, although this craving and, in parallel, increase attentional bias. Our was based on observation, not on a significant effect of results, in contrast, are consistent with a study by Munafò ATPD on the expenditure-related index of Omax. The CPT and colleagues (Munafò et al. 2007), which reported a results are consistent with studies reporting state-dependent reduction in attentional bias following ATPD among female effects on other behavioral economic indices of drug use smokers in acute withdrawal. Dopamine blockade using motivation (Field et al. 2006;Giordanoetal.2002; has been found to attenuate attentional bias to Mitchell 2004) and research indicating a dissociation heroin-related cues but not subjective craving (Franken between subjective craving and objective indices of moti- et al. 2004). In another study, combined dopaminergic and vation (e.g., Perkins et al. 1997). blockade using reduced smoking It is important to highlight two issues. First, the amino cue-elicited craving for cigarettes (Hutchison et al. 2004). acid mixture used in this study (90-g mixture developed by These findings are consistent with the hypothesis that McTavish et al. 1999b) lacked histidine, which is the dopaminergic neurotransmission may serve to signal the precursor of in the brain. expectation of reward (de la Fuente-Fernandez et al. 2002), Therefore, it is possible that the effects observed in our with conditioned cues stimulating dopaminergic release as a study are due to reduced synthesis of histamine in addition result of their association with addictive substances such as to dopamine. Second, participants in this study, although nicotine (Di Chiara et al. 2004). Thus, attentional bias for selected on the basis of cigarettes smoked per day (10 or smoking-related cues may reflect expectation of reward and more), were relatively young smokers. There is debate reduced dopaminergic neurotransmission via ATPD may regarding the processes that lead to attentional bias for attenuate this effect. Moreover, attentional bias may not be drug-related cues. Incentive salience theories regard such a simple cognitive correlate of subjective craving. It may be bias as a consequence of the ascription of incentive salience necessary to distinguish between phasic and tonic dopa- to initially neutral cues through repeated pairing with drug mine neurotransmission, with the former mediating cue (Franken et al. 2005) so that attentional bias reactivity and attentional bias to drug cues (Zhang and should increase with exposure. Reward probability theories Sulzer 2004), and the latter mediating withdrawal symp- (Fiorillo et al. 2003), in contrast, argue that dopamine toms associated with acute abstinence (Mansvelder et al. encodes prediction error (i.e., the discrepancy between 2003). A reduction in central dopamine levels may therefore predicted and actual reward), so that attentional bias should serve to reduce tonic dopamine levels (thereby elevating be greater when prediction error is greater (i.e., in in- withdrawal symptoms such as craving) while attenuating experienced drug users) compared with cues that have been phasic dopamine release (thereby reducing and learned to robustly predict drug reinforcement (i.e., in attentional bias). experienced users). Although our data do not allow us to One qualification of the modified Stroop effect is that it evaluate these two possibilities directly, use of ATPD in was observed only among the subset of participants who inexperienced vs experienced smokers might offer a means completed the BAL challenge first. The differential effect to test these competing theories on the basis that it should as a function of condition order (BAL first but not TYR/ attenuate dopamine-mediated effects. The issue is complex, PHE-free first) likely reflects a combination of an atten- however—the difference may instead reflect attention to uation of the modified Stroop effect with practice (i.e., conditioned cues vs the reward itself, while the difference Psychopharmacology between phasic and tonic dopamine release may be ventral and expectation of reward. Behav Brain Res relevant. It has also been suggested that dopamine may 136:359–363 Di Chiara G (2000) Role of dopamine in the behavioural actions of mediate both the rewarding and aversive effects of nicotine, nicotine related to addiction. Eur J Pharmacol 393:295–314 such that blockade of mesolimbic dopamine signaling Di Chiara G, Bassareo V, Fenu S, De Luca MA, Spina L, Cadoni C, may selectively reduce the aversive properties of nicotine Acquas E, Carboni E, Valentini V, Lecca D (2004) Dopamine and (Laviolette and van der Kooy 2003). These possibilities drug addiction: the nucleus accumbens shell connection. Neuro- 47(Suppl 1):227–241 will require detailed investigation to elucidate the exact Epping-Jordan MP, Watkins SS, Koob GF, Markou A (1998) Dramatic relationship between dopaminergic neurotransmission and decreases in brain reward function during nicotine withdrawal. cigarette-smoking behaviors. Nature 393:76–79 In conclusion, this study provides preliminary evidence Field M, Santarcangelo M, Sumnall H, Goudie A, Cole J (2006) Delay discounting and the behavioural economics of cigarette purchases that acute dopaminergic depletion (and possibly histidine in smokers: the effects of nicotine deprivation. Psychopharma- depletion) via ATPD may influence smoking behavior, as cology 186:255–263 well as cognitive and behavioral factors reflecting moti- Fiorillo CD, Tobler PN, Schultz W (2003) Discrete coding of reward vational drive to smoke in nonabstaining daily smokers. probability and uncertainty by dopamine neurons. Science 299:1898–1902 These effects were not captured by self-reported craving as Franken IH, Hendriks VM, Stam CJ, Van den Brink W (2004) A role measured using the QSU-Brief, and they occurred in the for dopamine in the processing of drug cues in heroin dependent absence of acute changes in negative mood. patients. Eur Neuropsychopharmacol 14:503–508 Franken IH, Booij J, van den Brink W (2005) The role of dopamine in Acknowledgements This study was funded by a National Institute human addiction: from reward to motivated attention. Eur J – on Drug Abuse Mentored Clinical Scientist Development Award (K08 Pharmacol 526:199 206 DA017145) to Brian Hitsman. We are grateful to Karin Mogg and Furst P, Pollack L, Graser TA, Godel H, Stehle P (1990) Appraisal of Brendan Bradley for supplying the code for the original modified four pre-column derivatization methods for the high-performance Stroop task used in this study. liquid chromatographic determination of free amino acids in biological materials. J Chromatogr 499:557–569 George TP, Vessicchio JC, Termine A, Jatlow PI, Kosten TR, O’Malley SS (2003) A preliminary placebo-controlled trial of References selegiline hydrochloride for smoking cessation. Biol Psychiatry 53:136–143 Giordano LA, Bickel WK, Loewenstein G, Jacobs EA, Marsch L, Benwell ME, Balfour DJ (1992) The effects of acute and repeated Badger GJ (2002) Mild deprivation increases the degree nicotine treatment on nucleus accumbens dopamine and locomotor that opioid-dependent outpatients discount delayed heroin and activity. Br J Pharmacol 105:849–856 money. Psychopharmacology 163:174–182 Berridge KC, Winkielman P (2003) What is an unconscious : Hamilton M (1967) Development of a rating scale for primary the case for unconscious ‘liking’. Cogn Emot 17:181–211 depressive illness. Br J Soc Clin Psychol 6:278–296 Brody AL, Mandelkern MA, Lee G, Smith E, Sadeghi M, Saxena S, Harmer CJ, McTavish SF, Clark L, Goodwin GM, Cowen PJ (2001) Jarvik ME, London ED (2004) Attenuation of cue-induced Tyrosine depletion attenuates dopamine function in healthy cigarette craving and anterior activation in volunteers. Psychopharmacology 154:105–111 bupropion-treated smokers: a preliminary study. Psychiatry Res Heatherton TF, Kozlowski LT, Frecker RC, Fagerstrom KO (1991) 130:269–281 The Fagerstrom test for nicotine dependence: a revision of the Caroll JB, Davies P, Richman B (1971) The American heritage word Fagerstrom Tolerance Questionnaire. Br J Addict 86:1119–1127 frequency book. American Heritage, New York Hodos W (1961) Progressive ratio as a measure of reward strength. Casey KF, Benkelfat C, Young SN, Leyton M (2006) Lack of effect of Science 134:943–944 acute dopamine precursor depletion in nicotine-dependent smokers. Houtsmuller EJ, Thornton JA, Stitzer ML (2002) Effects of selegiline Eur Neuropsychopharmacol 16:512–520 (L-deprenyl) during smoking and short-term abstinence. Psycho- Caskey NH, Jarvik ME, Wirshing WC (1999) The effects of pharmacology 163:213–220 dopaminergic D2 stimulation and blockade on smoking behavior. Hughes JR, Hatsukami D (1986) Signs and symptoms of tobacco Exp Clin Psychopharmacol 7:72–78 withdrawal. Arch Gen Psychiatry 43:289–294 Caskey NH, Jarvik ME, Wirshing WC, Madsen DC, Iwamoto-Schaap Hursh SR, Raslear TG, Shurtleff D, Bauman R, Simmons L (1988) A PN, Eisenberger NI, Huerta L, Terrace SM, Olmstead RE (2002) cost–benefit analysis of demand for food. J Exp Anal Behav Modulating rates by dopaminergic stimulation 50:419–440 and blockade. Nicotine Tob Res 4:259–266 Hutchison KE, Rutter MC, Niaura R, Swift RM, Pickworth WB, Corrigall WA, Coen KM, Adamson KL (1994) Self-administered Sobik L (2004) Olanzapine attenuates cue-elicited craving for nicotine activates the mesolimbic dopamine system through the tobacco. Psychopharmacology 175:407–413 . Brain Res 653:278–284 Jacobs EA, Bickel WK (1999) Modeling drug consumption in the Cox LS, Tiffany ST, Christen AG (2001) Evaluation of the brief clinic using simulation procedures: demand for heroin and questionnaire of smoking urges (QSU-brief) in laboratory and cigarettes in opioid-dependent outpatients. Exp Clin Psychophar- clinical settings. Nicotine Tob Res 3:7–16 macol 7:412–426 Dawe S, Gerada C, Russell MA, Gray JA (1995) Nicotine intake in Jarvik ME, Caskey NH, Wirshing WC, Madsen DC, Iwamoto-Schaap smokers increases following a single dose of haloperidol. PN, Elins JL, Eisenberger NI, Olmstead RE (2000) Bromocrip- Psychopharmacology 117:110–115 tine reduces cigarette smoking. Addiction 95:1173–1183 de la Fuente-Fernandez R, Phillips AG, Zamburlini M, Sossi V, Calne Koob GF, Le Moal M (1997) Drug abuse: hedonic homeostatic DB, Ruth TJ, Stoessl AJ (2002) Dopamine release in human dysregulation. Science 278:52–58 Psychopharmacology

Laviolette SR, van der Kooy D (2003) Blockade of mesolimbic Munafò MR, Hitsman B (2006) Neurocircuitry of attentional processes dopamine transmission dramatically increases sensitivity to the in addictive behaviours. In: Munafò MR, Albery I (eds) Cognition rewarding effects of nicotine in the ventral tegmental area. Mol and addiction. Oxford University Press, Oxford Psychiatry 8:50–59 Munafò MR, Mogg K, Roberts S, Bradley B, Murphy M (2003) Leyton M, Young SN, Pihl RO, Etezadi S, Lauze C, Blier P, Baker Selective processing of smoking-related cues in current smokers, GB, Benkelfat C (2000) Effects on mood of acute phenylalanine/ ex-smokers and never-smokers on the modified Stroop task. J tyrosine depletion in health women. Psychopharmacol 17:310–316 22:52–63 Munafò MR, Mannie ZN, Cowen PJ, Harmer CJ, McTavish SF (2007) Leyton M, Dagher A, Boileau I, Casey K, Baker GB, Diksic M, Gunn Effects of acute tyrosine depletion on subjective craving and R, Young SN, Benkelfat C (2004) Decreasing amphetamine- selective processing of smoking-related cues in abstinent cigarette induced dopamine release by acute phenylalanine/tyrosine smokers. J Psychopharmacol (in press) depletion: a PET/[11C]raclopride study in healthy men. Neuro- Murphy JG, MacKillop J (2006) Relative reinforcing efficacy of alcohol psychopharmacology 29:427–432 among college student drinkers. Exp Clin Psychopharmacol MacKillop J, Murphy JG (2007) A behavioral economic measure of 14:219–227 demand for alcohol predicts brief intervention outcomes. Drug Oldendorf WH, Szabo J (1976) Amino acid assignment to one of Alcohol Depend 89:223–233 three blood–brain barrier amino acid carriers. Am J Physiol Mackillop J, Menges DP, McGeary JE, Lisman SA (2007) Effects of 230:94–98 craving and DRD4 VNTR genotype on the relative value of Pardridge WM (1977) Kinetics of competitive inhibition of neutral alcohol: an initial human laboratory study. Behav Brain Funct amino acid transport across the blood–brain barrier. J Neurochem 3:11 28:103–108 Mansvelder HD, De Rover M, McGehee DS, Brussaard AB (2003) Perkins KA, Grobe J, Fonte C (1997) Influence of acute smoking Cholinergic modulation of dopaminergic reward areas: upstream exposure on the subsequent reinforcing value of smoking. Exp and downstream targets of nicotine addiction. Eur J Pharmacol Clin Psychopharmacol 5:277–285 480:117–123 Ratcliff R (1993) Methods for dealing with reaction time outliers. McNair D (1971) Profile of mood states manual. Educational and Psychol Bull 114:510–532 Industrial Testing Service, San Diego Robinson TE, Berridge KC (1993) The neural basis of drug craving: McTavish SF, Callado L, Cowen PJ, Sharp T (1999a) Comparison of an incentive-sensitization theory of addiction. Res Rev the effects of alpha-methyl-p-tyrosine and a tyrosine-free amino 18:247–291 acid load on extracellular noradrenaline in the rat in Robinson TE, Berridge KC (2000) The psychology and neurobiology vivo. J Psychopharmacol 13:379–384 of addiction: an incentive-sensitization view. Addiction 95 McTavish SF, Cowen PJ, Sharp T (1999b) Effect of a tyrosine-free (Suppl 2):S91–S117 amino acid mixture on regional brain synthesis Sayette MA, Shiffman S, Tiffany ST, Niaura RS, Martin CS, Shadel and release. Psychopharmacology 141:182–188 WG (2000) The measurement of drug craving. Addiction 95 Mitchell SH (2004) Effects of short-term nicotine deprivation on (Suppl 2):S189–S210 decision-making: delay, uncertainty and effort discounting. Spielberger CD, Gorusch RL, Lushene R, Vagg PR, Jacobs GA (1983) Nicotine Tob Res 6:819–828 Manual for the state trait anxiety inventory. Consulting Psycholo- Mogg K, Bradley BP (2002) Selective processing of smoking-related gists Press, Mountain View cues in smokers: manipulation of deprivation level and compar- Tiffany ST, Carter BL (1998) Is craving the source of compulsive drug ison of three measures of processing bias. J Psychopharmacol use? J Psychopharmacol 12:23–30 16:385–392 Tiffany ST, Conklin CA (2000) A cognitive processing model of Montgomery AJ, McTavish SF, Cowen PJ, Grasby PM (2003) alcohol craving and compulsive alcohol use. Addiction 95 Reduction of brain dopamine concentration with dietary tyrosine (Suppl 2):S145–S153 plus phenylalanine depletion: an [11C]raclopride PET study. Am Uslaner JM, Acerbo MJ, Jones SA, Robinson TE (2006) The J Psychiatry 160:1887–1889 attribution of incentive salience to a stimulus that signals an Mooney ME, Sofuoglu M (2006) Bupropion for the treatment of intravenous injection of . Behav Brain Res 169:320–324 nicotine withdrawal and craving. Expert Rev Neurother 6:965– Zhang H, Sulzer D (2004) Frequency-dependent modulation of 981 dopamine release by nicotine. Nat Neurosci 7:581–582