INTRODUCTION

Cocaine is among the most commonly abused illicit substances worldwide, with cocaine dependence (CD) contributing significantly to societal burdens such as health care, criminal justice, and public safety costs and lost productivity (Degenhardt and Hall, 2012). Despite relatively high treatment-seeking rates and ongoing efforts to improve therapeutic approaches, however, clinical success rates remain low, and no pharmacotherapies have been approved to date (Haile et al, 2012). A better understanding of neurobiological mechanisms contributing to CD could advance treatment strategies and improve clinical outcomes.

Among the main targets of investigation is the brain dopamine (DA) system, which not only is directly modulated by the acute action of cocaine, but also has a pivotal role in the addiction and withdrawal phenotypes (Melis et al, 2005). Decades of preclinical research, now supported by human neuroimaging findings, suggest that DA transmission at D2/3 receptors is diminished in addicted individuals abstinent 2 weeks (Melis et al, 2005), with higher relapse risk in those with the lowest transmission (Martinez et al, 2011; Wang et al, 2012). This evidence has prompted the view that low D2 is a useful biomarker for relapse risk, and that increasing DA transmission to ameliorate a hypo-active DA system may improve treatment success; however, this approach has yet to prove clinically efficacious.

At the same time, a novel line of research focused on the D3 DA receptor—a member of the D2-like family with unique features relevant to addiction (Sokoloff et al, 2001)—points to D3 as a potential new treatment target. The anatomical distribution of this receptor overlaps with key addiction neurocircuitry, including midbrain projections to limbic forebrain regions (amygdala, bed nucleus of stria terminalis, nucleus accumbens shell) mediating motivation, inhibitory control, emotion, and learning (Sokoloff et al, 2001), which can modulate activity in cortical loops involved in motivation, salience attribution, conditioned responses, and compulsive behavior (Cole et al, 2012). Moreover, preclinical studies demonstrate D3 upregulation following dopaminergic drug regimens, with corresponding behavioral sensitization, increased motivation, and drug seeking (LeFoll et al, 2005; Sokoloff et al, 2001). In line with these data, post-mortem human studies have reported elevated D3 levels in striatum of cocaine overdose fatalities (Segal et al, 1997; Staley and Mash, 1996), and neuroimaging findings from our laboratory point to similar effects in vivo, showing D3 elevation in D3-rich regions (substantia nigra (SN), globus pallidus (GP), ventral pallidum (VP)) in methamphetamine dependence (MD), which (in SN) relates to drug wanting (Boileau et al, 2012b). D3 contributions may even generalize to behavioral addictions, as we have also reported a relationship between SN D3 and symptom severity in pathological gambling (Boileau et al, 2012a).

Importantly, D3 receptor antagonists have shown promise in reversing acquisition and expression of drug-seeking and cue-induced relapse in animals (Heidbreder and Newman, 2010; Heidbreder et al, 2005), and clinical trials of nicotine dependence and food reward are translating this idea to humans (Mugnaini et al, 2013; Nathan et al, 2012). Taken together, the evidence suggests that activity at the D3 receptor is pathologically increased in addiction (in contrast to D2 downregulation), and, by extension, that a therapeutic strategy aimed at normalizing low transmission at D2 may undermine clinical outcome by exacerbating already-exaggerated D3 processes.

Positron emission tomography (PET) neuroimaging provides a non-invasive method to investigate DA receptors in human brain in vivo, but the evidence regarding D3 is sparse, owing to a lack (until recently) of suitable radioligands. Most DA PET studies have used [11C]raclopride to measure D2/3 receptors in striatum, but this ligand binds non-selectively to D2 and D3, so that binding primarily reflects D2, given the relative predominance of this receptor over D3. [11C]-(+)-propyl-hexahydro-naphtho-oxazin ([11C]-(+)-PHNO; Wilson et al, 2005)), on the other hand, is a recently developed D2/3 radioligand with preferential affinity for D3 in vivo, allowing for measurement of this receptor in D3-rich areas (whereas in D3-devoid regions, signal reflects D2; Tziortzi et al, 2011). Using [11C]-(+)-PHNO and [11C]raclopride in the same individuals, therefore, allows for a more detailed understanding of dopaminergic targets for pharmacological strategies.

In a previous study using [11C]-(+)-PHNO, we reported higher D3 receptor availability in the SN and marginally lower D2/3 in the striatum of methamphetamine-dependent polydrug-abusing subjects compared with controls (Boileau et al, 2012b). The present study, conducted in an independent sample of minimally comorbid cocaine-dependent subjects (many actively using, but all abstinent with negative urine screens on scan day), tested the hypothesis that [11C]-(+)-PHNO binding to D3 would be elevated in CD and related to the addiction behavioral phenotype, whereas [11C]raclopride binding was expected to be low.

MATERIALS AND METHODS

Subjects

All procedures were approved by the Centre for Addiction and Mental Health Research Ethics Board and complied with ethical standards of the Helsinki Declaration (1975; updated 1989). Fifteen CD volunteers were recruited from a community program (‘Getting Started,’ an ambulatory twice-weekly psychoeducation and support program aimed at engaging clients and increasing readiness for intensive treatment, but itself providing no therapeutic intervention). For comparison, 15 healthy control (HC) volunteers were recruited from the community via flyers and advertisements. After complete description of the study, all volunteers gave written informed consent.

Subjects in the CD group were required to (1) self-report cocaine as the primary currently abused drug; (2) meet DSM-IV criteria for CD (as per SCID); and (3) provide a scalp hair sample and/or urine screen positive for cocaine/metabolites. HC subjects were required to have no history of cocaine use, and no more than five lifetime occasions of recreational drug use (except cannabis). All subjects were required to be 18–55 years old and meet the following criteria: (1) no current Axis I disorder (as per SCID), except CD or substance-induced mood disorder in CD; (2) no lifetime history of alcohol or substance dependence (except caffeine or nicotine, and cocaine in CD; cannabis abuse was not exclusionary); (3) no medical conditions likely to affect the brain; (4) no current use of antidepressant or psychotropic medications; and (5) no PET or magnetic resonance image (MRI) contraindications (eg, radiation exposure exceeding guidelines, metal implants, pregnancy/lactation).

Study Procedure and Outcome Measures

Screening session

Screening for inclusion/exclusion consisted of the SCID for DSM-IV, detailed medical history, and questionnaires assessing alcohol and drug use patterns and depressive symptomatology (Beck Depression Inventory (Beck and Steer, 1984), Snaith–Hamilton Pleasure Scale (Snaith et al, 1995)). For CD subjects, we additionally assessed drug use history, severity, and toxicology (via scalp hair (US Drug Testing Laboratory) to assess drug use in recent months, and a comprehensive urine screen (9-Drug Test Panel, BTNX; broad spectrum, GC–MS) to test for prescription and illicit drugs.

Behavioral battery

Subjects underwent a brief battery of tests aimed at characterizing the CD behavioral phenotype. Tests included: the Halsted-Reitan Trail-Making Task (Reitan, 1955) to assess attention and set-shifting, Continuous Performance (Conners, 1994) and Go/No-Go tasks to assess attention and inhibitory control, the Balloon Analogue Risk Task (Lejuez et al, 2002) and Game of Dice Task (Brand et al, 2005) to assess risky decision making, the Kirby Delay Discounting Task (Kirby, 2000) to assess temporal discounting of reward, the Wisconsin Card Sorting Test to assess perseverative responding, the Barratt Impulsiveness Scale (Patton et al, 1995) and a Finger Tapping task to assess psychomotor function.

Neuroimaging session

Neuroimaging consisted of two PET scans, [11C]-(+)-PHNO and [11C]raclopride, conducted on the same day ([11C]raclopride first), along with a MRI to help delineate regions of interest (ROI) in PET analyses. CD subjects were not admitted to an inpatient unit for withdrawal management, but were asked to discontinue all drug use (except cigarette smoking) at least 10 days before the PET session. Compliance was assessed with a urine drug screen on scan day, requiring a negative result for cocaine. To avoid nicotine withdrawal on PET day, cigarette smokers were advised to smoke to satiation before each scan (up until 1 h before each injection). At the beginning of the PET day, CD subjects completed assessments of cocaine craving and withdrawal (Cocaine Selective Severity Assessment (Kampman et al, 1998), Cocaine Urge Questionnaire (adapted from Bohn et al (1995), Desire for Cocaine Scale (adapted from James et al (2004)).

[11C]-(+)-PHNO and [11C]raclopride synthesis and image acquisition protocols are described in detail elsewhere (Graff-Guerrero et al, 2008). PET scanning was performed on a high-resolution PET camera system (CPS-HRRT, Siemens Medical Imaging, Knoxville, TN), using a custom fitted thermoplastic mask to reduce head movement (TruScan Imaging, Annapolis, USA). Following a 15-min transmission scan, a bolus injection of [11C]-(+)-PHNO or [11C]raclopride was given into an antecubital vein ([11C]-(+)-PHNO mean dose=9.04 mCi, specific activity=1345.73 mCi/μmole, mass=2.21 μg; [11C]raclopride mean dose=9.2 mCi, specific activity=1862.31 mCi/μmole, mass=1.9 μg; no group differences). [11C]raclopride data were acquired for 60 min, and [11C]-(+)-PHNO data for 90 min, as 80 min of data acquisition have been shown to yield stable binding estimates (Ginovart et al, 2007). The MRI (Signa 1.5T, General Electric Medical Systems, Milwaukee, WI) consisted of a standard proton density sequence acquired over the whole brain.

Image Analysis and Statistical Approach

Delineation of ROIs including whole striatum, sensorimotor striatum (SMST), associative striatum (AST), limbic striatum (LST), VP, GP, SN, and cerebellar cortex (excluding vermis and lobules IX and X) is described in Martinez et al (2003) and Boileau et al (2012b). SN, VP, and GP ROIs were included because they are rich in D3 and a high proportion of [11C]-(+)-PHNO signal (100%, 75%, and 65%, respectively) reflects D3 binding (Tziortzi et al, 2011). [11C]-(+)-PHNO and [11C]raclopride time activity curves were obtained from dynamic data using ROMI (Rusjan et al, 2006). Specific binding (BPND) in ROIs was estimated using the simplified reference tissue method (Lammertsma and Hume, 1996), implemented in PMOD (v 2.8.5; PMOD Technologies Ltd, Zurich, Switzerland); this method is most suitable for simplified analysis, as it minimizes BPND under-estimation and inter-individual variability (Ginovart et al, 2007).

Group comparisons of [11C]-(+)-PHNO and [11C]raclopride BPND were performed using repeated-measures ANOVAs or ANCOVAs (with ROIs as the repeated measure and CD vs HC as the between-groups factor), using Greenhouse-Geisser sphericity corrections when indicated. Regional group differences were examined for significance using t-tests, Bonferroni corrected for planned comparisons. Relationships between PET measures and continuous behavioral variables were analyzed with Pearson product moment correlations and Spearman’s Rank tests for categorical data.

RESULTS

Subject Characteristics

Subject characteristics and demographic information are described in Table 1. CD and HC groups were matched for age, sex, ethnicity, body mass, and cigarette and alcohol use, but CD subjects had lower education levels and consumed more cannabis. CD also reported more depressive symptomatology and anhedonia than HC, and two CD subjects met criteria for a current substance-induced depressive episode at study entry. Control analyses for these factors (see below) confirmed that they did not influence the results.

Table 1 Participant Characteristics

Cocaine use characteristics in the CD group are described in Table 2. During the screening/intake session, 10 out of 15 CD provided a cocaine-positive urine test, confirming recent cocaine use, and hair analysis confirmed cocaine use in the remaining five subjects (and all other subjects with scalp hair, 12 out of 15). In a minority of the sample, toxicology screening (performed during the screening/intake session) also revealed (limited) use of other drugs: opiate metabolites in hair (3 out of 15) and urine (1 out of 15), cannabis in urine (4 out of 15), and benzodiazepine in hair (1 out of 15). Although subjects were asked to remain at least 10 days abstinent between screening and PET visits, two subjects only achieved 7 days abstinence between visits; however, their urine screens still met the requirement of testing negative for cocaine on scan day.

Table 2 Cocaine Use Patterns and Severity in CD Sample (N=15)

PET Results

All subjects completed both PET scans, but one CD [11C]raclopride scan was lost due to radiochemistry problems. [11C]raclopride and [11C]-(+)-PHNO BPND correlated (controlling for group) in SMST (where both tracers are thought to predominantly measure D2; r(26)=0.45, p=0.02), but not AST or LST (both p>0.25). Neither ROI volumes nor standard uptake values for cerebellum differed significantly between groups (all p>0.19).

A repeated-measures ANOVA investigating regional differences in [11C]-(+)-PHNO BPND between groups yielded a significant Group × ROI interaction (F(2.7,70.8)=3.63, p=0.02). Pairwise contrasts revealed that CD subjects had higher [11C]-(+)-PHNO BPND in the SN than HC (+24%, p=0.06, Cohen’s d=0.71; Figure 1a). Group differences were not statistically significant in other D3-rich regions (VP: +5%; GP: −9%) or the striatum (−4 to 5%).

Figure 1
figure 1

Individual binding potential (BPND) values across regions of interest for each PET tracer in cocaine-dependent (CD) and healthy control (HC) subjects. (a) [11C]-(+)-PHNO: as predicted, CD showed higher [11C]-(+)-PHNO BPND than HC in the substantia nigra (SN; #p=0.06). Ventral pallidum (VP), another D3-rich region, also showed this pattern, but this effect was not statistically significant. No group differences were found in globus pallidus (GP), striatal subregions, or whole striatum. (b) [11C]Raclopride: CD showed lower [11C]raclopride BPND in the striatum as a whole (*p<0.05), but not individual subregions. AST, associative striatum; LST, limbic striatum; SMST, sensorimotor striatum.

PowerPoint slide

A second ANOVA investigating [11C]raclopride BPND found lower binding in CD than HC (Figure 1b); this effect was not region dependent, but occurred in the striatum as a whole (−11%; F(1,27)=4.32, p<0.05). Group differences were not statistically significant in individual striatal subdivisions (AST: −11%, p=0.20, Cohen’s d=0.49; LST: −10%, p=0.14, Cohen’s d=0.57; SMST: −11%, p=0.06; Cohen’s d=0.72).

Several follow-up tests were performed to rule out potentially confounding factors. In all ROIs, PET measures (BPND) did not differ between CD subjects who did and did not (n=5) provide a cocaine-positive urine at study entry (indicating recent use vs abstinence; all p>0.25), between those who did (n=4) and did not currently abuse cannabis (all p>0.25), or between those who did (n=2) and did not experience current drug-induced depressive symptoms (all p>0.1). Further, removing abstinent subjects and subjects co-morbid for cannabis/depression from analyses did not affect the ROI × Group effect in [11C]-(+)-PHNO BPND (recency: F(2.6,54.6)=3.50, p=0.03; cannabis: F(2.7,58.7)=4.06, p=0.01; depression: F(2.9,69.4)=2.56, p=0.06), or the group difference in [11C]raclopride BPND (recency: F(1,23)=5.50, p=0.03; cannabis: F(1,22)=3.02, p=0.10; depression: F(1,23)=5.50, p=0.03). PET findings did not correlate with addiction severity measures, and testing whether time from last use influenced binding showed no correlation between days abstinent and [11C]raclopride (0.06<r<0.30, all p>0.29) or [11C]-(+)-PHNO (−0.32<r<0.07, all p>0.24) BPND in any of the ROIs.

Behavioral Battery and Relationships with PET Measures

As the behavioral battery was aimed at characterizing the CD phenotype, all CD subjects but not all HC subjects, completed the battery, so that group comparisons are not presented here (means are available in Supplementary Table 1). Across CD subjects, assessing correlations between [11C]-(+)-PHNO BPND in SN (index of brain D3 levels) and task performance identified a relationship with number of risky choices on the Game of Dice Task (r=0.51, p=0.05), and with commission errors on the Continuous Performance Task (r=0.52, p=0.05; Figure 2).

Figure 2
figure 2

Correlations across cocaine-dependent (CD) subjects (N=15) illustrating relationships between [11C]-(+)-PHNO BPND in substantia nigra (SN; index of brain D3 levels) and (a) commission errors during the Continuous Performance Task (CPT), indicative of behavioral impulsivity; (b) number of risky choices during the Game of Dice Task (GDT), indicative of risky decision making.

PowerPoint slide

DISCUSSION

The present study supports our predictions of high D3 (compared with low D2) receptor availability in CD. The findings add to a small but growing human neuroimaging literature pointing to D3 as a novel biomarker and potential treatment target, and are also in line with the extensive body of research showing downregulation of D2. In addition to extending our understanding of neurobiological and behavioral components of addiction, our findings may have important clinical implications: pharmacological strategies aimed at normalizing D2 deficiencies may inadvertently exacerbate an exaggerated D3 response, and D3 antagonism may provide a viable alternative or adjunct.

The main finding of the study is that brain D3 levels may be heightened in CD (vs HC), as indicated by higher [11C]-(+)-PHNO binding in the SN (where 100% of signal reflects D3; Tziortzi et al, 2011). The pattern is reiterated in the D3-rich VP (where 75% of signal reflects D3), but this difference was not statistically significant. Importantly, we also identified a relationship between SN [11C]-(+)-PHNO binding and impulsivity/risky decision making, suggesting that D3 is relevant to behavioral contributors to addiction. The finding is consistent with post-mortem data showing elevated D3 levels in cocaine abuse (Segal et al, 1997; Staley and Mash, 1996), as well as a sizeable body of preclinical data, which, although not entirely consistent (Richtand et al, 2001), mainly demonstrates D3 upregulation and related locomotor sensitization, drug cue reactivity, drug-seeking, and motivation to take drugs (LeFoll et al, 2005; Sokoloff et al, 2001). Our finding of 24% higher SN D3 binding in CD vs HC is remarkably in line with a preliminary report that reported 22% greater [11C]-(+)-PHNO BPND in SN (Matuskey et al, 2011). It also echoes our previous neuroimaging studies of MD (Boileau et al, 2012b), where we showed heightened SN [11C]-(+)-PHNO binding that related to self-reported drug wanting, and of pathological gambling (Boileau et al, 2012a), where SN [11C]-(+)-PHNO binding related to impulsiveness and gambling severity. It should be noted that the magnitude of SN D3 elevation (relative to respective HC groups) differed between our CD and MD samples (Boileau et al, 2012b; 24% here vs 46% in MD), which could be explained by age differences between samples (mean 42 years here vs 28 years in MD), or differences between the two drugs (eg, mechanism of action, duration of effects, route of administration).

An interesting (although controversial) point is that in our previous report, we calculated relative D3 levels, ie, individual D3-to-D2 ratios estimated as [11C]-(+)-PHNO BPND in SN (100% of signal thought to reflect D3) to [11C]-(+)-PHNO BPND in dorsal striatum (devoid of D3 so signal is thought to reflect D2 (Tziortzi et al, 2011)), and showed that this measure was greater in stimulant dependent than control subjects (Boileau et al, 2012b). Here, too, we find that CD subjects had a 32% greater D3-to-D2 binding fraction than HC (p=0.01), possibly providing further evidence for heightened D3 receptor expression (outside of SN). However, the biological relevance of this index is under debate, and the implications of this finding therefore highly speculative.

Together, [11C]-(+)-PHNO PET studies suggest that D3 elevation may be a consistent and predictive biomarker for addiction, with potential use for clinical innovation. Indeed, the clinical utility of targeting D3 in addiction treatment is currently under intensive investigation: Preclinical studies with highly selective D3 receptor antagonists have observed attenuation of drug-seeking, self-administration, and cue- and stress-induced reinstatement in animal models of addiction (Heidbreder and Newman, 2010; Heidbreder et al, 2005; LeFoll et al, 2007), and in humans, clinical trials have reported D3 antagonist effects on food reward and nicotine craving in smokers (Mugnaini et al, 2013; Nathan et al, 2012), supporting the potential clinical efficacy of D3 antagonism.

Mechanisms underlying the paradoxical upregulation of D3 (contrasting with D2 downregulation) and its downstream effects are still unclear (see Boileau et al (2012b) for discussion). Briefly, upregulation is thought to result from repeated stimulation of D1 receptors during cocaine use, leading to release of brain-derived neurotrophic factor (BDNF), which is, in turn, linked to increased D3 expression (Guillin et al, 2001), and is elevated in CD (D’Sa et al, 2011). Although D3 receptors are expressed on all DA neurons in the SN, their physiological role there remains under debate (Davila et al, 2003); moreover, D3 receptors are both reciprocal autoreceptors and heteroreceptors (Sokoloff et al, 1990), and it is unknown whether the D3 upregulation occurs on dopaminergic or other (eg, GABAergic) neurons (Bordet et al, 1997; Guillin et al, 2001). Increased transmission at D3 appears to modify limbic outputs and functional connectivity between orbitofrontal cortex and networks involved in cognitive control and goal-directed behavior (Cole et al, 2012), thereby modulating motivation to use drugs. Future studies will be important in determining the effects of D3 antagonism on these outcomes.

A secondary finding of the study is that [11C]raclopride binding in striatum was lower in CD than HC subjects, suggesting low D2 availability (although [11C]raclopride cannot distinguish between D2 and D3). This effect was statistically significant only when considering the striatum as a whole, but the 10–11% difference we observed in individual subregions is within the range of what has been previously reported (−6–17%; Martinez et al, 2004, 2011; Narendran et al, 2011), and effect size estimates are in the medium range, suggesting that our failure to find significant group differences in individual subregions may have been due to low statistical power. The pattern of low D2/3 signal is consistent with preclinical and other PET imaging evidence, supporting the notion that low D2 availability is a hallmark of addiction across a range of substances, reflecting a hypodopaminergic state that can contribute to relapse to drug use (Martinez et al, 2011; Wang et al, 2012). Our finding therefore confirms that our CD sample has comparable dopaminergic characteristics to previous study samples (Martinez et al, 2011; Wang et al, 2012), suggesting that the D3 finding reflects a previously unobserved process rather than sample differences.

One question raised by our use of two radiotracers is why low striatal D2/3 levels were detected with [11C]raclopride but not [11C]-(+)-PHNO, including in dorsal striatum, where [11C]-(+)-PHNO binding is thought to predominantly reflect D2. Although low statistical power is a possibility, there are also several biological explanations. It is possible that our failure to detect group differences with [11C]-(+)-PHNO reflects an ectopic upregulation of D3 in dorsal stratum, which has been observed in animals (Bordet et al, 1997) and would mask lowered D2 receptor binding. Alternatively, as an agonist radiotracer, [11C]-(+)-PHNO is more sensitive to endogenous DA levels (Shotbolt et al, 2012; Willeit et al, 2008), so that low tonic DA levels in CD subjects could have led to more available receptor sites, masking lowered D2 levels. Last, [11C]-(+)-PHNO only labels receptors in the G-protein-coupled high-affinity state (D2HIGH), and as animal literature suggests that D2HIGH proportion is relevant to addiction (Seeman et al, 2007), it is possible that our failure to detect group differences with [11C]-(+)-PHNO reflects a higher D2HIGH fraction in CD. We tested for this possibility by comparing [11C]-(+)-PHNO binding to [11C]raclopride (which binds D2HIGH and D2LOW indiscriminately) in SMST (presumably devoid of D3). This analysis is highly theoretical and additionally complicated by potential ectopic upregulation of D3, but revealed a (statistically non-significant, p=0.18) 14% greater D2HIGH fraction in CD than HC. This failure to find an effect is in line with a previous study using [11C]NPA (Narendran et al, 2011), but nonetheless could have contributed to the discrepancy between the [11C]raclopride and [11C]-(+)-PHNO findings in the present report. However, it is not possible to distinguish between these possibilities here.

Several limitations of the study should be noted. The first concerns the study sample, which included CD subjects with a wide range of abstinence durations (which was not supervised) and in some cases co-morbid drug use or mood symptoms. Our inclusion of 10 out of 15 actively using (albeit urine-negative) CD subjects distinguishes our studies from previous reports, which imposed a monitored/supervised abstinence period of at least 14 days (eg, Martinez et al, 2004, 2011; Wang et al, 2012). We conducted explicit follow-up analyses to test for differences between actively using and abstinent subjects in our sample, and found no modulating effects of abstinence (or mood/comorbid drug use), although these had low power and require confirmation in larger samples. At the same time, we excluded other co-morbidities to strengthen our ability to attribute effects to CD, limiting generalizability to the broader CD population. Another set of limitations concerns the radiotracer [11C]-(+)-PHNO, including previously noted limitations (see Rabiner and Laruelle (2010)) such as scanning at non-tracer doses (Shotbolt et al, 2012), using a reference tissue with known specific binding (cerebellum (Murray et al, 1994)—although vermis and lobules 9 and 10 were excluded), and long wash-out times in D3-rich regions that may limit adequate quantification of receptor binding (Willeit et al, 2006). In addition, as SN is the only region where 100% of [11C]-(+)-PHNO signal reflects D3 binding, with all other ROIs reflecting some mixture of D2 and D3 (Tziortzi et al, 2011), interpretation of findings is anatomically limited. Finally, as in all cross-sectional studies of drug abuse, we acknowledge that group differences could have predated drug use, and that longitudinal designs must be employed to distinguish cause and effect.

Despite these limitations, the present study is consistent with preclinical and post-mortem data, replicates our previous finding of heightened SN D3 levels in psychostimulant addiction, and is the first to address the combination of D3 and D2 receptor states in CD. D3 upregulation is emerging as a biomarker of the addicted state, and the newly gained ability to examine pharmacological features in living humans with [11C]-(+)-PHNO holds promise for the guidance of clinical investigations, and the development of novel D3 antagonist strategies in the treatment of addictions.

FUNDING AND DISCLOSURE

This project was funded in part by the National Institute of Health (NIH) NIDA R01DA025096 (SK/IB). IB also received funds from the Ontario Mental Health Foundation (OMHF), the Canadian Institute of Health Research (CIHR), and the Parkinson’s Society of Canada. SK received a fee to provide expert witness testimony regarding the adverse effects of methamphetamine. TG received consulting fees from Pfizer and Novartis, and grant/contract support from Pfizer, NIH, the Canada Foundation for Innovation, CIHR, and OMHF. PS received consulting fees from, Johnson & Johnson Consumer Health Care Canada, Pfizer Inc. Canada, Pfizer Global, Sanofi-Synthelabo Canada, GSK Canada, Genpharm and Prempharm Canada, NABI Pharmaceuticals, V-CC Systems Inc. and E-Health Behavior Change Software Co. Astra Zeneca Canada Inc, grant/contract support from Health Canada, Smoke Free Ontario, MHP, CTCRI, CIHR, Alberta Health Services (formerly Alberta Cancer Board), Vancouver Coastal Authority, Pfizer, OLA, ECHO, NIDA, CCS, and speakers Bureau/Honoraria from Johnson & Johnson Consumer Health Care Canada; Pfizer Inc., Canada; Pfizer Global; Sanofi-Synthelabo, Canada; GSK, Canada; Genpharm and Prempharm, Canada; NABI Pharmaceuticals. AW received funds from OMHF. SH received funds from the Canada Foundation for Innovation and Ontario Research Funds. DP, AB, PR, and JT report no competing interests.