Imaging the Neurochemistry of Alcohol and Substance Abuse

Home , DASB, Dihydrotetrabenazine, Iomazenil

539

NEUROIMAGING CLINICS OF NORTH AMERICA

Neuroimag Clin N Am 17 (2007) 539–555 Imaging the Neurochemistry of Alcohol and Substance Abuse

Diana Martinez, MDa,*, Jong-Hoon Kim, MDa, John Krystal, MDb, Anissa Abi-Dargham, MDa,c

- Cocaine dependence Behavioral correlates of low D2/3 receptor Dopamine D2/3 receptors and dopamine binding potential transmission Alcohol dependence and presynaptic Functional signiﬁcance of low D2/3 dopamine receptor binding in cocaine Alcohol dependence and the dopamine dependence transporter Behavior and dopamine transmission Serotonin and alcohol dependence Imaging cue-induced craving in cocaine Measures of GABA in alcohol dependence dependence Opioids and alcohol dependence Cocaine dependence and the dopamine - Heroin dependence transporter - Methamphetamine abuse Imaging studies of cocaine dependence - Methylenedioxymethamphetamine and other neurotransmitters (Ecstasy) abuse - Alcohol dependence - Hallucinogens - Dopamine D2/3 receptor and alcohol Summary dependence - References

Positron emission tomography (PET) and single we brieﬂy overview the concepts that are needed photon emission computed tomography (SPECT) to interpret these studies. The PET radiotracers use radiotracers to image molecular targets in the most frequently used in substance abuse research human brain. These techniques have been applied are those that label the dopamine type 2/3 (D2/3) over the last decade to study addiction and provide receptors of the striatum, such as the antagonists an important body of knowledge about the neuro- 18F-N-methylspiroperidol (labeled with a posi- chemical alterations associated with drug and tron-emitting ﬂuorine) and 11 C-raclopride (labeled alcohol dependence. with a positron emitting carbon). Other radio- Although the techniques of PET and SPECT tracers that are available include those that label molecular imaging have been reviewed elsewhere, the dopamine transporter, serotonin transporter

a Department of Psychiatry, Columbia University College of Physicians and Surgeons, New York State Psychiatric Institute, 1051 Riverside Drive, Box #31, New York, NY 10032, USA b Yale University School of Medicine, VA Connecticut Healthcare System (116-A), 950 Campbell Avenue, West Haven, CT 06516, USA c Department of Radiology, Columbia University College of Physicians and Surgeons, New York State Psychiatric Institute, 1051 Riverside Drive, Box #31, New York, NY 10032, USA * Corresponding author. E-mail address: [email protected] (D. Martinez).

1052-5149/07/$ – see front matter. Published by Elsevier Inc. doi:10.1016/j.nic.2007.07.004 neuroimaging.theclinics.com 540 Martinez et al

and receptors, GABA receptor, and opioid receptors. been performed in cocaine-dependent subjects, so The outcome measure in PET studies is termed re- the authors’ review of the literature begins with ceptor availability or binding potential (BP), which this disorder. is equivalent in pharmacologic terms to the product of receptor density and affinity of the radiotracer for Cocaine dependence the receptor. In addition to the BP, PET and the radiotracer Dopamine D2/3 receptors and dopamine 11 C-raclopride can also be used to measure striatal transmission dopamine transmission. 11 C-raclopride binding is As early as 1990, studies with PET have shown that sensitive to the endogenous dopamine in the brain: cocaine addiction is associated with alterations in increases in extracellular dopamine (produced with dopamine receptors. Using PET and 18F-N-methyl- a psychostimulant such as methylphenidate or spiroperidol, Volkow and colleagues [3] demon- amphetamine) decrease 11 C-raclopride binding strated that cocaine dependence was associated (Fig. 1). In the same individual, therefore, a com- with a decrease in D2/3 receptor availability in the parison of baseline (ie, pre-stimulant) and post- striatum compared with healthy control subjects. stimulant BP provides an indirect measure of Subsequent studies performed using similar dopamine transmission. The mechanism underly- methods have consistently shown decreases of ing the decrease in radiotracer binding is believed 11% to 15% in striatal D2/3 receptors in this same to be competition between extracellular dopamine population using 11 C-raclopride [4–6]. In addition, and the radiotracer for the D2/3 receptor, although Volkow and colleagues [4] reported that the de- other mechanisms, such as receptor affinity state, crease in D2/3 receptors persisted in a group of internalization, or polymerization may also play cocaine-dependent subjects rescanned after 3 a critical role [1,2]. months of inpatient rehabilitation. Several PET studies in human drug- and alcohol- Although this decrease in D2/3 receptor BP was dependent subjects have been performed using first measured in cocaine dependence, a similar de- these techniques. Most of these studies have mea- crease has been seen in several other addictions, sured dopamine receptors and dopamine release such as heroin addiction [7], alcohol dependence in the striatum, and thus are the focus of this review. [8,9], methamphetamine abuse [10], and even obe- In addition, most PET studies of addiction have sity [11]. These findings suggest that low D2/3

Fig. 1. Comparison of amphetamine-induced [11C]raclopride displacement in a healthy control (top row) and cocaine dependent subject (bottom row). Panel A shows the healthy control in the pre-amphetamine condition, panel B shows the post-amphetamine [11C]raclopride binding in the control. Panel C shows the cocaine dependent subject in the pre-amphetamine condition, and panel D shows the post-amphetamine condition in the same subject. In the pre- amphetamine condition, there is higher D2/3 receptor binding in the healthy control compared to the cocaine dependent subject (panel A versus panel B). In the post-amphetamine condition, there is a notable decrease in [11C]raclopride binding in the healthy control, whereas there is little difference in the cocaine dependent subjects. The decrease in radiotracer binding in the healthy control subject correlates with greater dopamine release, whereas the cocaine dependent subject has blunted dopamine transmission. Imaging the Neurochemistry of Addiction 541 receptor availability might be a general risk factor subjects [6]. In the study of dopamine transmis- for addiction, and it has been hypothesized that sion, cocaine dependence was associated with 11 low D2/3 receptor BP is associated with a low sensi- a marked reduction in amphetamine-induced C- tivity to naturally occurring reinforcers and a pro- raclopride displacement in each of the functional pensity to depend on pharmacologic stimulation subregions [18]. These studies thus demonstrate to experience reward [12,13]. that in cocaine dependence the deficits in dopa- In addition to baseline D2/3 receptor BP, PET mine neurotransmission are similar across the ven- studies have been used to investigate dopamine tral and dorsal subdivisions of the striatum. transmission in cocaine dependence. Volkow and colleagues [5] demonstrated that cocaine depen- Functional significance of low D2/3 receptor dence was associated with less displacement of binding in cocaine dependence 11 C-raclopride following methylphenidate (0.5 The observation of decreased D2/3 receptor BP in mg/kg IV) compared with control subjects, suggest- cocaine-dependent subjects raises the question of ing that cocaine dependence is associated with whether this finding may serve as a risk factor for a loss of dopamine transmission. Malison and col- cocaine dependence. In fact, a series of studies leagues [14] reported similar results using SPECT have suggested that a high level of D2/3 BP may 123 and I-IBZM (which also labels the D2/3 receptor) be protective against developing dependence. A using amphetamine (0.3 mg/kg IV) to increase en- recent study reported that nonaddicted siblings of dogenous dopamine. Both studies thus suggest cocaine abusers had a higher D2/3 receptor BP that cocaine dependence is associated with a de- compared with their cocaine-dependent siblings crease in presynaptic dopamine function. This hy- [19]. Because the siblings presumably had similar pothesis is supported by a PET study showing that risk factors for dependence, this finding suggests cocaine-dependent subjects (abstinent 11–30 that elevated D2/3 receptor BP may be a neurobio- days) had lower uptake of the levodopa analog logic marker of resilience. Volkow and colleagues 18 6- F-fluoro-L-DOPA compared with healthy con- [20,21] reported that high striatal D2/3 receptor BP trol subjects, suggesting that cocaine dependence in healthy control subjects was predictive of an un- is associated with a decrease in the dopamine stores pleasant experience following administration of the of the presynaptic neuron [15]. psychostimulant methylphenidate, and conversely, Because of the resolution of PET (and SPECT) lower D2/3 BP was associated with a pleasurable ex- cameras, these studies measured dopamine recep- perience. Insofar as a pleasurable experience with tors and dopamine transmission in the striatum a drug indicates a risk for substance abuse, these re- as a whole. In other words, the resolution did not sults indicate that high D2/3 receptor BP may be allow for differentiation of the signal emitted protective. from the caudate, putamen, and ventral striatum. The authors recently investigated the correlation With a higher-resolution device, the substructures between low D2/3 receptor BP and the choice to of the striatum can be measured separately self-administer cocaine in human cocaine-depen- [16,17]. The authors recently published two studies dent subjects [6]. As described, these subjects had in human cocaine-dependent subjects and matched a decrease in D2/3 receptor availability throughout healthy control subjects using a high-resolution the subdivisions of the striatum compared with PET camera [16,18]: one study measured D2/3 re- a group of matched healthy control subjects. Fol- ceptors and the other measured dopamine lowing the PET scans, the cocaine-dependent vol- transmission using 11 C-raclopride and a psychosti- unteers underwent cocaine self-administration mulant challenge (amphetamine 0.3 mg/kg IV). sessions in which participants were given the choice In these studies, the striatum was subdivided into between doses of smoked cocaine and monetary the ventral striatum (which contains the nucleus vouchers. This study showed no correlation be- accumbens), associative striatum (which includes tween D2/3 receptor BP and the positive effects of the caudate and anterior putamen and is largely cocaine or the choice to self-administer cocaine involved in cognition), and the sensorimotor [6]. Although low D2/3 receptor availability might striatum (which contains the posterior putamen be related to a pleasurable psychostimulant experi- and receives input from motor and premotor ence in control subjects, this phenomenon thus areas). does not seem to be present in addicted subjects. The results of the authors’ study investigating One way to interpret these data is that low D2/3 re- D2/3 receptor BP showed a significant reduction in ceptor BP may correlate with a positive response to all three striatal subdivisions in the cocaine-depen- a psychostimulant and is thus a risk factor for co- dent subjects (decreases of 15% in the limbic and caine dependence. Of the individuals who become associative striatum and 17% in the sensorimotor addicted (ie, choose to self-administer cocaine), striatum) compared with the healthy control most therefore have lower than average D2/3 542 Martinez et al

receptor binding. Within the cohort of cocaine with a greater risk for relapse. In contrast, the data abusers, however, low D2/3 receptor BP does not described demonstrated that subjects with the low- predict drug-seeking behavior. est amphetamine-induced dopamine release are more likely to self-administer cocaine, such that Behavior and dopamine transmission greater deficits in dopamine release may indicate Given the consistent finding of low dopamine risk for relapse. The reason for this difference is transmission in cocaine dependence, the authors not clear, although it has been suggested that set- recently completed a study designed to investigate shifting depends on dopamine transmission in the correlation between drug-seeking behavior the dorsal striatum and reversal learning is medi- and blunted dopamine transmission. As described, ated by dopamine in the ventral striatum [27]. Do- cocaine dependence was associated with a decrease pamine transmission in the ventral versus dorsal in amphetamine-induced 11 C-raclopride displace- striatum may thus play a critical role in relapse. ment compared with healthy control subjects [18]. Following the scans, the cocaine-dependent Cocaine dependence and the dopamine participants were given the choice between smoked transporter cocaine and monetary vouchers in self-administra- A study using SPECT and the dopamine transporter tion sessions. The results of this study showed (DAT) radiotracer 123I-beta-CIT reported a 20% in- that blunted dopamine transmission in the ventral crease in DAT availability in cocaine abusers who striatum was predictive of the choice for cocaine had been abstinent for only 96 hours [28].Two over the choice for money [18]. The self-administra- studies using the radiotracer 11 C-cocaine to label tion sessions were developed as a laboratory model the DAT, however, showed no difference in BP be- of relapse based on animal studies showing that tween healthy control subjects and cocaine abusers a priming dose of cocaine reinstates cocaine self-ad- who had been abstinent 5Æ8 days or 42Æ7 days ministration [22–24]. This finding thus suggests [29,30]. Together these studies suggest that the that cocaine-dependent subjects who are the most DAT is elevated in very early abstinence but does vulnerable to relapse are those with the lowest pre- not seem to differ from control subjects after this synaptic dopamine function. This is in agreement time point. with the hypothesis put forth by Melis and colleagues [13], who have proposed that a hypodopa- Imaging studies of cocaine dependence minergic state in addiction is associated with and other neurotransmitters a decreased interest in nondrug-related cues and ex- Despite that cocaine directly affects the serotonin cessive interest in drugs of abuse. system and that altering serotonin transmission in human subjects modulates the subjective effects of Imaging cue-induced craving in cocaine cocaine [31–33], only one SPECT study has been dependence performed to measure the serotonin transporter. Two recent studies have investigated the effect of This study showed that the transporter was in- drug-related cues on 11 C-raclopride binding in co- creased in the brainstem in cocaine abusers who caine dependence [25,26]. Both studies used a video had been abstinent for 3.7Æ3.8 days [34]. The dura- of persons using cocaine compared with a neutral tion of abstinence was brief, and it is not known if video (nature scenes). Volkow and colleagues [25] this elevation is long-lasting. showed a decrease in 11 C-raclopride BP in the dor- Two PET studies of cocaine dependence using the sal striatum (caudate and putamen) following the m-agonist receptor radiotracer 11 C-carfentanil have cocaine video compared with the neutral video. been performed. The first of these imaged co- Wong and colleagues [26] showed a significant caine-dependent subjects after 1 to 4 days of absti- decrease in BP in the left anterior putamen in the nence and again at 4 weeks of abstinence cocaine subjects who craved cocaine, whereas there compared with control subjects [35]. m Receptor was no significant change in cocaine abusers who BP was increased in the anterior cingulate, frontal did not crave cocaine. In both studies, the magni- and temporal cortex, caudate, and thalamus in the tude of 11 C-raclopride displacement correlated cocaine abusers in early abstinence and increased with craving for cocaine. in the cingulate, frontal cortex, caudate, and thala- In addition, Volkow and colleagues [25] showed mus after 4 weeks of abstinence. In a later study, that cue-induced changes in dopamine correlated this same group measured m receptor BP in a group with the severity of addiction, such that greater do- of cocaine abusers at three time points: after 1 day, pamine release in the dorsal striatum correlated 1 week, and 12 weeks of abstinence [36]. The results with higher scores of severity. This finding suggests showed that m receptor BP was increased in the an- that dopamine release in response to a cue corre- terior cingulate, frontal cortex, and temporal cortex lates with craving for drug and might thus correlate after 1 day of abstinence and remained elevated in Imaging the Neurochemistry of Addiction 543 the anterior cingulate and anterior frontal cortex up dependent subjects and healthy control subjects to 12 weeks [36]. In both of these studies, m receptor were performed measuring the striatum as a whole binding positively correlated with self- only and used SPECT rather than PET [41–43]. reports of craving for cocaine, suggesting that mod- These differences in imaging methodology, how- ulation of the endogenous opioid system may play ever, are unlikely to explain the differences in the a role in cocaine addiction and relapse. results. One of these studies followed the alcohol- dependent subjects for 3 months and found that Alcohol dependence the subjects who relapsed within this time frame had higher D2/3 receptor BP compared with those Dopamine D2/3 receptor and alcohol who did not relapse [42]. dependence The studies measuring D2/3 receptor BP described As with cocaine, most radiotracer imaging studies in this review thus have been limited to measuring in alcohol dependence have focused on dopamine the striatum. More recently, high-affinity radio- transmission using the same techniques as those tracers have been used to measure D2/3 receptors described. As shown in Table 1, eight PET and outside the striatum, where the receptor concentra- SPECT studies have measured D2/3 receptor BP in tions are low, such as within the temporal lobe. In alcohol dependence: six showed a decrease in alcohol dependence, only one group has measured D2/3 receptor BP [8,9,37–40] and two showed no extra-striatal D2/3 receptor BP using the high-affin- significant difference between alcohol-dependent ity SPECT radiotracer 123I-epidepride. In their first subjects and healthy control subjects [41,42]. The report, Repo and colleagues [41] reported no differ- studies showing a decrease in D2/3 receptor BP ence in D2/3 BP between alcohol-dependent sub- were performed with PET, and the two earliest stud- jects and healthy control subjects. In a subsequent ies measured the striatum as a whole [8,9]. The reanalysis of these data, Kuikka and colleagues other four PET studies used a high-resolution cam- [43] reported that the alcohol-dependent subjects era and showed a decrease in D2/3 receptors of had lower D2/3 receptor BP values in the left tempo- a similar magnitude to that seen in cocaine depen- ral pole compared with healthy control subjects dence in the caudate [37,38], putamen [37–40], (the reanalysis consisted of measuring the right and ventral striatum [38–40]. The two studies and left temporal poles separately). Although the showing no difference between the alcohol- alcohol-dependent subjects in this study included

Table 1: Comparison of studies investigating D2/3 receptor availability in alcohol dependence

D2/3 receptor BP Withdrawal (alcohol-dependent versus Reference period Radioligand healthy control subjects) Hietala et al, [8] 23Æ62 weeks 11C-raclopride (PET) Decreased in striatum (20%) Volkow et al, [9] 52Æ48 days 11C-raclopride (PET) Decreased in striatum (22%) Repo et al, [41] 1 week to 4 years 123I-epidipride (SPECT) Nonsignificant decreased BP in striatum (5%) Guardia et al, [42] 8 to 10 days 123I-IBZM (SPECT) Nonsignificant decreased BP in striatum (5%) Volkow et al, [37] 6 to 20 weeks 11C-raclopride (PET) Decreased in caudate (14%) and putamen (18%) in early detoxification (6 weeks), decreased in caudate in late detoxification (1–4 months later) Heinz et al, [39] 2 to 4 weeks 18F-desmethoxyfallypride Decreased in the putamen (PET) and ventral striatum Heinz et al, [40] 36Æ22 days 18F-desmethoxyfallypride Decreased in the putamen (PET) and ventral striatum Martinez et al, [38] 14 days 11C-raclopride (PET) Decreased in the caudate (21%), putamen (21%), and ventral striatum (17%)

Abbreviation: BP, binding potential. 544 Martinez et al

subjects who had a wide range of abstinence family history. In addition, the investigators re- (1 week to 4 years), this study did not include ported that higher metabolism in the prefrontal sufficient subjects at the extremes of the range of ab- cortex (measured with 18F-fludeoxyglucose) was stinence to address the question of whether D2/3 associated with higher D2/3 receptor BP in subjects recovers after a prolonged period of abstinence. who had family histories of alcohol dependence Of note, in the study of Volkow and colleagues [19]. Based on these findings, the investigators [37], alcohol-dependent subjects were scanned at postulated that high D2/3 receptor BP is protective 6 weeks and again at 1 to 4 months later, with no for subjects who have a vulnerability for alcohol recovery of striatal D2 receptor BP within this time dependence (ie, family history), and that this pro- frame. No studies have been performed investigat- tective factor is also represented by an increase in ing the recovery of D2/3 receptor BP after longer activation of cortical brain regions that mediate periods of abstinence, and this question remains behaviors such as inhibition control and restraint unanswered. over alcohol consumption. Munro and colleagues, however [46], found no difference in D2/3 receptor Behavioral correlates of low D2/3 receptor BP values between social drinkers with and with- binding potential out a positive family history. This study was simi- Several studies have investigated the behavioral sig- lar to that of Volkow and colleagues [19] with nificance of reduced D2/3 receptor BP in alcohol- regard to PET methods, enrollment numbers, dependent subjects. Heinz and colleagues [39] and subject demographics (eg, age, alcohol in- reported a correlation between low D2/3 receptor take). Although there were some differences in BP in the ventral striatum and activation in the the numbers of alcohol-dependent relatives re- medial prefrontal and anterior cingulate cortex in quired for the two studies, all subjects who had response to alcohol-related cues using functional a positive family in both studies included alco- magnetic resonance imaging (fMR imaging). In ad- hol-dependent fathers, so that it is unlikely that dition, this study showed that low D2/3 receptor this factor alone could explain the discrepancy in binding in the ventral striatum also correlated findings. with a greater craving for alcohol. Heinz and colleagues [40] also showed that craving for alcohol at the time of the PET scan correlated significantly Alcohol dependence and presynaptic with alcohol intake at the 6-month follow-up. dopamine The authors recently demonstrated a significant Presynaptic dopamine function has been studied in inverse correlation between 11 C-raclopride BP and alcohol dependence using three methods: (1) imag- the average daily quantity of alcohol consumed, ing the neuronal uptake of the radiotracer suggesting that a low BP may be associated with 18F-DOPA, which provides a measure of presynaptic greater severity of disease [38]. This finding is in dopamine stores, (2) (1)18F-dihydrotetrabenazine, agreement with studies showing lower D2 receptor which labels the type-2 vesicular monoamine trans- density in the caudate-putamen and nucleus ac- porters of the dopamine vesicles, and (3) 11 C-raclopr- cumbens of alcohol-preferring compared with ide with an amphetamine challenge. Two studies non-alcohol–preferring rats, before exposure to al- performed with 18F-DOPA reported an increase and cohol [44,45]. In addition, Thanos and colleagues no difference between alcohol-dependent subjects [45] showed that adenovirus-induced overexpres- and healthy control subjects [40,47].Tiihonenand sion of the D2 receptor in rats trained to self-admin- colleagues [47] demonstrated an increase in ister alcohol reduced their intake and their 18F-DOPA uptake in the putamen and caudate in preference for alcohol. Together these data suggest alcohol-dependent subjects compared with healthy that the low D2/3 receptor BP observed in alcohol- control subjects, which suggests that alcoholic sub- dependent individuals is associated with greater se- jects have increased presynaptic dopamine function. verity of disease (more craving, greater attention to Alternatively, Heinz and colleagues [40] showed no alcohol-related cues, and higher alcohol consump- difference in 18F-DOPA uptake in alcohol-dependent tion), and may also be associated with an increased subjects who had been abstinent for 36Æ22 days.Low risk for relapse. 18F-DOPA uptake in the putamen, however, corre- As described for cocaine dependence, a recent lated with greater craving for alcohol, suggesting study suggests that increased D2/3 receptor BP that alcohol-dependent subjects who have deficits may be protective against alcohol dependence. in presynaptic dopamine may be more susceptible Volkow and colleagues [19] recently reported to the reinforcing effects of alcohol [40]. that social drinkers who have a strong family his- In a PET study using the radioligand (1)18F-dihy- tory of alcohol dependence had higher D2/3 recep- drotetrabenazine, Gilman and colleagues [48] tor BP compared with those who had no such reported a decrease in striatal type-2 vesicular Imaging the Neurochemistry of Addiction 545 monoamine transporters (VMAT2) in the caudate dependence, and they have reported conflicting re- (6%) and putamen (13%) of alcohol-dependent sults. Heinz and colleagues [51,53] reported a 30% subjects, although this only reached significance reduction of serotonin transporter (SERT) binding in the putamen. Levels of VMAT2 were not specifi- in the midbrain in men but not women who were cally measured in the ventral striatum. These results alcohol-dependent using SPECT and the radiotracer suggest that alcohol dependence is associated with 123I-beta-CIT. SERT binding in men was associated a loss of presynaptic dopamine stores. with lifetime alcohol consumption and severity of The authors recently completed a study using depression [51,53]. In a subsequent study, Heinz 11 C-raclopride and an amphetamine challenge to and colleagues [54] reported a decrease in midbrain investigate dopamine transmission in alcohol de- SERT only in alcoholic subjects who were homozy- pendence [38]. Dopamine transmission was re- gous for the long allele of the promoter of the SERT duced exclusively in the ventral striatum in the gene. alcohol-dependent subjects compared with healthy These studies were conducted with the SPECT control subjects; no differences in the associative radiotracer 123I-beta-CIT, which allows the mea- and sensorimotor striatum were found between surement of SERT in the midbrain, yet this radio- the two groups. In contrast to a study of cocaine de- tracer has a low ratio of specific to nonspecific pendence that showed a reduction in amphet- binding in other brain regions. In contrast, some amine-induced 11 C-raclopride displacement in PET ligands for SERT produce a better signal- each of the striatal subdivisions, blunted dopamine to-noise ratio in other brain regions. The earliest transmission was seen only in the ventral striatum PET ligand that was available to measure the SERT in alcohol dependence. It is thus possible that was 11 C-McN5652. Using this radiotracer, Szabo addiction in general is associated with decreased and colleagues [55] reported a significant decrease dopamine function in the limbic striatum, whereas in the distribution volume (DV) of 11 C-McN5652 stimulant dependence is associated with a more in most brain regions (midbrain, thalamus, amyg- widespread reduction. dala, pons, cingulate, orbitofrontal cortex, and cerebellum) studied in alcohol dependence. This Alcohol dependence and the dopamine outcome measure (DV) differs from BP. BP is a mea- transporter sure of specific binding, whereas the DV is a measure PETand SPECTstudies have been conducted to mea- of specific binding and nonspecific binding com- sure the dopamine transporter in alcohol depen- bined. The region of reference in this study was dence. Tiihonen and colleagues [49] used SPECT the cerebellum, with a significant difference in non- and 123I-beta-CIT to measure DAT in violent alco- specific binding between the alcoholic subjects and holic subjects, nonviolent alcoholic subjects, and healthy control subjects. When the specific binding healthy control subjects. This study showed a signif- of 11 C-McN5652 to SERT alone was measured, the icant reduction in DAT in nonviolent alcoholic sub- only significant difference seen between the alco- jects and an increase in DAT binding in the violent holic subjects and control subjects was in the mid- alcoholic subjects relative to control subjects. A sub- brain [55]. sequent study by Laine and colleagues [50] reported More recently the radiotracer 11 C-DASB has been that alcohol dependence was associated with a de- used to image SERT because of its improved signal- crease in DAT binding only at 1 to 4 days of absti- to-noise ratio [56]. Brown and colleagues [57] nence; the same subjects scanned after 4 weeks of used this radiotracer to investigate alterations in abstinence did not differ from control subjects. Us- SERT in 30 alcohol-dependent subjects (abstinent ing PET and 11 C-d-threo methylphenidate, Volkow 14Æ2 days before scanning) and 18 healthy con- and colleagues [9] reported no difference in alco- trol subjects. Given the high correlation between hol-dependent subjects who had been abstinent alcoholism, aggression, and alterations in seroto- 52Æ48 days in comparison with healthy control nin transmission, the alcohol-dependent subjects subjects. Similarly, Heinz and colleagues [51,52] re- were divided into two groups based on their scores ported no difference in DAT in alcohol-dependent of hostility. No significant difference in SERT BP subjects scanned after 3 to 5 weeks of abstinence. It was found in any brain region (including mid- thus seems that DAT binding may be transiently de- brain, frontal and medial temporal cortex, stria- creased in alcohol withdrawal, but that long-term tum, thalamus, and cerebellum) between the two changes in the dopamine transporter are unlikely alcohol-dependent groups and healthy control to play a key role in alcohol dependence. subjects. Overall the main conclusion from these studies Serotonin and alcohol dependence of SERT in alcohol dependence is that two groups Only three groups have published studies investi- have shown a decrease in midbrain SERT in alco- gating alterations in the serotonin system in alcohol holic subjects, whereas one group has not. The 546 Martinez et al

reason behind this discrepancy is not clear. The recently Staley and colleagues [64] studied 23 alco- studies by Heinz and colleagues [54] suggest that hol-dependent men and 15 healthy control subjects this decrease may be present only in men who using the SPECT radiotracer 123I-iomazenil. The al- have a given serotonin promoter genotype. The cohol-dependent subjects were scanned at two time study of Szabo and colleagues [55] did not include points: at 1 week and at 4 weeks of abstinence. At 1 sufficient female subjects for comparison, and the week of abstinence, 123I-iomazenil uptake was in- study of Brown and colleagues [57] reported no ef- creased in the frontal, cingulate, temporal, insular, fect of gender on SERT binding. No studies outside parietal, and occipital cortices in the alcohol-depen- that of Heinz and colleagues have looked at the ef- dent subjects who were nonsmokers compared fect of genotype. In addition, the duration of absti- with nonsmoking control subjects and compared nence was varied among studies, from 2 weeks to with the alcohol-dependent subjects who were 27 years. For the duration of abstinence to explain smokers. No differences were seen between any of the difference in the study results, it is thus neces- the groups at 4 weeks. sary to postulate that SERT availability early in ab- The reasons behind these discrepancies are not stinence does not differ from that found in entirely clear, but likely relate to differences in sam- control subjects, but then significantly decreases ple size and duration of abstinence. Both studies and remains decreased for many years. that showed no difference between alcohol-dependent and control subjects included small numbers Measures of GABA in alcohol dependence of subjects, and only the study of Staley and col- Although imaging studies have largely focused on leagues scanned subjects after a short interval of ab- a range of neurotransmitter systems in alcohol de- stinence. Taken together, these studies thus suggest pendence, GABAA receptors probably play a more that GABAA benzodiazepine receptor binding shifts direct role in ethanol intoxication and withdrawal. across the different stages of alcohol dependence To date, the PET and SPECT radiotracers available to and may be increased in very early abstinence, but image the GABAA receptor all label the benzodiaze- more prolonged abstinence may be associated pine site (11 C-flumazenil for PET and 123I-iomaze- with a decrease in binding [65]. This hypothesis is nil for SPECT). Seven studies have examined this supported by the finding of Staley and colleagues parameter in alcohol dependence. Three of these [64] showing that GABAA receptor binding was pos- studies have shown that alcohol dependence is itively associated with severity of withdrawal and associated with a decrease in the GABAA receptor days since the last alcoholic drink. One must keep binding. Gilman and colleagues [58] reported a de- in mind, however, that the initial increase seems crease in the DV of 11 C-flumazenil in the medial to be present only in alcohol-dependent non- frontal lobe and cingulate of nine alcohol-depen- smokers (who tend to be in the minority in this dis- dent subjects, and a decrease in these same regions order) and the decrease in binding associated with and the cerebellum in eight alcoholic subjects who a longer duration of abstinence may be limited to had cerebellar degeneration. Using 123I-iomazenil, men. Abi-Dargham and colleagues [59] reported a lower There are also some methodologic issues with DV in the frontal cortex, anterior cingulate, and cer- imaging studies of the GABAA receptor in alcohol ebellum in alcoholic patients abstinent 1 to 6 dependence. The first issue is that alcohol depen- months. Finally, Lingford-Hughes and colleagues dence is associated with a known decrease in corti- [60] reported a decrease in GABAA receptor binding cal gray matter [66], and a loss of gray matter could in the frontal, parietal, and temporal cortices in al- theoretically result in a decrease in the GABAA ra- cohol-dependent men compared with healthy con- diotracer uptake, even without a true difference in 123 trol subjects using the radiotracer I-iomazenil. GABAA receptor availability. For instance, the study Two studies have reported no difference in GA- of Abi-Dargham and colleagues [59] reported a de- BAA receptor binding between alcohol-dependent crease in the size of the cortical regions studied, and subjects and healthy control subjects [61,62]. One further showed that the group of alcoholic subjects of these included a small number of subjects [61] had significantly reduced gray matter in the pre- and the other study included only women alco- frontal cortex, anterior cingulate, and cerebellum. hol-dependent subjects [62]. An additional two These were also the brain regions that had a signifi- studies have shown that alcohol dependence is as- cant decrease in GABAA receptor availability. sociated with an increase in GABAA receptor bind- Lingford-Hughes and colleagues, however [60], ing. Jalan and colleagues [63] reported an increase reported a decrease in binding of 123I-iomazenil in 11 C-flumazenil binding in the cortex, cerebellum, in cortical regions in which gray matter atrophy and the basal ganglia in alcohol-dependent subjects was absent, suggesting that gray matter loss alone who had cirrhosis and hepatic encephalopathy does not explain the decrease in GABAA receptor compared with healthy control subjects. More availability. Imaging the Neurochemistry of Addiction 547

In addition, because there are no regions of refer- subjects compared with control subjects using 11 ence for the GABAA receptor, the outcome measure C-carfentanil. This was a preliminary study, how- used in these studies is the DV, which reflects spe- ever, that included three heroin-dependent sub- cific binding and nonspecific binding. In other jects. Kling and colleagues [71] performed a study words, none of the studies controlled for possible using 18F-cyclofoxy, an opioid antagonist that labels differences in nonspecific radiotracer binding m-and k-opioid receptors, in methadone main- between alcohol-dependent subjects and healthy tained heroin-dependent subjects. This study control subjects. showed a 19% to 32% decrease in receptor availability in the thalamus, amygdala, caudate, anterior Opioids and alcohol dependence cingulate, and putamen compared with healthy Recent studies investigating the efficacy of the m re- control subjects. Because the subjects were taking ceptor antagonist as an effective treatment for alco- methadone at the time of the scan, these results sug- hol dependence have shown mixed results [67]. The gest that methadone treatment is associated with m receptor is thus an interesting target for imaging a low occupancy of opioid receptors. The heroin- in this disorder. Two PET imaging studies have dependent subjects, however, were not scanned at been preformed using the m receptor selective radio- baseline (off medication), and so the actual occu- tracer 11 C-carfentanil in alcohol dependence. pancy is not known. In other words, occupancy is Bencherif and colleagues [68] studied eight alco- a comparison of radiotracer binding in the baseline hol-dependent men after 4 days of abstinence and scan with radiotracer binding following drug ad- reported a decrease in m receptor BP in the prefron- ministration. Because the drug binds to the same re- tal cortex compared with control subjects. This ceptor, the post-medication scan shows lower group of investigators also reported that this finding radiotracer uptake. In this way PET can be used to was associated with a higher score for craving. No measure the percent of receptors occupied by a med- BP differences between the two groups were seen ication at a therapeutic dose. In the study of Kling in the parietal, temporal, or orbitofrontal cortex. and colleagues [71], it was assumed that the base- In contrast, a recent study by Heinz and colleagues line (off methadone) receptor availability in the [69] included 25 alcohol-dependent subjects who heroin-dependent subjects was similar to that of were scanned at 1 to 3 weeks and 5 to 7 weeks of the control subjects. If heroin dependence is associ- abstinence and 10 matched healthy control sub- ated with a baseline increase in opioid receptor BP, jects. In this study, alcohol dependence was associ- however (as suggested by Zubieta and colleagues), ated with an increase in m receptor BP in the ventral the occupancy of methadone in this study will be striatum of the alcohol-dependent group at both underestimated. time points compared with the control subjects, In a subsequent study Greenwald and colleagues and no difference in BP was seen in the other brain [72] investigated the occupancy of therapeutic regions studied (caudate, putamen, prefrontal, and doses of buprenorphine, a partial m-agonist and parietal cortex). In addition, higher m receptor BP k-antagonist, which has been successfully used to correlated with a greater craving for alcohol. These treat heroin addiction. In this study, five heroin- studies thus showed opposing findings in different dependent subjects were scanned after maintenance brain regions. Although both studies used PET, on 2-, 16-, and 32-mg dosages of buprenorphine, however, there were significant differences in the which resulted in whole-brain m receptor occupan- methods. The study of Bencherif and colleagues cies of 41%, 80%, and 84%, respectively. In addi- [68] included a smaller number of subjects, the tion, subjects were given a hydromorphone control subjects were not enrolled specifically for challenge, and increased receptor occupancy corre- this study (drawn from a pre-existing database), lated with a decrease in its subjective effects. This and this group did not report on m receptor avail- study suggests that high m receptor occupancy is ability in the ventral striatum. As yet, no overall reached with therapeutic doses of buprenorphine, conclusion can thus be made from these imaging and that this level of occupancy is needed to atten- studies. uate the subjective effects of other opioids. In a follow-up study, this group showed that 50% to 60% occupancy of the m receptor with buprenorphine Heroin dependence is effective at blocking symptoms of withdrawal in To date, only four PET studies have been performed opiate-dependent subjects [73]. measuring neurochemistry in heroin addiction. Only one study has been performed to measure Two of these imaged opioid receptors. Zubieta the D2/3 receptor BP in heroin dependence. Wang and colleagues [70] reported an increase in m recep- and colleagues [7] performed PET imaging with tor availability in the ventral striatum, inferofrontal 11 C-raclopride in 11 heroin-dependent subjects be- cortex, and anterior cingulate in heroin-dependent fore and after naloxone-precipitated withdrawal. 548 Martinez et al

D2/3 receptor BP in both conditions was compared also found that decreased DAT availability corre- with healthy control subjects. Heroin dependence lated with years of abuse and impairment in motor was associated with an 18% decrease in D2 receptor and memory tasks. Despite that these studies used availability, and there was no change in 11 C-raclopr- a wide range of methods (ie, different radiotracers, ide binding following withdrawal. As described, duration of abstinence, and numbers of subjects en- this reduction in D2/3 receptor BP is of the same rolled), they are largely in agreement, suggesting magnitude of that seen in other addictions, that the decrease in the dopamine transporter in although the behavioral significance of this de- methamphetamine abuse is a robust finding. crease in this population is not known. These ata are important because it is believed that methamphetamine may be toxic to dopamine neurons. PET and postmortem studies in nonhuman Methamphetamine abuse primates have shown that methamphetamine expo- In addition to cocaine, alcohol, and heroin-depen- sure results in a decrease in the DAT in addition to dent subjects, low D2/3 receptor BP values have other markers of dopaminergic neurons [79–81]. been reported in methamphetamine abusers. Vol- There is evidence, however, that this phenomenon kow and colleagues [10] showed a decrease of is reversible. Chou and colleagues [77] used SPECT 16% in D2/3 receptor BP in the putamen and of to scan five methamphetamine abusers in acute 10% in the caudate in methamphetamine abusers withdrawal and again after 2 weeks of abstinence compared with control subjects using 11 C-raclopr- and reported partial recovery of DAT binding. Two ide. In this study, D2/3 receptor BP was not specifi- studies in nonhuman primates suggest that a reduc- cally measured in the ventral striatum. tion in markers of dopamine neuron viability (DAT 18 Furthermore, low D2/3 receptor BP was associated and F-DOPA uptake) seems to be reversible after with a decrease in glucose metabolic rate in the or- prolonged abstinence [80,82]. Similarly a postmor- bitofrontal cortex in the methamphetamine abusers tem study demonstrated that although there was and in the control subjects. This finding is similar to a reduction in DAT binding in the human striatum, a study showing that low D2/3 receptor BP was asso- there was no decrease in measures of DOPA-decar- ciated with low orbitofrontal activity in cocaine boxylase and the vesicular monoamine transporter, abusers [4]. Together these findings suggest that suggesting a loss of DAT without frank cell death of dysregulation of these brain regions may mediate the dopaminergic neurons [83]. Nevertheless, PET the loss of inhibitory control over compulsive studies in human subjects raise the concern of neu- drug-taking in these addictions. rotoxicity, and these studies demonstrated a reduc- Alterations in DAT also have been demonstrated tion in DAT even in subjects who reported in methamphetamine dependence. Four notable a significant time of abstinence. imaging studies demonstrated significant decreases in DAT binding using PET or SPECT in metham- Methylenedioxymethamphetamine phetamine abusers compared with control subjects. (Ecstasy) abuse McCann and colleagues [75] reported on six methamphetamine abusers and four methcathinone To date most imaging studies investigating the users using 11 C-WIN-35,428, a selective DAT radio- effects of 3,4-methylenedioxymethamphetamine ligand. The methamphetamine abusers had a de- (MDMA) on neurochemistry have focused on the crease in DAT availability of approximately 25% serotonin system. Studies in animals have demon- in the putamen and caudate compared with control strated the toxicity of MDMA on serotonergic neu- subjects, and a similar reduction was seen in the rons and further have reported that the toxic methcathinone abusers [75]. Using the same radio- effects seem to largely occur at the synaptic termi- tracer, Iyo and colleagues [76] reported decreased nals, leaving the cell bodies of the midbrain intact DAT binding of 20% in the caudate/putamen [84]. Because SERT sites are located on the synaptic (20%) and ventral striatum (26%) in 11 metham- terminal of the neurons, imaging of this transporter phetamine abusers and nine control subjects. A can be used as a marker of neuronal integrity of the SPECT study by Chou and colleagues [77] showed synaptic terminals. SERT can be imaged with PET or DAT binding reductions of a similar magnitude. SPECT. The SPECT radiotracer 123I-beta-CIT labels Volkow and colleagues [78] studied 15 metham- the serotonin transporter in the midbrain (in the phetamine abusers who had 2 weeks of monitored striatum this radiotracer labels the dopamine trans- abstinence before scanning with 11 C-d-threo-meth- porter) [85,86]. The earliest PET radiotracer avail- ylphenidate. Compared with control subjects, the able to label the SERT was 11 C-(1)McN 5652, methamphetamine abusers demonstrated a de- which was followed by 11 C-DASB, a ligand with crease in DAT availability of 28% in the caudate a higher ratio of specific-to-nonspecific binding and 21% in the putamen [78]. The investigators and shorter scanning time compared with Imaging the Neurochemistry of Addiction 549

11 C-(1)McN 5652 [87]. Although the SPECT radio- and 11 C-DASB, in 23 MDMA users and matched tracer only allows reliable measurement of the SERT healthy control subjects. A decrease in the DV mea- in the raphe nucleus of the midbrain (and to some sure was seen in the amygdala, thalamus, dorsolat- extent the thalamus), the PET radiotracers allow eral prefrontal cortex, orbitofrontal, cingulate, measurement of the SERT in midbrain, thalamus, parietal, temporal, and occipital cortex with both striatum, and cortex (medial temporal lobe and radiotracers. Additional decreases in SERT binding cingulate). were seen in the hippocampus and striatum with Five studies in human MDMA users have been 11 C-McN 5652 but not with 11 C-DASB. Buchert performed measuring SERT in active MDMA users, and colleagues [93] published a large PET study in- as summarized in Table 2. The first of these including 117 subjects (30 current users, 29 ex-users, cluded 14 MDMA users with a lifetime use of 29 MDMA naı¨ve drug users, and 29 drug-naı¨ve more than 25 tablets [74]. Subjects were scanned control subjects). In this dataset, SERT BP was sig- with 11 C-(1)McN 5652 and 11 C-(-)McN 5652 nificantly reduced in current MDMA users in the (the inactive enantiomer of the radiotracer) to midbrain and thalamus, left caudate, hippocam- measure nonspecific binding. The results of this pus, posterior cingulate, and occipital and tempo- study showed that MDMA abuse was associated ral lobes compared with the other groups. No with a decrease in radiotracer distribution volumes difference in BP was seen in any region between in the midbrain, thalamus, striatum, pons, cingu- the ex-users and comparison subjects. Notably late, frontal, occipital, and parietal cortex in addi- this decrease was greater in women subjects relation to the cerebellum. These data, however, tive to men. Methodologic concerns included: (1) showed an unusually high variability, and an un- although the average duration of abstinence in conventional approach to the statistical analysis the ex-users was 1.4 years, some had had as little was used to interpret the results [84]. Two subse- as 29 days of cessation (the current users ranged quent studies used SPECT and the radiotracer from 4 to 60 days since last use, so that there 123I-beta-CIT [88,89] in MDMA abusers. The first was some overlap in time of abstinence between of these included a small sample (n510) of men these two groups), and (2) some subjects were MDMA users with a lifetime use of more than 50 scanned after a very short period of abstinence. tablets and matched healthy control subjects Reanalysis of the data, however, showed that exclu- [89]. The MDMA abusers demonstrated a signifi- sion of the subjects who had been abstinent less cant decrease in 123I-beta-CIT uptake in the occip- that 14 days did not significantly change the out- ital, calcarine, and posterior cingulate cortex and come: a decrease in BP was still seen in all the no difference in the midbrain and thalamus com- regions mentioned except for the caudate and pos- pared with control subjects. The second SPECT terior gyrus. study included a much larger sample divided Subsequently Buchert and colleagues [94] re- into three groups: heavy users (reported lifetime scanned 24 of the MDMA users 1 year later; 15 use greater than 50 tablets), moderate users (re- were current MDMA users and 9 were former ported lifetime use less than 50 tablets), and ex- MDMA users. This study showed that current users in addition to control subjects [88]. This MDMA users had an increase in midbrain SERT, study showed a significant decrease in 123I-beta- which the investigators attributed to a decrease in CIT specific binding in the midbrain and thalamus the magnitude of MDMA use over the year-long in- in addition to the frontal, temporal, occipital, and terval. SERT binding also increased in the thalamus parietal cortex exclusively in heavy women users. and striatum, although this did not reach signifi- There were some methodologic concerns, however, cance. In the ex-MDMA users, however, an increase including the use of 123I-beta-CIT to measure SERT in 11 C-McN5652 binding was seen in the thala- in the cortex. At equilibrium, the binding of this mus, which the investigators hypothesized was radiotracer in the cortical regions is minimally caused by an overshoot in SERT following long- higher than that seen in the region used to esti- term abstinence. Although this study suggests mate nonspecific binding (reference region). The that the MDMA-induced decrease in SERT may signal of the 123I-beta-CIT binding in the cortical be reversible, it should be kept in mind that the regions that represent SERT is thus too low to pro- test–retest reliability of these measures of SERT vide reliable estimates [86,90,91]. The most robust with PET over the course of a year has not been finding from these studies therefore is the decrease shown. in SERT in the midbrain in heavy women MDMA Together the results of these imaging studies sug- users. gest that MDMA use is associated with a decrease Subsequently two additional PET studies have in SERT binding, and that this decrease is more sig- been performed. McCann and colleagues [92] nificant in women users. In addition, the studies studied two SERT radiotracers, 11 C-(1)McN 5652 that included ex-users showed no difference 550 Martinez et al

Table 2: Comparison of studies measuring the serotonin transporter in MDMA abuse Correlation Extent of of radiotracer use (MDMA binding with Study Sample size users) Abstinence SERT binding MDMA use McCann et al, 14 (9 M/5 F) >50 tablets 19 weeks Decreased in Inverse [74] PET (1/À) current users; (range, all brain correlation [11C]McN5652 15 MDMA 3–147) regions studied with lifetime naı¨ve (including use (9 M/6 F) midbrain) in males and females Buchert et al, 30 current >50 tablets 4–60 days Decreased in Inverse [93] PET users, 29 (MDMA midbrain, correlation [11C]McN5652 ex-users, 29 users), thalamus, with mean MDMA naı¨ve 1.4 years and cortical MDMA drug users, ex-users regions dose and 29 drug naı¨ve control subjects McCann et al, MDMA: 23 >50 tablets 4.7Æ8.7 Decreased in Inverse [92] PET (13 M/10 F) months amygdala, correlation [11C]DASB and Control: 19 thalamus, and with time [11C]McN5652 (8 M/11 F) cortical regions of abstinence (MDMA naı¨ve measured. No from MDMA with minimal decrease in other drug midbrain. exposure) Semple et al, MDMA: 10 >50 tablets, 18Æ8 Signiﬁcant Inverse [89] SPECT (10 M) days decrease in correlation [123I]beta-CIT HC: 10 some posterior with time of (10 M) cortical abstinence regions; no from MDMA difference in midbrain Reneman Heavy Heavy: >50 21 days Decreased in Inverse et al, [88] users: tablets female heavy correlation SPECT users only between [123I]beta-CIT (midbrain and lifetime MDMA cortical use and SERT in regions); no female heavy difference users (not in in other groups other groups) 23 (12M/11F) Light: <50 tablets Moderate Ex: >50 tablets, users: but not in the last year 15 (9M/6F) Ex-users: 16 (8M/8F) HC: 15 (7M/8F)

Abbreviations: F, female; M, male; MDMA, methylenedioxymethamphetamine; SERT, serotonin transporter.

between this group and control subjects, suggest- who decreased their MDMA use. In addition, sev- ing that the decrease in SERT is reversible with ces- eral studies showed a correlation between SERT sation of use. This hypothesis is supported by the and duration of abstinence, also supporting that follow-up study of Buchert and colleagues [94], the MDMA-associated decrease in SERT may be re- which showed that SERT increased in subjects versible with time. Imaging the Neurochemistry of Addiction 551

Although MDMA also affects the dopamine has been observed during abstinence, and may transmission, only two studies have looked at do- represent a mixture of vulnerability and toxicity fac- pamine transmission and both measured the dopa- tors, although more research is needed to better mine transporter. Reneman and colleagues [95] characterize the relative contribution of these vari- used 123I-beta-CIT to measure striatal DAT in ables to the development of addiction. MDMA abusers who used the drug alone and Methamphetamine use has been consistently as- MDMA users who also used amphetamine. DAT sociated with low DAT levels in the striatum that binding was lower in the MDMA-amphetamine seem to be long-lasting, suggesting an additional group relative to control subjects but higher in the component of neurotoxicity that may not be shared MDMA-only group compared with the control sub- with other drugs of abuse. Alternatively, although jects. Using the same radiotracer, Semple and col- ecstasy use may be associated with low SERT in dif- leagues [89] showed no differences in DAT ferent brain regions, this decrease may be reversible binding between MDMA users and control subjects with cessation of use. in the striatum. Imaging of other neurotransmitter systems, such as the GABA, serotonergic, and opiate systems, has contributed some interesting insights into the path- Hallucinogens ophysiology of alcohol abuse and dependence but There is a paucity of literature investigating alter- has not yet yielded a consistent picture. ations in neurochemistry in hallucinogen abusers. Finally, preclinical research has suggested a prom- One study used the PET radiotracer 11 C-NNC 112 inent role for the glutamatergic system in the estab- to measure D1 receptor BP in chronic ketamine lishment of compulsive addictive behaviors by users and reported a 25% increase in D1 receptor strengthening certain frontostriatal synapses that availability in the dorsolateral prefrontal cortex may underlie the perseverative aspects of drug- compared with healthy control subjects [96].No taking behaviors. No tracers are available at this significant differences were seen in other cortical re- point to explore this system in the human brain, gions or the striatum. In addition, the increase in and much progress in ligand development is D1 receptor BP correlated with the number of vials needed to provide the tools necessary for such in- of ketamine used per week. This finding is consis- vestigations. In addition, some drugs of abuse tent with the animal literature suggesting that re- have not been studied with radioligand PET imag- peated exposure to N-methyl-D-aspartic acid ing, such as marijuana, prescription drugs, inhal- (NMDA) antagonists leads to reduced prefrontal ants, and most of the hallucinogens. So although dopaminergic function and a compensatory upre- much work has been done, we are at the beginning gulation of the D1 receptor, which is the main do- of an exciting phase in the field in which exponen- paminergic receptor in the cortex [97]. Because tial progress will be made with the availability of decreased dopamine transmission in the dorsolat- a wider range of radiotracers for different molecular eral prefrontal cortex is associated with deficits in targets implicated in addiction. We need to better working memory and executive function, neurocog- understand the pathophysiology and the longitudi- nitive function was examined in these subjects. No nal course of addiction, with studies targeting difference was seen in tests of working memory and different phases of addiction taking into account executive function between the ketamine users and particular genetic risk factors to develop better pre- healthy control subjects, and no correlation was ventive and therapeutic interventions. seen with D1 receptor BP. References Summary [1] Laruelle M. Imaging synaptic neurotransmission There is convergence from multiple lines of evi- with in vivo binding competition techniques: dence that addiction is associated with low D2/3 a critical review. J Cereb Blood Flow Metab transmission in the striatum, including reduced 2000;20(3):423–51. D2/3 receptors, dopamine release, and dopamine [2] Logan J, Fowler JS, Dewey SL, et al. A consider- synthesis, all in the presence of normal DAT, and ation of the dopamine D2 receptor monomer- furthermore that this decrease may predominantly dimer equilibrium and the anomalous binding properties of the dopamine D2 receptor ligand, affect the ventral striatum. This trait seems to be N-methyl spiperone. J Neural Transm 2001; common to multiple drugs of abuse but has been 108(3):279–86. most extensively studied and documented thus far [3] Volkow ND, Fowler JS, Wolf AP, et al. Effects of for cocaine and alcohol. Low dopamine transmis- chronic cocaine abuse on postsynaptic dopa- sion stems from presynaptic and postsynaptic mine receptors. Am J Psychiatry 1990;147(6): factors, is present in the chronic stages of addiction, 719–24. 552 Martinez et al

[4] Volkow ND, Fowler JS, Wang GJ, et al. Decreased markedly blunted in cocaine dependence and dopamine D2 receptor availability is associated predictive of the choice to self administer with reduced frontal metabolism in cocaine cocaine. Am J Psychiatry 2007;164(4):622–9. abusers. Synapse 1993;14(2):169–77. [19] Volkow ND, Wang GJ, Begleiter H, et al. High [5] Volkow ND, Wang GJ, Fowler JS, et al. Decreased levels of dopamine D2 receptors in unaffected striatal dopaminergic responsiveness in detoxi- members of alcoholic families: possible protec- fied cocaine-dependent subjects. Nature 1997; tive factors. Arch Gen Psychiatry 2006;63(9): 386:830–3. 999–1008. [6] Martinez D, Broft A, Foltin RW, et al. Cocaine de- [20] Volkow ND, Wang GJ, Fowler JS, et al. Prediction pendence and D2 receptor availability in the of reinforcing responses to psychostimulants in functional subdivisions of the striatum: relation- humans by brain dopamine D2 receptor levels. ship with cocaine-seeking behavior. Neuropsy- Am J Psychiatry 1999;156(9):1440–3. chopharmacology 2004;29(6):1190–202. [21] Volkow ND, Wang GJ, Fowler JS, et al. Brain DA [7] Wang GJ, Volkow ND, Fowler JS, et al. Dopamine D2 receptors predict reinforcing effects of stimu- D2 receptor availability in opiate-dependent sub- lants in humans: replication study. Synapse jects before and after naloxone-precipitated with- 2002;46(2):79–82. drawal. Neuropsychopharmacology 1997;16(2): [22] Self DW, Barnhart WJ, Lehman DA, et al. Oppo- 174–82. site modulation of cocaine-seeking behavior by [8] Hietala J, West C, Syvalahti E, et al. Striatal D2 D1- and D2-like dopamine receptor agonists. dopamine receptor binding characteristics in Science 1996;271(5255):1586–9. vivo in patients with alcohol dependence. Psy- [23] Khroyan TV, Barrett-Larimore RL, Rowlett JK, chopharmacology (Berl) 1994;116(3):285–90. et al. Dopamine D1- and D2-like receptor mech- [9] Volkow ND, Wang GJ, Fowler JS, et al. Decreases anisms in relapse to cocaine-seeking behavior: in dopamine receptors but not in dopamine effects of selective antagonists and agonists. transporters in alcoholics. Alcohol Clin Exp Res J Pharmacol Exp Ther 2000;294(2):680–7. 1996;20(9):1594–8. [24] Shaham Y, Shalev U, Lu L, et al. The reinstate- [10] Volkow ND, Chang L, Wang GJ, et al. Low level ment model of drug relapse: history, methodol- of brain dopamine D2 receptors in methamphet- ogy and major findings. Psychopharmacology amine abusers: association with metabolism in (Berl) 2003;168(1–2):3–20. the orbitofrontal cortex. Am J Psychiatry 2001; [25] Volkow ND, Wang GJ, Telang F, et al. Cocaine 158(12):2015–21. cues and dopamine in dorsal striatum: mecha- [11] Wang GJ, Volkow ND, Logan J, et al. Brain dopa- nism of craving in cocaine addiction. J Neurosci mine and obesity. Lancet 2001;357(9253): 2006;26(24):6583–8. 354–7. [26] Wong DF, Kuwabara H, Schretlen DJ, et al. In- [12] Volkow ND, Fowler JS, Wang GJ. Role of dopa- creased occupancy of dopamine receptors in mine in drug reinforcement and addiction in human striatum during cue-elicited cocaine crav- humans: results from imaging studies. Behav ing. Neuropsychopharmacology 2006;31(12): Pharmacol 2002;13(5–6):355–66. 2716–27. [13] Melis M, Spiga S, Diana M. The dopamine hy- [27] Voorn P, Vanderschuren LJ, Groenewegen HJ, pothesis of drug addiction: hypodopaminergic et al. Putting a spin on the dorsal–ventral divide state. Int Rev Neurobiol 2005;63:101–54. of the striatum. Trends Neurosci 2004;27(8): [14] Malison RT, Mechanic KY, Klummp H, et al. 468–74. Reduced amphetamine-stimulated dopamine [28] Malison RT, Best SE, van Dyck CH, et al. Elevated release in cocaine addicts as measured by striatal dopamine transporters during acute co- [123I]IBZM SPECT. J Nucl Med 1999;40(Suppl 5): caine abstinence as measured by [123I] beta-CIT 110P. SPECT. Am J Psychiatry 1998;155(6):832–4. [15] Wu JC, Bell K, Najafi A, et al. Decreasing striatal [29] Volkow ND, Wang GJ, Fowler JS, et al. Cocaine 6-FDOPA uptake with increasing duration of uptake is decreased in the brain of detoxified co- cocaine withdrawal. Neuropsychopharmacology caine abusers. Neuropsychopharmacology 1996; 1997;17(6):402–9. 14(3):159–68. [16] Mawlawi O, Martinez D, Slifstein M, et al. Imag- [30] Wang GJ, Volkow ND, Fowler JS, et al. Co- ing human mesolimbic dopamine transmission caine abusers do not show loss of dopamine with PET: I. Accuracy and precision of D2 transporters with age. Life Sci 1997;61(11): parameter measurements in the ventral stria- 1059–65. tum. J Cereb Blood Flow Metab 2001;21(9): [31] Aronson SC, Black JE, McDougle CJ, et al. Seroto- 1034–57. nergic mechanisms of cocaine effects in humans. [17] Drevets WC, Gautier C, Price JC, et al. Amphet- Psychopharmacology (Berl) 1995;119(2): amine-induced dopamine release in human ven- 179–85. tral striatum correlates with euphoria. Biol [32] Satel SL, Krystal JH, Delgado PL, et al. Trypto- Psychiatry 2001;49(2):81–96. phan depletion and attenuation of cue-induced [18] Martinez D, Narendran R, Foltin RW, et al. craving for cocaine. Am J Psychiatry 1995; Amphetamine-induced dopamine release is 152(5):778–83. Imaging the Neurochemistry of Addiction 553

[33] Walsh SL, Preston KL, Sullivan JT, et al. Fluoxe- alcoholism. Alcohol Clin Exp Res 2006;30(7): tine alters the effects of intravenous cocaine in 1143–51. humans. J Clin Psychopharmacol 1994;14(6): [47] Tiihonen J, Vilkman H, Rasanen P, et al. Striatal 396–407. presynaptic dopamine function in type 1 alco- [34] Jacobsen LK, Staley JK, Malison RT, et al. Ele- holics measured with positron emission tomog- vated central serotonin transporter binding raphy. Mol Psychiatry 1998;3(2):156–61. availability in acutely abstinent cocaine-depen- [48] Gilman S, Koeppe RA, Adams KM, et al. De- dent patients. Am J Psychiatry 2000;157(7): creased striatal monoaminergic terminals in 1134–40. severe chronic alcoholism demonstrated with [35] Zubieta JK, Gorelick DA, Stauffer R, et al. In- (1)[11 C]dihydrotetrabenazine and positron creased m opioid receptor binding detected by emission tomography. Ann Neurol 1998;44(3): PET in cocaine-dependent men is associated 326–33. with cocaine craving. Nat Med 1996;2(11): [49] Tiihonen J, Kuikka J, Bergstro¨m K, et al. Altered 1225–9. striatal dopamine re-uptake site densities in [36] Gorelick DA, Kim YK, Bencherif B, et al. Imaging habitually violent and non-violent alcoholics. brain m-opioid receptors in abstinent cocaine Nat Med 1995;1(7):654–7. users: time course and relation to cocaine crav- [50] Laine TP, Ahonen A, Rasanen P, et al. Dopamine ing. Biol Psychiatry 2005;57(12):1573–82. transporter availability and depressive symptoms [37] Volkow ND, Wang GJ, Maynard L, et al. Effects of during alcohol withdrawal. Psychiatry Res 1999; alcohol detoxification on dopamine D2 recep- 90(3):153–7. tors in alcoholics: a preliminary study. Psychiatry [51] Heinz A, Ragan P, Jones DW, et al. Reduced cen- Res 2002;116(3):163–72. tral serotonin transporters in alcoholism. Am [38] Martinez D, Gil R, Slifstein M, et al. Alcohol de- J Psychiatry 1998;155(11):1544–9. pendence is associated with blunted dopamine [52] Heinz A, Goldman D, Jones DW, et al. Genotype transmission in the ventral striatum. Biol Psychi- influences in vivo dopamine transporter avail- atry 2005;58(10):779–86. ability in human striatum. Neuropsychopharma- [39] Heinz A, Siessmeier T, Wrase J, et al. Correlation cology 2000;22(2):133–9. between dopamine D(2) receptors in the ventral [53] Heinz A, Jones DW, Bissette G, et al. Relation- striatum and central processing of alcohol cues ship between cortisol and serotonin metabolites and craving. Am J Psychiatry 2004;161(10): and transporters in alcoholism. [correction of al- 1783–9. coholism]. Pharmacopsychiatry 2002;35(4): [40] Heinz A, Siessmeier T, Wrase J, et al. Correlation 127–34. of alcohol craving with striatal dopamine syn- [54] Heinz A, Jones DW, Mazzanti C, et al. A rela- thesis capacity and D2/3 receptor availability: tionship between serotonin transporter geno- a combined [18F]DOPA and [18F]DMFP PET type and in vivo protein expression and study in detoxified alcoholic patients. Am J Psy- alcohol neurotoxicity. Biol Psychiatry 2000; chiatry 2005;162(8):1515–20. 47(7):643–9. [41] Repo E, Kuikka JT, Bergstrom KA, et al. Dopa- [55] Szabo Z, Owonikoko T, Peyrot M, et al. Positron mine transporter and D2-receptor density in emission tomography imaging of the serotonin late-onset alcoholism. Psychopharmacology transporter in subjects with a history of alcohol- (Berl) 1999;147(3):314–8. ism. Biol Psychiatry 2004;55(7):766–71. [42] Guardia J, Catafau AM, Batlle F, et al. Striatal [56] Wilson AA, Ginovart N, Hussey D, et al. In vitro dopaminergic D(2) receptor density measured and in vivo characterisation of [11 C]-DASB: by [(123)I]iodobenzamide SPECT in the predic- a probe for in vivo measurements of the seroto- tion of treatment outcome of alcohol-depen- nin transporter by positron emission tomogra- dent patients. Am J Psychiatry 2000;157(1): phy. Nucl Med Biol 2002;29(5):509–15. 127–9. [57] Brown AK, George DT, Fujita M, et al. PET [43] Kuikka JT, Repo E, Bergstrom KA, et al. Specific [(11 )C]DASB imaging of serotonin transporters binding and laterality of human extrastriatal in patients with alcoholism. Alcohol Clin Exp dopamine D2/D3 receptors in late onset type 1 Res 2007;31(1):28–32. alcoholic patients. Neurosci Lett 2000;292(1): [58] Gilman S, Koeppe RA, Adams K, et al. Positron 57–9. emission tomographic studies of cerebral benzo- [44] McBride WJ, Chernet E, Dyr W, et al. Densities of diazepine-receptor binding in chronic alco- dopamine D2 receptors are reduced in CNS holics. Ann Neurol 1996;40:163–71. regions of alcohol-preferring P rats. Alcohol [59] Abi-Dargham A, Krystal JH, Anjilvel S, et al. Al- 1993;10(5):387–90. terations of benzodiazepine receptors in type II [45] Thanos PK, Volkow ND, Freimuth P, et al. Over- alcoholic subjects measured with SPECT and expression of dopamine D2 receptors reduces [I-123]iomazenil. Am J Psychiatry 1998;155(11): alcohol self-administration. J Neurochem 2001; 1550–5. 78(5):1094–103. [60] Lingford-Hughes AR, Acton PD, Gacinovic S, [46] Munro CA, McCaul ME, Oswald LM, et al. Stria- et al. Reduced levels of GABA-benzodiazepine tal dopamine release and family history of receptor in alcohol dependency in the absence 554 Martinez et al

of grey matter atrophy. Br J Psychiatry 1998;173: [72] Greenwald MK, Johanson CE, Moody DE, et al. 116–22. Effects of buprenorphine maintenance dose on [61] Litton JE, Neiman J, Pauli S, et al. PET analysis of m-opioid receptor availability, plasma concentra- [11 C]flumazenil binding to benzodiazepine re- tions, and antagonist blockade in heroin-depen- ceptors in chronic alcohol-dependent men and dent volunteers. Neuropsychopharmacology healthy controls. Psychiatry Res 1993;50(1): 2003;28(11):2000–9. 1–13. [73] Greenwald M, Johanson CE, Bueller J, et al. Bu- [62] Lingford-Hughes AR, Acton PD, Gacinovic S, prenorphine duration of action: m-opioid recep- et al. Levels of gamma-aminobutyric acid-benzo- tor availability and pharmacokinetic and diazepine receptors in abstinent, alcohol-depen- behavioral indices. Biol Psychiatry 2007;61(1): dent women: preliminary findings from an 101–10. 123I-iomazenil single photon emission tomogra- [74] McCann U, Szabo Z, Scheffel U, et al. Positron phy study. Alcohol Clin Exp Res 2000;24(9): emission tomographic evidence of toxic effect 1449–55. of MDMA (‘‘Ecstasy’’) on brain serotonin neu- [63] Jalan R, Turjanski N, Taylor-Robinson SD, et al. rons in human beings. Lancet 1998;352(9138): Increased availability of central benzodiazepine 1433–7. receptors in patients with chronic hepatic en- [75] McCann U, Wong D, Yokoi F, et al. Reduced cephalopathy and alcohol related cirrhosis. Gut striatal dopamine transporter density in absti- 2000;46(4):546–52. nent methamphetamine and methcathinone [64] Staley JK, Gottschalk C, Petrakis IL, et al. Cortical users: evidence from positron emission tomogra- gamma-aminobutyric acid type A-benzodiaze- phy studies with [11 C]WIN-35,428. J Neurosci pine receptors in recovery from alcohol depen- 1998;18(20):8417–22. dence: relationship to features of alcohol [76] Iyo M, Sekine Y, Mori N. Neuromechanism of dependence and cigarette smoking. Arch Gen developing methamphetamine psychosis: a neu- Psychiatry 2005;62(8):877–88. roimaging study. Ann N Y Acad Sci 2004;1025: [65] Krystal JH, Staley J, Mason G, et al. Gamma-ami- 288–95. nobutyric acid type A receptors and alcoholism: [77] Chou YH, Huang WS, Su TP, et al. Dopamine intoxication, dependence, vulnerability, and transporters and cognitive function in metham- treatment. Arch Gen Psychiatry 2006;63(9): phetamine abuser after a short abstinence: 957–68. a SPECT study. Eur Neuropsychopharmacol [66] Pfefferbaum A, Lim KO, Zipursky RB, et al. Brain 2007;17(1):46–52. gray and white matter volume loss accelerates [78] Volkow ND, Chang L, Wang GJ, et al. Associa- with aging in chronic alcoholics: a quantitative tion of dopamine transporter reduction with MRI study. Alcohol Clin Exp Res 1992;16(6): psychomotor impairment in methamphetamine 1078–89. abusers. Am J Psychiatry 2001;158(3):377–82. [67] Gueorguieva R, Wu R, Pittman B, et al. New [79] Villemagne V, Yuan J, Wong D, et al. Brain dopa- insights into the efficacy of naltrexone based mine neurotoxicity in baboons treated with on trajectory-based reanalyses of two negative doses of methamphetamine comparable to clinical trials. Biol Psychiatry 2007;61(11): those recreationally abused by humans: evidence 1290–5. from [11 C]WIN-35,428 positron emission [68] Bencherif B, Wand GS, McCaul ME, et al. m-opi- tomography studies and direct in vitro determi- oid receptor binding measured by [11 C]carfenta- nations. J Neurosci 1998;18(1):419–27. nil positron emission tomography is related to [80] Melega WP, Raleigh MJ, Stout DB, et al. Recovery craving and mood in alcohol dependence. Biol of striatal dopamine function after acute am- Psychiatry 2004;55(3):255–62. phetamine- and methamphetamine-induced [69] Heinz A, Reimold M, Wrase J, et al. Correlation neurotoxicity in the vervet monkey. Brain Res of stable elevations in striatal m-opioid receptor 1997;766(1–2):113–20. availability in detoxified alcoholic patients with [81] Harvey DC, Lacan G, Melegan WP. Regional het- alcohol craving: a positron emission tomogra- erogeneity of dopaminergic deficits in vervet phy study using carbon 11-labeled carfentanil. monkey striatum and substantia nigra after Arch Gen Psychiatry 2005;62(1):57–64. methamphetamine exposure. Exp Brain Res [70] Zubieta J, Greenwald MK, Lombardi U, et al. Bu- 2000;133(3):349–58. prenorphine-induced changes in m-opioid recep- [82] Harvey DC, Lacan G, Tanious SP, et al. Recovery tor availability in male heroin-dependent from methamphetamine induced long-term ni- volunteers: a preliminary study. Neuropsycho- grostriatal dopaminergic deficits without sub- pharmacology 2000;23(3):326–34. stantia nigra cell loss. Brain Res 2000;871(2): [71] Kling MA, Carson RE, Borg L, et al. Opioid recep- 259–70. tor imaging with positron emission tomography [83] Wilson JM, Kalasinsky KS, Levey AI, et al. Striatal and [(18)F]cyclofoxy in long-term, methadone- dopamine nerve terminal markers in human, treated former heroin addicts. J Pharmacol Exp chronic methamphetamine users. Nat Med Ther 2000;295(3):1070–6. 1996;2(6):699–703. Imaging the Neurochemistry of Addiction 555

[84] Reneman L, de Win MM, van den Brink W, et al. [91] Ricaurte GA, McCann UD. Assessing long-term Neuroimaging findings with MDMA/ecstasy: tech- effects of MDMA (Ecstasy). Lancet 2001; nical aspects, conceptual issues and future pros- 358(9296):1831–2. pects. J Psychopharmacol 2006;20(2):164–75. [92] McCann UD, Szabo Z, Seckin E, et al. Quantita- [85] Neumeyer JL, Wang SY, Milius RA, et al. [123I]-2 tive PET studies of the serotonin transporter in beta-carbomethoxy-3 beta-(4-iodophenyl)tro- MDMA users and controls using [11 C]McN5652 pane: high-affinity SPECT radiotracer of mono- and [11 C]DASB. Neuropsychopharmacology amine reuptake sites in brain. J Med Chem 2005;30(9):1741–50. 1991;34(10):3144–6. [93] Buchert R, Thomasius R, Wilke F, et al. A voxel- [86] Heinz A, Jones DW. Serotonin transporters in based PET investigation of the long-term effects ecstasy users. Br J Psychiatry 2000;176:193–5. of ‘‘Ecstasy’’ consumption on brain serotonin [87] Huang Y, Hwang DR, Narendran R, et al. Com- transporters. Am J Psychiatry 2004;161(7):1181–9. parative evaluation in nonhuman primates of [94] Buchert R, Thomasius R, Petersen K, et al. Revers- five PET radiotracers for imaging the serotonin ibility of ecstasy-induced reduction in serotonin transporters: [11 C]McN 5652, [11 C]ADAM, transporter availability in polydrug ecstasy users. [11 C]DASB, [11 C]DAPA, and [11 C]AFM. J Cereb Eur J Nucl Med Mol Imaging 2006;33(2):188–99. Blood Flow Metab 2002;22(11):1377–98. [95] Reneman L, Booij J, Lavalaye J, et al. Use of am- [88] Reneman L, Booij J, de Bruin K, et al. Effects of phetamine by recreational users of ecstasy dose, sex, and long-term abstention from use on (MDMA) is associated with reduced striatal do- toxic effects of MDMA (ecstasy) on brain seroto- pamine transporter densities: a [123I]beta-CIT nin neurons. Lancet 2001;358(9296):1864–9. SPECT study—preliminary report. Psychophar- [89] Semple DM, Ebmeier KP, Glabus MF, et al. Re- macology (Berl) 2002;159(3):335–40. duced in vivo binding to the serotonin trans- [96] Narendran R, Frankle WG, Keefe R, et al. Altered porter in the cerebral cortex of MDMA prefrontal dopaminergic function in chronic rec- (‘‘ecstasy’’) users. Br J Psychiatry 1999;175: reational ketamine users. Am J Psychiatry 2004; 63–9. 162(12):2352–9. [90] Laruelle M, Wallace E, Seibyl JP, et al. Graphical, [97] Guo N, Hwang DR, Lo ES, et al. Dopamine de- kinetic, and equilibrium analyses of in vivo pletion and in vivo binding of PET D1 receptor [123I] beta-CIT binding to dopamine trans- radioligands: implications for imaging studies porters in healthy human subjects. J Cereb Blood in schizophrenia. Neuropsychopharmacology Flow Metab 1994;14(6):982–94. 2003;28(9):1703–11.