Nora D. Volkowa,b,1, Gene-Jack Wanga, Frank Telanga, Joanna S. Fowlerc,1, David Alexoffc, Jean Logand, Millard Jaynea, Christopher Wonga, and Dardo Tomasia

Laboratory of Neuroimaging, National Institute on Alcohol Abuse and Alcoholism, Rockville, MD 20857; National Institute on Drug Abuse, Rockville, MD  20857; Biosciences Department, Brookhaven National Laboratory, Upton, NY 11973; and Department of Radiology, New York University Langone Medical Center, New York, NY 10016

Contributed by Joanna S. Fowler, June 20, 2014 (sent for review April 9, 2014; reviewed by Bertha Madras, Harvard University Medical School, and Karen Berman, National Institute of Mental Health)

Moves to legalize marijuana highlight the urgency to investigate effects of chronic marijuana in the human brain. Here, we challenged  48 participants (24 controls and 24 marijuana abusers) with methylphenidate (MP), a drug that elevates extracellular dopamine (DA) as a surrogate for probing the reactivity of the brain to DA stimulation. We compared the subjective, cardiovascular, and brain  DA responses (measured with PET and [11Craclopride) to MP between controls and marijuana abusers.

Although baseline (placebo) measures of striatal DA D2 receptor availability did not differ between groups, the marijuana abusers showed markedly blunted responses when challenged with MP. Specifically, compared with controls, marijuana abusers had significantly attenuated behavioural (“self-reports” for high, drug effects, anxiety, and restlessness), cardiovascular (pulse rate and diastolic blood pressure), and brain DA [reduced decreases in distribution volumes (DVs) of [11Craclopride, although normal reductions in striatal nondis placeable binding potential  (BPND)] responses to MP. In ventral striatum (key brain reward region),M P-induced reductions in DVs and BPND (reflecting DA increases) were inversely correlated with scores of negative emotionality, which were significantly higher for marijuana abusers than controls. In marijuana abusers, DA responses in ventral striatum were also inversely correlated with addiction severity and craving.

The attenuated responses to MP, including reduced decreases in striatal DVs, are consistent with decreased brain reactivity to the DA stimulation in marijuana abusers that might contribute to their negative emotionality (increased stress reactivity and irritability) and addictive behaviors.

Despite the high prevalence of marijuana consumption, the effects of marijuana abuse in the human brain are not well understood. Marijuana, like other drugs of abuse, stimulates brain dopamine (DA) signalling in the nucleus accumbens (1, 2), which is a mechanism believed to underlie the rewarding effects of drugs (3–5) and to trigger the neuroadaptations that result in addiction (reviewed in ref. 6). Indeed, in humans, imaging studies have shown that drugs of abuse increase DA release in striatum (including the nucleus accumbens), and these increases have been associated with the subjective experience of reward (7–9).

However, for marijuana, the results have been inconsistent: One study reported striatal DA increases during intoxication (10); two studies showed no effects (11, 12); and one study reported DA increases in individuals with a psychotic disorder and in their relatives, but not in controls (13). Imaging studies of the brain DA system in marijuana abusers have also shown different findings from those reported for other types of substance abusers. Specifically, substance abusers (cocaine, methamphetamine, alcohol, heroin, and nicotine), but not marijuana abusers (14–16), show reduced baseline availability of DA D2 receptors in striatum (reviewed ref. 6). Similarly, cocaine abusers (17, 18) and alcoholics (19, 20),but not marijuana abusers (16), show attenuated DA increases instriatum when challenged with a stimulant drug, although marijuana abusers with comorbid schizophrenia or risk for schizophrenia showed blunted DA increases to stimulants (21) and to stress (22). However, prior studies are limited by their small sample sizes (ranging from six to 16 subjects). Also, prior studies did not control for the potential confounds that the changes in cerebral vascular resistance associated with marijuana abuse (23– 25) could have on the delivery of the radiotracer to the brain when using a stimulant drug as pharmacological challenge, because stimulants decrease cerebral blood flow (26). Thus, the extent to which there are changes in brain DA signalling in marijuana abusers is still unclear. Here, we compared brain DA reactivity in healthy controls and marijuana abusers on a larger sample than that in prior studies and measured arterial concentration of non metabolized radiotracer to control for differences in radiotracer delivery to brain. We used PET and [11Craclopride (radioligand that binds to D2/D3 receptors not occupied by DA) to assess the effects of methylphenidate (MP) on the nondisplaceable binding potential [BPND; ratio of the distribution volume (DV) in striatum to that in cerebellum], which is the most frequent model parameter used to estimate DA changes (27), in 24 healthy controls and 24 marijuana abusers.We also quantified the DV,  which corresponds to the equilibrium measurement of the ratio of the concentration of the radiotracer in tissue to that in  arterial plasma, to control for potential changes in radiotracer delivery that could confound group comparisons of stimulant-induced changes in BPND. We used MP, which is a stimulant drug that blocks DA transporters, because it induces robust and reproducible DA increases in the human brain (28, 29). We predicted that MP’s behavioral effects in marijuana abusers would be attenuated, consistent with preclinical findings (30), and that decreased DA reactivity in ventral striatum would be associated with higher scores in negative emotionality (neuroticism), which mediates genetic risk for marijuana dependence (31), and with addiction severity.

Results

Participant Characteristics.

Tobacco smoking was more prevalent in marijuana abusers than controls; otherwise, there were no differences in demographics between groups (Table 1). However, the groups differed significantly in personality measures; marijuana abusers had significantly lower scores in positive emotionality (P = 0.05) and higher scores in negative emotionality (P = 0.002) than controls (Table 1).

Correlation analysis between scores in negative emotionality and history of marijuana abuse showed a negative correlation between age of I  the scores. The correlations with reported daily doses of marijuana and negative emotionality were not correlations with positive emotionality and history of marijuana abuse were not significant.

Plasma Concentrations of MP and Behavioral and Cardiovascular Effects. 

MP concentrations in plasma (nanograms per milliliter) did not differ between groups at 10 min (controls, 195 ± 51; abusers, 194 ± 45), 25 min (controls, 125 ± 24; abusers, 121 ± 19), or 40 min (controls, 102 ± 25; abusers, 94 ± 15). MP had significant behavioral effects, and these effects were attenuated in marijuana abusers compared with controls (Fig. 1A). Specifically, MP significantly increased scores on self-reports (averaged measures), and the effects differed between groups, with controls reporting a more robust “high” (drug effect: F = 92, P = 0.0001; interaction: F = 6.2, P = 0.02), “restlessness” (F = 35, P = 0.0001; interaction: F = 5.8, P = 0.02), “anxiety” (F = 7, P = 0.01; interaction: F = 5.8, P = 0.02), and “drug effects” (F = 100, P = 0.0001; interaction F = 4, P = 0.05) than marijuana abusers.

Also, comparisons of “peak” behavioral effects to MP were significantly stronger for controls for high (P = 0.01), restlessness (P = 0.003), anxiety (P=0.03),and drug effects

(P=0.02), than for the marijuana abusers. The potency of MP was also reported to be stronger by the controls than by the marijuana abusers (8.3 ± 2 vs. 5.8 ± 3; t = 3.4, P = 0.002). In marijuana abusers, MP increased self-reports of marijuana craving (Placebo: 4.0 ± 3–MP: 6.3 ± 3; P = 0.006) and tobacco craving (Placebo: 2.4 ± 2–MP: 3.8 ± 4; P = 0.05).

PLUS

MP increased heart rate

(F = 98, P = 0.0001) and systolic (F = 153, P = 0.0001) and diastolic (F = 65, P = 0.0001) blood pressure in both groups, and MP’s effects differed between groups for heart rate (interaction effect; F = 4.6, P = 0.04) and diastolic blood pressure (interaction effect: F = 4.0, P = 0.05), but not for systolic blood pressure (Fig. 1B). Post hoc t tests revealed that MP-induced increases in heart rate and diastolic pressure were significantly stronger (P < 0.05) in controls than in marijuana abusers.

Effects of MP on the DVs of [11Craclopride.

The SPM analysis showed no group differences in baseline measures of DV. It also showed that MP significantly decreased DV in brain and that the effects were significantly larger in controls than in marijuana abusers (Fig. 2). Individual plotting of MP-induced changes in DV showed that MP-induced changes in cerebellum were decreased in controls but not in marijuana abusers and that there were larger decreases of MP-induced changes in striatum in controls than in marijuana abusers (Fig. 2). The ROI analysis corroborated that MP decreased the DV in cerebellum and striatum and that the effects were larger for controls than abusers.

For cerebellum, the drug (F = 15, P = 0.0004) and drug × group interaction (F = 8.2, P = 0.007) were significant; post hoc t tests showed larger decreases in controls (13 ± 11%) than abusers (1.4 ± 16%) (P = 0.01). For caudate, the drug (F = 41, P = 0.0001) and interaction (F = 4.8, P = 0.04) were significant; post hoc t tests revealed larger decreases in controls (22 ± 18%) than abusers (9 ± 22%) (P = 0.05).

For putamen, drug (F = 93, P = 0.0001) and interaction (F = 6.9, P = 0.02) were significant; post hoc t tests showed larger decreases in controls (30 ± 16) than abusers (16 ± 21%) (P = 0.02). For ventral striatum, drug (F = 56, P = 0.0001) and interaction (F = 7.3, P = 0.01) were significant; post hoc t tests showed greater decreases in controls (25 ± 18%) than abusers (11 ± 25%) (P = 0.02). A group (controls vs. abusers) by region (delta DV in caudate, putamen, ventral striatum, and cerebellum) comparison revealed that group differences differed between regions (F = 3.5, P = 0.02); post hoc analysis showed that group differences in cerebellum were larger than in putamen (P = 0.02) and ventral striatum (P = 0.02), and showed a trend in caudate (P = 0.07).

This finding is significant; it confounds group comparisons of BPND because the latter measure is normalized to the DV in cerebellum. Note that attenuated decreases in cerebellar DV with MP in the marijuana abusers could result in an overestimation of their DA increases, reflecting an apparent lower striatal-DV/cerebellar-DV ratio (BPND) with MP (see below).

Correlations Between MP-Induced Changes in DV and Clinical Measures.

Correlation analysis revealed that MP-induced decreases in DV in ventral striatum were negatively associated with scores in negative emotionality (r = 0.51, P = 0004), and weaker correlations were observed in putamen (r = 0.37, P = 0.02) and caudate

(r = 0.35, P = 0.02) such that the larger the DV decreases, the lower were the scores of negative emotionality. Correlation with positive emotionality and constraint were not significant. MP-induced craving for marijuana in the marijuana abusers was negatively associated with DV decreases in putamen (r = 0.46, P = 0.03) and ventral striatum (r = 0.51, P = 0.01) such that participants with the smallest decreases had the most intense craving.

Baseline Measures of D2/D3 Receptor Availability (BPND).

For the baseline (placebo) measures, the SPM analysis revealed no group differences in BPND (D2/D3 receptor availability). When we decreased the threshold of significance to uncorrected P < 0.05, SPM showed lower values in marijuana abusers than in controls in ventral striatum (0, −2, −8; statistical t values = 2.59, P uncorrected = 0.007). The ROI analysis also showed a nonsignificant trend toward lower baseline BPND in marijuana abusers than in controls in ventral striatum (controls, 3.20 ± 0.3; abusers, 2.97 ± 0.59; P = 0.11) and no differences in caudate (controls, 2.80 ± 0.36; abusers 2.76 ± 0.57) or putamen (controls, 3.42 ± 0.41; abusers, 3.35 ± 0.57).

Effects of MP on BPND. 

The SPM analysis revealed significant decreases in BPND with MP compared with placebo (interpreted as reflecting DA increases) in striatum in both controls and marijuana abusers (Fig. 3 and Table 2). The SPM analysis revealed no group differences in MP-induced decreases in BPND in striatum but unexpectedly revealed larger BPND decreases in marijuana abusers than in controls in midbrain (region centered in susbtantia nigra that also encompassed sub thalamic nucleus; center of cluster left: 12, −14, −10, and 132 voxels, t = 3.1; center of cluster right: 14, −18, −8, and 27 voxels; t = 2.9; PFWE < 0.05; SVC = 10 voxels) (Fig. 3 and Table 2). The ROI analysis corroborated a significant group × drug interaction in midbrain (F = 14, P = 0.0006), and post hoc t test analyses showed that whereas in marijuana abusers, MP decreased BPND in midbrain (−3.5 ± 8%; F = 5.4, P = 0.03), MP increased BPND in controls (4 ± 6%; F = 9.2, P = 0.006).

Correlations Between MP-Induced Changes on BPND and Clinical Measures.

Voxel-wise correlation analysis revealed that MP-induced decreases in BPND in ventral striatum were inversely associated with scores in negative emotionality (Fig. 3 B and C) such that the larger the BPND decreases, the lower the scores. The striatal correlations with positive emotionality and constraint were not significant. Because the SPM revealed a significant group difference in MP-induced changes in midbrain BPND, we also performed correlations with this brain region and showed a significant correlation with positive emotionality (r = 0.42, P = 0.003) such that the greater the BPND decreases, the lower the scores. In the marijuana abusers, MP-induced decreases in BPND in midbrain were correlated with increases in marijuana (r = 0.40, P = 0.05) and tobacco (r = 0.45, P = 0.03) craving, as well as with the dependency scores (r = 0.43, P = 0.04), such that the greater the decreases in BPND, the higher was the craving triggered by MP and the higher were the dependency scores.

Discussion

Here, we show that marijuana abusers had attenuated behavioural and cardiovascular responses and blunted reductions in striatal DV (although normal reductions in BPND) when challenged with MP compared with controls, which is consistent with decreased brain reactivity to DA stimulation. We also corroborate prior findings (14–16) of no significant differences in baseline striatal D2/D3 receptor availability between controls and marijuana abusers and provide preliminary evidence of abnormal midbrain

DA reactivity in marijuana abusers. DA D2/D3 Receptor Availability in Striatum.

Only four brain imaging studies (totalling 42 marijuana abusers) have measured DA D2/ D3 receptors (14–16, 42). These studies showed no differences in striatal D2/D3 receptors between marijuana abusers and controls, but their generalizability is limited by the small sample sizes (samples ranged from n = 6 to n = 16). Thus, our results showing no differences in D2/D3 receptor availability (except for a trend in ventral striatum), using a larger sample (24 marijuana abusers) than that used for studies that identified reductions in striatal D2/D3 receptors in alcoholics and cocaine abusers, indicate that marijuana abusers, different from other drug abusers, do not show significant striatal D2/D3 receptor reductions. This difference could reflect marijuana’s agonist properties at cannabinoid 1 (CB1) receptors, which heteromerize with D2 receptors, antagonizing their effects (43). Both CB1 and D2 receptors couple to Gi-o proteins and inhibit adenylyl-cyclase, whereas their co-stimulation results in Gs protein-dependent activation of adenylylcyclase (44, 45).

Moreover, CB1 receptor agonists and antagonists counteract and potentiate, respectively, D2 receptor agonist effects (46–49), although D2 and CB1 receptor interactions might differ between rodents and primates (50, 51). It is therefore possible that in marijuana abusers, chronic CB1 receptor stimulation prevented the striatal D2/D3 receptor down-regulation observed < 0.005) and group comparisons for the effects of MP (ΔBPND) (P < 0.01, cluster size of 10 voxels). The contrast MA > HC indicates that MP induced with repeated drug use (reviewed in ref. 6). However, it should be noted that the marijuana abusers studied in the present and prior studies have been at least 10 y younger than the cocaine abusers and alcoholics studied by prior PET studies, which is relevant because striatal D2/D3 receptors decrease with age (52), and it is hypothesized that drugs accelerate the effects of brain aging (53). Thus, studies in older marijuana abusers are needed to clarify this.

MP-Induced Changes in DV.

In controls but not in marijuana abusers, MP reduced cerebellar DV. To ensure that the DV responses in the controls were consistent with prior findings, we performed a secondary analysis on the effects of MP on the cerebellar DV in an independent cohort of controls, which showed a 12% reduction, and in a sample of adults with attention deficit hyperactivity disorder (ADHD), which also showed an 11% reduction (for controls of the current cohort, the cerebellar DV decrease was 13 ± 11%). The mechanism underlying the lack of an effect of MP in cerebellar DV in abusers is unclear but could reflect the effects of chronic marijuana on cerebrovascular reactivity (increased cerebral vascular resistance) (23–25), which might have prevented MP-induced vasoconstriction and associated reductions in radiotracer delivery to the brain.

The attenuated decreases in DV with MP in the marijuana abusers were observed throughout the brain but were most accentuated in cerebellum. The higher sensitivity of the cerebellum to what we interpret to reflect changes in vascular reactivity with marijuana abuse is consistent with clinical findings that report strokes associated with marijuana abuse are more frequently localized in the posterior circulation and ischemia is most frequently observed in cerebellum (25, 54–56). Cerebellar arteries express CB1 receptors in the smooth muscle layer (57), but because comparisons with arteries in other brain regions have not been done, it is not possible to determine if higher levels of CB1 receptors in cerebellar arteries underlie their higher sensitivity to vascular effects from marijuana.

However, CB1 receptors in cerebellum are also expressed in neurons and glia (58), and the cerebellum is a region that is affected in marijuana abusers (59–61); thus, we cannot rule out the possibility that other factors contribute to the lack of an effect of MP on the cerebellar DV in the marijuana abusers. MP also decreased the DV in striatum to a greater extent in controls than in abusers (Fig. 2). In ventral striatum, these decreases were associated with negative emotionality and with marijuana craving such that the lower the response, the higher the negative emotionality and the craving. This would suggest that these attenuated responses might reflect reduced striatal DA reactivity in marijuana abusers compared with controls even though there were no group differences in MP-induced decreases in BPND (see below). This is consistent with findings from an imaging study with [18F]-dopa that reported lower than normal DA synthesis capacity in the striatum of marijuana abusers (62).

MP-Induced Changes in BPND.

We showed no group differences in MP-induced changes in BPND in striatum, which is the standard measure for assessing DA changes. Similarly, a prior study reported no differences in amphetamine-induced decreases in BPND between marijuana abusers and controls (16). However, the significant group differences in MP’s effects on the DV in cerebellum confound the findings because BPND uses the cerebellum as a reference region to normalize for nonspecific binding. Because the DV in cerebellum was not decreased by MP in marijuana abusers but was decreased in controls, this would result in an overestimation of the decrease in BPND with MP (cerebellar denominator would have a relatively larger value) and an overestimation of DA increases in marijuana abusers compared with controls.

Interestingly, an imaging study comparing DA increases using BPND and 4-propyl-9-hydroxynaphthoxazine ([11C]PHNO)

(radiotracer with >20-fold higher affinity for D3 over D2 receptors, and presumably more sensitive to competition with endogenous DA) (63, 64) in response to a stressor in individuals at high risk for schizophrenia showed that those who abused marijuana had a blunted response, consistent with decreased DA signalling (22). Because the study used cognitive stress as a challenge, it was not confounded by potential group differences in stimulant-induced changes in cerebellar radiotracer delivery.  Unexpectedly, SPM revealed that MP decreased BPND in midbrain (centered in substantia nigra) in marijuana abusers but not in controls. Although the mechanism(s) underlying this group difference is unclear, we speculate that because the midbrain has a high concentration of D3 receptors (65), which are more sensitive to endogenous DA than D2 receptors (66), it could reflect up-regulation of D3 receptors in marijuana abusers.

Indeed, in rodents, chronic Δ (9)-tetrahydrocannabinol (THC; the main psychoactive ingredient of marijuana) increased D3 receptors in midbrain (30). In the marijuana abusers, an MP-induced decrease in midbrain BPND correlated with craving and with dependency scores. A similar finding was reported in methamphetamine abusers, in whom up-regulation of D3 receptors in midbrain (assessed with [11C]PHNO) correlated with amphetamine-induced craving (30, 67). This, along with preclinical studies showing that D3 receptor antagonists interfere with drug seeking and cue- and receptor signalling in midbrain might contribute to drug craving and to decreased sensitivity to reward in marijuana abusers (see below).

However, because the midbrain finding was unexpected, we report it as a preliminary finding in need of replication.

Blunted Behavioral and Cardiovascular Responses to MP in Marijuana Abusers.

Behavioral and cardiovascular effects of MP have been associated with MP-induced DA increases in striatum (9, 69), so the blunted responses in the marijuana abusers are also consistent with decreased striatal reactivity to DA signaling. Although, to our  knowledge, this is the first clinical report of an attenuation of the effects of MP in marijuana abusers, a preclinical study had reported that rats treated chronically with THC exhibited attenuated locomotor responses to amphetamine (2.5 mg/kg

administered i.p.) (30). Such blunted responses to MP could reflect neuroadaptations from repeated marijuana abuse, such as downregulation of DA transporters (70). The attenuation of MP’s effects could also reflect abnormal D2 receptor function, as was previously suggested to explain findings in marijuana-abusing schizophrenic patients, who, despite displaying low DA release, showed increases in psychotic symptoms when challenged with amphetamine (21). Finally, it is also possible that the attenuated responses reflect blunting of MP’s noradrenergic effects because MP blocks both DA and norepinephrine transporters. Our findings of blunted responses to MP in marijuana abusers have clinical implications because they suggest that individuals with ADHD who abuse marijuana might be less responsive to the therapeutic benefits derived from stimulant medications.

Reduced Positive Emotionality and Increased Negative Emotionality in Marijuana Abusers.

Marijuana abusers showed lower scores on positive emotionality and higher scores on negative emotionality than controls, consistent, on the one hand, with lower reward sensitivity and motivation and, on the other hand, with increased stress reactivity and irritability. These characteristics overlap with the amotivational syndrome (71) and with the enhanced sensitivity to stress associated with marijuana abuse and other addictions (72, 73). Positive emotionality was inversely associated with MP induced increases in midbrain DA, which could reflect the fact that in midbrain, D2 and D3 are autoreceptors; therefore, their stimulation would result in decreased DA release in striatum (including accumbens) (74), leading to decreased sensitivity to reward and amotivation (75). In contrast, MP-induced DA increases in ventral striatum were negatively associated with scores on negative emotionality, which is consistent with the protective role of DA signalling in negative emotions (76). The association between negative emotionality and age of initiation of marijuana abuse is consistent with prior findings of worse outcomes with earlier initiation of marijuana abuse (77).

Study Limitations.

The main limitation of this study was the inadequacy of BPND for comparing the DA increases between controls and marijuana abusers due to the group differences on  the effects of MP on cerebellar DV. Also, [11Craclopride cannot distinguish between D2 and D3 receptors, so studies with D3 receptor ligands are needed to determine if the increased midbrain DA response in marijuana abusers reflects D3 receptor upregulation. The relatively poor spatial resolution of PET limits accuracy in the quantification of small brain regions, such as midbrain. Our study cannot ascertain if group differences reflect chronic use of marijuana rather than premorbid differences, and whether marijuana abusers will recover with detoxification. Although attenuation of the effects of MP could reflect interference from CB1 receptor stimulation by marijuana, this is unlikely because marijuana abusers reported that their last use of marijuana was 1–7 d before the study when cannabinoids in plasma are still detectable but at concentrations unlikely to have pharmacological effects (78). However, future studies done after longer periods of withdrawal are needed to control for potential confounds from THC and its metabolites in plasma and to determine if the blunted responses recover.

We did not obtain MRI scans on the participants. However, this is unlikely to have affected the results because measures of [11Craclopride binding are equivalent when using a region extracted from an MRI scan or from the [11Craclopride scan

(79), and there is no evidence that marijuana abusers have striatal or cerebellar atrophy (reviewed in ref. 80). Finally, the groups differed in smoking status, but this is unlikely to account for the group differences because CO levels were used as a covariate in the analysis and there were no differences in the effects of MP between marijuana abusers who smoked cigarettes and those who did not.

Conclusions

The significantly attenuated behavioral and striatal DV response to MP in marijuana abusers compared with controls, indicates reduced brain reactivity to DA stimulation that in the ventral striatum might contribute to negative emotionality and drug craving.

Source:  http://www.pnas.org/content/111/30/E3149