Cracking the Code: Dopamine Signaling in Cocaine Dependence

Decades of research have established the role of dopamine signaling in the modulation of such critical brain functions as cognition, motivation, and reward and, when gone awry, in behaviors relevant to drug addictions. Preclinical (and to some extent human neuroimaging) research suggests that dopamine signaling is relevant to the reinforcing (rewarding) effects, mediating incentive salience of reward cues, as well as motivational aspects of drug use and abuse. Cocaine blocks synaptic reuptake of dopamine, and human neuroimaging studies (mostly using positron emission tomography [PET]) have consistently demonstrated impaired (reduced) striatal dopamine signaling in volunteers with chronic cocaine dependence. On the basis of its extensive work in this area, one group has proposed that reduced dopamine function in cocaine-dependent subjects might result in decreased sensitivity to non-drug-related stimuli (including natural reinforcers) and disruption in prefrontal cortical (inhibitory) functioning, both of which ultimately manifest in the impulsive and compulsive drug use characteristic of this disorder (2). Despite this progress, much work remains in reconciling the sometimes divergent preclinical and human findings and clarifying the exact role of dopamine in cocaine dependence (3).

In this issue, Martinez and colleagues (1) describe their latest in a series of studies over the past decade systematically investigating limbic striatal dopamine signaling in cocaine-dependent volunteers by means of [¹¹C]raclopride and PET. The strength of this study is the elegant extension of hypothesis-driven and well-established experimental methods from the “bench” to the clinical world. Further, the behavioral intervention (contingency management with the community reinforcement approach) was thoughtfully conducted, enabling a successful study of a challenging cohort while lending itself well to the research methods.

Basing the study on their previous work, the authors set out to investigate whether response to a behavioral intervention involving contingency management with monetary reinforcement was associated with reduced limbic striatal dopamine signaling in a treatment-seeking cohort of cocaine-dependent volunteers. They compared nondisplaceable binding potential (BP_ND) of dopamine receptor types 2 and 3 (D_2/3) and methylphenidate-induced change in D_2/3 receptor binding potential (a measure of presynaptic release of dopamine) determined with a radio-labeled D_2/3 antagonist, [¹¹C]raclopride, and PET in subjects with cocaine dependence and matched healthy comparison subjects, before and after the 12-week behavioral protocol. There was also a further 12-week period of clinical monitoring, which provided valuable information about longer-term outcomes.

The main and exciting finding of this study was the robust replication of the association of reduced limbic striatal presynaptic dopamine release and a model of relapse (involving choices between cocaine and a monetary reward) in a clinical cohort of cocaine-dependent subjects who were treated with a specific behavioral intervention (involving similar choices between cocaine and monetary reward). It is interesting that there was no difference in presynaptic dopamine release between the responders and comparison subjects, suggesting that the responders' dopamine signaling was unimpaired. Remarkably, these responders maintained clinical improvement and abstinence at the 6-month follow-up (unlike the nonresponders, who relapsed soon after continuation), further highlighting the importance of this association.

Several methodological issues are relevant to the interpretation of these findings. It is important for the clinical expert to understand what exactly was measured in this study. Briefly, D_2/3 receptor binding potential (BP_ND) reflects D_2/3 receptor availability (which depends on both the concentration of the D_2/3 receptors available and their affinity to the D_2/3 receptor radioligand [¹¹C]raclopride, which is sensitive to competition with endogenous dopamine). Thus, several factors might influence changes in this measure, making interpretation challenging (see reference 3 for specific discussion). Methylphenidate-induced change in D_2/3 receptor binding is considered a measure of presynaptic dopamine release, though other possible mechanisms (such as receptor internalization or dimerization) exist (4).

Second, questions about the interpretation of the findings of impaired (reduced) dopamine signaling in cocaine dependence remain. Both the vulnerability model (the view that impaired dopamine signaling is a vulnerability factor in initiation or maintenance of addictive behavior) and the toxic effect model (the conception of impaired dopamine signaling as a consequence of long-term cocaine use) have support from animal and human studies. In this study, nonresponders had a significantly longer duration of use than responders. As well, the demonstrations of reduced D_2/3 receptor availability in other addictive and psychiatric disorders suggest that the specificity of these findings to cocaine dependence remains to be established. However, the main finding was the replication of the association between reduced presynaptic release and cocaine-taking behavior (treatment response), and this might be more relevant to cocaine dependence than the reduction in D_2/3 receptor availability.

Third, multiple brain systems (e.g., cognition, motivation, reward), neurotransmitters (e.g., glutamate, γ-aminobutyric acid), and associated regions (e.g., orbitofrontal, cingulate) are likely to be involved in cocaine dependence. The strength of conducting a focused direct investigation of limbic striatal dopamine signaling using [¹¹C]raclopride is also an inherent limitation, in that other relevant brain regions are not included. Multimodal imaging strategies might help optimize the yield from such studies but would obviously add to the complexity and expense.

While not the main focus of this study, the clinical intervention used, although well worked out and implemented, is not trivial in its requirements of resources and effort (e.g., monetary vouchers, clinic visits, and urine screens), raising questions about the feasibility of such treatments in real-world clinical settings.

This study provides robust evidence that reduced presynaptic dopamine release is associated with poorer response to a validated behavioral intervention (involving a monetary reward) in treatment-seeking cocaine-dependent subjects. While we cannot infer causality or a mechanistic explanation from this design, this is an important finding in a clinical setting that adds to existing evidence implicating impaired (reduced) dopamine signaling in the (in)ability to make adaptive choices (in the setting of competing rewards) in cocaine dependence, and it merits further investigation. Further, it raises questions about the feasibility of using such novel methods to predict which patients respond to a specific intervention and whether these changes are reversible with long-term abstinence, though the methods used in this study do not allow us to make predictions about individual subjects.

The authors' suggestion of combining behavioral strategies with adjunctive enhancement of dopamine signaling to help enhance response is consistent with the multipronged strategy proposed by Volkow and colleagues, which consists of using behavioral and pharmacologic methods aimed at reducing the reward value of cocaine, increasing the value of nondrug reinforcers, weakening learned positive associations with drugs and drug cues, and strengthening frontal control (2). A few such strategies are being actively investigated, and preliminary studies show promising results. For example, a recent study by Schmitz and colleagues (5) found that adding an l-dopa/carbidopa combination (versus placebo) to three similar contingency management interventions focusing on different behaviors (attendance, medication compliance, and abstinence judged by cocaine-free urine samples) showed an l-dopa-enhancing effect only for the abstinence intervention. This led the authors to suggest that the saliency of nondrug rewards is facilitated by enhanced dopamine signaling but only when coupled with targeting a behavior (abstinence with contingency management) that reduces the saliency of cocaine. This emergence of hypothesis-driven clinical studies combining specific behavioral interventions and adjunctive agents directly or indirectly modulating dopamine signaling (see example of disulfiram [6, 7]) or other brain mechanisms (see reference 2 for discussion) relevant to cocaine (drug) dependence is exciting and promising.

Sophisticated neuroscience research methods, including neuroimaging, promise the discovery of biomarkers that might guide us in improving diagnostic subtyping, predicting response, and providing new targets for interventions. For a multitude of reasons, these remain important research tools not ready for routine clinical use. However, we must continue to build on our significant progress in the form of hypothesis-driven “next-generation” translational research using genetic and neuroimaging methods in clinical trials of addiction, along the lines suggested by the National Institute of Mental Health's Research Domain Criteria project (8). Studies such as the one by Martinez and colleagues in this issue are a step in the right direction and hold much promise for patients and clinicians alike.

Editorial accepted for publication March 2011.