Addiction research has become increasingly focused on how addictive drugs affect neural signaling within the brain's reward circuitry. New work shows that the frustration of not receiving drugs when they are expected can also affect the inner workings of reward circuits.
The past decade has brought major insights into how addictive drugs affect intracellular signaling in the brain circuitry that controls drug-seeking habits. At almost every level of neuronal function—second messenger cascades, gene expression, receptor and transporter trafficking, ion channel permeability, neuronal discharge rate, microtubule expression and dendrite sprouting— repeated drug experiences alter the reward system 1. These alterations, however, are seen in animals that receive the drug passively as well as in animals that self-administer the drug. Many of these neuroadaptations may turn out to be mere correlates or consequences of addiction, rather than the neuroadaptations that are responsible for addiction. It is the neuroadaptations of learning and memory 2 that seem essential for forming drug habits. In turn, it is compulsive drug habits that lead to the chronic intoxication that is essential for the neuroadaptations that differentiate the addicted brain from the non-addicted brain. In the January 2 issue of Nature, Sutton et al.3 begin to tease apart some of the molecular mechanisms underlying the processes of learning and memory in relationship to drug taking.
In an important demonstration of drug-associated learning, Sutton et al.3 report that levels of GluR1 (a subunit of -amino-3-hydroxy-5-methyl-4-isoxazol propionate (AMPA) glutamate receptors) increase within the nucleus accumbens of rats following extinction of cocaine self-administration, but not following passive withdrawal from cocaine. Extinction entails experimental sessions in which a drug-trained habit—lever-pressing in this study3—goes unrewarded by drug injections. In contrast, passive withdrawal entails time during which the opportunity to perform the drug-trained habit is not available. There is a crucial distinction between the consequences of these two conditions: extinction training decreases drug-seeking responses over time, whereas passive withdrawal can actually increase4 drug-seeking once the original testing situation is restored. It is especially important that this neuroadaptation occurs within the nucleus accumbens, because this brain region has a critical role in the rewarding effects of cocaine and other drugs of abuse5. In rats that underwent extinction training, GluR1 levels in the nucleus accumbens were inversely correlated with the degree of drug-seeking behavior: the more GluR1, the less drug seeking. Levels of GluR2/3 (reflecting GluR2, another AMPA receptor subunit) also increased within the nucleus accumbens. But this neuroadaptation seems more general, associated with the presence of some cues (the test environment) formerly associated with cocaine self-administration rather than the failure of a discrete action (lever press) to produce the expected results (cocaine). Importantly, Sutton et al.3 were able to mimic critical aspects of extinction training by implanting extra copies of the gene encoding GluR1 into the nucleus accumbens using viral-mediated gene transfer6.
Although we are far from understanding the molecular mechanisms of the work of Sutton et al.3, it has considerable significance. Foremost, it identifies a neuroadaptation that is not caused by the binding of a powerful drug to its receptor, but rather by the failure of an expected drug injection to result from an established drug-seeking act. It reflects a neuroadaptation that is the result of a powerful psychological experience—the frustration of an expectation—rather than a powerful pharmacological experience. One interpretation is that extinction-induced elevations in GluR1 and GluR2 within the nucleus accumbens reflect the memory of associations between the drug and the actions taken to get it in the past, and the effect of this memory on brain biology. The mechanisms underlying these elevations are not known, but they likely involve differences in post-transcriptional processes (protein redistribution, degradation), because they are not caused by elevated mRNA expression. They may involve midbrain dopamine systems, which are known to be affected by stress or the 'surprise' of not receiving what is expected7. Regardless, increases in these subunits would be expected to cause increased numbers of AMPA glutamate receptors within the nucleus accumbens, and subsequent increases in the sensitivity of nucleus accumbens neurons to the excitatory actions of glutamate.
Past work has led to the simple hypothesis that treatments that decrease the excitability of the nucleus accumbens are rewarding5, whereas treatments that increase excitability are aversive8 (Fig. 1). Elevated excitability and its accoutrements (including increased flux of calcium into nucleus accumbens neurons) may trigger the activation of transcription factors and genes that lead directly to aversive states2,
9. At least in rats, treatments associated with aversive states such as severe drug withdrawal tend to decrease rather than increase the likelihood of drug-seeking behaviors, whereas administration of small amounts of drug tends to whet the appetite for more10.
Figure 1. The nucleus accumbens and reward-related states in normal rats and in rats with extinguished drug self-administration habits: a working model.
a, Nucleus accumbens neurons tonically inhibit reward-related processes. Under normal circumstances, a balance exists between cortical (PFC, AMG) excitatory (+) influences mediated by glutamate actions at AMPA and NMDA receptors, and midbrain (VTA) inhibitory (-) influences mediated by dopamine actions at D2-like receptors5. Nucleus accumbens neurons have low baseline rates of firing, and depolarization-mediated influx of Ca2+ through NMDA receptors and calcium channels is not sufficient to alter gene expression. b, Extinction-induced elevations of AMPA (G1/G2) receptors render nucleus accumbens neurons hypersensitive to glutamatergic inputs. This increases basal firing rates and inhibition of reward-related or drug-seeking states10. Additionally, increased depolarization leads to elevated Ca2+ influx, which may trigger local activation of transcription factors and genes associated with aversive states9. AMG, amygdala; CREB, cAMP response element binding protein; D2, dopamine D2-like receptor; DA, dopamine; G1/G2, AMPA glutamate receptor containing GluR1 and GluR2; GABA, -aminobutyric acid; GLU, glutamate; PFC, prefrontal cortex; VTA, ventral tegmental area.
The findings of Sutton et al.3 are consistent with what many addiction researchers have long suspected: behavioral approaches that incorporate extinction-like processes may have efficacy in the treatment of cocaine addiction, either on their own or as an adjunct to more traditional strategies involving pharmacotherapies. Theoretically, the current study may also spark interest in the development of pharmacotherapies that selectively regulate GluR1 levels in the nucleus accumbens, although such specificity is currently unprecedented. However, what students of the brain may find most intriguing about this work is that it shows an effect on brain biology that we can, for now, categorize as being the result of a psychological event: the frustration11 resulting from unmet expectations. We already have many examples of how brain biology can affect behavior and mental function. The Sutton et al.3 findings offer insight on the other side—the less-studied side—of the mind−brain interaction.
Nestler, E.J. Molecular neurobiology of addiction. Am. J. Addict.10, 201–217 (2001). | Article | PubMed | ISI | ChemPort |
Berke, J.D. & Hyman, S.E. Addiction, dopamine, and the molecular mechanisms of memory. Neuron25, 515–532 (2000). | Article | PubMed | ISI | ChemPort |
Sutton, M.A. et al. Extinction-induced upregulation in AMPA receptors reduces cocaine-seeking behaviour. Nature42, 70–75 (2003). | Article | ChemPort |
Grimm, J.W., Hope, B.T., Wise, R.A. & Shaham, Y. Incubation of cocaine craving during withdrawal. Nature412, 141–142 (2001). | Article | PubMed | ISI | ChemPort |
Carlezon, W.A. Jr. & Wise, R.A. Rewarding actions of phencyclidine and related drugs in nucleus accumbens shell and frontal cortex. J. Neurosci.16, 3112–3122 (1996). | PubMed | ISI | ChemPort |
Carlezon, W.A. Jr. Nestler, E.J. & Neve, R.L. Viral-mediated gene transfer as a tool for neuropsychiatric research. Crit. Rev. Neurobiol.14, 47–68 (2000). | PubMed | ISI | ChemPort |
Schultz, W. Getting formal with dopamine and reward. Neuron36, 241–263 (2002). | Article | PubMed | ISI | ChemPort |
Kelz, M.B. et al. Expression of the transcription factor FosB in the brain controls sensitivity to cocaine. Nature401, 272–276 (1999). | Article | PubMed | ISI | ChemPort |
Pliakas, A.M. et al. Altered responsiveness to cocaine and increased immobility in the forced swim test associated with elevated cAMP response element binding protein in nucleus accumbens. J. Neurosci.21, 7397–7403 (2001). | PubMed | ISI | ChemPort |
Stewart, J. & Wise, R.A. Reinstatement of heroin self-administration habits: Morphine prompts and naltrexone discourages renewed responding after extinction. Psychopharmacology108, 79–84 (1992). | PubMed | ISI | ChemPort |
Amsel, A. & Roussel, J. Motivational properties of frustration: I. Effect on a running response of the addition of frustration to the motivational complex. J. Exp. Psychol.43, 363–368 (1952). | PubMed | ISI | ChemPort |
-amino-3-hydroxy-5-methyl-4-isoxazol propionate (AMPA) glutamate receptors) increase within the nucleus accumbens of rats following extinction of cocaine self-administration, but not following passive withdrawal from cocaine. Extinction entails experimental sessions in which a drug-trained habit—lever-pressing in this study
-aminobutyric acid; GLU, glutamate; PFC, prefrontal cortex; VTA, ventral tegmental area.
FosB in the brain controls sensitivity to cocaine. Nature 401, 272–276 (1999). | 
