Neurobiology of addiction and implications for treatment


Drug and alcohol misuse result in immense harm at both individual and societal level. Our understanding of the neuropharmacology of these disorders is increasing through the use of approaches such as neuroimaging and gene targeting and the availability of specific receptor agonists and antagonists. Our aim here is to describe some interesting new findings that are likely to inform advances in treatment.



Over the past 20 years there has been immense interest in the mesolimbic dopaminergic system; most drugs of misuse (except benzodiazepines) increase dopamine here. It is widely accepted that increased levels of dopamine in the nucleus accumbens are key in mediating the rewarding effects or positive reinforcement of drugs of misuse (Koob & Le Moal, 2001). Evidence is still accruing to support this. For instance, alcohol and morphine are no longer rewarding in mice lacking the D2 receptor (D2 knockout mice; Maldonado et al, 1997; Risinger et al, 2000). In humans, Volkow et al (1999) showed in a series of neuroimaging studies using cocaine or methylphenidate that increased dopamine levels in the brain were associated with euphoria and pleasure. Interestingly, low levels of dopamine D2 receptors were associated with pleasure after methylphenidate in drug-naïve individuals, whereas high receptor levels were associated with unpleasant feelings. This study gives us an insight into the role of neurobiology in explaining why drug use for some people is pleasurable and likely to be repeated and for others is unpleasant and not repeated.


The role of dopamine in addiction is now recognised as critical in anticipation and withdrawal as well. In an elegant series of experiments, Schultz (2001) found that in primates trained to associate a cue with a pleasurable experience (food), increased dopaminergic activity was seen in response to the cue and not to the food. If the food was not then presented, dopaminergic function dropped. Reduced dopaminergic function is thought to be associated with negative affect (e.g. dysphoria). Thus, an individual with an addiction may see a ‘ cue’ (e.g. a public house, mirror or needle) and if their drug of choice is not available may feel dysphoric, which is likely to increase the drive to obtain the drug.


Reduced dopaminergic function has been seen in withdrawal and early abstinence from many drugs of misuse. Neuroimaging studies in cocaine, opiate and alcohol addictions have revealed reduced levels of dopamine D2 receptors, which may recover to some extent during abstinence, but have been shown to persist for months (Volkow et al, 1999). Early stages of abstinence are associated with elevated levels of craving, drug-seeking and risk of relapse, and it is likely that hypodopaminergic function plays a mediating role. Presumably the release of dopamine produced by the drug of choice provides relief from withdrawal, although this has not yet been studied.

Pharmacotherapy (Table 1)

View this table:
Table 1

Molecular targets of drugs of misuse and pharmacological approaches (current and theoretical) directed at these

Because of the pre-eminence of the dopaminergic reward system in addiction, this has been a target for pharmacotherapy, but with mixed results. One strategy, for instance, has been to block the binding of cocaine to the dopamine transporter site (Nutt, 1993). In cocaine addiction, the development of dopaminergic partial agonists at the D3 receptor, such as BP-897, currently holds some promise. In rats, BP-897 inhibits cocaine-seeking behaviour in response to cues (Pilla et al, 1999). As a partial agonist, this drug stimulates the D3 receptor enough to keep withdrawal at bay, but not enough to cause a ‘high’ or to be rewarding. It is currently in phase 1 trials.

One drug that affects the dopaminergic system and has proven efficacy in the treatment of nicotine addiction is bupropion (Jorenby et al, 1999). The exact mechanism underlying this effect still has to be fully characterised; however, it has been shown that bupropion increases dopamine and noradrenaline levels by acting as an uptake inhibitor (Ascher et al, 1995).

Related systems involved in reward

Our understanding of other neurotransmitter systems that are involved in reward and that may modulate dopaminergic activity provides further targets for pharmacotherapy.


The opioid system has three receptor subtypes: mu, kappa and delta. The mu subtype appears to be key in opiate addiction: for mice lacking this receptor, morphine is no longer rewarding or reinforcing (Kieffer, 1999). In addition, a morphine withdrawal syndrome is not seen in these animals. Neuroimaging studies suggest that alterations in mu opiate receptor levels may be fundamental to addiction. Using [11C]-carfentanil positron emission tomography (PET) to label mu opiate receptors in the brain, Zubieta et al (2000) found increased receptor levels in the anterior cingulate in recently abstinent humans addicted to cocaine or opiates. This may reflect elevated mu opiate receptor levels or decreased endogenous opioid levels. In either case, craving may result.

Roles for kappa and delta opiate receptors in addiction are also evident. Unlike mu receptors, kappa receptor stimulation reduces dopamine function in the nucleus accumbens. This may possibly result in dysphoria. In animal models, delta antagonists can reduce self-administration of alcohol, suggesting that this receptor also plays a key role in reinforcement.

Naltrexone is a long-acting opiate antagonist. Its use in opiate addiction is based on its ability to antagonise any effects of opiates. However, in alcoholism the efficacy of naltrexone is thought to be a consequence of its ability to block the actions of endorphins that are released by alcohol and that mediate pleasure (Herz, 1997).


Glutamate is the brain's principal excitatory neurotransmitter for which there are three receptors — the ion channels N-methyl-D-aspartate (NMDA), alpha-amino-3-hydroxy-5-methyl-isoxazole-4-propionate (AMPA) and kainate — and also another receptor family which is coupled to G-proteins and the second (metabotropic) messenger system. Glutamatergic neurons from the prefrontal cortex and amygdala project onto the mesolimbic reward pathway, from which reciprocal dopaminergic projections arise (Louk et al, 2000). There is evidence that the glutamatergic projection from the prefrontal cortex to the nucleus accumbens plays a role in the reinstatement of stimulant-seeking behaviour.

The NMDA receptor has been implicated in nicotine, ethanol, benzodiazepine and cannabinoid addiction (Wolf, 1998). For example, NMDA antagonists inhibit sensitisation (i.e. enhanced responses) to stimulants such as cocaine and amphetamine and the development of opioid dependence. Not all NMDA antagonists are clinically useful, owing to their psychomimetic properties (cf. ketamine, phencyclidine). Nevertheless, memantine is a non-competitive NMDA receptor antagonist, used to treat neurological disorders, which has recently been shown to attenuate naloxone-precipitated withdrawal in humans addicted to opiates (Bisaga et al, 2001).

There is recent evidence to suggest an important role for other glutamate receptors, such as the metabotropic receptor, that may be independent of the dopaminergic system. In mice lacking the mGlu5 subtype of the metabotropic glutamatergic receptor, cocaine still increases dopamine in the nucleus accumbens; but the mice do not self-administer cocaine or show increased locomotor activity (Chiamulera et al, 2001).


Opioids and cannabinoids share some pharmacological properties producing effects such as sedation, hypothermia and anti-nociception. In addition, there is increasing recognition that opiate—cannabinoid interactions are important in drug addiction, although their precise nature remains to be characterised. The most potent cannabinoid in cannabis is Δ 9-tetrahydrocannabinol (Δ9-THC) (Ashton, 2001). Cannabinoids have been shown to increase opioid synthesis and/or release (Manzanares et al, 1999). This may explain why opiate antagonists block some effects of cannabis and induce withdrawal in Δ9-THC-dependent rats or, conversely, why marijuana may reduce opiate withdrawal.

There are two cannabinoid receptors: CB1 in the brain, for which the endogenous compound is anandamide, and CB2 on immune cells. CB1 receptors are widely distributed throughout the brain, but particularly in the cerebral cortex, hippocampus, cerebellum, thalamus and basal ganglia (Ameri, 1999). In mice lacking the CB1 receptor, rewarding and withdrawal responses to morphine and cannabinoids but not to cocaine are reduced (Ledent et al, 1999; Martin et al, 2000). This suggests that the CB1 receptor is involved in dependence on not only cannabinoids but also opiates. As a result, CB1 agonists may have clinical utility in treating opiate addiction.

The development of a CB1 receptor antagonist, SR141716A (Rinaldi-Carmona et al, 1995), not only accelerated research into cannabinoids but also provided a possible treatment. This antagonist blocks both the physiological and psychological effects of smoked marijuana and therefore could be to cannabis what naltrexone is to heroin.


The neurobiology of alcoholism involves many different neurotransmitters, but key are the gamma-aminobutyric acid (GABA)-ergic system and the glutamatergic system (Nutt, 1999). In alcohol withdrawal, increased glutamatergic NMDA function is present and is thought to be involved in seizures and cell death, by means of increased Ca2+ influx through its channel and low Mg2+. The hippocampus appears to be a critical site for such glutamatergic hyperactivity. Acamprosate, a taurine derivative, is increasingly used to maintain abstinence from alcohol as it has been shown to double abstinence rates. How acamprosate achieves its therapeutic effect has not yet been fully characterised; it antagonises the NMDA receptor (possibly through the polyamine site). Acamprosate also reduces glutamate levels and may be neuroprotective (Dahchour & De Witte, 2000). If such neuroprotection occurs in humans, this would have important implications for the treatment of alcoholism; currently some workers advocate starting acamprosate with detoxification.


As described above, the mu opiate receptor plays a key role in opiate reward, but many of the mechanisms underlying opiate tolerance, dependence and withdrawal remain elusive. As the opiate receptor may not change with chronic opiate exposure, changes ‘downstream’ of the receptor may be more critical. For example, noradrenergic overactivity is seen in opiate withdrawal and can be treated with α2 agonists such as lofexidine or clonidine (Strang et al, 1999).

In the treatment of opiate addiction, methadone is the most commonly prescribed drug, although the use of buprenorphine is increasing. Methadone (like heroin) is a full agonist at the mu receptor, whereas buprenorphine is a mu partial agonist. Partial agonists give lower levels of response at maximal receptor occupancy. Also, when a partial agonist occupies receptors, fewer are available for a full agonist (e.g. heroin). The partial agonist is therefore acting as an antagonist. Consequently, buprenorphine will stimulate the mu opioid receptor, but not maximally (hence, there is less risk of respiratory depression in overdose), and will also prevent the effects of heroin taken ‘ on top’. In addition, its longer half-life allows less than daily dosing, an advantage in supervised consumption.


Ecstasy (3,4-methylenedioxymethamphetamine or MDMA) and its derivatives MDA (Adam) and MDEA (Eve) have both stimulant and hallucinogenic properties. Acutely, MDMA increases 5-hydroxytryptamine (5-HT or serotonin) levels, and, to a lesser extent, dopamine levels, by stimulating release and inhibiting uptake.

Animal studies have revealed ecstasy and its derivatives to be neurotoxic to serotonergic neurons (MDA>MDMA>MDEA), but it is controversial whether and to what extent the same occurs in man (Boot et al, 2000). Neuroimaging studies using PET and single photon emission tomography (SPET) to measure 5-HT transporter levels in persons who are regular heavy ecstasy users report reduced levels. However, methodological questions about the tracer, contribution of blood flow and choice of subjects necessarily limit these conclusions (Semple et al, 1999; Reneman et al, 2001). There is some evidence for cognitive impairments in individuals using ecstasy which may persist after a period of chronic use, and it is not clear how reversible these are with time. In animal models, fluoxetine has been shown to be neuroprotective, apparently by blocking ecstasy uptake into 5-HT neurons, but it is unknown whether this protective effect occurs in humans.


The most widely misused group of drugs acting on this system are the benzodiazepines. These modulate the GABA—benzodiazepine receptor, increasing the action of GABA, and so result in greater inhibitory activity in the brain (Nutt & Malizia, 2001). In contrast to other drugs of misuse, benzodiazepines do not increase dopamine release in the mesolimbic system. Misuse of these drugs is probably driven by the development of tolerance leading to withdrawal if these drugs are not taken. Benzodiazepine dependence in the context of drug addiction, where large doses of benzodiazepines are taken, is distinct from dependence in the context of long-term use of a prescribed benzodiazepine for anxiety.

Gamma-hydroxybutyrate (GHB) is a short-chain fatty acid which, among other effects, enhances GABAergic function. GHB inhibits central nervous system activity and is a sedative but is also euphorigenic, presumably being linked to an increase in dopamine (Nicholson & Balster, 2001). It is increasingly used as a ‘ recreational club drug’ and there is growing concern about its safety, particularly when combined with alcohol to render women vulnerable to sexual assault.


This is an exciting time in addiction as the neurobiology of addiction disorders becomes clearer. Such characterisation not only provides a greater understanding of why people become addicted and what happens to the brain after a period of substance misuse, but also allows better understanding of current pharmacotherapies and, we hope, the development of new treatments.

  • Received January 22, 2002.
  • Revision received May 22, 2002.
  • Accepted May 29, 2002.