Pharmacotherapy, pharmacogenomics, and the future of alcohol ...

CLINICAL REVIEWS

Alcohol dependence treatment, part 2

CLINICAL REVIEWS

Pharmacotherapy, pharmacogenomics, and the future of alcohol dependence treatment, part 2 GEORGE A. KENNA, JOHN E. MCGEARY, AND ROBERT M. SWIFT Serotonin receptors and potential agents Alcoholics and experimental animals that consume large amounts of alcohol have comparatively lower levels of serotonin than nonalcoholics.88 Considerable experimental evidence suggests that serotonin plays a crucial role in the impulsivity and craving often seen in alcoholics89 and is at least partly responsible for alcohol dependence.90,91 Hypotheses for a relationship between serotonin and alcoholism91 note (1) the impulsive use of alcohol associated with hyposerotoninergic levels and increased dopaminergic activity,90 (2) the possible relationship between symptoms of anxiety and hyposerotoninergic states leading to the use of alcohol to self-medicate,93 (3) the ability of alcohol to increase serotoninergic activity in the hyposerotoninergic brain,94 (4) the interaction of serotonin with dopamine in contributing to the rewarding properties of alco-

Purpose. The neurobiological basis of alcohol dependence, established pharmacotherapies for alcohol dependence, pharmacotherapies under investigation, and obstacles to treatment are discussed. Summary. Alcohol binds to hydrophobic pockets of proteins, changing their threedimensional structure and their function. Proteins that are particularly sensitive to alcohol include ion channels, neurotransmitter receptors, and enzymes involved in signal transduction. Established pharmacologic treatments, notably disulfiram and naltrexone, combined with behavioral therapies, may reduce the amount of drinking, the risk of relapse, the number of days of drinking, and craving in some alcohol-dependent individuals. For many patients, however, these treatments are not effective. Recent advances in molecular and behavior genetics are guiding the development of new drugs; these efforts seek to identify pharmacologic pathways relevant to alcohol dependence and to more effectively match treatments to individuals according to their genetic characteristics. Efficacy and safety concerns for acamprosate have been satisfied; the drug

GEORGE A. KENNA, PH.D., B.S.PHARM., is Fellow in Community Health, Center for Alcohol and Addiction Studies (CAAS), Brown University, Providence, RI, and Clinical Pharmacist, Kent County Memorial Hospital, Warwick, RI. JOHN E. MCGEARY, PH.D., is Postdoctoral Fellow in Community Health, CAAS. ROBERT M. SWIFT, M.D., PH.D., is Associate Chief of Staff for Research and Education, Providence VA Medical Center, Providence; Associate Director, CAAS; and Professor of Psychiatry, Brown Medical School, Providence.

2380

Am J Health-Syst Pharm—Vol 61 Nov 15, 2004

was recently released for marketing in the United States. Medications such as sertraline, ondansetron, topiramate, and aripiprazole represent novel lines of research and are currently being tested for use in the treatment of alcoholism. Even with more efficacious medications, however, a transformation must occur in how alcoholism treatment is viewed, not only by the public but also by clinicians. Conclusion. In addition to existing drug treatments for alcohol dependence, many other medications are under investigation, particularly for specific types of alcoholism. Pharmacogenomics is expected to play an important role in this research effort. Index terms: Acamprosate; Alcohol deterrents; Alcoholism; Alcohols, ethyl; Anticonvulsants; Antidepressants; Antiemetics; Antipsychotic agents; Aripiprazole; Behavior therapy; Binding; Dependence; Disulfiram; Mechanism of action; Naltrexone; Ondansetron; Opiate antagonists; Pharmacogenetics; Research; Sertraline; Topiramate; Toxicity Am J Health-Syst Pharm. 2004; 61:2380-90

Address correspondence to Dr. Kenna at 457-West, Roger Williams Medical Center, 825 Chalkstone Avenue, Providence, RI 02908 ([email protected]). This is article 204-000-04-013-H01 in the ASHP Continuing Education System; it qualifies for 1.5 hours of continuing-education credit. See page 2389 or http://ce.ashp.org for the learning objectives, test questions, and instructions. Part 1 of this article appeared in the November 1, 2004, issue.

CLINICAL REVIEWS Alcohol dependence treatment, part 2

hol, and (5) the possibility that medications that affect serotonin might have an effect on appetitive behaviors through the renin–angiotensin system.92 Differences in the response to alcohol may also be related to subtypes of serotonin receptors. The pharmacology of serotonin and its many receptors has been found to contribute to alcohol consumption in animal studies.88 Moreover, in humans, in addition to and operating concurrently with dopamine release, alcohol has many actions on the brain through serotoninergic systems. 9 However, to complicate matters, there are many types of serotonin receptors. 5-HT1A-receptors may be associated with alcohol consumption, intoxication, and the development of tolerance; 5-HT2-receptors have been found to contribute to reward systems; and 5-HT3-receptors are implicated in the development of reinforcement through synaptic projections to the mesolimbic dopaminergic system.88 Alcohol increases brain levels of serotonin affecting these receptors. In addition, the concentration of the serotonin metabolite 5-hydroxyindoleacetic acid may be lower in the cerebrospinal fluid of individuals with suicidal ideation, impulsivity, and depression, as well as alcoholism, suggesting some potential for dysregulation of serotonin.95,96 Perhaps because of the multiplicity of types of serotonin receptors, studies of serotoninergic medications have provided mixed results. For example, although buspirone is a D2-receptor antagonist, its primary action is as a 5-HT1A-receptor partial agonist that exhibits anxiolytic properties. Studies have found that groups receiving buspirone, while also suffering from coexisting anxiety disorders, scored significantly better than placebo recipients on alcohol-use measures. 97-99 When these anxiety disorders were controlled for, however, buspirone was

not superior to placebo,49 rendering efficacy for alcohol dependence treatment alone questionable. 55,100 Ritanserin, a 5-HT2-receptor antagonist, also failed to clearly demonstrate usefulness as a pharmacotherapy for alcohol dependence in a doubleblind, placebo-controlled trial.101 The findings of clinical trials of specific serotonin-reuptake inhibitors (SSRIs) are also uncertain.55 Despite reductions in drinking in animal studies 102 and some limited human studies, 51,102,103 SSRIs have not reduced drinking or any other measures of alcohol consumption in most double-blind, placebocontrolled studies. However, frequencies of genetic variation in the serotonin transporter differ between persons with alcohol dependence and those with comorbid alcohol dependence and depression, suggesting that patients with this comorbidity may constitute an important subpopulation that might respond differently to treatment.104 Recent research suggests that SSRIs do not appear to be an efficacious approach to treating heterogeneous groups of alcoholics,105 since subtypes of alcoholics may respond differently to SSRIs. Although Kranzler et al.50 observed that fluoxetine hydrochloride 60 mg was no better than placebo in reducing alcohol consumption, two types of alcoholics, A and B, responded differentially to fluoxetine: While the type A alcoholics drank the same amount as the group receiving placebo, the type B alcoholics drank significantly more than those receiving placebo. Persons with type A alcoholism have a later onset of dependence (after 25 years of age), few pathological problems, and few drinking-related problems and childhood risk factors. Type B alcoholism, an early-onset form occurring before the age of 25, involves a more severe psychopathology, antisocial and impulsive tendencies, childhood risk factors, and significant problems with alcohol.

In a follow-up to the Kranzler et al.50 study, Pettinati and colleagues52 used sertraline hydrochloride 200 mg and classified subjects as having early-onset or late-onset alcoholism. This study confirmed that those with early-onset alcoholism actually drank significantly more when taking SSRIs, but, unlike the Kranzler et al.50 study, also found that those with late-onset alcoholism drank significantly less. It appears that a subgroup of alcohol-dependent individuals who lack certain predisposing factors, such as family history, may achieve some success with SSRIs. Rationale for ondansetron Functional control of the serotoninergic system may be regulated by genetic differences in the serotonin transporter gene mediating serotoninergic neurotransmission by controlling its strength and duration. The serotonin transporter 5HTT polymorphism, 5HTTLPR, is a common variable number of tandem repeats in the 5HTT promoter region that alters transcription.14 There are two major variants of 5HTT, the long (LL) and short (SS) versions. Since humans have two copies of each chromosome, three possible genotypic combinations result: homozygosity (two copies) for the short version (SS), homozygosity for the long version (LL), and heterozygosity (SL). While the importance of the SL version of the gene is unknown, the LL version is associated with three times more serotonin uptake from platelets and lymphoblasts than the SS version. Persons with the LL variant may have increased numbers of 5HTT-receptors and reduced levels of intersynaptic serotonin, affecting in particular 5-HT3-receptors in persons with early-onset alcoholism as a result.106 These individuals may be more likely to have the LL genotype, resulting in more rapid serotonin removal and attenuated dopaminergic functioning. Removing serotonin


2381

CLINICAL REVIEWS


from the synapse more vigorously than normal may contribute to upregulation of 5-HT3-receptors. These postsynaptic receptors project to dopaminergic neurons in the ventral tegmental area, which in turn project to the nucleus acccumbens via the mesolimbic tract. As a result of alcohol use, an increased release of dopamine associated with facilitated reinforcement may occur. A 5-HT3-receptor antagonist decreases dopamine release and thus reinforcement.82 The clinical outcome of this effect is hypothesized to be down-regulation of the dopaminergic neurons that project to the nucleus accumbens, enhanced dopaminergic functioning, and decreased reward from alcohol use.53 The 5-HT3-receptor antagonist ondansetron currently carries FDAapproved labeling for use in the treatment of cancer chemotherapyinduced and postoperative nausea. Ondansetron, which has effects functionally opposite those of SSRIs and blocks serotonin, may potentially be a new treatment for the same biologically predisposed population of type B alcoholics who do not respond to SSRIs. While further evidence is needed, early studies suggest that ondansetron may be an important adjunctive treatment for early-onset (type B) alcoholism. Sellers et al.94 found that patients taking ondansetron hydrochloride 4 µ g/kg twice daily drank significantly less alcohol than groups given 1 or 16 µg/ kg twice daily or placebo, suggesting a U-shaped dose–response relationship. Johnson et al.53 divided their subjects into those with early-onset and late-onset alcoholism consistent with previous studies.50,52,94 The patients with early-onset alcoholism taking ondansetron hydrochloride 4 µg/kg twice daily reported a lower cumulative percentage of drinking days than placebo recipients. In summary, as demonstrated in the studies of SSRIs, the variability of

2382

responses with individual factors, such as time of onset of alcoholism, suggests a role of different serotoninreceptor types in the variable response of alcoholism to treatment. Future studies are needed to further explicate the underlying differences among forms of alcoholism so that reasonable predictions about the outcomes of SSRI treatment can be made.105 Dopaminergic receptors and potential agents Current research related to the neurobiology of substance abuse supports the role of dopaminergic involvement in motivation, reward, and reinforcement.107 In a review of antidipsotropics, Litten and Allen108 suggest that the potential of dopaminergic agents to reduce alcohol consumption remains unexplored. A strong body of evidence indicates that alcohol activates dopamine release at the nucleus accumbens and extended amygdala.109,110 The action of alcohol on the mesolimbic pathway is considered to be strongly associated with susceptibility to alcoholism111 and the development of craving and loss of control.112 Indeed, the ability of medications such as naltrexone and even ondansetron to reduce drinking appears to be mediated by dopamine.113,114 During alcohol withdrawal in alcohol-preferring rats, reductions in dopamine levels and in the population of D2-receptors have been identified.115,116 Dopamine levels in alcohol-preferring rats increase when they are placed in the same specific environment in which they previously received alcohol (i.e., dopamine levels rise in anticipation of receiving alcohol).117 Alternatively, reduced D 2-receptor function in alcoholpreferring rats may favor alcohol self-administration.115 Imaging studies corroborate the results of these animal studies by showing that populations of D2-receptors are small in alcoholics118 and may contribute to addiction to alcohol and other


drugs.119 In contrast, administration of a D2-receptor adenovirus vector into the nucleus accumbens is associated with a significant reduction in alcohol intake in alcohol-preferring rats and rats not preferring alcohol, suggesting that overexpression of D2receptors may indeed be a protective factor against alcohol abuse and presumably other forms of drug abuse.120 The expression of D2-receptors can also be modulated by environmental factors (e.g., social dominance in monkeys121), pointing to the importance of developing alternative, nonmedicinal strategies to increase the expression of D2-receptors in drugabuse treatment.122 Strategies aimed at increasing D2-receptor levels (upregulation at the nucleus accumbens) could be beneficial, particularly during continued abstinence from alcohol. Neuroleptics regulate dopamine occupancy at D2-receptors, putatively causing their up-regulation.123,124 Neuroleptics. Similar dopaminergic dysfunction seems to be shared by reward associated with alcohol cues influencing dopamine release by the mesolimbic pathway and irrelevant stimuli interpreted as positive symptoms of schizophrenia (e.g., delusions).125 Neuroleptic regulation of D2-receptor occupancy by dopamine may cause their up-regulation,123,124 which could potentially reduce the positive symptoms of schizophrenia124 and reduce substance use.120 This relationship could be a strong factor in the frequent coincidence of substance abuse and schizophrenia.5 Neuroleptics such as haloperidol, 126 tiapride,127 olanzapine,128,129 and clozapine130 have all demonstrated various degrees of efficacy in reducing alcohol craving or consumption or in increasing abstinence. In a doubleblind, placebo-controlled trial, 16 patients diagnosed with alcohol abuse or dependence were pretreated with placebo or the D2-receptor antagonist haloperidol.126 Measures of craving and impaired control were significantly attenuated in the exper-


imental group. In another clinical trial, drinking among 54 alcoholdependent patients treated with tiapride, a D2-receptor antagonist, was significantly reduced compared with a group given placebo.127 Olanzapine, an antagonist of D2-, D4-, and 5-HT2-receptors, also reduces craving for alcohol. Although 5 mg of olanzapine did not influence the rewarding effects of alcohol in one study, it did moderate alcohol cues and the urge to drink.128 In a subsequent study examining whether a variable number of tandem repeats in the gene coding for the D4-receptor acts as a moderator of craving, olanzapine, compared with cyproheptadine—a sedating active placebo—reduced craving after exposure to alcohol cues and after a priming dose of alcohol in participants with at least one copy of the long version (seven repeats) of the gene compared with those who had copies of only the short version (four repeats).131 Clozapine, the first atypical antipsychotic, is a medication with multiple actions at D1-, D2-, and D4receptors, as well as serotoninergic, adrenergic, cholinergic, and histaminergic systems. In a retrospective study of 151 schizophrenic patients with alcohol and drug-use disorders, 36 of whom received clozapine, 79% of the patients given clozapine were diagnosed as being in remission from alcohol-use disorder for six months or longer, compared with 33% of those not taking clozapine.132 Additional information showing success in the use of quetiapine for treatment of comorbid schizophrenia, alcohol, and drug dependence suggests that other dopaminergic agents may play important roles in substance-abuse treatment as well.133 To date, studies involving antipsychotics have been limited by their adverse effects: Haloperidol may cause neurologic movement disorders; tiapride may induce neurologic adverse effects, neuroleptic malignant syndrome,132 and seizures134; olanza-

pine may cause significant weight gain,135 increased triglyceride levels,136 and diabetes137; and clozapine may induce weight gain135 and agranulocytosis.138 Furthermore, the use by schizophrenics of certain antipsychotics with substantial adverse effects may actually contribute to greater substance use in an effort to self-medicate.139,140 Dopamine’s importance in addiction, therefore, suggests that an important area for future research may involve dopaminergic medications with more modest adverse effects. Partial dopamine agonists. A fairly new class of drugs, partial dopamine agonists (PDAs), have a high affinity for dopamine receptors but low intrinsic activity.141 The result is that these drugs act as antagonists under conditions of hyperdopaminergic availability and as agonists under conditions of hypodopaminergic availability.142 In short, their activity is flexible, depending on the substrate and the level of dopaminergic synaptic activity. A rough analogy is that dopaminereceptor antagonists act to completely shut a door—preventing dopamine stimulation at the dopamine receptor, while PDAs wedge the door partway open, allowing some but not all possible dopaminergic neurotransmission. When there is a hyperdopaminergic state, the agonist blocks a majority of available dopamine, while during a hypodopaminergic state, the agonist facilitates some stimulation of the receptors. The actions of PDAs on substance use have been studied in animals. For example, terguride acts as a dopamine-receptor antagonist in rats trained to self-administer cocaine 143 and blocks amphetamine self-administration in rats 144 and cocaine- and food-maintained behavior in monkeys. 145 Consistent with the reduction of reinforcement associated with other drugs, terguride also significantly reduces ethanol intake in rats.143 The unique

mechanism of PDAs may become even more salient in conditions of rebound reduction in dopaminergic activity in the mesolimbic area that may be associated with ethanol withdrawal.146,147 Some researchers suggest that PDAs may provide a means of reversing dopamine depletion during ethanol withdrawal and, because of its flexible activity, may represent a novel strategy for normalizing dopaminergic neurotransmission in the behavioral continuum of alcoholism.143 Multiple receptors and potential agents Aripiprazole. Like most other atypical neuroleptics, aripiprazole binds with high affinity to the D2 family of receptors.148,149 The action of aripiprazole contrasts with currently marketed dopaminergics, however, since it may exert its effects through partial agonism of dopamine autoreceptors at D2L-receptors.150,151 Reduced D2-receptor function may play a key role in alcoholism, a hypothesis supported by studies showing that alcohol-preferring rats have moderate to low D2-receptor levels152,153 and are considered to be strains with low brain dopamine activity.154-157 Moreover, systemic administration of dopamine agonists in these strains markedly reduces ethanol intake.158,159 Aripiprazole, therefore, has the propensity to act as both a dopamine agonist in areas of hypoactivity and as an antagonist in areas of hyperactivity. Administration of this drug in rats results in decreased extracellular dopamine levels and dopamine release in the frontal lobe and striatum consistent with presynaptic agonism.160 Aripiprazole also displays dopamine antagonistic properties and has been found to block apomorphine-induced stereotypy and locomotor activity in rats. 149 While no studies in animals or humans have tested the utility of aripiprazole for the treatment of alcohol dependence, other PDAs have


2383

CLINICAL REVIEWS


been shown to reduce stimulant and alcohol use in animals.143-145 The prefrontal cortex contains a robust population of both 5-HT1Areceptors and 5-HT2A-receptors located principally on pyramidal neurons. 151 Aripiprazole is a partial agonist of 5-HT1A-receptors, an effect associated with anxiolytic action,161 and an antagonist of 5-HT2Areceptors, an effect associated with improvements in cognition and control of agitation, aggression,162, 163 and depression.164 5-HT2A-receptors are abundant in the limbic forebrain, frontal cortex, and nucleus accumbens.165 Long-term administration of most antidepressants with 5HT2A-receptor antagonist activity induces down-regulation of 5-HT2Areceptors, which is believed to be an important mechanism in the action of antidepressants.165 Aripiprazole is also an inverse agonist of 5-HT2Breceptors151 found to be important in the treatment of migraine headaches.166 More important, agonism of this receptor was previously implicated in valvular heart disease associated with the recalled weight-loss drug dexfenfluramine.167 Because of its mechanism of action, aripiprazole has no such reports. Multiple types of serotonin receptors have been implicated in the reinforcing effects of alcohol. For example, amperozide, a 5-HT2A-receptor antagonist, was found to cause persistent and irreversible suppression of alcohol intake in both cyanamideinduced and genetically selectedalcohol preferring and high-alcoholdrinking rats.168 In rats, drugs that are considered second-generation amperozide-like drugs, which are both 5-HT2-receptor and 5-HT1Areceptor agonists, have also demonstrated the ability to decrease reinforcement by alcohol.164,169,170 Alcohol drinking was significantly reduced in rats that were administered the mixed 5-HT 1A -receptor agonist/ 5-HT2A-receptor antagonist FG 5893. Moreover, another mixed 5-HT1A2384

receptor agonist/5-HT2A-receptor antagonist, FG 5938, also significantly reduced drinking in alcohol-preferring rats. These data suggest that compounds that are both 5-HT1A-receptor agonists and 5-HT2A-receptor antagonists may be important agents for the treatment of alcoholism. The activity of aripiprazole is presumed to be primarily due to the parent drug and less to its metabolite, dehydroaripiprazole, which has an affinity for D2-receptors similar to that of aripiprazole. The mean halflives of aripiprazole and the metabolite are 75 and 94 hours, respectively. Steady state is reached in 14 days. Elimination occurs primarily via cytochrome P-450 (CYP) isoenzyme 2D6 and CYP3A4. Because of the involvement of the P-450 isoenzyme system, drug–drug interactions could occur, and dosage adjustments may be necessary. However, no adjustment of the aripiprazole dosage is required for patient age, sex, race, smoking status, or hepatic or renal function. Unlike the previously noted dopaminergic agents used in alcohol treatment studies, aripiprazole, because of its unique mechanism of action, causes few extrapyramidal adverse effects. In both short (4-week) and long (52-week) trials, aripiprazole was associated with fewer extrapyramidal adverse effects and better tardive dyskinesia scores than haloperidol.171,172 Aripiprazole also lacks strong histaminergic activity that may cause sedation.173 The drug causes relatively little elevation of prolactin levels or QTc-interval prolongation.171,174 One major benefit of aripiprazole in clinical practice is that, unlike the 5-HT2C-receptor antagonists clozapine and olanzapine, it has no propensity to induce weight gain. Weight gain has been postulated to be associated with a polymorphic allele of the gene for the 5-HT2Creceptor175; aripiprazole is a partial agonist of this receptor, an action associated with antiobsessional and anorectic effects in humans.151 In a


28-week trial, patients treated with aripiprazole lost weight on average, and patients given olanzapine gained weight.176 Clozapine, while an important antipsychotic of last choice, has a major limitation, the risk of agranulocytosis. The risk of agranulocytosis with aripiprazole is no greater than that with the typical neuroleptics.173 Bristol-Myers Squibb is funding a double-blind, placebo-controlled trial of aripiprazole in treatmentseeking, alcohol-dependent patients. Treatment-seeking subjects generally must be abstinent for at least 72 hours, but not more than 14 days, prior to randomization. By contrast, non-treatment-seeking individuals, while expressing a desire to reduce their alcohol consumption, are not asked to stop their alcohol consumption prior to randomization. Topiramate. Various studies implicate the involvement of multiple neurotransmitters in the reward circuit that becomes dysregulated during the development of alcohol dependence,177 including GABAA,178 glutamate,179 and dopamine.180 Individuals affected by alcoholism might have more glutamate-binding sites, including AMPA and kainate receptors, in the hippocampus—an effect that could further result in facilitated dopaminergic neurotransmission in the midbrain.54,179,181 Topiramate is an antiepileptic with multiple mechanisms of action, including enhanced GABAA-receptor inhibition,182 resulting in decreased facilitation of the activity of dopamine in the midbrain—an effect believed to be of potential benefit in the treatment of addiction.183 While there is no effect on the NMDA receptor, there is antagonism of kainate to activate kainate and AMPA receptors184 and inhibition of types II and IV carbonic anhydrase isoenzymes. A medication that could simultaneously address multiple receptors affected by chronic alcoholism would represent a significant advance.


Little research on the use of topiramate for alcoholism is available. Anecdotal reports indicate that topiramate curbed the craving of some patients for alcohol.185 Johnson and colleagues54 used a dosage escalating from 25 to 300 mg/day or matching placebo in 150 male and female nontreatment-seeking alcoholics during the first 8 weeks of a 12-week study period. Patients continued to receive the dosage achieved at 8 weeks over the final 4 weeks of the study. To be included in the study, men had to report drinking at least 35 drinks per week and women at least 21 drinks per week during the 90-day period preceding randomization for this double-blind trial. Topiramate recipients reported significantly fewer drinks per day, drinks per day of drinking, and drinking days; significantly more days of abstinence; and significantly less craving than placebo recipients. Given topiramate’s desirable adverse-effect profile, coadministration with other medications, such as naltrexone, acamprosate, ondansetron, or SSRIs, may produce enhanced efficacy outcomes, particularly for certain types of alcoholism. Furthermore, since abstinence was not a goal at the start of the topiramate study by Johnson et al.,54 the medication may be more beneficial during the abstinence-initiation phase of treatment.186 Peak plasma topiramate concentration in humans occurs about two hours after oral administration. The mean plasma elimination half-life is 21 hours, and steady state is reached in about four days. Since the drug is renally excreted, dosages must be adjusted downward by 50% for patients with a creatinine clearance of