Cocaine

Chris-Ellyn Johanson and Charles R. Schuster

Cocaine remains a major drug of abuse in the view of the public and is blamed for many of the nation's ills, including the deterioration of families and cities and increases in violent crime. Neuropsychopharmacological research has focused on cocaine, and several reviews of its actions have been written in the last decade (34, 48, 84, 96). The present review concentrates on areas of research where significant advances have been made since the Third Generation of Progress was written and assumes that the reader is familiar with earlier literature.

At the time of the writing of the Third Generation of Progress, cocaine use in the United States was near peak levels. National surveys conducted over the last two decades indicated that past-year and past-30-day prevalence rates reached a peak in 1985 and then steadily declined, reaching the lowest levels in the most recent National Household Survey on Drug Abuse (NHSDA) in 1992. For example, the 1985 NHSDA reported that almost 6 million people had used cocaine (including crack) at least once in the past 30 days. By 1991, this number had decreased by over 70% to less than two million. This decline was found at all age levels, including the 12-to 17-year-old age group. On the other hand, among respondents reporting use in the past year, the numbers of those who had used it at least once a week or more remained the same (between 650,000 to 800,000) over this same time period. Thus, the more frequent, problematic use of cocaine remains at peak levels.

Another United States national survey of drug use, called Monitoring the Future (MTF), which has been conducted since 1975, uses high school seniors as respondents (63).†2 This survey has also shown that the use of cocaine (including crack) has declined over the same period. It must be recognized, however, that the prevalence rates for the use of cocaine are most likely higher among those who dropped out of school. Unfortunately, systematic data on dropouts are not available, although the NHSDA provides evidence that they are a vulnerable population. Nevertheless, it is encouraging that the overall trend noted by both the MTF and the NHSDA indicates a decline in cocaine use in adolescents.

In addition to decreases in overall prevalence, there has been a change in the demographic characteristics of those who use cocaine today, compared to the late 1970s and early 1980s. Although cocaine was then the "drug of choice" among the "elite," its use has become increasingly prevalent in other portions of the population. Data from the 1991 NHSDA showed that the past-year and past-month prevalence rates for cocaine use were significantly higher in those who had not completed high school compared to those who had received a high school diploma, supporting the concern that the MTF may be underestimating actual numbers of users. These differences were particularly striking for crack cocaine use, where there was a fourfold greater past-month prevalence rate among those who did not finish high school in comparison to college graduates. This is in contrast to the situation in 1982 when high school and college graduates were two to three times more likely to have used cocaine in the past 30 days than age-matched respondents who had not finished high school. Furthermore, in 1991, cocaine use in the past month among the unemployed was 4.6 times higher than in the employed. This difference was even greater with crack cocaine. Recent statistics from the National Institute of Justice show that the rates of cocaine-positive urines among both male and female arrestees in major cities remain near peak levels of 40-80%. It would thus appear that the decreasing trend in prevalence rates of cocaine use observed in the general population is absent or significantly less in the educationally, economically, and socially disadvantaged. These prevalence rates must be put in proper perspective, however, by noting that these disadvantaged groups are relatively small in numbers. Thus, it remains the case that the majority of cocaine users are employed high school graduates who would not be characterized as disadvantaged.

Although the yearly trends in the NHSDA and in the MTF may very well be accurate, it is likely that both surveys underestimate the actual numbers of regular crack cocaine users. Newspaper and magazine accounts estimate the number to be somewhere between 500,000 and one million. Regardless of the exact numbers, however, the impact on the health delivery systems in already overcrowded urban hospitals has been overwhelming. Lindenbaum et al. (54) reported that more than half of the patients at the Albert Einstein Medical trauma center tested positive for cocaine and approximately 50% had been involved in a violent crime. Publicly funded treatment programs have experienced a sixfold increase from 1985 to 1990 in the numbers of people seeking treatment for their cocaine problem (3). The number of reported emergency room (ER) admissions reported to the Drug Abuse Warning Network (DAWN) associated with cocaine use has also increased dramatically over this same time span. The total number of cocaine-related ER mentions in 1992 was 119,800 (28% of ER drug-related admissions) compared to approximately 28,800 (9% of ER drug-related admissions) in 1985. At the present time, cocaine ranks a close second to alcohol-in-combination-with-other-drugs in ER mentions.

Although prevalence rates for the use of cocaine in the past 30 days were significantly lower in women than in men, the use of cocaine by women in their child-bearing years is of vital concern. Estimates of the numbers of women who use cocaine during pregnancy have varied widely. Furthermore, there are many reasons to believe that these estimates are unreliable, not the least of which is that women may deny drug use in states where they face criminal sanctions or the loss of custody of their newborn for cocaine use during pregnancy (84). That there is a significant number of fetuses exposed to cocaine during pregnancy is attested to by a recent large-scale prospective drug screening study (65). Meconium samples from 3010 neonates born to women coming to an urban obstetrics service were analyzed for the presence of cocaine. Although only 11% of the women reported the use of illicit drugs during their pregnancy, 31% of the infants' meconium tested positive for cocaine.

There has been an increasing effort to probe the meaning of differences in prevalence rates in drug use in groups varying in demographic characteristics. One demographic characteristic that has received attention in this regard is race/ethnicity. In both the 1988 and 1990 NHSDA, lifetime prevalence rates for crack cocaine use were over twice as high among black Americans compared to white Americans. However, analyses of these data designed to understand underlying factors of etiologic significance demonstrated that when the survey respondents were grouped into neighborhood clusters, thereby holding shared characteristics such as drug availability and social conditions constant, the odds of crack cocaine use did not differ by race/ethnicity (7, 53). Thus, it appears that neighborhood characteristics that promote the use of crack cocaine are equally effective, independently of racial or ethnic status. There is, however, a disproportionately larger percentage of black Americans than white Americans living in neighborhoods with these characteristics, thus leading to the higher prevalence rates. It must be remembered, however, that because of the relatively smaller size of the black American population compared to the white American population in the United States (11% versus 81%), the majority of crack users are white. Differences in reported use of crack have also been noted in Hispanic populations. In 1988, the lifetime prevalence rate for Hispanic Americans was 2.1% compared to a percentage of 1.0 for white Americans. However, there was no difference in odds ratios when neighborhood factors were controlled (53). In addition, in the 1990 NHSDA the lifetime prevalence rate of Hispanic Americans was found to have declined to 1.6%, which was not significantly different from the prevalence rate of white Americans. In fact, when Chilcoat et al. (7) adjusted for neighborhood, Hispanic Americans were found to have a significantly lower odds ratio for being crack users.

In addition to environmental influences, a variety of studies have suggested that individual-level characteristics play a role in the etiology of cocaine abuse. Studies of community samples have shown that the diagnosis of alcoholism or drug abuse/dependence is associated with higher prevalence rates of a wide variety of psychiatric disorders (43). In a recent study that evaluated the psychiatric status of individuals seeking treatment for cocaine abuse or dependence (80), both high current (55.7%) and lifetime (73.5%) prevalence rates for psychiatric disorders other than substance abuse were reported. The principal types of coexisting psychiatric disorders were: major depression, minor bipolar conditions, anxiety disorders, antisocial personality, and a history of attention-deficit hyperactivity disorder (ADHD). Prospective studies have shown that adolescents who show residual symptoms of ADHD have higher rates of substance abuse problems (58). In follow-up studies into adulthood, those diagnosed with antisocial personality showed increased rates of drug abuse (58). However, for those who reached adulthood and no longer showed symptoms of ADHD or antisocial personality, the odds of being a drug abuser were no different than those in controls (58).

One of the complications in determining the role of psychiatric disorders in the development of substance abuse problems is that drug use itself can cause psychiatric sequelae. While there is continuing debate about the etiologic significance of psychiatric disorders in the development of substance abuse, there is general agreement that substance abuse and other psychiatric problems coexist and that substance abuse treatment clients with coexistent psychiatric disorders have poorer treatment outcome. Appropriate pharmacological or psychotherapeutic treatment of the comorbid psychiatric disorder has been shown to improve the outcome of substance abuse treatment (98).

Another risk factor for the initiation of the use of cocaine is the use of other psychoactive substances that are legal, less costly, or more easily obtained. Epidemiological studies have provided evidence that there is a sequence of drug use stages leading up to the use of cocaine and opiates. Kandel and Faust (41) showed that the use of legal drugs, such as beer and cigarettes, preceded the use of marijuana, which preceded the use of cocaine. There had been a concern that young people might initiate crack cocaine smoking earlier in the developmental sequence than previously found for intranasal or intravenous cocaine because of its availability, ease of ingestion, and relatively cheap price. However, a recent study by Kandel and Yamaguchi (42) reported that crack smoking was the last member of the chain of substances to be used, after powdered cocaine itself. Furthermore, Chilcoat and Schutz (7) showed that peak use of crack cocaine occurs in individuals over 20 years of age, with very little crack use being reported by individuals under 20. Although previous drug use per se may not be an etiologic factor, early use of licit substances, such as beer and cigarettes, can be used as a marker for identifying children who are at risk of developing problematic drug use, including the use of cocaine.

The pharmacokinetics of cocaine has been extensively investigated because of the purported differences in the dependence potential and toxicity associated with different routes of administration. Although the intranasal route of administration (insufflation) remains the most common route, the smoked and intravenous routes are associated with a higher frequency of ER mentions (DAWN) and number of people entering treatment for cocaine abuse or dependence.

Pharmacokinetic studies of intravenous and intranasal cocaine have been reviewed previously (34). Following nasal insufflation of 96 mg of cocaine, peak venous plasma levels between 150 and 200 ng/ml were reached in approximately 30 min. Intravenous administration of 32 mg of cocaine produced peak venous plasma levels of approximately 250-300 ng/ml after 4 min. A similar rise to between 200 and 250 ng/ml was found in venous plasma levels following the smoking of 50 mg of cocaine base (18). Regardless of the route of administration, the elimination half-life was approximately 40 min, although longer half-lives of 60 min have been reported.

Recent studies of the pharmacokinetics of smoked and intravenously administered cocaine have compared venous and arterial plasma levels (15). When a dose of cocaine base of 50 mg was smoked or an intravenous injection of 32 mg of cocaine hydrochloride (HCl) was given, arterial plasma levels reached a peak within a few seconds whereas venous plasma levels did not reach their peak until after several minutes. Furthermore, the peak arterial levels reached by either route of administration were substantially higher than the peak venous plasma levels. Thus, the levels of cocaine reaching the heart and brain after either smoking cocaine base or the intravenous administration of cocaine HCl were much greater than would have been predicted from prior data on venous plasma levels.

Cocaine is metabolized by cholinesterases present in the plasma and liver into two principal metabolites, benzoylecognine and ecgonine methyl ester, which are excreted in the urine. The major metabolites of cocaine can be found in the urine for periods up to 36 h after the last administration of the drug. Cocaine can also be measured in saliva and in hair. Furthermore, cocaine and its metabolites have been found in the meconium of infants born to woman who have used cocaine during pregnancy (65). The smoking of cocaine base produces a pyrolysis product, anhydroecgonine methyl ester, which can be detected in the urine and thereby serve as a marker for the use of cocaine by the smoking route. These various means of detecting drug use each have their own advantages and disadvantages for monitoring of patients, for drug testing in the workplace, and for forensic purposes.

Although cocaethylene was identified in 1978 in urine samples that contained both cocaine and alcohol, its spectrum of pharmacological activity and significance for dependence on (and toxicity of) cocaine and alcohol combinations has only recently been investigated (25). It is estimated that between 60% and 90% of cocaine abusers consume alcohol concomitantly with cocaine. It has been assumed that the reinforcing effects of alcohol and cocaine combinations could be understood as the summation of their separate reinforcing properties or due to alcohol's sedative effects counteracting excessive stimulation from cocaine. Cocaethylene, however, has been found to be a psychoactive substance with a pharmacological profile similar to cocaine but with a significantly longer duration of action. In addition, the LD50 in mice for cocaethylene is significantly lower than that for cocaine (26). The combination of cocaine and alcohol also has been shown to produce greater changes in heart rate and blood pressure than either drug alone (17). These experimental studies may help explain the fact that many cases of death apparently associated with cocaine use have reported very low blood levels of cocaine. That is, it is likely that these individuals had also consumed alcohol, leading to the formation of cocaethylene which subsequently produced excess toxicity. In fact, recent studies (28) of postmortem blood and tissue levels of cocaine and cocaethylene have reported cases of low levels of cocaine and very high levels of cocaethylene.

The interaction of cocaine and alcohol is further complicated by the fact that alcohol has been shown to increase the blood levels of cocaine administered intransally (67). Whether this is an effect of alcohol on the absorption of cocaine through the nasal mucosa, secondary to alcohol-induced vasodilation, remains to be determined. Interestingly, in that same study, cocaethylene levels in venous plasma rose slowly and were increasing at the time that the subjective and physiological effects of cocaine were returning to baseline. It would thus appear that the formation of cocaethylene after a single insufflation of cocaine (1.25 or 1.9 mg/kg) following the ingestion of alcohol (0.85 g/kg) does not contribute to the physiological or subjective effects observed over the next 2 h. However, one could speculate that with repeated administrations within a relatively short time period, cocaethylene levels could continue to rise and reach toxic levels.

A major consequence of the cocaine epidemic and enhanced public concern about its abuse has been an acceleration in the number of investigations designed to increase the understanding of cocaine's central nervous system (CNS) effects, particularly those related to its ability to produce dependence. Although it has been known for a number of years that dopamine (DA) plays a major role in the behavioral actions of cocaine (e.g., see refs. 34, 48, and 96), research in the 1990s has provided even more definitive evidence of DA's role in mediating the reinforcing effects of cocaine and other psychomotor stimulant drugs, as well as drugs of abuse from other classes (12). This has given rise to the so-called "dopamine hypothesis" of cocaine's reinforcing actions (51). The implications of the DA hypothesis are extremely important because it suggests the target neurochemical system for medications for the treatment of cocaine dependence.

There are several areas of research where advances have occurred in the 1990s that have contributed to an increased confidence that DA mediates cocaine's dependence-related behavioral actions. These include the identification and characterization of its site of action at the dopamine transporter, the cloning of the dopamine transporter, additional findings of cocaine's acute and chronic effects on DA neurotransmission, particularly those that have utilized in vivo techniques, and, finally, further evidence of the role of DA in mediating the behavioral effects of cocaine related to dependence. Because there are reviews of the earlier findings in these areas (34, 48), this review will concentrate exclusively on more recent ones.*3 (See The Development in Brain and Behaviour, this volume, for related discussion).

Cocaine blocks the uptake of DA, 5-hydroxytryptamine (5-HT), and norepinephrine (NE) in the CNS. However, the determination of which of these actions is associated with its reinforcing effects has only recently been elucidated. There is a significant positive correlation between the potencies of cocaine and some related compounds as DA uptake blockers and their ability to serve as reinforcers in self-administration studies in the rhesus monkey (76). In contrast, significant correlations between the reinforcing effects and blockade of uptake of NE and 5-HT have not been found. These data strongly suggest that it is the blockade of the uptake of DA that is an essential step in the mediation of the reinforcing effects of cocaine. Further support for this DA hypothesis comes from evidence supporting the conclusion that the cocaine "receptor" and the DA transporter are identical proteins. Particularly strong evidence has come from cloning experiments. Transfection of COS cells, which do not take up DA or bind cocaine, with a single cDNA for the DA transporter confers both DA uptake and cocaine binding activity simultaneously on the cells (86). Other studies have shown that several DA uptake inhibitors, such as high-affinity analogues of cocaine, mazindol, and several analogues of GBR 12909, bind to a single common site and interact competitively, which leads to the parsimonious conclusion that they bind to the dopamine transporter (5, 74). Furthermore, Grilli et al. (24) have shown that the expression of cocaine binding sites and dopamine uptake sites occurs at the same time during in vitro cellular development. Another indication that cocaine binding sites and the dopamine transporter are intimately related is the fact that cocaine binding sites are distributed within the CNS in areas of high concentrations of DA nerve terminals. While these data indicate a prominent role for DA, it remains likely that significant interactions between different neurochemical mediator systems will be discovered that modulate the reinforcing actions of cocaine and related compounds. A recent PET study by Volkow et al. (90a) lends support to this proposition. In this study, a dose of methyphenidate that produced occupation of 80% of the dopamine transporter (DAT) population did not attenuate the "high" produced by a second injection of methylpheniadate. Further, there was no positive correlation between the degree of DAT occupancy produced by methlphenidate and the intensity of the high experienced. In fact, some individuals experienced no high, despite the fact that 80% of the DAT were occupied by methylphenidate. The authors interpreted these data as suggesting that other transmitters in addition to DA are involved in mediating the mood altering effects of methylphenidate and presumably cocaine.

Progress has also been made in elucidating the characteristics of the cocaine structure that are significant for its binding activity at the DA transporter. The important structural features include a levorotatory configuration, a beta-oriented substituent at C-2 and C-3, and a benzene ring at the C-3 carbon (75). The effects of changes in these structural features, in terms of reduction in pharmacological activity, are described by Carroll et al. (5).

In addition to the characterization of the structure of cocaine relevant to its binding, several laboratories have been involved in characterizing the DA transporter protein itself. These studies led to the development of techniques that eventually resulted in the cloning and expression of cDNA for the cocaine-sensitive DA transporter (44, 86). This finding provides an opportunity to elucidate the molecular sequelae that result in DA uptake and to determine how this process is disrupted by cocaine. Ultimately this knowledge may be useful for the development of medications for treating cocaine dependence. One approach, stemming from this research, that appears promising for the development of a cocaine antagonist is based upon the possibility that the binding sites for DA and cocaine on the DA transporter are overlapping, but not identical. By altering the DA transporter through site-directed mutagenesis, it is possible to determine whether the changes differentially alter the binding of cocaine and DA. The finding that an aspartic acid residue lying within a particular region is crucial for both DA transport and cocaine binding whereas other areas are only important for DA transport supports the possibility of being able to develop a cocaine antagonist that does not interfere with normal DA transport (45) (see also The Dopamine Transporter: Potential Involment in Neuropsychiatric Disorders, this volume).

With repeated exposure to a fixed dose, the locomotor and stereotypic effects of cocaine have been shown to increase in magnitude, a phenomenon known as sensitization. Sensitization was an early observation and has been shown with a variety of psychomotor stimulants besides cocaine (78). Conditioning processes appear to play an important role in the development of sensitization, at least under certain circumstances, because it is not always found if animals are tested for the occurrence of sensitization following repeated cocaine administration in a novel environment--that is, a place where they have never received an injection of cocaine before (71). Because the DA systems mediating the locomotor effects of cocaine appear to be the same as those for its reinforcing effects, and because sensitization is a long-lasting effect (78), an understanding of sensitization and conditioned sensitization has been suggested to have relevance for cocaine abuse and the difficulty of treating abusers (39).

Recent years have witnessed a number of studies on the neurobiology of sensitization (for a review see ref. 39). In vivo methods, such as microdialysis and voltammetry, which allow the monitoring of behavior changes in extracellular DA levels simultaneously, have made a major contribution to our understanding of the CNS actions of repeatedly administered cocaine. Studies employing standard techniques (synaptosomes, tissue slices, and various electrophysiological methods) have also contributed substantially to this area of research.

Initially, in vivo techniques were used to demonstrate that acute administrations of cocaine increased extracellular DA levels in several brain regions, including the nucleus accumbens. Of more relevance to sensitization, Pettit et al. (70) showed that the elevation of synaptic levels of DA in response to cocaine was augmented to an even greater extent in the nucleus accumbens after repeated administration. Kalivas et al. (39, 40) measured both DA efflux and locomotor activity simultaneously in the same animals and found a correlation between changes in these two measures, both of which increased over a period of repeated cocaine treatment. In a subsequent study which examined more carefully the time course of the two effects, Kalivas and Duffy (37) found that following a 5-day regimen of 15 mg/kg cocaine, locomotor activity became augmented as evidenced by increased levels 24 hr after the regimen was terminated. In addition, similar levels of augmented activity were observed 4, 10, and 20 days later when cocaine was again tested. However, elevated levels of DA in the nucleus accumbens in response to an injection of cocaine were only observed 10 and 20 days post regimen, not during the earlier tests. A study by Koob and his associates (94) also provided information about changes in drug-induced DA efflux as well as in basal levels of DA following repeated cocaine. This study showed that basal DA levels were increased on day 1 following a 10-day regimen of 10 mg/kg or 30 mg/kg cocaine in rats but returned to saline-treated levels by day 7 post repeated cocaine regimen. These investigators also noted that while absolute levels of cocaine-induced DA were augmented on day 1, the percentage increase relative to baseline was less compared to chronically treated saline rats given cocaine because of this difference in baseline levels. The authors postulated that this difference may have been due to increased activation of presynaptic autoreceptors (94).

In addition to changes in DA efflux, Ng et al. (62) also provided evidence that DA uptake rate was increased with repeated cocaine administration. Using the in vitro technique of examining 3H-DA uptake in tissue slices, others have also found evidence of an increase in DA uptake rate following chronic cocaine (e.g., ref. 101), while others have found persistent inhibition, at least in the nucleus accumbens (31).

Kalivas and Duffy (38) placed dialysis probes in the area of DA cell bodies (A10) of the VTA whose projections terminate in the nucleus accumbens. They found that levels of extracellular DA in the VTA and locomotor activity in response to an injection of cocaine both were augmented 24 hr after the drug regimen was terminated. Thirteen days later, however, the locomotor response still showed sensitization to an injection of cocaine, but the levels of extracellular DA in the VTA did not. At this time, however, increased levels of DA in the terminal fields are seen. Based on evidence of this type, Kalivas and his colleagues have postulated that changes in DA processes related to sensitization involve several mechanisms that occur sequentially (39). Initially, changes in the DA-containing cell bodies in the VTA underlie sensitization. The mechanism involves increased somatodendritic release of DA in the VTA. As a consequence of this increased release, there is a decrease in the sensitivity of D2 autoreceptors that regulate impulse generation. Because activation of these autoreceptors normally inhibits neuronal firing, their decreased sensitivity results in increased neuronal firing and subsequent increases in extracellular DA in the nucleus accumbens (40).

Several studies using local administration of substances into the VTA have provided supporting evidence for the importance of the cell bodies in the VTA in the initiation of sensitization. For instance, injections of stimulants directly into the VTA result in augmented responses to systemically administered stimulants, whereas injections into the nucleus accumbens do not (39, 40). Furthermore, the administration of a DA antagonist into the VTA prevents the development of sensitization to the repeated administration of cocaine (88). Using electrophysiological techniques, White and his colleagues (27) have also shown that inhibitory impulse-regulating somatodendritic D2 autoreceptors in the VTA become subsensitive with repeated cocaine administration. This subsensitivity, which results in increased levels of neuronal firing, lasts less than 8 days (27) and thus supports the conception that these changes may be important in the initiation of sensitization, but not in its long-term maintenance. Electrophysiological studies by White and his colleagues have also shown the involvement of postsynaptic elements in the nucleus accumbens in sensitization. More specifically, DA receptors become supersensitive to the effects of extracellular DA (27). This increased sensitivity appears to be limited to D1 receptors (27).

Changes in the 5-HT system related to cocaine sensitization have also been demonstrated. Cunningham et al. (10) have described changes in serotonergic systems using behavioral, electrophysiological, and autoradiographic techniques that occur after chronic administration and appear to be correlated with behavioral sensitization. One consequence of uptake blockade of 5-HT in the dorsal raphe by acute administration of cocaine is decreased spontaneous 5-HT neuronal firing. This decrease appears to be a result of enhanced inhibition due to increased stimulation of 5-HT1A impulse-modulating autoreceptors (10). Furthermore, this decreased firing in response to cocaine is augmented over a 7-day chronic cocaine regimen that concurrently produces behavioral sensitization (10). These authors have speculated that the decreased neuronal firing would result in decreased release of 5-HT in neuronal projections to areas such as the VTA and nucleus accumbens. As a result, the inhibitory influence of 5-HT would be diminished, further increasing DA neurotransmission and thus contributing to behavioral sensitization.

It appears that the mechanism(s) underlying sensitization are complex. A recent highly informative review by Kalivas et al. (39) gives a more complete description of the entire circuitry that may be involved in both the initiation and maintenance of sensitization. In addition, the chapter by Nestler et al. in this volume describes cellular changes, such as decreased G-protein coupling, that are related to sensitization. As Kalivas points out, a clear picture of neurobiological changes which produce sensitization has not yet emerged, in part because researchers have used different treatment regimens, different techniques to measure changes, and different times of assessment. Each of the mechanisms proposed to mediate sensitization is supported by empirical evidence. However, additional studies are needed to differentiate and determine the interactions among these changes in the developmental course and long-term expression of sensitization. Nevertheless, a great deal has already been learned about the cascade of neurobiological events that affect behavioral responses following repeated cocaine treatment. Given the progress over the last 5 years (see ref. 34), it is likely that the next few years will yield a more definitive picture. It is also likely that this research, which is spurred in large part by an interest in cocaine abuse, will reveal facts about the molecular biology of sensitization which have importance for much broader areas of mental health disorders.

Cocaine is a robust positive reinforcer, and it is generally agreed that its reinforcing effects are mediated by mesolimbic/mesocortical dopaminergic neuronal systems. Several reviews on the behavioral determinants of cocaine self-administration and the important role of DA in the neurochemical mediation of these actions are available (34, 35, 48). To briefly summarize,*4 other DA uptake blockers and D2 DA agonists support self-administration behavior. Second, both D1 and D2 DA antagonists have been shown to modify cocaine self-administration, whereas noradrenergic antagonists do not. Third, depletions of DA produced by injections of the neurotoxin 6-OHDA into the mesolimbic/mesocortical dopaminergic neuronal pathway, including the VTA, nucleus accumbens, and ventral pallidum, attenuate cocaine self-administration, whereas depletions of norepinephrine do not. Furthermore, Dworkin and Smith (13) showed that DA depletions in the nucleus accumbens, while attenuating cocaine self-administration, did not have any effects on food- and water-maintained responding. Finally, direct injections of cocaine into the medial prefrontal cortex, a rostral projection of the mesolimbic/mesocortical pathway, support self-administration, although direct injections into the nucleus accumbens and VTA do not (23). In addition, the reinforcing effects of direct injections of cocaine into the prefrontal cortex are blocked by DA antagonists. However, depletions of DA in the medial prefrontal cortex produced by the neurotoxin 6-OHDA have produced a variety of effects that are difficult to reconcile.

In summary, although the data are not completely consistent (see ref. 34), the confluence of evidence from this earlier research supports the view that mesolimbic/mesocortical DA pathways mediate the initial steps in the cascade of neural events underlying cocaine's reinforcing effects. More recent evidence supporting this view has come from microdialysis studies. Pettit and Justice (68), using a microdialysis probe located in the nucleus accumbens, showed that extracellular DA levels were increased when cocaine was self-administered. For individual animals the levels remained constant across sessions, suggesting the attainment of an optimal level of DA (68). However, the DA levels attained were dose-dependent and correlated with increased levels of cocaine intake that resulted as dose was increased (69).

In contrast to Pettit and Justice (68, 69), Hurd et al. (30) have reported that with repeated exposure to cocaine during self-administration sessions (9-10 daily sessions), the increases in extracellular DA following cocaine self-administration were diminished relative to those following initial administrations. The authors speculated that this tolerance-like effect may have been due to changes in uptake, release mechanisms, or postsynaptic receptor sensitivity. Others have suggested that the disparity between these studies was due to methodological differences (69). Thus, as with sensitization, there still remains the need for additional studies to clarify the role of DA in the long-term maintenance of cocaine self-administration.

The role of D1 and D2 receptors in the mediation of the reinforcing effects of cocaine has also been studied. As reviewed by Johanson and Fischman (34), much of the earlier self-administration work using DA agonists and antagonists pointed to a prominent role for postsynaptic D2 receptors, although the specificity of this role had been questioned (100). However, it has been reported that SCH 23390, a relatively specific D1 antagonist, also blocks cocaine's reinforcing effects in rats when delivered systemically, as well as directly into the nucleus accumbens (49, 57), although in the rhesus monkey, Woolverton (99) failed to show that SCH 23390 affected cocaine self-administration. Furthermore, studies in rats using progressive ratio schedules have concluded that both D1 and D2 antagonists specifically decrease the reinforcing effects of cocaine, suggesting that both receptor subtypes are important in cocaine reinforcement (29, 77). There is also a recent study demonstrating that at least one D1 agonist, SKF 81297, has reinforcing effects in rhesus monkeys (92). Finally, there is recent evidence that D3 receptors also have a role in mediating the reinforcing effects of cocaine (4). Thus, at the present time, the interactions among the various DA receptors mediating cocaine's reinforcing effects have not been completely delineated. Undoubtedly, all three receptors, and perhaps others as well, play some role, but there are differences among studies as a function of species and paradigms.

Although self-administration paradigms are the most direct means for evaluating reinforcing effects, drug discrimination studies can provide complementary information, based upon the conception that they are related to the subjective effects of drugs in humans. Johanson and Fischman (34) reviewed earlier studies that indicated that cocaine's discriminative stimulus (DS) effects, like its reinforcing effects, are mediated by DA. Recent studies have augmented these findings and have begun to elucidate the relative importance of different DA receptor subtypes. Spealman et al. (87) demonstrated that cocaine analogues that were more potent than cocaine in binding to cocaine binding sites and in inhibiting uptake of DA were also more potent substitutes for cocaine as DS in squirrel monkeys. However, neither D1 nor D2 agonists, nor a combination of the two, completely substituted for cocaine, although both D1 and D2 antagonists blocked its DS effects. Spealman et al. (87) suggested that the activation of both receptor subtypes was important in cocaine's actions. Similar results were found in pigeons (33). However, studies in rats and rhesus monkeys (46, 97) have not replicated all of these findings, suggesting that there may be species differences in the DA receptor subtypes mediating cocaine's DS effects.

In addition to the evidence of a role for DA in cocaine reinforcement and discrimination studies, there have also been some recent studies suggesting at least a modulatory role for 5-HT. Cunningham and Callahan (9), using a drug discrimination paradigm, showed that fluoxetine, a 5-HT uptake blocker, shifted the cocaine dose-response function to the left, indicative of potentiation. In a drug discrimination study with pigeons, both fluoxetine and 8-OH-DPAT, a 5-HT1A agonist, partially substituted for, and the putative 5-HT1A antagonist NAN-190 partially blocked, the DS effects of cocaine (33). In addition to evidence of 5-HT's modulatory effects from drug discrimination studies, self-administration studies have also indicated a role for 5-HT. Loh and Roberts (55) showed that depletions of 5-HT in the medial forebrain bundle or amygdala, induced by injections of the neurotoxin 5,7-dihydroxytryptamine, increased the reinforcing effects of cocaine as indicated by an increase in break-point under a progressive ratio schedule. Similarly, Carroll et al. (6) reported that rate of cocaine self-administration was reduced by fluoxetine. However, these investigators wisely concluded that this change may have been due to either a decrease (blockade) or an increase in the reinforcing effects of cocaine. Given the finding by Ritz et al. (76) that there is an inverse correlation between binding of compounds to 5-HT binding sites and their ability to maintain self-administration, it is possible that 5-HT influences cocaine self-administration. It is also clear that further research is needed to clarify the nature of its modulatory effect.

Other neurotransmitters may also affect cocaine self-administration as a consequence of their interaction with dopaminergic systems. Injections of APV (2-amino-5-phosphonovaleric acid), a selective N-methyl-D-aspartate (NMDA) receptor antagonist, into the nucleus accumbens decreases the effective reinforcing dose of cocaine, and MK-801, a noncompetitive NMDA antagonist, also has been shown to interfere with both the acquisition and maintenance of cocaine self-administration (72, 83). Because there is evidence of glutamatergic projections to the nucleus accumbens which interact with dopamine systems, these results suggest a modulatory role for glutamate on the reinforcing effects of cocaine.

It is also well known that endogenous opioid and DA systems interact in the CNS (48). An interesting series of studies on the effects of opiate agonists/antagonists on cocaine self-administration has been instrumental in shaping a new approach to the treatment of cocaine dependence, at least in clients that are also dependent on opiates. Mello et al. (60) showed that buprenorphine, a mixed opioid agonist/antagonist, specifically reduced the self-administration of cocaine in rhesus monkeys, and this reduction was maintained over time. The specificity, durability, and interpretation of this effect as indicating an antagonism of cocaine's reinforcing effects by buprenorphine has been questioned (95). In an open clinical trial, buprenorphine was reported to significantly reduce both opiate and cocaine abuse by patients who had been abusing these drugs for over 10 years (19). However, in a controlled clinical trial that demonstrated that buprenorphine was effective in reducing heroin abuse, no decreases in coexistent cocaine abuse were noted (36). Further research is necessary to determine whether buprenorphine has any clinical effectiveness in the treatment of cocaine abuse and dependence.

In 1985, Dackis and Gold (11) postulated that the prolonged use of cocaine resulted in a depletion of brain DA. This depletion resulted in symptoms of anergia, anhedonia, depression, and cocaine-craving during cocaine abstinence. Further detailed descriptions of a cocaine withdrawal syndrome were provided by Gawin and Kleber (21). These clinicians interviewed outpatients who retrospectively described the phasic nature and sequence of withdrawal symptoms following abrupt discontinuation of cocaine. Gawin and Kleber (21), like Dackis and Gold (11), attributed these withdrawal symptoms to disturbances in DA function. Furthermore, they hypothesized that these withdrawal symptoms would lead to craving for cocaine and relapse to cocaine use. Therefore, pharmacotherapies which would increase DA brain levels or activate DA receptors should alleviate the mood changes during withdrawal, thereby decreasing craving for cocaine and the probability of relapse (11). Later studies, conducted under controlled conditions (i.e., in an inpatient unit), however, have failed to find evidence of a phasic cocaine withdrawal pattern and, in fact, have only reported mild changes in mood and cocaine-craving during withdrawal (e.g., see ref. 82). It should be noted that these studies were done with inpatients where environmental influences on withdrawal (i.e., conditioned cues) were not present.

Although the idea that brain levels of DA are depleted as a result of chronic cocaine administration and that the functional consequences of this depletion becomes manifest upon the termination of cocaine use continues to be accepted by many clinical investigators, the support for this conception is weak. There are data in humans demonstrating that prolactin release, which is under central inhibitory DA control, is increased during the first weeks of cocaine abstinence (11, 61). However, the controlled study of cocaine withdrawal by Satel et al. (82) showed only modest neuroendocrine changes indicative of DA depletion, and these were not correlated with craving for cocaine. On the other hand, position emission tomography (PET) studies of human cocaine abusers have provided support for the role of dopaminergic dysfunction during cocaine withdrawal. Volkow et al. (90) showed decreased cerebral blood flow in the prefrontal cortex in humans with a history of cocaine use. Although this deficit might be a consequence of a cerebral accident, the authors suggested that this effect was due to changes in neuronal functioning indicative of neurotoxic changes in DA neurons. A subsequent study using PET with [18F]fluorodeoxyglucose showed that glucose metabolism was increased in the orbitofrontal cortex and basal ganglia shortly following cessation of cocaine use in humans, which is also consistent with decreased DA activity (89). Activity levels were normal in subjects that were tested 2-4 weeks after cessation when symptoms of withdrawal had decreased.

Although the neuroendocrine and PET studies in humans are cited to support the idea that DA is depleted as a consequence of repeated cocaine exposure, these measures of DA levels are indirect. More definitive evidence can be obtained from animal experiments where DA levels can be measured directly. In fact, long-term depletion of brain DA and 5-HT has been found following the repeated administration of methamphetamine (MA) in a variety of animal species (85). Because MA and cocaine are pharmacologically similar, evidence of depletions following MA gives credence to the idea that brain DA may also be depleted by cocaine. However, while other MA-like drugs have also been shown to have similar long-lasting neurotoxic effects, cocaine does not decrease brain levels of DA or its metabolites in animal experiments (e.g., see ref. 47).

Although there does not appear to be evidence from animal studies of DA depletion, there have been recent data from animal studies indicating a functional DA depletion. That is, there is evidence, obtained using microdialysis techniques, of a decrease in basal levels of extracellular DA following chronic cocaine, at least at certain times periods. For instance, although Parsons et al. (66) showed no change in basal extracellular DA levels immediately after a 10-day regimen of 20 mg/kg/day cocaine, after 10 days of abstinence these levels were significantly reduced, which the investigators attributed to reduced release, not increases in uptake. Rossetti et al. (79), using a high-dose regimen of 15 mg/kg twice daily for 16 days, found a similar time course of effects and also reported decreases in basal extracellular DA following other drugs of abuse. Rossetti et al. (79) related these decreases in basal DA levels to the findings of Markou and Koob (59), who demonstrated that immediately following a period of prolonged cocaine self-administration, there was an increase in the threshold for electrical intracranial self-stimulation which was prolonged. Furthermore, the magnitude and duration of the increase were dose- and time-dependent. On the other hand, other investigators have shown increased levels of basal extracellular DA immediately following a chronic regimen, which, however, were not maintained over time (94). However, when expressed as a percent of basal DA levels, the chronic cocaine regimen resulted in a diminished increase in DA levels extracellularly in response to cocaine (94). Most provocative is the finding that within hours following a short regimen of intravenous cocaine self-administration, there were decreases in basal extracellular DA levels (93); these decreases were directly related to the duration of self-administration. These authors also related these findings to the changes reported in threshold for intracranial electrical stimulation reinforcement.

Although the exact nature and importance of DA dysfunction in the maintenance of cocaine dependence and withdrawal is not clear, the possibility of DA dysfunction has had a major impact on the strategy for the development of medications for the treatment of cocaine abuse. Furthermore, the evidence that cocaine's reinforcing effects are mediated by DA provides a rationale for pursuing the use of dopaminergic agonists or antagonists in the treatment of cocaine dependence, in a manner analogous to the use of methadone or naltrexone for the treatment of heroin addiction (see Amyotropic Lateral Sclerosis, Glutamate, and Oxydative Stress, this volume). Hence, both bromocriptine, a dopamine agonist, and amantadine, which presumably releases DA and blocks its uptake, have been assessed in clinical trials for their ability to decrease withdrawal symptoms. Although there are reports that these medications decrease craving for cocaine, more recent double-blind placebo-controlled trials have failed to confirm a positive effect (e.g., see ref. 91). In addition, a variety of other types of dopaminergic agonists (e.g., pergolide, mazindol, methylphenidate) have been suggested for the treatment of cocaine dependence. Unfortunately controlled clinical trials have failed to establish their usefulness in promoting cocaine abstinence. Despite evidence that D1 and D2 receptor antagonists decrease the reinforcing effects of cocaine (see previous section), there have been no controlled trials using antagonists. There has been a report of success in an open trial with the blocker flupenthixol decanoate, but as yet this drug has not been evaluated in a more rigorous trial (20). Furthermore, because this antagonist differentially blocks inhibitory autoreceptors at low doses, the authors attribute its success to increasing DA activity, which has been decreased due to autoreceptor supersensitivity (14). On the other hand, Gawin et al. (22) have demonstrated the efficacy of desipramine, a tricyclic antidepressant, for treating cocaine dependence in a carefully conducted clinical trial. Gawin et al. (22) have speculated that desipramine's effectiveness in promoting abstinence is based on its ability to reverse the neurochemical changes produced by chronic exposure to cocaine, perhaps as a result of its ability to block DA uptake. In fact, a meta-analysis of nine placebo-controlled studies of the effectiveness of desipramine has claimed that, although this drug did not improve retention in treatment, it did promote greater rates of abstinence than placebo (52). The meta-analysis, however, included only two reports which showed that desipramine was more effective than placebo in promoting abstinence, one of which was an interim report of an ongoing study (64). When the study was completed, the investigators concluded that desipramine was not more effective than placebo (1). Thus, the conclusions of the meta-analysis by Levin and Lehman (52) may no longer be tenable. Furthermore, a later 12-week study comparing amantadine, desipramine, and placebo in different groups of methadone-maintained patients showed no differences across groups in numbers of cocaine-free urines (50). It would thus appear that the effectiveness of desipramine for the treatment of cocaine dependence is not established, although it may be useful in cocaine-dependent people with coexisting depression. It is also interesting to note that in a human self-administration study, desipramine appeared to have no effect on cocaine self-administration, although there were indications that it increased cocaine's aversive effects (16).

In summary, despite optimism based in part upon the success in the treatment of opiate addiction with agonists or antagonists and early indications of success in open clinical trials and a few controlled trials, to date, there is no pharmacological agent that appears to be effective for the treatment of cocaine dependence. The failure to find an effective medication for the treatment of cocaine dependence may be due to our limited knowledge of the neurobiological basis of cocaine reinforcement and dependence. It may also be the case, however, that the neural substrates of cocaine reinforcement and dependence are more difficult to selectively manipulate pharmacologically than those underlying opiate reinforcement and dependence. Thus, limiting treatment research to pharmacotherapy of any type may not be the most useful strategy (see Obesity, Fat Intake and Chronic Disease, this volume).

Clinical studies conducted since the Third Generation of Progress have confirmed earlier observations that cocaine abuse produces a wide variety of toxic effects, including: toxic psychosis; increased incidence of panic disorder and suicide; convulsions; brain damage; cardiovascular problems; liver toxicity; hyperthermia; muscle damage leading to rhabdomyolysis; deviated nasal septum from snorting cocaine, as a result of its vasoconstrictive properties; and burns of the larynx from crack cocaine smoking. Cocaine has also been implicated as a cause of traumatic accidents in urban areas (54). Several recent reviews on organ toxicity are available (2), and therefore the coverage in this section will be selective.

By far the most common toxic effects of cocaine involve the cardiovascular system. There have been numerous case reports of neurovascular complications associated with the use of cocaine by all routes of administration (32). These include cerebral parenchymal hemorrhages, subarachnoid hemorrhages, and ischemic cerebral manifestations. In a large percentage of cases, there was evidence of underlying vascular abnormalities which, in association with the systemic hypertension produced by cocaine, may have led to hemorrhagic stroke. The pathological changes in the vasculature that place individuals at increased risk for neurovascular accidents with cocaine include arteriolar thickening, increased perivascular deposits of collagen and glycoprotein, and inflammatory cellular infiltrates (8).

The cardiovascular toxicity of cocaine and the mechanisms underlying the pathological changes have been extensively reviewed elsewhere (8). This review concluded that more than one mechanism accounts for the deleterious effects of cocaine on the myocardium and that a subgroup of the population may be more vulnerable to the cardiotoxic effects of cocaine. It is difficult to determine at present, for example, whether the atherosclerotic changes seen in young cocaine users are a function of their cocaine use or a preexisting condition. This preexisting condition is then exacerbated by cocaine use, resulting in a cardiac emergency that brings the individual to the attention of researchers.

There have been a number of recent reports showing an association between nontraumatic rhabdomyolysis and cocaine abuse. Although the exact mechanism responsible for this association is not known, it is important clinically for several reasons. First, chest pain associated with rhabdomyolysis may be misdiagnosed as cocaine-induced myocardial infarction; second, rhabdomyolysis may lead to fatal kidney failure (81).

The use of cocaine by women during pregnancy is a major public health concern, although the evidence for teratogenic and mutagenic effects is controversial. Furthermore, estimates of the incidence of neonatal exposure to cocaine in utero vary widely and are increasingly difficult to obtain (84). In addition, much of the early clinical research giving rise to the concern about "crack babies" was seriously flawed and failed to consider that pregnant cocaine users are often polydrug abusers, nutritionally deprived, and infected with sexually transmitted diseases, and that they rarely obtain adequate prenatal care. A meta-analysis of studies on the effects of cocaine on fetal development found few perinatal effects, other than low birth weight, that could be specifically attributed to the mother's use of cocaine (56). Importantly, the low birth weight seen in many neonates that are born to cocaine-using mothers can be significantly attenuated if the mother receives adequate prenatal care (73). It is thus of the utmost importance that in our attempts to protect the developing fetus from cocaine exposure, we do not have policies which will lead drug-using women to avoid contact with medical services. In addition, it is now time to focus research efforts on developmental outcomes because even if there are no discernible perinatal effects, the long-term outcome of exposure to cocaine in utero as well as subsequent exposure to a drug-using lifestyle is unknown.

Because it is extremely difficult to determine the specific effects of cocaine on pregnancy outcome and later postnatal development, animal research would seem to be the obvious answer for controlling many of the confounding variables. The use of animal models for studying the effects of drugs of abuse on fetal development, however, raises its own set of difficulties. For a discussion of these problems, see Schuster and Gust (84).

Although the use of cocaine in the general population has declined markedly in the last few years, the abuse of cocaine by regular users has not. Furthermore, since the introduction of crack cocaine in the mid-1980s, there has been a tremendous increase in (a) the number of ER mentions in association with cocaine (DAWN) and (b) the numbers of people presenting for treatment with cocaine dependence as their primary problem. During this same period of time, rapid advances in our understanding of the molecular, cellular, physiological, and behavioral bases of cocaine-dependence-producing effects have been made. This research has clearly shown the importance of DA systems in mediating these effects. It is also clear, however, that a full understanding of the dependence-producing properties will not emerge until the complex interactions among the multiple neuronal systems that cocaine affects are understood. Furthermore, these neuronal mechanisms must be studied in conjunction with environmental manipulations which modify cocaine's reinforcing efficacy. This research will also provide insights into the interactions of neurochemical systems and their interaction with environmental determinants of behavior that will have implications far beyond the phenomenon of cocaine dependence. These insights may also aid us in our search for a medication to decrease the reinforcing efficacy of cocaine, decrease withdrawal signs and symptoms that are related to relapse, and help maintain longer-term abstinence. Such medications would help to make the cocaine-dependent individual more "available" for behavioral and psychosocial interventions which are essential if the individual is to achieve a cocaine-free lifestyle.

This chapter was prepared during the same period of time that another chapter on the same topic was being prepared by CEJ. That chapter will appear in Handbook of Experimental Pharmacology, Volume 116: Pharmacological Aspects of Drug Dependence: Toward an Integrated Neurobehavioral Approach, edited by C. R. Schuster and M. Kuhar, publisher Springer-Verlag. The authors are M. W. Fischman and C.-E. Johanson. Portions of the two chapters overlap and all authors have agreed to this arrangement. In addition, the editors of both books have been informed.

published 2000