Mesocorticolimbic Dopaminergic Neurons

Functional and Regulatory Roles

Michel Le Moal

In 1978, Moore and Bloom (26) observed that progress had been slow in elucidating the role of the dopaminergic (DA) neurons and, in particular, their fundamental functional property. While still valid, few neuronal systems have provoked as much investigative efforts as the DA neurons, as noted in other chapters in this volume. DA neurons are the source of hypotheses concerning sensorimotor and psychotic defects, the pathophysiology of drug abuse, and favored targets for drug development and grafting replacement therapy (see Animal Models of Drug Addiction, New Developments in Dopamine and Schizophrenia, Parkinson’s Disease, Cocaine, and Pathophysiology of Tobacco Dependence).

The nonstriatal projections, generally referred to as a separate set often labeled "mesolimbic" or A10, have received somewhat less attention than the "nigrostriatal" A9 DA neurons. The mesolimbic neurons—later the "mesocorticolimbic" group—have subsequently been divided into numerous subsets denoted in the terms of the region of the projection, of which about 20 have been noted. The cellular subgroups are intermingled within the ventral mesencephalon while some limbic projections have their origin in the substantia nigra and vice versa such that selective inactivation of each system is not only difficult, but has caused confusion in interpreting pharmacological manipulations. Consequently the list of the "functions" attributed to the "mesolimbic" neurons is as varied as the behavioral paradigms used to study them. Furthermore, no disease or clinical syndrome with anatomical or biochemical abnormalities primarily involving these structures has yet been demonstrated. In this chapter, I extend a working hypothesis formulated in 1984 and extended recently (21). The conclusions can be summarized briefly as follows: (i) DA neurons do not have specific functions; (ii) they regulate and enable integrative functions in the neuronal systems onto which they project; (iii) lesion of their terminals induces neuropsychological deficits that are characteristic of the functions of the neuronal systems they regulate; and (iv) the deficits observed depend on the behavioral situations or the tasks used to explore them.

The DA cells are embedded within dense tracts of ascending and descending fibers (among them the medial forebrain bundle) through which they communicate with large regions, such as the limbic–forebrain and the limbic–midbrain structures, defined by Nauta (27). The ventral mesencephalon, which encompasses the ventral tegmental area (VTA), is an anterior part of the reticular formation. The DA cells receive signals of all sorts, and numerous peptides and transmitters have been found in this area (15).

Destruction of the VTA with radio-frequency (RF) lesions produces a behavioral syndrome [see (21) for references] characterized by: (i) a high level of locomotor activity and hypoexploration; (ii) profound deficits in behavioral suppression capacity (10); (iii) disappearance of the behavioral patterns essential to the survival of the individual or the species [e.g., social, maternal (9), and hoarding behaviors (44)]; (iv) deficits defined operationally in terms of (a) the facilitation of active (one way) avoidance learning, (b) deficits in approach behaviors with continuous or fixed-interval schedules of reinforcement due to a lack of behavioral suppression ability, and (c) deficits in intracranial self-stimulation and in self-administration of psychostimulant drugs; and (v) profound deficits in attentional and representational cognitive processes as measured by delayed alternation tasks without deficits in a visual discrimination (10). The hypokinetic–hypoattentional deficits are correlated with a decrease in dopamine in the anteromedian frontal cortex and to a lesser extent in the nucleus accumbens, whereas no correlation was observed with the dopamine levels in the nucleus caudatus or the serotonin and noradrenaline levels in the forebrain.

Local injections of 6-hydroxydopamine (6-OHDA) confirmed some of the results obtained after RF lesions: hyperactivity and disruption of behavioral capacities (18, 21, 31, 32), but without the aphagia, adipsia, or motor disorders. Moreover, these deficits were unlike those resulting from local destruction of DA terminals within the nucleus accumbens or within the frontal cortex (see ref. 21). Extended investigations have led to a clarification of five deficits.

Extensive studies (19, 20) using RF techniques and different doses of 6-OHDA in the cell bodies and accumbens regions as well as in combined lesions led to the following conclusions: (i) large 6-OHDA lesions of the VTA or of the nucleus accumbens produce hypoactivity, a complete blockade of the locomotor stimulating effects of amphetamine, and a profound supersensitive response to apomorphine; (ii) smaller VTA lesions produce significant increases in spontaneous daytime and nocturnal locomotor activity, with the largest effect occurring at the lowest dose; (iii) the hyperactivity produced by RF lesions to the VTA is unresponsive to amphetamine, but it decreases after apomorphine and is blocked by the addition of a 6-OHDA lesion to the nucleus accumbens; and (iv) these combined lesions produce a blockade of the stimulating effects of amphetamine and a potentiated response to apomorphine which was identical to that observed with a nucleus accumbens lesion alone. These results suggest that dopamine may play an essential role in the expression of both spontaneous and stimulant-induced activity (28). Furthermore, the much larger increase in spontaneous activity observed in the VTA-RF group than in the VTA-6-OHDA groups suggests that an as yet unidentified powerful inhibitory influence is exerted on the DA neurons within the mesencephalon or at the level of the terminal fields.

Aphagia and adipsia result from lesions of the nigrostriatal neurons or from added damage of mesocorticolimbic neurons and of a trigeminal sensory component. Extensive studies (see ref. 21) allow us to conclude the following: (i) the most effective site for producing aphagia and adipsia after RF lesion to the mesencephalon is an intermediate zone between the substantia nigra and the VTA; (ii) 6-OHDA in this intermediate zone led to a less severe feeding deficit than that observed with RF lesions, suggesting that both DA and non-DA neurons are involved in the mesencephalic aphagic syndrome; (iii) rats with RF lesions of the sensory trigeminal nucleus are aphagic; but when they recover, their deficit is less severe than after RF lesion; and (iv) a 6-OHDA lesion of the mesencephalic intermediate zone combined with a RF lesion of the trigeminal sensory nucleus led to the more severe deficits observed. These results demonstrate that the DA neurons projecting to the cortical and limbic regions contribute to the aphagic syndrome but are not essential for it.

After 6-OHDA infusion in various terminal areas and combined lesions of these terminal areas, only rats with combined 90% total depletion of dopamine showed a severe deficit in initiation and incentive to respond in an active avoidance task (19). These rats showed no response to a low or high dose of amphetamine, and they remained cataleptic as if receiving high doses of neuroleptics for the duration of the experiment but rapidly recovered from transient aphagia and adipsia (less than 10 days post lesion). These findings show that major psychomotor deficits are prevalent after the total lesion of the DA fibers in the whole forebrain and that the existence of interactions between DA projections is a prerequisite for the initiation and development of a fully adaptive response. This cooperation supports the hypothesis that interrelationships exist between the various terminal fields of the DA network.

Dozens of studies, mainly using DA receptor blockade, suggest a role for dopamine in sensorimotor integration as well as in goal-directed response and motivational arousal. It is generally assumed that the cortical and limbic projections—that is, the neurons from the VTA—do not modulate sensorimotor integration. However, limbic and frontal neglect syndromes have been described when at least one-third of the system is destroyed. The extent of the resulting neglect was correlated (see ref. 21) with the overall damage to the substantia nigra and VTA, rather than to any individual region within this DA continuum. Damage to the corticolimbic projections potentiates the severity of the neglect produced by nigrostriatal lesions. Thus, the involvement of the individual subclasses of mesotelencephalic DA neurons in the neglect syndrome is more widespread than was previously thought to be the case. Do these changes result from deficits in motivation, impaired stimulus perception, or inadequate sequence of motor responses? Numerous tests under various levels of motivation have shown (32) that rats are insensitive to motivational stimuli only when DA lesions are placed in the medial ventral mesencephalon.

The disruption of instrumental conditioning after blockade of central DA neurotransmission with drugs or lesions has been explained as an inability to initiate movements rather than as a disruption of learning mechanisms per se. However, an attentional impairment following disruption of mesofrontal or mesoseptal DA systems has been inferred (30) from performance on an alternation hole board learning task (conditioned blocking). Here increases in behaviors collateral to task performance were observed after dopamine-depleting lesions of the frontal cortex, septum, and medial ventral mesencephalon. The level of DA activity (or the balance between the DA and noradrenaline projections) in frontal and limbic regions can contribute to both efficient associative conditioning and the normal ability of rats not to ignore a redundant stimulus (see ref. 21).

Studies on the ventral striatum have led to the hypothesis that this region mediates communications between neural systems involved in motivation, emotion, and movement or ongoing behavior (4, 25). Several behavioral investigations (see ref. 21) led to the conclusion that DA terminal lesions induce hypoexploration, failure to inhibit response strategies, and, more generally, a perseveration syndrome with reduced distraction caused by irrelevant information and decreased behavioral switching and flexibility and paradoxical locomotor disinhibition in an emotional context. In addition, these lesioned animals exhibited an enhanced latency in the initiation of motor responses, disturbances in the acquisition of spatial discrimination, and great difficulty in reversing previously learned habits. However, after acquisition has taken place, subsequent lesions do not impair retention.

The behavioral changes caused by these lesions appear to result partly from an inability to switch from one behavioral activity to another and to organize complex behaviors, such as hoarding activities (see ref. 21). Such lesions also disrupt the acquisition of a displacement activity, such as schedule-induced polydipsia. DA neurons in the ventral striatum have been implicated in psychostimulant self-administration; both acquisition and retention are suppressed after lesion of these neurons. Also, a direct relationship has been found to exist between dopamine utilization in the ventral striatum and the intensity of physical, environmental or social stimuli: The more stressful the signal, the more altered the DA activity (see ref. 21).

However, it is difficult to destroy the mesoaccumbens DA projections without affecting the other DA forebrain projections that course through this area. Vigorous sprouting of remaining axons associated with increased rates of transmitter synthesis and an enhancement of postsynaptic dopamine receptor plasticity occur in this region (see A Critical Analysis of Neurochemical Methods for Monitoring Transmitter Dynamics in the Brain). In addition, anatomical and functional considerations suggest that the prefrontal cortex exerts descending influences upon subcortical regions, nucleus accumbens, septum, and the ventral mesencephalon itself. For instance, both dopamine receptors and turnover increase in accumbens after selective lesion in the prefrontal cortex.

Data from many groups suggest that the corticofrontal region is important for the representation and the evaluation of situations as being potential stress. In a conditioned fear situation (i.e., after exposure to an environment previously paired with footshocks), increased dopamine utilization is a general phenomenon related to various types of anxiety, adaptive situations, and representations (see ref. 21). A 3-min electric footshock session more effectively enhanced dopamine utilization in the frontal cortex of isolated rats when compared to those living in groups. However, long-term isolation in the rat greatly reduces dopamine utilization in the mesofrontal DA transmission (12 weeks of isolation), but it increases utilization in the mesoaccumbens or mesostriatal projections. After rats that had been isolated for 8 weeks were kept in groups for 4 weeks, a relative increase in the DA metabolites occurred, with values between those of the controls and those of rats isolated for 12 weeks (2). These results also suggested the existence of interneuronal regulations between the various DA projections, because an increase in dopamine utilization in a given projection field was accompanied by a decrease in another field and suggested that the meso-prefrontal DA neurons may exert inhibitory effects on the activity of DA neurons innervating subcortical structures (21). Further support for this hypothesis comes from studies of electrolytic lesions of the ventral tegmentum (see ref. 21) where the correlation between the increase in locomotor activity and the decrease in dopamine content in the frontal cortex and nucleus accumbens suggests that DA neurons play a prominent and inhibitory role in prefrontal cortex functioning.

DA mesoaccumbens activation may indirectly reflect the functioning of the prefrontal cortex in adaptive, representational, or stressful situations and may explain why in some situations the mesoaccumbens DA projection is also activated. A mild, 8-min tail pressure causes a large and sustained (longer than 2 hr) increase in dopamine utilization in the nucleus accumbens but not in the prefrontal cortex, as measured by extracellular DOPAC using in vivo voltammetry (22). A specific increase in DOPAC levels was also observed in the nucleus accumbens 1 hr after stress in postmortem biochemical assays (6). Moreover, the increase in DA metabolites was antagonized by pretreatment with benzodiazepine or other anxiolytic drugs. It is interesting to note that restraint associated with cold increases DOPAC levels in the frontal cortex but not in the nucleus accumbens (21), whereas electric footshock stress increases DOPAC in both brain regions (21, 47), and immobilization stress increases dopamine utilization selectively in the nucleus accumbens (48). Carlson et al. (5) showed that 24 hr of food deprivation in rats produced an increase in DOPAC levels in the medial prefrontal cortex but not in the ventral or dorsal striatum. Explanations for these differences may include the anatomical relationships of the frontal cortex. The frontal cortex may be considered as a visceral cortex involved in the recognition of autonomic signals and reciprocally connected with the hypothalamus. Stimulation of the medial frontal cortex decreases gastric motility, and the prefrontal cortex seems to have an inhibitory pathway to the medial hypothalamus, which has also been implicated in the regulation of feeding.

It is now generally agreed that the prefrontal neuronal system mainly includes the prefrontal cortex and the anteromedial part of the striatum which receives afferents not only from the mediodorsal thalamic nucleus but also from the prefrontal cortex. The prefrontal cortex in the rat is defined anatomicofunctionally as the region to which efferent fibers project from the mediodorsal thalamic nucleus. However, some VTA DA neurons project to the anteromedial part of the caudate nucleus (see ref. 26).

Delayed-response tasks are generally considered to be particularly susceptible to cognitive deficits after lesions of the prefrontal system in all species of mammals. The mesoprefrontal DA projections are also involved in cognitive processes, especially those required for performing delayed alternation tasks. 6-OHDA lesion of the terminals in the prefrontal cortex or in the anterior and medial parts of the striatum induced a deterioration of the preoperative performance of lesioned rats (see ). During the postoperative learning period, treated rats exhibited collateral behaviors (suggestive of enhanced distractability such as rearing, stopping, and returning) before entering the chosen arm of the T maze, but these same rats were able to learn normally a visual discrimination task for food reward. Similar deficits were observed after prefrontal cortex and anteromedial striatum lesions, resembling the effects of a large lesion of the frontal cortex. Thus, the deficits observed characterized a kind of DA disconnection syndrome for the region to which the DA neurons projected. However, the exact nature of the impaired function is more difficult to explain.

In support of this hypothesis, 6-OHDA lesions of DA terminals in monkey prefrontal association cortex impaired performance of delayed alternation spatial tasks (3, 12). This impairment was nearly as severe as that caused by surgical ablation of the same area, but could be temporarily reversed by systemic injections of either L-dopa or apomorphine. Dopamine augments the activity of prefrontal neurons involved in temporal integration of external cues and motor performance, including spatial short-term memory and predictive processes (42). Moreover, an activation of D1-type dopamine receptors appears to be responsible for the increased activity in prefrontal neurons related to the performance of the task. In this respect, it is interesting that the frontal syndrome has also been interpreted in terms of impaired selective attention. The increased locomotor activity that occurs after 6-OHDA-selective or RF lesion of the medial prefrontal cortex or of the DA cells in the ventral mesencephalon may be related to attentional deficits, and the hyperactivity–hypoexploration syndrome caused by RF lesion of the mesencephalic cell bodies is at least partly due to a complex imbalance within the DA cortical system (see ref. 21).

Taken as a whole, data from frontal cortex, nucleus accumbens, and septal and amygdala regions (for details, see ref. 21) demonstrate that the functional effects of DA terminal lesions are as varied as the number of projections investigated. However, lesions of the cell bodies induce profound adaptive disability, presumably reflecting deficits in all the projection areas. Selective lesions of the different DA projections provoke different behavioral syndromes which characterize the deafferentated neuronal system. The various projections function in an interdependent manner and are stimulated either from the bottom (i.e., the reticular formation) or from the top (i.e., the various integrative regions where they project). This integrated set defines the network ().

These considerations confirm a nonspecific activational role of the DA transmission, necessary but not sufficient for organization of adaptive behaviors. DA neurons are necessary because if they are not present—and only minute amounts may be necessary—the neurophysiological functions and integrative processes do not operate. They are not sufficient because if the other systems are pushed or if the external or internal environments have urgent needs, the systems in imbalance can temporarily recover. DA neurons do not appear to regulate specific function or behavior, and the patterns of activity of these neurons do not support such an integrative capability.

When DA neurons are stimulated—for instance, by high doses of psychostimulant drugs—the behavioral response rate also increases but, in parallel, the responses show less variation or adaptability, leading to stereotypy (repetition of invariant sequences of behavior). This drug response phenomenon yields an inverted "U-shaped" dose–response curve to increasing doses of psychostimulant. As described elegantly by Lyon and Robbins (23), a shift occurs from a progressive facilitation of response sequencing to a progressively disorganized and fragmented behavior, from complex patterns to simple ones, and from a sensorimotor facilitation and environment exploration, flexibility, and adaptation to a channeling of sensorimotor integration and eventually to a massive activation of the motor output system which closes the organism to the environment and binds the subject to stereotypy and routines (23) (see ). It can be inferred that the more these executive systems are stimulated, the greater the number of routines that are selected and the fewer the number of environmental controls that are possible—that is the greater the number of extrapyramidal output systems that are involved, repeating motor schemes over and again (39).

Based on local infusion of agonists and lesion studies, psychostimulant-induced locomotion may result from a facilitation of the DA transmission in the nucleus accumbens and not of dorsal striatum (21, 23). For example, DA transmission blockade or 6-OHDA lesion of the dorsal striatum blocks amphetamine-induced stereotypy, while lesion of the ventral striatum has no such effect, but blocks the locomotor activation produced by amphetamines. However, this anatomical functional dichotomy—that is, dorsal striatum stereotypy versus ventral striatum locomotion—has probably been oversimplified. DA activation of the dorsal striatum might be a necessary but certainly not sufficient condition for amphetamine-induced stereotypy. Intra-accumbens injection of psychostimulants dose-dependently increases locomotion as well as stereotypy, so the ventral striatum may have a more important role than the dorsal striatum in the initiation of the locomotor and stereotyped effects of the psychostimulants than was previously thought (see ref. 21).

The whole striatum operates under the control of the neuronal systems involved in motivation, arousal, and cognitive processes. These processes modulate investigative behaviors initiated by motivation and trigger activity in the striatal DA system essential for appropriate sensorimotor coordination (14). At lower psychostimulant drug doses, stimulation of the DA mesoaccumbens pathway appears first, modifying the ventral striatum–ventral pallidum output system. Then the ventral striatum promotes further activity in the DA projection to the anteromedial and anteroventral striatum and lastly then enhances the activation of the mesostriatal (dorsolateral caudatus) pathway. These successive instructions are translated into behavioral sequences: locomotion, head and oral movements, then stereotyped features. As the dose of stimulant increases, the behavioral responses are occluded, arising from the occurrence of intense head and oral movements, sniffing, and the disappearance of locomotor hyperactivity. Of course, this succession of actions also involves non-DA output pathways and a diffuse set of regions and subsystems (see ref. 21). Various chemical stimulations and lesions have been found to replicate or reduce the psychostimulant-induced locomotor or stereotyped responses: sniffing and biting are induced via the substantia nigra reticulata, oral stereotypies are obtained via the superior colliculus, the thalamus is involved in posture, the reticular formation is involved in head turns, and the ventral pallidum and the mesencephalic locomotor regions mediate some aspects of the locomotor response.

In conclusion, the DA projections of the ventral and dorsal striatum may serve as filtering and gating mechanisms for signals from the limbic regions (i.e., for basic biological drives and motivational variables) and from the neocortex (i.e., for signals of a cognitive nature), which have to be synchronized and eventually translated into motor acts through the pallidal and pontine motor nuclei (9, 25). This synchronization of sensorimotor integration (dorsal striatum) and motivational and energizing processes (ventral striatum) allows the initiation of specific responses appropriate to the moment-to-moment changes in the environment as the behavioral sequence progresses. Multiple motor subsystems exist () which function normally at a low dose of psychostimulant in a coordinated manner, and dopamine regulates the entity which dissolves into dysfunctional parts when the neuroregulator fails to act in a coordinated fashion.

Any change in the functioning of the DA neurons will have repercussions on adaptive capacities and their brain–body indexes. For example, schedule-induced behaviors have frequently been used as behavioral models of adaptation. Dopamine depletion in the ventral striatum blocks the acquisition of water intake response during a scheduled food delivery test. Lesion of the DA mesoseptal projection leads to the reverse effect—that is, to an increase in the acquisition rate with the same type of behavioral response (21). Thus, animals placed under scheduled food delivery conditions have significantly increased plasma corticosterone, brain endorphin levels, and analgesia compared to rats given nonscheduled food. Conversely, rats which were free to engage in excessive drinking or other displacement activities showed a significant reduction in these hormonal responses (45, 46). Interestingly, a DA terminal lesion in the ventral striatum abolished the increase in corticosterone levels in rats subjected to a schedule of food delivery. This suggests that the DA pathways are involved at a more general level in this kind of adaptive response.

Recent data have indicated that the DA neurons possess receptors to corticosterone (13) which might make them receptive to arousing or stressful situations. Glucocorticoids may potentiate the activation of biogenic amine response to stress and then help the neuronal system to oppose the neural action of this potentially deleterious stress and restore the homeostatic balance. Repeated intermittent treatment with the psychostimulant or stress and induced increases in plasma corticosterone produced enduring changes in the response of DA neurons and that of the pituitary to subsequent stress. Stress clearly changes the dopamine metabolism in the ventral striatum as well as in the frontal cortex (21, 47).

The linking of DA functioning, adaptive capacities, and coping is particularly evident with stress that cannot be controlled. For example, rats receiving an identical amount of stress from electric footshock, but allowed to control its duration, displayed much less stereotypy than did stressed rats (21). These data are integrated between several well-established findings: (i) the inducing and precipitating role of psychostimulants in psychopathological states; (ii) the role of stress in the relapse and precipitation of these psychopathological states; (iii) the sensitizing effects of either psychostimulants or stress—both of which are interchangeable—on the DA neurons; and (iv) the importance of individual differences as regards the vulnerability and the sensitivity to the drug or stress precipitating effects.

These findings have led to the hypothesis that the effects of stress or sensitization to psychostimulants may depend on these individual coping mechanisms. As described above, displacement activity such as schedule-induced polydipsia stabilizes an organism that is disrupted by conflicts or by being prevented from reaching a goal and is adaptive roles through its de-arousing effects as shown by the decrease of costicosterone level and blockade of endogenous pain inhibitory systems (45, 46). Moreover, repeated administrations of psychostimulants and DA neuron manipulation facilitate the development of displacement activities, and coping with conflicting situations is related to enhanced activity in the dopaminergic neurons. Stress-reducing properties of adjunctive behavior reduces responsiveness to psychostimulants. Some rats learn the task and drink whereas others never learn, even after repeated sessions. Only the polydipsic rats displayed a lower behavioral activation in response to a low dose of amphetamine, whereas the nonpolydipsic rats and the rats without access to water and which were not able to engage in a displacement activity did not; that is, a reduced activation and desensitization of the DA neurons were revealed by the lower response to amphetamine. These results support the hypothesis that the effects of stress are determined by an individual's self-perceived ability to cope with stressors.

Two other sources of individual differences involve brain asymmetry and sex hormones. Rotational or turning preferences or asymmetrical orienting responses and other lateralized activities in normal animals may be related to differences in DA activity between the left and right basal ganglia, the left and right frontal cortex, and hippocampus (8, 21).

Inheritable or epigenetic determinants and sexual hormones such as testosterone also have been shown to contribute to lateralization as well as to sex-related differences in responses to drugs acting on the DA neurons which seem sexually dimorphic. The in vivo presence or absence of gonadal steroids modulate the release of dopamine measured in vitro from striatal tissue fragments obtained from female, but not male, rats (21). Similar differences have been reported in dopamine turnover and fluctuation across the estrous cycle (see ref. 21 for references). The effects of gonadal steroids on the behavioral responses modulated by the DA neurons have been extensively documented, and estrous cycle has been found to influence DA mechanisms in the frontal cortex and possibly in the tuberoinfundibular system (16). Female rats are more sensitive to amphetamine, and they have higher dopamine levels after drug administration in the striatum contralateral to the dominant direction of rotation. Thus, the general reaction of the organism under stress is partly dependent on individual differences to which DA sex-related neuronal asymmetry may contribute, although these individual differences are not yet frequently taken into account.

Chronic administration of indirect DA agonists induces reverse tolerance, or behavioral sensitization (37), to many of the effects on these drugs such as hypothermia, hyperactivity, stereotypy, and even seizures at the higher doses. When the same dose of stimulant is repeated, a progressive increase in locomotor activity occurs. This phenomenon has several characteristics: (i) it reflects a long-lasting change in responsivity which persists weeks and months after withdrawal of the drug; (ii) it involves changes in both the magnitude and the triggering of hyperactivity and stereotypy; (iii) females seem to be more sensitive than males; and (iv) it has been observed in a great variety of animals. Some studies have shown the environmental context and conditioning to play an important role in the process of sensitization and to be crucial for the ultimate degree of expression of drug-induced behavior. However, besides this obviously associative form of sensitization, a more subtle and complex form of long-lasting, time-dependent sensitization has been demonstrated (38). This kind of sensitization is a ubiquitous phenomenon, is demonstratable with a broad spectrum of agents, does not require daily treatment, can be obtained after a simple exposure to a low-to-moderate dose of various psychopharmacologically active drugs or to a stressful situation, is subject to cross-sensitization, and becomes more visible with the passage of time. Dopamine receptor antagonists have been found to block the development of sensitization but not to block it once it has developed (i.e., when given on the test day), suggesting that dopamine transmission is necessary to the occurrence and development of, but not to the expression of, stimulant-induced sensitization (40).

Controversies exist about the respective role of drug–environment conditioning and nonassociative sensitization in the development of behavioral sensitization (40). However, both drug–environment conditioning and sensitization, whether they be considered independently or together, involve long-lasting changes in the neural systems that mediate the behavioral responses measured.

Numerous data show enhanced dopamine metabolism in animals to which a subsequent challenge injection of amphetamine has been administered (for reviews, see refs. (40 and 49), and these changes may be at least partly responsible for behavioral sensitization. These results suggest that repeated exposure to the abnormally high concentration of transmitter released after repeated drug administration causes autoreceptors situated on presynaptic terminals, cell body, and/or dendrites of neurons to become subsensitive. Because these receptors are thought to control dopamine synthesis and release and the discharge rate (see Electrophysiological Properties of Midbrain Dopamine Neurons and Dopamine Autoreceptor Signal Transduction and Regulation), any subsensitivity of these receptors would trigger enhanced dopamine release by reducing the negative feedback. However, electrophysiological data here are somewhat contradictory (40). The electrophysiological effects dissipate quickly, unlike the behavioral and biochemical effects; they do not appear after a single drum administration and are often opposite to those predicted by the hypothesis. Also, in vitro, release of dopamine elicited by amphetamine is not modulated by presynaptic autoreceptors. More research is needed to explore the cascade of cellular events underlying the observed behavioral and neurochemical consequences of sensitization.

Electrical stimulation of specific restricted regions of the brain, particularly the medial forebrain bundle and its connections, can initiate and maintain operant behaviors and has provided a powerful means of studying of the neural mechanisms involved in reinforcement and appetitively or aversively motivated learning. Given that DA neurons modulate the activity of most of these regions through the medial forebrain bundle, it is not surprising that considerable modifications of brain stimulation effects have been observed after manipulating the DA transmissions.

Two hypotheses were developed for the role of dopamine in neuronal mechanisms. The first argued for a unitary system, and the DA neurons were said to be a critical link and were misleadingly said to form the "pleasure centers," leading to the "DA theory of reward" (50). The second strategy (multiple brain substrates for reward) was based on the fact that self-stimulation has been observed from a large diversity of structures, from all regions of the brain. The highest rates of self-stimulation obtained in the brain are in the VTA and even more in the mesencephalic and pontine raphe nuclei (21), regions through which the dense ascending and descending fibers linking the limbic forebrain–midbrain areas pass.

Furthermore, self-stimulation of many regions of the brain persists after lesion of the DA neurons or of their terminals. The existence of numerous areas from which such self-stimulation can be obtained without any DA influence is incompatible with the concept of a unitary substrate for brain stimulation reward. DA neurons were neither sufficient nor necessary (see ref. 21 for references) for self-stimulation, but may appear to be involved only because they modulate regions which are the authentic neural substrates for reinforcement or for performance (33).

Finally, the use of a quantitative technique for measuring changes in glucose utilization showed that self-stimulation of the medial forebrain bundle including DA fibers activates large neuronal systems including the dorsomedial thalamus, septum, ventral striatum, prefrontal cortex, amygdala, and hippocampus (36), implying that integration across any neuronal brain region plays a more important role than any specific neurotransmitter.

Less artificial than the self-stimulation phenomenon, intravenous drug self-administration (21, 50, and 51) has been studied routinely for the last two decades. Research in this field has direct important strategic implications for the study of drug-seeking addiction and dependence. In one view, a given drug may have a limited neurobiological specificity but can activate neuronal circuitries linked to one of the brain's reinforcement systems. Another view promotes the existence of a unique brain mechanism for the reward processes, of which the DA neurons are taken to be the most important component. Alternatively, the various classes of drugs may each have their own impact on specific receptors and synaptic transmissions, in which addiction results from a complex interplay between these separate networks (17). In this view, the DA pathways would be viewed as a more specific substrate for psychostimulant drugs.

In support of this hypothesis, nearly 20 years ago low doses of a dopamine receptor antagonist were reported to increase the response rate for amphetamine and cocaine self-administration in rats (52). These paradoxical results were interpreted as a reduction of the reinforcing effects of the self-administered drug. The partial blockade of the dopamine receptors by low doses of antagonist led the animal to respond more strongly, because higher doses of the drug were now required to maintain the same level of drug reward. The increase in self-administration rate, observed after the dopamine receptor blockade, was qualitatively similar to the effects of decreasing the dose of drug per injection. Higher doses of antagonist induced a complete blockade of the transmission and caused complete extinction of self-administration (52). Large 6-OHDA lesions of the mesocorticolimbic DA system can also produce extinction of already established selfadministration (21).

However, in contrast to studies in rats trained for drug self-administration, in drug-naive rats RF or 6-OHDA lesions of the mesocortical or mesolimbic DA cell bodies produce an increase in the rate of acquisition of amphetamine intravenous self-administration (21). This result indicated that the lesioned animals were more sensitive to the effects of the psychostimulant. These studies had three important characteristics which provide some unique insights into the role of the mesocorticolimbic DA system in psychostimulant reinforcement: (i) The dose of amphetamine available per lever-press or nose-poke was very small (7.5 mg/kg was generally used as compared to the usually available dose of at least 100 mg/kg); such low doses make it difficult, if not impossible, for a normal rat to acquire drug self-administration; (ii) most of the other studies (see ref. 21 for references) used an experimental paradigm in which the animals were experimentally manipulated (lesion or receptor blockade) after acquisition and stabilization of the self-administration response rate; and (iii) a two-nose-poke procedure was used during acquisition of drug self-administration, so that it is possible to study the rat's capacity to discriminate between the drug and the vehicle. These results suggest that partial lesions of the DA system may increase vulnerability to acquire self-administration of psychostimulants, perhaps via release of dopamine from remaining terminals acting on supersensitive postsynaptic receptors (see A Critical Analysis of Neurochemical Methods for Monitoring Transmitter Dynamics in the Brain).

Lesions of neurons containing the transmitter, serotonin, can also facilitate the acquisition of amphetamine self-administration. The role of serotonin can be explained through changes in dopamine turnover in other regions. After lesion of the raphe, dopamine utilization is enhanced in the nucleus accumbens but decreases in the prefrontal cortex (see ref. 21). Chronic rearing isolation and DA lesion in the amygdala induced the same biochemical pattern and also induced an increased sensitivity to the self-administered drug (7, 43). For example, rats housed in groups reliably fail to self-administer cocaine, whereas weanling rats housed under isolated conditions readily acquired an operant conditioning task to receive infusions of cocaine. Stress or food deprivation also induced changes in stimulant self-administration (20). These data suggest that environmental factors which induce changes in dopamine utilization in some parts of the DA network determine individual differences in the propensity to self-administer the drug. Conversely, a near total lesion of the noradrenergic dorsal bundle leads to the reverse pattern: a decrease in dopamine turnover at the cortical level without any changes in the nucleus accumbens level (see review in ref. 21).

During the last few years, research has focused on individual differences in vulnerability to drug intake and dependence, a fact largely emphasized by clinicians (29), and on factors which may predict such vulnerability. For instance, vulnerability and locomotor reactivity to stressful situations differentiate individuals, and such parameters predict individual vulnerability to drug selfadministration (34). Such sensitive animals present a specific pattern of DA activity: The transmitter is reduced in the prefrontal cortex and increased in the nucleus accumbens (35) (); moreover, these vulnerable animals present a higher and longer stress-induced increase in dopamine concentrations in the nucleus accumbens (41).

DA neurons have homeostatic and regulatory roles in that they allow the forebrain and cortical neuronal systems to function normally. The DA neurons are closely interrelated and controlled by so many feedbacks that it may be incorrect to conclude that the symptoms observed after restricted manipulations are directly and solely due to a unique set of neurons. A local lesion can bring about increases or decreases in dopamine utilization in the other regions and provokes a new pattern of regulation within the network. A lesion of the DA neurons disturbs many of the brain integrative functions not directly related to the sensory and motor processes (i.e., learning and memory, cognitive functions, and reinforcement processes). These deficits explain why organisms cannot normally survive without dopamine and why DA regulation is essential for the adaptative psychobiological capacities. After lesion these capacities are paradoxically virtually present but latent (i.e., not overtly expressed). They can be restored by L-dopa or various pharmacological treatments or by changes in the external (more than internal) environment (e.g., stress, strong emotional stimulation, arousing situations or stimuli, and cues previously paired with a reinforcer).

Manipulations of the DA neurons may help us to better understand the neurobiological function of the regions to which they project. Dopamine, besides the signal it translates, has a general role in activation; more dopamine, within the physiological limits, increases the adaptative capabilities and the vigor and probability of responses. Thus, a functional concept of the dopamine signal is to activate the final common pathway of several integrative processes.

published 2000