Activation of Corticotropin-Releasing Factor in the Limbic System During Cannabinoid Withdrawal

See allHide authors and affiliations

Science  27 Jun 1997:
Vol. 276, Issue 5321, pp. 2050-2054
DOI: 10.1126/science.276.5321.2050


Corticotropin-releasing factor (CRF) has been implicated in the mediation of the stress-like and negative affective consequences of withdrawal from drugs of abuse, such as alcohol, cocaine, and opiates. This study sought to determine whether brain CRF systems also have a role in cannabinoid dependence. Rats were treated daily for 2 weeks with the potent synthetic cannabinoid HU-210. Withdrawal, induced by the cannabinoid antagonist SR 141716A, was accompanied by a marked elevation in extracellular CRF concentration and a distinct pattern of Fos activation in the central nucleus of the amygdala. Maximal increases in CRF corresponded to the time when behavioral signs resulting from cannabinoid withdrawal were at a maximum. These data suggest that long-term cannabinoid administration alters CRF function in the limbic system of the brain, in a manner similar to that observed with other drugs of abuse, and also induces neuroadaptive processes that may result in future vulnerability to drug dependence.

Cannabis continues to be a major drug of abuse, and as many as 9% of Cannabis users may meet criteria for substance dependence (1). Short-term exposure to Cannabis derivatives (hashish, marijuana) produces subjective emotional responses ranging from mild relaxation to panic reactions (1, 2); long-term use of Cannabismay result in mental lethargy and anhedonia (3). A clear-cut abstinence syndrome is rarely reported, presumably because of the long half-life of cannabinoids, which precludes the emergence of abrupt abstinence symptoms (1), although nervousness, tension, restlessness, sleep disturbances, and anxiety have been described in humans, monkeys, and rats after termination of long-term cannabinoid administration (4). A distinct abstinence syndrome can, however, be elicited in animals treated with cannabinoids over a long period (5) by administering a competitive cannabinoid antagonist (6). This antagonist-precipitated withdrawal may unmask the development of underlying neuroadaptive processes that contribute to the development of cannabinoid dependence. The neurobiological substrates of cannabinoid-induced emotional responses remain to be elucidated, although they are likely to be mediated by activation of CB1 cannabinoid receptors, which are present in the limbic system and brain nuclei that have been implicated in stress responses (7). Psychotropic cannabinoids are potent activators of the hypothalamic-pituitary-adrenal (HPA) axis (8), and this property may contribute to the unpleasant side effects described by users of Cannabis.

A common element of withdrawal from drugs of abuse is a negative affective state that is characterized in humans by dysphoria and anxiety and in animals by a reward deficit and enhanced behavioral reactivity to stressors (9). We report here that cannabinoid withdrawal, induced by administration of a cannabinoid CB1antagonist, results not only in enhanced behavioral responses to stressors but also in increased release of CRF and induction of c-fos in the central nucleus of the amygdala. Our data reveal an unambiguous neurochemical response in the limbic system, attributable to long-term cannabinoid exposure, similar to that produced by other major drugs of abuse (9, 10). This finding supports the hypothesis that cannabinoids can set in motion neuroadaptive processes in the brain that contribute to the development of substance dependence.

Among the various brain neurochemical and neuroendocrine systems that participate in the mediation of motivational aspects of drug dependence, CRF may hold a prominent position (11). Recent evidence suggests that hypothalamic and extrahypothalamic CRF systems have a role in mediating cannabinoid-induced anxiety (12). Short-term treatment with the CB1 cannabinoid receptor agonist (−)-Δ8-tetrahydrocannabinol dimethyl heptyl (HU-210) activates the HPA axis in rats (12). In addition to activation of the pituitary-adrenal axis by hypothalamic CRF neurons (13), brain CRF systems, particularly in the central nucleus of the amygdala, appear to mediate behavioral responses to stressors. CRF neurons and receptors in the central nucleus of the amygdala participate in the arousal-enhancing properties of psychostimulants as well as in behavioral sensitization (14) and play a key role in anxiety reactions observed during ethanol withdrawal (10,11). A CRF antagonist, α-helical CRF(9–41), can also attenuate the anxiogenic behavioral effects of HU-210. Moreover, studies with intracranial microdialysis indicate that immobilization stress as well as ethanol and cocaine withdrawal result in elevated extracellular concentrations of CRF in the central nucleus of the amygdala (10).

We exploited the availability of a CB1 cannabinoid receptor antagonist, SR 141716A, to evaluate the role of the central amygdaloid CRF system in the effects of short-term and long-term cannabinoid exposure as well as in cannabinoid withdrawal (15). We used intracranial microdialysis (16, 17) to examine changes in extracellular CRF in the rat central nucleus of the amygdala in response to a single administration of either HU-210 or SR 141716A (Fig. 1D). Release of CRF was also monitored in rats after long-term (2 weeks) exposure to HU-210 and during a behavioral withdrawal syndrome induced by injection of SR 141716A after long-term exposure to HU-210. We studied the temporal profile of the abstinence syndrome as well as anxiety-like responses in different experimental groups with observational measures (18) and a defensive withdrawal test (19) previously characterized for use with cannabinoids (12). In addition, we examined activation of the HPA stress response by measuring plasma corticosterone concentrations (20). Finally, we examined the anatomical distribution of cannabinoid-responsive brain areas by analyzing the appearance of Fos immunoreactivity (21), which has been an effective tool for mapping neural activity after stress (22, 23), CRF administration (24), and drug withdrawal (25).

Figure 1

(A) Effects of a single injection of HU-210 (100 μg/kg) on CRF release from the central nucleus of the amygdala. Statistical analysis (one-way analysis of variance for repeated measures) revealed that HU-210 lowered CRF release [F 11,66 = 1.99, P < 0.05 (*),N = 7]. Vehicle injections did not alter CRF efflux (F 11,55 = 0.45, P = 0.93,N = 6). Administration of SR 141716A did not modify CRF release (31). (B) Effects of SR 141716A (3 mg/kg) on CRF release from the central amygdaloid nucleus in animals pretreated for 14 days with HU-210 (100 μg/kg). Cannabinoid withdrawal induced by SR 141716A was associated with increased CRF release [F 11,77 = 3.54, P < 0.005 (*), N = 8]. Vehicle injections did not alter CRF efflux (F 11,66 = 0.69, not significant, N= 7). Data in (A) and (B) were standardized by transforming dialysate CRF concentrations into percentages of baseline values based on averages of the first four fractions. (C) Mean ± SEM of summed cannabinoid withdrawal scores 0, 10, 30, and 60 min after injection of SR 141716A in rats treated for 14 days with HU-210 or its vehicle. The cannabinoid antagonist induced a mild behavioral syndrome in drug-naı̈ve rats receiving long-term pretreatment with vehicle (SR 141716A) and a clear withdrawal syndrome in animals pretreated with HU-210 (long-term HU-210 + SR 141716A) [F 2,18 = 33.49, P < 0.0001 (**)]. Rats pretreated with cannabinoid (long-term HU-210) that received vehicle on the test day did not exhibit withdrawal signs. Drug-naı̈ve control animals that received vehicle injections were indistinguishable from the long-term HU-210 treatment group, and cannabinoid-naı̈ve rats did not exhibit observable changes in behavior after a single injection of HU-210 (31). (D) Anatomical location of the active region of microdialysis probes (outer diameter, 0.5 mm) in animals subjected to SR 141716A–induced cannabinoid withdrawal.

A single injection of HU-210 decreased the amount of CRF released from the central nucleus of the amygdala (Fig. 1A). The inhibitory effects of HU-210 on CRF release were still apparent 24 hours after completion of the long-term cannabinoid treatment regimen (26). In contrast, induction of withdrawal by SR 141716A after long-term exposure to HU-210 had the opposite effect and increased CRF efflux (Fig. 1B). The increase in extracellular CRF concentration paralleled the progression of behavioral withdrawal symptoms over time (Fig. 1C). The behavioral changes after administration of the cannabinoid antagonist in animals receiving long-term treatment with HU-210 were also reflected in anxiety-like responses in the defensive withdrawal test (27) and activation of the HPA axis as revealed by increased plasma corticosterone concentrations (Table1). In addition, the changes in extracellular CRF concentration were accompanied by the appearance of increased Fos immunoreactivity in the central nucleus of the amygdala (Table 2 and Fig. 2).

Figure 2

Brain sections showing Fos immunoreactivity in cell nuclei of the central amygdaloid nucleus of rats after a single injection of vehicle (A), HU-210 (B), or SR 141716A (C) and during SR 141716A–induced withdrawal in rats receiving long-term HU-210 pretreatment (D). Fos immunoreactivity in rats undergoing long-term HU-210 treatment that received vehicle was indistinguishable from saline controls in (A) (31). Fos-immunopositive cell count (mean ± SEM) was as follows: 1.7 ± 1.7 [vehicle (N = 4)], 36.7 ± 6.7 [HU-210 (N = 3)], 8.0 ± 4.6 [SR 141716A (N = 4)], and 32.5 ± 4.4 [SR 141716A after long-term HU-210 (N = 4)]. Scale bar, 50 μm.

Table 1

Defensive withdrawal test results (seconds spent in the chamber; mean ± SEM) and plasma corticosterone concentrations (mean ± SEM) in rats after a single dose of HU-210 or SR 141716A and after long-term (14 days) exposure to HU-210 alone or followed by a single dose of SR 141716A.

View this table:
Table 2

Distribution of Fos-immunopositive cells after a single dose of HU-210 or SR 141716A and after long-term (14 days) exposure to HU-210 alone or followed by a single dose of SR 141716A. Each group contained three or four animals. Number of immunopositive cells is indicated as follows: 0 (−), 1 to 10 (+), 11 to 20 (++), 21 to 30 (+++), 31 to 50 (++++), >50 (+++++).

View this table:

The distribution of Fos immunoreactivity indicated that induction of withdrawal by SR 141716A activated not only the central nucleus of the amygdala but also other stress-responsive brain sites that receive projections from this nucleus, such as the shell of the nucleus accumbens, the bed nucleus of the stria terminalis (BNST), the paraventricular nucleus of the hypothalamus (PVN), and brainstem structures involved in autonomic responses to stress, including the nucleus of the solitary tract (NTS). Cannabinoid withdrawal was associated with activation of stress-related mid- and hindbrain nuclei, including the ventral tegmental area, locus coeruleus, central gray, NTS, and, especially, the area postrema (Table 2), brain regions that are also activated during opiate withdrawal (25) and are recognized as critical for conveying stress information to the PVN, which ultimately triggers the HPA response (28).

Cannabinoid withdrawal was also associated with increased Fos immunoreactivity in extrapyramidal motor regions that are rich in CB1 receptors (7), such as the caudate-putamen, ventral pallidum, and substantia nigra. This profile of neural activation is consistent with the dominant behavioral symptoms of cannabinoid withdrawal in rats (compulsive grooming and scratching, forepaw treading, and rubbing of the face), and it points toward involvement of the basal ganglia in the motor component of cannabinoid withdrawal. The greatest Fos immunoreactivity was found in the piriform cortex, a cortical area involved in limbic kindled seizures (29), which were observed in 2 of 10 animals.

Blockade of CB1 receptors with SR 141716A in cannabinoid-naı̈ve rats did not alter release of CRF in the amygdala and produced a different pattern of neuronal activation than in animals undergoing long-term HU-210 treatment. The antagonist increased Fos immunoreactivity mainly in the PVN, accumbens shell, and central gray matter, and it induced anxiety-like responses in both the defensive withdrawal test (Table 1) and the elevated plus-maze tests (30). However, SR 141716A did not increase plasma corticosterone concentrations in drug-naı̈ve animals (Table 1). In conjunction with earlier findings that SR 141716A can increase arousal and disrupt the sleep-waking cycle in rats (6), this observation suggests that endogenous “cannabinoid tone” may have a role in normal behavioral function without affecting HPA activity.

Together, our results provide in vivo neurochemical, endocrinological, and immunocytochemical evidence that long-term exposure to cannabinoids leads to neuroadaptive changes that result in enhanced release of CRF in the central amygdala as well as activation of stress-responsive nuclei during cannabinoid withdrawal. These changes are consistent with the irritability and discomfort that have been described to occur after cessation of long-term consumption of marijuana (4). Moreover, the neurobiological consequences of cannabinoid withdrawal, in particular the alteration in amygdaloid CRF function (9,10), are similar to those observed during withdrawal from ethanol, cocaine, and opiates as well as during exposure to environmental stressors. Thus, activation of the amygdaloid CRF system may have a motivational role of mediating the stress-like symptoms and negative affect that accompany withdrawal and, therefore, may be a common element in development of dependence on drugs of abuse. The demonstration that long-term exposure to a cannabinoid agonist evokes neuroadaptive processes in the limbic system that resemble those associated with other major drugs of abuse may provide a neurobiological basis for the gateway hypothesis. Cannabinoid abuse, by activating CRF mechanisms, may lead to a subtle disruption of hedonic systems in the brain that are then “primed” for further disruption by other drugs of abuse (9).


View Abstract


Navigate This Article