Brain cannabinoid CB2 receptors modulate cocaine's actions ... - Nature

7 downloads 0 Views 856KB Size Report
Jul 24, 2011 - Furthermore, activation of CB2 receptors by 2-arachidonoylglycerol, JWH015 or JWH133 inhibits locomotion10,11, morphine-6-glucuronide–.
a r t ic l e s

Brain cannabinoid CB2 receptors modulate cocaine’s actions in mice

© 2011 Nature America, Inc. All rights reserved.

Zheng-Xiong Xi1, Xiao-Qing Peng1,3, Xia Li1,3, Rui Song1–3, Hai-Ying Zhang1, Qing-Rong Liu1, Hong-Ju Yang1, Guo-Hua Bi1, Jie Li1 & Eliot L Gardner1 The presence and function of cannabinoid CB 2 receptors in the brain have been the subjects of much debate. We found that systemic, intranasal or intra-accumbens local administration of JWH133, a selective CB 2 receptor agonist, dosedependently inhibited intravenous cocaine self-administration, cocaine-enhanced locomotion, and cocaine-enhanced accumbens extracellular dopamine in wild-type and CB 1 receptor knockout (CB 1−/−, also known as Cnr1 −/−) mice, but not in CB 2−/− (Cnr2 −/−) mice. This inhibition was mimicked by GW405833, another CB 2 receptor agonist with a different chemical structure, and was blocked by AM630, a selective CB 2 receptor antagonist. Intra-accumbens administration of JWH133 alone dose-dependently decreased, whereas intra-accumbens administration of AM630 elevated, extracellular dopamine and locomotion in wild-type and CB 1−/− mice, but not in CB 2−/− mice. Intra-accumbens administration of AM630 also blocked the reduction in cocaine self-administration and extracellular dopamine produced by systemic administration of JWH133. These findings suggest that brain CB 2 receptors modulate cocaine’s rewarding and locomotorstimulating effects, likely by a dopamine-dependent mechanism. The behavioral and psychoactive effects of cannabinoids are mediated by activation of brain cannabinoid receptors1,2. Two major cannabinoid receptors (CB1 and CB 2) have been identified. Given that CB 1 receptors are highly expressed in the brain 2,3 and CB 2 receptors are found primarily in the periphery4,5, it has generally been believed that the behavioral and psychotropic effects of cannabinoids are CB1-mediated1,2 and that CB2 receptor ligands have no psychoactive effects 6. However, the purported lack of brain CB 2 receptors has been challenged by recent reports of low densities of CB2 receptors on microglia7 and neuronal8–11 cells in several brain regions, ­including the anterior olfactory nucleus, cerebral cortex, cerebellum, hippocampus, striatum and brainstem. Furthermore, activation of CB 2 receptors by 2-arachidonoylglycerol, JWH015 or JWH133 ­inhibits locomotion 10,11, morphine-6-glucuronide– induced emesis 11 and neuropathic pain 12,13, while stimulating neural progenitor ­proliferation 14 and producing neuroprotective effects 15,16. More recent studies have suggested that CB 2 receptor activation inhibits neuronal firing in ­dorsal-root ganglia and spinal cord 17,18 and GABAergic transmission in rat cerebral cortex 19. These data indicate that functional CB 2 receptors may be expressed on CNS neuronal cells, prompting us to re-­examine the role of CB 2 receptors in drug reward and ­addiction. To this end, we used highly selective CB 2 receptor agonists and antagonists, combined with specific CB1 receptor knockout (CB1−/−) and CB 2 receptor knockout (CB 2−/−) mice, to investigate the possible involvement of brain CB 2 receptors in cocaine’s behavioral and neurochemical effects.

RESULTS JWH133 inhibits intravenous cocaine self-administration To determine whether CB2 receptor activation alters intravenous cocaine self-administration, we used JWH133, a highly selective CB2 receptor agonist (200-fold selectivity for CB2 versus CB1)20,21, and AM630, a highly selective CB2 receptor antagonist (160-fold selectivity for CB2 versus CB1)20,21, as pharmacological tools. We found that over 50% of wild-type (20 of 34) and CB2−/− (22 of 36) mice, but only about 30% of CB1−/− (10 of 36) mice, acquired stable intravenous cocaine self-administration, defined as 20 or more infusions per 3-h session, with a regular pattern of self-administration achieved after 10 d of training (Supplementary Fig. 1). Notably, CB1−/− mice displayed a significant reduction in both the total number (P < 0.05, two-way ANOVA; Supplementary Fig. 1a) and rate (infusions per h, P < 0.05; Supplementary Fig. 1b) of cocaine infusions on days 1–5, as compared with wild-type or CB2−/− mice. In addition, the ­majority of CB1−/− mice (7 of 10) ­displayed a distinct ‘burst-like’ drug-­taking pattern, with long inter-burst intervals, whereas wild-type and CB2−/− mice displayed evenly paced drug-taking without a significant difference between the two strains (Supplementary Fig. 1c). These findings suggest that deletion of CB1 receptors may lower cocaine’s rewarding efficacy, leading to a compensatory increase in drug intake during each individual drug-taking episode. This is further supported by the finding that CB1−/− mice displayed a significant reduction in break-point level for cocaine self-administration under progressive-ratio reinforcement, as ­compared with wild-type mice (P < 0.05; Supplementary Fig. 1d). Given that the progressive-ratio

1Intramural

Research Program, National Institute on Drug Abuse, Baltimore, Maryland, USA. 2Beijing Institute of Pharmacology and Toxicology, Beijing, China. 3These authors contributed equally to this work. Correspondence should be addressed to Z.-X.X. ([email protected]). Received 25 April; accepted 13 June; published online 24 July 2011; doi:10.1038/nn.2874

1160

VOLUME 14 | NUMBER 9 | SEPTEMBER 2011  nature NEUROSCIENCE

a r t ic l e s

**

10

*

0

CB1–/–

Wild type

CB2–/–

d

20

###

10

20

40

*

*

*

60 80 100 120 Time (min)

30

Cocaine infusions per h

n=9

* (n = 9)

n=9

20

**

10

**

0

Ve AM h + + V e 63 JW h 0 + H13 JW 3 A AM M6 H13 3 25 30 + 1 V + JW eh AM H 25 133 1 + Ve h

–24 1 24 48 Time (h) after 20 mg per kg JWH133

Ve h

Cocaine infusions per h

2

break point, defined as the maximal work performed by the animal to get a cocaine infusion, is cocaine dose dependent and positively correlated to reward strength22, the reduction in progressive-ratio break point that we observed in CB1−/− mice suggests that there is a reduction in cocaine’s reward strength and/or motivation for cocainetaking behavior. This is consistent with previous findings that CB1 receptor deletion impairs cocaine’s rewarding, locomotor-stimulating and dopamine (DA)-elevating effects23,24. Intraperitoneal administration of JWH133 (10 or 20 mg per kg of body weight) produced a significant and dose-dependent reduction in cocaine self-administration and cocaine intake in both wild-type (P < 0.001, one-way ANOVA) and CB1−/− (P < 0.05) mice, but not in CB2−/− mice (P = 0.58; Fig. 1a). This ­inhibition lasted for no longer than 24 h after JWH133 administration (20 mg per kg; Fig. 1b,c). Pretreatment with AM630, a selective CB2 receptor antagonist, but not with AM251, a selective CB1 receptor antagonist25, significantly attenuated JWH133-induced inhibition of cocaine self-administration (P < 0.001, one-way ANOVA; Fig. 1d). This suggests that JWH133’s attenuating effect is mediated

20 15 10 5 0

* 6 4

*** 6

0 3 10 GW405833 (mg per kg)

120

c 25

Wild type

100 80 60 40 20

9

*** 15

*** 11

6

***

0

Cocaine infusions per h

b

Wild type

V JW ehi H cle 1 pe 0 m JW r k g H g 20 AM per mg 25 kg 1 pe 3 m rk g g

25

Break-point (% baseline)

a Cocaine infusions per h

© 2011 Nature America, Inc. All rights reserved.

4

0

c 30

0

6

20

by activation of CB2, rather than by CB1, receptors. This ­conclusion is further supported by the ­ finding that systemic ­ administration of GW405833 (3 or 10 mg per kg, intraperitoneal), another highly selective, but structurally distinct, CB2 receptor agonist26, also inhibited cocaine self-administration in wild-type mice (P < 0.001; Fig. 2a). To determine whether JWH133-induced attenuation of cocaine self-administration was a result of a reduction in cocaine’s rewarding efficacy, we studied JWH133’s effect on intravenous cocaine selfadministration under progressive ratio reinforcement. We found that systemic administration of JWH133 (10, 20 mg per kg, intraperitoneal) significantly lowered the progressive-ratio break point for cocaine selfadministration in wild-type mice (P < 0.01; Fig. 2b), suggesting that there was a reduction in cocaine’s reward strength and/or motivation for drug-taking behavior after JWH133 administration. We previously showed that CB1 receptor blockade by AM251 substantially lowered the progressive-ratio break point for cocaine self-­administration in rats27. We therefore also tested AM251 and found that AM251 (3 mg per kg) lowered the progressive-ratio break point for cocaine selfadministration in wild-type mice (P < 0.001; Fig. 2b). These data suggest that the JWH133-induced reduction in cocaine self-administration resulted from a reduction in cocaine’s rewarding efficacy. JWH133 inhibits cocaine intake via brain CB2 receptors To further determine whether JWH133’s action was mediated by activation of brain or peripheral CB2 receptors, we first studied the effects

d 25

Wild type n = 10

15

***

10 5 0

0 50 100 Intranasal JWH133 (µg per side)

20

Wild type n=9

15 10 5 0

0 100 200 Intravenous JWH133 (µg)

e 30 25 20

Wild type

f n=3

n=9

*

15 10 5 0

Intra-NAc JWH133 (µg per side)

Cocaine infusions per h

(n = 6)

A AM M6 63 30 JW 0 + H

20

8

0 3 1 3

10

(n = 8)

0.

(n = 9)

Figure 1  Effects of JWH133 on cocaine self-administration. (a) Systemic administration of JWH133 (10 and 20 mg per kg, intraperitoneal, 30 min before testing) inhibited cocaine self-administration under FR1 reinforcement in wild-type (one-way ANOVA, F2,16 = 13.09, P < 0.001) and CB1−/− (F2,10 = 5.01, P < 0.05) mice, but not in CB2−/− (F2,14 = 0.56, P = 0.58) mice. (b) Time course of JWH133’s attenuation of cocaine self-administration in wild-type mice on the test day. (c) Time course of recovery of cocaine self-administration in wild-type mice after JWH133 administration. (d) In wild-type mice, JWH133-induced attenuation of cocaine self-administration was prevented by pretreatment with the CB2 receptor antagonist AM630 (10 mg per kg, intraperitoneal, 30 min before JWH133), but not by pretreatment with the CB1 receptor antagonist AM251 (3 mg per kg, intraperitoneal) (F5,40 = 6.31, P < 0.001). Neither AM630 nor AM251 altered cocaine self-administration in wild-type mice. Data are means ± s.e.m. *P < 0.05, **P < 0.01 compared with vehicle (veh) control groups, ###P < 0.001, compared with pre-JWH133 (−24 h) condition.

Vehicle JWH133 (10 mg per kg) JWH133 (20 mg per kg)

Cocaine infusions per h

30

b

Vehicle JWH133 (10 mg per kg) JWH133 (20 mg per kg)

Cocaine infusions per h

40

Cocaine infusions

Cocaine infusions per h

a

CB2–/– 25 20

n=6

15 10 5 0

0 3 Intra-NAc JWH133 (µg per side)

Figure 2  Effects of GW405833 or JWH133 on cocaine self-administration. (a) GW405833 (3 and 10 mg per kg, intraperitoneal) dose-dependently inhibited cocaine self-administration under FR1 reinforcement in wild-type mice (one-way ANOVA, F2,6 = 20.03, P < 0.01). n is indicated for each bar. (b) Treatment with JWH133 (10 and 20 mg per kg) or AM251 (3 mg per kg, intraperitoneal) significantly lowered the cocaine self-administration break-point under progressive-ratio reinforcement in wild-type mice (F3,37 = 13.83, P < 0.001). (c) Intranasal microinjections of JWH133 (50 and 100 µg per nostril) dose-dependently inhibited cocaine self-administration under FR1 reinforcement (F2,18 = 14.34, P < 0.001). (d) Intravenous injection of the same micro-quantity (100 and 200 µg) of JWH133 as was administered intranasally had no effect on cocaine self-administration (F2,16 = 1.59, P = 0.23). (e) Intra-NAc microinjections of JWH133 (0.3, 1 and 3 µg per side) dose-dependently inhibited cocaine self-administration under FR1 reinforcement in wild-type mice. This inhibition was blocked by intra-NAc co-administration of AM630 (3 µg per side, F3,24 = 4.49, P < 0.05). (f) IntraNAc administration of JWH133 (3 µg per side) had no effect on cocaine self-administration in CB2−/− mice (F1,10 = 2.37, P = 0.15). Data are means ± s.e.m. *P < 0.05, ***P < 0.001, compared with vehicle control group.

nature NEUROSCIENCE  VOLUME 14 | NUMBER 9 | SEPTEMBER 2011

1161

a r t ic l e s a 8,000

Wild type

b 2,000

Vehicle JWH133 (10 mg per kg) JWH133 (20 mg per kg)

CB1–/–

c

Vehicle JWH133 (10 mg per kg) JWH133 (20 mg per kg)

8,000

Vehicle JWH133 (10 mg per kg) JWH133 (20 mg per kg)

6,000 n=9

JWH133

2,000

Distance

4,000

1,000

******

0

JWH133

Cocaine (10 mg per kg) –1

2

2,000 0

Cocaine (10 mg per kg) –1

n=7

JWH133

4,000

***

–1,000

1

0

n = 10

***

0

**

–2,000

Distance

6,000 Distance

CB2–/–

0

1

Cocaine (10 mg per kg)

–2,000 –1

2

0

Time (h)

Time (h)

1

2

Time (h)

JWH133 itself has no reinforcing or aversive effects We further examined whether JWH133 itself has cocaine-like rewarding effects. To address this issue, we first trained mice to acquire

10,000

Vehicle JWH133 (10 mg per kg) JWH133 (20 mg per kg)

b

8,000

10,000 8,000

6,000 4,000

Distance

n=8 n = 8–10

* n = 8 –10

2,000

2,000

**

0 Wild type

CB1–/–

4,000

CB2–/–

c

Intra-NAc Vehicle Intra-NAc JWH133 (1 µg per side) Intra-NAc JWH133 (3 µg per side) n=8

6,000 n=8

*

n=7

15,000

Vehicle AM630 (10 mg per kg) AM630 (20 mg per kg)

–/–

CB1

d

15,000

n=8

10,000 n=8 5,000

Wild type 10,000

*

CB1–/–

*

CB2–/–

* n=8

5,000

n=8

n=7

n=7

**

0 Wild type

JWH133 inhibits cocaine-enhanced locomotion To determine whether JWH133’s effect on cocaine self-administration generalizes to other cocaine actions, we investigated the effects of JWH133 on cocaine-enhanced locomotion. Systemic administration of cocaine (10 mg per kg) produced a robust increase in locomotion in all three mouse strains (Fig. 3). Pretreatment with JWH133 (10 and 20 mg per kg, 30 min before cocaine) dose­dependently attenuated cocaine-enhanced locomotion in wild-type (Fig. 3a) and CB1−/− (Fig. 3b) mice, but not in CB2−/− (Fig. 3c) mice.

Distance

a

stable cocaine self-administration, and then cocaine was replaced by JWH133 (1 mg per kg per infusion) or vehicle. We found that neither JWH133 nor vehicle sustained stable self-administration in mice that were previously trained to self-administer cocaine (Supplementary Fig. 2a). In fact, the self-administration behavior underwent gradual extinction over the 5 d of substitution testing. This extinction pattern was essentially identical to that seen when vehicle was substituted for cocaine. However, when JWH133 or vehicle was replaced by cocaine, self-administration behavior returned to the levels observed during stable cocaine self-administration. In addition, we also found that cocaine (10 and 20 mg per kg, intraperitoneal) produced a substantial and robust conditioned place preference, whereas JWH133, at the same doses, produced neither conditioned place preference nor place aversion in wild-type mice (Supplementary Fig. 2b). These findings suggest that JWH133 has no cocaine-like reinforcing or aversive effects in mice.

Distance

of intranasal microinjections of JWH133 on intravenous cocaine selfadministration. Extensive studies have shown that a wide variety of compounds that cannot penetrate the blood-brain barrier can be delivered directly from nose to brain28. We found that intranasal microinjections of JWH133 (50 and 100 µg per 10 µl per side) dose-dependently inhibited intravenous cocaine self-administration (Fig. 2c). To explore the possibility that the effects of intranasal JWH133 might be mediated by drug absorption into the nasal vasculature with subsequent venous delivery of drug to pharmacological site(s) of action, we observed the effects of intravenous injection of the same micro-quantities of JWH133 as were administered intranasally on cocaine self-administration. We found that intravenous microinjections of JWH133 (100 and 200 µg) had no effect on cocaine self-administration (Fig. 2d). These data suggest that intranasal JWH133–induced pharmacological effects are mediated by activating brain rather than peripheral CB2 receptors. To further explore this issue, we observed the effects of local administration of JWH133 into the nucleus accumbens (NAc) on cocaine selfadministration. We found that intra-NAc microinjections of JWH133 (0.3, 1 and 3 µg per side) significantly and dose-dependently inhibited cocaine self-administration in wild-type mice (P < 0.05, one-way ANOVA; Fig. 2e), but not in CB2−/− mice (P > 0.05; Fig. 2f). This inhibition was blocked by intra-NAc co-administration of AM630 (3 µg per side).

Distance

© 2011 Nature America, Inc. All rights reserved.

Figure 3  Systemic administration of JWH133 (10 and 20 mg per kg, intraperitoneal, 30 min before cocaine) dose-dependently inhibited cocaine-enhanced locomotion in wild-type (a, two-way ANOVA for repeated measures over time, F2,16 = 14.45, P < 0.001) and CB1−/− (b, F2,18 = 12.57, P < 0.001) mice, but not in CB2−/− (c, F2,12 = 0.17, P = 0.85) mice. Data are means ± s.e.m. **P < 0.01, ***P < 0.001, compared with vehicle treatment group.

CB2–/–

0 Wild type

–/–

CB1

CB2–/–

0 0 1 310 A+J 0 310 0 3 10 Intra-NAc AM630 (µg per side)

Figure 4  Effects of systemic or local intra-NAc administration of JWH133 or AM630 on locomotor activity. (a) Systemic administration of JWH133 (10 or 20 mg per kg, intraperitoneal) dose-dependently inhibited locomotion in wild-type (one-way ANOVA, F2,24 = 8.03, P = 0.002) and CB1−/− (F2,25 = 13.44, P < 0.001) mice, but not in CB2−/− (F2,14 = 3.36, P > 0.05) mice. (b) Intra-NAc microinjections of JWH133 (1 or 3 µg per side) significantly inhibited locomotion in wild-type (F2,14 = 4.17, P < 0.05) and CB1−/− (F2,12 = 4.91, P < 0.05) mice, but not in CB2−/− (F2,14 = 0.04, P > 0.05) mice. (c) Systemic administration of AM630 did not alter locomotion in any strain of mice. (d) Intra-NAc administration of AM630 (1, 3 or 10 µg per side) significantly augmented locomotion in wild-type (F3,21 = 4.62, P < 0.05) and CB1−/− (F2,12 = 10.57, P < 0.01) mice, but not in CB2−/− (F2,14 = 0.05, P > 0.05) mice. A+J, AM630 plus JWH133 (3 µg per side). Data are means ± s.e.m. *P < 0.05, **P < 0.01, compared with vehicle control group.

1162

VOLUME 14 | NUMBER 9 | SEPTEMBER 2011  nature NEUROSCIENCE

a r t ic l e s

150

b 150

Vehicle (n = 8) JWH133 (10 mg per kg, n = 10) AM630 (10 mg per kg, n = 8) AM630 + JWH133 (n = 6)

CB1–/–

**

50

0

d

** ***** ** *** *** *** ** ** ** *** *** *** *** *** *** ********* *** *** JWH133 *************** –1

0

1

DA (% baseline)

4

JWH133 (10 mg per kg, n = 5)

100

e

0

1

*** ** ##

100

–2

–1

4

JWH133 (10 mg per kg, n = 6)

#

0

1

50

f

JWH133 (20 mg per kg, n = 6)

JWH133

–1

0

1

2 3 Time (h)

4

5

Vehicle (n = 11)

300

JWH133 (10 mg per kg, n = 6) JWH133 (20 mg per kg, n = 8)

*** ***

200

JWH133

**

**

#

100

0

JWH133 (20 mg per kg, n = 9)

CB2–/–

*** 200

–2

–1

Time (h)

0 Time (h)

1

2

JWH133

*** *** * *

100

Cocaine (10 mg per kg) 2

JWH133 (10 mg per kg, n = 12)

CB2

100

0

5

Vehicle (n = 7)

Cocaine (10 mg per kg) 0

2 3 Time (h)

Vehicle (n = 5) –/–

150

300

*** JWH133

–1

CB1–/–

JWH133 (20 mg per kg, n = 8)

200

** **** * ***** *** *** ** *** *** *** *** JWH133 ***

50

0

5

Vehicle (n = 10)

Wild type 300

© 2011 Nature America, Inc. All rights reserved.

2 3 Time (h)

DA (% baseline)

100

DA (% baseline)

DA (% baseline)

JWH133 (20 mg per kg, n = 9)

c

DA (% baseline)

Vehicle (n = 11) JWH133 (3 mg per kg, n = 6) JWH133 (10 mg per kg, n = 7)

Wild type

DA (% baseline)

a

Cocaine (10 mg per kg) 0 Time (h)

Figure 5  Effects of systemic JWH133 and/or AM630 on NAc DA. (a–f) Systemic administration of JWH133 (3, 10 or 20 mg per kg, intraperitoneal) dose-dependently inhibited basal (a–c) or cocaine-enhanced (d–f) extracellular NAc DA in wild-type (a, two-way ANOVA for repeated measures over time, F3,29 = 25.97, P < 0.001; d, F2,19 = 4.47, P < 0.05) and CB1−/− (b, F3,28 = 10.07, P < 0.001; e, F2,16 = 4.78, P < 0.05) mice, but not in CB2−/− (c, F2,23 = 0.10, P > 0.05; f, F2,22 = 1.53, P > 0.05) mice. AM630 alone (10 mg per kg, intraperitoneal) did not alter NAc DA in CB1−/− mice, whereas AM630 pretreatment (10 mg per kg, intraperitoneal) prevented JWH133-induced inhibition of NAc DA in CB1−/− mice (b). Data are means ± s.e.m. *P < 0.05, **P < 0.01, ***P < 0.001, compared with pre-drug baseline. #P < 0.05, ##P < 0.01, compared with vehicle treatment group.

Systemic administration of the same doses of JWH133 alone also significantly inhibited locomotion in a dose-dependent ­manner in wild-type (P < 0.05) and CB1−/− mice (P < 0.01), but not in CB2−/− mice (Fig. 4a), suggesting that there is an effect mediated by activation of CB2 receptors. As the same doses of JWH133 alone did not alter locomotor performance on a fast-running rotarod device in all three mouse strains (Supplementary Fig. 3), we infer that JWH133’s ­inhibition of cocaine self-administration or locomotion is not produced by nonspecific impairment of locomotor capacity. To further determine whether such locomotor inhibition is mediated by activation of brain CB2 receptors, we observed the effects of intra-NAc JWH133 and/or AM630 on locomotion. We found that intra-NAc microinjections of JWH133 (1 and 3 µg per side) significantly inhibited locomotion in wild-type and CB1−/− mice (P < 0.05), but not in CB2−/− mice (Fig. 4b), in a dose-dependent manner similar to systemic administration (Fig. 4a). Notably, systemic administration of AM630 did not alter locomotion in any of the mouse strains that we tested (Fig. 4c). However, when locally administered into the NAc, AM630 (1, 3 and 10 µg per side) significantly increased locomotor activity (P < 0.05; Fig. 4d) in wild-type and CB1−/− mice, but not in CB2−/− mice. These data suggest that CB2 receptors may tonically modulate locomotion. A higher level of AM630 in the brain, achieved by local rather than by systemic administration, appears to be required to block endocannabinoid action on brain CB2 receptors. JWH133 inhibits cocaine-enhanced extracellular NAc DA Given the crucial role of the mesolimbic DA system in cocaine self-administration and modulation of locomotion29, we further investigated the effects of JWH133 on basal and cocaine-enhanced nature NEUROSCIENCE  VOLUME 14 | NUMBER 9 | SEPTEMBER 2011

DA in the NAc by in vivo microdialysis. We did not see significant ­ ifferences in basal levels of extracellular NAc DA between wild-type d (0.25 ± 0.05 nM, n = 26) and CB2−/− (0.24 ± 0.05 nM, n = 22) mice (P > 0.05; Supplementary Fig. 4). However, CB1−/− mice displayed a significant basal reduction (0.13 ± 0.04 nM, n = 28), as compared with wild-type mice (P < 0.05; Supplementary Fig. 4). Consistent with our findings in cocaine self-administration and locomotion, systemic administration of JWH133 (3, 10 and 20 mg per kg, ­intraperitoneal) also significantly (P < 0.001) and dose-dependently lowered extracellular NAc DA in wild-type (Fig. 5a) and CB1−/− (Fig. 5b) mice, but not in CB2−/− (Fig. 5c) mice. This reduction in NAc DA was blocked by AM630 (10 mg per kg, intraperitoneal) in CB1−/− mice (Fig. 5b), suggesting that JWH133’s DA-inhibiting effect is mediated by activation of CB2 receptors. Moreover, pretreatment with the same doses of JWH133 also significantly attenuated (P < 0.05) cocaineenhanced NAc DA in wild-type (Fig. 5d) and CB1−/− (Fig. 5e) mice, but not in CB2−/− (Fig. 5f) mice. To determine whether this inhibition is mediated by activation of brain or peripheral CB2 receptors, we also observed the effects of intranasal or intra-NAc local administration of JWH133 on extracellular DA. We found that intranasal administration of JWH133 (100 µg per nostril) produced a robust reduction in extracellular NAc DA in wild-type and CB1−/− mice, but not in CB2−/− mice (Fig. 6a). Similarly, intra-NAc local perfusion of JWH133 (1–1,000 µM) robustly lowered extracellular DA in both wild-type and CB1−/− mice, but not in CB2−/− mice (Fig. 6b). In fact, an unexpected increase in extracellular DA was observed in CB2−/− mice after local administration of JWH133. The underlying mechanisms are unclear. One possibility is that JWH133 may bind to other (non-CB2) receptors 1163

b

Wild type (n = 5) 200

CB1–/– (n = 8) –/–

CB2

200

* 100

0

* * * Intranasal JWH133 (100 µg per side) –1

CB1–/– (n = 7)

CB2–/– (n = 7)

(n = 5)

150

50

Wild type (n = 5) 250

DA (% baseline)

a DA (% baseline)

* * ** ** 1

0

* ***

*** ** ***

2

0

c

d

CB1–/– (n = 7)

*** *** *** *** *** ****** ** ** ** ** ****** ** * ** * * * * * * * ** * * *

–/–

400 300 200

CB2

* *

* *

100 µM 1,000 µM

2

3

4

Time (h)

Wild type (n = 7) 500

10 µM 1

Time (h)

*

***** ********* ****** ***

Intra-NAc 1 µM JWH133: –1

* *

* *

100

0

3

*

150

50

Wild type (n = 7) –/–

200

CB1

(n = 7)

(n = 8)

100 Intra-NAc 0 AM630: 1 µM 10 µM –100

in CB2−/− mice, producing an increase in –1 0 1 extracellular DA. Congruent with this finding, intra-NAc local perfusion of AM630 (1, 10 and 100 µM) elevated extracellular DA in a concentrationdependent manner in both wild-type and CB1−/− mice, but not in CB2−/− mice (Fig. 6c), suggesting that endocannabinoids tonically modulate NAc DA release by activating CB2 receptors in the brain. Furthermore, AM630-enhanced extracellular DA appeared more robust in CB1−/− mice than in wild-type mice (Fig. 6c), suggesting that there is a higher endocannabinoid basal activity on CB2 receptors in the brains of CB1−/− mice. Moreover, intra-NAc local perfusion of AM630 also blocked the reduction in extracellular NAc DA that was produced by systemic administration of JWH133 in wild-type and CB1−/− mice (Fig. 6c,d), suggesting that JWH133-induced inhibition of DA release is mediated by activation of NAc CB2 receptors. The microdialysis probes and microinjection cannulae were located in the NAc (Supplementary Fig. 5).

JWH133 (10 mg per kg) 100 µM

DA (% baseline)

Figure 6  Effects of intranasal or intra-NAc local perfusion of JWH133 or AM630 on extracellular NAc DA. (a) Intranasal administration of JWH133 (100 µg per nostril) significantly lowered extracellular DA in wild-type and CB1−/− mice, but not in CB2−/− mice (two-way ANOVA for repeated measures over time, F2,15 = 10.81, P = 0.001). (b) Intra-NAc local perfusion of JWH133 lowered extracellular DA in wild-type and CB1−/− mice in a dose-dependent manner and elevated extracellular DA in CB2−/− mice (F2,18 = 47.00, P < 0.001). (c) Intra-NAc local perfusion of AM630 elevated extracellular DA in wild-type and CB1−/− mice, but not in CB2−/− mice (F2,18 = 12.13, P < 0.001). Furthermore, AM630-enhanced extracellular DA appeared to be more robust in CB1−/− mice than in wild-type mice (F1,12 = 7.50, P < 0.05). (d) Renormalized data over new baselines 1 h before JWH133 administration from the data in c, illustrating that intra-NAc local perfusion of AM630 blocked JWH133’s action on extracellular DA in wild-type and CB1−/− mice. Data are means ± s.e.m. *P < 0.05, **P < 0.01, ***P < 0.001, compared with pre-drug baseline.

DA (% baseline)

© 2011 Nature America, Inc. All rights reserved.

a r t ic l e s

150

100

JWH133 (10 mg per kg)

50

Intra-NAc AM630: 100 µM 0

2 3 Time (h)

4

5

6

–1

0

1 Time (h)

2

3

DISCUSSION We found that systemic administration of the CB2 receptor agonist JWH133 robustly and dose-dependently inhibited intravenous cocaine self-administration under both fixed ratio 1 (FR1) and progressive ratio reinforcement and inhibited cocaine-enhanced locomotion and extracellular NAc DA in wild-type and CB1−/− mice, but not in CB2−/− mice. This effect was mimicked by GW405833 (another selective CB2 receptor agonist) and blocked by AM630, a selective CB2 receptor antagonist, but not by AM251, a selective CB1 receptor antagonist, suggesting that the effect is mediated by activation of CB2 receptors. Furthermore, intranasal microinjections of JWH133, but not intravenous injections of the same micro-quantities of JWH133 as were injected intranasally, robustly and dose-dependently inhibited intravenous cocaine self-administration, suggesting that the effect was mediated by activation of brain and not peripheral CB2 receptors. This is further supported by the finding that local intra-NAc administration of JWH133 robustly inhibited cocaine self-administration in a dose-dependent manner, an effect that was blocked by intra-NAc co-administration of AM630. In addition, intra-NAc local administration of JWH133 dose-dependently lowered, whereas AM630 elevated, basal levels of locomotion and extracellular NAc DA. Intra-NAc local perfusion of

AM630 blocked the reduction in cocaine self-­administration and NAc DA produced by systemic ­ administration of JWH133. These data ­ suggest that both the behavioral and neurochemical effects of JWH133 are mediated by activation of brain CB2 receptors. Systemic administration of AM630 did not alter, whereas intra-NAc local administration of AM630 elevated, extracellular DA and locomotion, suggesting that local AM630 administration is more effective than systemic administration. This may be related to AM630’s relatively poor pharmacokinetic properties and/or blood-brain barrier passage. In addition, intra-NAc AM630 robustly ­elevated extracellular DA and locomotion, but did not alter cocaine self-administration. This may be related to previous findings that locomotion is largely DA dependent30, whereas cocaine self-administration is dependent on both DA and non-DA mechanisms31. Pharmacological blockade of NAc CB2 receptors elevated, whereas genetic deletion of CB2 receptors did not alter, basal extracellular DA in the NAc, for reasons that are unclear. One possibility is that CB 2 receptor deletion–induced disinhibition of NAc DA release may be compromised by actions in other brain loci that modulate the mesolimbic DA system. Another possibility is that neuroadaptative processes may antagonize CB2 receptor inactivation–induced DA neuronal disinhibition after CB2 receptor deletion. Whatever the exact mechanism(s), our data strongly suggest that brain CB2 receptors functionally modulate the mesolimbic DA system and DA-related functions. Activation of brain CB2 receptors by JWH133 inhibited both the behavioral and neurochemical effects of cocaine. Given that JWH133 neither altered locomotor performance, as assessed by the rotarod test, nor produced drug rewarding or aversive effects, as assessed by intravenous self-administration and conditioned place preference, JWH133-induced inhibition of cocaine self-administration is most likely mediated by attenuation of cocaine’s rewarding efficacy secondary to the reduction in cocaine-enhanced NAc DA rather than by nonspecific locomotor impairment or malaise. We fully recognize that these findings challenge the currently accepted opinion that selective CB2 receptor agonists have no CNS effects. This opinion is largely based on previous reports that the

1164

VOLUME 14 | NUMBER 9 | SEPTEMBER 2011  nature NEUROSCIENCE

© 2011 Nature America, Inc. All rights reserved.

a r t ic l e s selective CB2 receptor agonist AM1241 neither inhibits locomotion or rotarod performance, nor produces catalepsy or ­hypothermia in rats or mice32. In addition, AM1241 also fails to alter brain functional activity, as assessed by pharmacological magnetic resonance imaging33. The ineffectiveness of AM1241 may be related to the relatively lower doses (30 µg per kg to 3.3 mg per kg) used in those studies, relatively poor selectivity and species differences in CB 2 receptor response to AM1241 (refs. 34–36). For example, AM1241 is reported to act as a full or partial agonist at human CB2 ­receptors35, whereas it acts as an inverse agonist at rodent CB 2 ­ receptors 36. Furthermore, the analgesic effects produced by AM1241 are reported to be blocked by the opioid receptor antagonist naloxone 37, suggesting that AM1241 may interact with other non-cannabinoid receptors. However, the CB 2 receptor agonist GW405833, at high doses (30–100 mg per kg), produces substantial CNS effects, such as analgesia, sedation and catalepsy 26, consistent with our finding that GW405833 (3–10 mg per kg) substantially inhibited cocaine self-administration in mice. The presence of CB2 receptors in the CNS, particularly on neurons, has been the subject of much debate10,38. Previous studies using in situ hybridization and northern blot assays failed to detect CB2 receptor mRNA in brain5,39,40. However, recent studies with more sensitive reverse transcription PCR and immunolabeling techniques have found substantial CB2 receptor expression in microglia and subpopulations of neuronal cells in brain7–11. Using highly sensitive and specific Taqman probes, we recently identified two CB 2 receptor isoforms (CB2A and CB2B) in both brain and peripheral tissues, which show substantial species differences in both structure and expression between humans, rats and mice41. It is now well accepted that CB2 receptors are expressed on microglia and a subset of neurons, with levels increasing under certain pathological conditions, such as neuroinflammation and brain injury38. There are two possible explanations for our findings. First, a low density of CB2 receptors may be expressed on mesolimbic DA neurons. As CB 2 receptors are Gi/o coupled42, activation of CB2 receptors on DA neurons in the midbrain ventral tegmental area (VTA) may directly inhibit VTA DA neurons and decrease NAc DA release, and therefore inhibit intravenous cocaine self-administration and cocaine-enhanced locomotion, as we observed. Although direct evidence of CB2 receptor expression in the mesolimbic DA neurons is currently lacking, functional CB 2 receptors are found on other neurons. For example, CB2 receptor mRNA is expressed on striatal GABAergic neurons in non-human primates43, and activation of CB2 receptors inhibits GABAergic neurotransmission in the medial entorhinal cortex of the rat19. In addition, CB2 receptors are also found on neurons in the dorsal root ganglion and spinal cord44,45, and activation of CB2 receptors on dorsal root ganglion–spinal cord neurons inhibits neuronal response to noxious stimuli45,46, thereby contributing to the antinociceptive effects of CB2 receptor agonists47. The second possibility is that activation of CB2 receptors located on microglial cells or astrocytes in the VTA and/or NAc may indirectly inhibit NAc DA release by releasing cytokines and inflammatory factors48, thereby inhibiting cocaine self-­administration and cocaine-enhanced locomotion as we observed. Whatever the mechanisms, our findings suggest for the first time, to the best of our knowledge, that activation of brain CB2 receptors inhibits cocaine’s rewarding and psychomotor-stimulating effects, which is congruent with a rapidly expanding corpus of published reports implicating brain CB2 receptors in modulating a variety of CNS functions, such as locomotion10, pain13,47, emesis11, neurogenesis14 and neuroprotection15. This finding not only challenges current views that CB2 receptors are absent from the CNS and that CB2 nature NEUROSCIENCE  VOLUME 14 | NUMBER 9 | SEPTEMBER 2011

receptor ligands lack CNS effects, but also suggests that brain CB2 receptors may be a target for the pharmacotherapy of drug abuse and addiction. Methods Methods and any associated references are available in the online ­version of the paper at http://www.nature.com/natureneuroscience/. Note: Supplementary information is available on the Nature Neuroscience website. Acknowledgments We thank Y. Shaham and E.A. Stein of the Intramural Research Program of the National Institute on Drug Abuse, and K. Mackie of Indiana University for their helpful comments on this manuscript. This research was supported by the Intramural Research Program of the National Institute on Drug Abuse. AUTHOR CONTRIBUTIONS Z.-X.X. developed the original research proposal, designed and supervised all of the experiments, analyzed all of the data and wrote the manuscript. X.-Q.P., X.L. and R.S. conducted the cocaine self-administration experiments. X.L., G.-H.B. and H.-Y.Z. conducted the in vivo microdialysis experiments. X.L., H.-J.Y., R.S. and J.L. conducted the locomotor behavioral experiments. X.-Q.P., R.S. and H.-J.Y. conducted the conditioned place preference/aversion experiments. Q.-R.L. contributed to the original research proposal. E.L.G. contributed to the original idea of this work and was responsible for overall supervision of the research and for revisions and modifications to the manuscript. COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests. Published online at http://www.nature.com/natureneuroscience/. Reprints and permissions information is available online at http://www.nature.com/ reprints/index.html. 1. Parolaro, D. & Rubino, T. The role of the endogenous cannabinoid system in drug addiction. Drug News Perspect. 21, 149–157 (2008). 2. Mackie, K. Cannabinoid receptors: where they are and what they do. J. Neuroendocrinol. 20 (suppl. 1), 10–14 (2008). 3. Glass, M., Dragunow, M. & Faull, R.L. Cannabinoid receptors in the human brain: a detailed anatomical and quantitative autoradiographic study in the fetal, neonatal and adult human brain. Neuroscience 77, 299–318 (1997). 4. Griffin, G., Tao, Q. & Abood, M.E. Cloning and pharmacological characterization of the rat CB2 cannabinoid receptor. J. Pharmacol. Exp. Ther. 292, 886–894 (2000). 5. Munro, S., Thomas, K.L. & Abu-Shaar, M. Molecular characterization of a peripheral receptor for cannabinoids. Nature 365, 61–65 (1993). 6. Malan, T.P. Jr. et al. CB2 cannabinoid receptor agonists: pain relief without psychoactive effects? Curr. Opin. Pharmacol. 3, 62–67 (2003). 7. Stella, N. Cannabinoid and cannabinoid-like receptors in microglia, astrocytes, and astrocytomas. Glia 58, 1017–1030 (2010). 8. Baek, J.H., Zheng, Y., Darlington, C.L. & Smith, P.F. Cannabinoid CB2 receptor expression in the rat brainstem cochlear and vestibular nuclei. Acta Otolaryngol. (Stockh.) 128, 1–7 (2008). 9. Gong, J.-P. Cannabinoid CB2 receptors: immunohistochemical localization in rat brain. Brain Res. 1071, 10–23 (2006). 10. Onaivi, E.S. et al. Discovery of the presence and functional expression of cannabinoid CB2 receptors in brain. Ann. NY Acad. Sci. 1074, 514–536 (2006). 11. Van Sickle, M.D. et al. Identification and functional characterization of brainstem cannabinoid CB2 receptors. Science 310, 329–332 (2005). 12. Guindon, J. & Hohmann, A.G. Cannabinoid CB2 receptors: a therapeutic target for the treatment of inflammatory and neuropathic pain. Br. J. Pharmacol. 153, 319–334 (2008). 13. Jhaveri, M.D. et al. Evidence for a novel functional role of cannabinoid CB2 receptors in the thalamus of neuropathic rats. Eur. J. Neurosci. 27, 1722–1730 (2008). 14. Goncalves, M.B. et al. A diacylglycerol lipase-CB2 cannabinoid pathway regulates adult subventricular zone neurogenesis in an age-dependent manner. Mol. Cell. Neurosci. 38, 526–536 (2008). 15. Viscomi, M.T. et al. Selective CB2 receptor agonism protects central neurons from remote axotomy-induced apoptosis through the PI3K/Akt pathway. J. Neurosci. 29, 4564–4570 (2009). 16. Sagredo, O. et al. Cannabinoid CB2 receptor agonists protect the striatum against malonate toxicity: relevance for Huntington’s disease. Glia 57, 1154–1167 (2009). 17. Elmes, S.J.R., Jhaveri, M.D., Smart, D., Kendall, D.A. & Chapman, V. Cannabinoid CB2 receptor activation inhibits mechanically evoked responses of wide dynamic range dorsal horn neurons in naive rats and in rat models of inflammatory and neuropathic pain. Eur. J. Neurosci. 20, 2311–2320 (2004). 18. Sagar, D.R. et al. Inhibitory effects of CB1 and CB2 receptor agonists on responses of DRG neurons and dorsal horn neurons in neuropathic rats. Eur. J. Neurosci. 22, 371–379 (2005).

1165

© 2011 Nature America, Inc. All rights reserved.

a r t ic l e s 19. Morgan, N.H., Stanford, I.M. & Woodhall, G.L. Functional CB2 type cannabinoid receptors at CNS synapses. Neuropharmacology 57, 356–368 (2009). 20. Ashton, J.C., Wright, J.L., McPartland, J.M. & Tyndall, J.D.A. Cannabinoid CB1 and CB2 receptor ligand specificity and the development of CB2-selective agonists. Curr. Med. Chem. 15, 1428–1443 (2008). 21. Huffman, J.W. CB2 receptor ligands. Mini Rev. Med. Chem. 5, 641–649 (2005). 22. Richardson, N.R. & Roberts, D.C.S. Progressive ratio schedules in drug selfadministration studies in rats: a method to evaluate reinforcing efficacy. J. Neurosci. Methods 66, 1–11 (1996). 23. Soria, G. et al. Lack of CB1 cannabinoid receptor impairs cocaine self-administration. Neuropsychopharmacology 30, 1670–1680 (2005). 24. Li, X. et al. Attenuation of basal and cocaine-enhanced locomotion and nucleus accumbens dopamine in cannabinoid CB1 receptor knockout mice. Psychopharmacology (Berl.) 204, 1–11 (2009). 25. Thakur, G.A., Nikas, S.P. & Makriyannis, A. CB1 cannabinoid receptor ligands. Mini Rev. Med. Chem. 5, 631–640 (2005). 26. Valenzano, K.J. et al. Pharmacological and pharmacokinetic characterization of the cannabinoid receptor 2 agonist, GW405833, utilizing rodent models of acute and chronic pain, anxiety, ataxia and catalepsy. Neuropharmacology 48, 658–672 (2005). 27. Xi, Z.-X. Cannabinoid CB1 receptor antagonists attenuate cocaine’s rewarding effects: experiments with self-administration and brain-stimulation reward in rats. Neuropsychopharmacology 33, 1735–1745 (2008). 28. Costantino, H.R., Illum, L., Brandt, G., Johnson, P.H. & Quay, S.C. Intranasal delivery: physicochemical and therapeutic aspects. Int. J. Pharm. 337, 1–24 (2007). 29. Wise, R.A. Dopamine, learning and motivation. Nat. Rev. Neurosci. 5, 483–494 (2004). 30. Schwarting, R.K.W. & Huston, J.P. Behavioral and neurochemical dynamics of neurotoxic meso-striatal dopamine lesions. Neurotoxicology 18, 689–708 (1997). 31. Bardo, M.T. Neuropharmacological mechanisms of drug reward: beyond dopamine in the nucleus accumbens. Crit. Rev. Neurobiol. 12, 37–67 (1998). 32. Malan, T.P. Jr. et al. CB2 cannabinoid receptor–mediated peripheral antinociception. Pain 93, 239–245 (2001). 33. Chin, C.-L. et al. Differential effects of cannabinoid receptor agonists on regional brain activity using pharmacological MRI. Br. J. Pharmacol. 153, 367–379 (2008). 34. Mukherjee, S. et al. Species comparison and pharmacological characterization of rat and human CB2 cannabinoid receptors. Eur. J. Pharmacol. 505, 1–9 (2004).

35. Yao, B.B. et al. In vitro pharmacological characterization of AM1241: a protean agonist at the cannabinoid CB2 receptor? Br. J. Pharmacol. 149, 145–154 (2006). 36. Bingham, B. et al. Species-specific in vitro pharmacological effects of the cannabinoid receptor 2 (CB2) selective ligand AM1241 and its resolved enantiomers. Br. J. Pharmacol. 151, 1061–1070 (2007). 37. Ibrahim, M.M. et al. CB2 cannabinoid receptor activation produces antinociception by stimulating peripheral release of endogenous opioids. Proc. Natl. Acad. Sci. USA 102, 3093–3098 (2005). 38. Atwood, B.K. & Mackie, K. CB2: a cannabinoid receptor with an identity crisis. Br. J. Pharmacol. 160, 467–479 (2010). 39. Brown, S.M., Wager-Miller, J. & Mackie, K. Cloning and molecular characterization of the rat CB2 cannabinoid receptor. Biochim. Biophys. Acta 1576, 255–264 (2002). 40. Schatz, A.R., Lee, M., Condie, R.B., Pulaski, J.T. & Kaminski, N.E. Cannabinoid receptors CB1 and CB2: a characterization of expression and adenylate cyclase modulation within the immune system. Toxicol. Appl. Pharmacol. 142, 278–287 (1997). 41. Liu, Q.-R. et al. Species differences in cannabinoid receptor 2 (CNR2 gene): identification of novel human and rodent CB2 isoforms, differential tissue expression and regulation by cannabinoid receptor ligands. Genes Brain Behav. 8, 519–530 (2009). 42. Bayewitch, M. et al. The peripheral cannabinoid receptor: adenylate cyclase inhibition and G protein coupling. FEBS Lett. 375, 143–147 (1995). 43. Lanciego, J.L. et al. Expression of the mRNA coding the cannabinoid receptor 2 in the pallidal complex of Macaca fascicularis. J. Psychopharmacol. 25, 97–104 (2010). 44. Zhang, J. et al. Induction of CB2 receptor expression in the rat spinal cord of neuropathic but not inflammatory chronic pain models. Eur. J. Neurosci. 17, 2750–2754 (2003). 45. Anand, U. et al. Cannabinoid receptor CB2 localization and agonist-mediated inhibition of capsaicin responses in human sensory neurons. Pain 138, 667–680 (2008). 46. Ross, R.A. et al. Actions of cannabinoid receptor ligands on rat cultured sensory neurones: implications for antinociception. Neuropharmacology 40, 221–232 (2001). 47. Beltramo, M. et al. CB2 receptor-mediated antihyperalgesia: possible direct involvement of neural mechanisms. Eur. J. Neurosci. 23, 1530–1538 (2006). 48. Haydon, P.G., Blendy, J., Moss, S.J. & Rob Jackson, F. Astrocytic control of synaptic transmission and plasticity: a target for drugs of abuse? Neuropharmacology 56 (suppl. 1), 83–90 (2009).

1166

VOLUME 14 | NUMBER 9 | SEPTEMBER 2011  nature NEUROSCIENCE

ONLINE METHODS

© 2011 Nature America, Inc. All rights reserved.

Animals. Male wild-type and CB1−/− mice with C57BL/6J genetic backgrounds were bred at the National Institute on Drug Abuse from three CB1+/− ­breeding pairs49 that were generously donated by A. Zimmer (National Institute of Mental Health). CB2−/ − mice with C57BL/6J genetic backgrounds were bred at the National Institute on Drug Abuse from three CB2+/− breeding pairs50 that were generously donated by G. Kunos (National Institute on Alcohol Abuse and Alcoholism). Genotyping was performed by Charles River Laboratories before the experiments. About 50% of used mice were re-genotyped in our own laboratory for genotype confirmation after completion of experiments. All of the animals were matched for age (8–14 weeks) and weight (25–35 g). They were housed individually in a climate-controlled animal colony room on a reversed light-dark cycle (lights on at 7:00 p.m., lights off at 7:00 a.m.) with free access to food and water. All experimental procedures were conducted in accordance with the Guide for the Care and Use of Laboratory Animals of the US National Research Council and were approved by the Animal Care and Use Committee of the National Institute on Drug Abuse of the US National Institutes of Health. Cocaine self-administration. Mice were prepared for experimentation by surgical catheterization of the right external jugular vein. Catheterization was performed under an anesthetic mixture of ketamine (60 mg per kg, intraperitoneal) and xylazine (12 mg per kg, intraperitoneal) using aseptic surgical technique. A 6.0 cm length of MicroRenathane tubing (inner diameter, 0.012 inches; outer diameter, 0.025 inches; Braintree Scientific) was inserted 1.2 cm into the right jugular vein and anchored to a 24-gauge steel cannula (Plastics One) that was bent at a right angle and mounted to the skull with cyanoacrylate glue and dental acrylic. A 2.5-cm extension of flexible tubing was connected to the distal end of the cannula. For intracranial microinjections, two guide cannula (MAB 4.15.IC, SciPro) were surgically implanted into the NAc (anterior-posterior, 1.44 mm; medial-lateral, ±1.5 mm; dorsal-ventral, –3.8 mm; from bregma, with an angle of 8° from vertical) in separate groups of mice. To keep the implanted catheters patent, they were flushed daily with 0.05 ml of saline solution containing 20 IU ml−1 heparin and 0.33 mg ml−1 gentamycin. To avoid cocaine overdose, each animal was limited to a maximum of 30 cocaine injections per 3 h session. Intravenous cocaine self-administration experiments were conducted in operant response test chambers (Model ENV-307A, Med Associates). Each test chamber had two levers located 2.5 cm above the floor, one active and one inactive. A cue light and a speaker were located 5 cm above the active lever. A house chamber light was on during each 3-h session. Depression of the active lever activated the infusion pump; depression of the inactive lever was counted, but had no consequence. Each drug infusion was paired with a conditioned cue light and a cue sound (tone). After recovery from surgery, each mouse was placed into a test chamber and allowed to lever-press for intravenous cocaine (1 mg per kg per infusion) delivered in 0.015 ml over 4.2 s, on an FR1 reinforcement schedule. During the initial 3–5 d, all animals received five free cocaine infusions in a 10-min time interval to prime the animal for drug-seeking and drug-taking behavior. These five free drug infusions were subtracted from the total number of drug infusions for data analysis. During the 4.2-s injection period, additional responses on the active lever were recorded, but did not lead to additional infusions. Each session lasted 3 h. After 1–2 weeks of cocaine self-administration, the cocaine dose was switched from 1 mg per kg per infusion to 0.5 mg per kg per infusion for an additional 1–2 weeks of cocaine self-administration until stable day-to-day self-administration was established, which was defined as ≥20 cocaine infusions per session with a steady self-administration pattern for at least 3 consecutive days. Subjects randomly received one dose of JWH133 (10 or 20 mg per kg, intraperitoneal), GW405833 (3 or 10 mg per kg, intraperitoneal), AM630 (10 mg per kg, intraperitoneal) or vehicle (Tocrisolve 100) 30 min before cocaine selfadministration. After each test, animals received an additional 5–7 d of cocaine self-administration until baseline response rate was re-established before testing the next dose of drug. To determine whether the effect of JWH133 on cocaine self-administration was induced by activation of brain or peripheral CB2 receptors, we tested the effects of intranasal (50, 100 µg/10 µl/nostril) or intra-NAc (0.3, 1 and 3 µg per µl per side) microinjection of JWH133 on cocaine self-administration. Intranasal drug administration was performed under inhalant isoflurane anesthesia using the Fluovac System (Harvard Apparatus). To determine whether the effects of

doi:10.1038/nn.2874

intranasal JWH133 on cocaine self-administration might be mediated by absorption into the nasal vasculature with subsequent venous delivery to peripheral sites of action, the same micro-quantities of JWH133 as were administered intranasally (100 and 200 µg) were injected intravenously in a separate experimental session via the implanted jugular catheter 30 min before cocaine self-administration. Additional groups of animals were initially trained under FR1 reinforcement as outlined above. After stable cocaine self-administration under FR1 reinforcement was established, animals were switched to progressive ratio reinforcement, under which the work requirement (lever presses) to receive a cocaine infusion was progressively raised within each test session22. The progressive ratio breakpoint was defined as the maximal number of lever presses completed for the last cocaine infusion before a 1 h period during which no infusions were obtained by the animal. Animals were allowed to continue daily sessions of cocaine selfadministration under progressive ratio reinforcement conditions until day-today variability in break-point fell within 1-2 ratio increments for 3 consecutive days. Once a stable break-point was established, subjects randomly received one dose of JWH133 (10 or 20 mg/kg, intraperitoneal), vehicle (Tocrisolve 100), or AM251 (3 mg/kg, intraperitoneal) 30 min before progressive ratio cocaine selfadministration testing. After stable cocaine self-administration under FR1 reinforcement was established for at least 3 consecutive days, the animals were divided into two groups and cocaine was replaced by JWH133 (1 mg per kg per infusion) or by vehicle (Tocrisolve-100) for 5 d. Given that animals might take several days to acquire self-administration for a novel reinforcer, each replacement test was repeated for 5 d. Conditioned place preference or aversion. Four groups of wild-type mice (n = 12 each group) were used to study cocaine-induced (10 or 20 mg per kg) or JWH133-induced (10 or 20 mg per kg) conditioned place preference or aversion. A three-chamber place preference apparatus (Med Associates) was used in this study. The apparatus consisted of two large compartments (16.8 × 12.7 × 12.7 cm) and one small compartment (7.2 × 12.7 × 12.7 cm), which separated the large compartments. The two large compartments had different visual and tactile cues. One compartment was black with a stainless steel rod floor. The other compartment was white with a stainless steel mesh floor. The small compartment was gray with a smooth polyvinyl chloride floor. The apparatus had a clear Plexiglas top with a light on it. During the preconditioning phase (days 1–2), mice were placed in the small compartment and were allowed to freely explore the three compartments for 15 min daily. The time spent in each compartment was recorded. We used an unbiased conditioned place preference procedure. Mice spending over 500 s in either compartment were excluded. The next 10 d (days 3–12) constituted the conditioning phase, with one session per day. Each mouse received an intraperitoneal injection of the same dose of test drug (cocaine or JWH133) on days 3, 5, 7, 9 and 11, and was then confined in one large ­compartment for 30 min. On days 4, 6, 8, 10 and 12, each mouse received an intraperitoneal injection of vehicle (saline or Tocrisolve-100) and was then confined in the other large compartment for 30 min. Test drug was paired with either white or black compartment in a counterbalanced manner. On the test day (24 h after the last injection), mice were allowed to freely explore the three ­compartments for 15 min without injections, and the time spent in each ­compartment was recorded. All behavioral testing was performed during the light phase of the light/dark cycle. Locomotor activity. Before drug administration, each animal was placed in a locomotor detection chamber (Accuscan Instruments) for 3 d (4 h per d) for environmental habituation. On each test day, mice were placed in the chamber for 1 h of habituation, and then removed and given either saline or one dose (10 or 20 mg per kg intraperitoneal, 30 min before cocaine) of JWH133. Animals were then placed back into the locomotor chambers to observe the effects of JWH133 alone on locomotion for 3 h. Additional groups of animals were used to study the effects of JWH133 pretreatment on cocaine-enhanced locomotion. Cocaine (10 mg per kg, intraperitoneal) was given 30 min after JWH133 administration. The mice were then placed back into the locomotor chambers to monitor locomotor activity for 3 h. To determine whether intracranial microinjections of CB2 ligands produce similar effects as systemic administration, two separate groups of mice with intracranial guide cannula implantation received intra-NAc local administration of JWH133 (1 or 3 µg per side) or AM630 (1, 3 or 10 µg per side),

nature NEUROSCIENCE

a mixture of JWH133 and AM630 (3 µg each per side), or vehicle (Tocrisolve100). After each test, animals received an additional 3–5 d of locomotor habituation before testing the next dose of drug. The order of testing for the various doses of the drugs was counterbalanced. Data were collected in 10-min intervals using the VersaMax data analysis system (Accuscan Instruments). Total distance was used to evaluate the effects of different treatments on locomotor behavior.

© 2011 Nature America, Inc. All rights reserved.

Rotarod performance. A four-station mouse rotarod device (AccuScan Instruments) was used to study the effects of JWH133 on operant locomotor performance. The speed of rotation of the rotarod was increased from 2.5 to 20 rpm over 2 min and the time the animal remained on the rod was determined as the mean of three trials. After 5–7 d of habituation and training on the rotarod device, animals randomly received either vehicle or one dose (10 or 20 mg per kg, intraperitoneal) of JWH133 before rotarod testing began. After the drug injection, animals were placed on the rotarod device and their locomotor performance was assessed every 30 min for 3 h. After each test, animals then received 3–5 d of additional rotarod habituation until baseline response was re-stabilized before testing the next dose of drug. The various drug doses were given randomly and counterbalanced. In vivo microdialysis. All mice were surgically implanted with intracranial guide cannulae (MAB 4.15.IC, SciPro) into the NAc. The surgery was performed under an anesthetic mixture of ketamine hydrochloride (80 mg ml−1) and xylazine

nature NEUROSCIENCE

hydrochloride (12 mg ml−1), using standard aseptic surgical and stereotaxic technique. The stereotaxic coordinates for the NAc were +1.4 mm anterior-posterior, ±1.5 mm medial-lateral, –3.8 mm dorsal-ventral from bregma, with an angle of 8° from vertical. The guide cannulae were fixed to the skull with dental acrylic. In vivo brain microdialysis was performed after at least 7 d following surgery. The procedures for the microdialysis and the subsequent dialysate DA quantification were identical to those we have reported previously24. Drugs. Cocaine HCl (Sigma-Aldrich) was dissolved in physiological saline. JWH133, AM251 and AM630 were obtained from Tocris Bioscience. They were dissolved in Tocrisolve-100 (vehicle). Data analyses. All data are presented as mean ± s.e.m. One-way ANOVA was used to analyze the effects of JWH133 or other drugs on cocaine self-administration. Two-way ANOVA for repeated measures over time was used to analyze the effects of JWH133 on locomotion or extracellular DA. Individual group comparisons were carried out using the Student-Newman-Keuls or Bonferroni methods. 49. Zimmer, A., Zimmer, A.M., Hohmann, A.G., Herkenham, M. & Bonner, T.I. Increased mortality, hypoactivity, and hypoalgesia in cannabinoid CB1 receptor knockout mice. Proc. Natl. Acad. Sci. USA 96, 5780–5785 (1999). 50. Buckley, N.E. et al. Immunomodulation by cannabinoids is absent in mice deficient for the cannabinoid CB2 receptor. Eur. J. Pharmacol. 396, 141–149 (2000).

doi:10.1038/nn.2874