REVIEW Mechanisms of CB1 Receptor Signaling: Endocannabinoid Modulation of Synaptic Strength

Home , Cannabinoid receptor

International Journal of Obesity (2006) 30, S19–S23 & 2006 Nature Publishing Group All rights reserved 0307-0565/06 $30.00 www.nature.com/ijo REVIEW Mechanisms of CB1 receptor signaling: endocannabinoid modulation of synaptic strength

K Mackie

Department of Anesthesiology, University of Washington School of Medicine, Seattle, WA, USA

The CB1 cannabinoid receptor has attracted much recent interest because of the observation that CB1 receptor antagonists have efficacy in treating metabolic syndrome and obesity. CB1 receptors also mediate most of the psychotropic effects of D9- tetrahydrocannabinol (D9THC), the principal psychoactive component of cannabis. In addition, they are one component of an interesting and widespread paracrine signaling system, the endocannabinoid system. The endocannabinoid system is comprised of cannabinoid receptors, endogenous cannabinoids, and the metabolic pathways responsible for their synthesis and degradation. The details of the endocannabinoid system have been most thoroughly studied in the brain. Here it has been shown to be intimately involved in several forms of neuronal plasticity. That is, activation of CB1 receptors by endocannabinoids produces either short- or long-term changes in the efficacy of synaptic transmission. The behavioral consequences of these changes are many, but some of the most striking and relevant to the current symposium are those associated with endogenous reward and consumptive behavior. International Journal of Obesity (2006) 30, S19–S23. doi:10.1038/sj.ijo.0803273

Keywords: cannabinoid; neuronal plasticity; retrograde signalling; unsaturated fatty acid

Introduction presynaptic terminals. In this region of the brain CB1 receptor and the neuropeptide, cholecystokinin (CCK) Cannabinoid receptors are members of the G protein- expression is highly correlated.7,8 Generally, in the hippo- coupled receptor (GPCR) superfamily of cell surface, hepti- campus, more than 80% of the neurons expressing high helical receptors.1 Thus far, the cloning of two cannabinoid levels of CB1 also express CCK, while in cortex the receptors, CB1 and CB2, has been reported. In addition, co-expression is about 50%. Similarly, most large diameter there is strong pharmacological evidence for additional CCK containing neurons also express CB1. Although in cannabinoid receptors.2–4 CB2 receptors are primarily found many forebrain and midbrain structures CB1 receptors are on immune cells, particularly cells of macrophage lineage.1 expressed on GABAergic neurons, there is functional As these receptors only indirectly affect neuronal function, evidence for CB1 receptor expression on some glutamatergic they will not be considered further in this review. The CB1 neurons in the forebrain.9 In addition, CB1 receptors are receptor was cloned and identified as a cannabinoid receptor abundantly expressed by other glutamatergic neurons such nearly 15 years ago.5 CB1 receptors are members of the as cerebellar granule cells.1 Gi/Go-linked GPCR family. Thus, they inhibit voltage- sensitive calcium channels and adenylyl cyclase and activate inwardly rectifying potassium channels and MAP kinase.1 CB1 receptors are among the most abundant GPCRs in the CB1 receptors inhibit neurotransmitter release brain, and their levels are comparable to those of ionotropic glutamate receptors.6 This abundance comes about The strong presynaptic localization of CB1 receptors and because of the very high levels of CB1 receptor expression their inhibition of voltage-dependent calcium channels and in a relatively restricted number of neurons. In the fore- adenylyl cyclase suggest that a primary function of CB1 brain, CB1 receptors are particularly enriched on axons and receptors might be to inhibit neurotransmitter release. Indeed, this appears to be the case. Numerous studies with cultured cells, isolated organs, and brain slice preparations Correspondence: Dr K Mackie, Box 356540, Department of Anesthesiology, have found that activation of presynaptic CB1 receptors University of Washington School of Medicine, 1959 NE Pacific St, Seattle, WA 98195-6540, USA. leads to an attenuation of neurotransmission. At most E-mail: [email protected] synapses, this appears to be primarily because of inhibition Endocannabinoid modulation of synaptic strength K Mackie S20 of voltage-dependent calcium channels, although at times while MAGL is mostly found presynaptically,14 implying activation of potassium channels has been suggested.10 At distinct sites of degradation for these different families of most GABAergic synapses, there appears to be little role for endocannabinoids. direct inhibition of the vesicular release machinery by CB1 receptors. This coincides with the observation that despite high levels of CB1 expression in presynaptic terminals, they 11 Interaction of endogenous and exogenous appear to be absent from the active zone of the synapse. cannabinoids

The presence of a cannabinoid receptor that binds D9THC Endogenous cannabinoids suggests that cannabis might produce its psychoactive effects in the same way that opium extracts do. Namely, that by binding to and activating the CB1 receptor, D9THC mimics The presence of a receptor implies the existence of the actions of endogenous cannabinoids. While this idea is endogenous ligand(s). The cloning of the CB1 receptor attractive, it appears to be too simplistic. While under most intensified the search for such compounds that might mimic conditions, release of endogenous opioids is low – this does D9-tetrahydrocannabinol (D9THC). Two classes of endogen- not seem to be the case for endocannabinoids.15,16 Thus, in ous cannabinoids (endocannabinoids) have been identified ‘normal’ CNS tissue, opioids receptors are relatively unoccu- and thoroughly studied.12 The first to be identified was pied and exogenously administered opiates can bind to and anandamide. This compound is the amide of arachidonic activate them. In contrast, the continuous production acid and ethanolamine. Anandamide is one member of a of endocannabinoids results in CB1 receptors having a rather large family of fatty acid amides. Many, but not all, of greater level of occupancy and activation compared to these fatty acid amides bind CB1 receptors. The second opioids receptors. Therefore, the effects of exogenous family of endocannabinoids identified was the glycerol cannabinoids, such as D9THC, will depend on the level of esters, typified by 2-arachidonyl glycerol (2-AG). A striking CB1 occupancy, as well the nature of endocannabinoid difference between the endocannabinoids and many classic binding to the receptor. neurotransmitters is that endocannabinoids appear to be Most studies have found that D9THC and anandamide are ‘formed on demand’ rather than presynthesized and stored partial agonists with similar efficacy, while 2-AG has a in synaptic vesicles. This has significant implications for the substantially greater efficacy.17 Thus, in one way, D9THC and regulation of endocannabinoid signaling. A number of other anandamide can be thought of being roughly equivalent – fatty acid-containing compounds have also been identified activating CB1 receptors to a similar extent. In this way, as potential endocannabinoids. These include the amino- D9THC will be mimicking the effects of anandamide where acid amides, virodhamine, and noladin ether.13 These are this endocannabinoid predominates. In contrast, at synapses discussed in more detail in other reviews in this volume. where 2-AG is the major endocannabinoid, the partial Despite the relatively similar structure of the two families agonism of D9THC could antagonize 2-AG’s action, causing of endogenous cannabinoids, substantial differences distin- a decrease in CB1-mediated effects. This scheme predicts that guish the two classes. The first difference is the route of the psychoactive effects of D9THC would not necessarily be synthesis. The primary pathway of anandamide formation is fully antagonized by a CB1 receptor antagonist. Indeed, this cleavage of a N-acyl phosphatidyl enthanolamine by a appears to be the case. In a study investigating the ability of specific phospholipase D. In contrast, 2-AG is typically the CB1 receptor antagonist, rimonabant, to reverse the formed by the sequential action of phospholipase C (forming effects of D9THC, Marilyn Heustis and her colleagues found diacyl glycerol) and of an sn-1 diacyl glycerol lipase, forming that even very high doses of rimonabant (90 mg) were only 2-AG. The localization and regulation of these enzymes are modestly effective in reducing most of the subjective completely different.14 Thus, the production of anandamide components of the ‘high’ elicited by smoking of a cannabis and 2-AG might reasonably be expected to occur at distinct cigarette.18 Interestingly, the autonomic effect (tachycardia) locations and following different stimuli. The actions of elicited by cannabis was more effectively antagonized by 2-AG and anandamide at CB1 receptors are also divergent. rimonabant than the subjective effects,18 suggesting that the Most studies have found 2-AG to be a full agonist, while former might be mediated by fatty acid amides and the latter anandamide has a lower intrinsic efficacy. Thus, at synapses mediated by 2-AG. with relatively low CB1 receptor densities, 2-AG will act as a full agonist while anandamide might only be a partial agonist. Finally, the catabolic pathways for these two endocannabinoids are distinct. Anandamide and related Endocannabinoid release mediates short-term fatty acid amides are primarily degraded by fatty acid amino plasticity at specific synapses hydrolase (FAAH), while 2-AG appears to be primarily degraded by a monoacyl glycerol lipase (MAGL). Fatty acid The observation that endocannabinoids are produced fol- amino hydrolase tends to be post-synaptically localized lowing increases in intracellular calcium or increases in

International Journal of Obesity Endocannabinoid modulation of synaptic strength K Mackie S21 cAMP suggested that they might be produced during periods the depolarized neuron. This is because a number of of intense CNS activity. This appears to be the case. experiments have shown that the retrograde transmitter Endocannabinoids mediate a process called DSI (i.e. depolar- released following depolarization acts only over a distance of ization-induced suppression of inhibition). Depolarization- 10 mm, or so. However, if the soma of a neuron adjacent to induced suppression of inhibition occurs after a neuron, for the depolarized neuron expresses CB1 receptors, activation example a hippocampal pyramidal cell or cerebellar Purkinje of these receptors by endocannabinoids can open inwardly neuron, is strongly depolarized. The depolarization robustly rectifying potassium channels, decreasing the spontaneous attenuates the inhibitory input onto that neuron. A similar firing of the adjacent neuron. This will potentially decrease phenomenon occurs with some excitatory inputs; the neurotransmitter release over a wide area, limited only by process here is referred to as DSE (depolarization-induced the extent of the adjacent neuron’s axonal network. This sort suppression of excitation). Both DSI and DSE were thought of widespread influence of local endocannabinoid release to involve a retrograde messenger as depolarization of the has been found in both cerebellum and cortex.24,25 A final post-synaptic cell was required, but the phenomenon was consideration is that stimuli other than step depolarizations manifested by a decrease in neurotransmitter release from can produce DSI and DSE. For example, near physiological the presynaptic terminal – thus a signal passed from the post- firing patterns in cerebellum cause DSE and muscarinic and synaptic to the presynaptic cell. In a pair of studies published metabotropic glutamate receptor activation can cause DSI in in 2001, Rachel Wilson and Roger Nicoll elegantly demon- the hippocampus. Taking these considerations into account, strated that endocannabinoids are involved in hippocampal it is clear that DSI and DSE are mechanisms that allow a quite DSI.19,20 Endocannabinoids were soon implicated in hippo- flexible degree of control of short-term, but not necessarily campal DSE and cerebellar DSE. These initial studies were local, neuronal plasticity. followed by numerous other studies, all with the general conclusion that if CB1 receptors are expressed presynaptically to a neuron, then DSI or DSE could usually be evoked at that synapse.12,21 Of particular relevance to consumptive Endocannabinoid involvement in long-term behavior is that leptin inhibits DSI in lateral hypothalamic plasticity slices; furthermore, DSI was prolonged in leptin-deficient, ob/ob mice.22 The above discussion highlighted the involvement of Endocannabinoid involvement in DSI and DSE has gained endocannabinoids in a form of short-term neuronal plasti- widespread acceptance. However, there are several features of city. More recently, convincing evidence has emerged the process and experimental details that need to be implicating endocannabinoids in a form of long-term considered. The first is that endocannabinoids are widely plasticity, termed long-term depression, or LTD.26 Long-term believed to be the retrograde messenger involved in DSI and depression typically results following sustained low-fre- DSE. Yet, as of now, there is no direct demonstration that quency stimulation of a neuronal pathway and can have endocannabinoids made post-synaptically diffuse presynap- many causes. Endocannabinoids have been found to be tically to cause DSI or DSE. It is conceivable that the post- involved in LTD of two excitatory pathways, from cortex to synaptic cell makes an, as yet unknown, retrograde messen- the dorsal striatum and from prefrontal cortex to the nucleus ger that diffuses presynaptically eliciting endocannabinoid accumbens and two inhibitory circuits, one in the hippo- production. The second is that glial cells, particularly campus and the other in the amygdala.26 Given the role astrocytes and microglial cells, but likely oliogodendrocytes of some of these pathways (e.g., nucleus accumbens as well, are prolific producers of endocannabinoids.23 Thus, and amygdala) in reward and hedonistic behaviors, it is the involvement of glia in these schemes needs to be not surprising that a growing literature supports the notion considered. The third is that CB1 activation not only inhibits that endocannabinoids are intimately involved in these GABA and glutamate release but also inhibits release of behaviors.27 neuropeptides from CB1 receptor-containing nerve term- The hallmark of LTD is that the strength of neurotransmis- inals. In the case of forebrain CB1 receptors, this would often sion decreases following a particular pattern of synaptic mean the inhibition of CCK release.11 Since CCK is generally activity, typically prolonged low-frequency stimulation. That excitatory in nature, decreasing its release will complicate is, the quantity of neurotransmitter released is attenuated the perhaps overly simplistic interpretation that activation following the LTD stimulus. This phenomenon has been best of CB1 receptors in DSI solely decreases inhibition. Rather studied in the nucleus accumbens and hippocampus and a the net effect will be a balance of the decreased inhibition fairly complete description of the mechanism is possible. because of less activation of GABA A receptors (GABA B Prolonged low-frequency stimulation of excitatory inputs receptors do not seem to be involved, at least in hippocam- leads to a substantial release of glutamate. These act at post- pal DSI) and decreased excitation because of less activation synaptic group I metabotropic glutamate receptors causing of CCK receptors. A fourth consideration is the spatial the synthesis of endocannabinoids. The endocannabinoids influence of DSI and DSE. As described above, DSI and DSE (likely 2-AG) then diffuse retrogradely, activating presynap- will diminish release only from nerve terminals adjacent to tic CB1 receptors. Through a mechanism that is not well

International Journal of Obesity Endocannabinoid modulation of synaptic strength K Mackie S22 understood, sustained activation of these presynaptic CB1 4 Hajos N, Freund TF. Distinct cannabinoid sensitive receptors receptors triggers a change in the terminal that leads to a regulate hippocampal excitation and inhibition. Chem Phys Lipids long-term inhibition of neurotransmitter release. Activation 2002; 121: 73–82. 5 Matsuda LA, Lolait SJ, Brownstein MJ, Young AC, Bonner TI. of cannabinoid receptors is only needed for the initiation of Structure of a cannabinoid receptor and functional expression of LTD. Once established, CB1 receptor antagonists do not the cloned cDNA. Nature 1990; 346: 561–564. block LTD. Note that although prolonged low-frequency 6 Herkenham M, Lynn AB, Johnson MR, Melvin LS, de Costa BR, stimulation of excitatory neurons is needed for LTD to Rice KC. Characterization and localization of cannabinoid receptors in rat brain: a quantitative in vitro autoradiographic occur, the synapses undergoing LTD are those that have CB1 study. J Neurosci 1991; 11: 563–583. receptors. These may or may not be glutamatergic. For 7 Katona I, Sperlagh B, Sik A, Kafalvi A, Vizi ES, Mackie K et al. example, in nucleus accumbens, it is the persistently firing Presynaptically located CB1 cannabinoid receptors regulate glutamate terminals from prefrontal cortex that undergo GABA release from axon terminals of specific hippocampal interneurons. J Neurosci 1999; 19: 4544–4558. LTD, while in hippocampus the glutamate input is provided 8 Marsicano G, Lutz B. Expression of the cannabinoid receptor CB1 by the Schaffer collaterals, but it is the terminals of the CCK- in distinct neuronal subpopulations in the adult mouse forebrain. positive GABAergic basket cells that undergo LTD. Another Eur J Neurosci 1999; 11: 4213–4225. form of LTD sometimes involving endocannabinoids is a 9 Marsicano G, Goodenough S, Monory K, Hermann H, Eder M, 28 Cannich A et al. CB1 cannabinoid receptors and on-demand specific form of (spike) timing-dependent LTD. This type of defense against excitotoxicity. Science 2003; 302: 84–88. LTD occurs if a postsynaptic cell depolarizes (at times a sub- 10 Hoffman AF, Lupica CR. Mechanisms of cannabinoid inhibition threshold depolarization is sufficient) just prior to a synaptic of GABA(A) synaptic transmission in the hippocampus. J Neurosci terminal release of glutamate release in a repetitive fashion. 2000; 20: 2470–2479. This form of LTD involves endocannabinoids, and is distinct 11 Hajos N, Katona I, Naiem SS, MacKie K, Ledent C, Mody I et al. Cannabinoids inhibit hippocampal GABAergic trans- from the previously discussed form of LTD in that presynap- mission and network oscillations. Eur J Neurosci 2000; 12: tic NMDA glutamate receptors are required. 3239–3249. 12 Freund TF, Katona I, Piomelli D. Role of endogenous cannabinoids in synaptic signaling. Physiol Rev 2003; 83: 1017–1066. 13 Walker JM, Krey JF, Chu CJ, Huang SM. Endocannabinoids and Summary related fatty acid derivatives in pain modulation. Chem Phys Lipids 2002; 121: 159–172. 14 Gulyas AI, Cravatt BF, Bracey MH, Dinh TP, Piomelli D, Boscia F Ample evidence has accumulated over the past decade that et al. Segregation of two endocannabinoid-hydrolyzing endocannabinoids serve as mediators of short- and long- enzymes into pre- and postsynaptic compartments in the rat term neuronal plasticity. Thus, in addition to being the hippocampus, cerebellum and amygdala. Eur J Neurosci 2004; 20: 9 441–458. target for D THC, the principal psychoactive component of 15 Walker JM, Huang SM, Strangman NM, Tsou K, Sanudo-Pena MC. cannabis, CB1 cannabinoid receptors are also central to Pain modulation by release of the endogenous cannabinoid important normal behavioral pathways. The behaviors anandamide. Proc Natl Acad Sci USA 1999; 96: 12198–12203. involved are diverse and include: movement, sensory 16 Giuffrida A, Parsons LH, Kerr TM, Rodriguez de Fonseca F, Navarro M, Piomelli D. Dopamine activation of endogenous learning, analgesia, anxiety, and appetitive behaviors, to cannabinoid signaling in dorsal striatum. Nat Neurosci 1999; 2: name a few. At the current time, the most relevant of these 358–363. for therapeutic manipulation are obesity (with involvement 17 Luk T, Jin W, Zvonok A, Lu D, Lin XZ, Chavkin C et al. of both central and peripheral mechanisms – alternatively Identification of a potent and highly efficacious, yet slowly desensitizing CB1 cannabinoid receptor agonist. Br J Pharmacol homeostatic and hedonic mechanisms) and craving-based 2004; 142: 495–500. disorders, such as alcohol and tobacco dependency. As our 18 Huestis MA, Gorelick DA, Heishman SJ, Preston KL, Nelson RA, understanding of the basic biology of the endocannabinoid Moolchan ET et al. Blockade of effects of smoked marijuana by system develops, it is likely that additional therapeutically the CB1-selective cannabinoid receptor antagonist SR141716. Arch Gen Psychiatry 2001; 58: 322–328. beneficial manipulations of this system will be developed, as 19 Wilson RI, Kunos G, Nicoll RA. Presynaptic specificity of we have seen with the CB1 receptor antagonist, rimonabant. endocannabinoid signaling in the hippocampus. Neuron 2001; 31: 453–462. 20 Wilson RI, Nicoll RA. Endogenous cannabinoids mediate retrograde signalling at hippocampal synapses. Nature 2001; References 410: 588–592. 21 Alger BE. Retrograde signaling in the regulation of synaptic 1 Howlett AC, Barth F, Bonner TI, Cabral G, Casellas P, Devane WA transmission: focus on endocannabinoids. Prog Neurobiol 2002; et al. International Union of Pharmacology. XXVII. Classification 68: 247–286. of cannabinoid receptors. Pharmacol Rev 2002; 54: 161–202. 22 Jo YH, Chen YJ, Chua Jr SC, Talmage DA, Role LW. Integration of 2 Jarai Z, Wagner JA, Varga K, Lake KD, Compton DR, Martin BR endocannabinoid and leptin signalling in an appetite-related et al. Cannabinoid-induced mesenteric vasodilation through an neural circuit. Neuron 2005; 48: 1055–1066. endothelial site distinct from CB1 or CB2 receptors. Proc Natl Acad 23 Stella N. Cannabinoid signaling in glial cells. Glia 2004; 48: Sci USA 1999; 96: 14136–14141. 267–277. 3 Breivogel CS, Griffin G, Di Marzo V, Martin BR. Evidence for a 24 Bacci A, Huguenard JR, Prince DA. Long-lasting self-inhibition of new G protein-coupled cannabinoid receptor in mouse brain. Mol neocortical interneurons mediated by endocannabinoids. Nature Pharmacol 2001; 60: 155–163. 2004; 431: 312–316.

International Journal of Obesity Endocannabinoid modulation of synaptic strength K Mackie S23 25 Kreitzer AC, Carter AG, Regehr WG. Inhibition of interneuron 27 Lupica CR, Riegel AC, Hoffman AF. Marijuana and cannabinoid firing extends the spread of endocannabinoid signaling in the regulation of brain reward circuits. Br J Pharmacol 2004; 143: cerebellum. Neuron 2002; 34: 787–796. 227–234. 26 Gerdeman GL, Lovinger DM. Emerging roles for endocannabi- 28 Sjostrom PJ, Turrigiano GG, Nelson SB. Neocortical LTD via noids in long-term synaptic plasticity. Br J Pharmacol 2003; 140: coincident activation of presynaptic NMDA and cannabinoid 781–789. receptors. Neuron 2003; 39: 641–654.

International Journal of Obesity