I. J. Lever • A. S. C. Rice (✉)
Pain Research Group, Department of Anaesthetics, Intensive Care and Pain Medicine, Imperial College London, Chelsea and Westminster Hospital Campus, 369 Fulham Road, London SW10 9NH, UK
Abstract Convincing evidence from preclinical studies demonstrates that cannabinoids can reduce pain responses in a range of inﬂammatory and neuropathic pain models. The anatom- ical and functional data reveal cannabinoid receptor-mediated analgesic actions operating at sites concerned with the transmission and processing of nociceptive signals in brain, spinal cord and the periphery. The precise signalling mechanisms by which cannabinoids produce analgesic effects at these sites remain unclear; however, signiﬁcant clues point to cannabinoid modulation of the functions of neurone and immune cells that mediate nociceptive and inﬂammatory responses. Intracellular signalling mechanisms engaged by cannabinoid receptors—like the inhibition of calcium transients and adenylate cyclase, and
pre-synaptic modulation of transmitter release—have been demonstrated in some of these cell types and are predicted to play a role in the analgesic effects of cannabinoids. In contrast, the clinical effectiveness of cannabinoids as analgesics is less clear. Progress in this area requires the development of cannabinoids with a more favourable therapeutic index than those currently available for human use, and the testing of their efﬁcacy and side-effects in high-quality clinical trials.
Keywords CB1 · CB2 · Endocannabinoid · Spinal cord · Sensory neurone · G protein-coupled receptor
The CB Signalling System
Although the psychotropic and therapeutic properties of the marijuana plant (Cannabis sativa) have been documented for thousands of years, it was not until 1964 that Gaoni and Mechoulam ﬁrst identiﬁed Δ9 tetrahydrocannabinol
(Δ9THC) as a major psychoactive constituent of C. sativa and elucidated its
structure. Since then, more than 60 bioactive components have been identiﬁed
(Howlett et al. 2002; Mechoulam 2000). These, including synthetic molecules and endogenous compounds derived from animal tissues, are collectively known as the cannabinoids (CBs). Two membrane receptors for CBs have been identiﬁed and cloned, events which preceded the discovery of several endogenous CB ligands that bind to these receptors. CB pharmacology is still in the process of identifying all the signalling pathways that mediate the cellu- lar actions of CBs and the processes governing production and degradation of endogenous CB ligands, which together make up the endocannabinoid system.
This chapter will outline our current knowledge of the endocannabinoid signalling system. It will then focus on evidence for the analgesic properties of CBs; initially from the perspective of animal models and the insights that such
research has provided into possible analgesic mechanisms for CBs operating at different sites along the pain signalling pathway, from the peripheral nervous system to the brain. It will then discuss evidence from the clinical use of CBs,
which leads to perspectives on the likely therapeutic usefulness of CBs as analgesics.
The ﬁrst evidence for the existence of CB-responsive receptors came from radioligand binding studies that reported the presence of saturable, stereos- elective, high-afﬁnity CB binding sites in mammalian brain (Devane et al.
1988). Subsequently, the ﬁrst CB receptor subtype, CB1, was cloned from a rat complementary DNA (cDNA) library (Matsuda et al. 1990). It is 473 amino acids in length and hasa molecular weight of 53 kDa, although variants of 59
and 64 kDa also exist. Cloning of human CB1 (hCB1) (Gerard et al. 1991) and mouse (Chakrabarti et al. 1995) homologues that share close sequence homol- ogy (>97%) with rat CB1 followed. Two splice variants of human CB1 (CB1a and CB1b) have been identiﬁed. Both have truncated amino terminals, confer- ring altered ligand binding and activation properties, but have relatively low tissue abundance compared to full-length CB1 (Shire et al. 1995; Ryberg et al.
2005). The second CB receptor, CB2, was ﬁrst identiﬁed in a human promye-
locytic leukaemia cell line (Munro et al. 1993). Subsequently, the murine and
rat homologues of CB2 have been cloned, revealing 82% (Shire et al. 1996) and 81% (Brown et al. 2002; Grifﬁn et al. 2000) sequence homology with hCB2 respectively and 90% homology between them. CB1 and CB2 receptors are both seven trans-membrane domain, G protein-coupled receptors (GPCRs) with C and N terminals. The distinction between them is ﬁrstly based on their predicted amino acid sequences: hCB2 is shorter than hCB1 (360 amino acids, 40 kDa) with only 44% sequence homology (rising to 68% in the trans- membrane regions) (Munro et al. 1993). The two receptor subtypes are also distinguished by their signalling mechanisms (Sect. 1.2) and tissue distribu- tion (Sect. 1.3). CB1 (Ledent et al. 1999; Zimmer et al. 1999), CB2 (Buckley et al.
2000) and CB1/CB2 (Járai et al. 1999) knock-out mice have been created and
Residual pharmacological activity of CBs in CB receptor knock-out mice or following the administration of receptor antagonists in rodents, has sug- gested the existence of additional CB receptors (Begg et al. 2005; Howlett et al.
2002). Bioassays from mesenteric artery preparations (Járai et al. 1999) and binding studies in CNS tissue (Breivogel et al. 2001) have uncovered resid- ual effects of both endogenous and synthetic CBs. These are thought to be mediated by several pharmacologically distinct non-CB1/CB2 GPCR systems. One is known to be expressed on blood vessel endothelium and is respon- sive to the endocannabinoid anandamide (AEA) (Sect. 126.96.36.199) and abnormal cannabidiol (Abn-CBD), a synthetic form of the phytocannabinoid cannabid- iol (Járai et al. 1999; Offertaler et al. 2003). Another is distinctively responsive to the vanilloid capsaicin, as well as to the synthetic CBs WIN55,212-2 and CP55,940 (Sect. 1.4; Breivogel et al. 2001; Hájos and Freund 2002). In addition, the anti-inﬂammatory effects of the endogenous cannabimimetic compound palmitoylethanolamide (PEA) are not likely to be explained by its binding to either CB1 or CB2 receptors (Grifﬁn et al. 2000). Instead, research in mice suggests these effects may be mediated by activation of the nuclear receptor
peroxisome proliferator-activated receptor-α (PPARα) (Lo et al. 2005). CBs are also reported to have agonist activity at two of the transient receptor poten- tial (TRP) type of ligand-gated cation channels that are expressed by sensory
neurones. The capsaicin receptor TRPV1 and the TRPA1 receptor are respec- tively implicated in the transduction of noxious heat and noxious cold sensory stimuli by primary afferents (Bandell et al. 2004). Several studies have demon-
strated the ability of anandamide to induce membrane currents and increase
intracellular calcium in both rat and human cells expressing TRPV1 receptors (Zygmunt et al. 1999; Smart et al. 2000; Ross et al. 2001; Roberts et al. 2002; Dinis et al. 2004; Ross; 2004). More recently, Jordt et al. (2004) described the ability of THC and cannabidiol to activate TRPA1-expressing transfected cells and rat sensory neurones in culture, providing an explanation for the non- CB1/CB2 mediated excitatory effects of AEA on perivascular sensory nerves from TRPV1 knock-out mice (Zygmunt et al. 1999).
Molecular and Cellular Consequences of CB Receptor Activation
CB1 receptors are coupled to pertussis toxin (PTX)-sensitive G (αi) and G (αo) (Gi/o) signalling proteins that function to recruit signal transduction pathways and engage various effector mechanisms within cells (Pertwee 1997; Howlett et al. 2002; McAllister and Glass 2002). The intracellular C-terminal domain of the receptor mediates signalling functions (Nie and Lewis 2001a), controls re- ceptor internalisation/recycling in the membrane after ligand binding (Coutts et al. 2001) and regulates desensitisation after prolonged (>2 h) agonist expo- sure (Kouznetsova et al. 2002). The proximal part of this domain is critical for G protein binding (Nie and Lewis 2001b). CB1 receptor signalling via Gi /Go inhibits cyclic adenosine monophosphate (cAMP) production by adenylate cyclase (Howlett et al. 1988) and modulates ion channel function; including inhibition of voltage gated calcium channels (VGCC) (Mackie and Hille 1992; Twitchell et al. 1997; Chemin et al. 2001). Forskolin-induced production of cAMP [measured in cultured dorsal root ganglion (DRG) cells] can be reduced by CBs (Ross et al. 2001; Oshita et al. 2005), and CB1-mediated inhibition of
the N-type VGCC is implicated in the reduction of stimulus-evoked Ca2+ in-
ﬂux and rises in [Ca2+]i produced by CB agonists applied to cultured sensory neurones (Ross et al. 2001; Khasabova et al. 2004). CB1 activity also modulates
K+ channel conductances, enhancing the activity of both A-type and inwardly rectifying potassium channels (Kir current) (Deadwyler et al. 1995; Mackie et al. 1995; Shen et al. 1996). Where CB1 receptors exist on pre-synaptic nerve
terminals, the effect of these actions is to reduce the probability of activity- regulated transmitter release. Exogenous CBs act pre-synaptically to inhibit glutamatergic (Shen et al. 1996) and γ-aminobutyric acid (GABA)ergic (Ka- tona et al. 1999) transmission in various brain regions including hippocampus,
cerebellum, striatum and nucleus accumbens. CB2 receptors also couple to in- hibitory G proteins to modulate adenylate cyclase activity but not ion channel function. This difference may be explained by the receptor’s low afﬁnity for the Go protein subtype compared to CB1 (McAllister and Glass 2002). CB re- ceptors also couple to signalling pathways related to cell proliferation, which in turn can recruit mitogen activated protein (MAP) kinase and protein ki- nase (PK)B signalling systems (Derkinderen et al. 2001; Bouaboula et al. 1999 see Howlett et al. 2002). For example, PKB activation has been implicated in
the CB receptor-mediated survival of oligodendrocytes after trophic factor withdrawal (Molina-Holgado et al. 2002).
Tissue Expression of CB Receptors in Nociceptive Pathways
CB1 receptors are found primarily in CNS neurones (Egertova et al. 2000; Herkenham et al. 1991; Matsuda et al. 1993; Tsou et al. 1998; Mailleux and Vanderhaeghen 1992), and the majority of CB2 receptors are expressed by cells with inﬂammatory and immune response functions, including glia (Pertwee
1997; Howlett et al. 2002; Walter and Stella 2004). However, there are reports of CB1 receptor expression on peripheral neurones (Bridges et al. 2003) as well as on glial (Salio et al. 2002b; Molina-Holgado et al. 2002) and immune cell types (Galiègue et al. 1995), although in non-neuronal tissues such as spleen, levels of CB1 messenger RNA (mRNA) are lower than CB2 receptor mRNA (Galiègue et al. 1995; Carlisle et al. 2002; Walter and Stella 2004). CB2 mRNA has also been detected in CNS (spinal cord) tissue from injured but not naïve rats (Zhang et al. 2003). CB receptors have been identiﬁed at tissue sites associated with the transmission and processing of nociceptive information. These are the putative cellular targets responsible for mediating the analgesic effects of CB treatment.
CB Receptors in Brain
CB1 receptors are expressed at their highest levels in brain and are particu- larly enriched in cerebral cortex, hippocampus, basal ganglia and cerebellum (Herkenham et al. 1991; Mailleux and Vanderhaeghen 1992; Masuda et al. 1993; Glass et al. 1997). The receptors have also been located in pain-modulating re- gions like periaqueductal grey (PAG), rostro-ventromedial medulla (RVM) and thalamus (Herkenham et al. 1991; Tsou et al. 1998). In the lateral and basal nuclei of the amygdala, CB1 receptors have been localised to a population of cholecystokinin (CCK)-containing interneurones which are activated to in- hibit pre-synaptic release of GABA (Katona et al. 2001). There are relatively low levels of CB1 mRNA in thalamus (Mailleux and Vanderhaeghen 1992; Masuda et al. 1993); however, one study reports an up-regulation of CB1 receptors in this area following injury to peripheral nerve (Siegling et al. 2001).
CB Receptors in Spinal Cord
CB1 receptors have been localised in spinal cord tissue (Farquhar-Smith et al.
2000; Hohmann et al. 1999a; Ong et al. 1999; Sanudo-Pena et al. 1999) where
they are distributed in areas that are important for nociceptive processing. Mi- croarray experiments reveal a 2.9-fold greater abundance of the CB1-encoding
gene in the dorsal compared with the ventral spinal horn (Sun et al. 2002). De- tailed analysis of CB1 immunoreactivity (CB1-ir) in the spinal cord (Farquhar- Smith et al. 2000) revealed a concentration in superﬁcial dorsal horn (laminae I and II), the termination area for nociceptive primary afferent ﬁbres, as well as the dorso-lateral funiculus and lamina X (Farquhar-Smith et al. 2000; Tsou et al. 1998). In lamina II, CB1-ir has been localised to the inner part of lam- ina II (IIi) – the termination zone of non-peptidergic C-ﬁbre afferents – using antibodies to the C-terminal of the receptor (Farquhar-Smith et al. 2000). Alter- natively, using antibodies raised to N-terminal regions of the receptor, CB1-ir is described in the outer part (Ilo) of lamina II, which is the termination zone of peptidergic C-ﬁbres (Tsou et al. 1998; Salio et al. 2002a, b). CB1-ir in the superﬁcial dorsal horn is likely to be located at both pre- and post-synaptic sites, with the major population being expressed by intrinsic spinal neurones (Hohmann et al. 1999a; Hohmann and Herkenham 1999b; Salio et al. 2001; Salio et al. 2002a).
Ultrastructural analysis identiﬁed CB1 receptors on unmyelinated afferent ﬁbre terminals (as well as on astrocytic cells) within superﬁcial spinal cord laminae (Salio et al. 2002a, b). This report of pre-synaptic CB1 receptors would appear to correlate well with the 50% reduction to CB binding sites in the cord reported after an extensive unilateral dorsal rhizotomy (Hohmann and Herken- ham 1999a). However, more restricted lumbar rhizotomy experiments report only a less than 5% reduction to the levels of CB1-ir in the superﬁcial dorsal horn (Farquhar-Smith et al. 2000). In accordance with this study, a modest 16% reduction in CB receptor binding was attributed to TRPV1-responsive afferents ablated by neonatal capsaicin treatment (Hohmann and Herkenham 1998).
The expression of CB1 in the dorso-lateral funiculus (Farquhar-Smith et al.
2000) is consistent with observations that endocannabinoids are involved in
descending modulation of nociceptive processing (Meng et al. 1998; Meng and Johansen 2004). Using immunohistochemistry, a progressive ipsilateral up-regulation of CB1 receptors in the spinal cord has been reported following a chronic constriction injury to rat sciatic nerve, in a mechanism that putatively involves the MAP kinase and PKC signalling pathways (Lim et al. 2003). The in- duction of CB2 receptor mRNA expression in the ipsilateral spinal cord also oc- curs in this neuropathy model, where the location and timing of expression cor- relate with the appearance of activated microglial markers (Zhang et al. 2003).
De-afferentation studies suggest that the contribution of central primary af- ferent terminals to the total CB1-ir in the spinal cord is likely to be small, making it likely that the receptor is expressed on intrinsic neurones, in the spinal cord. Co-staining with markers for intrinsic dorsal horn neurones shows that CB1 receptors are indeed expressed on both excitatory interneurones containing PKCγ (Farquhar-Smith et al. 2000) as well as inhibitory ones containing GABA (Hohmann et al. 1999a; Salio et al. 2002a). Another population co-expresses CB1
and type 1 μ-opioid receptors and is likely to exist separately from GABAergic
interneurones (Salio et al. 2001; Kemp et al. 1996). Enkephalinergic interneu-
rones also contain CB1 mRNA, and it is possible that a proportion of the CB1-ir in the spinal cord exists on the efferent terminals of supraspinal neu- rones, including CB1-expressing neurones from the RVM and PAG that send descending inputs to the spinal cord (Hohmann et al. 1999a; Tsou et al. 1998; Farquhar-Smith et al. 2000). In summary, the anatomical evidence suggests that CBs could potentially operate at both pre- and post-synaptic loci in order to modulate neurotransmission at nociceptor synapses in the dorsal horn.