Truth Seeker
New Member
Abstract
Cannabinoid CB1 and CB2 receptors are located at key sites involved in the relaying and processing of noxious inputs. Both CB1 and CB2 receptor agonists have analgesic effects in a range of models of inflammatory and neuropathic pain. Importantly, clinical trials of cannabis-based medicines indicate that the pre-clinical effects of cannabinoid agonists may translate into therapeutic potential in humans. One of the areas of concern with this pharmacological approach is that CB1 receptors have a widespread distribution in the brain and that global activation of CB1 receptors is associated with adverse side effects. Studies of the endogenous cannabinoids (endocannabinoids) have demonstrated that they are present in most tissues and that in some pain states, such as neuropathic pain, levels of endocannabinoids are elevated at key sites involved in pain processing. An alternative approach that can be used to harness the potential therapeutic effects of cannabinoids is to maximise the effects of the endocannabinoids, the actions of which are terminated by re-uptake and metabolism by various enzymes, including fatty acid amide hydrolase (FAAH), monoacylglycerol lipase (MAGL) and cyclooxygenase type 2 (COX2). Preventing the metabolism, or uptake, of endocannabinoids elevates levels of these lipid compounds in tissue and produces behavioural analgesia in models of acute pain. Herein we review recent studies of the effects of inhibition of metabolism of endocannabinoids versus uptake of endocannabinoids on nociceptive processing in models of inflammatory and neuropathic pain.
Introduction
The analgesic effects of the cannabinoid agonists acting at both cannabinoid type 1 (CB1) and CB2 receptors have been widely described and reviewed (Pertwee, 2001; Walker and Huang, 2002; Hohmann and Suplita, 2006). One of the major drawbacks of this approach to analgesia is the widespread distribution of the CB1 receptor in the brain and the associated side effects such as sedation, dependence, cognitive impairment and psychosis (Thomas, 1996; Kalant, 2004; Pacher et al., 2006) arising from their global activation. The recent report that CB1 receptors present on nociceptors play a major role in cannabinoid-induced anti-nociceptive effects suggests that peripherally acting CB1 receptor agonists have potential therapeutic effects (Agarwal et al., 2007). Studies describing anti-nociceptive effects of CB2 receptor agonists in models of inflammatory (Clayton et al., 2002; Quartilho et al., 2003; Nackley et al., 2003a, 2003b; Elmes et al., 2004, 2005; Valenzano et al., 2005; Whiteside et al., 2005) and neuropathic pain (Malan et al., 2001; Ibrahim et al., 2003; Elmes et al., 2004; Scott et al., 2004; Sagar et al., 2005; Valenzano et al., 2005; Whiteside et al., 2005) support the further investigation of this receptor as a potential analgesic target.
The endocannabinoids (ECs), anandamide (AEA; Devane et al., 1992) and 2-arachidonoylglycerol (2AG; Mechoulam et al., 1995; Sugiura et al., 1995) and N-arachidonoyl dopamine (Bisogno et al., 2000; Huang et al., 2002) act at CB1 and CB2 receptors to modulate physiological responses, including nociception (for review see Cravatt and Lichtman, 2004a). Recent studies have investigated the targeting of the ECs, rather than the receptors, as an alternative approach to achieving analgesia in the absence of side effects. This approach should have the benefit of cannabinoid receptor activation at sites of high EC turnover, as opposed to global CB1 receptor activation that can result in side effects. Increased neuronal activity such as that associated with noxious stimulation of pain pathways is believed to drive the synthesis of ECs (Figure 1). There are multiple targets within peripheral tissue at which increased ECs may act including CB1, and potentially CB2, receptors as well as pro-nociceptive transient receptor potential vanilloid subfamily member 1 (TRPV1) receptors present on peripheral nerves and immune cells (Figure 1). The rapid termination of the biological actions of the ECs, by metabolism and transport, however, limits the level of analgesia that is achieved by ECs.
AEA is predominantly metabolised by fatty acid amide hydrolase (FAAH), whereas 2AG is predominantly metabolised by monoacylglycerol lipase (MAGL), nevertheless there is evidence that 2AG is also metabolised by FAAH and both AEA and 2AG can be metabolised by cyclooxygenase type 2 (COX2) (Yu et al., 1997; Kozak et al., 2000, 2001, 2003, 2004), see also review by Fowler in this themed issue. In addition, both AEA and 2AG are believed to be transported into cells (Day et al., 2001; Fegley et al., 2004; Ligresti et al., 2004; Ortega-Gutierrez et al., 2004; McFarland and Barker, 2004; Hillard and Jarrahian, 2005; Hermann et al., 2006), which also contributes to the termination of the biological activity of ECs. Here, we review recent studies of the impact of noxious stimulation on levels of ECs in peripheral and central nervous system (CNS) tissue involved in nociceptive processing and the ability of inhibition of metabolism, or transport of ECs to produce anti-nociceptive effects in models of inflammatory and neuropathic pain.
The relationship between nociceptive processing and levels of ECs
The impact of noxious stimulation on levels of ECs has been investigated in different models of persistent pain. ECs are present in the rat paw skin (Calignano et al., 1998; Beaulieu et al., 2000; Guindon et al., 2006; Maione et al., 2007). There are, however, variable reports on the changes in levels of ECs in the paw skin following inflammation induced by intraplantar injection of formalin (Beaulieu et al., 2000; Maione et al., 2007). Beaulieu et al. (2000) reported that the levels of AEA, 2AG and N-palmitoylethanolamine (PEA) in the paw skin were not altered 1 h after intraplantar injection of formalin in rats. By contrast, levels of AEA, 2AG and PEA have been reported to be reduced in the hindpaw of both mice and rats, 1 h following intraplantar injection of formalin (Maione et al., 2007). We have shown that levels of AEA (Richardson et al., 2007) and PEA (unpublished observation) are significantly reduced in the hindpaw of carrageenan-inflamed rats compared with vehicle control, corroborating the work by Maione et al. (2007). Hindpaw inflammation induced by intraplantar injection of formalin has also been shown to increase levels of AEA in the periaquaductal grey matter (PAG) (Walker et al., 1999), an important area involved in the descending modulation of nociceptive processing (Fields et al., 2006).
There have been a small number of studies in patients on the effects of inflammation on ECs. Increased levels of AEA, but not 2AG, have been reported in inflamed tissue in the colon of patients suffering from ulcerative colitis, a finding replicated in a mouse model of colonic inflammation (D'Argenio et al., 2006). The converse was described in a model of allergic inflammation, which was associated with increased levels of 2AG, but not AEA (Oka et al., 2005).
Overall, levels of ECs and related compounds are differentially altered depending on the nature of inflammation and the tissue under investigation. These differences may arise due to the impact of the pathological condition on levels of the enzymes contributing to the metabolism of ECs and the differential role of these enzymes to the metabolism of different members of the EC and related family of lipids. Thus, levels of FAAH, COX and lipoxygenase (LOX) may be differentially altered in a manner that is specific to the pathological condition. It is also feasible that changes in levels of cannabinoid receptors, which are known to be modulated in pathological conditions, may also impact on levels of ECs. The presence of additional cell types during inflammatory responses may provide a further source of EC synthesis and metabolism, which will impact on levels of ECs present in tissue. An obvious cell type that may contribute to changes in levels of ECs is the mast cell; however, the presence of the EC system in these cells is currently unclear. Although CB1 and CB2 mRNA and protein have been identified in mast cells in vitro and ex vivo (Samson et al., 2003; Stander et al., 2005), functional evidence for cannabinoid receptors was not found in two further studies using a human mast cell line (Maccarrone et al., 2000a) or rat peritoneal mast cells (Lau and Chow, 2003). Another study showed the presence of CB2 receptors on mast cells (Facci et al., 1995); however, when AEA and PEA were tested, only PEA, which displays little or no activity at cannabinoid receptors, produced a functional response. Furthermore, there have been no reports of EC biosynthesis and release from mast cells, although there is one report of an active membrane transport uptake mechanism present in a human mast cell line which was inhibited by the transport inhibitor AM404 (Maccarrone et al., 2000a). Interestingly, this study also showed FAAH-mediated metabolism of AEA in the presence of a 5-LOX inhibitor, but not COX inhibitors, which was subsequently inhibited by PEA (Maccarrone et al., 2000a).
In neuropathic pain states, changes in the levels of ECs and related compounds have been reported in several regions of ascending and descending pain pathways. In general, the effects of neuropathic pain on levels of ECs are more consistent between studies. We have shown that the levels of AEA and N-oleoylethanolamine, but not 2AG, were higher in the ipsilateral hindpaw of neuropathic rats, compared to sham-operated controls (Figure 2a). When considering the peripheral nerves involved in the relay of nociceptive inputs to the spinal cord, increased levels of AEA and 2AG have been reported in the L5 dorsal root ganglion of neuropathic rats (Mitrirattanakul et al., 2006). Similarly, levels of AEA and 2AG, but not PEA, were increased in the spinal cord, PAG and rostral ventromedial medulla in neuropathic rats (Petrosino et al., 2007). In the dorsal raphe nucleus, which has reciprocal projections to the PAG (Jansen et al., 1998), levels of AEA, but not 2AG, were increased in the chronic constriction injury model of neuropathic pain (Palazzo et al., 2006; Petrosino et al., 2007). Collectively, these studies suggest that the impact of inflammation on levels of ECs is variable, depending on the nature of the inflammatory stimulus and experimental protocol. By contrast, in models of neuropathic pain, levels of ECs are consistently reported to be increased at multiple sites involved in the nociceptive processing. It is currently unclear whether these changes in levels of ECs reflect changes in synthesis and/or metabolism. Studies using selective inhibitors of the various catabolic enzymes in models of persistent pain have provided further insight into this question.
Functional effects of inhibition of FAAH in models of pain
A number of studies have shown that FAAH is the predominant enzyme responsible for the hydrolysis of AEA (Cravatt et al., 2001; Kathuria et al., 2003; Lichtman et al., 2004; Cravatt and Lichtman, 2004a; Fegley et al., 2005). Indeed, levels of AEA in FAAH knockout mice are 15-fold higher than levels in wild-type mice and FAAH knockout mice have reduced sensitivity to pain in models of acute and tonic inflammatory pain (Cravatt et al., 2001). There appears to be a complex relationship between the different enzymes involved in the metabolism of ECs in particular under pathological conditions, when LOX is upregulated (reviewed in Phillis et al., 2006). FAAH metabolism of AEA and 2AG has been shown to be limited by LOX-mediated production of hydroperoxides from ECs, which are competitive inhibitors of FAAH in human mast cells (Maccarrone et al., 2000b), although, to date, this has not been shown to be physiologically relevant in other systems.
A tissue-specific FAAH knockout mouse with deletion of the gene coding for FAAH in peripheral tissues (CNS-spared) has provided further insight into the role of peripheral versus central FAAH in nociceptive processing (Cravatt et al., 2004b). Comparison of the CNS-spared FAAH knockout with complete FAAH knockout mice, has shown that CNS-spared FAAH knockout mice do not show hypoalgesia in the tail immersion test but do show a reduction in carrageenan-induced inflammation similar to that seen with the complete knockout. These data suggest that there is an important role of peripheral FAAH in modulating inflammatory nociceptive processing (Cravatt et al., 2004b).
Inhibition of FAAH using pharmacological agents such as URB597 or OL135 produces analgesia and reduces inflammation in models of acute-inflammatory pain (Holt et al., 2005; Chang et al., 2006; Jayamanne et al., 2006). We have studied the influence of FAAH inhibition on the levels of ECs and related compounds in the carrageenan model of inflammatory pain. Intraplantar injection of URB597 (25, 100 μg in 50 μl) dose relatedly increased the levels of AEA and 2AG in the ipsilateral hindpaw of carrageenan-inflamed rats (Richardson et al., 2007). By contrast, the effects of URB597 on inflammatory hyperalgesia were not dose-related (Robinson et al., 2006). The lower dose of URB597 (25 μg in 50 μl) attenuated hyperalgesia, whereas the higher dose of URB597 (100 μg in 50 μl) produced a transient pro-nociceptive effect, despite increasing levels of AEA and 2AG in the hindpaw. These data show that increased levels of ECs do not always translate into anti-nociceptive effects. Studies using a FAAH inhibitor with combined TRPV1 antagonist activity, N-arachidonoyl-serotonin, have indicated anti-nociceptive effects in the formalin test, mediated by activation of CB1 receptors and blockade of TRPV1 receptors (Maione et al., 2007).
Studies of the effects of FAAH inhibitors on aberrant pain behaviour in models of neuropathic pain have produced conflicting reports. To date, inhibition of FAAH has been reported to produce no effect (Jayamanne et al., 2006) or anti-nociceptive effects (Chang et al., 2006; Russo et al., 2007) in models of neuropathic pain. Given this disparity in the literature, it is worthwhile to consider these studies in more detail. In the study by Jayamanne et al. (2006), a single intraperitoneal injection of URB597 (0.3 mg kg−1) was ineffective at reducing mechanical allodynia in the partial sciatic nerve ligation model of neuropathic pain in rats. This same dose of URB597 did, however, attenuate mechanical allodynia in the Complete Freund's Adjuvant model of inflammatory pain, suggesting that there is a shift in the sensitivity to URB597 in neuropathic rats (Jayamanne et al., 2006). A single higher dose (10 mg kg−1 po) of URB597 or repeated oral dosing of URB597 (1—50 mg kg−1 po for 4 days) at the peak of neuropathic pain behaviour dose dependently reduced mechanical hyperalgesia in the mouse chronic constriction injury model (Russo et al., 2007). On the basis of antagonist selectivity, the effects of URB597 were suggested to be mediated by both CB1 and CB2 receptors (Russo et al., 2007). Similarly, intraperitoneal injection of another FAAH inhibitor, OL135, dose dependently (ED50, 5.3 mg kg−1) inhibited mechanical allodynia in the spinal nerve ligation model of neuropathic pain in rats (Chang et al., 2006). N-Arachidonoyl-serotonin has also been shown to attenuate pain behaviour in the chronic constriction injury model of neuropathic pain in rats (Maione et al., 2007).
Given the differential role of peripheral and spinal FAAH in controlling nociceptive processing, identified by the study using FAAH CNS-sparing knockout mice (Cravatt et al., 2004b), we compared the effects of peripheral versus spinal inhibition of FAAH on nociceptive processing in spinal nerve ligation model of neuropathic rats. We reported that a fourfold higher dose of URB597 into the hindpaw was required to inhibit mechanically evoked responses of wide dynamic range spinal cord dorsal horn neurons in neuropathic rats, compared with sham-operated controls (Jhaveri et al., 2006). In this study, intraplantar injection of 100 μg (in 50 μl), but not 25 μg (in 50 μl), URB597 significantly inhibited both innocuous and noxious mechanically evoked responses of wide dynamic range neurons in neuropathic rats. In contrast, intraplantar injection of the lower dose of URB597 significantly attenuated mechanically evoked responses of wide dynamic range neurons in sham-operated rats. Nevertheless, the degree of inhibition produced by the higher dose (100 μg in 50 μl) of URB597 in neuropathic rats was comparable with that produced by the lower dose (25 μg in 50 μl) in sham-operated rats. The functional effects of URB597 in sham-operated rats were associated with significant increases in levels of AEA, 2AG and N-oleoylethanolamine in the hindpaw (Figure 2b). By contrast, neither dose of URB597 increased levels of ECs in the hindpaw of neuropathic rats (Jhaveri et al., 2006). The effects of inhibition of spinal FAAH on neuronal responses in neuropathic rats were also investigated. In this case, spinal administration of URB597 increased levels of ECs and attenuated mechanically evoked responses of spinal neurons to a similar extent in both neuropathic and sham-operated rats (Jhaveri et al., 2006). Thus, despite the reported increased levels of ECs in the spinal cord of neuropathic rats (see earlier), inhibition of FAAH did not produce any greater effect on responses of wide dynamic range neurons in neuropathic rats, compared with sham-operated rats. Our data suggest that there are differential effects of nerve injury on the contribution of FAAH to the metabolism of ECs at peripheral and spinal sites of nociceptive processing. These differences may arise as a result of the presence of immune cells, changes in pH (Paylor et al., 2006) in areas of tissue damage and changes in the expression of FAAH (Lever et al., 2006).
Effects of inhibition of MAGL in models of pain
The predominant route of catabolism of 2AG is described as being via MAGL, which has a localised distribution in the brain (Dinh et al., 2002). There is, however, mounting in vitro (Cravatt et al., 1996; Bisogno et al., 1997; Goparaju et al., 1998, 1999; Ueda et al., 1998; Lang et al., 1999; Fowler et al., 2001; Ueda, 2002) and in vivo (Jhaveri et al., 2006; Maione et al., 2006) evidence suggesting the involvement of FAAH in the metabolism of 2AG. To date, there have been far fewer studies of the effects of inhibitors for MAGL on physiological responses, including nociceptive processing, compared with inhibitors of FAAH because of a lack of selective pharmacological agents. URB602 has been described as a selective, but low potency, inhibitor of MAGL, which does not alter the catabolism of AEA (Hohmann et al., 2005; see however Vandevoorde et al., 2007). Local administration of URB602 into the periaqueductal grey area has been shown to increase levels of 2AG, but not AEA, in the midbrain and enhance non-opioid-mediated stress-induced analgesia (Hohmann et al., 2005). The effects of MAGL inhibition are not limited to central sites of action as local hindpaw administration of URB602 attenuates the late phase of the formalin response, a model of inflammatory pain, with these effects blocked by a CB2 receptor antagonist (Guindon et al., 2007). Although the analgesic effects associated with inhibition of MAGL suggests that this approach has therapeutic potential for the development of novel analgesics, the selectivity of URB602 has recently been questioned. Despite initial work reporting that URB602 inhibited MAGL, but not FAAH (Hohmann et al., 2005), URB602 has subsequently been reported to inhibit both MAGL and FAAH in vitro (Muccioli et al., 2007; Vandevoorde et al., 2007). Thus, whether increasing levels of 2AG produce anti-nociceptive effects remains to be determined once selective inhibitors of MAGL are available.
COX2 metabolism of ECs: increased importance in models of inflammatory pain?
Although hydrolysis of AEA and 2AG by FAAH and MAGL has been most extensively studied, AEA and 2AG can also be oxidised by cyclooxygenase that may have increased importance under conditions associated with an upregulation of the mainly inducible isoform COX2 (reviewed by Fowler in this themed issue). AEA and 2AG interact with the conserved side pocket of the COX2 isoform of the enzyme to produce a range of AEA and 2AG metabolites, such as the prostaglandin ethanolamides and prostaglandin glycerol esters, respectively (Kozak et al., 2001, 2003). COX2 is constitutively expressed in the spinal cord (Ghilardi et al., 2004) but is upregulated both peripherally and spinally in inflammatory and neuropathic pain states (for a review see Svensson and Yaksh, 2002). The analgesic and anti-inflammatory effects of COX2 inhibitors such as meloxicam, SC-58125 and celecoxib in models of inflammatory (Zhang et al., 1997; Yaksh et al., 2001; Francischi et al., 2002) and neuropathic pain (Bingham et al., 2005; Takahashi et al., 2005; Matsunaga et al., 2007) are well described.
The contribution of the cyclooxygenase system to inflammatory and neuropathic hyperalgesia, coupled with the limited literature on the effects of inhibition of COX2 on levels of ECs in vivo, led us to examine the role of COX2 metabolism of ECs under conditions of inflammation. We have shown that intraplantar administration of the COX2 inhibitor nimesulide increased levels of AEA and PEA in carrageenan-inflamed hindpaw, alongside a significant decrease in carrageenan-induced changes in weight bearing an index of hyperalgesia (Robinson et al., 2006, 2007). Although a number of COX inhibitors also inhibit FAAH, previous studies have shown that this is not the case for COX2-selective nimesulide (Fowler et al., 2003). Our data suggest that there is a contribution of COX2 to the metabolism of AEA and, indirectly, PEA under conditions of inflammation in vivo and that blocking this effect can produce analgesia. The effects of another COX2 inhibitor, rofecoxib, which is also a weak inhibitor of FAAH, on levels of ECs has also been reported (Guindon et al., 2006). In this study, peripheral injection of rofecoxib into the rat hindpaw did not alter levels of AEA, N-oleoylethanolamine or PEA in the formalin model of persistent pain. This discrepancy may be due to differences in the importance of COX2- and FAAH-mediated mechanisms and the extent of inflammation induced in the models used. It is clear from these initial studies that further studies of the role of COX2 in the metabolism of ECs in vivo are warranted.
Although the majority of the evidence for COX2 metabolism of AEA and 2AG to prostaglandin ethanolamides and the corresponding prostaglandin glycerol esters has been shown in vitro (Yu et al., 1997; Kozak et al., 2002a), there is evidence that this route of metabolism is relevant in vivo under certain conditions. To date, there have been no reports of detection of AEA or 2AG oxidative metabolites in naive tissue. However, oxidative metabolites of AEA were detected following exogenous administration of AEA in FAAH knockout mice (Weber et al., 2004). This suggests that, although FAAH appears to be the predominant metabolic enzyme of the ECs, there is an enhanced role of alternative metabolic pathways when the EC system is activated while FAAH activity is compromised. Consistent with the idea of functionally relevant EC metabolism by COX2, bimatoprost a structural analogue of prostaglandin-F2α ethanolamide, and prostaglandin-E2 glycerol ester have been shown to produce pharmacological effects in vivo (Woodward et al., 2004; Sang et al., 2006). It has been suggested that bimatoprast, has therapeutic potential as an antiglaucoma agent due to its effects on ocular hypotension (Woodward et al., 2004). Prostaglandin-E2 glycerol ester has been shown to modulate inhibitory synaptic transmission in primary hippocampal neurons (Sang et al., 2006). Although the receptors mediating the effects of these COX2 metabolites of ECs are unknown, it is likely that the role of COX2 in vivo is greater than just terminating the actions of ECs.
Despite little direct evidence in terms of changes in levels of ECs to support their role in COX2-mediated analgesia, links between the two systems are now well established (reviewed by Fowler in this themed issue). Synergistic analgesic actions of ECs and COX2 inhibitors have been shown (Guindon et al., 2006), and although not COX2-specific, several non-selective cyclooxygenase inhibitors inhibit the activity of FAAH in a pH-dependent manner (Holt et al., 2001; Fowler et al., 2003) highlighting the importance of the balance and interactions between metabolic pathways. There is some evidence that ECs are able to modulate COX2 activity, which would provide an interesting feedback control of COX2 metabolism of these compounds. Both AEA and 2AG have been shown to increase cyclooxygenase and 5-LOX activity in intact human neuroblastoma CHP100 cells (Maccarrone et al., 2000b). Similarly, AEA produced a concentration-dependent increase in the expression of COX2 in cerebral microvascular endothelium cells (Chen et al., 2005). By contrast, PEA has been shown to reduce carrageenan inflammation-induced increases in COX activity, which would be expected to produce anti-inflammatory effects (Costa et al., 2002). Although current work continues to focus on the interactions between COX2- and FAAH-mediated mechanisms involved in the analgesic effects of ECs, it is interesting to speculate as to the effects of joint COX/LOX inhibitors on the EC system. On the basis of in vitro evidence that both COX2 and LOX metabolise AEA and 2AG (Yu et al., 1997; Moody et al., 2001; Kozak et al., 2002a, 2002b), it would be predicted that dual inhibition of these enzymes will alter levels of ECs and metabolites. This may involve increasing levels of ECs, or pushing them towards a FAAH-mediated route of metabolism, dependent on the availability and efficiency of FAAH under the pathophysiological conditions.
An area that has been overlooked to date is the impact of the various metabolic pathways on other ECs, such as N-arachidonoyl dopamine and the associated fatty acid amides PEA and N-oleoylethanolamine. Nevertheless, it is clear that the interactions of ECs with these associated compounds have profound effects on the EC system, and therefore, metabolism of these compounds must also be considered alongside AEA and 2AG.
Targeting the cellular uptake of AEA: anti-nociceptive effects in vivo?
In addition to metabolism, removal of synaptic ECs, which involves trans-membrane transport into cells followed by metabolism, regulates their effects. Thus, an alternative approach to increasing levels of ECs is to block their transport. Although the identity of the AEA transporter is still debated (Glaser et al., 2003, 2005), a number of inhibitors of the putative AEA transport process are available, including arvanil, AM404, VDM11, UCM707, OMDM-1 and OMDM-2. UCM707 (Lopez-Rodriguez et al., 2001), unlike its predecessors, has little affinity for TRPV1 channels (Lopez-Rodriguez et al., 2001, 2003) and CB1 receptors (de Lago et al., 2002). This compound does, however, bind to the CB2 receptor (Lopez-Rodriguez et al., 2003), which might compromise interpretations of its effects. The effects of selective inhibitors of EC uptake on nociceptive processing have been investigated in some models of pain. In general terms, the anti-nociceptive effects of inhibitors of the putative AEA transport process are not as robust as the effects of FAAH inhibitors. For example, systemic administration of UCM707 alone did not alter acute nociception in the hot-plate test. It did, however, potentiate the effects of exogenously administered AEA (de Lago et al., 2002). Inhibitors of the putative AEA transport process do appear to have greater effects in models of inflammatory/neuropathic pain. For example, UCM707 significantly inhibited formalin-evoked pain behaviour (La Rana et al., 2006). Systemic administration of AM404, which is a TRPV1 agonist (Rawls et al., 2006), a FAAH inhibitor (Jarrahian et al., 2000) and an inhibitor of the putative AEA transport process, attenuates thermal and mechanical hyperalgesia in models of acute, inflammatory and neuropathic pain (Costa et al., 2006; La Rana et al., 2006) and spinal expression of Fos-immunoreactivity in neuropathic rats (Rodella et al., 2005). AM404 also binds to CB1 and CB2 receptors, albeit with low affinity (Khanolkar et al., 1996). Indeed, the inhibitory effects of AM404 have been shown to be partially mediated via the CB1, CB2 and TRPV1 receptors, with administration of a combination of all three antagonists producing a complete block of the effects of AM404 (Costa et al., 2006). AM404 is also reported to inhibit COX activity in vitro (Hogestatt et al., 2005), which may contribute to the reported functional effects. Other more recently developed inhibitors of the putative AEA transport process, such as OMDM-2, have been reported to be more efficacious in increasing levels of ECs, compared to UCM707 (Ortar et al., 2003; de Lago et al., 2005). The effects of these compounds on nociceptive processing, however, remain to be determined.
In conclusion, despite the multiple enzymes involved in the metabolism of ECs, selective inhibition of FAAH increases levels of ECs that are associated with CB1 and CB2 receptor-mediated anti-nociceptive effects in models of acute and inflammatory pain. Importantly, increase in levels of ECs has not been associated with the adverse side effects associated with direct agonists. The impact of persistent pain states on levels of ECs and the enzymes responsible for their metabolism has yet to be fully addressed. This is particularly relevant in models of inflammatory pain that are associated with an upregulation of COX2. The consensus of studies in neuropathic rats is that levels of ECs are increased at multiple levels following this type of injury, but the effects of inhibition of FAAH are less clear-cut. Further studies of the effects of inhibitors of MAGL and transport mechanisms in models of persistent pain are required to identify their potential therapeutic effect.
Source, Graphs and Figures: Endocannabinoid metabolism and uptake: novel targets for neuropathic and inflammatory pain
Cannabinoid CB1 and CB2 receptors are located at key sites involved in the relaying and processing of noxious inputs. Both CB1 and CB2 receptor agonists have analgesic effects in a range of models of inflammatory and neuropathic pain. Importantly, clinical trials of cannabis-based medicines indicate that the pre-clinical effects of cannabinoid agonists may translate into therapeutic potential in humans. One of the areas of concern with this pharmacological approach is that CB1 receptors have a widespread distribution in the brain and that global activation of CB1 receptors is associated with adverse side effects. Studies of the endogenous cannabinoids (endocannabinoids) have demonstrated that they are present in most tissues and that in some pain states, such as neuropathic pain, levels of endocannabinoids are elevated at key sites involved in pain processing. An alternative approach that can be used to harness the potential therapeutic effects of cannabinoids is to maximise the effects of the endocannabinoids, the actions of which are terminated by re-uptake and metabolism by various enzymes, including fatty acid amide hydrolase (FAAH), monoacylglycerol lipase (MAGL) and cyclooxygenase type 2 (COX2). Preventing the metabolism, or uptake, of endocannabinoids elevates levels of these lipid compounds in tissue and produces behavioural analgesia in models of acute pain. Herein we review recent studies of the effects of inhibition of metabolism of endocannabinoids versus uptake of endocannabinoids on nociceptive processing in models of inflammatory and neuropathic pain.
Introduction
The analgesic effects of the cannabinoid agonists acting at both cannabinoid type 1 (CB1) and CB2 receptors have been widely described and reviewed (Pertwee, 2001; Walker and Huang, 2002; Hohmann and Suplita, 2006). One of the major drawbacks of this approach to analgesia is the widespread distribution of the CB1 receptor in the brain and the associated side effects such as sedation, dependence, cognitive impairment and psychosis (Thomas, 1996; Kalant, 2004; Pacher et al., 2006) arising from their global activation. The recent report that CB1 receptors present on nociceptors play a major role in cannabinoid-induced anti-nociceptive effects suggests that peripherally acting CB1 receptor agonists have potential therapeutic effects (Agarwal et al., 2007). Studies describing anti-nociceptive effects of CB2 receptor agonists in models of inflammatory (Clayton et al., 2002; Quartilho et al., 2003; Nackley et al., 2003a, 2003b; Elmes et al., 2004, 2005; Valenzano et al., 2005; Whiteside et al., 2005) and neuropathic pain (Malan et al., 2001; Ibrahim et al., 2003; Elmes et al., 2004; Scott et al., 2004; Sagar et al., 2005; Valenzano et al., 2005; Whiteside et al., 2005) support the further investigation of this receptor as a potential analgesic target.
The endocannabinoids (ECs), anandamide (AEA; Devane et al., 1992) and 2-arachidonoylglycerol (2AG; Mechoulam et al., 1995; Sugiura et al., 1995) and N-arachidonoyl dopamine (Bisogno et al., 2000; Huang et al., 2002) act at CB1 and CB2 receptors to modulate physiological responses, including nociception (for review see Cravatt and Lichtman, 2004a). Recent studies have investigated the targeting of the ECs, rather than the receptors, as an alternative approach to achieving analgesia in the absence of side effects. This approach should have the benefit of cannabinoid receptor activation at sites of high EC turnover, as opposed to global CB1 receptor activation that can result in side effects. Increased neuronal activity such as that associated with noxious stimulation of pain pathways is believed to drive the synthesis of ECs (Figure 1). There are multiple targets within peripheral tissue at which increased ECs may act including CB1, and potentially CB2, receptors as well as pro-nociceptive transient receptor potential vanilloid subfamily member 1 (TRPV1) receptors present on peripheral nerves and immune cells (Figure 1). The rapid termination of the biological actions of the ECs, by metabolism and transport, however, limits the level of analgesia that is achieved by ECs.
AEA is predominantly metabolised by fatty acid amide hydrolase (FAAH), whereas 2AG is predominantly metabolised by monoacylglycerol lipase (MAGL), nevertheless there is evidence that 2AG is also metabolised by FAAH and both AEA and 2AG can be metabolised by cyclooxygenase type 2 (COX2) (Yu et al., 1997; Kozak et al., 2000, 2001, 2003, 2004), see also review by Fowler in this themed issue. In addition, both AEA and 2AG are believed to be transported into cells (Day et al., 2001; Fegley et al., 2004; Ligresti et al., 2004; Ortega-Gutierrez et al., 2004; McFarland and Barker, 2004; Hillard and Jarrahian, 2005; Hermann et al., 2006), which also contributes to the termination of the biological activity of ECs. Here, we review recent studies of the impact of noxious stimulation on levels of ECs in peripheral and central nervous system (CNS) tissue involved in nociceptive processing and the ability of inhibition of metabolism, or transport of ECs to produce anti-nociceptive effects in models of inflammatory and neuropathic pain.
The relationship between nociceptive processing and levels of ECs
The impact of noxious stimulation on levels of ECs has been investigated in different models of persistent pain. ECs are present in the rat paw skin (Calignano et al., 1998; Beaulieu et al., 2000; Guindon et al., 2006; Maione et al., 2007). There are, however, variable reports on the changes in levels of ECs in the paw skin following inflammation induced by intraplantar injection of formalin (Beaulieu et al., 2000; Maione et al., 2007). Beaulieu et al. (2000) reported that the levels of AEA, 2AG and N-palmitoylethanolamine (PEA) in the paw skin were not altered 1 h after intraplantar injection of formalin in rats. By contrast, levels of AEA, 2AG and PEA have been reported to be reduced in the hindpaw of both mice and rats, 1 h following intraplantar injection of formalin (Maione et al., 2007). We have shown that levels of AEA (Richardson et al., 2007) and PEA (unpublished observation) are significantly reduced in the hindpaw of carrageenan-inflamed rats compared with vehicle control, corroborating the work by Maione et al. (2007). Hindpaw inflammation induced by intraplantar injection of formalin has also been shown to increase levels of AEA in the periaquaductal grey matter (PAG) (Walker et al., 1999), an important area involved in the descending modulation of nociceptive processing (Fields et al., 2006).
There have been a small number of studies in patients on the effects of inflammation on ECs. Increased levels of AEA, but not 2AG, have been reported in inflamed tissue in the colon of patients suffering from ulcerative colitis, a finding replicated in a mouse model of colonic inflammation (D'Argenio et al., 2006). The converse was described in a model of allergic inflammation, which was associated with increased levels of 2AG, but not AEA (Oka et al., 2005).
Overall, levels of ECs and related compounds are differentially altered depending on the nature of inflammation and the tissue under investigation. These differences may arise due to the impact of the pathological condition on levels of the enzymes contributing to the metabolism of ECs and the differential role of these enzymes to the metabolism of different members of the EC and related family of lipids. Thus, levels of FAAH, COX and lipoxygenase (LOX) may be differentially altered in a manner that is specific to the pathological condition. It is also feasible that changes in levels of cannabinoid receptors, which are known to be modulated in pathological conditions, may also impact on levels of ECs. The presence of additional cell types during inflammatory responses may provide a further source of EC synthesis and metabolism, which will impact on levels of ECs present in tissue. An obvious cell type that may contribute to changes in levels of ECs is the mast cell; however, the presence of the EC system in these cells is currently unclear. Although CB1 and CB2 mRNA and protein have been identified in mast cells in vitro and ex vivo (Samson et al., 2003; Stander et al., 2005), functional evidence for cannabinoid receptors was not found in two further studies using a human mast cell line (Maccarrone et al., 2000a) or rat peritoneal mast cells (Lau and Chow, 2003). Another study showed the presence of CB2 receptors on mast cells (Facci et al., 1995); however, when AEA and PEA were tested, only PEA, which displays little or no activity at cannabinoid receptors, produced a functional response. Furthermore, there have been no reports of EC biosynthesis and release from mast cells, although there is one report of an active membrane transport uptake mechanism present in a human mast cell line which was inhibited by the transport inhibitor AM404 (Maccarrone et al., 2000a). Interestingly, this study also showed FAAH-mediated metabolism of AEA in the presence of a 5-LOX inhibitor, but not COX inhibitors, which was subsequently inhibited by PEA (Maccarrone et al., 2000a).
In neuropathic pain states, changes in the levels of ECs and related compounds have been reported in several regions of ascending and descending pain pathways. In general, the effects of neuropathic pain on levels of ECs are more consistent between studies. We have shown that the levels of AEA and N-oleoylethanolamine, but not 2AG, were higher in the ipsilateral hindpaw of neuropathic rats, compared to sham-operated controls (Figure 2a). When considering the peripheral nerves involved in the relay of nociceptive inputs to the spinal cord, increased levels of AEA and 2AG have been reported in the L5 dorsal root ganglion of neuropathic rats (Mitrirattanakul et al., 2006). Similarly, levels of AEA and 2AG, but not PEA, were increased in the spinal cord, PAG and rostral ventromedial medulla in neuropathic rats (Petrosino et al., 2007). In the dorsal raphe nucleus, which has reciprocal projections to the PAG (Jansen et al., 1998), levels of AEA, but not 2AG, were increased in the chronic constriction injury model of neuropathic pain (Palazzo et al., 2006; Petrosino et al., 2007). Collectively, these studies suggest that the impact of inflammation on levels of ECs is variable, depending on the nature of the inflammatory stimulus and experimental protocol. By contrast, in models of neuropathic pain, levels of ECs are consistently reported to be increased at multiple sites involved in the nociceptive processing. It is currently unclear whether these changes in levels of ECs reflect changes in synthesis and/or metabolism. Studies using selective inhibitors of the various catabolic enzymes in models of persistent pain have provided further insight into this question.
Functional effects of inhibition of FAAH in models of pain
A number of studies have shown that FAAH is the predominant enzyme responsible for the hydrolysis of AEA (Cravatt et al., 2001; Kathuria et al., 2003; Lichtman et al., 2004; Cravatt and Lichtman, 2004a; Fegley et al., 2005). Indeed, levels of AEA in FAAH knockout mice are 15-fold higher than levels in wild-type mice and FAAH knockout mice have reduced sensitivity to pain in models of acute and tonic inflammatory pain (Cravatt et al., 2001). There appears to be a complex relationship between the different enzymes involved in the metabolism of ECs in particular under pathological conditions, when LOX is upregulated (reviewed in Phillis et al., 2006). FAAH metabolism of AEA and 2AG has been shown to be limited by LOX-mediated production of hydroperoxides from ECs, which are competitive inhibitors of FAAH in human mast cells (Maccarrone et al., 2000b), although, to date, this has not been shown to be physiologically relevant in other systems.
A tissue-specific FAAH knockout mouse with deletion of the gene coding for FAAH in peripheral tissues (CNS-spared) has provided further insight into the role of peripheral versus central FAAH in nociceptive processing (Cravatt et al., 2004b). Comparison of the CNS-spared FAAH knockout with complete FAAH knockout mice, has shown that CNS-spared FAAH knockout mice do not show hypoalgesia in the tail immersion test but do show a reduction in carrageenan-induced inflammation similar to that seen with the complete knockout. These data suggest that there is an important role of peripheral FAAH in modulating inflammatory nociceptive processing (Cravatt et al., 2004b).
Inhibition of FAAH using pharmacological agents such as URB597 or OL135 produces analgesia and reduces inflammation in models of acute-inflammatory pain (Holt et al., 2005; Chang et al., 2006; Jayamanne et al., 2006). We have studied the influence of FAAH inhibition on the levels of ECs and related compounds in the carrageenan model of inflammatory pain. Intraplantar injection of URB597 (25, 100 μg in 50 μl) dose relatedly increased the levels of AEA and 2AG in the ipsilateral hindpaw of carrageenan-inflamed rats (Richardson et al., 2007). By contrast, the effects of URB597 on inflammatory hyperalgesia were not dose-related (Robinson et al., 2006). The lower dose of URB597 (25 μg in 50 μl) attenuated hyperalgesia, whereas the higher dose of URB597 (100 μg in 50 μl) produced a transient pro-nociceptive effect, despite increasing levels of AEA and 2AG in the hindpaw. These data show that increased levels of ECs do not always translate into anti-nociceptive effects. Studies using a FAAH inhibitor with combined TRPV1 antagonist activity, N-arachidonoyl-serotonin, have indicated anti-nociceptive effects in the formalin test, mediated by activation of CB1 receptors and blockade of TRPV1 receptors (Maione et al., 2007).
Studies of the effects of FAAH inhibitors on aberrant pain behaviour in models of neuropathic pain have produced conflicting reports. To date, inhibition of FAAH has been reported to produce no effect (Jayamanne et al., 2006) or anti-nociceptive effects (Chang et al., 2006; Russo et al., 2007) in models of neuropathic pain. Given this disparity in the literature, it is worthwhile to consider these studies in more detail. In the study by Jayamanne et al. (2006), a single intraperitoneal injection of URB597 (0.3 mg kg−1) was ineffective at reducing mechanical allodynia in the partial sciatic nerve ligation model of neuropathic pain in rats. This same dose of URB597 did, however, attenuate mechanical allodynia in the Complete Freund's Adjuvant model of inflammatory pain, suggesting that there is a shift in the sensitivity to URB597 in neuropathic rats (Jayamanne et al., 2006). A single higher dose (10 mg kg−1 po) of URB597 or repeated oral dosing of URB597 (1—50 mg kg−1 po for 4 days) at the peak of neuropathic pain behaviour dose dependently reduced mechanical hyperalgesia in the mouse chronic constriction injury model (Russo et al., 2007). On the basis of antagonist selectivity, the effects of URB597 were suggested to be mediated by both CB1 and CB2 receptors (Russo et al., 2007). Similarly, intraperitoneal injection of another FAAH inhibitor, OL135, dose dependently (ED50, 5.3 mg kg−1) inhibited mechanical allodynia in the spinal nerve ligation model of neuropathic pain in rats (Chang et al., 2006). N-Arachidonoyl-serotonin has also been shown to attenuate pain behaviour in the chronic constriction injury model of neuropathic pain in rats (Maione et al., 2007).
Given the differential role of peripheral and spinal FAAH in controlling nociceptive processing, identified by the study using FAAH CNS-sparing knockout mice (Cravatt et al., 2004b), we compared the effects of peripheral versus spinal inhibition of FAAH on nociceptive processing in spinal nerve ligation model of neuropathic rats. We reported that a fourfold higher dose of URB597 into the hindpaw was required to inhibit mechanically evoked responses of wide dynamic range spinal cord dorsal horn neurons in neuropathic rats, compared with sham-operated controls (Jhaveri et al., 2006). In this study, intraplantar injection of 100 μg (in 50 μl), but not 25 μg (in 50 μl), URB597 significantly inhibited both innocuous and noxious mechanically evoked responses of wide dynamic range neurons in neuropathic rats. In contrast, intraplantar injection of the lower dose of URB597 significantly attenuated mechanically evoked responses of wide dynamic range neurons in sham-operated rats. Nevertheless, the degree of inhibition produced by the higher dose (100 μg in 50 μl) of URB597 in neuropathic rats was comparable with that produced by the lower dose (25 μg in 50 μl) in sham-operated rats. The functional effects of URB597 in sham-operated rats were associated with significant increases in levels of AEA, 2AG and N-oleoylethanolamine in the hindpaw (Figure 2b). By contrast, neither dose of URB597 increased levels of ECs in the hindpaw of neuropathic rats (Jhaveri et al., 2006). The effects of inhibition of spinal FAAH on neuronal responses in neuropathic rats were also investigated. In this case, spinal administration of URB597 increased levels of ECs and attenuated mechanically evoked responses of spinal neurons to a similar extent in both neuropathic and sham-operated rats (Jhaveri et al., 2006). Thus, despite the reported increased levels of ECs in the spinal cord of neuropathic rats (see earlier), inhibition of FAAH did not produce any greater effect on responses of wide dynamic range neurons in neuropathic rats, compared with sham-operated rats. Our data suggest that there are differential effects of nerve injury on the contribution of FAAH to the metabolism of ECs at peripheral and spinal sites of nociceptive processing. These differences may arise as a result of the presence of immune cells, changes in pH (Paylor et al., 2006) in areas of tissue damage and changes in the expression of FAAH (Lever et al., 2006).
Effects of inhibition of MAGL in models of pain
The predominant route of catabolism of 2AG is described as being via MAGL, which has a localised distribution in the brain (Dinh et al., 2002). There is, however, mounting in vitro (Cravatt et al., 1996; Bisogno et al., 1997; Goparaju et al., 1998, 1999; Ueda et al., 1998; Lang et al., 1999; Fowler et al., 2001; Ueda, 2002) and in vivo (Jhaveri et al., 2006; Maione et al., 2006) evidence suggesting the involvement of FAAH in the metabolism of 2AG. To date, there have been far fewer studies of the effects of inhibitors for MAGL on physiological responses, including nociceptive processing, compared with inhibitors of FAAH because of a lack of selective pharmacological agents. URB602 has been described as a selective, but low potency, inhibitor of MAGL, which does not alter the catabolism of AEA (Hohmann et al., 2005; see however Vandevoorde et al., 2007). Local administration of URB602 into the periaqueductal grey area has been shown to increase levels of 2AG, but not AEA, in the midbrain and enhance non-opioid-mediated stress-induced analgesia (Hohmann et al., 2005). The effects of MAGL inhibition are not limited to central sites of action as local hindpaw administration of URB602 attenuates the late phase of the formalin response, a model of inflammatory pain, with these effects blocked by a CB2 receptor antagonist (Guindon et al., 2007). Although the analgesic effects associated with inhibition of MAGL suggests that this approach has therapeutic potential for the development of novel analgesics, the selectivity of URB602 has recently been questioned. Despite initial work reporting that URB602 inhibited MAGL, but not FAAH (Hohmann et al., 2005), URB602 has subsequently been reported to inhibit both MAGL and FAAH in vitro (Muccioli et al., 2007; Vandevoorde et al., 2007). Thus, whether increasing levels of 2AG produce anti-nociceptive effects remains to be determined once selective inhibitors of MAGL are available.
COX2 metabolism of ECs: increased importance in models of inflammatory pain?
Although hydrolysis of AEA and 2AG by FAAH and MAGL has been most extensively studied, AEA and 2AG can also be oxidised by cyclooxygenase that may have increased importance under conditions associated with an upregulation of the mainly inducible isoform COX2 (reviewed by Fowler in this themed issue). AEA and 2AG interact with the conserved side pocket of the COX2 isoform of the enzyme to produce a range of AEA and 2AG metabolites, such as the prostaglandin ethanolamides and prostaglandin glycerol esters, respectively (Kozak et al., 2001, 2003). COX2 is constitutively expressed in the spinal cord (Ghilardi et al., 2004) but is upregulated both peripherally and spinally in inflammatory and neuropathic pain states (for a review see Svensson and Yaksh, 2002). The analgesic and anti-inflammatory effects of COX2 inhibitors such as meloxicam, SC-58125 and celecoxib in models of inflammatory (Zhang et al., 1997; Yaksh et al., 2001; Francischi et al., 2002) and neuropathic pain (Bingham et al., 2005; Takahashi et al., 2005; Matsunaga et al., 2007) are well described.
The contribution of the cyclooxygenase system to inflammatory and neuropathic hyperalgesia, coupled with the limited literature on the effects of inhibition of COX2 on levels of ECs in vivo, led us to examine the role of COX2 metabolism of ECs under conditions of inflammation. We have shown that intraplantar administration of the COX2 inhibitor nimesulide increased levels of AEA and PEA in carrageenan-inflamed hindpaw, alongside a significant decrease in carrageenan-induced changes in weight bearing an index of hyperalgesia (Robinson et al., 2006, 2007). Although a number of COX inhibitors also inhibit FAAH, previous studies have shown that this is not the case for COX2-selective nimesulide (Fowler et al., 2003). Our data suggest that there is a contribution of COX2 to the metabolism of AEA and, indirectly, PEA under conditions of inflammation in vivo and that blocking this effect can produce analgesia. The effects of another COX2 inhibitor, rofecoxib, which is also a weak inhibitor of FAAH, on levels of ECs has also been reported (Guindon et al., 2006). In this study, peripheral injection of rofecoxib into the rat hindpaw did not alter levels of AEA, N-oleoylethanolamine or PEA in the formalin model of persistent pain. This discrepancy may be due to differences in the importance of COX2- and FAAH-mediated mechanisms and the extent of inflammation induced in the models used. It is clear from these initial studies that further studies of the role of COX2 in the metabolism of ECs in vivo are warranted.
Although the majority of the evidence for COX2 metabolism of AEA and 2AG to prostaglandin ethanolamides and the corresponding prostaglandin glycerol esters has been shown in vitro (Yu et al., 1997; Kozak et al., 2002a), there is evidence that this route of metabolism is relevant in vivo under certain conditions. To date, there have been no reports of detection of AEA or 2AG oxidative metabolites in naive tissue. However, oxidative metabolites of AEA were detected following exogenous administration of AEA in FAAH knockout mice (Weber et al., 2004). This suggests that, although FAAH appears to be the predominant metabolic enzyme of the ECs, there is an enhanced role of alternative metabolic pathways when the EC system is activated while FAAH activity is compromised. Consistent with the idea of functionally relevant EC metabolism by COX2, bimatoprost a structural analogue of prostaglandin-F2α ethanolamide, and prostaglandin-E2 glycerol ester have been shown to produce pharmacological effects in vivo (Woodward et al., 2004; Sang et al., 2006). It has been suggested that bimatoprast, has therapeutic potential as an antiglaucoma agent due to its effects on ocular hypotension (Woodward et al., 2004). Prostaglandin-E2 glycerol ester has been shown to modulate inhibitory synaptic transmission in primary hippocampal neurons (Sang et al., 2006). Although the receptors mediating the effects of these COX2 metabolites of ECs are unknown, it is likely that the role of COX2 in vivo is greater than just terminating the actions of ECs.
Despite little direct evidence in terms of changes in levels of ECs to support their role in COX2-mediated analgesia, links between the two systems are now well established (reviewed by Fowler in this themed issue). Synergistic analgesic actions of ECs and COX2 inhibitors have been shown (Guindon et al., 2006), and although not COX2-specific, several non-selective cyclooxygenase inhibitors inhibit the activity of FAAH in a pH-dependent manner (Holt et al., 2001; Fowler et al., 2003) highlighting the importance of the balance and interactions between metabolic pathways. There is some evidence that ECs are able to modulate COX2 activity, which would provide an interesting feedback control of COX2 metabolism of these compounds. Both AEA and 2AG have been shown to increase cyclooxygenase and 5-LOX activity in intact human neuroblastoma CHP100 cells (Maccarrone et al., 2000b). Similarly, AEA produced a concentration-dependent increase in the expression of COX2 in cerebral microvascular endothelium cells (Chen et al., 2005). By contrast, PEA has been shown to reduce carrageenan inflammation-induced increases in COX activity, which would be expected to produce anti-inflammatory effects (Costa et al., 2002). Although current work continues to focus on the interactions between COX2- and FAAH-mediated mechanisms involved in the analgesic effects of ECs, it is interesting to speculate as to the effects of joint COX/LOX inhibitors on the EC system. On the basis of in vitro evidence that both COX2 and LOX metabolise AEA and 2AG (Yu et al., 1997; Moody et al., 2001; Kozak et al., 2002a, 2002b), it would be predicted that dual inhibition of these enzymes will alter levels of ECs and metabolites. This may involve increasing levels of ECs, or pushing them towards a FAAH-mediated route of metabolism, dependent on the availability and efficiency of FAAH under the pathophysiological conditions.
An area that has been overlooked to date is the impact of the various metabolic pathways on other ECs, such as N-arachidonoyl dopamine and the associated fatty acid amides PEA and N-oleoylethanolamine. Nevertheless, it is clear that the interactions of ECs with these associated compounds have profound effects on the EC system, and therefore, metabolism of these compounds must also be considered alongside AEA and 2AG.
Targeting the cellular uptake of AEA: anti-nociceptive effects in vivo?
In addition to metabolism, removal of synaptic ECs, which involves trans-membrane transport into cells followed by metabolism, regulates their effects. Thus, an alternative approach to increasing levels of ECs is to block their transport. Although the identity of the AEA transporter is still debated (Glaser et al., 2003, 2005), a number of inhibitors of the putative AEA transport process are available, including arvanil, AM404, VDM11, UCM707, OMDM-1 and OMDM-2. UCM707 (Lopez-Rodriguez et al., 2001), unlike its predecessors, has little affinity for TRPV1 channels (Lopez-Rodriguez et al., 2001, 2003) and CB1 receptors (de Lago et al., 2002). This compound does, however, bind to the CB2 receptor (Lopez-Rodriguez et al., 2003), which might compromise interpretations of its effects. The effects of selective inhibitors of EC uptake on nociceptive processing have been investigated in some models of pain. In general terms, the anti-nociceptive effects of inhibitors of the putative AEA transport process are not as robust as the effects of FAAH inhibitors. For example, systemic administration of UCM707 alone did not alter acute nociception in the hot-plate test. It did, however, potentiate the effects of exogenously administered AEA (de Lago et al., 2002). Inhibitors of the putative AEA transport process do appear to have greater effects in models of inflammatory/neuropathic pain. For example, UCM707 significantly inhibited formalin-evoked pain behaviour (La Rana et al., 2006). Systemic administration of AM404, which is a TRPV1 agonist (Rawls et al., 2006), a FAAH inhibitor (Jarrahian et al., 2000) and an inhibitor of the putative AEA transport process, attenuates thermal and mechanical hyperalgesia in models of acute, inflammatory and neuropathic pain (Costa et al., 2006; La Rana et al., 2006) and spinal expression of Fos-immunoreactivity in neuropathic rats (Rodella et al., 2005). AM404 also binds to CB1 and CB2 receptors, albeit with low affinity (Khanolkar et al., 1996). Indeed, the inhibitory effects of AM404 have been shown to be partially mediated via the CB1, CB2 and TRPV1 receptors, with administration of a combination of all three antagonists producing a complete block of the effects of AM404 (Costa et al., 2006). AM404 is also reported to inhibit COX activity in vitro (Hogestatt et al., 2005), which may contribute to the reported functional effects. Other more recently developed inhibitors of the putative AEA transport process, such as OMDM-2, have been reported to be more efficacious in increasing levels of ECs, compared to UCM707 (Ortar et al., 2003; de Lago et al., 2005). The effects of these compounds on nociceptive processing, however, remain to be determined.
In conclusion, despite the multiple enzymes involved in the metabolism of ECs, selective inhibition of FAAH increases levels of ECs that are associated with CB1 and CB2 receptor-mediated anti-nociceptive effects in models of acute and inflammatory pain. Importantly, increase in levels of ECs has not been associated with the adverse side effects associated with direct agonists. The impact of persistent pain states on levels of ECs and the enzymes responsible for their metabolism has yet to be fully addressed. This is particularly relevant in models of inflammatory pain that are associated with an upregulation of COX2. The consensus of studies in neuropathic rats is that levels of ECs are increased at multiple levels following this type of injury, but the effects of inhibition of FAAH are less clear-cut. Further studies of the effects of inhibitors of MAGL and transport mechanisms in models of persistent pain are required to identify their potential therapeutic effect.
Source, Graphs and Figures: Endocannabinoid metabolism and uptake: novel targets for neuropathic and inflammatory pain