The analgesic properties of the synthetic cannabinoid WIN55,212-2 were investigated in a model of neuropathic pain. In male Wistar rats, bilateral hind limb withdrawal thresholds to cold, mechanical and noxious thermal stimuli were measured. Following this, unilateral L5 spinal nerve ligation was performed. Seven days later, sensory thresholds were reassessed and the development of allodynia to cold and mechanical stimuli and hyperalgesia to a noxious thermal stimulus confirmed. The effect of WIN55,212-2 (0.1 – 5.0 mg kg−1, i.p.) on the signs of neuropathy was then determined; there was a dose related reversal of all three signs of painful neuropathy at doses which did not generally alter sensory thresholds in the contralateral unligated limb. This effect was prevented by co-administration of the CB1 receptor antagonist SR141716a, but not by co-administration of the CB2 receptor antagonist SR144528, suggesting this action of WIN55,212-2 is mediated via the CB1 receptor. Administration of SR141716a alone had no affect on the observed allodynia and hyperalgesia, which does not support the concept of an endogenous analgesic tone. These data indicate that cannabinoids may have therapeutic potential in neuropathic pain, and that this effect is mediated through the CB1 receptor. Keywords: Cannabinoid, analgesia, neuropathic pain, hyperalgesia, allodynia, WIN55,212-2, CB1 receptor, SR141716a Introduction There have been suggestions that extracts of Cannabis sativa have medicinal properties for thousands of years, but until recently, little sound scientific data has been available to support this hypothesis. However, the structure of Δ9-tetrahydrocannabinol (THC), the major psychoactive component of the 66 known cannabinoids found in marijuana, was elucidated in the early 1960s and since then THC has been shown to be associated with a number of pharmacological effects, including analgesia (Gaoni & Mechoulam, 1964). A further breakthrough in the elucidation of cannabinoid pharmacology came in 1988 with the discovery of a cannabinoid receptor in rat brain (CB1, Devane et al., 1988). This receptor was subsequently cloned in 1990 (Matsuda et al., 1990) and was shown to have a seven transmembrane G-protein coupled structure (Howlett et al., 1986). A second cannabinoid receptor (CB2), with 44% sequence homology to CB1, was identified and cloned in 1993 (Munro et al., 1993) and was found to be predominantly expressed by cells of the immune system. Also in 1992, the endogenous cannabinoid anandamide (AEA) was extracted from the brain and spinal cord (Devane et al., 1992). More recently palmitoylethanolamide (PEA) and 2-arachidonyl glycerol (2-AG) have also been identified as potential endogenous cannabinoids (Mechoulam et al., 1995; Sugiura et al., 1995). The synthesis of the specific high affinity receptor antagonists, SR141716a (SR1) at CB1 (Rinaldi-Carmona et al., 1995; Welch et al., 1998) and SR144528 (SR2) at CB2 (Rinaldi-Carmona et al., 1998; Griffin et al., 1999), and the development of CB1 (Ledent et al., 1999) and CB2 ‘knockout mice' (Buckley et al., 2000) afforded new techniques for the elucidation of CB receptor mediated effects. Several studies have demonstrated an analgesic effect with synthetic and endogenous cannabinoids in inflammatory pain models (for example Jaggar et al., 1998a, 1998b; Calignano et al., 1998; Richardson et al., 1998.) Neuropathic pain is defined as ‘pain initiated or caused by a primary lesion or dysfunction in the nervous system' (Merskey & Bogduk, 1994) and is an area of largely unmet therapeutic need. Tricyclic antidepressants and certain anticonvulsants are the mainstay of clinical therapy for neuropathic pain (McQuay et al., 1996; Sindrup & Jensen, 1999). However, systematic reviews reveal that only between 30 – 50% of patients suffering from neuropathic pain achieve clinically significant (>50%) pain relief with any available single therapy (McQuay et al., 1996; Sindrup & Jensen, 1999). Furthermore, side effects of these therapies often limit their usefulness (McQuay et al., 1996). Although controversial, it is generally accepted that opioids are less effective in neuropathic than inflammatory pain. (Rowbotham, 1999). One explanation for this observation is that spinal opioid receptor expression decreases after peripheral nerve injury (Besse et al., 1992). A recent study has demonstrated little decrease of spinal CB receptor binding when compared to μ-opioid receptor binding after neonatal capsaicin treatment (Hohmann & Herkenham, 1998). Furthermore, it has also been demonstrated using immunocytochemistry that, following dorsal rhizotomy, spinal CB1 receptor expression remains unaltered (Farquhar-Smith et al., 2000). This sparing of CB1 receptors following peripheral nerve injury could indicate a potential therapeutic advantage of cannabinoids over opioids in the treatment of neuropathic pain. This hypothesis is further supported by a recent study (Mao et al., 2000), which demonstrated an anti-hyperalgesic effect of THC in an animal model of neuropathic pain and suggested a therapeutic advantage of THC over opioids in painful neuropathy. Various animal models of neuropathic pain have been developed (Kim et al., 1997). The three most commonly used models share partial injury of the sciatic nerve and a subsequent alteration in hind limb withdrawal thresholds to sensory stimuli as a common feature. In particular, hyperalgesia to a noxious thermal stimulus and allodynia to cold and mechanical stimuli are usually observed. The first model described was the chronic constriction injury (CCI) model in which loose chromic gut ligatures are placed around the sciatic nerve (Bennett & Xie, 1988). A more recent modification was the partial sciatic nerve ligation (PNL) model, in which a tight ligation is placed around 1/3 to 1/2 of the sciatic nerve trunk (Seltzer et al., 1990). A third model is the spinal nerve ligation (SNL) model, in which the L5 and L6 spinal nerves are tightly ligated (Kim & Chung, 1992). A direct comparison of various features of these three models of neuropathic pain has been reported (Kim et al., 1997). In this study, the latency and duration of signs of painful neuropathy were examined, and the authors also examined the effect of sympathectomy on these signs. This study showed that the partial nerve injury evoked signs with roughly the same onset in all three models, but that qualitatively the mechanical and cold allodynia were greatest in SNL. Ongoing pain, as determined by non-weight bearing behaviour, was greatest in CCI. Surgical lumbar sympathectomy caused a reduction in pain behaviour in all three models, with the largest decrease in SNL, suggesting a greater sympathetic nervous system involvement in this model. There are two reports of the effectiveness of cannabinoids in an animal model of neuropathic pain. One study reported that the synthetic cannabinoid WIN55,212-2 alleviated the allodynia and hyperalgesia associated with the CCI model of neuropathic pain (Herzberg et al., 1997). It was confirmed that this effect was mediated through the CB1 receptor and occurred at a dose that was not associated with obvious side effects. This study also demonstrated an increase in thermal hyperalgesia and mechanical allodynia by administration of SR141716a alone, leading the authors to infer the presence of an endogenous tone of cannabinoid analgesia in neuropathy. However, the evidence for such tone during inflammation is controversial (Beaulieu et al., 2000). A separate study reported that THC, administered intrathecally, also alleviated the hyperalgesia of the CCI model (Mao et al., 2000). This effect was shown to be CB1 receptor mediated. It has been suggested that the neuropathy in the CCI model is largely dependent upon an inflammatory reaction (Wagner et al., 1998) and therefore the anti-inflammatory effects of cannabinoids may have obfuscated the true effect on neuropathy in this model. Both the PNL and SNL are associated with a lesser inflammatory component. Therefore, in this study we examined whether the anti-allodynic and anti-hyperalgesic effects of cannabinoids were found in a model of neuropathic pain associated with less of an inflammatory component than the CCI model, namely the SNL model. We also investigated the concept of endogenous tone in this model.