Despite significantly increasing brain and spinal cord tissue levels of FAA in SCI rats, seven day systemic administration of the FAAH inhibitor URB597 did not demonstrate antinociception. Acute treatment with the FAAH inhibitor PF-3845, which also significantly increased CNS levels of FAA, led to a moderate, non-dose dependent antinociception in SCI rats. By contrast, WIN 55,212-2 robustly ameliorated below-level cutaneous hypersensitivity in SCI rats for the duration of the seven-day treatment period. Markedly decreased tissue levels of OEA and PEA were observed at the epicenter and tissue caudal to the epicenter in SCI rats. While both FAAH inhibitors increased tissue FAA in SCI rats, only PF-3845 increased AEA, OEA and PEA at the epicenter, which in part could underlie PF-3845's modest antinociceptive effect. Neither FAAH inhibition nor CB receptor activation over seven days improved hind limb motor function. While direct activation of CB receptors leads to a significant and sustained antinociception, the current data indicate that increasing CNS FAA levels alone may not be sufficient for ameliorating below-level neuropathic SCI pain.

Changes in FAA levels in spinal tissue segments that may be crucial in the initiation and maintenance of neuropathic pain following a SCI were reported in the current study. The current study confirmed a previous finding of no change in levels of AEA and PEA in spinal tissue rostral to the epicenter and at the epicenter four weeks after a SCI [26]. However, the current study observed that PEA was decreased caudal to the epicenter and that OEA was decreased both rostrally and caudally to the epicenter as well as within the epicenter. It is speculated that the loss of tonic FAA-meditated inhibition following SCI could contribute to enhanced excitability of spinal dorsal horn neurons either rostral or caudal to the epicenter [2], [27]. In the current study, slight decreases (p>0.05; Table 1) in FAA were observed in the lumbar spinal cord, which receives primary afferents that innervate the hind paw. The mechanism that leads to an apparent decrease of FAA in lumbar segment, several segments from the epicenter, has yet to be elucidated, but a consequence could be dorsal horn neuron hyperexcitability. However, lumbar spinal cord FAA were significantly increased following URB597 treatment without being accompanied by significant antinociception. Likewise, FAA levels in lumbar spinal cord were increased following PF-3845, but only a slight antinociceptive effect was obtained. Convincing evidence that the loss of CB₁ receptor-mediated tonic inhibition alone can lead to hyperexcitable lumbar dorsal horn neurons is lacking, but perhaps it is a fusion of diminished functioning of several neurotransmitter systems (e.g. GABA) that leads to the neuropathic pain state following SCI [28], [29].

Changes in supraspinal function could also be crucial in the maintenance of neuropathic SCI pain [16], [19]. The current study found no significant change in whole brain FAA levels following SCI, but perhaps there are changes in specific pain-related brain nuclei following SCI. It is further speculated that processes involved in cellular function and intercellular signaling could be altered in the brain as well as in the spinal cord following injury [30], [31]. Changes in the brain opioid system have been suggested to underlie reduced efficacy of supraspinal injection of morphine in SCI rats [32], [33]. By contrast, supraspinal injection of WIN 55,212-2 is fully efficacious in SCI rats [19]. Changes in either brain CB receptors or FAA level have yet to be identified following SCI. Nonetheless, despite a robust increase in whole brain FAA following FAAH inhibition, there was a lack of robust antinociception; a role for brain FAA in SCI pain will need further elucidation.

A number of preclinical neuropathic pain studies have demonstrated marked antinociception following both acute and repeated activation of CB receptors, which appears to be in part mediated through the CB₁ receptor [3], [20], [34]–[38]. It is possible that attenuation of specific symptoms, such as hypersensitivity to noxious stimulation (hyperalgesia), could be ameliorated through other CB receptors or non-CB receptors [11], [39].

The current study confirms previous findings in neuropathic pain models of a lack tolerance to the antinociceptive effect following repeated treatment with a CB receptor agonist [20], [35], [36]. By contrast, a loss of efficacy following repeated treatment with morphine in SCI rats has been observed [20]. Thus, a potential advantage of CB1 receptor agonists over, for example, opioid receptor agonists is sustained efficacy over the course of the treatment period. Tolerance to CB receptor agonists does appear however, under specific circumstances. In the case of an acute pain state, as modeled by the hot plate test in uninjured animals, tolerance to the antinociceptive effect appears within five days of daily treatment [20], [40]. The mechanism by which CB receptor agonists retain efficacy in some pain states and not in others is currently unknown. While it is possible that the injury state could underlie the development of antinociceptive tolerance, the test stimulus could also influence whether or not tolerance is obtained.

One issue raised by CB receptor agonists is the presence of motoric effects at or near antinociceptive doses. In the case of WIN 55,212-2, the 50% antinociceptive dose is 0.9 mg/kg (s.c., 30 min post-dosing) whereas the dose needed to produce 50% disruption on the rotarod test (s.c., 3 hrs. post-dosing) is about the same [14], [20]. In a previous study, a sedating dose of the anxiolytic drug diazepam, did not increase hind paw withdrawal thresholds in SCI rats, indicating that the behavioral endpoint is sensitive to antinociceptive treatment rather than to a treatment that merely disrupt motor function [18]. Thus, the increase in withdrawal thresholds observed with WIN 55,212-2 in the current study cannot be entirely attributed to catalepsy or motor dysfunction.

While targeting CB₁ receptors is a promising therapeutic strategy for ameliorating neuropathic SCI pain, prominent psychomotor side effects are associated with CB₁ receptor activation [14], [20]. Alternatively, elevating synaptic levels of FAA is antinociceptive and does not appear to induce adverse side effects as the antinociception appears to be mediated through the CB₂ receptor, and possibly through other non-CB mediated mechanisms, such as through activation of the TRPV1 channel, as well as through the CB₁ receptor [9]–[11], [41]. URB597 is a potent FAAH inhibitor—50% of brain FAAH activity is inhibited at a dose of 0.15 mg/kg (i.p.) [10]. At doses higher than 0.3 mg/kg, 1 and 3 mg/kg (i.p.), levels of FAA do not appear to be greatly increased, thus a FAAH inhibition plateau is reached at about 1 mg/kg [42]. In vivo, URB597 inhibits the degradation of AEA within 15 min of injection and significant inhibition of FAAH activity is observed up to 12 hours post-injection [10]. Importantly, no deleterious effect on motor function, as assessed in the rotarod test, was observed at doses of up to 5 mg/kg of URB597 and robust behavioral effects (e.g. antinociception, anxiolytic-like effects) were observed at very low doses (0.1–0.5 mg/kg, i.p.) [13], [43]. In the current study, significant spinal tissue levels of FAA were obtained with 3 mg/kg URB597 in both SCI and uninjured rats. Similarly, both doses of PF-3845 significantly raised FAA levels in spinal tissues, including the epicenter of spinal cord from SCI rats as well as spinal tissue from uninjured rats. Thus, the current study demonstrates that it is possible to increase FAA in injured spinal tissue.

Despite significantly increasing CNS FAA levels, no antinociceptive effect was observed with URB597 in SCI rats, following either a single dose (day 1) or multiple doses (at either day 3 or day 7). A single dose of URB597 in peripheral nerve-injured mice resulted in a small antinociceptive effect, but four days of treatment, however, led to an even greater antinociceptive effect [22]. One significant difference between the previous studies showing efficacy with URB597 and the current study is that previous studies utilized models of peripheral neuropathy whereas the current study utilized a model of central pain following a SCI. There are a number of drugs that are efficacious in peripheral neuropathic pain models but are not efficacious in neuropathic SCI models [17], [18]. Why this is so is not entirely clear, but the differential pharmacology between peripheral and central neuropathic pain states highly suggests that there are meaningful differences in the substrates that mediate the symptoms presented by these pain states. The differential substrates hypothesized to exist in preclinical models is indirectly supported by clinical findings, as a similar pharmacological divergence also exists between clinical neuropathic SCI and peripheral neuropathic pains [2].

Whereas URB597 lacked efficacy, PF-3845 led to a partial, non-dose dependent amelioration of neuropathic SCI pain. In addition to inhibitory effects on neurons, FAA concurrently suppress the release of pro-inflammatory mediators from immune cells and facilitate the release of anti-inflammatory substances [26]. It is possible that PF-3845 is more potent than URB597 on FAAH expressed in microglia and macrophage or that PF-3845 better penetrates damaged tissue as found at the epicenter to reach these cell types. In fact, 3 mg/kg (p.o.) PF-3845 led to a greater increase of AEA in CNS tissues, including the epicenter, compared to 3 mg/kg (i.p.) URB597. It has been reported elsewhere that FAA levels in mice treated with PF-3845 were higher compared to mice treated with URB597, despite the same dose and route [12]. Following PF-3845 treatment, FAA and anti-inflammatory substances from the epicenter could diffuse to nearby spinal dorsal horn neurons that are hyperexcited following SCI [44]. Perhaps it is the increase in anti-inflammatory substances within and adjacent to the epicenter that is crucial in suppressing hyperexcited central neurons. The antinociceptive effect observed with PF-3845 in neuropathic SCI rats could be partially mediated through a combination of CB receptor-dependent and CB receptor-independent mechanisms [45], [46].

While no robust effect was observed on below-level pain in the current study, it is possible that above-level and at-level SCI pain could be attenuated by FAAH inhibition. In addition to significant hyperexcitation of central neurons, there appears to be significant primary afferent hyperexcitation following SCI, which could underlie cutaneous hypersensitivity and spontaneous pain at and above the SCI [44], [47], [48]. These afferents could be sensitive to FAA through CB receptor or non-CB receptor mediated mechanisms or both. Increased sodium channel expression has been observed in injured peripheral nerves following injury—perhaps a similar increase in sodium channel expression in peripheral nerves following a SCI, as reported in the CNS, underlies at-level pain [49], [50]. Indeed, SCI patients have reported relief from at-level pain following application of either capsaicin, a TRPV1 agonist, or lidocaine, a sodium channel blocking local anesthetic [51], [52]. Perhaps the slight increase in FAA obtained with PF-3845 would have been sufficient to reduce at-level neuropathic pain. Further studies, however, are needed to characterize the extent of the involvement of primary afferent neurons in SCI pain.

A closer examination of studies that tested FAAH inhibitors on peripheral neuropathic cutaneous hypersensitivity indicates weak to no efficacy. As mentioned earlier, a single treatment of URB597 (10 mg/kg, p.o.) in mice with a chronic constriction injury of the sciatic nerve led to a small antinociceptive effect [22]. In rats with a spinal nerve ligation, URB597 (0.3 mg/kg, i.p.) treatment led to a 42% reversal of hind paw tactile hypersensitivity [42]. It is possible that higher doses of URB597 could yield greater antinociception, even though Karbarz et al. (2009) reported no further increases in FAA with higher doses. In rats with a partial sciatic nerve ligation, no effect was obtained with URB597 (0.3 mg/kg, i.p.), but the same dose was significantly efficacious in rats with a peripheral inflammation [13]. Likewise, significant antinociception and brain FAA levels were obtained in rats with a hind paw inflammation at 3 mg/kg PF-3845 (p.o.) [12]. In FAAH deficient mice, brain AEA levels were about 10-fold higher than that of wild-type mice [53]. Following a sciatic nerve injury, FAAH knockout mice demonstrated marked ipsilateral hind paw hypersensitivity, indicating that increased CNS AEA did not prevent the onset of nerve injury-induced nociception. However, the degree of hind paw carrageenan-induced hypersensitivity in FAAH deficient mice was not as severe as that seen in wild-type mice. While the findings indicate some antinociception following increased CNS FAA in the neuropathic state, the antinociception is not as robust as compared to that obtained in the painful inflammatory state. It is speculated that a SCI that is accompanied by significant inflammation could be sensitive to the effect of FAAH inhibition. Perhaps early treatment, around the time of injury or soon thereafter, will be efficacious.

A potential limitation of the current study is the dose of URB597 used. Perhaps a much greater dose of URB597 would have shown efficacy, even though it appears that there is no further gain in FAA levels in the CNS [42]. In mice with a peripheral nerve injury antinociceptive effects of URB597 were observed at 10 mg/kg (i.p.) and at 50 mg/kg (p.o.) [22], [54]. The antinociceptive effects were attenuated with CB₁ receptor antagonist, indicating mediation through CB₁ receptors. Interestingly, Russo et al. (2007) saw only partial attenuation of the effect of URB597 with CB₂ receptor antagonist treatment whereas Kinsey et al. (2009) observed full attenuation. The different behavioral endpoints and potential off-target activity of a high dose could in part explain the differences between the two findings. Nonetheless, perhaps much higher doses of URB597 in the SCI rats could lead to robust antinociception.

A recent clinical study of the FAAH inhibitor PF-04457845 in osteoarthritis pain [55] showed a lack of analgesic efficacy despite demonstration of significant elevation of FAA in patients' plasma and significant efficacy in preclinical studies in rodent models of non-neuropathic pain [56]. The disconnection between the preclinical findings and clinical outcomes raises issues of translatability, of whether clinical experimental methods in fact parallel preclinical experimental methods and whether preclinical models accurately model the clinical state. On one hand, it is not entirely clear if peripherally measured levels of the biomarker of interest are adequate in order to make predictions on behaviors that are CNS mediated. On the other hand, there may be a tenuous relationship of some preclinical animal models (e.g. the rat hind paw formalin test) to long-standing pathologies such as osteoarthritis. The current study utilized a rat model of neuropathic SCI pain, in which analgesics used for clinical neuropathic SCI pain appear antinociceptive [18]. The current findings suggest that increasing FAA alone may not be an efficacious treatment specifically for below-level SCI neuropathic pain.

While the current study suggests at most a weak antinociceptive effect of FAAH inhibition alone, greater antinociceptive effects could be obtained by combining an FAAH inhibitor with a known analgesic drug, thereby enhancing the efficacy of the analgesic drug [57]. The combination of opioids with AEA leads to a synergistic antinociception [58], [59]. A metabolite of acetaminophen has been shown to enhance synaptic AEA by blocking its transporter [60]. Although acetaminophen alone is not antinociceptive in SCI rats, adding it to analgesic drugs led to synergistic antinociception which was CB₁ receptor-mediated [61]. Combination drug therapy, with FAAH inhibitors at the core, could be of great value for neuropathic SCI pain, for which few effective analgesic treatments exist [9], [62].