A Novel Neuroimmune Mechanism in Cannabinoid-mediated Attenuation of Nerve Growth Factor–induced Hyperalgesia

Anandamide and palmitoylethanolamide were purchased from Tocris Cookson, Bristol, United Kingdom. Anandamide was presented as a suspension in soya emulsion at a concentration of 10 mg/ml. A control soya emulsion was a gift of Dr. Clive Washington, Ph.D., Lecturer in Pharmaceutical Sciences, School of Pharmaceutical Sciences, University of Nottingham, Nottingham, United Kingdom. Palmitoylethanolamide was dissolved in dimethyl sulfoxide (Sigma, Poole, United Kingdom) saline (2:3 ratio) to a concentration of 5 mg/ml. SR141716A was dissolved in dimethyl sulfoxide/saline (2:3 ratio) to a concentration of 1 mg/ml. SR144528 was also dissolved in dimethyl sulfoxide/saline (2:3 ratio) to a concentration of 1 mg/ml. Capsazepine was dissolved in 0.9% NaCl to a concentration of 100 μg/ml (Sigma). Hexadecyltrimethylammonium bromide was obtained from Sigma, and K-blue substrate was purchased from Neogen Corporation (Lexington, KY). OCT embedding compound and Superfrost slides were obtained from BDH Laboratory Supplies (Poole, United Kingdom). Toluidine blue was supplied by Sigma and made up to a 0.5% solution by dissolution in distilled water.

In total, 159 male Wistar rats were used (weight: 180–315 g). Eighty-five animals were used in the behavioral experiments, and 74 were used in the investigations into neutrophil accumulation. All experiments conformed to British Home Office regulations.

Hind limb withdrawal latencies to a noxious thermal stimulus were measured to demonstrate a thermal hyperalgesia. The latency of limb withdrawal was measured using a Hargreaves device (Ugo Basile, Varese, Italy). 27A beam of infrared radiation at a wavelength of 50 nm was applied to the plantar surface of the paw. To prevent tissue damage, the maximum temperature delivered is 46°C and the maximum time of application is 21.4 s. The system automatically cuts out when the animal withdraws its limb, and a latency (in seconds) is displayed. The animals were acclimatized to the testing environment (Plexiglas box 23 × 18 × 14 cm mounted on an infrared lucent dry glass pane) before baseline limb withdrawal latencies of left and right hind paws were measured. For these and all subsequent measurements, the mean of three tests was taken as the withdrawal latency. There was a delay of at least 1 min before retesting the same animal and 3 min between testing the same paw. Further latencies were measured at 15, 30, 60, 120, and 180 min after intraplantar injection of NGF and were all made by a single blinded observer. Latencies were expressed as difference from baseline; thus, a negative latency is indicative of a relative hyperalgesia.

After baseline testing, animals were randomly assigned to treatment groups, with each comprising five animals. All intraplantar injections were administered via  a 26-gauge needle to the mid-plantar aspect of the left hind paw while animals were gently restrained.

Forty-five animals were randomized into nine groups. Six groups of animals received a careful intraplantar injection of 0.05 ml NGF (20 μg/ml), immediately followed by intraperitoneal treatments. Animals were treated with either soya emulsion (vehicle for anandamide) and 10 or 25 mg/kg anandamide or with dimethyl sulfoxide/saline at a 2:3 ratio (vehicle for palmitoylethanolamide) and 10 or 25 mg/kg palmitoylethanolamide. A final control group of animals received intraperitoneal 0.9% NaCl. The volume of control solution given was equivalent to the volume of the higher dose of agonist.

To ascertain the contribution of locomotor effects of the endocannabinoids on withdrawal latencies, two control groups were examined, which received 0.05 ml NaCl, 0.9%, via  intraplantar injection to the plantar surface of the left hind paw via  a 26-gauge needle and intraperitoneal administration of 25 mg/kg anandamide or palmitoylethanolamide.

A further 40 animals were also randomized into groups; after baseline testing, they also received a careful intraplantar injection of 0.05 ml NGF (20 μg/ml), immediately followed by intraperitoneal treatments. Anandamide at a dose of 25 mg/kg was coadministered with 1 mg/kg of the CB¹receptor antagonist SR141716A or 1 mg/kg of the CB²receptor antagonist SR144528, or 25 mg/kg palmitoylethanolamide was coadministered with 1 mg/kg SR141716A or SR144528.

To investigate the contribution of the vanilloid TRPV1 receptor (formerly VR1) on any putative action of anandamide, 25 mg/kg anandamide was coadministered with the selective TRPV1 antagonist capsazepine at a dose of 0.1 mg/kg. This dose was chosen because it has been shown to antagonize hyperalgesia associated with capsaicin administration effectively. 28To identify any intrinsic activity of capsazepine in this model, the same dose was administered alone to another group of animals that had received NGF.

Finally, to assess intrinsic inverse agonist activity of the CB¹and CB²receptor antagonists, these agents were given alone to two additional groups. Intraplantar NGF was administered as described previously, followed by intraperitoneal administration of 1 mg/kg SR141716A or SR144528.

All latencies were expressed in seconds as difference from baseline. Therefore, negative values denoted a relative hyperalgesia. Differences of treatment groups compared with control were assessed across all time points using a two-way ANOVA (post hoc  Dunnett). To appraise the influence of anandamide or palmitoylethanolamide on the withdrawal threshold per se  in the groups that received intraplantar 0.9% NaCl, withdrawal latencies at each time point were compared with baseline by a one-way ANOVA (post hoc  Dunnett). Significance was taken at P < 0.05.

The majority of tissue myeloperoxidase originates from neutrophils, and its measurement is commonly used to quantify neutrophil accumulation (inter alia ). 29The concentration of myeloperoxidase correlates well with other measures of neutrophil cell accumulation. 30Because myeloperoxidase activity is proportional to the concentration of the indicator (K blue substrate) and the amount of indicator is proportional to the absorbance, absorbance was taken as the measurement of myeloperoxidase activity and, hence, neutrophil accumulation.

Seventy male Wistar rats were randomized into 14 groups (n = 5 in each group). All animals were anesthetized with isoflurane, 0.5–1%, in N²0 and oxygen (70:30 ratio), and the dorsum of the left hind paw was injected with 0.05 ml NGF (20 μg/ml in 0.9% NaCl). Injection in the dorsum of the paw compared with the plantar surface allowed more accurate subsequent removal of the area of injected skin. Furthermore, the skin at this location is of more uniform thickness and delineates an exact plane of tissue dissection, ensuring complete harvest of the tissue. Treatments (intraperitoneal) were administrated immediately after the injection of NGF. Three control groups were examined and received dimethyl sulfoxide/saline at a 2:3 ratio, soya emulsion, or 0.9% NaCl in equivalent volumes to their active treatment counterparts. Anandamide was administered at doses of 10 and 25 mg/kg, and 25 mg/kg was coadministered with 1 mg/kg SR141716A or 1 mg/kg SR144528. Similarly, palmitoylethanolamide was administered at doses of 10 and 25 mg/kg, and 25 mg/kg palmitoylethanolamide was coadministered with 1 mg/kg SR141716A or 1 mg/kg SR144528. SR141716A and SR144528 were also given alone at a dose of 1 mg/kg. A final group received 0.1 mg/kg capsazepine.

Three hours after administration of NGF, animals were terminally reanesthetized with 1.5 g/kg urethane (intraperitoneal) and transcardially perfused with 100 ml NaCl, 0.9%, at 4°C. Skin from the dorsum of both hind paws was removed, and a 9-mm diameter punch biopsy was taken before being frozen and stored at −20°C until use.

Skin samples were chopped, added to a 5-ml aliquot container, and homogenized (Kinematica Polytron homogenizer; Phillip Harris Scientific, Staffordshire, United Kingdom) for 1 min in 1 ml 50-mm phosphate buffer containing 0.5% hexadecyltrimethylammonium bromide (detergent buffer) to which 0.9% NaCl was added. Buffer at a dose of 1 ml was used to ensure recovery of homogenate from the distal end of the homogenizer. Cell debris was removed by centrifugation (20000 g  for 20 min at 4°C), and the supernatant was frozen at −20°C until required.

After thawing, 100 μl skin homogenate (myeloperoxidase source) was added to 100 μl phosphate buffer containing 0.5% hexadecyltrimethylammonium bromide and 400 μl K-blue substrate (stabilized 3,3′ and 5,5′ tetramethylbenzidine and H²O²; Neogen Corporation) at room temperature. K blue substrate develops a deep blue color in the presence of myeloperoxidase; this indicator is detected by absorbance measured by spectrometry. The optical density at 620 nm (proportional to the concentration of myeloperoxidase in the given volume of homogenate) was measured at 25 min (Unicam 5625 Spectrometer; ATI Unicam, Cambridge, United Kingdom). The optical density reading at 25 min was taken as the endpoint.

To provide representative images of neutrophil accumulation in skin after palmitoylethanolamide treatments, skin from four separate male Wistar rats that had received either dimethyl sulfoxide/saline control or 25 mg/kg palmitoylethanolamide or 25 mg/kg palmitoylethanolamide with either 1 mg/kg SR141716A or SR144528 was prepared as previously described. Skin was postfixed in 4% paraformaldehyde in 0.1 m phosphate buffer for 2 h and stored in 20% sucrose in 0.1 m phosphate buffer for 12 h before being frozen in OCT embedding compound and stored at −70°C. Sections 20 μm thick were cut and thaw mounted on slides before exposure to the vital stain, 0.5% toluidine blue, and subsequent dehydration with graded alcohols. Images were captured by a Hammamatsu 3CCD C5810 camera (Hammamatsu, Hammamatsu City, Japan).

The final absorbances or optical densities (at 25 min) of the various treatment groups were compared with those of control (one-way ANOVA, post hoc  Dunnett). Significance was taken at P < 0.05.

Intraplantar injection of 1 μg NGF reduced the mean withdrawal latency to a noxious thermal stimulus, consistent with a thermal hyperalgesia. There was no significant difference between vehicle control groups of soya emulsion (vehicle for anandamide groups), dimethyl sulfoxide/saline at a 2:3 ratio (vehicle for palmitoylethanolamide groups), or 0.9% NaCl (vehicle for capsazepine groups) over all times (two-way ANOVA). Therefore, these control data were pooled into a single control group (fig. 1).

Administration of 25 mg/kg anandamide effectively attenuated the NGF-associated hyperalgesia, demonstrated by the mean differences in limb withdrawal latency from baseline being close to 0 (fig. 2A). Over all time points, there was a significant difference between control and 25 mg/kg anandamide treatment (two-way ANOVA, post hoc  Dunnett). The lower dose of 10 mg/kg was not significantly different from control (fig. 2A; two-way ANOVA).

The higher dose of palmitoylethanolamide (25 mg/kg) prevented NGF-induced reduction of withdrawal latency (fig. 2B). Considered across all time points, the palmitoylethanolamide-treated group was significantly different from control (two-way ANOVA, post hoc  Dunnett). At a dose of 10 mg/kg, palmitoylethanolamide also significantly reduced this hyperalgesia (fig. 2B; two-way ANOVA, post hoc  Dunnett).

Coadministration of 25 mg/kg anandamide with 1 mg/kg CB¹receptor antagonist SR141716A prevented the antihyperalgesic effect of anandamide (fig. 3A). This treatment group was not significantly different from control (two-way ANOVA). Coadministration of 25 mg/kg anandamide with 1 mg/kg CB²receptor antagonist SR144528 did not alter the antihyperalgesic effect of anandamide (fig. 3B). This treatment group was significantly different from control (two-way ANOVA, post hoc  Dunnett).

The administration of 1 mg/kg SR171614A did not alter the antihyperalgesic effect of 25 mg/kg palmitoylethanolamide (fig. 4A). This group was also different from control (two-way ANOVA, post hoc  Dunnett). Conversely, coadministration of 1 mg/kg SR144528 abrogated the antihyperalgesic effect of palmitoylethanolamide, and this group was not different from control (fig. 4B; two-way ANOVA).

To show that the action of anandamide and palmitoylethanolamide at these doses on withdrawal latency is not influenced by factors other than an inflammatory hyperalgesia, the higher dose of these agonists was administered in the absence of inflammation. Instead of NGF, an equal volume of 0.9% NaCl was injected into the plantar surface of the hind paw. Neither 25 mg/kg anandamide nor 25 mg/kg palmitoylethanolamide altered the mean difference in withdrawal latency compared with baseline (data not shown, one-way ANOVA).

To assess a putative TRPV1 receptor–mediated antihyperalgesic action of anandamide, the TRPV1 antagonist capsazepine (0.1 mg/kg) was coadministered with 25 mg/kg anandamide. This group was significantly different from control, demonstrating that the antihyperalgesic action of anandamide was not significantly reduced by capsazepine. (fig. 5A; two-way ANOVA, post hoc  Dunnett). Across all time points, capsazepine alone revealed no significant difference from control (fig. 5B; two-way ANOVA); however, there was a trend toward a reduction in the relative NGF-induced hyperalgesia with time.

Administration of 1 mg/kg SR141716A did not significantly affect an NGF-induced hyperalgesia compared with controls (data not shown, two-way ANOVA). After 1 mg/kg SR144528, there was a trend toward an attenuation of NGF-induced reduction in withdrawal latencies, but this did not reach statistical significance (data not shown, two-way ANOVA).

Three hours after administration of NGF, punch biopsies were taken. This corresponds to the time of the last measurement in the behavioral experiment. Using tissue homogenates taken from the ipsilateral paw (into which the NGF had been administered) in the vehicle control groups, mean absorbance (SEM) at 620 nm increased with time, reflecting increased myeloperoxidase concentration proportional to neutrophil accumulation. The endpoint was taken as 25 min; at that point, absorbance was 1.84 (0.31) for dimethyl sulfoxide/saline, 2.92 (0.40) for the soya control, and 2.26 (0.72) for 0.9% NaCl groups (data not shown). There was no significant difference between groups, and the data were pooled. The corresponding contralateral absorbencies were 0.052 (0.017), 0.012 (0.003), and 0.040 (0.006), respectively.

Administration of 10 and 25 mg/kg anandamide resulted in a mean (SEM) final absorbance of 2.61 (0.9) and 1.46 (0.37), respectively, and neither result was significantly different from control (fig. 6A; ANOVA). Administration of 25 and 10 mg/kg palmitoylethanolamide resulted in a final absorbance of 0.87 (0.13) and 1.00 (0.25), respectively, which were both significantly different from control (fig. 6B; ANOVA, post hoc  Dunnett).

Coadministration of 25 mg/kg anandamide and 1 mg/kg SR141716A or SR144528 resulted in absorbencies that were not significantly different from control (fig. 7A; ANOVA). The reduction in neutrophil accumulation after 25 mg/kg palmitoylethanolamide was unaffected by coadministration with 1 mg/kg SR141716A (significantly different from control, ANOVA, post hoc  Dunnett) yet was abolished by coadministration of SR144528 (fig. 7B; not significantly different from control, ANOVA). These data are indicative that the antineutrophil accumulation effect of palmitoylethanolamide is mediated by an SR144528 reversible mechanism. Figure 8presents representative sections taken from ipsilateral skin sections.

Administration of the receptor antagonists alone was not associated with a change in absorbance compared with control (data not shown, ANOVA).

The administration of 0.1 mg/kg capsazepine results in an absorbance not significantly different from control (data not shown, t  test). Capsazepine does not influence neutrophil accumulation. Nevertheless, comparison with behavioral data taken from figure 5Bat the same time point as the tissue for the myeloperoxidase assay was taken (180 min) shows that the same treatment causes a significant reduction in hyperalgesia (data not shown; P < 0.05, t  test). Thus, capsazepine causes a reduction in hyperalgesia (at 180 min) that is independent of neutrophil accumulation.

Intraplantar administration of 1 μg NGF resulted in a thermal hyperalgesia that was attenuated by anandamide and palmitoylethanolamide. Anandamide was associated with antihyperalgesic effects at a dose of 25 mg/kg, whereas 10 mg/kg was not significantly different from control. The antihyperalgesic action of anandamide was abrogated by the CB¹receptor antagonist SR141716A, indicating a CB¹receptor–mediated mechanism. Recent investigation has discovered activity of anandamide at the TRPV1 receptor. 31Here, the antihyperalgesic action of anandamide was unaffected by the TRPV1 antagonist capsazepine, which does not support a TRPV1 receptor–mediated action for anandamide in this scenario.

An antihyperalgesic action of palmitoylethanolamide was also demonstrated, and 10 and 25 mg/kg doses were effective. The CB²receptor antagonist SR144528 attenuated a palmitoylethanolamide-induced antihyperalgesia, yet the lack of affinity of palmitoylethanolamide at the CB²receptor requires further explanation as to the nature of its action. Administration of SR141716A and SR144528 failed to show any intrinsic activity.

NGF stimulated local neutrophil accumulation as measured by increased myeloperoxidase activity. Anandamide at a dose of 25 mg/kg was associated with a nonsignificant reduction in myeloperoxidase activity, whereas palmitoylethanolamide effectively reduced myeloperoxidase activity and thus neutrophil accumulation. This action was SR144528 sensitive. Myeloperoxidase activity was not affected by administration of either antagonist alone. Therefore, anandamide and palmitoylethanolamide attenuate an NGF-induced thermal hyperalgesia. The action of anandamide is independent of neutrophil accumulation and is mediated via  the CB¹receptor. Conversely, the antihyperalgesic effect of palmitoylethanolamide is associated with a reduction in neutrophil accumulation. These actions of palmitoylethanolamide are sensitive to the CB²receptor antagonist SR144528. These data show two separate mechanisms of a cannabinoid-induced antihyperalgesia: a CB¹receptor–mediated action of anandamide that is independent of an antiinflammatory effect or attenuation of neutrophil accumulation and the antihyperalgesic action of palmitoylethanolamide that is associated with antiinflammatory effects and inhibition of neutrophil accumulation. Although peripheral actions of cannabinoids have been previously suggested, we now show an in vivo  antihyperalgesic action of palmitoylethanolamide that is mediated by modulation of a peripheral neuroimmune inhibition of neutrophil accumulation.

NGF has direct and indirect prohyperalgesic actions. NGF can directly sensitize primary afferent nociceptors as well as inducing a de novo  sensitivity. 11NGF can also regulate sensitivity of neurones to other algogenic systems such as bradykinin 12and the vanilloid TRPV1 receptor. 32Prohyperalgesic actions of NGF also involve neuroimmune interactions, notably NGF-induced mast cell degranulation that liberates more proinflammatory mediators, including NGF 13(fig. 9). These mechanisms contribute to the development of an NGF-induced hyperalgesia, because pretreatment with the mast cell degranulating compound 48/80 attenuates an NGF-induced hyperalgesia at least in part. 8,10Neutrophil accumulation has been demonstrated to be integral to the development of an NGF-induced hyperalgesia. 9Hyperalgesia is directly related to the degree of neutrophil accumulation, and after removal of neutrophils from the circulation, NGF fails to evoke hyperalgesia. 9It is unclear as to why a direct action of NGF on nociceptor sensitization does not incur hyperalgesia regardless of the lack of neutrophils. Although NGF can sensitize nociceptors directly and indirectly, direct action may not be enough to induce a thermal hyperalgesia, which requires neutrophil accumulation. Mast cell degranulation is implicated as a major impetus to neutrophil accumulation. Neutrophil accumulation after immune complex–induced peritonitis is reduced by 50% in mast cell–deficient mice. 33A number of compounds released from mast cells may contribute to neutrophil accumulation, including NGF 34and lipoxygenase products of arachidonic acid metabolism such as leukotriene B⁴. 35Indeed, peritonitis-induced peritoneal exudates in mast cell-deficient mice contained half of the leukotriene B⁴contained in normal animals. 33Nevertheless, 50% of neutrophil accumulation seems to be independent of mast cell action. Local administration of NGF results in a thermal hyperalgesia that is associated with increased leukotriene B⁴concentration. 7Both hyperalgesia and the increase in leukotriene B⁴are attenuated by inhibition of 5-lipoxygenase. 7Similarly, 5-lipoxygenase inhibition significantly reduces the thermal hyperalgesia and neutrophil accumulation induced by NGF. 9Carrageenan-induced neutrophil accumulation (which is NGF driven) is unresponsive to nonsteroidal antiinflammatory drugs such as indomethacin. 36The efficacy of compounds in reduction of the corresponding outcomes in this experimental paradigm would offer a novel antiinflammatory therapy.

The release of further NGF from mast cell degranulation provides a powerful proinflammatory amplification mechanism that would be damaging if it were unchecked. Cannabinoids have been cited as a possible mitigator of this amplification. 37Indeed, exogenous administration of cannabinoids has been shown to be antihyperalgesic in several disparate models of inflammatory pain in which NGF has been shown to play a pivotal role. 23,26Evidence for the existence of CB¹receptors in the periphery on primary afferent neurones as well as in the spinal cord offers an a priori  reason for cannabinoid analgesic potential. Indeed, the gene encoding the CB¹receptor exists within NGF-dependent primary afferent neurones. 38Moreover, the expression of cannabinoid receptors by cells involved in neuroimmune interactions employed in an inflammatory hyperalgesia, such as the mast cell 24and neutrophil, 25supports the premise that inflammatory pain may be especially sensitive to actions of cannabinoids. 39Peripheral administration of anandamide attenuates a carrageenan-induced hyperalgesia in rats 23and the first phase of the formalin test in mice, actions that are both mediated by the CB¹receptor. 21It is conceivable that the reduction of an NGF-induced hyperalgesia by anandamide utilizes a peripheral mechanism at the level of the primary afferent neuron (fig. 9). Because increases in cyclic adenosine monophosphate are implicated in an inflammation-induced hyperalgesia, 40reduction of intracellular cyclic adenosine monophosphate provides a feasible mechanism for the CB¹receptor–mediated antihyperalgesic actions of anandamide. 39Anandamide could also act at a spinal level, however, because spinally administered anandamide reduces a carrageenan-induced hyperalgesia and reduces capsaicin-induced calcitonin gene–related peptide release in the isolated spinal cord. 41 

Palmitoylethanolamide is also antihyperalgesic, yet this activity is associated with attenuation of neutrophil accumulation. This situation is confounded by ignorance of the receptor at which palmitoylethanolamide exerts its actions. Although the selective CB²receptor antagonist SR144528 mitigates its antihyperalgesic actions, implicating a CB²receptor–mediated action, palmitoylethanolamide does not bind appreciably to the CB²receptor. 42Other workers have also found the antihyperalgesic actions of palmitoylethanolamide to be SR144528 sensitive. 21,43Metabolites of palmitoylethanolamide that have agonist activity at the CB²receptor could explain this apparent discrepancy. Previous studies have suggested that levels of palmitoylethanolamide remain unchanged for 6 h after administration, implying that these actions are not metabolite driven. 44An increase in endogenous CB²receptor agonists by a palmitoylethanolamide-mediated interference with uptake and metabolism (an “entourage” effect) presents another possible explanation, 43but the identity of this putative endogenous CB²agonist is unclear. Alternatively, palmitoylethanolamide acts at a hitherto uncharacterized CB²-like receptor at which SR144528 has antagonist properties. Expression of CB²-like receptors on peripheral capsaicin-insensitive primary afferent neurones has been suggested. 42 

Palmitoylethanolamide significantly reduced NGF-induced neutrophil accumulation. Palmitoylethanolamide could act at any of the processes involved in neutrophil accumulation (fig. 9), although its systemic administration cannot unequivocally claim a peripheral action of its attenuation of peripheral neutrophil accumulation. Palmitoylethanolamide has been previously shown to reduce serotonin release from mast cell–like RBL-2H3 cells, 24and oral administration of palmitoylethanolamide prevented substance P–induced mast cell degranulation quantified by immunocytochemical identification. 16,44This inhibitory action of palmitoylethanolamide on mast cells would be at a potent upstream position and would prevent the inflammatory amplification resulting from further NGF release. Nevertheless, other studies have disputed the ability of cannabinoids to reduce mast cell activation. These discrepancies may be partly explained by the heterogeneity of mast cell populations such as cutaneous and peritoneal types. 43Alternatively, because 50% of neutrophil accumulation is independent of mast cells, palmitoylethanolamide may act directly on neutrophils to inhibit the processes that lead to their influx. For example, mice treated with tetrahydrocannabinol analogs display reduced leukocyte adhesion. 45Palmitoylethanolamide thus seems to be a potent novel antihyperalgesic that acts by modulation of NGF-driven inflammatory processes impervious to other antiinflammatory medication.

Palmitoylethanolamide has antihyperalgesic action mediated in part by reducing the number of neutrophils that contribute to and maintain an NGF-induced hyperalgesia. The mechanism by which neutrophils perform this task is unknown, but products of lipoxygenase metabolism have been implicated (fig. 9). Neutrophils generate a number of hydroperoxyeicosatetraenoic acids, hydroxyeicosatetraenoic acids, and dihydroxyeicosatetraenoic acids. 46Some of these compounds have not only been implicated in hyperalgesia 47but are also active at the TRPV1 receptor. 48The involvement of the TRPV1 receptor in the maintenance of an NGF-induced hyperalgesia is supported here by an antihyperalgesic action of capsazepine 3 h after NGF administration that is independent of neutrophil influx.

Anandamide and palmitoylethanolamide attenuate an NGF-induced hyperalgesia. The action of palmitoylethanolamide is mediated by a reduction in the local neutrophil accumulation that is involved in the generation and maintenance of this hyperalgesia. Such activity is clinically relevant, because currently available nonsteroidal antiinflammatory drugs are ineffective in this model of inflammatory pain. 8Anandamide is likely to act on neurones expressing CB¹receptors. Although the effect of palmitoylethanolamide is sensitive to the CB²receptor antagonist SR144528, the lack of affinity of palmitoylethanolamide for the CB²receptor obscures the site of action. These data suggest a novel antihyperalgesic mechanism of cannabinoids, although one cannot categorically claim a peripheral effect for systemic administration of palmitoylethanolamide. Nevertheless, this reduction of neutrophil accumulation would demonstrate a peripheral site of action that provides a rationale for the development of therapeutic cannabinoids devoid of central psychoactive side effects. Furthermore, the efficacy of palmitoylethanolamide in the reduction of hyperalgesia offers neuroimmune modulation as a potential site for the development of other novel therapeutic agents devoid of central nervous system–mediated side effects.