Chronic vehicle treatment did not alter the ipsilateral or contralateral paw withdrawal thresholds (PWTs) in sham-operated rats (data not shown). By contrast, SNL rats chronically treated with vehicle had a significantly (P < 0.001) lower ipsilateral PWT, compared to the PWT of the contralateral hindpaw of SNL rats and ipsilateral hindpaw of sham-operated rats (data not shown), consistent with the development of mechanical allodynia in these rats (Fig 1). Chronic daily treatment with minocycline (30 mg/kg, ip for 14 days) starting at 1 hr before SNL surgery significantly reduced the development of mechanical allodynia at days 5, 10 and 14 post-surgery compared to vehicle-treated SNL rats (Fig 1). Chronic minocycline treatment (30 mg/kg, ip) in sham-operated rats did not alter ipsilateral or contralateral PWT (data not shown).

SNL surgery significantly increased OX-42 immunoreactivity in the L4-L6 sections of ipsilateral spinal cord of chronic vehicle-treated rats, compared to the contralateral spinal cord, indicative of spinal microglia activation (Fig 2, OX-42 immunoreactivity, ipsilateral-vehicle, 30.2 ± 0.5; contralateral-vehicle, 17.0 ± 0.3). Chronic minocycline treatment (30 mg/kg, ip) significantly attenuated OX-42 immunoreactivity in the ipsilateral (P < 0.001) and contralateral (P < 0.01) spinal cord of SNL rats, compared to vehicle controls (Fig 2). Thus, effects of minocycline treatment on neuropathic pain behaviour were associated with decreased levels of a marker for activated microglia in the ipsilateral spinal cord of SNL rats. Nevertheless, levels of OX-42 in the ipsilateral spinal cord of SNL rats treated with minocycline were still elevated compared to the contralateral spinal cord of minocycline-treated SNL rats (total: ipsilateral-minocycline, 14.9 ± 1.3; contralateral-minocycline, 6.5 ± 0.7).

In SNL rats with chronic vehicle treatment, levels of AEA were significantly increased (P < 0.01) in the ipsilateral spinal cord, compared to the contralateral spinal cord (Table 1). Following chronic minocycline treatment in SNL rats, levels of AEA in the ipsilateral spinal cord were also elevated, compared to the contralateral spinal cord (Fig 3). The elevated ipsilateral levels of AEA in minocycline-treated rats were not different from the levels in vehicle-treated SNL rats.

There were no significant differences in levels of OEA in the ipsilateral and contralateral spinal cord of vehicle-treated SNL rats (Table 1). Levels of OEA in ipsilateral and contralateral spinal cord of SNL rats were unaltered by minocycline treatment (Fig 3).

By contrast, chronic minocycline treatment of SNL rats had a marked impact on levels of 2-AG and PEA in the ipsilateral spinal cord, compared to the contralateral spinal cord (Fig. 3) and the ipsilateral spinal cord from vehicle-treated animals. In vehicle-treated SNL rats, levels of PEA were decreased (P < 0.01) in the ipsilateral spinal cord, compared to contralateral spinal cord (Table 1). However levels of PEA were increased (P < 0.01) in the ipsilateral spinal cord (31.59 ± 5.61 nmol/g) of minocycline-treated SNL rats compared to their contralateral spinal cord (5.01 ± 0.48 nmol/g) (Fig 3). Although there were no significant differences in levels of 2-AG in the ipsilateral and contralateral spinal cord of vehicle-treated SNL rats (Table 1), levels of 2-AG were significantly decreased (P < 0.01) in the ipsilateral spinal cord (4.53 ± 0.93 nmol/g) compared to contralateral spinal cord (47.66 ± 4.38 nmol/g) of minocycline treated SNL rats (Fig 3).

Chronic minocycline treatment of sham-operated rats did not significantly alter levels of AEA, OEA or PEA in the ipsilateral spinal cord, compared to the contralateral spinal cord (Fig 3). By contrast, levels of 2-AG in the ipsilateral spinal cord of minocycline treated sham-operated rats were significantly (P < 0.05) lower than levels in the contralateral spinal cord (Fig 3; ipsilateral: 39.79 ± 4.38 nmol/g, contralateral: 62.61 ± 7.91 nmol/g).

Levels of ECs and related compounds in the contralateral spinal cord of SNL rats treated with minocycline were compared to levels in the contralateral spinal cord of SNL rats treated with vehicle. Although levels of AEA in the contralateral spinal cord of minocycline treated SNL rats were not significantly different to levels in the contralateral spinal cord of vehicle-treated SNL rats, there was a trend towards an increase in the minocycline treated group (53.72 ± 2.33 pmol/g vs 28.42 ± 0.86 pmol/g respectively). Levels of OEA in the contralateral spinal cord of minocycline-treated rats were comparable to levels in the contralateral spinal cord of vehicle-treated rats (1.51 ± 0.13 nmol/g vs 1.43 ± 0.08 nmol/g). In contrast to the marked effects of minocycline treatment on levels of PEA and 2-AG in the ipsilateral spinal cord of SNL rats, this treatment did not significantly influence levels of PEA and 2AG in the contralateral spinal cord. There were no significant differences between levels of 2-AG or PEA in contralateral spinal cord of minocycline-treated SNL rats compared to vehicle-treated SNL rats (2-AG, 47.66 ± 4.38 nmol/g vs 54.40 ± 4.02 nmol/g; PEA, 5.01 ± 0.48 nmol/g vs 8.80 ± 0.97 nmol/g).

Our in vivo data demonstrate that neuropathic pain has a differential effect on the levels of AEA and PEA in the spinal cord, which may reflect the presence of microglia in the spinal cord. Here, we investigated the hypothesis that microglia have a differential effect on the hydrolysis of AEA and PEA, which may account for our in vivo observations in neuropathic rats. We report marked differences in the time courses for PEA and AEA hydrolysis in intact BV-2 microglial cells (Fig 4.) Specifically, the capacity for hydrolysis of AEA by the BV-2 cells was limited, in that, after 10 minutes, there was no further metabolism of AEA, as shown by the plateau in accumulation of water-soluble product. In marked contrast, the hydrolysis of PEA continued for at least 30 minutes (Fig 4).

The spinal nerve ligation model of neuropathic pain was used in this study. Spinal nerves L5-L6 were ligated according to the procedures described by 42. Male Sprague Dawley rats (101 – 135 g) were anaesthetised using isoflurane (3% induction, 1–1.5% maintenance in 33% O₂/67% N₂O) and placed in a prone position. A midline incision was made at the L3-S2 level and the left paraspinal muscles at L4-S2 level were separated from spinal processes. Part of the L6 transverse process was removed with fine rongeurs and the L4-L6 nerves identified. The L5-L6 spinal nerves were isolated and tightly ligated (one suture per nerve) distal to the dorsal root ganglia and proximal to the sciatic nerve formation with 6-0 silk. The wound was closed in two layers using absorbable sutures and wound clips, following complete haemostasis. A similar procedure was performed for the sham surgery, except spinal nerves were not ligated. Post-surgery, the sham-operated and the SNL rats were group-housed (5 rats per cage) and their posture and behaviour were closely monitored for 48 hours.

Effects of chronic intraperitoneal (ip) injections of minocycline (Sigma, UK, 30 mg/kg, n = 16) or vehicle (2 ml/kg sterile water, n = 16) on mechanical allodynia were measured over 14 days post-surgery in sham-operated and SNL rats. Minocycline (30 mg/kg, ip) or vehicle was injected into rats 1 hour before sham- or SNL surgery and subsequently daily for 14 days, approximately 15 hours prior to behavioural testing.

From post-operative day 1 onwards (days 1, 3, 5, 7, 10 and 14), behavioural testing was performed to assess the development of mechanical allodynia, all testing was carried out between 8 and 11 am. Rats were placed in perspex cubicles with wire mesh grid floors and allowed to acclimatise prior to behavioural testing. Mechanical sensitivity of the ipsilateral and contralateral hind paw was assessed by measuring the paw withdrawal threshold (PWT) at which foot withdrawal to normally innocuous mechanical punctate stimuli was observed. Stimuli were delivered, from below, to the plantar surface of the foot using 0.2, 2, 4, 6, 8, 10 and 15 g von Frey hair stimuli for a maximum of 5 seconds per application. Each trial consisted of the application of a single von Frey hair, starting with a 4 g stimulus, and either increasing or decreasing in intensity until the PWT was observed, with the stimulus either side of the withdrawal weight subsequently repeated. The paw withdrawal threshold was recorded as the lowest von Frey stimulus to elicit a response.

On post-operative day 15, approximately 15 hours after the last injection of minocycline (30 mg/kg, ip) or vehicle, rats were decapitated prior to a standardised and rapid dissection of the ipsilateral and contralateral lumbar (L4-L6) spinal cord. The whole process was complete within 3 minutes and samples were placed immediately on dry ice and then stored at -80°C to minimise post-mortem changes in levels of endocannabinoids (ECs) and related compounds. A validated lipid extraction method 43 was employed for measurement of AEA and 2-AG and related compounds, PEA and N-oleoylethanolamine (OEA). In brief, tissue was homogenised in an ethyl acetate/hexane mixture with internal standards (0.42 nmol d8-AEA, 1 nmol d8-2-AG), followed by repeated centrifugation and supernatant layer collection stages. Solid phase extraction was subsequently performed to purify samples. Simultaneous measurement of ECs and related compounds was then performed using liquid chromatography-tandem mass spectrometry (LC-MS/MS). Analysis was carried out on an Agilent 1100 system (Agilent Technologies, Waldbrunn, Germany) coupled to a triple quadrupole Quattro Ultima MS (Waters, Manchester, UK) recording in electrospray positive mode. Analytes were separated chromatographically on a HyPurity Advance C8 column and pre-column (100 × 2.1 mm internal diameter, 3 μm particle size; Thermo Fisher Scientific, Runcorn, UK) with a mobile phase flow rate of 0.3 ml/min. A gradient elution was used, with mobile phases consisting of A (water, 1 g/L ammonium acetate, 0.1% formic acid) and B (acetonitrile, 1 g/L ammonium acetate, 0.1% formic acid). Samples were injected from a cooled auto sampler maintained at 4°C. Multiple reaction monitoring of individual compounds, using specific precursor and product mass to charge (m/z) ratios allowed simultaneous measurement of AEA, 2-AG, PEA and OEA.

In a separate series of experiments, groups of vehicle-treated SNL rats or minocycline-treated SNL rats (n = 4 per group) were prepared for immunohistochemical studies of microglia activation. The development of mechanical allodynia in vehicle-treated SNL rats and the effect of minocycline treatment on mechanical allodynia were consistent with behavioural changes observed in vehicle-treated SNL and minocycline-treated SNL rats employed for measurement of levels of ECs. On post-operative day 15, rats were transcardially perfused with saline (0.9% w/v sodium chloride) followed by paraformaldehyde (PFA, Sigma, UK, 4% w/v) in phosphate-buffered saline (PBS, 3.2 mM Na₂HPO₄, 0.5 mM KH₂PO₄, 1.3 mM KCl, 135 mM NaCl, pH 7.4) under terminal anaesthesia induced by pentobarbital sodium (Euthetal^®, Merial Animal Health Ltd, UK, 800 mg/kg, ip). The spinal cord was dissected and stored in PFA for approximately 2 hrs at 4°C, and then transferred to a cryoprotectant solution (30% w/v sucrose, 0.1% w/v sodium azide in PBS).

Spinal cords were frozen in O.C.T. embedding medium (VWR, UK) and 14 μm serial sections were cut onto Superfrost Plus slides (VWR) using a cryostat (CM3050S, Leica Microsystems, UK Ltd). Sections were air dried and then treated with 3% hydrogen peroxide in PBS for 30 minutes at room temperature, washed with PBS and then blocked with 5% normal goat serum (Sigma, UK) in PBS/0.2% Triton X-100. They were then incubated overnight with mouse monoclonal anti-OX-42 (AbD Serotec, UK), diluted 1:1000 in PBS/0.2% Triton X-100. This antibody recognizes the complement receptor type 3(CR3) and specifically labels the plasma membrane of the microglial cells. Sections were washed in PBS and incubated with a biotinylated goat anti-mouse IgG secondary (Vector Laboratories, UK) diluted 1:500 in PBS/0.2% Triton X-100 for 2 hours at RT. They were washed again, treated with avidin-biotin complex (Vector Laboratories, UK) according to the manufacturer's instructions, rinsed again and incubated with diaminobenzidine/nickel (Vector Laboratories, UK). The reaction was stopped with PBS and slides were dehydrated and coverslipped. Sections obtained from the eight rats were processed at the same time, and the experiments were repeated three times. Images were acquired using a Leica DMRA2 microscope (Leica, Nussloch, Germany), and digital images were captured using a Retiga 1300 12-bit camera (QImaging, Burnaby, BC, Canada) and QCapture 1.1.6 software (QImaging) running on a Macintosh G4 computer. Images were then assembled into figures using Photoshop 7.0 (Adobe Systems, San Jose, CA). OX-42 optical density was analysed at the level of L4, L5 and L6. To quantify, we applied Image J Software based analysis: nine equivalent boxes (areas of interest, measuring 180 × 180 squared μm) were placed within each dorsal horn covering the medial to lateral extent of Rexed laminae I-V, and the mean grey saturation measured within each area of interest. The mean value of three equivalent boxes outside each section was then subtracted to normalize background. Three sections were analysed at each level. Results (mean ± SEM) are expressed as arbitrary units.

The AEA/PEA hydrolysis assay was based on Paylor et al. 44. In brief, BV2 mouse microglial cells (a generous gift from N. Stella, University of Washington), passage number <8, grown to confluency in 24 well plates were pre-incubated for 10 min at 37°C in 400 μl HEPES-containing physiological buffer, pH 8.0, containing 0.1% fatty acid-free bovine serum albumin. The reaction was initiated by the addition of 250 nM AEA or PEA containing [³H]-AEA or [³H]-PEA (approx 60,000 dpm/well), respectively. After varying periods of incubation (2 to 30 minutes) at 37°C, the buffer was rapidly removed by aspiration and replaced with 400 μl ice-cold methanol. Wells were scraped and the entire contents of each well collected. Following additions of 400 μl chloroform and 200 μl H₂O, the phases were separated by centrifugation at 5,000 g for 2 minutes. Radioactivity in an aliquot of the aqueous, upper phase containing [³H]-ethanolamine was quantified by liquid scintillation spectroscopy.

Behavioural data were statistically analysed using two-way ANOVA with treatment and time as factors, followed by Bonferroni's post-hoc test, and are presented as mean ± SEM of mechanical PWT in grams. Levels of ECs and related compounds were calculated as the ratio of the EC peak area with its internal standard on the chromatogram and are expressed as mol/g wet weight of tissue. Levels of ECs and related compounds were excluded from analysis if the value was greater than 2 standard deviations from the mean, giving n = 6 for all groups except: n = 5 for contralateral vehicle-saline and ipsilateral vehicle-sham. Ipsilateral and contralateral spinal cord levels of AEA, 2-AG, OEA and PEA were statistically compared using a Mann-Whitney test and are presented as mean % of contralateral spinal cord values, or raw mean ± SEM). For comparison of levels of AEA, 2-AG, OEA and PEA between treatment groups, data were statistically compared using Kruskal-Wallis test. Immunohistochemical data were analysed using the Kruskal-Wallis non-parametric test and are presented as mean ± SEM of density of OX-42 staining. Hydrolysis of AEA & PEA was expressed as pmoles/well and was analysed as means ± SEM from six different cell passages investigated in triplicate.