Neurobiology of Disease 62 (2014) 394–406

Contents lists available at ScienceDirect

Neurobiology of Disease journal homepage: www.elsevier.com/locate/ynbdi

Riluzole attenuates neuropathic pain and enhances functional recovery in a rodent model of cervical spondylotic myelopathy Eun Su Moon a,b,1, Spyridon K. Karadimas a,c,1, Wen-Ru Yu a, James W. Austin a,c, Michael G. Fehlings a,c,d,e,⁎ a

Division of Genetics & Development, Toronto Western Research Institute, and Spinal Program, Krembil Neuroscience Centre, University Health Network, Toronto, Ontario M5T 2S8, Canada Department of Orthopaedic Surgery, Yonsei University College of Medicine, Seoul, Republic of Korea Institute of Medical Sciences, Faculty of Medicine, University of Toronto, Ontario, Canada d Neuroscience Program, University of Toronto, Toronto, Ontario M5S 1A8, Canada e Department of Surgery, Division of Neurosurgery, University of Toronto, Toronto, Ontario M5T 2S8, Canada b c

a r t i c l e

i n f o

Article history: Received 4 September 2012 Revised 4 October 2013 Accepted 22 October 2013 Available online 29 October 2013 Keywords: Cervical spondylotic myelopathy CSM Riluzole Neuropathic pain Spinal cord Spine Rodent model Drug treatment

a b s t r a c t Cervical spondylotic myelopathy (CSM) is the commonest cause of spinal cord impairment worldwide and despite surgical treatment, it is commonly associated with chronic neuropathic pain and neurological impairment. Based on data suggesting a key role of sodium and glutamate mediated cellular injury in models of spinal cord compression, we examined whether riluzole, a sodium channel/glutamate blocker, could improve neurobehavioral outcomes in a rat model of CSM. To produce chronic progressive compression of the cervical spinal cord, we used an established model of graded mechanical cord compromise developed in our laboratory. The chronic (8 weeks) mechanical compression of the cervical spinal cord resulted in persistent mechanical allodynia and thermal hyperalgesia at 8 weeks. Moreover, we found increased expression of phosphorylated NR1 and NR2B in the dorsal horns as well as astrogliosis and increased microglia expression in the dorsal horns after mechanical compression. Following daily systemic administration for 7 weeks after the induction of compression, riluzole (8 mg/kg) significantly attenuated forelimb and hindlimb mechanical allodynia and alleviated thermal hyperalgesia in the tail. Importantly, riluzole led to a decrease in swing phase duration, an increase in hind leg swing speed and an increase paw intensity in gait analysis. Riluzole also decreased the number of phosphorylated NR1 and phosphorylated NR2B positive cells in the dorsal horns and the microglia activation in the dorsal horns. Together, our results indicate that systemic riluzole administration during chronic cervical spinal cord compression is effective at protecting spinal cord tissue, preserving neurobehavioral function and alleviating neuropathic pain, possibly by decreasing NMDA receptor phosphorylation in astrocytes and by eliminating microglia activation. As such, riluzole represents a promising clinical treatment for CSM. © 2013 Elsevier Inc. All rights reserved.

Introduction Cervical spondylotic myelopathy (CSM) is a chronic progressive disorder in which the cervical spinal cord undergoes chronic compressive injury secondary to degenerative disk disease, ossification of the posterior longitudinal ligament (OPLL) or frank intervertebral disk disruption (Baptiste and Fehlings, 2006; Bohlman and Emery, 1988). CSM is the world's most common cause of spinal cord impairment in adults

⁎ Corresponding author at: Gerald and Tootsie Halbert Chair in Neural Repair and Regeneration, Toronto Western Hospital, University Health Network, West Wing, 4th Floor, Room 4W-449, 399 Bathurst Street, Toronto, Ontario M5T 2S8, Canada. Fax: +1 416 603 5298. E-mail address: [email protected] (M.G. Fehlings). URL: http://www.drfehlings.ca/ (M.G. Fehlings). Available online on ScienceDirect (www.sciencedirect.com). 1 These authors have equal contribution. 0969-9961/$ – see front matter © 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.nbd.2013.10.020

(Young, 2000). The gold standard of care for CSM is surgical decompression of the spinal cord. While successful surgical intervention can arrest the progression of disease, most patients are left with significant residual neurological impairment, including weakness, numbness, spasticity and neuropathic pain (Kalsi-Ryan et al., 2013). Neuropathic pain is characterized by allodynia and hyperalgesia which initially arise in the primary zone at the site of injury and, over time, may spread to areas that are not directly damaged by the injury (secondary hyperalgesia). Neuropathic pain has also been characterized in animal models of CSM (Karadimas et al., 2013; Lee et al., 2012). Due to its chronic nature and difficulty in management, neuropathic pain has been regarded as one of the most obstinate clinical symptoms of CSM. Importantly, neuropathic pain in CSM patients remains a major clinical problem and a therapeutic challenge because existing treatments are often ineffective and can cause serious side effects. Glutamate N-methyl-D-aspartate receptors (NMDARs), especially those located in the dorsal horn of the spinal cord, are critically involved

E.S. Moon et al. / Neurobiology of Disease 62 (2014) 394–406

in nociceptive transmission and synaptic plasticity. Moreover, they have long been considered a target for the treatment of neuropathic pain (Gao et al., 2005; Guo et al., 2002; Miki et al., 2002; Zou et al., 2002). NMDARs are ionotropic channels — heteromers composed of NR1 and one or more of the four NMDA2 (NR2A-D) or two NMDA3 (NR3A and B) subunits (Villmann and Becker, 2007). The NR1 subunit is essential for the function of the NMDA receptor and phosphorylation of the NR1 is a major mechanism of modulating channel activity and trafficking to the neuronal surfaces (Chen and Roche, 2007; Chen et al., 2006). In vitro electrophysiologic studies showed that riluzole inhibited NMDA or kainic acid evoked — currents in Xenopus oocytes expressing rat NMDA or kainate glutamate receptors (Debono et al., 1993). Additionally, Azbill et al. combining in vitro and in vivo approaches demonstrated that riluzole increases the glutamate uptake as measured in rat spinal cord synaptosomes (Azbill et al., 2000). In other animal studies, riluzole decreased levels of glutamate in the CSF, possibly by increasing glutamate transporter activity (Bellingham, 2011). Together, this suggests that riluzole has a number of molecular targets involved in the neural mechanism of pain, including SCI-induced pain. Using a clinically relevant model of gradual chronic compression of the cervical spinal cord developed in our laboratory (Lee et al., 2012), we hypothesized that daily systemic administration of riluzole would decrease neuropathic pain and protect against functional loss during chronic spinal cord compression. Previous studies have shown the efficacy of riluzole in a rat model of SCI (Hama and Sagen, 2011), in a rat model of spasticity induced by cutaneous stimulation and in other pain models (Kitzman, 2009; Munro et al., 2007; Sung et al., 2003). However, riluzole has yet to be tested in models of CSM. In the present paper, we demonstrate for the first time that the sodium–glutamate blocker riluzole attenuates neuropathic pain in a model of CSM. Material and methods Animal care A total of 41 female Sprague–Dawley rats (weight 300–400 g, average weight 349 g; Charles River Laboratories, Wilmington, MA) were used for the experiments. All experimental protocols of this study were approved by the animal care committee of the University Health Network to ensure an ethical study, in accordance with the policies established in the guide on the care and use of experimental animals prepared by the Canadian Council of Animal Care.

395

area (C2–T2), and the skin and superficial muscles were retracted. The rats underwent a C6–C7 laminectomy and the rod of the chronic compression device (CCD) was inserted into the C2 and the T2 spinous processes. A threaded screw with an extradural plate fixed to the tip was advanced through the CCD rod. To ensure the stability of the CCD we followed the procedure described by Lee et al. The screw was advanced very precisely 0.2 mm (one half turn) using a microscope at an oblique angle. The sham-operated animals underwent identical surgical procedure but without having cord compression. The surgical wounds were sutured, and the animals were given post-operative analgesia and saline (0.9%; 5 ml) to prevent dehydration. Animals were allowed to recover and housed in standard rat cages with absorbent bedding at a temperature of 26 °C. In addition, the animals received Clavamox (amoxicillin and clavulanic acid). Gradual mild chronic compression starting 1 week post-surgery was achieved by advancement of the screw by 0.4 mm (one turn) weekly up to 3 weeks. Experimental groups and treatments At first week post-operatively, all injured rats were divided blindly and randomly into three experimental groups: 1) sham (no compression, n = 6), (2) control group (HBC in saline injection, n = 18), and (3) riluzole group (riluzole injection, n = 17). A summary of the experimental protocol is depicted in Fig. 1. Micro-computed tomography The extent of the compression was quantitatively evaluated using micro-computed tomography (micro-CT; GE Locus Ultra MicroCT at the STTARR facility, University Health Network). This system offers 150 μm3 resolution, scan times as low as 1 s, and maximum transaxial and longitudinal fields of view of 14 and 10 cm, respectively. Under isoflurane anesthesia, micro-CT was used to image the spinal canal at 4 weeks postsurgery. Based on the acquired mid-sagittal images from micro-CT, the Compression Ratio was calculated using the following formula (Fehlings et al., 1999): Compression Ratio (%) = {1 − 2c / (a + b)} × 100, where ‘c’ is the anteroposterior canal diameter at the level of maximum compression, ‘a’ is the anteroposterior canal diameter at the nearest normal level above the site of compression, and ‘b’ is the anteroposterior canal diameter at the nearest normal level below the level of the site of the compression. The distances were calculated using MicroView software (2D and 3D image view 2.1.2., GE Healthcare, Little Chalfont, U.K.). Neurobehavioral assessments

Drug preparation and administration The experimental design involved random allocation of treatment. Drug concentration was chosen based on the pharmacodynamic and kinetic properties of systemically administered riluzole (R116, Sigma, St. Louis MO, USA). Riluzole was initially dissolved in the media, 30% 2hydroxypropyl-β-cyclodextrin solution (HBC, Sigma, H107), resulting in a concentration of 8 mg/ml. The final dosage is 8 mg/kg, diluted with saline to final volume of 1 ml. The HBC solution without riluzole was used as control. To investigate the neuroprotective effects of riluzole on chronic cervical spinal cord compression, we started daily intraperitoneal administration of 8 mg/kg of riluzole as treatment or HBC solution as control after the onset of cord compression (1 week post-initial surgery) and we terminated the treatment at 8 weeks post-surgery. Surgical procedures & experimental groups Chronic compression For the purpose of this study we exploited a clinically relevant model of CSM which has been recently characterized by our laboratory (Lee et al., 2012). Briefly, under halothane anesthesia (1–2%) and a 1:1 mixture of O2/N2O, the surgical area was shaved and disinfected with 70% ethanol and betadine. A midline incision was made at the cervical

The effect of riluzole on neuropathic pain during chronic compression of the cervical spinal cord was characterized using assessments of mechanical allodynia and thermal hyperalgesia. To measure functional deficits, automated computerized gait assessment was done using the CatWalk system (Hamers et al., 2001; Koopmans et al., 2005). Importantly, the assessment of outcomes was performed by two blinded observers. Assessment of mechanical allodynia by von Frey filament testing Cutaneous sensitivity to innocuous mechanical stimulation of both forepaws and hindpaws was assessed weekly in all rats using a series of filaments of varying thicknesses (von Frey filaments). Filaments were applied to the mid-plantar surface of forepaws and hindpaws, one paw at a time. We used 14 Touched-Test von Frey filaments, number 5–16 (North Coast Medical, Inc., CA, USA) with a regularly calibrated stiffness corresponding to 0.16, 0.4, 0.6, 1.0, 1.4, 2.0, 4.0, 6.0, 8.0, 10, 15, and 26 g. Probing was only performed when the animal's four paws were in contact with the floor. Each probe was applied to the foot until it was bent. A minimum of three withdrawals of the tested paw out of five filament applications was considered a positive response. Filaments were applied in ascending order, and the smallest filament that elicited a positive response was considered the threshold stimulus.

396

E.S. Moon et al. / Neurobiology of Disease 62 (2014) 394–406

Fig. 1. Experimental protocol. Illustration of the compression time points and the experimental timing of microcomputed tomography (micro-CT) and neurobehavioral studies. The animals were divided into three groups: (1) sham (S, no compression, n = 6), (2) control (C, 1.4 mm stenosis, n = 18), with advancement of the screw by 0.4 mm (one turn) weekly up to 3 weeks, and (3) riluzole (R, 1.4 mm stenosis, n = 17), with advancement of the screw by 0.4 mm (one turn) weekly up to 3 weeks. von Frey filament test (VFT) was performed weekly to examine the mechanical allodynia in all the experimental groups. The tail flick (TF) test and CatWalk were used at 8 weeks post-surgery to assess the thermal hyperalgesia and the gait adaptations to neuropathic pain, respectively. At 8 weeks after surgery, all animals were sacrificed. HBC = 2-hydroxypropyl-β-cyclodextrin.

Automated gait analysis (CatWalk) To examine the effects of riluzole on gait, we performed gait analysis using the CatWalk system (Noldus Information Technology, Wageningen, Netherlands). Gait disturbances related to chronic compression were analyzed by collecting data at 8 weeks post-surgery. The CatWalk analysis has been described in detail (Hamers et al., 2001; Koopmans et al., 2005). Briefly, the system consists of a horizontal glass plate and video capturing equipment placed underneath, connected to a computer. In our work, for correct analysis of the gait adaptations, after standardizing the crossing speed, the following criteria concerning walkway crossing were used: (1) the rat needed to cross the walkway, without any interruption and (2) a minimum of three correct crossings per animal were required. Labeling of the footprints of each run was performed by one observer blinded to the treatment groups. With the CatWalk, a vast variety of static and dynamic gait parameters can be measured. In the present study, one static and one dynamic individual paw parameters were used: • Swing phase duration (expressed in seconds): time that the paw is not in contact with the glass floor. • Swing speed (expressed in pixels/second): speed of the paw during the swing • Paw intensity: a measure for the mean pressure exerted by the paw during floor contact. Data analysis was performed with a threshold value of 40 (arbitrary units, a.u., possible range 0–250), i.e. all pixels brighter than 40 were used to generate the mean paw intensity. Assessment of thermal hyperalgesia by the tail flick test The thermal nociceptive response was evaluated in all animals by recording the latency to withdrawal of the tail in response to noxious skin heating (the tail flick test). Because the tail flick test requires animal restriction, we performed it once at 8 weeks post-surgery in order to avoid destabilizing the CCD. In short, the dorsal surface of the tail between 4 and 6 cm from the tip was exposed to a beam of light generated from an automated analgesia meter (IITC Life Science, Woodland Hills, CA). The timer was stopped when the animal flicked its tail away from the beam of light. Tail-flick latency was measured at 5-minute intervals until a stable baseline is obtained over 3 consecutive trials. The mean latency was used as a measure to indicate thermal hyperalgia.

Tissue processing Animal perfusion Animals were deeply anesthetized with sodium pentobarbital (80 mg/kg i.p.) and then perfused transcardially with cold PBS followed by 4% paraformaldeyde (PFA) in PBS, pH 7.4. A 2 cm length of the spinal cord centered at the epicenter was dissected and processed for different procedures as follows. Frozen sections For cryotomy, the spinal cords were postfixed in the perfusing solution plus 10% sucrose overnight at 4 °C, and then cryoprotected in 20% sucrose in PBS for 48 h at 4 °C. Then the spinal cord centered at the compression site was dissected and embedded in mounting media (HistoPrep, Fischer Scientific) on dry ice. Cryostat sections (20 μm) were cut and stored in −80 °C. Histopathology Serial transverse spinal cord sections were cut at a thickness of 20 μm. To assess the potential effects of riluzole on gray matter preservation and scar tissue, we systematically sampled tissue sections in each animal, every 120 μm over a distance of 4000 μm and stained them with the cellular stain hematoxylin–eosin (H&E) and myelinselective pigment LFB. Scar tissue was defined as tissue that exhibited a fibrous, inconsistent tissue matrix. Tissue section areas were obtained using the Cavalieri method of Stereo Investigator (MBF Bioscience, Williston, VT, USA). The percentage of scar tissue for each section was calculated using the following formula: % of scar tissue and cavity of tissue section = area of scar tissue and cavity of tissue section/total area of section. The percentage of the preserved gray matter was calculated using the following formulas: % of preserved gray matter of tissue section = area of gray matter of tissue section/total area of tissue section. Immunohistochemical procedures and image analysis For all immunohistochemical staining, the blocking solution contained 5% nonfat milk, 1% bovine serum albumin, and 0.3% Triton

E.S. Moon et al. / Neurobiology of Disease 62 (2014) 394–406

X-100 in 0.1 M PBS unless otherwise stated. For single immunofluorescence staining, the frozen sections were air dried at room temperature for 10 min and then were washed with PBS for 10 min. Sections were blocked and then incubated overnight at 4 °C with the following primary antibodies: rabbit anti-phosphorylated NMDAR1 (1:100 Cell Signalling Technology) for phosphorylated NR1 (pNR1) subunits, rabbit anti-phosphorylated NMDNR2B (Biovision) for phosphorylated NR2 (pNR2B) subunits, rabbit anti-GFAP (1:1000, Millipore Bioscience Research Reagents), and rabbit anti-Iba-1 (1:1000, Wako) for microglia. The slides were washed in PBS three times and then incubated with fluorescent Alexa 568 goat anti-mouse and 488 goat-anti-rabbit secondary antibody (Invitrogen 1:400), respectively for 1 h. The slides were coverslipped with Mowiol mounting medium containing DAPI to counterstain the nuclei. For quadruplicate immunofluorescence labeling, the following primary antibodies were applied together on the sections at 4 °C overnight: rabbit anti-phosphorylated NMDAR1 (1:100 Cell Signalling Technology) subunits and mouse anti-CD11b (Chemicon; 1:100). The following day, slides were incubated with fluorescent Alexa 647 goat anti-rabbit and 350 goat-anti-mouse secondary antibody (Invitrogen 1:400), respectively for 1 h. Then the mouse anti-GFAP (GA5) Alexa Fluor (R) 555 conjugate (1:100, Cell Signaling) and the mouse antineuronal nucleus (NeuN) Alexa Fluor (R) 488 conjugate (1:100; Millipore Bioscience Research Reagents) antibodies were applied for 2 h. Finally, the slides were coverslipped with Mowiol mounting medium. Images were taken using a Zeiss 510 laser confocal microscope or Leica epifluorescence microscope. Phosphorylated NMDA subunits have been linked with the development and the maintenance of the neuropathic pain (Gao et al., 2005; Guo et al., 2002; Miki et al., 2002; Zou et al., 2002). To determine changes in pNR1and pNR2B expression as well as the accumulation of

397

microglia cells in the dorsal horns, pNR1, pNR2B and Iba-1 labeled cells in the dorsal horns (lamina i to v) were counted in 5 immunohistochemical sections taken at approximately 1.5 mm caudal to the compression epicenter from all animals in each experimental group. Cells were considered pNR1 positive or pNR2B or Iba-1 positive if they were labeled with both pNR1 or pNR2B or Iba-1 specific markers and the nuclear stain DAPI. All sections used for cell quantification were photographed at 10× using a computer assisted fluorescence microscope equipped with image tiling and stitching software (Stereo Investigator). GFAP and Iba-1 expression was assessed through semi-quantitative immunodensity measurements of the dorsal white matter tracts and the dorsal horns (gray matter). Briefly, tissue section images were captured using a 10× objective and Stereo Investigator software. Using ImageJ software (NIH), the approximate area of the dorsal columns (DC) and dorsal horns (DH) was traced and the integrated density of GFAP and Iba-1 was calculated for each tissue section. A total of 5 tissue sections centered at the compression epicenter were analyzed in each animal. Western blotting 1 cm long spinal cord samples centered at the compression epicenter from sham (n = 5), control (n = 8) and riluzole (n = 7) groups were individually homogenized in a RIPE buffer (Thermo Fisher Scientific, Ottawa, Canada) at 4 °C, resolved (20–50 μg per lane) in a 12% SDS-polyacrylamide gel at 200 V and finally transferred to a nitrocellulose membrane. Membranes were then blocked with 5% nonfat milk for 1 h and incubated with galectin-3 (Abcam) antibody. Membranes were incubated with mouse rabbit (1:2000) secondary antibody conjugated to horseradish peroxidase. Reaction products were visualized using an

Fig. 2. Assessment of spinal cord stenosis by micro-CT. (A) Picture of the chronic compression device (CCD) shows the level of plate compression on the cervical spinal cord. (B) Mid-sagittal (right) micro-computed tomography (micro-CT) shows the representative extent of canal stenosis. Image taken from a rat in the control compressed group (no riluzole). (C) Compression Ratio (CR) was calculated as: CR(%) = [1 − 2c / (a + b)] × 100. Typically, chronic compression with the CCD apparatus resulted in a CR between 30 and 40%. Importantly, the CR was not altered in the control and riluzole treated rats (t-test, p = 0.78). Data are presented as mean ± SEM (n = 15 per group).

398

E.S. Moon et al. / Neurobiology of Disease 62 (2014) 394–406

ECL Western Blot Detection kit (Amersham Biosciences Inc.) and Gel Pro analysis software (Media Cybernetics, Silver Spring, MD) to quantify the amount of protein. Densitometric values were normalized to those of β-actin (1:400; Sigma). Statistical analysis Statistical analyses were done using SigmaStat software (Sigma Plot 11.0, Aspire Software International, Ashburn, VA, USA). Mechanical allodynia and neuroanatomical statistical analyses were performed using a two-way ANOVA, followed by a Bonferroni post-hoc pairwise multiple comparisons. For the tail flick test and gait analyses, a oneway ANOVA followed by a Bonferroni post-hoc pairwise multiplecomparison test was performed. A t-test was used when two groups were compared, as in Compression Ratio calculation and CatWalk analysis. Data are reported as means ± SEMs, and p b 0.05 was considered significant. Results Micro-CT assessment of the extent of cervical spinal stenosis The mean CT Compression Ratio (CR), a reflection of the cervical canal anteroposterior (AP) diameter, was measured at 4 weeks postsurgery in order to assure identical canal compromise between the riluzole and control groups (Fig. 2). Canal diameter was decreased due to progressive compression from turning the screw of the CCD. Fig. 2B demonstrates a representative sagittal midline CT image from a compressed animal in the control group. The mean CR ratios are presented in Fig. 2C. There was no difference found in the CR between the riluzole and control groups (t-test, p = 0.78). Riluzole attenuates the development of mechanical allodynia and thermal hyperalgesia in a rodent model of CSM Mechanical allodynia Assessment of forepaw (Fig. 3A). There was an overall difference in forepaw mechanical allodynia, as measured by the sensitivity to mechanical stimuli, between the three groups (two-way ANOVA, p = 0.008) (Fig. 3A). Sensitivity to mechanical stimuli (decreased withdrawal threshold) was not significantly altered in the sham-operated animals over the course of the study. Also, at one week post-surgery, animals of all the three groups did not demonstrate any sensitivity to mechanical stimulation compared to pre-surgery values. Furthermore, there was no difference between the groups at 1 week post-surgery (Bonferroni post-hoc, p = 0.769, p = 1.0, p = 1.0 for sham vs. control, sham vs. riluzole and control vs. riluzole, respectively). The control group demonstrated significant increases in sensitivity compared to the sham animals at weeks 2 to 8 (p = 0.003 for week 2 and p b 0.001 for weeks 3 to 8). Treatment with riluzole led to decreases in compression related sensitivity compared to the control animals at weeks 2, 6, 7 and 8 (p = 0.012, p = 0.025, p = 0.039, p b 0.001). The riluzole treatment group only demonstrated increased sensitivity compared to the sham animals at weeks 4 and 5 (p = 0.002 and p = 0.018, respectively). Assessment of hindpaw (Fig. 3B). There was an overall difference in hindpaw mechanical allodynia between the 3 groups (two-way ANOVA, p b 0.001). Similar to the forepaw, there was no difference in sensitivity to mechanical stimuli in the sham operated animals over the course of the study. At one week post-surgery, the control rats demonstrated increased sensitivity (i.e. lower threshold) as compared to the assessments made at the pre-surgery time point and in the sham operated and riluzole treated rats. There was a statistically significant increase in sensitivity in the control group compared to the sham group at weeks 2 to 8 post-surgery (Bonferroni post-hoc p b 0.001 for each

Fig. 3. Riluzole significantly reduces mechanical allodynia and thermal hyperalgesia. Mechanical allodynia testing in the forepaw (A) and hindpaw (B) was carried out weekly for 8 weeks. Graphs illustrate the changes in withdrawal threshold to mechanical stimulation following treatment with riluzole and vehicle under chronic mechanical compression of the cervical spinal cord. Both fore and hindpaws of the rats in the riluzole group demonstrated significantly increased withdrawal thresholds compared to the control group (two way ANOVA, p b 0.05 for each). (**: S vs R, p b 0.05; ++: C vs R, p b 0.05; +++: C vs R, p b 0.001; ##: S vs C, p b 0.05; ###: sham vs con, p b 0.001). (C) The tail flick test for thermal hyperalgesia was carried out at 8 weeks post-surgery. The graph illustrates the changes in tail withdrawal latency to thermal noxious stimuli following treatment with riluzole and control. Animals of the riluzole group demonstrated increased tail withdrawal latency at 8 weeks compared to the control group (ANOVA, p b 0.05; *Bonferroni post-hoc p b 0.05). [n = 18 in the control group, 17 in the riluzole group and 6 in the sham group]. Error bars represent SEM.

week). Treatment with riluzole led to decreases in compression-related mechanical allodynia compared to the control animals at weeks 3 to 8 (p = 0.024 for week 3, p = 0.017 for week 4 and p b 0.001 for each week from 5 to 8). The riluzole treatment group only showed a statistically significant increase in sensitivity relative to the sham group at 4 weeks post-surgery (p = 0.001). Thermal hyperalgesia We also examined the effects of riluzole on thermal hyperalgesia using the tail flick analysis (Fig. 3C). The tail flick test was performed at 8 weeks after surgery, just prior to sacrifice in order to avoid potentially destabilizing the CCD. There was an overall difference in thermal hyperalgesia, as measured by latency for tail removal (tail flick), between the groups (ANOVA, F = 43.414 and p = 0.001). Sensitivity to thermal noxious stimuli was not significantly altered in the shamoperated animals at 8 weeks post-surgery compared to prior to surgery. Control animals demonstrated a significant decrease in latency compared to the sham animals (Bonferroni post-hoc p = 0.007). Treatment

E.S. Moon et al. / Neurobiology of Disease 62 (2014) 394–406

399

with riluzole significantly mitigated the compression-mediated reduction in latency compared to the control animals (p = 0.006). There was no difference in latency found between the riluzole and sham animals (p = 0.683). Riluzole improves gait performance in a rodent model of CSM To further characterize the effects of riluzole on neurobehavioral function, we performed gait analysis using the CatWalk system at 8 weeks post-surgery (Fig. 4). Riluzole administration significantly increased the swing phase of both forelimbs (Fig. 4A) and hindlimbs (Fig. 4B) compared to the control group (t-test, p b 0.05 for each). Furthermore, the mean intensity with which the fore and the hindpaws were placed on the glass surface was significantly decreased in the control group compared to the sham operated group (one-way ANOVA, F = 56.398 and p b 0.05 for forepaws and F = 59.714 and p b 0.05). Riluzole administration significantly increased paw intensity in both fore (Fig. 4C) and hindpaws (Fig. 4D). Of note, these gait analyses have also been recently used as a novel measure of mechanical allodynia in a model of chronic neuropathic pain (Vrinten and Hamers, 2003). There was no difference found between the riluzole and the sham group in either measure (p N 0.05). Additionally, hindlimb swing speed was significantly increased in the riluzole treated animals (Fig. 4E) compared to the control group (p b 0.05). Riluzole promotes gray matter preservation and reduction of scarring following cervical spinal cord compression Chronic compression led to significant increases in the percentage of scar tissue area. A representative H&E/LFB image from the control (left) and riluzole (right) groups is presented in Fig. 5A respectively. As shown in Fig. 5B, treatment with riluzole significantly reduced the area of scar tissue relative to the control group (two-way ANOVA, p = 0.033). Specifically, there was a statistically significant difference in the scar tissue area between the riluzole and the control group at 720 and 240 μm caudal to the injury epicenter (Bonferroni post-hoc p = 0.045 and 0.018, respectively), at the injury epicenter (p = 0.005) and at 240, 480, 720, 960, 1200 and 1140 μm rostral to the injury epicenter (p = 0.007, p = 0.019, p = 0.048, p = 0.032, p = 0.039 and p = 0.015, respectively). Chronic compression also led to the loss of gray matter. As shown in Fig. 5C, treatment with riluzole significantly preserved gray matter (two way ANOVA, p = 0.002). Specifically, treatment with riluzole resulted in significantly more spared gray matter area compared to the control group at 1440 μm (p = 0.032), 1680 μm (p = 0.025), 1920 μm (p = 0.033) and 2160 μm (p = 0.016) caudal to the epicenter. Riluzole administration attenuates microglia activation and astrogliosis elicited by constant compression of the cervical spinal cord in the dorsal horns of the gray matter Since microglia activation has been correlated with the development of neuropathic pain, initially we performed Iba-1 immunofluorescence staining which showed that under, the chronic compression of the cervical spinal cord, microglia cells exhibited a changed morphology characterized by big soma and short irregular processes indicating an activated state (Fig. 6A low and high magnification). Interestingly, microglia cells after chronic riluzole administration shift status. Our quantification showed that there were significant differences between the experimental groups in terms of Iba-1 immunodensity and Iba-1 positive cells in the dorsal horns (one-way ANOVA, F = 30.110 and p b 0.001 and F = 29.268 and p b 0.001, respectively). Specifically we found that the chronic compression of the cervical spinal cord induced microglia accumulation in the dorsal horns (Bonferroni, p b 0.001 for control vs sham group). As Figs. 6B & C show, statistical analysis demonstrated that daily riluzole administration significantly attenuated the

Fig. 4. Riluzole improves gait performance. Gait analysis with the Catwalk was measured at 8 weeks post-surgery. (A and B) Graphs illustrate the swing phase duration in the forelimb (A) and the hindlimb (B). The animals of the control group demonstrated a significant increase in the swing phase duration compared to the sham operated animals in forelimbs and hindlimbs. Treatment with riluzole significantly decreased the swing phase duration in both forelimbs and hindlimbs compared to the control group. (*t-test, p b 0.05). (C and D) Graphs illustrate the mean contact intensity with which the forepaw (C) and the hindpaw (D) were placed on the glass surface. Contact intensity was significantly decreased in the control group compared to the sham operated group. Riluzole administration significantly increased paw intensity in both fore and hindpaws (*Bonferroni post-hoc p b 0.001 for each). (E) Riluzole also significantly increased the hindlimb swing speed relative to control animals. (*t-test, p b 0.05). [n = 18 in the control group, 17 in the riluzole group and 6 in the sham group]. Error bars represent SEM.

number of microglia cells in the lumbar dorsal horns of the animals in the riluzole treated group when compared to the controls (p = 0.001). Then, we assessed the protein expression of galectin-3 in order to evaluate the levels of microglia activation in all the experimental groups. Western blotting revealed that riluzole administration attenuated the galectin-3 protein expression elicited by the chronic compression of the cervical spinal cord (Bonferroni, p = 0.013).

400

E.S. Moon et al. / Neurobiology of Disease 62 (2014) 394–406

Fig. 5. Riluzole treatment leads to neural tissue preservation. (A) Representative images of H–E & LFB staining in the epicenter of tissue sections coming from the control (left) and riluzole (right) groups. (B) Scar tissue area was found to be statistically significantly decreased in the riluzole group compared to the control groups. H–E & LFB stained sections illustrate the decrease of neural scar formation around the epicenter in the riluzole group rather than control group. (C) The preserved gray matter area was found to be significantly increased in the riluzole group compared to the control group. [two way ANOVA, p b 0.05; *Bonferroni post-hoc, p b 0.05, scale bar = 500 μm]. [n = 18 in the control group, 17 in the riluzole group and 6 in the sham group]. Error bars represent SEM.

Moreover, we used GFAP immunoreactivity (via fluorescence intensity measurements) to semi-quantitatively assess increases in astrocyte proliferation during spinal cord compression. Overall, there were significant differences in the three groups in terms of the CT, DC and DH GFAP expression (ANOVA, F = 19.349 and p b 0.001, F = 11.341 p = 0.005, and F = 14.389 and p b 0.001 respectively). As Fig. 7B demonstrates, we found a statistically significant increase in CT GFAP immunoreactivity in the control group compared to the sham operated animals (p b 0.001). Treatment with riluzole reduced compression-mediated increases in GFAP immunoreactivity compared to the control animals (p = 0.027). There was also a significant difference in GFAP immunoreactivity in the riluzole and sham animals (p = 0.027). Similar trends were also seen in the DC. Control animals contained significantly more GFAP immunoreactivity relative to sham animals (p = 0.006). Treatment with riluzole led to a non-significant reduction in compressionrelated GFAP increases compared to the control group (p = 0.09). There was no difference in GFAP immunoreactivity in the sham and riluzole animals (p = 0.216). Additionally, we found a statistically significant increase in DH GFAP immunoreactivity in the control group compared to the sham operated animals (p b 0.001). Treatment with riluzole reduced compression-mediated increases in DH GFAP immunoreactivity compared to the control animals (p = 0.014). There was no difference

in DH GFAP immunoreactivity between the riluzole and sham animals (p = 0.418). Riluzole decreases the number of p-NR1 and p-NR2B expressing cells in the dorsal horns of the gray matter following chronic mechanical compression of the cervical spinal cord As the phosphorylation of NMDA subunits has been implicated in the development and maintenance of neuropathic pain (Gao et al., 2005; Guo et al., 2002; Miki et al., 2002; Zou et al., 2002), we assessed increases in pNR1 and pNR2B expression through cell counts in the dorsal gray matter (lamina i to v). Representative images of pNR1 and pNR2B immunoreactivity taken 1.5 mm caudal to the epicenter are shown in Fig. 8A. As Fig. 8B demonstrates, we found significant differences in the number of pNR1 positive cells in the three treatment groups (ANOVA, F = 73.184, p = 0.003). There was a significant increase in pNR1 positive cells in the control group compared to the sham animals (p = 0.006). Treatment with riluzole resulted in a decrease in the number of NR1 positive cells in the dorsal horns compared to the control rats (p = 0.006). Additionally, there was no significant difference found between the sham and riluzole animals (p = 0.677). With regard to pNR2B expression, we found significant differences in

E.S. Moon et al. / Neurobiology of Disease 62 (2014) 394–406

401

Fig. 6. Riluzole reduces microglia reaction following chronic cervical spinal cord compression. (A) Iba-1 immunostaining in the sham, control and riluzole treated rats is shown. Low and high magnification images show increased numbers of microglia cells in the dorsal horns of the control group animals. Moreover, in the control group, microglia cells exhibit increased size of soma as well as few and short processes which indicate an activated state. (B & C) Control group showed increased Iba-1 immunodensity and increased number of Iba-1/DAPI positive cells in the dorsal horns compared to the sham operated animals. Iba-1 immunodensity and the number of Iba-1/DAPI positive cells were attenuated by continuous riluzole administration. n = 4 in the control group, 4 in the riluzole group and 4 in the sham group. (D) Representative images of Western blot bands. Densitometry analysis showed that the expression levels of the galectin-3 protein were statistically significantly increased in the control group compared to the sham operated group (post-hoc, p b 0.001). Finally, Western blotting analysis demonstrated that the expression pattern of the galectin-3 protein was statistically significantly decreased in the riluzole treated animals compared to the controls (post-hoc, p = 0.013). Data are presented as mean ± SEM.

the number of pNR2B positive cells in the three treatment groups (ANOVA, F = 91.394, p b 0.001). There was a significant increase in PNR2B positive cells in the control group compared to the sham animals (p b 0.001). Treatment with riluzole resulted in a decrease in the number of PNR2B positive cells in the dorsal horns compared to the control rats (p b 0.001). Additionally, there was a no significant difference found between the sham and riluzole animals (p = 0.085). In order to delineate which cell type exhibits injury and riluzolemediated regulation of pNR1 and pNR2B, we performed quadruplicate immunofluorescence labeling for pNR1 or pNR2B with NeuN, GFAP and with Iba-1. As Figs. 9A & B show, in CSM, pNR1and pNR2B expression is mainly induced in astrocytes and, to a lesser extent, in the neuronal population of the dorsal horns. Colocalization of pNR1 or pNR2B labeling and Iba-1 was not detected (data not shown). Interestingly, riluzole administration decreased the expression of pNR1 and pNR2B in astrocytes located in the dorsal horns of gray matter (Figs. 9A & B). Discussion We have demonstrated that riluzole attenuated mechanical allodynia and thermal hyperalgesia at and below the level of the lesion and enhanced gait performance in a clinically relevant model of CSM. Moreover, we demonstrated that chronic intraperitoneal riluzole administration led to decreased pNR1 and pNR2B subunit expression and reduced gliosis in the DC, CT and DH of the cervical spinal cord. We suggest that the reduced pNR1 and pNR2B cell counts and reduced

GFAP immunoreacitvity in the DH could be the mechanism by which riluzole reduces chronic neuropathic pain in CSM animals at the spinal level — through modulating glutamate excitotoxicity in the dorsal horns. To our knowledge, this is the first study to demonstrate that riluzole can reduce neuropathic pain in models of CSM. Since CSM is the most common cause of spinal cord impairment among adults over the age of 55 (Moore and Blumhardt, 1997; Young, 2000) and in light of the aging demographics of our society, our results could be of significant clinical relevance. As neuropathic pain is one of the most prominent clinical features of CSM, these data in combination with previous reports support the potential use of riluzole in patients with CSM. The use of riluzole in the CNS Riluzole (2-amino-6(trifluoromethoxy)benzothiazole; Rilutek®, Sanofi-Aventis Inc.) is the first and only drug approved for the treatment of ALS in the United States. Mechanistic studies have suggested that riluzole slows the neurodegeneration of motor neurons. An important mechanism appears to be decreasing the excitotoxic effect of excessive levels of extracellular glutamate by inhibiting glutamate release and postsynaptic glutamate receptor activation (Bellingham, 2011; Wang et al., 2004). Pre-clinical studies have demonstrated that riluzole decreases glutamate mediated excitotoxicity by increasing the activity of synaptic glutamate transporters and by antagonizing NMDA and nonNMDA receptors. In addition, synaptic release of glutamate is inhibited by blocking voltage-gated Na+ and Ca2+ channels (Azbill et al., 2000;

402

E.S. Moon et al. / Neurobiology of Disease 62 (2014) 394–406

Fig. 7. Riluzole reduces astrogliosis. (A) GFAP immunostaining in the sham, control and riluzole treated rats at the injury epicenter is shown. GFAP immunoreactivity was measured in the dorsal columns (DC), corticospinal tracts (CT) and dorsal horns (DH). High resolution images show representative DH GFAP expression in each group (right panels). (B) Compression groups (riluzole and control) showed astrogliosis in the DC, CT and DH. GFAP immunoreactivity in the DC, CT and DH of riluzole animals was significantly decreased when compared to the control group. Scale bar = 500 μm (larger images on the left), 25 μm (higher resolution images on the right). ANOVA, p b 0.001; *Bonferroni post-hoc, p b 0.05. n = 12 in the control group, 12 in the riluzole group and 6 in the sham group. Error bars represent SEM.

effect of riluzole. Interestingly, antinociception in this study was not obtained in uninjured rats following central administration, suggesting a peripherally mediated antinociception. In our experiments, using the von Frey filament test we examined the effects of riluzole in mechanical allodynia. The von Frey filament test is one of the most frequently used methods to measure mechanical allodynia in various animal models and it has been shown to be objective and reproducible (Bell-Krotoski and Tomancik, 1987). Our data indicate that continuous intraperitoneal riluzole administration for 7 weeks attenuated the induction as well as the maintenance of the chronic compression-induced mechanical allodynia. The CatWalk has been used as a powerful tool to study functional deterioration and recovery after spinal cord injury (Gensel et al., 2006; Hamers et al., 2001; Koopmans et al., 2005) in rats. Moreover, several dynamic and static parameters of CatWalk such as swing phase duration and paw intensity have also been used as a novel method to assess mechanical allodynia in a model of chronic neuropathic pain (Vrinten and Hamers, 2003), acute inflammatory pain models (Gabriel et al., 2007) and models of ankle–joint pain (Angeby-Moller et al., 2008). The paw intensity parameter has been used to assess the effects of neuropathic pain, including mechanical allodynia (Vrinten and Hamers, 2003). They found that, after a chronic constriction injury, the change in paw intensity showed a high degree of correlation with Von Frey thresholds. In a carrageenan-induced monoarthritis model, the effect of the analgesics Morphine and Rofecoxib on CatWalk gait parameters was examined. It was demonstrated that Morphine and Rofecoxib could partially reverse carrageenan induced changes in CatWalk parameters (Angeby-Moller et al., 2008). In the present study, we looked at changes in the paw intensity and duration of swing phases of a complete step cycle in rats with CSM. Paw intensity was decreased and duration of the swing phase was increased in CSM rats (Fig. 4). Importantly, we demonstrated that riluzole significantly improved the degree of contact between the paw and the glass plate, and the normal swing phase in CSM animals. Based on these studies, we expect that the swing phase parameter was not only affected by compression-related motor compromise but also could be affected by the increased sensitivity to mechanical stimulation in the paws of these rats. Moreover, our results showed that riluzole treatment decreased the maintenance of thermal hyperalgesia. It should be noted that our experimental design did not allow us to examine the potential effects on the induction of thermal hyperalgesia after the gradual spinal cord compression.

Mechanistic insights of riluzole and neuropathic pain Debono et al., 1993). In contrast to glutamate γ-amino-butyric acid (GABA) is the major inhibitory neurotransmitter throughout the CNS. It has been showed that, at a concentration which is unattainable in human subjects, riluzole blocked the uptake of GABA into striatal synaptosomes (Mantz et al., 1994). Riluzole's effects on neuropathic pain Our data with riluzole are consistent with other models of neuropathic pain. Specifically, in a rat model of neuropathic pain induced by chronic constriction sciatic nerve injury (CCI), it has been shown that systemic and intrathecal riluzole administration reduced the induction of mechanical allodynia and thermal hyperalgesia (Sung et al., 2003). Moreover, this study also showed that riluzole gradually reversed thermal hyperalgesia and mechanical allodynia in CCI rats when the treatment began on postoperative day 5 (Sung et al., 2003). In addition, Hama and Sagen showed that systemically administered riluzole was significantly antinociceptive on below-level cutaneous hypersensitivity in rats with a SCI (Hama and Sagen, 2011). Their data indicated that the acute antinociceptive effect of riluzole in SCI rats was mediated at the supraspinal but not the spinal level. Moreover, they also observed antinociception in uninjured rats, indicating a generalized antinociceptive

There is evidence to indicate the involvement of glutamatergic pathways in the injury-induced plastic changes and reorganization of the spinal nociceptive network after injury (al-Ghoul et al., 1993; Urban and Gebhart, 1999). It has been reported that spinal NMDA receptors play an important role in central sensitization and are closely linked to the development of allodynia and hyperalgesia under neuropathic conditions. Specifically, several studies suggest that phosphorylation of the NMDA receptor NR1 subunit is correlated to the presence of signs of neuropathy and to persistent pain following nerve injury (Gao et al., 2005; Roh et al., 2008; Ultenius et al., 2006). Moreover, it has been demonstrated that the phosphorylation of NR2B subunits of the NMDA glutamate receptor is increased in a neuropathic-pain model and that this phosphorylation is required for the maintenance of neuropathic pain by L5-spinal nerve transaction (Abe et al., 2005; Guo et al., 2002). In our experiments we examined whether the potent analgesic effect of riluzole administration is associated with the suppression of spinal NMDA receptor activity, as measured by a reduction in pNR1 and pNR2B subunit expression in the spinal cord dorsal horn of rats which underwent chronic mechanical compression of the cervical spinal cord. We found that chronic mechanical compression of the spinal cord results in overexpression of the pNR1 and pNR2B subunits in the

E.S. Moon et al. / Neurobiology of Disease 62 (2014) 394–406

403

Fig. 8. Riluzole reduces the phosphorylation of NR1 and NR2B in the dorsal horns of compressed rats. (A) Immunohistochemistry for pNR1 and pNR2B (red) is shown for each group. (B) Compression resulted in a significant increase in the number of pNR1 positive cells in the dorsal horns. Treatment with riluzole reduced the number of pNR1 positive cells compared to the control group. (C) Similarly, there was a significant increase in pNR2B positive cells in the control group compared to the sham animals in the dorsal horns. Treatment with riluzole resulted in a decrease in the number of pNR2B positive cells in the dorsal horns compared to the control rats. There was no difference found between the sham and riluzole animals for both pNR1 and pNR2B cell counts. ANOVA, p b 0.001; *Bonferroni post-hoc, p b 0.05. n = 5 in the control group, 5 in the riluzole group and 3 in the sham group. Error bars represent SEM. Scale bar = 25 μm.

spinal dorsal horns which was eliminated by the chronic riluzole administration (Fig. 7). It is now clear that neurons are not the only players that establish and maintain common clinical pain states (Scholz and Woolf, 2007). Multiple mechanisms, working alone or in concert, may contribute to the establishment of hypersensitivity. Microglia represents 5% of the glial cells and they represent “macrophages” of the central nervous system. A strong body of literature suggests that the microglial cell population is a key mediator of induction and maintenance of neuropathic pain behaviors after experimental peripheral nerve injury and SCI as well as after clinical cases of SCI (Burgos et al., 2012; Hains and Waxman, 2006; Inoue and Tsuda, 2009; Inoue and Tsuda, 2012; Tan et al., 2009; Xu et al., 2013). Moreover, our group has recently showed increased microglia activation in the spinal cord coming from human CSM cases. Activated microglia contribute to neuropathic pain by promoting the synthesis and secretion of proinflammatory cytokines and chemokines including

TNF-a (Burgos et al., 2012; Hama et al., 2012; Ishikawa et al., 2013; Leung and Cahill, 2010; Liang et al., 2012; Peng et al., 2006) which affects the intrinsic properties of neurons in the dorsal horns. Galectin-3, a member of the carbohydrate-binding proteins family, is highly expressed in activated microglia (Comte et al., 2011; Kriz et al., 2003; Lalancette-Hebert et al., 2009; Ohtaki et al., 2008; Satoh et al., 2011) but not in resting microglia cells (Lalancette-Hebert et al., 2012). Furthermore, Lalancette-Hebert et al. recently showed that galectin-3 is essential in the process of activation and proliferation of microglia after brain ischemia. Indeed, the use of galectin-3 as a marker of microglia activation has been increased (Lalancette-Hebert et al., 2012). In the present study, we found that the gradual and then constant compression of the cervical spinal cord induces microglia recruitment and activation in the lumbar dorsal horns as it was indicated by the increased number of Iba-1 positive cells and by the increased galectin-3 protein expression, respectively. Continuous

404

E.S. Moon et al. / Neurobiology of Disease 62 (2014) 394–406

Fig. 9. pNR1 and pNR2B cellular expression in the spinal dorsal horns. Triple-label immunostaining of pNR1 or pNR2B with GFAP and NeuN was carried out. For both pNR1 and pNR2B there was colocalization mainly in astrocytes suggesting that both receptors are located in the astrocytes. Riluzole treatment decreased the expression of astrocytic pNR1 and pNR2B under the chronic cervical spinal cord compression. The decrease astrocytic expression of pNR1 and pNR2B is associated with the decreased hypertrophy of astrocytes. Arrows shows the colocalization.

riluzole administration during the course of the compression decreased the number of microglia cells in the lumbar dorsal horns and shifted their status to a less activated state by decreasing the galectin-3 expression. The current results in combination with the existing body of

literature suggest that the microglia activation is implicated in the development and maintenance of the level of neuropathic pain in CSM. Finally, we provide a mechanistic insight regarding how riluzole modulates the microglia activation.

E.S. Moon et al. / Neurobiology of Disease 62 (2014) 394–406

Finally, there is also evidence supporting the involvement of astrocytes in neuropathic pain modulation and processing (Garrison et al., 1991; Raghavendra et al., 2004; Sweitzer et al., 1999; Hald et al., 2009; Obata et al., 2006). In this study, immunofluorescence experiments showed that the chronic mechanical compression of the cervical spinal cord leads to astrogliosis in the DH of the gray matter not only at the level of compression epicenter (Fig. 6) but also above and below that. Although the quantification of the immunofluorescence entails certain limitations, current results propose a potential implication of astrogliosis in the development of neuropathic pain after constant compression of the cervical spinal cord. The relationship between pNR1 expression and pNR2B expression changes in the microglia phenotype with the development and maintenance of neuropathic pain has been established in the literature; however, the data we provided in the present study regarding this relationship only allowed us to make an associative inference. As such, future work will be needed to confirm such a mechanism in CSM. Conclusion Riluzole improved functional recovery and preserved tissue in a rat model of mild gradual cervical spinal cord compression. Additionally, riluzole inhibited the induction and maintenance of neuropathic pain. Data from our study in combination with previously published reports suggest that riluzole might decrease the phosphorylation of the NR1 and NR2B receptors in the spinal dorsal horns which in turn might lead to attenuation of the glutamate excitotoxicity and to subsequent decrease of the neuronal sensitization. Moreover, riluzole decreased dorsal horn microglia recruitment and activation which has been previously shown to be positively correlated with the induction and the maintenance of neuropathic pain. Taken together, our results suggest that riluzole represents a novel therapeutic approach to attenuate neuropathic pain and protect against neurological loss in CSM. Given the high prevalence of CSM and the aging demographics of our population, these data could be of considerable clinical significance. Conflict of interest statement The authors state that they have no conflicts of interest to declare. References Abe, T., et al., 2005. Fyn kinase-mediated phosphorylation of NMDA receptor NR2B subunit at Tyr1472 is essential for maintenance of neuropathic pain. Eur. J. Neurosci. 22, 1445–1454. al-Ghoul, W.M., et al., 1993. Glutamate immunocytochemistry in the dorsal horn after injury or stimulation of the sciatic nerve of rats. Brain Res. Bull. 30, 453–459. Angeby-Moller, K., et al., 2008. Using the CatWalk method to assess weight-bearing and pain behaviour in walking rats with ankle joint monoarthritis induced by carrageenan: effects of morphine and rofecoxib. J. Neurosci. Methods 174, 1–9. Azbill, R.D., et al., 2000. Riluzole increases high-affinity glutamate uptake in rat spinal cord synaptosomes. Brain Res. 871, 175–180. Baptiste, D.C., Fehlings, M.G., 2006. Pathophysiology of cervical myelopathy. Spine J. 6, 190S–197S. Bellingham, M.C., 2011. A review of the neural mechanisms of action and clinical efficiency of riluzole in treating amyotrophic lateral sclerosis: what have we learned in the last decade? CNS Neurosci. Ther. 17, 4–31. Bell-Krotoski, J., Tomancik, E., 1987. The repeatability of testing with Semmes–Weinstein monofilaments. J. Hand. Surg. [Am.] 12, 155–161. Bohlman, H.H., Emery, S.E., 1988. The pathophysiology of cervical spondylosis and myelopathy. Spine 13, 843–846. Burgos, E., et al., 2012. Cannabinoid agonist WIN 55,212-2 prevents the development of paclitaxel-induced peripheral neuropathy in rats. Possible involvement of spinal glial cells. Eur J Pharmacol. 682, 62–72. Chen, B.S., Roche, K.W., 2007. Regulation of NMDA receptors by phosphorylation. Neuropharmacology 53, 362–368. Chen, B.S., et al., 2006. Regulation of NR1/NR2C N-methyl-D-aspartate (NMDA) receptors by phosphorylation. J. Biol. Chem. 281, 16583–16590. Comte, I., et al., 2011. Galectin-3 maintains cell motility from the subventricular zone to the olfactory bulb. J. Cell Sci. 124, 2438–2447. Debono, M.W., et al., 1993. Inhibition by riluzole of electrophysiological responses mediated by rat kainate and NMDA receptors expressed in Xenopus oocytes. Eur. J. Pharmacol. 235, 283–289.

405

Fehlings, M.G., et al., 1999. The optimal radiologic method for assessing spinal canal compromise and cord compression in patients with cervical spinal cord injury. Part II: Results of a multicenter study. Spine (Phila Pa 1976) 24, 605–613. Gabriel, A.F., et al., 2007. The CatWalk method: a detailed analysis of behavioral changes after acute inflammatory pain in the rat. J. Neurosci. Methods 163, 9–16. Gao, X., et al., 2005. Enhancement of NMDA receptor phosphorylation of the spinal dorsal horn and nucleus gracilis neurons in neuropathic rats. Pain 116, 62–72. Garrison, C.J., et al., 1991. Staining of glial fibrillary acidic protein (GFAP) in lumbar spinal cord increases following a sciatic nerve constriction injury. Brain Res. 565, 1–7. Gensel, J.C., et al., 2006. Behavioral and histological characterization of unilateral cervical spinal cord contusion injury in rats. J. Neurotrauma 23, 36–54. Guo, W., et al., 2002. Tyrosine phosphorylation of the NR2B subunit of the NMDA receptor in the spinal cord during the development and maintenance of inflammatory hyperalgesia. J. Neurosci. 22, 6208–6217. Hains, B.C., Waxman, S.G., 2006. Activated microglia contribute to the maintenance of chronic pain after spinal cord injury. J. Neurosci. 26, 4308–4317. Hald, A., et al., 2009. Differential activation of spinal cord glial cells in murine models of neuropathic and cancer pain. Eur. J. Pain 13, 138–145. Hama, A., Sagen, J., 2011. Antinociceptive effect of riluzole in rats with neuropathic spinal cord injury pain. J. Neurotrauma 28, 127–134. Hama, A.T., et al., 2012. The antinociceptive effect of the asthma drug ibudilast in rat models of peripheral and central neuropathic pain. J. Neurotrauma 29, 600–610. Hamers, F.P., et al., 2001. Automated quantitative gait analysis during overground locomotion in the rat: its application to spinal cord contusion and transection injuries. J. Neurotrauma 18, 187–201. Inoue, K., Tsuda, M., 2009. Microglia and neuropathic pain. Glia 57, 1469–1479. Inoue, K., Tsuda, M., 2012. P2X4 receptors of microglia in neuropathic pain. CNS Neurol. Disord. Drug Targets 11, 699–704. Ishikawa, T., et al., 2013. Differences between tumor necrosis factor-alpha receptors types 1 and 2 in the modulation of spinal glial cell activation and mechanical allodynia in a rat sciatic nerve injury model. Spine (Phila Pa 1976) 38, 11–16. Kalsi-Ryan, S., Karadimas, S.K., Fehlings, M.G., 2013. Cervical spondylotic myelopathy: the clinical phenomenon and the current pathobiology of an increasingly prevalent and devastating disorder. Neuroscientist 19 (4), 409–421. http://dx.doi.org/10.1177/ 1073858412467377 (Epub 2012 Nov 30, PubMed PMID: 23204243). Karadimas, S.K., et al., 2013. A novel experimental model of cervical spondylotic myelopathy (CSM) to facilitate translational research. Neurobiol. Dis. 54, 43–58. Kitzman, P.H., 2009. Effectiveness of riluzole in suppressing spasticity in the spinal cord injured rat. Neurosci. Lett. 455, 150–153. Koopmans, G.C., et al., 2005. The assessment of locomotor function in spinal cord injured rats: the importance of objective analysis of coordination. J. Neurotrauma 22, 214–225. Kriz, J., et al., 2003. Efficient three-drug cocktail for disease induced by mutant superoxide dismutase. Ann. Neurol. 53, 429–436. Lalancette-Hebert, M., et al., 2009. Live imaging of Toll-like receptor 2 response in cerebral ischaemia reveals a role of olfactory bulb microglia as modulators of inflammation. Brain 132, 940–954. Lalancette-Hebert, M., et al., 2012. Galectin-3 is required for resident microglia activation and proliferation in response to ischemic injury. J. Neurosci. 32, 10383–10395. Lee, J., Satkunendrarajah, K., Fehlings, M.G., 2012. Development and characterization of a novel rat model of cervical spondylotic myelopathy: the impact of chronic cord compression on clinical, neuroanatomical, and neurophysiological outcomes. J. Neurotrauma 29 (5), 1012–1027. http://dx.doi.org/10.1089/neu.2010.1709 (Epub 2011 Aug 29, PubMed PMID: 21561323). Leung, L., Cahill, C.M., 2010. TNF-alpha and neuropathic pain—a review. J. Neuroinflammation 7, 27. Liang, Y., et al., 2012. Behavioral and morphological evidence for the involvement of glial cells in the antinociceptive effect of najanalgesin in a rat neuropathic pain model. Biol. Pharm. Bull. 35, 850–854. Mantz, J., et al., 1994. Riluzole, a novel antiglutamate, blocks GABA uptake by striatal synaptosomes. Eur. J. Pharmacol. 257, R7–R8. Miki, K., et al., 2002. Changes in gene expression and neuronal phenotype in brain stem pain modulatory circuitry after inflammation. J. Neurophysiol. 87, 750–760. Moore, A.P., Blumhardt, L.D., 1997. A prospective survey of the causes of non-traumatic spastic paraparesis and tetraparesis in 585 patients. Spinal Cord 35, 361–367. Munro, G., et al., 2007. Pharmacological comparison of anticonvulsant drugs in animal models of persistent pain and anxiety. Neuropharmacology 53, 609–618. Obata, H., et al., 2006. Spinal glial activation contributes to postoperative mechanical hypersensitivity in the rat. J Pain 7, 816–822. Ohtaki, H., et al., 2008. Stem/progenitor cells from bone marrow decrease neuronal death in global ischemia by modulation of inflammatory/immune responses. Proc. Natl. Acad. Sci. U. S. A. 105, 14638–14643. Peng, X.M., et al., 2006. Tumor necrosis factor-alpha contributes to below-level neuropathic pain after spinal cord injury. Ann. Neurol. 59, 843–851. Raghavendra, V., et al., 2004. Complete Freunds adjuvant-induced peripheral inflammation evokes glial activation and proinflammatory cytokine expression in the CNS. Eur. J. Neurosci. 20, 467–473. Roh, D.H., et al., 2008. Depletion of capsaicin-sensitive afferents prevents laminadependent increases in spinal N-methyl-D-aspartate receptor subunit 1 expression and phosphorylation associated with thermal hyperalgesia in neuropathic rats. Eur. J. Pain 12, 552–563. Satoh, K., et al., 2011. Increase of galectin-3 expression in microglia by hyperthermia in delayed neuronal death of hippocampal CA1 following transient forebrain ischemia. Neurosci. Lett. 504, 199–203. Scholz, J., Woolf, C.J., 2007. The neuropathic pain triad: neurons, immune cells and glia. Nat. Neurosci. 10, 1361–1368.

406

E.S. Moon et al. / Neurobiology of Disease 62 (2014) 394–406

Sung, B., et al., 2003. Altered expression and uptake activity of spinal glutamate transporters after nerve injury contribute to the pathogenesis of neuropathic pain in rats. J. Neurosci. 23, 2899–2910. Sweitzer, S.M., et al., 1999. Acute peripheral inflammation induces moderate glial activation and spinal IL-1beta expression that correlates with pain behavior in the rat. Brain Res. 829, 209–221. Tan, A.M., et al., 2009. Early microglial inhibition preemptively mitigates chronic pain development after experimental spinal cord injury. J. Rehabil. Res. Dev. 46, 123–133. Ultenius, C., et al., 2006. Spinal NMDA receptor phosphorylation correlates with the presence of neuropathic signs following peripheral nerve injury in the rat. Neurosci. Lett. 399, 85–90. Urban, M.O., Gebhart, G.F., 1999. Central mechanisms in pain. Med. Clin. North Am. 83, 585–596.

Villmann, C., Becker, C.M., 2007. On the hypes and falls in neuroprotection: targeting the NMDA receptor. Neuroscientist 13, 594–615. Vrinten, D.H., Hamers, F.F., 2003. ‘CatWalk’ automated quantitative gait analysis as a novel method to assess mechanical allodynia in the rat; a comparison with von Frey testing. Pain 102, 203–209. Wang, S.J., et al., 2004. Mechanisms underlying the riluzole inhibition of glutamate release from rat cerebral cortex nerve terminals (synaptosomes). Neuroscience 125, 191–201. Xu, Z.Z., et al., 2013. Resolvin E1 inhibits neuropathic pain and spinal cord microglial activation following peripheral nerve injury. J. Neuroimmune Pharmacol. 8, 37–41. Young, W.F., 2000. Cervical spondylotic myelopathy: a common cause of spinal cord dysfunction in older persons. Am. Fam. Physician 62 (1064–70), 1073. Zou, X., et al., 2002. Role of protein kinase A in phosphorylation of NMDA receptor 1 subunits in dorsal horn and spinothalamic tract neurons after intradermal injection of capsaicin in rats. Neuroscience 115, 775–786.