1. Introduction
Pannexin 1 (panx1) is a protein that functions as a large-pore membrane channel with up to 500pS conductance [2]. It has been widely investigated since it was identified in the mammalian genome in the year 2000 by Panchin et al. [35]. Panx1 can be opened after mechanical stimulation [2], membrane depolarization [2,6], purinergic [27,36], and N-methyl-D-aspartate (NMDA) receptor [56] activation, intracellular calcium [27], or elevated extracellular potassium [44,46]. Panx1 mRNA has been found during development and in mature systems in many tissues in rats, mice, and humans [6]. In the central nervous system, panx1 is expressed both in neurons and astrocytes [50]. In neurons, panx1 has been observed in the postsynaptic density of hippocampal pyramidal cells in co-expression with the postsynaptic density 95 protein, suggesting a modulatory role in excitability of postsynaptic neurons [58]. At present, there is growing evidence that supports a role of panx1 channels in some pathologic conditions of the central nervous system, particularly in epilepsy, cerebral ischemia, and neuroinflammation [16,44,46,50,51].
Glutamate and adenosine triphosphatase (ATP) have been described as crucial molecules in acute pain signaling in the dorsal horn of the spinal cord, as well as in the process of developing and maintaining central sensitization underlying chronic pain [5,7,25,30]. Interestingly, panx1 is highly permeable to ATP [2,8,27], the main activator of the purinergic signaling, 1 of the pathways that mediates neuroplastic phenomena occurring in neuropathic pain [52]. Besides, ATP and glutamate can open panx1 channels via activation of P2X7 [20,36] and NMDA [50] receptors respectively, thus leading to the question of whether panx1 participates in pain signaling in the spinal cord of either normal animals and/or in neuropathic pain conditions. In this regard, a recent report showed that the gap junction blocker carbenoxolone attenuated mechanical hypersensitivity in a model of pathological pain induced in rat by partial transection of the infraorbital nerve [54]. However, carbenoxolone is a nonselective inhibitor of gap-junctions and hemichannels, and therefore, as pointed out by Wang et al. [54], a variety of different mechanisms (mostly occurring in glial cells) could account for the inhibitory effect of carbenoxolone on central sensitization produced by nerve injury, such as inhibition of intercellular Ca2+ waves in astrocytes via gap-junctions, modulation of astrocyte volume–regulated anion channels, inhibition of the expression of interleukin-23 in microglia, and suppression of the release of certain glial mediators via inhibition of hemichannels (for specific references, see Wang et al. [54]). Thus, because the role of panx1 in chronic pain remains largely unknown, we addressed this question in a rat model of neuropathic pain. We hypothesized that panx1 participates in dorsal horn mechanisms of hyperalgesia/allodynia accompanying chronic neuropathic pain, and that pharmacological blockade of panx1 would reduce mechanical hyperalgesia and some neuroplastic processes (eg, spinal wind-up activity) in neuropathic animals.
2. Methods
2.1. Animals
Male Sprague–Dawley rats (225–250g body weight) were used in this study. All animals were obtained from the facilities of the Faculty of Medicine of the University of Chile, held in a light–dark cycle of 12/12hours, starting at 8 AM, with food and water ad libitum. The experiments were conducted in accordance with the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health (NIH) and adhered to the guidelines of the Committee for Research and Ethical Issues of IASP [56]. Furthermore, the housing conditions and experimental procedures conformed to protocols approved by the Bioethics Committee of the University of Santiago of Chile.
2.2. Induction of neuropathy
Neuropathy was induced by using a modification of the spared nerve injury rat model described by Decosterd and Woolf [16], which results in early, prolonged, and robust changes in mechanical sensitivity and thermal responsiveness that closely mimic many features of clinical neuropathic pain. In the original description [16], 2 of the 3 terminal distal branches of the sciatic nerve were axotomized (tibial and common peroneal nerves), sparing 1 (sural nerve), whereas in the present version of the model, only the sural nerve was transected, sparing the tibial and common peroneal nerves. This procedure allowed us to generate a neuropathic pain model in which nociceptive reflexes to be recorded are preserved (eg, C-reflex activity), as the sural nerve contains almost no motor fibers [38].
Animals were anesthetized with 400mg/kg i.p. of 7% chloral hydrate solution (w/v). After shaving the right hindpaw at the level of sciatic nerve, a skin incision approximately 10mm long was made. The subcutaneous tissue was dissected, and the biceps femoris muscle was freed from the pelvic and vertebral heads to expose the sciatic nerve. The nerve path was then followed until its split into 3 branches: the sural, common peroneal, and tibial nerves. The sural nerve was cut 2mm from its emergence, and the overlying tissues were sutured in layers. During the 2days after surgery, animals were daily given 3mg/kg s.c. of the analgesic ketoprofen and 5mg/kg s.c. of the antimicrobial agent enrofloxacin. The neural lesion described above resulted in mechanical hyperalgesia of the hindpaw that persisted for at least 28 days (data not shown). Control (sham) rats received similar surgery but without severing the sural nerve.
2.3. Pharmacological blockage of panx1 in spinal cord
For studying the effect of panx1 channel blockade on mechanical hyperalgesia induced by neuropathy, panx1 channels of the lumbar spinal cord were challenged with the 10panx peptide (Trp-Arg-Gln-Ala-Ala-Phe-Val-Asp-Ser-Tyr), a mimetic peptide of the first extracellular loop domain of panx1 [36] that seems to act through a steric interference with channel function, obtained from Tocris (St Louis, MO). Neuropathic and sham rats received a single 10-μL i.t. injection of 10, 30, 100, or 300μmol/L 10panx, whereas the respective control groups received 10μL saline solution i.t. In addition, 2 other drugs described in the literature as pharmacological blockers of panx1 were also used: carbenoxolone (Cbx), a synthetic drug that blocks panx1 channels expressed in Xenopus oocytes [7], which was i.t. injected as 100μmol/L solution; and probenecid (Prb), an agent that exhibit inhibitory activity against panx1 [45], which was i.t. injected as 150μmol/L solution. Both Cbx and Prb were obtained from Sigma (St Louis, MO).
For the electrophysiological study of the effect of panx1 channel blockade on C-reflex wind-up activity, a single 10-μL i.t. injection of 300μmol/L 10panx, 100μmol/L Cbx, 150μmol/L Prb, or saline solution was administered to neuropathic and sham control animals. Single doses of 10panx, Cbx and Prb were adapted from the studies of Thompson et al. [50], Bruzzone et al. [7] and Ma et al. [28], respectively. Injections into the subarachnoid space were made under brief isoflurane anesthesia (2minutes), by direct percutaneous injection between lumbar vertebrae L5 and L6, and evidenced by the slight movement of the rat’s tail that results from mechanical stimulation when the needle penetrates the meninges of the spinal cord [29]. At the end of the behavioral and electrophysiological experiments, rats were euthanized with an overdose of chloral hydrate (1g/kg i.p.).
2.4. Electrophysiological assessment of the C-fiber–evoked nociceptive reflex and wind-up activity
The C-reflex was elicited in the right hindlimb of rats anesthetized with 1.5% to 1.8% isoflurane in oxygen using a latex diaphragm-modified rodent facemask, as described previously by Smith and Bolon [47]. Briefly, rectangular electric pulses of supramaximal strength and 2-millisecond duration were applied every 10seconds to the common peroneal and tibial nerve receptive field by means of 2 stainless steel needles inserted into the skin of the second and third toes (Grass S11 stimulator equipped with a Grass SIU 5 stimulus isolation unit and a Grass CCU 1A constant current unit; Astro-Med, West Warwick, RI). The C-fiber–evoked reflex response was recorded from the ipsilateral biceps femoris muscle by using another pair of stainless steel needles. After amplification (Grass P511 preamplifier; Astro-Med, West Warwick, RI), the electromyographic responses were digitized at 100KHz and integrated into a time-window from 150 to 450ms after the stimulus (Powerlab ML 820, ADInstruments, Castle Hill, NSW, Australia). Once stable C-reflex responses were obtained, the stimulus strength was lowered, and the current required for threshold activation of the C-reflex was determined. Integrated C-reflex responses, evoked by single stimuli with twice the intensity of the threshold stimulating current, were then recorded. Afterward, trains of 15 stimuli each, at 1Hz and twice the threshold intensity, were delivered to the toes to develop wind-up activity. In the C-reflex paradigm, wind-up consists of a stimulus frequency-dependent remarkable increment of the electromyographic integrated response. All responses were stored on hard disk for later analysis. Least-square regression lines were fitted among experimental points showing only incremental trend (before wind-up saturation at the seventh stimulus), discarding the remaining points (Origin 6.0 software; Microcal Software, Northampton, MA), as described elsewhere [14]. The slopes of the regression lines represent wind-up scores (Fig. 1A).
;)
Fig. 1:
Effect of intrathecal administration of 10panx or saline solution on spinal wind-up activity in sham control and neuropathic rats. (A) Figures depicting data processing for determining spinal cord wind-up scores in animals injected i.t. with saline solution or 300 μmol/L 10panx. C-reflex responses were elicited in sham animals by repetitive electric stimulation (1 Hz) of the second and third toes, before (circles) and T minutes after (squares) the i.t. injection of saline solution (left panel) or 10panx (right panel). The responses were integrated and plotted against the stimulus number, and the curves were normalized so that, before injecting the animals, the reflex gain at the seventh stimulus represented a 100% increase. Slopes of least-squares regression lines (dashed lines) represent wind-up scores. For the left and right panels, T = 15 minutes; that is, data were taken before and 15 minutes after i.t. injection of saline solution or 10panx. For the left panel, means of slopes ± standard error of the mean (SEM) of regression curves are 15.09 ± 0.43 (circles) and 15.59 ± 0.44 (squares), the difference between slopes being not statistically significant (n = 8, P > .05, 2-tailed paired Student t test). For the right panel, means of slopes ± SEM of regression curves are 15.81 ± 0.80 (circles) and 4.02 ± 0.58 (squares), the difference between slopes being statistically significant (n = 8, P = .0064, paired 2-tailed Student t test). Data from neuropathic rats were processed in a similar way to those from sham animals (not shown). (B) Time-course of wind-up activity (percent change) after i.t. administration of saline solution (circles) or 300 μmol/L 10panx (squares), in sham (open symbols) and neuropathic (closed symbols) rats. Wind-up scores at time T are expressed as percent change of the slope at time T compared to the slope before 10panx injection at time 0. Values are mean ± SEM, n = 8 rats per group. Two-way repeated-measures analysis of variance (ANOVA) revealed a “time” effect (intragroup analysis: F9,45 = 4.16, ANOVA P = .0006), and a “group” effect (intergroup analysis: F1,45 = 603.69, ANOVA P < 0.0001; * P < .05 with respect to corresponding saline solution values, Bonferroni multiple comparisons test).
The stimulation current required for threshold activation of the C-reflex was lower in neuropathic rats than in sham control animals (3.8±2.1mA vs 8.3±1,5mA, respectively); consequently, wind-up series in neuropathic rats were run with lower electric intensities than those series run in sham control animals, according to the protocol used (2× threshold stimulating current). To equalize the response in both groups, the integrated C-reflex responses were expressed in terms of percent variation compared with their previous condition (before the i.t. injection of drug or saline solution) and normalized so that responses elicited by the seventh stimulus in saline solution–injected animals represent a 100% increase with respect to basal responses obtained before applying the wind-up protocol (Fig. 1A). The experiments began with the measurement of a basal wind-up, prior to i.t. 10panx, Cbx, Prb or saline solution administration. Thereafter, 10μmol/L of 300μmol/L 10panx, 100μmol/L Cbx, 150μmol/L Prb, or saline solution was injected i.t. by direct percutaneous injection in both normal and neuropathic rats, as described previously [29], and wind-up determination was resumed for 240minutes every 30minutes. Wind-up scores obtained were plotted as time-course curves.
2.5. Behavioral assessment of mechanical hyperalgesia
Mechanical hyperalgesia in the hindlimb after neuropathy was assessed by measuring paw pressure threshold with an analgesymeter (Ugo Basile, Varese, Italy), as originally described by Randall and Selitto [40]. In brief, the behavioral testing consisted of adding an increasing pressure gradually applied to the right hindpaw until the withdrawal reflex of the paw occurred. To prevent injury of the hindpaw, a cut-off value of 570g was used [37]. Hindpaw pressure testing was performed on neuropathic and sham rats before (day 0), and 3, 7, and 10days after neuropathy or sham surgery, respectively. Thereafter, on day 10 after induction of the neuropathy, the rats received a single 10-μL i.t. injection of 10panx, Cbx, Prb, or saline solution. The 10panx was administered as a 10-, 30-, 100-, or 300-μmol/L solution. Cbx and Prb were injected as 100μmol/L and 150μmol/L solutions, respectively, and 10μL i.t. saline solution served as control. Paw pressure thresholds were measured before injecting drugs or saline solution, and at 15, 30, 60, 90, 120, 150, 180, 210, and 240minutes after the i.t. injection.
2.6. Expression of panx1 protein in spinal cord
Panx1 protein expression was measured following the protocol of Mylvaganam et al. [32]. Rats were sacrificed with an overdose of chloral hydrate (1g/kg, intraperitoneal, i.p.). For protein extraction, the posterior quadrants of the lumbar spinal cords (L1–L3 vertebrae level) ipsilateral to the lesion, taken from 5 normal and 5 neuropathic animals, were homogenized in ristocetin-induced platelet agglutination (RIPA) buffer containing 50mmol/L Tris (pH 8.0), 150mmol/L NaCl, 1% sodium dodecyl sulfate (SDS), 1% Igepal CA-630, 0.5% sodium deoxycholate, 1mmol/L NaF, 10lM/mL phenylmethylsulfonyl fluoride (PMSF), 1mmol/L Na3 VO4, and protease inhibitor cocktail (Hoffmann-La Roche, Mississauga, ON, Canada) on ice. Homogenates were incubated on ice for 30minutes and sonicated twice for 20seconds on ice. Protein concentrations were determined with the bicinchoninic acid protein assay reagent (Pierce/Thermo Fisher Scientific, Nepean, ON, Canada). Proteins were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE). A 30-μg quantity of each homogenate was loaded into each lane. For western blotting, the gel was transferred to polyvinylidene difluoride membrane (PVDF, BioRad) and blocked in 5% (w/v) nonfat dry milk and 0.05% Tween 20 in Tris-buffered saline solution (pH 7.6) for 1hour, and then incubated with either panx1 polyclonal antibody (1:5000 dilution; Millipore AB9886; lot LV140345) or β-actin monoclonal antibody (1:10,000 dilution; Stressgen, 905-733-100) in Tris-buffered saline solution with 3% nonfat dry milk and 0.05% Tween 20, overnight. Membranes were washed and incubated with antirabbit secondary antibody (Cedarlane, Burlington, ON, Canada) conjugated with horseradish peroxidase. The immunoreaction was detected with a chemiluminescence kit (Amersham Biosciences/GE Healthcare, Baie d’Urfe, QC, Canada). After the x-ray films were scanned, the signal intensities of the bands were analyzed with ImageJ software (NIH, Bethesda, MD).
2.7. Statistical analysis
All data were expressed as the mean±standard error of the mean (SEM), and statistical analyses were made by using Prism 5.0 software (GraphPad Software, San Diego, CA). P values of less than .05 were considered statistically significant. Intragroup and intergroup statistical analyses of electromyographic and behavioral time-course curves were carried out using 2-way repeated-measures analysis of variance (ANOVA) followed by the Bonferroni multiple comparisons test. Statistical analysis of area under dose–response curves obtained from behavioral testing was performed by using 1-way ANOVA followed by the Dunnett test. Expression levels of panx1 from neuropathic and control animals were compared by means of 2-tailed unpaired Student t test.
3. Results
3.1. Pharmacological blockage of panx1 depressed the spinal wind-up activity in neuropathic and sham rats
Because central sensitization results from neuroplastic changes in the excitability of pain-transmitting neurons in the spinal cord, we tested electrophysiologically, using the C-nociceptive reflex paradigm, whether different panx1 blockers could inhibit spinal wind-up activity in neuropathic rats, a form of plasticity in spinal dorsal horn that is observed during low frequency (eg, 1Hz) electrical stimulation of C fibers. Results showed that application of a train of electric pulses to toes, at 1Hz, evoked wind-up activity on C-fiber–mediated reflex responses in the ipsilateral biceps femoris muscle, in both sham and neuropathic sural nerve–transected rats. The i.t. administration of any of the panx1-blocking drugs, 10panx, Cbx, or Prb, was able to depress the wind-up activity in both neuropathic and sham rats. Fig. 1B shows that 300μmol/L 10panx, but not saline solution, significantly decreased wind-up scores in both group of animals (*P<.05, 2-way ANOVA followed by Bonferroni post hoc test). Fig. 2 shows that 100μmol/L Cbx (A) and 150μM Prb (B) also result in significant decreases of wind-up scores in neuropathic and sham rats (for both drugs, *P<.05, 2-way ANOVA followed by Bonferroni post hoc test). The fact that pharmacological inactivation of panx1 channels led to a significant reduction of the wind-up activity in spinal cord of neuropathic and sham animals supports the notion that panx1 channels are involved in the mechanisms underlying the wind-up phenomenon under nonsensitized and sensitized spinal cords in rats.
Fig. 2:
Effect of intrathecal administration of carbenoxolone (Cbx) or probenecid (Prb) on spinal wind-up activity in sham control and neuropathic rats. Data were processed in a similar way to those from 10panx-injected animals. Wind-up scores at time T are expressed as percent change of the slope at time T compared to the slope before drug (or saline solution) injection at time 0. Values are mean ± SEM, n = 8 rats per group. (A) Time-course of wind-up activity (percent change) after intrathecal administration of saline solution (squares) or 100 μmol/L Cbx (upward-pointing triangles) in sham (open symbols) and neuropathic (closed symbols) rats. Two-way analysis of variance (ANOVA) for repeated measures revealed a “time” effect (intragroup analysis: F9,36 = 4.17, ANOVA P < .001), and a “group” effect (intergroup analysis: F1,36 = 60.09, ANOVA P < .0015; * P < .05 with respect to corresponding saline solution values, Bonferroni multiple comparisons test). (B) Time-course of wind-up activity (percent change) after intrathecal administration of saline solution (squares) or 150 μmol/L Prb (downward-pointing triangles), in sham (open symbols) and neuropathic (closed symbols) rats. Two-way ANOVA for repeated measures revealed a “time” effect (intragroup analysis: F9,36 = 2.25, ANOVA P = .0413), and a “group” effect (intergroup analysis: F1,36 = 41.06, ANOVA P = .003; * P < .05 with respect to corresponding saline solution values, Bonferroni multiple comparisons test).
3.2. Pharmacological blockage of panx1 decreased mechanical hyperalgesia in neuropathic rats but did not affect mechanical nociception in sham animals
Transection of the sural nerve resulted in long-lasting hyperalgesia to mechanical stimuli, as evidenced by a significant reduction of the paw withdrawal threshold in the experimental groups. The hyperalgesic status began 3days after transecting the sural nerve and lasted for at least 10days (Figs. 3A and 4A, left panel, *P<.05, 2-way ANOVA followed by Bonferroni post hoc test). In contrast, the paw withdrawal threshold in the sham groups remained unchanged throughout the 10-day testing period (Figs. 3B and 4B, left panel).
Fig. 3:
Effect of intrathecal administration of 10panx on withdrawal threshold of neuropathic and sham control rats. (A) (Left panel) Time-course of changes in the mechanical nociceptive threshold (g/cm2) of neuropathic rats during the 10 days after surgery. Surgery was performed at time 0 (downward-pointing arrow). Two-way repeated-measures analysis of variance (ANOVA) revealed a “time” effect (intragroup analysis: F4,100 = 85.67, ANOVA P < 0.0001; * P < .05 with respect to withdrawal threshold before surgery, Bonferroni multiple comparisons test), but not a “group” effect (intergroup analysis: F4,100 = 0.13, ANOVA P = 0.9688). (Middle panel) Time-course of changes in the mechanical nociceptive threshold of neuropathic rats after intrathecal injection (upward-pointing arrow) of saline solution or 10, 30, 100, and 300 μmol/L 10panx on day 10 after surgery. Two-way ANOVA for repeated measures revealed a “time” effect (intragroup analysis: F8,200 = 32.49, ANOVA P = .0001) and a “group” effect (intergroup analysis: F4,200 = 20.42, ANOVA P = .0001; * P < .05 with respect to corresponding saline solution values, Bonferroni multiple comparisons test). (Right panel) Area under the curve of the middle panel graph; * P < .05, ** P < .01, as compared to saline solution (1-way ANOVA followed by Dunnett’s multiple comparisons test). (B) (Left panel) Time-course of changes in the mechanical nociceptive threshold (g/cm2) of sham control rats in the 10 days after sham surgery. Surgery was performed at time 0 (downward arrow). Two-way repeated-measures ANOVA did not reveal either a “time” effect (intragroup analysis: F5,50 = 1.22, ANOVA P = .3151) or a “group” effect (intergroup analysis: F4,50 = 1.42, ANOVA P = .2959). (Right panel) Time-course of changes in the mechanical nociceptive threshold of sham control rats after intrathecal injection (upward-pointing arrow) of saline solution or 10, 30, 100, and 300 μmol/L 10panx on day 10 after sham surgery. Two-way repeated-measures ANOVA did not reveal either a “time” effect (intragroup analysis: F8,80 = .87, ANOVA P = 0.5449) or a “group” effect (intergroup analysis: F4,80 = 0.12, ANOVA P = .9702).
Fig. 4:
Effect of intrathecal administration of carbenoxolone (Cbx) or probenecid (Prb) on withdrawal threshold of neuropathic and sham control rats. (A) (Left panel) Time-course of changes in the mechanical nociceptive threshold (g/cm2) of neuropathic rats in the 10 days after surgery. Surgery was performed at time 0 (downward-pointing arrow). Two-way repeated-measures analysis of variance (ANOVA) revealed a “time” effect (intragroup analysis: F4,48 = 91.92, ANOVA P < .0001; * P < .05 with respect to withdrawal threshold before surgery, Bonferroni multiple comparisons test), but not a “group” effect (intergroup analysis: F2,48 = 1.27, ANOVA P = .3628). (Middle panel) Time-course of changes in the mechanical nociceptive threshold of neuropathic rats after intrathecal injection (upward-pointing arrow) of saline solution, 100 μmol/L Cbx, or 150 μmol/L Prb on day 10 after surgery. Two-way repeated-measures ANOVA revealed a “time” effect (intragroup analysis: F8,72 = 20.48, ANOVA P < .0001) and a “group” effect (intergroup analysis: F2,72 = 16.83, ANOVA P = .0009; * P < .05 with respect to corresponding saline solution values, Bonferroni multiple comparisons test). (Right panel) Area under the curves of the middle panel graph; # P < .05, as compared to saline solution (1-way ANOVA followed by Dunnett multiple comparisons test). (B) (Left panel) Time-course of changes in the mechanical nociceptive threshold (g/cm2) of sham control rats in the 10 days after sham surgery. Surgery was performed at time 0 (downward-pointing arrow). Two-way repeated-measures ANOVA did not reveal either a “time” effect (intragroup analysis: F5,30 = 0.37, ANOVA P = .8625) or a “group” effect (intergroup analysis: F2,30 = 3.03, ANOVA P = .1234). (Right panel) Time-course of changes in the mechanical nociceptive threshold of sham control rats after intrathecal injection (upward-pointing arrow) of 100 μmol/L Cbx or 150 μmol/L Prb on day 10 after sham surgery. Two-way repeated-measures ANOVA did not reveal either a “time” effect (intragroup analysis: F8,64 = 2.72, ANOVA P = .119) or a “group” effect (intergroup analysis: F2,64 = 0.89, ANOVA P = .4483).
The i.t. injection of increasing doses of 10panx (10, 30, 100, or 300μmol/L 10panx), but not saline solution, induced a significant, dose-dependent antihyperalgesic effect on neuropathic animals, as both the peak of the effect and the duration of the antihyperalgesic response were directly proportional to the dose (Fig. 3A, middle panel, *P<.05, 2-way ANOVA followed by Bonferroni post hoc test). Furthermore, calculation of the area under the time-course curves (AUC), as a measure of the effect of 10panx during the complete period of observation, also showed that the antihyperalgesic effect of 10panx behaves in a dose-dependent manner (Fig. 3A, right panel, *P<0.05, **P<0.01, 1-way ANOVA followed by Dunnett post hoc test). The antihyperalgesic effect of 10panx i.t. was transient, and the mechanical nociceptive threshold returned to the initial hyperalgesic score in around 3 to 4hours. In contrast to the effect of 10panx on neuropathic rats, the i.t. injection of similar doses of 10panx to sham rats did not affect the paw withdrawal threshold in these animals (Fig. 3B, right panel). Similarly to 10panx, the i.t. injections of 100μmol/L Cbx and 150μmol/L Prb, but not saline solution, were able to induce significant, transient antihyperalgesic effects on neuropathic rats, as can be observed in time-course curves (Fig. 4A, middle panel, *P<.05, 2-way ANOVA followed by Bonferroni post hoc test) and AUCs (Fig. 4A, right panel, #P<.05, 1-way ANOVA followed by Dunnett post hoc test). However, similar doses of Cbx and Prb were ineffective in sham animals (Fig. 4B, right panel). As a whole, the behavioral results suggest that panx1 participates in the central sensitization process that develops after a neuropathic injury.
3.3. Panx1 is present, and is similarly expressed, in the lumbar spinal cord of neuropathic and sham rats
Panx1 protein expression in the lumbar spinal cord of neuropathic and control rats was characterized by western blots with commercially available antibodies and normalized to β-actin. Panx1 runs as a band of 48kDa. No significant difference in expression level for panx1 protein was observed when comparing normalized band densities from lumbar spinal cord of neuropathic and sham controls (data not shown), thus suggesting that neuropathic injury did not affect genetic or posttranslational panx1 regulation.
4. Discussion
To study the role of panx1 channels in chronic pain, we used a spared nerve injury model of rat neuropathy involving the lesion of 1 (the sural nerve) of the 3 terminal branches of the sciatic nerve, leaving the other 2 branches (tibial and common peroneal nerves) intact. Such a model resulted in mechanical hyperalgesia, as evidenced by a significant reduction of the paw withdrawal threshold to pressure stimuli lasting for at least 10days, which is in agreement with previous studies showing that transection of 1 or more terminal branches of the sciatic nerve results in early, prolonged, and robust hypersensitivity to mechanical stimuli with von Frey monofilaments [15,26]. Together with the reduced threshold to pressure stimulus observed in neuropathic rats, a decreased stimulating current was required for threshold activation of the C-reflex from cutaneous regions (second and third toes) neighboring the denervated area, which supports earlier results showing that the threshold for electrical activation of C fibers was significantly reduced in rats subjected to spared nerve injury [23].
In addition to their contribution to the physiology of nonsynaptic communication, there is convincing evidence supporting a role of panx1 channels in some pathologies of the central nervous system, such as epilepsy, stroke, and neuroinflammation [17,44,46,50,51]. Nevertheless, the participation of panx1 on pain transmission and/or modulation had not yet been addressed in the literature, with the use of panx1 channel blockers nor by using panx1 KO mice. Here, we show for the first time that panx1 channels play a significant role in neuropathic pain, because (i) blockage of panx1 channels with intrathecal 10panx, Cbx, or Prb significantly depressed the spinal wind-up activity both in neuropathic and sham rats; (ii) intrathecal 10panx, Cbx, or Prb decreased mechanical hyperalgesia in neuropathic rats without affecting mechanical nociception in sham animals; and (iii) panx1 protein was expressed in the dorsal horn of the lumbar spinal cord of neuropathic and sham rats. The antihyperalgesic effect of 10panx, Cbx, or Prb on the hindpaw pressure threshold of neuropathic rats was short and reversible, meaning that the sensitized response was maintained, at least in part, by an active, panx1-dependent spinal mechanism that was transiently counteracted by the i.t. injection of the panx1 blocker. This mechanism is likely to be sited in neurons and/or in glial cells of the spinal cord, as we found that panx1 is expressed in the lumbar spinal cord, and, in addition, it is known that in nervous tissue panx1 is expressed in neurons as well as in astrocytes [20,50]. As it is known, in neuropathic pain states, the spinal hypersensitivity is generated by sustained co-release of glutamate, ATP, and peptides from peripheral nerves, allowing activation of NMDA, purinergic, neurokinin, and TrkB receptors [4,53], but also by tonic activation of central descending pathways that facilitate pain transmission [3,34]. In addition, central sensitization is increased by non-neural glial cells in the spinal cord, which are activated after peripheral nerve injury, and causes these cells to enhance pain by releasing neuroexcitatory glial proinflammatory glutamate, ATP, and cytokines [5,15,18,22,31,48]. Accordingly, antagonists of the relevant neuronal transmitters and glial mediators have been shown to prevent or even to inhibit the already established sensitization process [10,11,13,21,52].
In spinal cord, central sensitization is created by immediate-onset, transcription-independent, NMDA receptor-mediated mechanisms [4], such as wind-up and long-term potentiation [43], and afterward maintained by late-onset, transcription-dependent mechanisms that include ERK signaling to the nucleus and subsequent phosphorylation of CREB [57], with increased expression of early genes codifying for c-fos, cyclooxygenase-2 (COX2), and the neurotrophin brain-derived neurotrophic factor (BDNF) [9,12], as well as for other late-response genes [42]. The fact that intrathecal injection of any of the 3 panx1 blockers used inhibited wind-up phenomena in normal rats suggests that panx1 channels play a role in some spinal pain-amplifying mechanism of the C-fiber–mediated responses, similarly to (or in combination with) that subserved by postsynaptic NMDA receptors existing in second-order nociceptive neurons. More specifically, panx1 channels could be working as a source of inward calcium currents secondary to the activation of NMDA receptors, as described by Thompson et al. [50] in the hippocampus. In this system, pharmacological or genetic inactivation of panx1 leads to a decrease in the calcium influx or dye uptake in hippocampus slices, thereby diminishing this secondary current [50]. Activation of NMDA receptors has been demonstrated as crucial in the process of generating hyperalgesia during neuropathy by many authors, and entry of extra calcium amounts to the postsynaptic densities in the spinal cord via panx1 channels may play a similar role. Two nonexclusive mechanisms have been proposed to explain the opening of panx1 channels after NMDA receptor activation in neurons co-expressing both transcripts; the first suggests that calcium entry via NMDA receptors can exert a calcium-mediated opening of panx1 [33], and the second shows that NMDA receptor activation can open panx1 channels via associated Src family kinases [55]. Nevertheless, in addition to NMDA receptor activation, opening of panx1 channels (and subsequent calcium entry) could also be explained by activation of P2X7 receptors in cells co-expressing the 2 membrane proteins [27,36], through a mechanism involving P2X7 receptor–panx1 coupling via Src tyrosine kinase pathway [20]. P2X7 receptor–panx1 interaction have been described in neurons and glia; however, our results using western blotting could not resolve whether panx1 channels in the spinal cord were expressed in neurons or glial cells. In addition, lumbar spinal cord samples from neuropathic rats showed similar expression levels of panx1 to those taken from sham control animals, thus suggesting that panx1 is not overexpressed in spinal cord tissue after nerve injury. However, this result awaits immunohistochemical confirmation because restricted regulation of panx1 expression in a small compartment of the lumbar spinal cord (ie, some specific dorsal horn layer) cannot be resolved by western blotting. In any case, irrespective of whether panx1 channels are or are not upregulated in some dorsal horn cell types, the present results showed that intrathecal administration of panx1 inhibitors depressed wind-up activity in rats with spare nerve injury. As a whole, the depressing effect of panx1 inhibitors on wind-up activity suggests that the panx1-dependent pain–amplifying spinal mechanism is still active in hyperalgesic neuropathic rats. As has already been pointed out, despite being a rapid-onset and short-lasting phenomenon, wind-up may contribute to long-term changes in the spinal cord, leading to central sensitization if wind-up is maintained over time, that is, when is produced by peripheral formalin injection [19,39]. This is also in agreement with the current view that after nerve injury, central sensitization is maintained by continuing input from the periphery and from central regions [49], where wind-up and long-term potentiation are the likely mechanisms by which the response to a given input is increased [3]. The fact that all the 3 panx1 channel blockers were able to depress the spinal wind-up activity, but not the behavioral nociceptive response in sham animals, could be related to the different protocols of noceptive stimulation used. In fact, although wind-up studies used repetitive suprathreshold activation of C-afferent fibers to the dorsal horn, thus recruiting NMDA receptors, behavioral nociception was tested by using threshold phasic paw pressure, which does not involve the NMDA [41] but rather the AMPA receptors [24]. Besides, the more long-lasting effect of panx1-blocker drugs in the electrophysiological setting than in the behavioral setting could be due to a differential sensitivity to panx1 blockers of neurons that respond to tonic electric stimulation (wind-up) versus those responding to phasic mechanical nociceptive stimulation (withdrawal reflex).
Finally, calcium entry into dorsal horn nociceptive neurons via activated panx1 channels not only could trigger short-term plastic mechanisms underlying enhanced neuronal excitability (such as wind-up or long-term potentiation) but also could contribute to set up transcription-dependent, long-term changes involved in later stages of the central sensitization process, as propagation of calcium signals into the nucleus is a major route for inducing changes in the expression of genes involved in neuroplasticity [1]. This suggests that panx1 could play a role not only in the triggering of central sensitization in spinal cord of neuropathic rats but also in the maintenance of the process, which is in line with the fact that, in the present study, intrathecal administration of any of the panx1 channel blockers was able to inhibit an already well-established process of mechanical hyperalgesia.
4.1. Conclusion
In conclusion, although the present experiments do not account for the physiological mechanism by which panx1 actually enhances the pain response, it seems likely that panx1 is part of the cellular machinery governing short-term neuroplastic responses in spinal cord, such as wind-up, and thereby the triggering and maintenance of the hyperalgesic response observed on neuropathic rats. Several lines of evidence have demonstrated that panx1 is activated during well-known cell death paradigms, such as ATP or NMDA receptor overactivation and ischemia. Whether or not panx1 channels are overexpressed or remain persistently active during neuropathic pain are currently open questions. More conclusive studies using panx1 KO animals or neurons are necessary in the future.
Conflict of interest statement
The authors declare no conflict of interest with regard to this work.
Acknowledgements
This work was supported by grant 1120952 from FONDECYT and grant FB0807 from CEDENNA. The authors thank Ms. Cristina Arenas for technical support and animal care.
References
[1]. Bading H. Nuclear calcium signalling in the regulation of brain function.
Nat Rev Neurosci. 2013;14:593-608.
[2]. Bao L, Locovei S, Dahl G. Pannexin membrane channels are mechanosensitive conduits for ATP.
FEBS Lett. 2004;572:65-68.
[3]. Baron R, Hans G, Dickenson AH. Peripheral input and its importance for central sensitization.
Ann Neurol. 2013;74:630-636.
[4]. Baron R. Mechanisms of disease: neuropathic pain—a clinical perspective.
Nat Clin Pract Neurol. 2006;2:95-106.
[5]. Bennett GJ. Update on the neurophysiology of pain transmission and modulation: focus on the NMDA-receptor.
J Pain Symptom Manage. 2000;19:S2-S6.
[6]. Bruzzone R, Hormuzdi SG, Barbe MT, Herb A, Monyer H. Pannexins, a family of gap junction proteins expressed in brain.
Proc Natl Acad Sci USA. 2003;100:13644-13649.
[7]. Bruzzone R, Barbe MT, Jakob NJ, Monyer H. Pharmacological properties of homomeric and heteromeric pannexin hemichannels expressed in
Xenopus oocytes.
J Neurochem. 2005;92:1033-1043.
[8]. Burnstock G. Purinergic receptors and pain.
Curr Pharm Des. 2009;15:1717-1735.
[9]. Castren E, Pitkanen M, Sirvio J, Parsadanian A, Lindholm D, Thoenen H, Riekkinen PJ. The induction of LTP increases BDNF and NGF mRNA but decreases NT-3 mRNA in the dentate gyrus.
NeuroReport. 1993;4:895-898.
[10]. Chizh BA, Headley PM. NMDA antagonists and neuropathic pain—multiple drug targets and multiple uses.
Curr Pharm Des. 2005;11:2977-2994.
[11]. Clark AK, Old EA, Malcangio M. Neuropathic pain and cytokines: current perspectives.
J Pain Res. 2013;6:803-814.
[12]. Cole AJ, Saffen DW, Baraban JM, Worley PF. Rapid increase of an immediate early gene messenger RNA in hippocampal neurons by synaptic NMDA receptor activation.
Nature. 1989;340:474-477.
[13]. Constandil L, Goich M, Hernández A, Bourgeais L, Cazorla M, Hamon M, Villanueva L, Cyclotraxin-B Pelissier T. A new TrkB antagonist, and glial blockade by propentofylline, equally prevent and reverse cold allodynia induced by BDNF or partial infraorbital nerve constriction in mice.
J Pain. 2012;13:579-589.
[14]. Constandil L, Hernández A, Pelissier T, Arriagada O, Espinoza K, Burgos H, Laurido C. Effect of interleukin-1beta on spinal cord nociceptive transmission of normal and monoarthritic rats after disruption of glial function.
Arthritis Res Ther. 2009;11:R105.
[15]. De Leo JA, Tawfik VL, LaCroix-Fralish ML. The tetrapartite synapse: path to CNS sensitization and chronic pain.
PAIN®. 2006;122:17-21.
[16]. Decosterd I, Woolf CJ. Spared nerve injury: an animal model of persistent peripheral neuropathic pain.
PAIN®. 2000;87:149-158.
[17]. Dvoriantchikova G, Ivanov D, Barakat D, Grinberg A, Wen R, Slepak VZ, Shestopalov VI. Genetic ablation of Pannexin1 protects retinal neurons from ischemic injury.
PLoS ONE. 2012;7:e31991.
[18]. Gwak YS, Kang J, Unabia GC, Hulsebosch CE. Spatial and temporal activation of spinal glial cells: role of gliopathy in central neuropathic pain following spinal cord injury in rats.
Exp Neurol. 2012;234:362-372.
[19]. Haley JE, Sullivan AF, Dickenson AH. Evidence for spinal N-methyl-D-aspartate receptor involvement in prolonged chemical nociception in the rat.
Brain Res. 1990;518:218-226.
[20]. Iglesias R, Locovei S, Roque A, Alberto AP, Dahl G, Spray DC, Scemes E. P2X7 receptor-pannexin1 complex: pharmacology and signaling.
Am J Physiol Cell Physiol. 2008;295:C752-C760.
[21]. Jaggi AS, Singh N. Therapeutic targets for the management of peripheral nerve injury-induced neuropathic pain.
CNS Neurol Disord Drug Targets. 2011;10:589-609.
[22]. Jarvis M. The neural–glial purinergic receptor ensemble in chronic pain states.
Trends Neurosci. 2010;33:48-57.
[23]. Kohno T, Moore KA, Baba H, Woolf CJ. Peripheral nerve injury alters excitatory synaptic transmission in lamina II of the rat dorsal horn.
J Physiol. 2003;548:131-138.
[24]. Kong LL, Yu LC. It is AMPA receptor, not kainate receptor, that contributes to the NBQX-induced antinociception in the spinal cord of rats.
Brain Res. 2006;1100:73-77.
[25]. Larsson M. Ionotropic glutamate receptors in spinal nociceptive processing.
Mol Neurobiol. 2009;40:260-288.
[26]. Lee BH, Won R, Baik EJ, Lee SH, Moon CH. An animal model of neuropathic pain employing injury to the sciatic nerve branches.
NeuroReport. 2000;11:657-661.
[27]. Locovei S, Wang J, Dahl G. Activation of pannexin 1 channels by ATP through P2Y receptors and by cytoplasmic calcium.
FEBS Lett. 2006;580:239-244.
[28]. Ma W, Hui H, Pelegrin P, Surprenant A. Pharmacological characterization of pannexin-1 currents expressed in mammalian cells.
J Pharmacol Exp Ther. 2009;328:409-418.
[29]. Mestre C, Pélissier T, Fialip J, Wilcox G, Eschalier A. A method to perform direct transcutaneous intrathecal injection in rats.
J Pharmacol Toxicol Methods. 1994;32:197-200.
[30]. Millan MJ. The induction of pain: an integrative review.
Prog Neurobiol. 1999;57:1-164.
[31]. Milligan ED, Watkins LR. Pathological and protective roles of glia in chronic pain.
Nat Rev Neurosci. 2009;10:23-36.
[32]. Mylvaganam S, Zhang L, Wu C, Zhang ZJ, Samoilova M, Eubanks J, Carlen PL, Poulter MO. Hippocampal seizures alter the expression of the pannexin and connexin transcriptome.
J Neurochem. 2010;112:92-102.
[33]. Orellana JA, Froger N, Ezan P, Jiang JX, Bennett MV, Naus CC, Giaume C, Sáez JC. ATP and glutamate released via astroglial connexin 43 hemichannels mediate neuronal death through activation of pannexin 1 hemichannels.
J Neurochem. 2011;118:826-840.
[34]. Ossipov MH, Lai J, Malan TP Jr, Porreca F. Spinal and supraspinal mechanisms of neuropathic pain.
Ann N Y Acad Sci. 2000;909:12-24.
[35]. Panchin Y, Kelmanson I, Matz M, Lukyanov K, Usman N, Lukyanov S. A ubiquitous family of putative gap junction molecules.
Curr Biol. 2000;10:R473-R474.
[36]. Pelegrin P, Surprenant A. Pannexin-1 mediates large pore formation and interleukin-1beta release by the ATP-gated P2X7 receptor.
EMBO J. 2006;25:5071-5082.
[37]. Pelissier T, Infante C, Constandil L, Espinosa J, Lapeyra CD, Hernández A. Antinociceptive effect and interaction of uncompetitive and competitive NMDA receptor antagonists upon capsaicin and paw pressure testing in normal and monoarthritic rats.
PAIN®. 2008;134:113-127.
[38]. Peyronnard JM, Charron L. Motor and sensory neurons of the rat sural nerve: a horseradish peroxidase study.
Muscle Nerve. 1982;5:654-660.
[39]. Pezet S, Marchand F, D’Mello R, Grist J, Clark AK, Malcangio M, Dickenson AH, Williams RJ, McMahon SB. Phosphatidylinositol 3-kinase is a key mediator of central sensitization in painful inflammatory conditions.
J Neurosci. 2008;28:4261-4270.
[40]. Randall L, Selitto J. A method for measurement of analgesic activity on inflamed tissue.
Arch Int Pharmacodyn Ther. 1957;111:409-419.
[41]. Ren K, Dubner R. NMDA receptor antagonists attenuate mechanical hyperalgesia in rats with unilateral inflammation of the hindpaw.
Neurosci Lett. 1993;163:22-26.
[42]. Ryan M, Ryan B, Williams J. Temporal profiling of gene networks associated with the late phase of long-term potentiation in vivo.
PLoS ONE. 2012;7:e40538.
[43]. Sandkühler J, Gruber-Schoffnegger D. Hyperalgesia by synaptic long-term potentiation (LTP): an update.
Curr Opin Pharmacol. 2012;12:18-27.
[44]. Santiago MF, Veliskova J, Patel NK, Lutz SE, Caille D, Charollais A, Meda P, Scemes E. Targeting pannexin1 improves seizure outcome.
PLoS ONE. 2011;6:e25178.
[45]. Silverman W, Locovei S, Dahl G. Probenecid, a gout remedy, inhibits pannexin 1 channels.
Am J Physiol Cell Physiol. 2008;295:C761-C767.
[46]. Silverman WR, De Rivero-Vaccari JP, Locovei S, Qiu F, Carlsson SK, Scemes E, Keane RW, Dahl G. The pannexin 1 channel activates the inflammasome in neurons and astrocytes.
J Biol Chem. 2009;284:18143-18151.
[47]. Smith JC, Bolon B. Isoflurane leakage from non-rebreathing rodent anaesthesia circuits: comparison of emissions from conventional and modified ports.
Lab Anim. 2006;40:200-209.
[48]. Sorge RE, Trang T, Dorfman R, Smith SB, Beggs S, Ritchie J, Austin JS, Zaykin DV, Vander Meulen H, Costigan M, Herbert TA, Yarkoni-Abitbul M, Tichauer D, Livneh J, Gershon E, Zheng M, Tan K, John SL, Slade GD, Jordan J, Woolf CJ, Peltz G, Maixner W, Diatchenko L, Seltzer Z, Salter MW, Mogil JS. Genetically determined P2X7 receptor pore formation regulates variability in chronic pain sensitivity.
Nat Med. 2012;18:595-599.
[49]. Suzuki R, Dickenson A. Spinal and supraspinal contributions to central sensitization in peripheral neuropathy.
Neurosignals. 2005;14:175-181.
[50]. Thompson RJ, Jackson MF, Olah ME, Rungta RL, Hines DJ, Beazely MA, MacDonald JF, MacVicar BA. Activation of pannexin-1 hemichannels augments aberrant bursting in the hippocampus.
Science. 2008;322:1555-1559.
[51]. Thompson RJ, Zhou N, MacVicar BA. Ischemia opens neuronal gap junction hemichannels.
Science. 2006;312:924-927.
[52]. Tsuda M, Tozaki-Saitoh H, Inoue K. Purinergic system, microglia and neuropathic pain.
Curr Opin Pharmacol. 2012;12:74-79.
[53]. Von Hehn CA, Baron R, Woolf CJ. Deconstructing the neuropathic pain phenotype to reveal neural mechanisms.
Neuron. 2012;73:638-652.
[54]. Wang H, Cao Y, Chiang CY, Dostrovsky JO, Sessle BJ. The gap junction blocker carbenoxolone attenuates nociceptive behavior and medullary dorsal horn central sensitization induced by partial infraorbital nerve transection in rats.
PAIN®. 2014;155:429-435.
[55]. Weilinger N, Tang PL, Thompson RJ. Anoxia-induced NMDA receptor activation opens pannexin channels via Src family kinases.
J Neurosci. 2012;32:12579-12588.
[56]. Zimmermann M. Ethical guidelines for investigations of experimental pain in conscious animals.
PAIN®. 1983;16:109-110.
[57]. Zimmermann M. Pathobiology of neuropathic pain.
Eur J Pharmacol. 2001;429:23-37.
[58]. Zoidl G, Petrasch-Parwez E, Ray A, Meier C, Bunse S, Habbes HW, Dahl G, Dermietzel R. Localization of the pannexin1 protein at postsynaptic sites in the cerebral cortex and hippocampus.
Neuroscience. 2007;146:9-16.
