Reactive oxygen species (ROS) play an important role in a... : PAIN (original) (raw)

1 Introduction

Reactive oxygen species (ROS) have been implicated in many degenerative neurological conditions such as Alzheimer's disease, Parkinson's disease, amyotropic lateral sclerosis, as well as brain dysfunction due to injury or aging (Balazs and Leon, 1994; Coyle and Puttfarcken, 1993; Gerlach, 1994; Götz et al., 1994; Hensley et al., 1997; Jenner, 1994; Lewen et al., 2000; Olanow, 1992).

There are multiple sources of ROS production in nervous tissue. For example, various normal enzyme reactions produce ROS as byproducts and the major source of intracellular ROS is leakage from mitochondria (Halliwell and Gutteridge, 1989; Schapira and Cooper, 1992). Normally, intracellular ROS is beneficial, such as protecting against invading pathogens, and ROS levels are precisely controlled by various enzymatic activities. In pathological conditions, however, intracellular ROS levels increase due to increased production or impaired removal, and cause cell damage ranging from cytoplasmic swelling to death. Consequently, removal of excessive ROS is often important for restoring normal conditions.

In spite of much work on ROS, studies showing the involvement of ROS in chronic pain are limited. In the chronic constriction injury (CCI) model of rat neuropathic pain, heat hyperalgesia was reduced by systemically injected antioxidants (Khalil et al., 1999; Tal, 1996). Although these studies suggest the involvement of free radicals in the generation of heat hyperalgesia, little attention has been paid to the critical role of ROS in neuropathic pain, perhaps because the effect was relatively small and detailed studies have not followed. The present study examines more extensively the involvement of ROS in neuropathic pain using potent ROS scavengers. ‘Spin-trap’ reagents are the most potent ROS scavengers, with phenyl-_N-tert_-butylnitrone (PBN) being especially noteworthy (Tizot et al., 2000). The present study examines the effect of three ROS scavengers, particularly PBN, in the spinal nerve ligation (SNL) model of neuropathic pain (Kim and Chung, 1992). These ROS scavengers are non-specific in that they scavenge all types of ROS indiscriminately, including superoxides, hydroxyl radicals, and peroxynitrites. Our data suggest that ROS play a major role in mechanisms of neuropathic pain in this model.

Preliminary data were presented in abstract form (Kim et al., 2002, 2003).

2 Materials and methods

2.1 Experimental animals

Male adult Sprague–Dawley rats (200–250 g) were used in this study. Animals were housed in groups of two or three in plastic cages with soft bedding and free access to food and water under a 12/12 h reversed light–dark cycle (dark cycle: 8:00 a.m.–8:00 p.m.). All animals were acclimated in their cages for 1 week before any experiments, which were carried out in accordance with the National Institute of Health's Guide for the Care and Use of Laboratory Animals.

2.2 Neuropathic surgery (spinal nerve ligation, SNL)

Rats were anesthetized with halothane (3% induction, 2% maintenance) in O2 and then the L5 spinal nerve was tightly ligated (Kim and Chung, 1992). Briefly, each animal was placed in the prone position and the left L6 vertebral transverse process was removed under the dissection microscope. The left L5 spinal nerve could then be identified, separated from the adjacent L4 spinal nerve and tightly ligated using 6–0 silk thread. The wound was treated with 10% povidone iodine solution and closed with wound clips. At the end of each experiment, the ligation sites were re-exposed and the location of the ligature confirmed.

2.3 Behavioral tests for mechanical thresholds

Behavioral tests were conducted blindly in that the experimenter who conducted the tests did not know the nature of the experimental manipulation. The behavioral tests measured foot withdrawal thresholds in response to mechanical stimuli applied to the left hind paw. For each test, the animal was placed in a plastic chamber (8.5×8.5×28 cm) and habituated for at least 15 min. The chamber was placed on top of a mesh screen so that mechanical stimuli could be administered to the plantar surface of left hind paw. Thresholds were determined by the up–down method (Baik et al., 2003; Chaplan et al., 1994; Dixon, 1980) using a set of von Frey monofilaments (von Frey numbers: 3.65, 3.87, 4.10, 4.31, 4.52, 4.74, 4.92 and 5.16; equivalent to: 0.45, 0.74, 1.26, 2.04, 3.31, 5.50, 8.32, and 14.45 g). A von Frey filament was applied perpendicularly to the most sensitive areas of the plantar surface at the base of the third or fourth toes with sufficient force to bend the filament slightly for 2–3 s. An abrupt withdrawal of the foot during stimulation or immediately after stimulus removal was a positive response. The first stimulus was always the 4.31 filament. When there was a positive response, the next lower filament was used, and when no response was obtained, the next higher filament was applied. This testing pattern continued until responses to six von Frey stimuli from the first change of response (either higher or lower than the first stimulus depending on whether the first response was negative or positive) were measured. The responses were then converted into a 50% threshold value using the formula: 50% threshold = 10(X+kd)/104, where X is the value of the final von Frey hair used in log units, k is the tabular value for the pattern of positive/negative responses, and d is the mean differences between stimuli in log units (0.22) (Dixon, 1980). When positive or negative responses were still observed at the end of a stimulus session, values of 3.54 or 5.27 were assigned, respectively, by assuming a value of ±0.5 for k in these cases. The behavioral data were plotted linearly in von Frey values.

2.4 Tests for sedation

To see whether ROS scavengers induce sedation or anesthesia, either of which interfere with posture and righting reflexes, as opposed to analgesia which does not, the following assessments were made (Devor and Zalkind, 2001).

Five-point scale for posture:

Five-point scale for righting reflexes:

2.5 ROS scavenging agents

ROS scavenging agents were obtained from Sigma Chemical Company (St Louis, MO, USA). Phenyl-_N-tert_-butylnitrone (PBN; MW, 177.24) was dissolved in saline and filtered using a paper of 0.2 μm pore size, 5,5-dimethylpyrroline-_N_-oxide (DMPO; MW, 113.16) was dissolved in saline, and nitrosobenzene (NBZ; MW, 107.11) was suspended in 1% Tween 80.

2.6 Intrathecal and intracerebroventricular injections

For spinal subarachnoid injections, intrathecal catheters were implanted at the time of SNL. Under halothane anesthesia, the hair was clipped and a posterior midline incision was made from T11 to L1. The paraspinal muscles were retracted and the posterior surfaces of the T12 and T13 vertebrae were exposed. The posterior articular processes and laminae of the T12 vertebra were removed with a pair of rongeurs to expose the spinal meninges. A small nick was made in the dura mater and a prepared catheter (sterilized PE 10 tubing filled with artificial cerebrospinal fluid, Becton Dickinson, Sparks, MD) was inserted into the subarachnoid space. The catheter was gently guided caudally until the tip reached the lumbar enlargement of the spinal cord (approximately 1 cm caudal to the initial insertion). The proximal end of the PE 10 tubing was then connected to PE 50 tubing, which was in turn fed subcutaneously to mid-thoracic levels with anchors by sutures to muscles and the tip was exposed in the dorsal midline. The external tip was then sealed and the incision closed. After recovery from anesthesia, rats were returned to their cages and housed individually.

Twenty-four rats were catheterized and at the end of the experiments, autopsies were performed and the spread of injected Evans blue dye (50 μl, 0.5%) was determined as an estimate of the spread of injected materials. Twenty of these 24 rats were satisfactory in that the tips of the catheters were at L5 levels and there was no obvious damage to the cord.

For intracerebroventricular injections, rats were anesthetized with halothane and placed in a stereotaxic apparatus. A small hole was drilled in the skull and a 30G needle was advanced into the right lateral ventricle (AP: −0.8 mm from the bregma; L: 1.6 mm from the midline; depth: 4 mm below the skull surface). The needle was connected to a Hamilton syringe with PE 10 tubing, and 50 μl of solution (saline or 1 mg of PBN in saline) was slowly delivered (22 min) by an infusion pump. The wound was then closed and anesthesia discontinued.

2.7 Ectopic discharge recording

Ectopic discharges were recorded using an in vivo recording set-up (Han et al., 2000). Five to 10 days after L5 SNL, rats were anesthetized by intraperitoneal injection of two doses of 0.6 g/kg urethane at 10-min intervals. Then the spinal cord was exposed by a laminectomy from L1 to L6. Rats were mounted on a spinal investigation frame and a heated mineral oil pool was made over the exposed tissue to prevent drying. The dura mater was opened and the L5 dorsal root was cut close to the cord. The cut dorsal root was placed on a mirror-based platform and teased into small fascicles with a pair of forceps under a dissecting microscope. Each teased fascicle was placed on two monopolar recording hook electrodes separated by 2–3 mm until a spontaneously active unit was found. If the recording contained more than one unit as evidenced by firing of units with different amplitudes, the fascicle was further divided. Single-unit activity was amplified with an AC-coupled amplifier (WPI, DAM-80), displayed on an oscilloscope and led to a window discriminator (Mentor N-750) whose output was used to generate histograms via a data acquisition system (CED 1401 with Spike 2). Conduction velocities of the recorded fibers were calculated by dividing the distance of two recording electrodes by the action potential delay between the two electrodes.

2.8 Data analysis

The overall design of the present experiments is the randomized Latin square design (Kirk, 1995). Data are presented as means± standard errors of the mean (SEMs) and analyzed using the SAS statistical program. Statistical analyses were done using one-way analyses of variance (ANOVAs), two-way repeated ANOVAs, or two-way ANOVAs with repeated time factors (Thomson, 1990), as appropriate, followed by Duncan post hoc tests.

3 Results

3.1 Effects of single injections of ROS scavengers

Fig. 1 shows changes in mechanical thresholds of neuropathic rats after intraperitoneal injections of graded doses of PBN (Fig. 1A) or DMPO (Fig. 1B) in accord with the randomized Latin square design. Thus, for PBN, SNLs were done on eight rats, which resulted in marked reductions in mechanical thresholds. On the third post-operative (PO) day, these eight rats were divided into four groups of two rats each and each group received either one of three doses of PBN or saline. Mechanical thresholds were determined for all animals for 24 h post-injection at the time points indicated (Fig. 1A). Then, all animals were rested for 24 h. On the fifth PO day, each group received either one of the PBN doses or saline, which they did not receive before, followed by the same paradigm of behavioral testing. This procedure was then repeated two more times, so the overall experiment took 10 days (injections on third, fifth, seventh, and ninth PO days followed by 24 h of testing after each injection). Thus all eight animals received all doses of PBN and saline and then had mechanical thresholds determined at the indicated times. Fig. 1A is a summary of these results, and each point represents tests on eight rats. Note that mechanical thresholds were markedly reduced by the third day after nerve ligation and then the 20 or 50 mg/kg of PBN raised these thresholds and the 100 mg/kg of PBN returned the thresholds essentially to normal.

F1-15

Fig. 1:

Effects of graded doses of systemic PBN (A) and DMPO (B) on mechanical thresholds in neuropathic rats. (A) After measurements of mechanical thresholds for 2 days, the L5 spinal nerves were ligated in eight rats and subsequently the thresholds were greatly reduced. On the third PO day, the rats were divided into four groups of two rats each, and one of three doses of PBN (20, 50 or 100 mg/kg) or saline was injected intraperitoneally, and behavioral tests were done for the next 24 h followed by 24 h of rest. The process was repeated three more times, so that every rat received all four treatments followed by behavioral testing after each. Thus each point represents the results from eight rats. Note the normal or almost normal thresholds after the 100 mg PBN injections. In B, the same paradigm was used but with DMPO (20, 50, and 100 mg/kg) and saline. The effects of DMPO are less in magnitude but longer lasting compared to PBN. Data are expressed as means±SEMs. Asterisks indicate significant differences from the saline control group by a two-way repeated ANOVA, followed by the Duncan post hoc test.

Fig. 1B shows the results of a similar study done with intraperitoneal injections of DMPO. The effects of DMPO lasted longer but were not as dramatic as for PBN.

3.2 A possible sedative or anesthetic effect of PBN

Since sedation or anesthesia could raise mechanical thresholds, it was necessary to determine if these contributed to the behavioral changes interpreted as analgesia. All rats shown in Fig. 1A were assessed for posture and righting reflexes based on the five-point scales described in methods, and all scored 0 at all time points, showing that sedation or anesthesia did not occur, so the behavioral changes likely represent analgesia.

3.3 Cumulative and long-term effects of PBN

To examine cumulative effects, 50 mg/kg of PBN were injected intraperitoneally three times at 4-h intervals in six rats on the sixth PO day (Fig. 2). Note that the effect of PBN is graded with repeated injections. Sedation scores were 0 in all these rats at all times showing that sedation is not a factor in these cumulative effects.

F2-15

Fig. 2:

Effects of repeated PBN injections on mechanical thresholds. After measurements of mechanical thresholds for 2 days, L5 spinal nerves were ligated in six rats. Six days after the ligations, PBN (50 mg/kg) was injected intraperitoneally every 4 h, and behavioral tests were repeated every hour after each injection. Note the steady increase in mechanical thresholds after each injection.

Long-term effects of repeated injections of PBN and DMPO were examined. Starting four to five days of post-lesion, either 10 mg/kg of PBN (Fig. 3A) or DMPO (Fig. 3B) were given intraperitoneally at 12 h intervals for 4 days (eight doses) (the control group received intraperitoneal injections of saline). Mechanical thresholds, measured 11–12 h after each injection but just before the following injection, were increased to a constant level as long as the drug injections were repeated but were returned to preinjection levels after the injections ceased. A similar paradigm, but starting 11.5 days of post-ligation, showed similar results. These data indicate that repeated injections of a low dose of PBN or DMPO (10 mg/kg) reduce mechanical thresholds with no signs of tolerance.

F3-15

Fig. 3:

Sustained effects of repeated PBN injections on neuropathic pain behaviors. In A, a low dose (10 mg/kg in saline) of PBN was given intraperitoneally (downward arrowheads) at 12-h intervals for 4 days, beginning 5.5 days after the neuropathic injury. Controls received the same volume of saline. Behavioral tests were conducted once a day. Note the great reduction in mechanical thresholds by 3 days post-ligation, that these thresholds are significantly elevated by 36 h after the first PBN dose, and that thresholds return to neuropathic levels within 24 h after the last PBN dose. Also note that there is no development of tolerance (loss of potency) and that the same analgesic effect is obtained if treatment stops and then resumes. In B, a similar experiment was done with DMPO (10 mg/kg). Data are expressed as means±SEMs. _n_=6 for A and B. Asterisks indicate values significantly (P<0.05) different from controls using a two-way ANOVA with a repeated time factor, followed by the Duncan post hoc test.

The efficacy of repeated injections of another ROS scavenger NBZ was also tested. At a dose of 10 mg/kg, the animals became lethargic and mildly cyanotic for the initial 0.5–1 h but then recovered to normal afterward. When tested 11–12 h after the first injection in neuropathic animals, the similar increases in mechanical thresholds was observed as with PBN and DMPO. This experiment was, however, discontinued due to the side effect.

3.4 Preemptive PBN treatment

To examine preemptive analgesic effects, PBN (10 mg/kg) was given intraperitoneally twice daily for 5 days starting 2.5 days prior to the neuropathic lesion (Fig. 4). Note the slow decline in mechanical thresholds after the lesion and that the thresholds remain elevated for several days after cessation of PBN. Thus there is a clear and relatively long-lasting preemptive effect.

F4-15

Fig. 4:

Effects of PBN pretreatment on the development of neuropathic pain behavior. Mechanical thresholds were measured, and then PBN (10 mg/kg in saline) was given intraperitoneally (downward arrowheads) twice a day for 5 days. Control animals received the same volumes of saline. Spinal nerve ligations were done 2.5 days after the first injection of PBN. Note that thresholds are significantly elevated in the PBN group compared to controls for 2 days after drug administration ceased (5 days after neuropathic surgery). Data are expressed as means±SEMs. _n_=6 and 7 for PBN and saline groups, respectively. Asterisks indicate values significantly (P<0.05) different from corresponding control values by using a two-way ANOVA with a repeated time factor, followed by the Duncan post hoc test.

3.5 Site of action: peripheral, spinal, and supraspinal sites

To examine peripheral actions of PBN, spontaneous single unit activity was recorded from teased L5 dorsal rootlets after L5 SNL (Fig. 5). Due to the long-lasting action of PBN, only one unit was tested from each rat. Recordings were made with two electrodes (separated by 2–3 mm) simultaneously and conduction velocity was calculated from the distance of the two electrodes and the delay of recorded activity between two electrodes. Intraperitoneal injections of PBN (100 mg/kg) did not reduce discharge rates (Fig. 5), suggesting that PBN does not act at the ligation site or the dorsal root ganglion where ectopic discharges originate.

F5-15

Fig. 5:

Effect of PBN on ectopic discharges. The L5 spinal nerve was ligated in 12 rats and ectopic discharges were recorded from the teased L5 dorsal root using an in vivo recording set-up. Recordings were made 5–10 days after spinal nerve ligation and only one unit was recorded per animal. Ten recorded units had conduction velocities in the Aβ fiber range (>14 m/s) while two others were Aδ fibers. After recording for 30 min, PBN (100 mg/kg) was injected intraperitoneally. Recording continued for the next 2 h. A shows ectopic discharges (1-s long traces) before, 30 min after, and 60 min after PBN injection. B shows normalized firing rates after PBN injection, the normalization being necessary due to the large variability of rates between individual units. The normalization was done by averaging 10 min firing rates for each individual unit and expressing them as the percent of the first 10 min of the basal rate and then the rates of all 12 units were averaged. Note that PBN did not significantly change the rate of ectopic discharges.

To examine spinal actions of spin-trap agents, intrathecal injections of PBN and DMPO were made in the randomized Latin square design. For PBN (Fig. 6A), intrathecal catheter implantations were done in nine rats followed by SNLs, which markedly reduced mechanical thresholds. On the seventh PO day, the nine rats were divided into four groups of two or three rats each and each group received one of the three intrathecal doses of PBN or saline. Mechanical thresholds were determined for all animals up to 24 h post-injection as indicated in Fig. 6A followed by a rest of 24 h. This procedure was repeated three times, with a different treatment each time, so the overall experiment took 8 days (injections on 7th, 9th, 11th, and 13th PO days followed by 24 h of testing after each injection) and each animal received each of the indicated substances and then had mechanical thresholds determined at the indicated times. Fig. 6A is a summary with each point representing responses from nine rats. Note that the 0.5 or 1.0 mg doses raised mechanical thresholds whereas the 0.1 mg dose did not. Higher doses (2.0 and 5.0 mg) did not produce any stronger effects than 1.0 mg (data not shown).

F6-15

Fig. 6:

Effects of graded doses of intrathecal PBN (A) and DMPO (B) on mechanical thresholds in neuropathic rats. (A) The L5 spinal nerve was ligated after measurements of mechanical thresholds for 2 days (P1 and P2) in nine rats. One week after surgery, thresholds were greatly reduced with von Frey values of 5.27 dropping to about 3.7. The rats were then divided randomly into four groups of two or three rats each and one of three doses of PBN (0.1, 0.5 or 1.0 mg in 50 μl of saline) or saline (50 μl) was injected intrathecally in each group and behavioral tests were repeated for the next 24 h followed by 24 h of rest. The procedure was repeated three times so that every rat received all treatments by the end of the experiment. Note the significant increases in von Frey thresholds. In B, the same paradigm was done with DMPO (0.1, 0.5, and 1.0 mg in 50 μl of saline) and saline in another nine rats with similar but less dramatic effects. Data are expressed as means±SEMs. Asterisks indicate significantly different levels as compared to corresponding values with the saline control group by a two-way repeated ANOVA, followed by the Duncan post hoc test.

Fig. 6B shows the results of a similar study done with intrathecal injections of DMPO. The effects for DMPO were similar but not as dramatic as for PBN.

Since intrathecal injections of ROS scavengers produced slightly less analgesia than systemic injections, the possibility of additional supraspinal sites that might be affected by ROS scavengers was tested by injecting PBN into the lateral ventricle. Fig. 7 shows that intraventricular injection of 1 mg of PBN (the most effective intrathecal dose) produced a significant but much smaller effect than intrathecal injection, thus showing a significant but small supraspinal effect of PBN in addition to the spinal effect.

F7-15

Fig. 7:

Effects of intracerebroventricular injections of PBN (1.0 mg) on mechanical thresholds in neuropathic rats. The L5 spinal nerve was ligated after measurements of mechanical thresholds for two days (P1 and P2) in 11 rats. One week after surgery, thresholds were greatly reduced with von Frey values of 5.27 dropping to about 3.7 (Base). The rats were then divided randomly into two groups and PBN (1.0 mg in 50 μl of saline, _n_=6) or saline (50 μl, _n_=5) was injected into the right lateral ventricle slowly (infusion for 22 min). Data are expressed as means±SEMs. Note that there is a significant rise in thresholds after PBN, but the effects are much less than after systemic or intrathecal administration. Asterisks indicate significantly different levels as compared to the corresponding values with the saline control group by a two-way ANOVA with a repeated time factor, followed by the Duncan post hoc test.

4 Discussion

This study shows that systemic ROS scavengers ameliorate the behavioral signs of mechanical allodynia in the SNL model of neuropathic pain. Based on tests for posture and righting reflexes, these compounds are not sedatives or general anesthetics for the doses we used, so the behavioral changes are interpreted as analgesia.

The site of action of ROS scavengers is an important question. Our findings are that systemic administration of the spin-trap reagents PBN and DMPO has the greatest analgesic effect, intrathecal administration has almost as much of an effect and intraventricular administration has a significant but much smaller effect. Our interpretation is that the reagents are working primarily at spinal levels but also somewhat at supraspinal levels. Accordingly a maximum effect will be produced by systemic administration of an agent that passes the blood–brain barrier and thus acts at both spinal and supraspinal levels.

A potential caveat is that the ROS scavenger PBN is also known to inhibit gene induction of inducible nitric oxide synthase (iNOS) and activate the transcription factor NFκB (Kotake, 1999). However, PBN's analgesic action has a rapid onset with a peak effect within 1 h of injection, which speaks against the analgesia being due to gene induction. Furthermore, three different ROS scavengers showed similar analgesic effects. Thus, we believe that the analgesia from these compounds in the SNL model is not due to gene transcription but to ROS scavenging.

If the compounds in the present study produce analgesia by scavenging ROS, this implies that excessive ROS is important in the generation of pain in the SNL model. This, then, raises a number of important questions. These include (1) which specific ROS are important for neuropathic pain, (2) what are the sources of the excessive ROS, and (3) by what mechanisms does ROS produce pain?

For the first point, there are a number of types of ROS that can damage neuronal function. Mitochondria normally produce superoxide as a part of an oxidative phosphorylation process and superoxide is readily converted to hydrogen peroxide by superoxide dismutase and then to highly toxic hydroxyl radical. In addition, superoxide and nitric oxide are produced in the cytoplasm by enzymatic reactions, which are activated by increased cytoplasmic calcium. Superoxide and nitric oxide can easily be converted to peroxynitrite, which is also highly toxic. A barrage of primary afferent injury discharges at the time of nerve injury followed by a steady flow of ectopic discharges into the spinal cord in the neuropathic condition may increase mitochondrial respiration as well as intracellular calcium and consequently lead to increased ROS production. It is also possible that excessive ROS builds up in glial cells, which then leaks to produce neuronal damage and dysfunction. In normal conditions, the level of intracellular ROS is tightly regulated by removing excessive ROS through enzymatic reactions, which convert ROS to harmless non-radicals. However, excessive ROS can accumulate intracellularly due to either overproduction or impairment of removal. Sorting out these possibilities is an important task.

Another important question is the mechanism by which excessive ROS produce pain. It is well established that sensitization of dorsal horn cells in the spinal cord (central sensitization) plays a fundamentally important role in neuropathic pain. Therefore, it is conceivable that excessive ROS affects central sensitization. Our working hypothesis is that ROS initiates factors already known to be involved in central sensitization, rather than triggering an independent additional mechanism. For example, many second messengers are involved in sensitization of dorsal horn neurons (Ali and Salter, 2001; Zhang et al., 2003) and it is possible that ROS triggers these second messengers. Another possibility is that ROS activates spinal glial cells, which in turn play an important role in chronic pain (Raghavendra et al., 2003).

One particular ROS, nitric oxide (NO), has been extensively studied in pain mechanisms. For example, inhibitors of nitric oxide synthase (NOS), the enzyme that produces NO, reduce both the NO increase and some of the pain-related behaviors that follow inflammation or neuropathic lesions (Choi et al., 1996; Haley et al., 1992; Hao and Xu, 1996; Meller et al., 1994; Salter et al., 1996; Tedesco et al., 2002; Yoon et al., 1998). The difficulty is that NO has many actions, so its action as a free radical is only one of several possible mechanisms of NO in neuropathic or inflammatory pain. Two studies showed ROS reductions mildly relieved neuropathic pain behaviors. Tal (1996) systemically injected the antioxidant, 4-hydroxy-2,2,6,6-tetramethylpiperidine-1-oxyl (TEMPOL), and found a significant reduction of heat hyperalgesia in CCI rats. TEMPOL acts as a catalyst and mimics the activity of superoxide dismutase (Tal, 1996), thereby removing one type of free radical, the superoxides. In addition, Khalil et al. (1999) showed that systemic tirilazad in CCI rats also significantly alleviated thermal hyperalgesia by improving peripheral microvascular blood flow in the area innervated by the injured nerve. Tirilazad is not a typical anitoxidant but has an anti-lipid peroxidation action and thereby stabilizes membranes (Kavanagh and Kam, 2001), but its action is likely to be limited to the periphery since it has difficulty crossing the blood–brain barrier (Kavanagh and Kam, 2001; Raub et al., 1993). However, these studies, although indicating the possible importance of free radicals in certain pain behaviors, did not (1) test for mechanical allodynia, which is the most debilitating aspect of neuropathic pain, (2) show details of the time course of the effects of antioxidants, (3) determine if tolerance develops, and (4) examine the site of action. Such data are essential for understanding the mechanisms of action of ROS scavengers and for demonstration of their usefulness in reducing clinical persistent pain.

Oxidative stress seems to be an important determinant in degenerative and sometimes painful peripheral nerve conditions (Wagner et al., 1998), and neurodegeneration in the aging brain and in some neurodegenerative diseases are widely regarded as at least a partial consequence of ROS damage associated with increased levels of pro-inflammatory cytokines (Coyle and Puttfarcken, 1993; Götz et al., 1994; In't Veld et al., 1998; Jenner, 1994). Not considered in these studies, however, is the possibility that oxidative stress in the spinal cord is an important underlying mechanism of neuropathic pain. Since the main action of the spin-trap reagents seems to be at the level of the spinal cord (see below), and since peripheral nerve lesions lead to central sensitization of dorsal horn neurons (Ji and Woolf, 2001; Lin et al., 1997; Woolf and Costigan, 1999), a likely possibility is that excess ROS is involved in production of central sensitization. An important point is that spin-trap reagents reverse behavioral signs of neuropathic pain, showing that the pain is a temporary dysfunction, and thus there is presumably no structural damage, such as spinal interneuron death. Our working hypothesis, therefore, is that the chronic pain in this model is the result of dysfunction but not death of dorsal horn neurons stressed by excessive spinal ROS following nerve lesions. This is a conceptual addition to known ROS actions in cell death in that less severe ROS actions can be manifested as neuronal stress followed by dysfunction.

The site of action of the spin-trap reagents was shown in two ways. First, ectopic discharges in damaged primary afferents drive central sensitization, which underlies neuropathic pain behaviors (Chung and Chung, 2002), but systemic PBN (100 mg/kg), which ameliorated the neuropathic pain behaviors, did not reduce the ectopic discharges. Since ectopic discharges in the SNL model originate either from the DRG or the nerve injury site (Chung and Chung, 2002; Liu et al., 1999, 2001), these places are not the targets of ROS scavengers for analgesia. Second, intrathecal injections of PBN or DMPO produce analgesia, which implicates the spinal cord as the major site of action, presumably by reducing central sensitization (Gracely et al., 1992; Willis, 1994; Woolf and Salter, 2000), even though the primary afferent input that drives the sensitization is not reduced. Action of a systemically injected compound in the spinal cord or brain obviously requires passing through the blood–brain barrier. Both compounds (PBN and DMPO) tested in this study have relatively low molecular weights and readily pass through the blood–brain barrier, less so for DMPO than PBN due to its polarized structure. For example, systemically injected PBN can be detected in the brain tissue as early as 15 min after injection (Chen et al., 1990). The lower penetrability of DMPO may explain the less powerful action of DMPO than PBN after systemic injections. However, intrathecally injected DMPO also produces less analgesia than PBN, suggesting PBN as a more effective scavenger of ROS than DMPO, and this difference may determine their effectiveness.

The half-life of intraperitoneally injected PBN in the blood is about 2 h (Trudeau-Lame et al., 2003), but the analgesic effect persists 6–8 h. This delay is probably due to the necessity of ROS to build-up to levels that produce neuropathic pain after the spin-trap agents are no longer present. This implies that this nerve lesion leads to a long-lasting production of toxic levels of ROS. Consistent with this is that preemptive PBN treatment delayed, but did not permanently prevent, the development of neuropathic pain behaviors. Fortunately, repeated small doses of ROS scavengers produce persistent and reproducible analgesia with no development of tolerance. Thus, if these agents are to be useful in therapy, they will presumably have to be given continuously or at relatively short intervals, but the absence of tolerance, at least within the framework of these experiments, indicates that such treatment could be continued over time without difficulty.

In conclusion, systemic administration of ROS scavengers ameliorates the marked decreases in mechanical thresholds that occur in a neuropathic pain model. This analgesia can be extended by repeated injections at short intervals with no development of tolerance. Furthermore, preemptive treatments with a ROS scavenger on neuropathic injury delay the development of pain behaviors. The ROS scavengers do not change the rate of ectopic discharges originating from the DRG or peripheral nerve, and intrathecal injection produces a strong analgesic effect suggesting the major site of action is the spinal cord. We thus conclude that ROS contributes importantly to both the generation and maintenance of neuropathic pain in the SNL model, and suggest that this syndrome, and possibly other neuropathic syndromes, would benefit from continuous administration of non-toxic spin-trap reagents and possibly other ROS scavengers.

Acknowledgements

This work was supported by NIH Grants NS 31680, NS 10161, NS 11255, and AG 13945, and a Research Development Grants from the Sealy Memorial Endowment Fund (2547–03 and 2570–01).

References

Ali DW, Salter MW. NMDA receptor regulation by Src kinase signalling in excitatory synaptic transmission and plasticity. Curr Opin Neurobiol. 2001;11:336-342.

Baik EJ, Chung JM, Chung K. Peripheral norepinephrine exacerbates neuritis-induced hyperalgesia. J Pain. 2003;4:212-221.

Balazs L, Leon M. Evidence of an oxidative challenge in the Alzheimer's brain. Neurochem Res. 1994;19:1131-1137.

Chaplan SR, Bach FW, Pogrel JW, Chung JM, Yaksh TL. Quantitative assessment of tactile allodynia in the rat paw. J Neurosci Methods. 1994;53:55-63.

Chen G, Bray TM, Janzen EG, McCay PB. Excretion, metabolism and tissue distribution of a spin trapping agent, Phenyl-_N_-_tert_-butyl-nitrone (PBN) in rats. Free Radic Res Commun. 1990;9:317-323.

Choi Y, Raja SN, Moore LC, Tobin JR. Neuropathic pain in rats is associated with altered nitric oxide synthase activity in neural tissue. J Neurol Sci. 1996;138:14-20.

Chung JM, Chung K. Importance of hyperexcitability of DRG neurons in neuropathic pain. Pain Pract. 2002;2:87-97.

Coyle JT, Puttfarcken P. Oxidative stress, glutamate, and neurodegenerative disorders. Science. 1993;262:689-695.

Devor M, Zalkind V. Reversible analgesia, atonia, and loss of consciousness on bilateral intracerebral microinjection of pentobarbital. Pain. 2001;94:101-112.

Dixon WJ. Efficient analysis of experimental observations. Annu Rev Pharmacol Toxicol. 1980;20:441-462.

Gerlach M, Ben Shachar D, Riederer P, Youdim MB. Altered brain metabolism of iron as a cause of neurodegenerative diseases? J Neurochem. 1994;63:793-807.

Götz ME, Künig G, Riederer P, Youdim MBH. Oxidative stress: free radical production in neural degeneration. Pharmacol Ther. 1994;63:37-122.

Gracely RH, Lynch SA, Bennett GJ. Painful neuropathy: altered central processing maintained dynamically by peripheral input. Pain. 1992;51:175-194.

Haley JE, Dickenson AH, Schachter M. Electrophysiological evidence for a role of nitric oxide in prolonged chemical nociception in the rat. Neuropharmacology. 1992;31:251-258.

Halliwell B, Gutteridge JM. Free radicals in biology and medicine. Oxford: Clarendon Press; 1989.

Han HC, Lee DH, Chung JM. Characteristics of ectopic discharges in a rat neuropathic pain model. Pain. 2000;84:253-261.

Hao JX, Xu XJ. Treatment of a chronic allodynia-like response in spinally injured rats: effects of systemically administered nitric oxide synthase inhibitors. Pain. 1996;66:313-319.

Hensley K, Pye QN, Tabatabaie T, Stewart CA, Floyd RA., 1997. Reactive oxygen involvement in neurodegenerative pathways. In: Wood PL, editor., Neuroinflammation: mechanisms and management. Humana Press, Totowa, NJ, pp. 265-281.

In't Veld BA, Laune LJ, Hoes AW, Ott A, Hofman A, Breteler MM, Stricker BH. NSAIDs and incident Alzheimer's disease. The Rotterdam Study. Neurobiol Aging. 1998;19:607-611.

Jenner P. Oxidative damage in neurodegenerative disease. Lancet. 1994;344:796-798.

Ji RR, Woolf CJ. Neuronal plasticity and signal transduction in nociceptive neurons: implications for the initiation and maintenance of pathological pain. Neurobiol Dis. 2001;8:1-10.

Kavanagh RJ, Kam PCA. Lazaroids: efficacy and mechanism of action of the 21-aminosteroids in neuroprotection. Br J Anaesth. 2001;86:110-119.

Khalil Z, Liu T, Helme RD. Free radicals contribute to the reduction in peripheral vascular responses and the maintenance of thermal hyperalgesia in rats with chronic constriction injury. Pain. 1999;79:31-37.

Kim SH, Chung JM. An experimental model for peripheral neuropathy produced by segmental spinal nerve ligation in the rat. Pain. 1992;50:355-363.

Kim HK, Park SK, Taglialatela G, Chung K, Coggeshall RE, Chung JM. Effects of spin-trap agents on rat models of neuropathic and arthritic pain. Neurosci Abstr. 2002;454.15.

Kim HK, Zhou J-L, Chung K, Chung JM. The site of action for free radical scavenger-induced analgesia. Neurosci Abstr. 2003;589.12.

Kirk RE., 1995. Experimental design: procedures for the behavioral sciences, 3rd ed. Brooks-Cole, Pacific Grove, CA, p. 37–40.

Kotake Y. Pharmacologic properties of phenyl _N_-_tert_-butylnitrone. Antioxid Redox Signal. 1999;1:481-499.

Lewen A, Matz P, Chan PH. Free radical pathways in CNS injury. J Neurotrauma. 2000;17:871-890.

Lin Q, Peng YB, Wu J, Willis WD. Involvement of cGMP in nociceptive processing by and sensitization of spinothalamic neurons in primates. J Neurosci. 1997;17:3293-3302.

Liu X, Chung K, Chung JM. Ectopic discharges and adrenergic sensitivity of sensory neurons after spinal nerve injury. Brain Res. 1999;849:244-247.

Liu X, Zhou J-L, Chung K, Chung JM. Ion channels associated with the ectopic discharges generated after segmental spinal nerve injury in the rat. Brain Res. 2001;900:119-127.

Meller ST, Cummings CP, Traub RJ, Gebhart GF. The role of nitric oxide in the development and maintenance of the hyperalgesia produced by intraplantar injection of carrageenan in the rat. Neuroscience. 1994;60:367-374.

Olanow CW. An introduction to the free radical hypothesis in Parkinson's disease. Ann Neurol. 1992;32:S2-S8.

Raghavendra V, Tanga F, Rutkowski MD, DeLeo JA. Anti-hyperalgesic and morphine-sparing actions of propentofylline following peripheral nerve injury in rats: mechanistic implications of spinal glia and proinflammaotry cytokines. Pain. 2003;104:655-664.

Raub TJ, Barsuhn CL, Williams LR, Decker DE, Sawada GA, Ho NFH. Application of a biophysical-kinetic model to understand the roles of protein binding and membrane partitioning on passive diffusion of highly lipophilic molecules across cellular barriers. J Drug Target. 1993;1:269-286.

Salter M, Strijbos PJLM, Neale S, Duffy C, Follenfant RL, Garthwaite J. The nitric oxide-cyclic GMP pathway is required for nociceptive signalling at specific loci within the somatosensory pathway. Neuroscience. 1996;73:649-655.

Schapira AH, Cooper JM. Mitochondrial function in neurodegeneration and aging. Mutat Res. 1992;275:1133-1143.

Tal M. A novel antioxidant alleviates heat hyperalgesia in rats with an experimental painful peripheral neuropathy. Neuroreport. 1996;7:1382-1384.

Tedesco LS, Fuseler J, Grisham M, Wolf R, Roerig SC. Therapeutic administration of nitric oxide synthase inhibitors reverses hyperalgesia but not inflammation in a rat model of polyarthritis. Pain. 2002;95:215-223.

Thomson S. Performance of multiple comparisons in repeated measures designs under nonsphericity. SAS Users Group Int Proc. 1990;15:1365-1370.

Tizot DA, Raimbaud E, Lockhart B, Lestage P, Goldstein S. Synthesis, structure, and neuroprotective properties of novel imidazolyl nitrones. J Med Chem. 2000;43:2165-2175.

Trudeau-Lame ME, Kalgutkar AS, LaFontaine M. Pharmacokinetics and metabolism of the reactive oxygen scavenger α-phenyl-_N_-_tert_-butylnitrone in the male Sprague–Dawley rat. Drug Metab Dispos. 2003;31:147-152.

Wagner R, Heckman HM, Myers RR. Wallerian degeneration and hyperalgesia after peripheral nerve injury are glutathione-dependent. Pain. 1998;77:173-179.

Willis WD., 1994. Central plastic responses to pain. In: Gebhart GF, Hammond DL, Jensen TS, editors., Proceedings of the Seventh World Congress on Pain, Progress in Pain Research and Management. IASP Press, Seattle, WA, pp. 301-321.

Woolf CJ, Costigan M. Transcriptional and posttranslational plasticity and the generation of inflammatory pain. Proc Natl Acad Sci USA. 1999;96:7723-7730.

Woolf CJ, Salter MW. Neuronal plasticity: increasing the gain in pain. Science. 2000;288:1765-1768.

Yoon YW, Sung B, Chung JM. Nitric oxide mediates behavioral signs of neuropathic pain in an experimental rat model. Neuroreport. 1998;9:367-372.

Zhang X, Wu J, Willis WD. The effects of protein phosphatase inhibitions on nociceptive behavioral responses of rats following intradermal injection of capsaicin. Pain. 2003;106:443-451.

Keywords:

Anti-allodynia; Free radicals; PBN; Spinal nerve ligation model

© 2004 Lippincott Williams & Wilkins, Inc.