The inhibition of nitric oxide-activated poly(ADP-ribose) synthetase attenuates transsynaptic alteration of spinal cord dorsal horn neurons and neuropathic pain in the rat

Pain 72 (1997) 355–366

The inhibition of nitric oxide-activated poly(ADP-ribose) synthetase attenuates transsynaptic alteration of spinal cord dorsal horn neurons and neuropathic pain in the rat Jianren Mao a ,*, Donald D. Price a, Jeipei Zhu b, Juan Lu a, David J. Mayer a a

Department of Anesthesiology, Medical College of Virginia, Virginia Commonwealth University, Richmond, VA, USA b Department of Anatomy, Medical College of Virginia, Virginia Commonwealth University, Richmond, VA, USA Received 30 January 1997; revised version received 27 March 1997; accepted 16 May 1997

Abstract Transsynaptic alteration of spinal cord dorsal horn neurons characterized by hyperchromatosis of cytoplasm and nucleoplasm (so-called ‘dark’ neurons) occurs in a rat model of neuropathic pain induced by chronic constriction injury (CCI) of the common sciatic nerve. The incidence of dark neurons in CCI rats has been proposed to be mediated by glutamate-induced neurotoxicity. In the present study, we examined whether the inhibition of the nitric oxide (NO)-activated poly(ADP-ribose) synthetase (PARS), a nuclear enzyme critical to glutamate-induced neurotoxicity, would both reduce the incidence of dark neurons and attenuate behavioral manifestations of neuropathic pain in CCI rats. Dark neurons were observed bilaterally (with ipsilateral predominance) within the spinal cord dorsal horn, particularly in laminae I–II, of rats 8 days after unilateral sciatic nerve ligation as compared to sham operated rats. The number of dark neurons in the dorsal horn was dose-dependently reduced in CCI rats receiving once daily intrathecal (i.t.) treatment with the PARS inhibitor benzamide (200 or 400 nmol, but not 100 nmol benzamide or saline) for 7 days. Consistent with the histological improvement, thermal hyperalgesia, mechanical hyperalgesia, and low threshold mechano-allodynia also were reliably reduced in CCI rats treated with either 200 or 400 nmol benzamide. Neither dark neurons nor neuropathic pain behaviors were reliably affected by i.t. administration of either 800 nmol novobiocin (a mono(ADP-ribose) synthetase) or 800 nmol benzoic acid (the backbone structure of benzamide), indicating a selective effect of benzamide. Intrathecal treatment with an NO synthase inhibitor NG-nitro-l-arginine methyl ester (40 nmol, but not its inactive d-isomer) utilizing the same benzamide treatment regimen resulted in similar reductions of both dark neurons and neuropathic pain behaviors in CCI rats. These results provide, for the first time, in vivo evidence indicating that benzamide is neuroprotective and that the PARS-mediated transsynaptic alteration of spinal cord dorsal horn neurons contributes to behavioral manifestations of neuropathic pain in CCI rats. These observations may have general implications beyond treatment of neuropathic pain in that PARS-mediated neuronal alterations may play a significant role in glutamate-mediated neurotoxicity under many other circumstances.  1997 International Association for the Study of Pain. Published by Elsevier Science B.V. Keywords: Nitric oxide; Benzamide; Transsynaptic degeneration; Nerve injury; Neuropathic pain; Poly(ADP-ribose) synthetase

1. Introduction Peripheral nerve injury often results in a pathological pain syndrome known as neuropathic pain (Bonica, 1979; Thomas, 1984). Features of post-injury neuropathic pain include hyperalgesia (exaggerated responses to noxious stimulation), allodynia (nociceptive responses to innocuous * Corresponding author. Department of Anesthesiology, Medical College of Virginia, Box 980516, Richmond, VA 23298–0516, USA. Tel.: +1 804 8285336; fax: +1 804 8284023.

stimulation), spontaneous pain, radiation of pain, and temporal summation of pain (Bonica, 1979; Thomas, 1984; Price et al., 1989). Mechanisms of neuropathic pain have been an area of extensive investigation since the description of a rat model of neuropathic pain induced by chronic constriction injury (CCI) of the common sciatic nerve (Bennett and Xie, 1988). It has been demonstrated that activation of ionotropic glutamate, particularly N-methyl-d-aspartate (NMDA), receptors contributes to the neuropathic pain syndrome, because hyperalgesia and spontaneous pain can be reduced by pre- or post-injury treatment with NMDA recep-

0304-3959/97/$17.00  1997 International Association for the Study of Pain. Published by Elsevier Science B.V. PII S0304- 3959 (97 )0 0063- 8

356

J. Mao et al. / Pain 72 (1997) 355–366

tor antagonists (Davar et al., 1991; Mao et al., 1992a,e, 1993a,b; Yamamoto and Yaksh, 1992; Tal and Bennett, 1993). Activation of NMDA receptors is associated with increases in intracellular Ca2+ concentrations (Collingridge and Singer, 1990; Mayer and Miller, 1990) followed by intracellular changes such as activation of the Ca2+-sensitive protein kinase C (Alkon and Rasmussen, 1988; Nishizuka, 1989) and the production of nitric oxide (NO), a novel neural messenger (Bredt and Snyder, 1992). Such intracellular changes have been shown to modulate synaptic efficacy (Kaczmarek, 1987; Olds et al., 1989; Collingridge and Singer, 1990; Madison et al., 1991; Numann et al., 1991; West et al., 1991) leading to increased neuronal excitability and thereby a central hyperactive state. Support for contributions of PKC- and/or NO-mediated neuronal hyperexcitability to neuropathic pain comes from studies demonstrating increased levels of spinal cord membrane-bound PKC (Mao et al., 1992d, 1993a) and elevated frequencies of spontaneous and stimulus-evoked discharges of spinal cord dorsal horn neurons (Palecek et al., 1992; Laird and Bennett, 1993), including spinothalamic tract neurons (Laird and Bennett, 1993), in this CCI model. Further, behavioral studies have indicated that thermal hyperalgesia can indeed be reduced by intrathecal administration of GM1 ganglioside (an inhibitor of PKC translocation) or NG-nitro-l-arginine methyl ester (an NO synthase inhibitor) in CCI rats (Mao et al., 1992c; Meller et al., 1992). In addition to neuronal hyperexcitability, however, PKC and NO have been shown to induce severe excitotoxic consequences including neuronal death (Vaccarino et al., 1987; Favaron et al., 1988; Moncada et al., 1992; Wallis et al., 1992; Zhang et al., 1994). That such excitotoxic consequences occur in CCI rats is supported by histological evidence indicating CCI-induced transsynaptic alteration of spinal cord dorsal horn neurons (Bennett et al., 1989; Sugimoto et al., 1989, 1990). These altered dorsal horn neurons exhibit pyknosis and hyperchromatosis of both cytoplasm and nucleoplasm resulting in their appearance as ‘dark’ neurons (Bennett et al., 1989; Sugimoto et al., 1989). The incidence of dark neurons within the spinal cord has been proposed to be induced by excitotoxicity resulting from excessive discharges of injured peripheral nerves (Bennett et al., 1989; Sugimoto et al., 1990). Thus, if the incidence of dark neurons is associated with nerve injury-induced neurotoxicity, and if such neurotoxicity contributes to behavioral manifestations of neuropathic pain in CCI rats, one would expect to see reductions in dark neurons and in signs of neuropathic pain in CCI rats by inhibiting NO and/or PKC-mediated intracellular processes leading to neuronal excitotoxicity. Recent in vitro studies have demonstrated that NOmediated activation of poly(ADP-ribose) synthetase (PARS) is a critical intracellular mechanism of neurotoxicity (Cosi et al., 1994; Zhang et al., 1994). PARS has been shown to be activated by NO (Zhang et al., 1994) and its

activation can lead to neuronal deterioration and eventually cell death (Cosi et al., 1994; Zhang et al., 1994). Benzamide is a PARS inhibitor (Raaphorst and Azzam, 1988; Rankin et al., 1989; Cosi et al., 1994; Zhang et al., 1994), which has been shown to protect neurons from glutamate and NOmediated neurotoxicity (Cosi et al., 1994; Zhang et al., 1994). Thus, in the present study we examined whether the inhibition of NO-activated PARS with benzamide would reduce both the incidence of dark neurons and signs of neuropathic pain in CCI rats. Although NO has been implicated in mechanisms of neuropathic pain in this CCI model (Meller et al., 1992), there is little information regarding intracellular processes contributing to the NO effect on the development of neuropathic pain. Since benzamide inhibits NO-activated nuclear enzyme PARS and thus prevents neurons from severe consequences of neuronal excitotoxicity, it would be possible in this study to examine specifically the role of NO-mediated neurotoxicity in mechanisms of neuropathic pain in this CCI model. We report here that the incidence of spinal cord dark neurons is reliably reduced in CCI rats treated with benzamide, and this reduction of dark neurons is associated with attenuation of mechanical hyperalgesia, low threshold mechano-allodynia, and thermal hyperalgesia.

2. Materials and methods 2.1. Experimental animals Adult male Sprague-Dawley rats (Hilltop) weighing 300–350 g were used. Animals were individually housed in cages with water and food pellets available ad libitum. The animal room was artificially illuminated from 7:00 to 19:00 h. Experimental procedures were approved by Institutional Animal Care and Use Committee of the Virginia Commonwealth University. 2.2. Surgical preparation of CCI rats Rats were anesthetized with intraperitoneal 50 mg/kg sodium pentobarbital. Nerve ligation was performed according to the method of Bennett and Xie (1988). Briefly, one side of the rat’s sciatic nerves was exposed, and a 5–7mm long nerve segment was then dissected. Four loose ligatures (4-0 chromic gut) were made around the dissected nerve with a 1.0–1.5-mm interval between each of them. The ligation was carefully manipulated such that the nerve was barely constricted, but no arrest of blood circulation through the epineural vasculature was viewed under a microscope. The skin incision was closed with 4-0 silk sutures. Rats in a sham operation group underwent the same surgical procedure except for sciatic nerve ligation. All animals received one post-operative injection of potassium penicillin (30 000 IU/rat) intramuscularly in order to prevent infection.

J. Mao et al. / Pain 72 (1997) 355–366

2.3. Intrathecal catheter implantation Each rat also was implanted with an intrathecal (i.t.) catheter immediately following nerve ligation or sham operation as described previously (Mao et al., 1992e). A gentamicin sulfate-flushed polyethylene (PE-10) tube was inserted into the subarachnoid space through an incision at the cisterna magna. The caudal end of the catheter was gently threaded to the lumbar enlargement. The rostral end was then secured with dental cement to a screw embedded in the skull. The skin wound was closed with wound clips. Approximately 5–10% of the operated rats exhibited postsurgical motor deficits (e.g., limb paralysis), and these rats were excluded from the experiment. 2.4. Behavioral assessments 2.4.1. Thermal hyperalgesia Thermal hyperalgesia to radiant heat was assessed by using a paw-withdrawal test (Hargreaves et al., 1988). The rat was placed in a plastic cylinder (diameter 18 cm × height 22 cm) on a 3-mm thick glass plate. The radiant heat source was from a projection bulb placed directly under the plantar surface of the rat’s hindpaw. The paw-withdrawal latency was defined as the time elapsed from the onset of radiant heat stimulation to withdrawal of the rat’s hindpaw. The radiant heat source was adjusted to result in baseline latencies of 9–10 s and a cut-off time of 18 s was preset to prevent possible tissue damage. Three test trials separated by a 2-min intertrial interval were made for each of the rat’s hindpaws, and scores from each hindpaw were averaged to yield a mean absolute paw-withdrawal latency. 2.4.2. Mechanical hyperalgesia The pinprick test was used to assess mechanical hyperalgesia in CCI rats as described previously (Tal and Bennett, 1994b). To carry out this test, a rat was placed in a cage. The bottom of the cage is made of a perforated metal sheet with many small square holes. This set-up allowed an experimenter to touch from the bottom of the cage the mid-plantar surface of a rat’s hindpaw with the tip of a safety pin. The mid-plantar area was selected for the test, because this area is convenient to work with and a previous report has demonstrated reliable responses to pinprick or von Frey filament stimulation applied to this plantar area in CCI rats (Tal and Bennett, 1994b). The duration of paw withdrawal (lifting from ground) in response to pinprick was recorded. The duration of paw withdrawal in non-ligated hindpaws was arbitrarily assigned as 1 s because their brief withdrawals made it impractical to record the duration. An average duration recorded from three such pinprick tests (with a 2–3-s interstimulus interval) was used for statistical analysis. 2.4.3. Mechano-allodynia Mechano-allodynia was assessed by using the von Frey filament test according to the method of Tal and Bennett

357

(1994b). Briefly, a rat was placed in the same cage used for the pinprick test, and an experimenter applied von Frey filaments in ascending order of bending force ranging from 0.07 to 76 g (see Tal and Bennett, 1994a for details) to the mid-plantar surface of the rat’s hindpaw. A von Frey filament was applied perpendicular to the skin and depressed slowly until it bent. A threshold force of response is defined as the first filament in the series that evoked at least one clear paw-withdrawal out of five applications. Each of these five stimuli was applied to slightly different areas of the mid-plantar surface with a 2–3-s interstimulus interval. Both ligated and non-ligated hindpaws were tested for comparisons. The von Frey filament test was always carried out prior to the pinprick test to avoid possible confounding effects of pinprick stimulation. All behavioral tests were made with the experimenter unaware of the group assignment. 2.5. Histological examination of dark neurons The histological procedure was the same as that described in previous reports (Bennett et al., 1989; Sugimoto et al., 1989). On day 8 after nerve ligation or sham operation, rats in all groups were anesthetized with pentobarbital within 2 h after the last behavioral tests. Each rat was perfused transcardially with phosphate-buffered saline followed by a fixative containing 1% glutaraldehyde, 1% formaldehyde, and 0.2 mM CaCl2 in 0.12 M phosphate buffer at pH 7.3. Lumbar spinal cords were taken out and kept in the fixative overnight. For each rat, two blocks (about 1.0–1.5 mm each) of lumbar spinal cord at the L4 –L5 level were osmicated in phosphate-buffered 2% OsO4 solution for 2 h and then dehydrated in graded ethanols and embedded in an epoxy resin. From each block, about ten 0.5-mm thick sections were sliced at 50-mm intervals and mounted to gelatincoated slides. These sections were stained with a dye solution containing 1% toluidine blue and 1% sodium borate at about 65°C for approximately 2.5 min. The sections were finally cover-slipped for microscopical examination. Sugimoto et al. (1989) have reported that degenerated spinal cord neurons often exhibit three distinctive features: (i) increased chromophilia throughout both cytoplasm and nucleoplasm (hence the name of ‘dark’ neurons); (ii) homogeneously and intensely stained nucleoplasm with virtually indiscernible heterochromatin; and (iii) irregular cellular outlines. Normal dorsal horn neurons sometimes exhibit enhanced cytoplasmic staining, but they do not show intensified nucleoplasmic staining. Glial cells (particularly oligodendrocytes) also may exhibit chromophilia. Unlike dark neurons, however, oligodendrocytes have overt aggregates of heterochromatin which are seen as clumps under light microscopic examination. As such, only those neurons that exhibit all three features were counted as dark neurons (Fig. 1). For microscopic examination, the spinal cord dorsal horn was divided into three regions, laminae I–II, III–IV, and V–

358

J. Mao et al. / Pain 72 (1997) 355–366

Fig. 1. The photomicrograph shows representative dark neurons (arrow), glial cells (arrow head), and normal neurons (open arrow). Dark neurons exhibit increased chromophilia throughout both cytoplasm and nucleoplasm with virtually indiscernible heterochromatin, and a lack of regular cellular outlines. Normal dorsal horn neurons can be easily distinguished from dark neurons, whereas glial cells are smaller and have overt aggregates of heterochromatin which are seen as intracellular clumps. Bar = 25 mm.

VI, based on the laminar delineation described in previous studies (Molander et al., 1984; Mao et al., 1993a). Each region was subdivided into medial and lateral portions. An examiner who was unaware of the group assignment counted the number of dark neurons in each subdivision under medium power magnification according to the criteria described above. When a potential dark neuron was in doubt, the neuron was then examined under higher power magnification to verify its identity. At least three randomly selected spinal cord sections were examined for each rat. A counting of each spinal cord section produced numbers of dark neurons for each subdivision (e.g., medial portion of laminae I–II). For an individual rat, numbers of dark neurons were averaged from all sampled spinal cord sections in a given subdivision. Each averaged number of dark neurons for a given division was then used for the following statistical analyses. 2.6. Experimental design Five groups of rats (n = 5–6/group) were used to examine the effect of benzamide on the occurrence of dark neurons. They included: (i) CCI rats treated with saline (CCISAL); (ii) CCI rats treated with 100 nmol benzamide (CCIB100); (iii) CCI rats treated with 200 nmol benzamide (CCIB200); (iv) CCI rats treated with 400 nmol benzamide (CCIB400); and (v) sham-operated rats treated with saline (SHAM-SAL). A potential control group of sham-operated rats treated with 400 nmol benzamide was not included, because sham-operated rats exhibited few dark neurons (see Results) and benzamide is unlikely to increase numbers

of dark neurons in sham-operated rats. Two additional control groups (n = 5/group) were included to examine the selectivity of the benzamide effect in blocking PARS, since benzamide may have minimal activity in blocking mono(ADP-ribose) synthetase. One group of CCI rats was given 800 nmol novobiocin, a selective mono(ADP-ribose) synthetase inhibitor (Zhang et al., 1994), and the other was given 800 nmol benzoic acid (the backbone structure of benzamide). Both agents have been shown in an in vitro study to be ineffective in protecting neurons from NMDA neurotoxicity, when the doses were twice as high as the highest effective dose of benzamide (Zhang et al., 1994). Thus, the dose of 800 nmol was selected for each agent as this dose doubles the highest effective dose of benzamide (400 nmol) used in this study. Novobiocin and benzoic acid were purchased from Sigma Company. Two groups (n = 5/group) of CCI rats (CCI-lNAME and CCI-dNAME) were used to examine whether inhibition of NO production itself with 40 nmol NG-nitro-l-arginine methyl ester (l-NAME) or 40 nmol of its inactive d-isomer (d-NAME) would affect the incidence of dark neurons. The selected dose of l-NAME has been shown to reliably attenuate thermal hyperalgesia in CCI rats (Meller et al., 1992). Benzamide was purchased from Sigma Chemical Company (St. Louis, MO) and l-NAME/d-NAME from Research Biochemicals International (Natick, MA). All agents were given i.t. through a microsyringe once per day for 7 days. The first treatment was given at 1 h after either nerve ligation or sham operation. The paw-withdrawal test was made before and 8 days after surgery to examine thermal hyperalgesia. After the completion of behavioral tests on day 8, rats were sacrificed for the histological examination as described above. Data obtained from histological experiments indicated that benzamide (200 or 400 nmol) treatment reduced both the incidence of dark neurons and thermal hyperalgesia in CCI rats. Thus, we further examined whether benzamide would also affect other signs of neuropathic pain in CCI rats, namely mechanical hyperalgesia and mechano-allodynia. Three additional groups of CCI rats were used: one group (n = 10) was treated with saline and the other two groups each with 200 nmol (n = 10) or 400 nmol (n = 5) benzamide. A control group of sham-operated rats (n = 5) also was used to determine the baseline response to the von Frey filament stimulus and noxious mechanical (pinprick) stimulus. Either saline or benzamide was given i.t. using the same regimen as described above, and behavioral tests were made before and on day 8 after surgery. 2.7. Data analysis 2.7.1. Behavioral data Difference scores of paw-withdrawal latencies (contralateral minus ipsilateral) obtained from the thermal hyperalgesia test were analyzed using a two-way analysis of variance (ANOVA) to detect differences across treatment

359

J. Mao et al. / Pain 72 (1997) 355–366 Table 1 Median numbers of dark neurons in the dorsal horn (ranges in parentheses) Ligated side

SHAM-SAL CCI-SAL CCI-B100 CCI-B200 CCI-B400 CCI-L40 CCI-D40 CCI-NOVO CCI-BENZ

Non-ligated side

I–II

III–IV

V–VI

I–II

0.7 (0.3–1.5) 12.5 (8.3–19.5)* 12.9 (11.2–14.5)* 7.7 (2.0–12.0)*,**,† 1.5 (1.0–4.0)**,†,†† 2.3 (2.0–10.7)** 11 (4.0–17.0)* 8.0 (5.0–18.0)* 11.0 (2.0–16.0)*

0.3 2.5 1.7 2.2 1.0 0.5 3.5 3.5 2.0

0 (0–0) 1.5 (0–2.2)* 0.7 (0.5–1.6) 0.2 (0–1.2) 0.2 (0–2.0) 0 (0–0) 2.0 (1.0–3.0)* 1.5 (1.0–2.0)* 1.0 (1.0–2.0)*

0.3 3.0 6.2 5.5 0.5 1.0 2.5 2.5 1.0

(0–1.0) (1.7–6.0)* (1.5–2.2)* (1.3–3.7)* (0–3.0)* (0–1.3) (2.0–6.0)* (2.0–6.0)* (2.0–4.0)*

(0–0.5) (0.5–10.2)*,*** (4.2–8.0)*,*** (1.0–7.5)* (0–2.0) (0.7–7.0) (1.0–10.0)*,*** (1.0–11.0)*,*** (1.0–5.0)*,***

III–IV

V–VI

0 (0–0) 1.3 (0.5–3.5)* 1.5 (1.0–1.7)* 1.2 (0–6.0)* 0.5 (0–2.0) 0 (0–0.7) 1.5 (1.0–5.0)* 2.5 (1.0–4.0)* 1.0 (0–3.0)*

0 (0–0) 0 (0–1.2) 0.5 (0–1.0) 0.2 (0–0.5) 0.2 (0–1.0) 0 (0–0) 1.0 (1.0–2.0) 1.0 (1.0–2.0) 0.5 (0–2.0)

CCI-B100, CCI-B200, CCI-B400, CCI-L40, CCI-D40, CCI-NOVO, CCI-BENZ, nerve-ligated rats treated with 100 nmol benzamide, 200 nmol benzamide, 400 nmol benzamide, 40 nmol l-NAME, 40 nmol d-NAME, 800 nmol novobiocin, 800 nmol benzoic acid, respectively; CCI-SAL, nerve-ligated rats treated with saline; SHAM-SAL, sham-operated rats treated with saline. *P , 0.05, significantly different from the SHAM-SAL group; **P , 0.05, significantly different from the CCI-SAL group; ***P , 0.05, significantly different from the ligated side of the same group; †P , 0.05, significantly different from the CCI-B100 group; ††P , 0.05, significantly different from the CCI-B200 group.

groups. When main effects were detected, a subsequent Waller-Duncan K-ratio t test was performed to determine the source of differences among groups. The pinprick data were also analyzed using a one-way ANOVA to determine differences in paw-withdrawal durations between salineand benzamide-treated CCI rats. The non-parametric Mann-Whitney U-test was used to determine differences among groups in threshold forces that elicited paw-withdrawals in response to the von Frey filament stimulus.

the ipsilateral spinal cord dorsal horn when examined on day 8 after surgery (Table 1; P , 0.05). These dark neurons were distributed topographically within the spinal cord dorsal horn. First, the incidence of dark neurons was most

2.7.2. Histological data The final countings of dark neurons (averaged from all sampled spinal cord sections) in each subdivision (i.e., medial versus lateral portion of a given region) were first analyzed using the Mann-Whitney test. Since the analysis showed no difference in numbers of dark neurons between medial and lateral portions of each sampled region (e.g., laminae I–II), we then pooled the data from medial and lateral portions of a given region. Thus, the total number of dark neurons from a sampled region was subsequently analyzed using the Mann-Whitney test to determine (i) regional differences in numbers of dark neurons among treatment groups and (ii) group differences in numbers of dark neurons among sampled dorsal horn regions.

3. Results 3.1. Effect of benzamide on the incidence of dark neurons The incidence of dark neurons in sham-operated rats (SHAM-SAL) was negligible. Very few dark neurons were observed from all three sampled regions of sham-operated rats (Table 1). In clear contrast to sham operation, CCI induced a reliable increase in dark neurons in all three sampled regions (i.e., laminae I–II, III–IV, and V–VI) of

Fig. 2. The camera lucida drawings show the incidence and distribution of dark neurons in the lumbar spinal cord dorsal horn of a representative sham rat (1) and CCI rats treated with saline (2), 100 nmol (3), 200 nmol (4), 400 nmol benzamide (5), or 800 nmol benzoic acid (6). The left side of the drawings is ipsilateral (ipsi.) to the ligated hindpaw, and the right side is contralateral (contra.) to the ligated hindpaw.

360

J. Mao et al. / Pain 72 (1997) 355–366

In contrast to the effect of benzamide, neither treatment with 800 nmol novobiocin (a selective mono(ADP-ribose) synthetase) inhibitor nor with 800 nmol benzoic acid (the backbone structure of benzamide) changed the incidence of dark neurons in CCI rats (Figs. 2 and 4). There were no reliable differences in numbers of dark neurons between the CCI-SAL group and the CCI-NOVO or CCI-BENZ group (Table 1), indicating a selective effect of benzamide on the incidence of dark neurons. Inhibition of NO production with an NO synthase inhibitor l-NAME, but not d-NAME, showed a similar pattern of reduction of dark neurons in CCI rats (Fig. 4). Numbers of dark neurons were reliably reduced in the ipsilateral laminae I–II of CCI rats receiving i.t. treatment with 40 nmol l-NAME (CCI-lNAME) utilizing the same regimen as that used in the benzamide experiments (Table 1; P , 0.05). Fig. 3. Once daily treatment with benzamide dose-dependently (400>200>100 nmol = saline) reduced numbers of dark neurons in the ipsilateral laminae I–II of CCI rats. a: P , 0.05, as compared to each saline and 100 nmol benzamide group; b: P , 0.05, as compared to the 200 nmol benzamide group. Numbers shown inside each column are the range of dark neurons in the corresponding group.

pronounced in the superficial laminae I–II as compared to laminae III–IV and V–VI of CCI rats (Fig. 2 and Table 1; P , 0.05). Numbers of dark neurons were not statistically different between laminae III–IV and V–VI (Table 1; P . 0.05). Second, unilateral CCI induced a reliable, albeit small, increase in dark neurons on the contralateral laminae I–II and III–IV as compared to corresponding contralateral laminae of sham-operated rats (Table 1; P , 0.05). Third, consistent with the primary innervation of the injured sciatic nerve to the ipsilateral dorsal horn, numbers of dark neurons were significantly higher in the ipsilateral than in the contralateral dorsal horn (Table 1; P , 0.05) indicating an ipsilaterally predominant increase in dark neurons following CCI. Such increases in dark neurons in CCI rats were dosedependently attenuated by i.t. treatment with a PARS inhibitor benzamide (400 . 200 . 100 nmol = saline; Figs. 2 and 3). Once daily i.t. treatment with 200 or 400 nmol (but not 100 nmol) benzamide for 7 days reliably reduced the number of dark neurons in ipsilateral laminae I–II of CCI rats (CCI-B200, CCI-B400) in comparison with the corresponding laminae of saline-treated CCI rats (CCI-SAL; Table 1; P , 0.05). The statistical analysis comparing dark neurons within other sampled regions (ipsilateral laminae III–VI and V–VI and contralateral laminae I–VI) between the CCI-SAL and CCI-B200 or CCI-B400 groups was inconclusive because there were only a small number of dark neurons in these regions of saline-treated rats (Table 1). Of three benzamide doses, 400 nmol resulted in a nearly complete inhibition of dark neurons (Figs. 2 and 3). As such, reliably fewer dark neurons were seen in the 400 nmol benzamide group as compared to either 200 nmol or 100 nmol benzamide group (each P , 0.05; Table 1 and Fig. 3).

3.2. Effect of benzamide on signs of neuropathic pain 3.2.1. Thermal hyperalgesia CCI rats treated with saline for 7 days exhibited thermal hyperalgesia to radiant heat stimulation when examined on day 8 after surgery. Difference scores of paw-withdrawal

Fig. 4. The camera lucida drawings show the incidence and distribution of dark neurons in the lumbar spinal cord dorsal horn of a representative sham rat (1) and CCI rats treated with saline (2), 40 nmol l-NAME (3), 40 nmol d-NAME (4), and 800 nmol novobiocin (5). The left side of the drawings is ipsilateral (ipsi.) to the ligated hindpaw, and the right side is contralateral (contra.) to the ligated hindpaw.

J. Mao et al. / Pain 72 (1997) 355–366

Fig. 5. Once daily i.t. treatment with 200 or 400 nmol (not 100 nmol) benzamide for 7 days attenuated thermal hyperalgesia as indicated by a reliable reduction of difference scores of paw-withdrawal latencies (contralateral minus ipsilateral). There were no statistical differences between groups of rats treated with 200 or 400 nmol benzamide. A similar result was seen in CCI rats treated with 40 nmol l-NAME. a: P , 0.05, as compared to the baseline difference score of the same group. b: P , 0.05, as compared to the CCI-SAL group.

latencies (contralateral minus ipsilateral) examined on day 8 were reliably higher than the baseline (before surgery) difference score in these rats (Fig. 5; P , 0.05). Paw-withdrawal latencies on the contralateral hindpaw of the same CCI rats were not statistically different before and 8 days after surgery (baseline: 8.74 ± 0.12 s; day 8: 8.59 ± 0.25 s; P . 0.05), nor were they different in both hindpaws of sham-operated rats (P . 0.05). The development of thermal hyperalgesia was significantly reduced in CCI rats treated with either 200 or 400 nmol (not 100 nmol) benzamide for 7 days (Fig. 5; P , 0.05) as compared to saline-treated CCI rats. Similarly, paw-withdrawal latencies on the contralateral hindpaw of the same CCI rats were not statistically different before and after benzamide treatment (P . 0.05). There was a trend of more reduction of thermal hyperalgesia in the 400-nmol group than in the 200-nmol group (Fig. 5), but the difference did not reach statistical significance partially due to a small sample size. Similarly, the 7-day treatment regimen with 40 nmol l-NAME (but not d-NAME) also reliably attenuated the development of thermal hyperalgesia in CCI rats (Fig. 5; P , 0.05). Neither novobiocin nor benzoic acid affected the degree of thermal hyperalgesia, mechanical hyperalgesia and mechanical allodynia (see below) in CCI rats (each P . 0.05). 3.3. Mechanical hyperalgesia While noxious mechanical stimulation applied to the plantar surface of a rat’s hindpaw before nerve ligation elicited a very brief (,1 s) lifting of the stimulated hindpaw,

361

the same stimulus applied to the ipsilateral hindpaw of CCI rats (CCI-SAL group) on day 8 resulted in a significantly longer withdrawal of the stimulated hindpaw (baseline: 1 s; day 8: 5 ± 0.2 s; Fig. 6; P , 0.05). This prolonged withdrawal of the ipsilateral hindpaw of CCI rats was often accompanied by licking, shaking, and/or scratching of the hindpaw, indicating an exacerbated response to the stimulus. The abnormally prolonged response to this noxious mechanical stimulus was not present in the contralateral hindpaw of CCI rats before and on day 8 after nerve ligation, nor was it present in sham-operated rats (SHAM-SAL group) before and on day 8 after sham operation. Benzamide (200 or 400 nmol, but not 100 nmol) treatment reliably attenuated the development of this hyperalgesic response to noxious mechanical stimulation in CCI rats. Thus, the duration of response in benzamide-treated CCI rats (200 or 400 nmol) was reliably shorter on day 8 as compared to that in saline-treated CCI rats (Fig. 6; P , 0.05) and was not statistically different from their baseline response durations (Fig. 6; P . 0.05). 3.3.1. Mechano-allodynia The lowest baseline threshold force of the von Frey filament stimulus that elicited at least one clear hindpaw withdrawal out of five applications was 6 g. About 50% of the tested rats responded to the stimulus force of 15 g (Fig. 7). While threshold forces were not reliably different before and on day 8 after surgery in both hindpaws of sham-operated rats (SHAM-SAL group; P . 0.05) or the contralateral hindpaw of CCI rats (CCI-SAL group; P . 0.05), threshold forces were reliably lowered in the ipsilateral hindpaw of the same CCI rats in comparison with the threshold forces

Fig. 6. Once daily i.t. treatment with 200 or 400 nmol benzamide for 7 days attenuated mechanical hyperalgesia as indicated by a reliably reduced duration of paw-withdrawal in response to the pinprick stimulus. As such, the duration of paw-withdrawal was not statistically different between baseline and day 8 after nerve ligation in benzamide-treated CCI rats. *P , 0.05, as compared to the baseline score of the corresponding group.

362

J. Mao et al. / Pain 72 (1997) 355–366

Fig. 7. Once daily i.t. treatment with 200 nmol benzamide for 7 days attenuated, while 400 nmol benzamide completely blocked, the development of low threshold mechano-allodynia. The cumulative response curve was shifted towards the right (closer to the curve of the contralateral hindpaw). There was no reliable difference in threshold forces of the contralateral hindpaw between CCI rats treated with saline or benzamide. a: P , 0.05, as compared to the contralateral (non-ligated) hindpaw of the same rats. b: P , 0.05, as compared to the ipsilateral (ligated) hindpaw of rats in the CCI-SAL group. c: P , 0.05, as compared to the ipsilateral (ligated) hindpaw of rats in the CCI-B200 group.

for the contralateral hindpaw (Fig. 7; P , 0.05). As such, the cumulative response curve was shifted toward the left, and the von Frey filament stimulus with a much lower force (e.g., 0.07 g) elicited clear hindpaw withdrawal in 20% of the tested CCI rats in this group (Fig. 7). Such a low threshold mechanical stimulus elicited only very weak tactile sensations when applied to the skin of the experimenter. As shown in Fig. 7, when the force was increased to the lowest baseline threshold force (i.e., 6 g), 80% of rats in this group responded as compared to only 20% responding in the same group before surgery indicating the presence of a low threshold mechano-allodynia in CCI rats. Benzamide treatments increased the threshold forces of the von Frey filament stimulus in the ipsilateral hindpaw of CCI rats as compared to the threshold forces of the ipsilateral hindpaw of saline-treated CCI rats (Fig. 7; P , 0.05). Consequently, the cumulative response curves of CCI-B200 and CCI-B400 groups were shifted back to the right (toward the curve of the contralateral hindpaw) (Fig. 7; P , 0.05), suggesting a specific reduction of low threshold (likely to be Ab-mediated) mechano-allodynia. The reduction of mechanical allodynia was reliably higher in the CCI-B400 group as compared to the CCI-B200 group (P , 0.05). This difference between the CCI-B200 and CCI-B400 group was that the lower dose of benzamide produced an incomplete reduction of allodynic responses to von Frey fila¨ ment stimulation, whereas the higher dose completely abolished such responses in CCI rats (Fig. 7; P , 0.05).

4. Discussion The major findings of the present study are: (i) the inci-

dence of dark neurons within the spinal cord dorsal horn of CCI rats was reduced dose-dependently by a PARS inhibitor benzamide; (ii) direct inhibition of NO production by an NO synthase inhibitor l-NAME also decreased the number of dark neurons in CCI rats; and (iii) in association with the reduced incidence of dark neurons, thermal hyperalgesia, mechanical hyperalgesia, and low threshold mechano-allodynia were clearly attenuated by benzamide treatment. These results provide, for the first time, in vivo evidence indicating that activation of the NO-activated nuclear enzyme PARS contributes to transsynaptic alterations of dorsal horn neurons following peripheral nerve injury in CCI rats. These findings support the notion that NO-associated neurotoxicity may be critical to neural mechanisms of neuropathic pain and strongly indicate that PARS inhibitors are neuroprotective from such neurotoxicity. In the present experiments, established criteria (Bennett et al., 1989; Sugimoto et al., 1989) were used to identify dark neurons and to differentiate dark neurons from normal neurons and glial cells with increased chromophilia. A recent electron microscopic study confirmed that CCIinduced dark neurons that were identified under a light microscope using the same established criteria are clearly degenerative neurons (neither glial cells nor normal neurons) with intracellular microstructural changes and atypical synaptic appearances (Hama et al., 1994). Thus, it is practical to distinguish dark neurons from either normal neurons or glial cells under light microscopic examination using the guidelines of previously established criteria. Conceivably, actual numbers of dark neurons could have been higher than those reported in Table 1, because any questionable dark neurons were not counted in this study. Additionally, as has been confirmed in previous studies (Bennett et al., 1989;

J. Mao et al. / Pain 72 (1997) 355–366

Sugimoto et al., 1989), the occurrence of dark neurons is not likely due to histological artifacts because an ipsilateral predominance of dark neurons was clearly shown, a feature consistent with the primarily ipsilateral innervation of the injured sciatic nerve to the spinal cord. One caveat regarding dark neurons is that the occurrence of dark neurons does not provide direct information about actual viability of the neurons in question, an issue that can be readily determined in cell cultures. Consequently, it is not yet clear whether dark neurons represent a population of truly dead neurons (permanent loss of cell viability) or neurons simply undergoing morphological changes (temporary changes in cell viability) after peripheral nerve injury. Nonetheless, both histologic changes revealed in a previous electron microscope study (Hama et al., 1994) and the relationship between dark neurons and neuropathic pain behaviors reported in the present study are in favor of the view that these dark neurons lack functional integrity (Bennett et al., 1989; Sugimoto et al., 1989). The distribution pattern of dark neurons within the dorsal horn is characterized by two important features. One is that the incidence of dark neurons is bilateral but with a clear ipsilateral predominance within the dorsal horn. Similar bilateral but ipsilateral predominant changes have been shown in the spinal cord of CCI rats in a number of studies, including the autoradiographic assay of 14C-2-deoxyglucose metabolism (Price et al., 1991; Mao et al., 1992b), [3H]phorbol ester (PKC) binding (Mao et al., 1992d, 1993a), opioid receptor binding (Stevens et al., 1991), immunocytochemical examination of PKC (Mao et al., 1995b) and c-fos (Kajander et al., 1990), as well as radioimmunoassay of calcitonin-gene-related peptide and substance P levels (Kajander and Xu, 1995). A general finding of previous studies (Mao et al., 1992b, 1993a) as well as the present study, was the association of functional-anatomical changes in the ipsilateral dorsal horn of CCI rats and nociceptive behaviors elicited from the ipsilateral hindpaw of CCI rats. Less clear is the association between nociceptive behaviors on the contralateral hindpaw and changes in dark neurons of the contralateral spinal cord of CCI rats. This lack of contralateral association may be due to a weak excitotoxic impact from sciatic nerve injury on the contralateral spinal cord and/or the inadequate sensitivity of behavioral tests utilized to detect subtle changes. The second feature is that the greatest density of dark neurons was seen in the superficial laminae of the dorsal horn. This topographic distribution of dark neurons indicates that neurons in laminae I–II are a main target of CCI-induced excitotoxicity. It also suggests that the loss of function of these neurons may be particularly critical to spinal cord mechanisms of neuropathic pain in CCI rats. This is consistent with the known importance of laminae I–II in spinal cord nociceptive transmission, a region containing nociceptive projection neurons as well as excitatory and inhibitory interneurons (Bennett et al., 1979, 1980; Price et al., 1979; Gobel et al., 1980) and actively partici-

363

pating in the balance of local excitatory and inhibitory activities and modulation of nociceptive information transmitted from the spinal cord to supraspinal structures (Willis, 1985; Price, 1988). Previous studies have suggested that excessive discharges from injured sciatic nerve may be a driving force for the initiation and progress of central changes in CCI rats (Bennett et al., 1989; Sugimoto et al., 1989). In this regard, peripheral nerve injury has indeed been shown to evoke injury discharges that contribute to the development of neuropathic pain in rats (Seltzer et al., 1991). In this CCI model, aberrant nerve discharges arising from ectopic action potential generating sites along the injured sciatic nerve occur as early as 1–3 days after nerve ligation (Kajander et al., 1992). Such abnormal nerve discharges may cause excessive release of glutamate/aspartate, activation of the NMDA and/or other subtypes of glutamate receptors, and subsequent activation of intracellular PKC and NO within the spinal cord (Davar et al., 1991; Mao et al., 1992c,e; Meller et al., 1992). The impact of such peripheral changes on the spinal cord appears to be focused on the superficial but not deeper laminae as shown in both previous (Kajander et al., 1990; Mao et al., 1992d, 1995b; Kajander and Xu, 1995) and present studies. On the one hand, the enhanced neural transmission as a result of direct glutamate-mediated central excitation may lead to sensitization of dorsal horn neurons, a mechanism known to contribute to neuropathic pain in a number of studies (Palecek et al., 1992; Laird and Bennett, 1993; Mao et al., 1995a). On the other hand, excessive production of NO, a synaptic messenger under non-pathological conditions, may trigger intracellular cascades leading to excitotoxicity. Such excitotoxicity may manifest as the PARS activation and resultant degenerative neuronal alterations. Our present study thus provides, for the first time, in vivo evidence that NO-mediated neuronal alterations are related to the development of neuropathic pain. The focus of this evidence is that the inhibition of the NO-activated nuclear enzyme PARS, shown to be crucial to glutamate-mediated neurotoxicity in in vitro studies (Cosi et al., 1994; Zhang et al., 1994), effectively attenuates both the incidence of dark neurons and neuropathic pain behaviors in CCI rats. Nitric oxide has been proposed to have both pre- and post-synaptic actions (Bredt and Snyder, 1992). This study provides new evidence suggesting a postsynaptic action of NO, which appears to take place within the same neuron that produces NO. The reduction of dark neurons by benzamide is unlikely to be due to non-specific effects for several reasons. First, there is a reliable reduction of dark neurons with benzamide (400 nmol . 200 nmol . 100 nmol = saline), indicating a dose-dependent pattern of benzamide actions. Second, dark neurons were not affected in relevant control groups (novobiocin, benzoic acid), indicating a selective effect of benzamide on poly(ADP-ribose) synthetase. Third, a similar reduction of dark neurons also occurs in the spinal cord of CCI rats when an NO synthase inhibitor l-NAME but not d-

364

J. Mao et al. / Pain 72 (1997) 355–366

NAME was used. This is consistent with the observation that PARS is activated by NO (Zhang et al., 1994), which in turn leads to excitotoxicity and neuronal death in cell cultures (Cosi et al., 1994; Zhang et al., 1994). Importantly, there is an association between benzamide treatment and the reduction in mechano-allodynia (i.e., a reliably greater reduction seen in 400 nmol than in 200 nmol benzamide treatment group). Thus, 200 nmol benzamide produced an incomplete but statistically reliable reduction of allodynic responses to von Frey filament stimulation, whereas 400 nmol benzamide completely abolished such responses in CCI rats. On the other hand, although 200 or 400 nmol benzamide significantly reduced mechanical and thermal hyperalgesia as compared to saline and 100 nmol benzamide groups, 400 nmol benzamide failed to produce a greater reduction in hyperalgesia than 200 nmol benzamide. The lack of a clear dose-related reduction in hyperalgesia tests within the dose range used in this study could be due to small numbers of rats in each group, a ceiling effect of 200 nmol benzamide on hyperalgesic behaviors, or both. The exact relationship between the incidence of dark neurons and the development of various signs of neuropathic pain in CCI rats is yet to be determined. The technical limitation for such investigations is that one cannot easily differentiate functional types of dark neurons (e.g., laminae I–II projection neurons versus interneurons and excitatory versus inhibitory interneurons). Given a vast majority of dark neurons in laminae I–II, one possibility is that most dark neurons are inhibitory interneurons and that the high incidence of dark neurons in CCI rats may reflect a persistent imbalance of the excitatory-inhibitory circuitry within the spinal cord dorsal horn, i.e., a loss of inhibition. Important contributions from a loss of spinal cord inhibition to the development of neuropathic pain are indicated by exacerbated thermal hyperalgesia in CCI rats when spinal cord gaminobutyric acid (GABA) receptors are blocked with their antagonists (Yamamoto and Yaksh, 1993). In patients with neuropathic pain, transcutaneous electrical stimulation of the lowest threshold axons supplying pathological zones of the skin or innocuous mechanical stimulation applied by gently moving a cotton swab across the skin, both of which presumably induce Ab-mediated inhibition of spinal cord nociceptive transmission in normal subjects, produced painful sensation (Price et al., 1992). Thus, under neuropathic conditions, low threshold mechanical stimulation not only fails to produce inhibition but somehow has gained access to spinal cord nociceptive mechanisms (hence allodynia). These changes may result, as least in part, from the loss of inhibitory interneurons secondary to peripheral nerve injury. Conceivably, the loss of inhibition would also contribute to exaggerated responses to noxious stimulation (hyperalgesia) in CCI rats. These explanations are in accordance with the reduction by benzamide of both dark neurons and signs of neuropathic pain in CCI rats. It is of interest to note that, unlike a PARS inhibitor that attenuates both hyperalgesia and allodynia, an NMDA receptor antagonist

inhibits thermal hyperalgesia with little effect on mechanoallodynia (Tal and Bennett, 1994a). This subtle difference between the two classes of agents may indicate a diversity of mechanisms underlying thermal hyperalgesia and allodynia. More explicitly, our data suggest that the histopathological alterations in the dorsal horn (dark neurons) secondary to NO-mediated PARS activation would be particularly important for the development of mechano-allodynia in CCI rats. The presence of a strong dose-response relationship between the reduction of dark neurons and mechanoallodynia in our study appears to support this view. In summary, the present study provides novel evidence for the involvement of the NO-activated nuclear enzyme PARS in transsynaptic alterations of spinal cord dorsal horn neurons in CCI rats. Thus, it appears that CCI-induced spinal cord hyperexcitability may result from at least two general central changes: (i) increases in synaptic efficacy between primary and second order nociceptive dorsal horn neurons, leading to enhanced spontaneous and stimulusinduced neuronal activity; and (ii) neurotoxic reduction of inhibitory mechanisms, possibly reflected by the presence of dark neurons. Clearly, the relationship between the development of dark neurons (potentially irreversible morphological changes) and neuropathic pain as demonstrated in the present study emphasizes the significance of early intervention in the successful management of neuropathic pain patients. Importantly, our observations may be more general than just neuropathic pain. These observations may relate to any phenomenon involving intracellular cascades beginning with, but not restricted to, NMDA receptor activation with resultant excessive production of NO. Thus, benzamide and related agents may be proven to be neuroprotective from a variety of in vivo glutamate-associated neurotoxic conditions.

Acknowledgements We wish to thank Dr. Gary J. Bennett for his valuable comments on this study. We also wish to thank Mr. Les Keniston and Ms. Rosa Dent for their technical assistance. This work was supported by PHS grants DA08835 and NS24009. References Alkon, D.L. and Rasmussen, H., A spatial-temporal model of cell activation, Science, 239 (1988) 998–1005. Bennett, G.J. and Xie, Y.K., A peripheral mononeuropathy in rat that produces disorders of pain sensation like those seen in man, Pain, 33 (1988) 87–107. Bennett, G.J., Hayashi, H., Abdelmoumene, M. and Dubner, R., Physiological properties of staked cells of substantial gelatinosa intracellularly stained with horseradish peroxidase, Brain Res., 164 (1979) 285– 289. Bennett, G.J., Abdelmoumene, M., Hayashi, H. and Dubner, R., Physiology and morphology of substantial gelatinosa neurons intracellularly

J. Mao et al. / Pain 72 (1997) 355–366 stained with horseradish peroxidase, J. Comp. Neurol., 194 (1980) 809– 827. Bennett, G.J., Kajander, K.C., Sahara, Y., Iadarola, M.J. and Sugimoto, T., Neurochemical and anatomical changes in the dorsal horn of rats with an experimental painful peripheral neuropathy. In: F. Cervero, G.J. Bennett and P.M. Headley (Eds.), Proceedings of Sensory Information in the Superficial Dorsal Horn of the Spinal Cord, Plenum Press, New York, 1989, pp. 463–471. Bonica, J.J., Causalgia and other reflex sympathetic dystrophies. In: J.J. Bonica, J.C. Liebeskind and D.G. Albe-Fessard (Eds.), Advances in Pain Research and Therapy, 3rd Edn, Raven Press, New York, 1979, pp. 141–166. Bredt, D. and Snyder, S.H., Nitric oxide, a novel neuronal messenger, Neuron, 8 (1992) 3–11. Collingridge, G.L. and Singer, W., Excitatory amino acid receptors and synaptic plasticity, Trends Pharmacol. Sci., 11 (1990) 290–296. Cosi, C., Suzuki, H., Milani, D., Facci, L., Menegazzi, M., Vantini, G., Kanai, Y. and Skaper, S.D., Poly(ADP-ribose) polymerase: early involvement in glutamate-induced neurotoxicity in cultured cerebellar granule cells, J. Neurosci. Res., 39 (1994) 38–46. Davar, G., Hama, A., Deykin, A., Vos, B. and Maciewicz, R., MK-801 blocks the development of thermal hyperalgesia in a rat model of experimental painful neuropathy, Brain Res., 553 (1991) 327–330. Favaron, M., Manev, H., Alho, H., Bertolino, M., Ferret, B., Guidotti, A. and Costa, E., Gangliosides prevent glutamate neurotoxicity in neuronal cultures, Proc. Natl. Acad. Sci. USA, 85 (1988) 7351–7355. Gobel, S., Falls, W.M., Bennett, G.J., Abdelmoumene, M., Hayashi, H. and Humphery, E., An EM analysis of the synaptic connections of horseradish peroxidase-filled stalked cells and islet cells in the substantial gelatinosa of the adult cat spinal cord, J. Comp. Neurol., 194 (1980) 781–807. Hama, A.T., Sagen, J. and Pappas, G.D., Spinal neurons in rats with unilateral constriction nerve injury: a preliminary study, Neurol. Res., 16 (1994) 297–304. Hargreaves, K., Dubner, R., Brown, F., Flores, C. and Jones, J., A new and sensitive method for measuring thermal nociception in cutaneous hyperalgesia, Pain, 32 (1988) 77–88. Kaczmarek, L.K., The role of protein kinase C in the regulation of ion channels and neurotransmitter release, Trends Neurosci., 10 (1987) 30– 34. Kajander, K.C., Wakisaka, S., Driasci, G. and Iadarola, M.J., Labeling of Fos protein increases in an experimental model of peripheral neuropathy in the rat, Soc. Neurosci. Abs., 16 (1990) 1281. Kajander, K.C., Wakisaka, S. and Bennett, G.J., Spontaneous discharge originates in the dorsal root ganglion at the onset of a painful peripheral neuropathy in the rat, Neurosci. Lett., 138 (1992) 225–228. Kajander, K.C. and Xu, J.X., Quantitative evaluation of calcitonin generelated peptide and substance P levels in rat spinal cord following peripheral nerve injury, Neurosci. Lett., 186 (1995) 84–88. Laird, J.M.A. and Bennett, G.J., An electrophysiological study of dorsal horn neurons in the spinal cord of rats with an experimental peripheral neuropathy, J. Neurophysiol., 69 (1993) 2071–2085. Madison, D.V., Malenka, R.C. and Nicoll, R.A., Mechanisms underlying long-term potentiation of synaptic transmission, Annu. Rev. Neurosci., 14 (1991) 379–397. Mao, J., Mayer, D.J., Hayes, R.L., Lu, J. and Price, D.D., Differential roles of NMDA and non-NMDA receptor activation in induction and maintenance of thermal hyperalgesia in rats with painful peripheral mononeuropathy, Brain Res., 598 (1992a) 271–278. Mao, J., Price, D.D., Coghill, R.C., Mayer, D.J. and Hayes, R.L., Spatial patterns of spinal cord 14C-2-deoxyglucose metabolic activity in a rat model of painful peripheral mononeuropathy, Pain, 50 (1992b) 89–100. Mao, J., Price, D.D., Hayes, R.L., Lu, J. and Mayer, D.J., Intrathecal GM1 ganglioside and local nerve anesthesia reduce nociceptive behaviors in rats with experimental peripheral mononeuropathy, Brain Res., 584 (1992c) 28–35. Mao, J., Price, D.D., Mayer, D.J. and Hayes, R.L., Pain-related increases in

365

spinal cord membrane-bound protein kinase C following peripheral nerve injury, Brain Res., 588 (1992d) 144–149. Mao, J., Price, D.D., Mayer, D.J., Lu, J. and Hayes, R.L., Intrathecal MK801 and local nerve anesthesia synergistically reduce nociceptive behaviors in rats with experimental peripheral mononeuropathy, Brain Res., 576 (1992e) 254–262. Mao, J., Mayer, D.J., Hayes, R.L. and Price, D.D., Spatial patterns of increased spinal cord membrane-bound protein kinase C and their relation to increases in 14C-2-deoxyglucose metabolic activity in rats with painful peripheral mononeuropathy, J. Neurophysiol., 70 (1993a) 470– 481. Mao, J., Price, D.D., Hayes, R.L., Lu, J., Mayer, D.J. and Frenk, H., Intrathecal treatment with dextrorphan or ketamine potently reduces pain-related behaviors in a rat model of peripheral mononeuropathy, Brain Res., 605 (1993b) 164–168. Mao, J., Price, D.D. and Mayer, D.J., Mechanisms of hyperalgesia and opiate tolerance: a current view of their possible interactions, Pain, 62 (1995a) 259–274. Mao, J., Price, D.D., Phillips, L.L., Lu, J. and Mayer, D.J., Increases in protein kinase C gamma immunoreactivity in the spinal cord dorsal horn of rats with painful peripheral mononeuropathy, Neurosci. Lett., 198 (1995b) 75–78. Mayer, M.L. and Miller, R.J., Excitatory amino acid receptors, second messengers and regulation of intracellular Ca2+ in mammalian neurons, Trends Pharmacol. Sci., 11 (1990) 254–260. Meller, S.T., Pechman, P.S., Gebhart, G.F. and Maves, T.J., Nitric oxide mediates the thermal hyperalgesia produced in a rat model of neuropathic pain in the rat, Neuroscience, 50 (1992) 7–10. Molander, C., Xu, Q. and Grant, G., The cytoarchitectonic organization of the spinal cord in the rat. I. The lower thoracic and lumbosacral cord, J. Comp. Neurol., 230 (1984) 133–141. Moncada, C., Lekieffre, D., Arvin, B. and Meldrum, B., Effect of NO synthase inhibition on NMDA- and ischaemia- induced hippocampal lesions, NeuroReport, 3 (1992) 530–532. Nishizuka, Y., The family of protein kinase C for signal transduction, J. Am. Med. Assoc., 262 (1989) 1826–1833. Numann, R., Catterall, W.A. and Scheuer, T., Functional modulation of brain sodium channels by protein kinase C phosphorylation, Science, 254 (1991) 115–118. Olds, J.L., Anderson, M.L., McPhie, D.L., Ataten, L.D. and Alkon, D.L., Imaging of memory-specific changes in the distribution of protein kinase C in the hippocampus, Science, 245 (1989) 866–869. Palecek, J., Paleckova, V., Dougherty, P.M., Carlton, S.M. and Willis, W.D., Responses of spinothalamic tract cells to mechanical and thermal stimulation of skin in rats with experimental peripheral neuropathy, J. Neurophysiol., 67 (1992) 1562–1573. Price, D.D., Hayashi, H., Dubner, R. and Ruda, M.A., Functional relationships between neurons of marginal and substantia gelatinosa layers of primate dorsal horn, J. Neurophysiol., 42 (1979) 1590–1608. Price, D.D., Psychological and Neural Mechanisms of Pain, Raven Press, New York, 1988. Price, D.D., Bennett, G.J. and Rafii, A., Psychophysical observations on patients with neuropathic pain relieved by a sympathetic block, Pain, 36 (1989) 273–288. Price, D.D., Mao, J., Coghill, R.C., d’Avella, D., Cicciarello, R., Fiori, M., Mayer, D.J. and Hayes, R.L., Regional changes in spinal cord glucose metabolism in a rat model of painful neuropathy, Brain Res., 564 (1991) 314–318. Price, D.D., Long, S. and Huitt, C., Sensory testing of pathophysiological mechanisms of pain in patients with reflex sympathetic dystrophy, Pain, 49 (1992) 163–174. Raaphorst, G.P. and Azzam, E.I., Poly(ADP-ribose) synthetase inhibitors increase radiation and thermal sensitivity but do not affect thermotolerance, Radiat. Res., 116 (1988) 442–452. Rankin, P.W., Jacobson, E.L., Benjamin, R.C., Moss, J. and Jacobson, M.K., Quantitative studies of inhibitors of ADP-ribosylation in vitro and in vivo, J. Biol. Chem., 264 (1989) 4312–4317.

366

J. Mao et al. / Pain 72 (1997) 355–366

Seltzer, Z., Cohn, S., Ginzgurg, R. and Berlin, B., Modulation of neuropathic pain behavior in rats by spinal disinhibition and NMDA receptor blockade of injury discharge, Pain, 45 (1991) 69–75. Stevens, C.W., Kajander, K.C., Bennett, G.J. and Seybold, V.S., Bilateral and differential changes in spinal cord mu, delta and kappa opioid binding in rats with a painful, unilateral neuropathy, Pain, 46 (1991) 315–326. Sugimoto, T., Bennett, G.J. and Kajander, K.C., Strychnine-enhanced transsynaptic degeneration of dorsal horn neurons in rats with an experimental painful peripheral neuropathy, Neurosci. Lett., 98 (1989) 139– 143. Sugimoto, T., Bennett, G.J. and Kajander, K.C., Trans-synaptic degeneration in the superficial dorsal horn after sciatic nerve injury: effects of a chronic constriction injury, transection, and strychnine, Pain, 42 (1990) 205–213. Tal, M. and Bennett, G.J., Dextrorphan relieves neuropathic heat-evoked hyperalgesia in the rat, Neurosci. Lett., 151 (1993) 107–110. Tal, M. and Bennett, G.J., Neuropathic pain sensations are differentially sensitive to dextrorphan, NeuroReport, 5 (1994a) 1438–1440. Tal, M. and Bennett, G.J., Extra-territorial pain in rats with a peripheral mononeuropathy: mechano-hyperalgesia and mechano-allodynia in the territory of an uninjured nerve, Pain, 57 (1994b) 275–282. Thomas, P.K., Clinical features and differential diagnosis of peripheral

neuropathy. In: P.J. Dyck, P.K. Thomas, E.H. Lambert and R. Bunge (Eds.), Peripheral Neuropathy, 2nd Edn., W.B. Saunders Company, Philadelphia, 1984, pp. 1169–1190. Vaccarino, F., Guidotti, A. and Costa, E., Ganglioside inhibition of glutamate-mediated protein kinase C translocation in primary cultures of cerebellar neurons, Proc. Natl. Acad. Sci. USA, 34 (1987) 8707. Wallis, R.A., Panizzon, K. and Wasterlain, C.G., Inhibition of nitric oxide synthase protects against hypoxic neuronal injury, NeuroReport, 3 (1992) 645–648. West, J.W., Numann, R., Murphy, B.J., Scheuer, T. and Catterall, W.A., A phosphorylation site in the Na+ channel required for modulation by protein kinase C, Science, 254 (1991) 866–868. Willis, W.D., The Pain System, Karger, New York, 1985. Yamamoto, T. and Yaksh, T.L., Spinal pharmacology of thermal hyperalgesia induced by constriction injury of sciatic nerve. Excitatory amino acid antagonists, Pain, 49 (1992) 121–128. Yamamoto, T. and Yaksh, T.L., Effects of intrathecal strychnine and bicuculline on nerve compression-induced thermal hyperalgesia and selective antagonism by MK-801, Pain, 54 (1993) 79–84. Zhang, J., Dawson, V.L., Dawson, T.M. and Snyder, S.H., Nitric oxide activation of poly(ADP-ribose) synthetase in neurotoxicity, Science, 263 (1994) 687–689.