Nerve injury-induced changes in opioid modulation of wide dynamic range dorsal column nuclei neurones

PII: S 0 3 0 6 - 4 5 2 2 ( 0 1 ) 0 0 6 1 7 - 0

Neuroscience Vol. 111, No. 1, pp. 215^228, 2002 ß 2002 IBRO. Published by Elsevier Science Ltd All rights reserved. Printed in Great Britain 0306-4522 / 02 $22.00+0.00

www.neuroscience-ibro.com

NERVE INJURY-INDUCED CHANGES IN OPIOID MODULATION OF WIDE DYNAMIC RANGE DORSAL COLUMN NUCLEI NEURONES R. SUZUKI* and A. H. DICKENSON Department of Pharmacology, University College London, Gower Street, London WC1E 6BT, UK

AbstractöIn the present study we investigated the e¡ects of spinal morphine on the electrically and naturally evoked responses of gracile nuclei neurones in a rat model of neuropathy, induced by the tight ligation of lumbar L5/6 spinal nerves. Two weeks after surgery, animals were prepared for electrophysiological recordings and neuronal responses were characterised to a range of controlled natural (brush, low- and high-intensity von Frey ¢laments, heat 45‡C) and peripheral electrical stimuli. Morphine (0.1, 0.25, 1 and 5 Wg) was applied spinally and its e¡ect was compared to that in sham-operated or naive animals. Following surgery, all neuropathic rats exhibited signs of mechanical allodynia. Nerve injury induced a signi¢cant increase in the receptive ¢eld size of gracile nuclei neurones, and also produced a non-signi¢cant increase in the proportion and level of spontaneous activity in these neurones. The baseline electrical and natural evoked responses remained unaltered. Spinal morphine reduced both the AN-¢bre- and C-¢bre-evoked responses of gracile nuclei neurones, and similarly inhibited the heat-evoked responses of neuropathic, sham-operated and naive rats. Morphine, however, produced only minor reductions ( 6 30% inhibition of pre-drug control responses) of the AL-¢bre- and brush-evoked responses of gracile nuclei neurones. These drug e¡ects were similar in all animal groups. In complete contrast, morphine produced a marked inhibition of the low-intensity punctate mechanical evoked responses (von Freys 2 and 9 g) after nerve injury, an e¡ect that was totally lacking in the sham-operated or naive animal groups. This dramatic shift was selective for the low-intensity punctate mechanical stimuli and such an e¡ect was not seen with the noxious mechanical punctate stimulus (von Frey 75 g) where there was a modest inhibition in all groups. Our results suggest that there is plasticity in the opioid modulation of dorsal column projection pathways following spinal nerve ligation and these alterations appear to interact with sensory pathways conveying low-threshold punctate stimuli. ß 2002 IBRO. Published by Elsevier Science Ltd. All rights reserved. Key words: gracile nuclei, neuropathic pain, electrophysiology, allodynia, morphine, wide dynamic range neurone.

(Koltzenburg et al., 1994; Ossipov et al., 1999) which are normally involved in the mediation of innocuous sensory input. Upon entry into the dorsal root entry zone, these large myelinated primary a¡erent ¢bres bifurcate into ascending and descending branches, and terminate in the dorsal column or spinal cord (Sorkin and Carlton, 1997). Contrary to the general belief that the dorsal column is mainly comprised of large-diameter myelinated ¢bres, it has been shown that C-¢bres can also send their collaterals through dorsal columns from the lumbar enlargement (see Sorkin and Carlton, 1997). In addition to inputs from primary a¡erent ¢bres, the dorsal column nuclei receive ipsilateral projections from the spinal cord through the postsynaptic dorsal column (PSDC) pathway (Willis and Coggeshall, 1991; Sorkin and Carlton, 1997). PSDC neurones respond to both noxious and innocuous stimulation (Angaut-Petit, 1975; Giesler and Cli¡er, 1985; Ferrington et al., 1988; Al-Chaer et al., 1998). Recently, it has been suggested that gracile nuclei (GN) neurones act as a relay for mechanical nociceptive input from the spinal cord through to the thalamic nuclei following peripheral nerve injury (Miki et al., 2000). This suggests an important role of the dorsal column in the transmission of innocuous sensory information but also noxious evoked activity. Peripheral nerve injury also induces neurochem-

Neuropathic pain arising from injury to peripheral nerves often produces sensory abnormalities characterised by spontaneous burning pain, sensory loss, hyperalgesia and allodynia. The mechanisms underlying neuropathic pain are complex and appear to involve various peripheral and central components of sensory systems. Plasticity reported in the spinal cord (Laird and Bennett, 1993; Chapman et al., 1998) and spinothalamic tract neurones (Palecek et al., 1992) following peripheral nerve injury appear to be insu⁄cient to explain all symptoms of the disorder. Many large-diameter ¢bres are known to project to the dorsal column nuclei and these neurones appear to be mainly involved in the relay of non-noxious mechanical sensory information from cutaneous receptors and so may relate to the allodynia seen after nerve injury. Evidence from clinical and animal studies to date suggests that dynamic brush-evoked allodynia is mediated mainly through unmyelinated large-diameter AL-¢bres

*Corresponding author. Tel. : +44-207-679-3737; fax: +44-207-6793742. E-mail address: [email protected] (R. Suzuki). Abbreviations : AP, action potential ; GN, gracile nuclei; PSDC, postsynaptic dorsal column ; SNL, spinal nerve ligation. 215

NSC 5452 8-4-02

216

R. Suzuki and A. H. Dickenson

ical alterations in the GN, which may contribute to some of the symptoms of neuropathic pain (Ma and Bisby, 1998a; Ma et al., 1999). These observations suggest signi¢cant supraspinal plasticity following peripheral nerve injury. Whilst the e⁄cacy of opioids in acute pain conditions is well established, there has been considerable debate over the issue of opioid sensitivity in neuropathic pain states (Dickenson and Suzuki, 1999). We have previously reported minor e¡ects of systemic morphine yet a clear e¡ect of spinal morphine on dorsal horn neurones after nerve injury (Suzuki et al., 1999) which agrees well with the ¢nding of only minimal plasticity within spinal opioid receptor systems following nerve injury (Stevens et al., 1991; Porreca et al., 1998; Go¡ et al., 1998). It is possible that changes in spinal opioid controls also impinge upon dorsal column nuclei. To date, only a very limited number of electrophysiological studies have investigated the e¡ect of peripheral nerve injury on the responses of dorsal column nuclei neurones (Miki et al., 1998; Miki et al., 2000). Hence the aim of this study was to study the responses of GN neurones in normal and neuropathic animals using in vivo electrophysiology, and furthermore, to investigate the e¡ect of intrathecal morphine on the responses of GN neurones after nerve injury. To our knowledge, there have been no electrophysiological studies investigating the e¡ect of pharmacological agents on the responses of GN neurones following peripheral nerve injury. Through the elucidation of spinal and supraspinal opioid receptor plasticity, we should gain a better insight to the mechanisms underlying neuropathic pain.

EXPERIMENTAL PROCEDURES

A total of 25 male Sprague^Dawley rats (Central Biological Services, University College London, UK) were used in this study. All experimental procedures were approved by the Home O⁄ce and follow the guidelines under the International Association for the Study of Pain (Zimmermann 1983). Surgery for spinal nerve ligation (SNL) Experiments were carried out on a sham-operated group (n = 9) and a lumbar L5/L6 SNL group (n = 9). Neuropathic surgery was carried out as previously described (Kim and Chung, 1992). Rats weighing between 140 and 150 g were used for the surgery. The left L5 and L6 spinal nerves were isolated and tightly ligated with 6-0 silk thread under halothane anaesthesia (50% O2 : 50% N2 O). The surgical procedure for the sham-operated group was identical to that of the SNL group, except that the L5/L6 spinal nerves were not ligated. Behavioural tests After surgery, the foot posture and general behaviour of the operated rats were monitored throughout the postoperative period. Behavioural testing was carried out over a 2-week period following procedures previously described to con¢rm the success of the surgery (Suzuki et al., 2000a). Mechanical sensitivity was assessed through the measurement of foot withdrawal frequencies to a series of calibrated von Frey ¢laments (1, 5 and 9 g; 9.9, 49.5, 89.1 mN, respectively) applied to the plantar surface of the foot. The occurrence of foot withdrawal for each trial was quanti¢ed as previously described (Suzuki et al., 2000a).

Electrophysiological studies Two weeks after surgery, the operated animals were used for in vivo electrophysiology as previously described (Suzuki et al., 1999). In addition, a separate group of unoperated naive animals (Sprague^Dawley, 200^250g; n = 7) was also prepared for electrophysiological recordings. In brief, anaesthesia was induced with 2.0^2.5% halothane (66% N2 O and 33% O2 ) and a tracheal cannulation was performed. The rat was placed in a stereotaxic frame to ensure stability during the whole of the electrophysiological recordings. The head of the rat was tilted forward in order to gain better access to the dorsal column, and the rat was secured in this position with a bar held against the skull. This preparation allowed stable recordings to be made from a single GN neurone for up to 7^10 h without any marked change in signal:noise ratios. The level of anaesthesia was subsequently reduced to 1.2^1.8% and the dorsal column nucleus was exposed through blunt dissection of the overlaying muscle. A laminectomy was performed to expose the spinal cord at L1^3 vertebral level so as to allow intrathecal drug administration. The cord was held rigid by clamps caudal and rostral to the exposed section. The core body temperature of the rat was monitored and maintained (36.5^37‡C) by means of a heating blanket connected to a rectal probe via an automatic feedback control unit. Extracellular recordings of ipsilateral GN neurones were made using parylene coated tungsten electrodes. All neurones employed in the study had de¢ned receptive ¢elds in the toe regions of the hindpaw. It is important to note that one of the selection criteria for the neurones employed in the study was the presence of a de¢ned peripheral receptive ¢eld on the plantar surface of the hindpaw, as well as a C-¢bre-evoked response. The area of the neuronal receptive ¢eld was mapped by probing the skin of the hindpaw with three di¡erent von Frey hairs of 9, 15 and 75 g. An area was considered to be within the receptive ¢eld if the application of the von Frey ¢lament-evoked responses greater than 0.5 Hz. The receptive ¢eld area for each von Frey hair was mapped on standard diagrams of the projected area of the plantar surface of the paw. The diagrams were subsequently copied to plain copier paper (80 g/m2 ) and marked areas were carefully cut out and weighed. The receptive ¢eld sizes were measured as the weight of the particular area and ¢nally, expressed as percentage of the mean weight of 10 control diagrams of the whole paw (81.0 þ 0.4 mg). The peripheral receptive ¢eld was stimulated electrically using two needles inserted to the receptive ¢eld of the ipsilateral hindpaw. A train of 16 transcutaneous electrical stimuli (2-ms wide pulses, 0.5 Hz) was applied at three times the threshold current for C-¢bres and a poststimulus histogram (PSTH) was constructed (Fig. 1, top panel). AL- (0^20 ms), AN- (20^90 ms) and C-¢bre-evoked neuronal responses (90^300 ms) were separated and quanti¢ed on the basis of latency (Fig. 1, top panel). The peripheral neuronal receptive ¢eld of the GN neurone was subsequently stimulated using a wide range of noxious and innocuous natural stimuli (mechanical/thermal) over a period of 10 s per stimulus. An example of the neuronal responses to mechanical and heat stimuli is shown in Fig. 1 (bottom panels A and B). Heat was applied with a constant water jet onto the centre of the receptive ¢eld. Data were captured and analysed by a CED 1401 interface coupled to a Pentium computer with Spike 2 software (PSTH and rate functions). Drug administration. Prior to drug administration, control responses to electrical and selected natural stimuli were established. Firstly, tests for electrical stimulation were conducted at three times the threshold current for C-¢bres. This was followed by a quanti¢cation of the spontaneous activity of spinal neurones, prior to the application of natural stimuli, over a period of 100^120 s, or until response stabilisation. The natural stimuli were subsequently applied in the following order: brush, von Frey 2 g, von Frey 9 g, von Frey 75 g, heat (45‡C). Each stimulus was separated by an interval of approximately 7^10 s. Morphine (Evans Medical) was applied directly onto the surface of the spinal cord in a volume of 50 Wl (cumulative doses 0.1, 0.25, 1 and 5 Wg). The e¡ect of the drug was followed over a

NSC 5452 8-4-02

Plasticity in opioid modulation after nerve injury

217

Fig. 1. (Top) An example of a poststimulus histogram of a single GN neurone following a train of 16 stimuli (at three times C-¢bre threshold, 0.5 Hz, 2-ms pulse width) in a SNL rat. The responses evoked by the di¡erent ¢bres were separated and quanti¢ed on the basis of latency measurements (AL-¢bre 0^20 ms; AN-¢bre 20^90 ms; C-¢bre 90^300 ms). Any neuronal responses occurring after the C-¢bre latency band (300^800 ms) were taken to be the postdischarge of the neurone. (Bottom) An example of a rate recording of the responses of a GN neurone to (A) a range of innocuous and noxious von Frey ¢laments (weights 1^75 g) and to (B) heat (32^50‡C) in a SNL rat. All stimuli were applied for 10 s onto the peripheral receptive ¢eld of the neurone located on the hindpaw. The neurone shows a graded response to increasing stimulus intensity. The duration of the stimulus application (10 s) is indicated as a bar above each neuronal response.

time course of 40 min after each dose and tests were carried out at 10-min intervals to determine the e¡ects of morphine on the electrical and natural evoked responses. Following the application of the highest dose of morphine, the e¡ects of morphine were reversed by intrathecal naloxone (50 Wg). Maximal e¡ects of all doses were seen between 20 and 30 min after drug admin-

istration, justifying the 40-min intervals between cumulative dosing. Drug e¡ects were assessed by calculating the mean e¡ect of the drug (expressed as % inhibition) at each dose. Percent inhibition = 1003[(response at x minutes post-drug)/(pre-drug control value)U100]. 100% implies a complete abolition of the

NSC 5452 8-4-02

218

R. Suzuki and A. H. Dickenson

Fig. 2. A schematic diagram of the dorsal column nuclei where recordings of GN neurones were made. Cu, Cuneate nucleus ; Sol, nucleus of the solitary tract. (Paxinos and Watson, 1986.)

responses, whilst a negative value implies a facilitation. ‘Maximum percent inhibition’ indicates the maximum inhibitory e¡ect obtained with morphine within the dose range employed in the study.

details of these behavioural alterations have been described in our previous publications (Suzuki et al., 2000a).

Statistical analysis. Data are presented as mean þ S.E.M. unless stated otherwise. Drug e¡ects are expressed as mean maximal percentage of the pre-drug control value. Drug e¡ects were analysed with non-parametric Wilcoxon signed rank test using Statview 4.5. Mann^Whitney test was employed for the comparison of drug e¡ects between animal groups. Level of signi¢cance was taken to be *P 6 0.05.

Electrophysiological studies

RESULTS

General behaviour and behavioural testing Following surgery, rats maintained good health, exhibiting normal weight gain and general level of activity with no signs of distress. SNL, but not sham-operated rats, exhibited guarding behaviour of the ipsilateral hindpaw. Consistent with previous studies, all SNL rats displayed behavioural signs of mechanical allodynia of the ipsilateral hindpaw (Suzuki et al., 2000a). An increase in the level of allodynia was observed over the postoperative period and this was maintained throughout the whole of the testing period. Neither the contralateral hindpaw of SNL rats nor the hindpaws of sham-operated rats developed modi¢ed mechanical sensitivity. Full

Changes in the response characteristics of GN neurones following SNL. A total of 25 GN neurones (SNL, n = 9; sham-operated, n = 9; naive, n = 7) were employed in the pharmacological study. Fig. 2 shows a schematic diagram of the recording site in the GN. All neurones had de¢ned receptive ¢elds in the toe/plantar region of the ipsilateral hindpaw (Fig. 3A) and displayed characteristics of wide dynamic range neurones, responding to both innocuous stimuli such as light touch or brush, as well as to noxious stimuli (pinch or heat). In addition, although not employed in the present pharmacological study, we encountered neurones which were solely responsive to innocuous stimuli. These low-threshold neurones did not give a sustained response to pinch or other noxious stimuli and failed to code for the intensity of the stimulus. The mean cell depth of GN neurones and the electrical thresholds for neuronal activation are given in Table 1. All GN neurones employed in the study were of similar depths from the surface of the brain stem. Likewise, the neuronal thresholds for C- and AL-¢bre activation were similar in all animal groups. Table 2 shows the baseline control responses of GN neurones to electrical and natural stimuli, prior to drug adminis-

Table 1. The mean cell depth of GN neurones and thresholds for AL- and C-¢bre activation

Cell depth (Wm) C-¢bre threshold (mA) C-¢bre latency (ms) AL-¢bre threshold (mA) AL-¢bre latency (ms)

SNL

Sham-operated

Naive

591 þ 108 2.3 þ 0.3 183 þ 12 0.17 þ 0.02 13 þ 1

540 þ 49 2.4 þ 0.2 189 þ 8 0.19 þ 0.16 12 þ 0.3

595 þ 56 2.4 þ 0.7 182 þ 19 0.12 þ 0.03 12 þ 0.5

NSC 5452 8-4-02

Plasticity in opioid modulation after nerve injury

219

Fig. 3. (A) An example of the receptive ¢eld sizes of a single GN neurone, mapped using von Frey ¢laments 9, 15 and 75 g in a SNL rat. (B) A comparison of the mean receptive ¢eld sizes of GN neurones in SNL, sham-operated and naive animals. The receptive ¢eld area is expressed as a percentage of the projected area of the plantar surface of the paw. *P 6 0.05, statistically signi¢cant di¡erence between study groups.

Table 2. Pre-drug control responses of GN neurones to peripherally applied electrical and natural stimuli in nerve injured, sham-operated and naive rats Pre-drug control responses (number of spikes)

AL-¢bre-evoked response AN-¢bre-evoked response C-¢bre-evoked response Brush-evoked response Mechanical punctate-evoked response :

von Frey 2 g von Frey 9 g von Frey 75 g

Heat-evoked response

SNL

Sham-operated

Naive

86 þ 7 51 þ 9 278 þ 41 403 þ 72 76 þ 19 116 þ 21 475 þ 75 424 þ 69

83 þ 19 37 þ 6 244 þ 34 439 þ 73 60 þ 12 108 þ 26 433 þ 78 459 þ 79

61 þ 13 31 þ 9 195 þ 29 315 þ 73 50 þ 13 89 þ 10 302 þ 58 313 þ 91

Responses to natural stimuli were quanti¢ed over a 10-s period.

NSC 5452 8-4-02

220

R. Suzuki and A. H. Dickenson

Fig. 4. (A) A comparison of the proportion of GN neurones exhibiting spontaneous activities in SNL, sham-operated and naive animals. (B) The magnitude of the ongoing activity in SNL, sham-operated and naive rats. Both the proportion and the magnitude of the spontaneous activity were increased in GN neurones following peripheral nerve injury.

tration. There was a tendency for the AN-¢bre-evoked responses to be greater after nerve injury, as compared to naive or sham controls, however, this was non-signi¢cant. All other pre-drug control responses to electrical and natural stimuli were similar in the three animal groups. These identical baselines allow the e¡ects of morphine to be compared in all animal groups. In marked contrast, nerve injury induced a signi¢cant increase in the receptive ¢eld size of GN neurones, 2 weeks after SNL (Fig. 3B). This was observed with von Frey ¢laments of both low- and high-intensities (von Frey 9 g: SNL vs sham-operated, P = 0.02; SNL vs naive, P = 0.04. von Frey 15 g: SNL vs sham-operated, P = 0.01; SNL vs naive, P s 0.05. von Frey 75 g: SNL vs sham-operated, P = 0.04; SNL vs naive, P = 0.03). Similarly, nerve injury induced a moderate increase in the magnitude and proportion of GN neurones exhibiting spontaneous activities (Fig. 4). Although the magnitude of the ongoing activity was greater after nerve injury, this di¡erence was non-signi¢cant. Thus overall, SNL produced an expansion of the receptive ¢eld sizes of GN neurones to both low- and high-intensity mechanical stimuli 2 weeks after surgery. Changes in pre-drug electrical and natural evoked responses were, however, surprisingly minor, although there appeared to be a tendency for the AN-¢bre-evoked response to be greater in SNL rats.

E¡ect of spinal morphine on the spontaneous activity of GN neurones Spinal morphine produced a signi¢cant reduction of the spontaneous activity of GN neurones with respect to control values (0.1 Wg, P = 0.003; 0.25 and 1 Wg, P = 0.02; 5 Wg, P = 0.04) in SNL rats (n = 6, maximal inhibition: 70 þ 11%). In sham-operated and naive animals, however, morphine produced non-signi¢cant e¡ects (maximal inhibitions: sham-operated, n = 3, 63 þ 17%; naive, n = 2, 56 þ 4%) (data not shown). E¡ect of spinal morphine on the electrical evoked responses of GN neurones Spinal morphine produced minor reductions ( 6 40^ 50% inhibition of pre-drug control value) of the C-¢bre- and AL-¢bre-evoked responses of GN neurones in all animal groups. The e¡ects of morphine on the C-¢bre-evoked response and AL-¢bre-evoked response were signi¢cant with respect to pre-drug control values in SNL (C-¢bre: 0.25 Wg, P = 0.01; 1 Wg and 5 Wg, P = 0.04; AL-¢bre: 1 Wg, P = 0.02; 5 Wg, P = 0.03) and sham-operated (C-¢bre: 1 Wg, P = 0.02; 5 Wg, P = 0.01; AL-¢bre: 1 Wg, P = 0.04; 5 Wg, P = 0.03) rats, but not in naive animals (P s 0.05). No signi¢cant di¡erences were seen in the drug e¡ects between animal groups for both

NSC 5452 8-4-02

Plasticity in opioid modulation after nerve injury

221

Fig. 5. The e¡ect of spinal morphine on the (A) C-¢bre-evoked response, (B) AL-¢bre-evoked response and (C) AN-¢breevoked response of GN neurones in SNL (closed circles), sham-operated (open squares) and naive rats (open triangles). Data are presented as a percentage inhibition of the pre-drug control value þ S.E.M.

neuronal measures (Fig. 5A, B). Similarly, morphine produced a signi¢cant inhibition of the AN-¢bre-evoked response with respect to pre-drug control in SNL (0.1 Wg, P = 0.01; 0.25 Wg and 1 Wg, P = 0.04; 5 Wg, P = 0.03) and sham-operated rats (5 Wg, P = 0.02), however, produced non-signi¢cant e¡ects in naive rats. Drug e¡ects were similar between all animal groups (P s 0.05, Fig. 5C). An example of the e¡ect of spinally administered morphine on the electrical evoked responses of a GN neurone is given in Fig. 6. E¡ect of spinal morphine on the natural evoked responses of GN neurones Spinal administration of morphine produced only minor reductions of the brush-evoked response of GN neurones, and these did not exceed 25%, even at the highest dose of the drug (Fig. 7A). Whilst the responses to brush were non-signi¢cantly reduced in SNL rats (P s 0.05), signi¢cant e¡ects were seen with morphine

in sham-operated (0.25 Wg, P = 0.02; 1 Wg, P = 0.04; 5 Wg, P = 0.01) and naive rats (0.1 Wg, P = 0.03). In contrast, the heat-evoked response of GN neurones was signi¢cantly reduced after morphine in SNL (0.1 Wg, 0.25 Wg, 1 Wg and 5 Wg, P = 0.03), sham-operated (0.1 Wg, P = 0.03; 0.25 Wg and 1 Wg, P = 0.01; 5 Wg, P = 0.03) and naive rats (0.25 Wg, P = 0.04), with respect to predrug controls (Fig. 7B). No di¡erences were seen in the drugs’ e¡ects between animal groups for either measure. In complete contrast to these relatively minor inhibitory e¡ects, a remarkable e¡ect was seen with the mechanical evoked response following the application of low-intensity von Frey ¢laments (Fig. 8A, B). Spinal morphine produced a robust inhibition of the 2 g and 9 g von Frey-evoked response of GN neurones in SNL rats. The magnitude of the reductions exceeded 70% and these e¡ects were clearly signi¢cant compared to pre-drug control values for both von Frey 2 g (0.1 Wg, P = 0.04; 0.25 and 1 Wg, P = 0.01; 5 Wg, P = 0.007) and von Frey 9 g (0.1

NSC 5452 8-4-02

222

R. Suzuki and A. H. Dickenson

Fig. 6. An example of the e¡ect of spinal morphine on the electrical evoked responses of a GN neurone. (A) A trace of the control poststimulus histogram before morphine administration. (B, C) Poststimulus histogram traces after the administration of 0.25 and 5 Wg morphine. Both the C-¢bre- and AN-¢bre-evoked responses are dose-dependently reduced by morphine, whilst the AL-¢bre-evoked responses remain relatively una¡ected.

Wg, P = 0.04; 0.25 Wg, P = 0.01; 1 and 5 Wg, P = 0.001). Interestingly, in complete contrast, spinal morphine produced little or no inhibitory e¡ect on these neuronal responses in sham-operated and naive rats (P s 0.05), and the low-intensity mechanical evoked responses tended to be facilitated by spinal morphine at all doses

(Fig. 8A, B). Overall, the magnitude of the reductions were signi¢cantly greater after nerve injury compared to sham-operated rats (von Frey 2 g: 0.1 Wg, P = 0.001; 0.25 Wg, P = 0.04; 1 Wg, P = 0.01; 5 Wg, P = 0.03; von Frey 9 g: 0.25 Wg, P = 0.01; 1 Wg and 5 Wg, P = 0.04) and naive rats (von Frey 2 g: 0.1 Wg and 0.25 Wg, P = 0.01; 1 Wg,

Fig. 7. The e¡ect of spinal morphine on the (A) brush- and (B) heat-evoked responses of GN neurones in SNL (closed circles), sham-operated (open squares) and naive rats (open triangles). Data are presented as a percentage inhibition of the pre-drug control value þ S.E.M.

NSC 5452 8-4-02

Plasticity in opioid modulation after nerve injury

223

Fig. 8. The e¡ect of spinal morphine on the (A) 2 g von Frey-evoked response, (B) 9 g von Frey-evoked response and (C) noxious 75 g von Frey-evoked response of GN neurones in SNL (closed circles), sham-operated (open squares) and naive rats (open triangles). Data are presented as a percentage inhibition of the pre-drug control value þ S.E.M. *Indicates signi¢cant di¡erence (P 6 0.05) of SNL rats, compared to sham-operated group. ^ Indicates signi¢cant di¡erence (P 6 0.05) of SNL rats, compared to naive rats.

P = 0.005; 5 Wg, P = 0.02; von Frey 9 g: 0.25 Wg, P = 0.03; 1 Wg, P = 0.01, 5 Wg, P = 0.04). Contrary to the marked shift seen with the low-intensity mechanical stimuli after nerve injury, spinal morphine produced similar reductions of the 75 g von Frey-evoked response in all animal groups. Neuronal responses to this noxious mechanical stimulus were signi¢cantly attenuated by morphine in SNL rats (0.25 Wg, P = 0.03; 1 Wg and 5 Wg, P = 0.02), whilst those in shamoperated and naive rats were non-signi¢cantly reduced (Fig. 8C). No group di¡erences were observed on the e¡ects of morphine. Interestingly, the overall magnitude of inhibition of the noxious 75 g von Frey-evoked response in SNL rats was smaller (maximal inhibition: 50 þ 13%) than that of the 2 g (maximal inhibition: 83 þ 6%) and 9 g von Frey-evoked responses (maximal inhibition: 70 þ 9%).

The inhibitions produced by spinal morphine were partially reversed by the administration of spinal naloxone (50 Wg). For example, the C-¢bre-evoked response and the low-threshold mechanical evoked responses were reversed to 68 þ 16%, 60 þ 20% (von Frey 2 g) and 58 þ 19% (von Frey 9 g) of pre-drug control values, respectively.

DISCUSSION

In the present study, we investigated the e¡ect of spinally administered morphine on the responses of GN neurones 2 weeks following peripheral nerve injury. To our knowledge, this is the ¢rst electrophysiological study investigating the e¡ectiveness of any pharmacological agent on the responses of GN neurones after neuropa-

NSC 5452 8-4-02

224

R. Suzuki and A. H. Dickenson

thy. All neurones had vigorous responses to noxious mechanical and thermal stimuli and received C-¢bre inputs ^ this was the main selection criterion for the neurones employed in the study. Many low-threshold neurones were encountered but not studied further for the pharmacological study. Comparison of the neuronal plasticity in the response characteristics of GN neurones and spinal cord dorsal horn neurones Overall, the response characteristics of GN neurones we report here, were surprisingly similar to those reported previously with spinal dorsal horn neurones (Chapman et al., 1998; Suzuki et al., 1999). Both the AL-¢bre (GN: 86 þ 7 action potentials (AP); spinal cord: 71 þ 4 AP) and C-¢bre-evoked responses (GN: 278 þ 41 AP; spinal cord: 229 þ 14 AP) of SNL rats were similar between GN neurones and spinal neurones. The graded response exhibited by GN neurones to increasing mechanical and thermal stimuli was also similar to that seen spinally with lamina V-type wide dynamic range neurones. Furthermore, both the magnitude (GN: 1.9 þ 0.4 Hz; spinal cord 2.2 þ 0.7 Hz) and proportion of neurones exhibiting spontaneous activity after neuropathy (GN: 67%; spinal cord 72%) were similar between the two neuronal groups. In contrast, there appeared to be a tendency for the electrical AL-¢bre and C-¢bre thresholds of SNL rats to be higher in GN neurones (AL-¢bre threshold: 0.17 þ 0.02mA; C-¢bre threshold: GN: 2.3 þ 0.3mA), when compared to the spinal cord (AL-¢bre threshold: 0.12 þ 0.02mA; C-¢bre threshold: 1.1 þ 0.1mA). The electrical threshold for C-¢bre activation, but not for AL-¢bre activation, was signi¢cantly higher in GN neurones (P = 0.02), as compared to spinal neurones. We have previously reported that nerve injury induces a number of alterations in the response characteristics of spinal neurones (Chapman et al., 1998; Suzuki et al., 1999; Suzuki et al., 2000b), which are characterised by a decrease in the magnitude of C-¢bre-evoked responses, an increase in the level of ongoing activity, as well as an increase in the receptive ¢eld size to low-intensity mechanical stimuli (von Frey 9 g). In GN neurones, we observed signi¢cant increases in the neuronal receptive ¢eld size, as well as a moderate increase in the level of spontaneous activity. The expansion of receptive ¢eld was much more pronounced in GN neurones, compared to that in the spinal cord, and signi¢cant changes were observed with both low- and high-intensity mechanical stimuli. Thus, in contrast to spinal neurones (Suzuki et al., 2000b), all mechanical responses exhibited increased receptive ¢elds in GN neurones, and this may be su⁄cient to produce both allodynia and hyperalgesia after nerve injury. No changes were observed in the magnitude of C-¢bre-evoked responses following nerve injury in GN neurones. Plasticity of GN neurones following peripheral nerve injury A previous electrophysiological study on dorsal col-

umn nuclei neurones demonstrated that there is an increase in the level of spontaneous activity, as well as in neuronal afterdischarge, following chronic constriction injury (Miki et al., 1998). Additionally, Miki et al. (1998) report that GN neurones respond to mechanical and thermal stimuli in the noxious and innocuous range. Consistent with these ¢ndings, we demonstrate that GN neurones respond to a wide range of peripheral stimuli (electrical, noxious and innocuous natural) in a graded manner. Additionally, we observed a signi¢cant increase in the receptive ¢eld size of GN neurones to low- and high-intensity mechanical stimuli, an e¡ect which was not reported by Miki et al. (1998). With exception to changes in receptive ¢eld size, overall there were no dramatic alterations in the evoked responses of GN neurones 2 weeks after peripheral nerve injury, and the changes we report here were relatively minor. Hence, similar to the e¡ect seen in spinal cord (Chapman et al., 1998; Suzuki et al., 1999), the electrophysiological changes that accompany neuropathy are relatively minor compared to the robust behavioural changes seen in these animals. It is yet unclear what underlies the behavioural manifestation of allodynia and hyperalgesia, however, the enlargement of receptive ¢eld size observed in GN and spinal neurones, together with the decreased mechanical threshold and increased spontaneous activity of spinal neurones, are all likely to play a role in the pathophysiology of neuropathic pain states. Consistent with this view, several lines of evidence have demonstrated neurochemical plasticity in GN neurones following peripheral nerve injury. Sciatic nerve transection induces a de novo expression of preprotachykinin mRNA in medium to large dorsal root ganglion cells that project to the GN, and increased SP is transported to the GN through large myelinated ¢bres (Noguchi et al., 1995). Similar increases in calcitonin gene-related peptide (CGRP) (Ma and Bisby, 1998a), brain-derived neurotrophic factor (BDNF) (Cho et al., 1998) and neuropeptide Y (Ma and Bisby, 1998b) have also been reported following partial and complete sciatic nerve injuries. Hence the increased expression of these excitatory pain-related neuropeptides in the GN may similarly contribute to the generation of some of the neuropathic pain behaviours. Interestingly, Molander et al. (1992) demonstrated that low-threshold electrical stimulation of injured nerves at AL-¢bre intensities produces increased c-fos expression in the GN following sciatic nerve transection. Under normal conditions where the nerves remain intact, only few GN neurones express Fos-like immunoreactivity, therefore nerve injury may alter the excitability of the dorsal column-medial lemniscus pathway (Tokunaga et al., 1999). E¡ect of spinal morphine on the responses of GN neurones Our present results demonstrate that spinal morphine is e¡ective in reducing the responses of GN neurones to electrical and natural (low- and high-intensity mechanical evoked response, heat-evoked response) stimuli in SNL rats. Overall, spinal morphine produced similar

NSC 5452 8-4-02

Plasticity in opioid modulation after nerve injury

moderate inhibitions of the electrical (AL-, AN- and C-¢bre) and high-threshold natural (von Frey 75 g, heat) evoked responses in SNL, sham-operated and naive rats (maximal inhibitions 6 50%), and these e¡ects were similar in all animal groups. Interestingly, despite the relative lack of inhibition of the AL-¢bre-evoked response, a remarkable di¡erence however, was seen with the responses to low-intensity mechanical punctate stimuli (von Freys 2 and 9 g), which were markedly inhibited by morphine following nerve injury (maximal inhibition s 70%). This may relate to the fact that these natural mechanical punctate stimuli, which may include a contribution from both AL- and AN-¢bres, produce a pattern of activity more amenable to inhibition than the suprathreshold synchronised responses elicited by the electrical activation of AL-¢bres. In complete contrast, the responses to low-intensity mechanical punctate stimuli remained relatively una¡ected by morphine in shamoperated and naive animals (maximal inhibition 6 10%). This dramatic shift appeared to be selective for lowintensity mechanical punctate stimuli since such an e¡ect was not seen with the noxious 75 g von Frey-evoked response. Our results would therefore suggest for an alteration in spinal opioid inhibitory control, particularly on the low-threshold signalling system following nerve injury. Not only were we somewhat surprised by the marked change in opioid sensitivity of the low-threshold mechanical punctate responses of GN neurones (von Freys 2 and 9 g), but the robust responses of the same GN neurones to noxious stimuli in all three groups were equally unexpected. The dorsal column nuclei receive ipsilateral input from primary a¡erent ¢bres, as well as ipsilateral projections from the spinal cord through the PSDC pathway (see Sorkin and Carlton, 1997). One possibility is that spinal morphine directly inhibits the PSDC pathway. Alternatively, spinal morphine can hyperpolarise nerve terminals of small primary a¡erent ¢bres and this could lead to the dampening of excitability in collateral ¢bres running to the GN, especially at branch points. Whilst C-¢bres are responsible for noxious sensory transmission (von Frey 75 g, heat), activity evoked by low-intensity mechanical punctate stimuli (von Freys 2 and 9 g) is likely to be carried by AL- and AN-¢bres. If AL-¢bres are solely responsible for the supraspinal transmission of low-intensity mechanical evoked response, it is unlikely that the robust inhibition of these responses has resulted from an action of spinal morphine on primary a¡erent terminals, since AL-¢bres appear not to express opioid receptors (Arvidsson et al., 1995). In the present study, spinal morphine produced little or no inhibitory e¡ect on the responses of GN neurones to low-intensity mechanical punctate stimuli in sham-operated and naive animals. After nerve injury, however, both the 2 and 9 g von Frey-evoked responses were markedly reduced following morphine administration. Hence the increased e¡ectiveness of spinal morphine seen after neuropathy may be due to an acquirement of de novo opioid sensitivity of AL-¢bres, possibly through a phenotypic switch. Similarly, it may also be possible that low-threshold projections to the dorsal col-

225

umn via the PSDC pathway acquire opioid sensitivity after nerve injury. To date, there is still no evidence to support this idea. However, numerous studies have provided evidence for neuronal plasticity in GN neurones following peripheral nerve injury (Noguchi et al., 1995; Ma and Bisby 1998a,b; Miki et al., 1998; Cho et al., 1998; Tokunaga et al., 1999). Another possibility is that AN-¢bres are also involved in the transmission of low-threshold mechanical punctate stimuli (von Freys 2 and 9 g) in addition to the largediameter AL-¢bres. AN-¢bres express opioid receptors on their nerve terminals (Beland and Fitzgerald, 2001) hence morphine could inhibit the excitability of AN-¢bre collaterals projecting supraspinally to the dorsal column, and attenuate the responses evoked by low-intensity mechanical punctate stimuli. There is evidence to suggest that there is an increase in soybean agglutinin (tracer for small-sized ¢bres) labelling after axotomy which may re£ect sprouting of C-¢bres or AN-¢bres in these areas (Shortland et al., 1999). If such sprouting occurs with AN-¢bres in the SNL model of neuropathy, it may relate to the enhanced inhibitory e¡ects of spinal morphine on the static low-threshold mechanical evoked responses (von Frey 2 and 9 g). Furthermore, punctate hyperalgesia in humans has been suggested to be AN-¢bre-mediated, and in animal behavioural studies, static allodynia is morphine-sensitive (Field et al., 1999). The e⁄cacy of opioids following peripheral nerve injury Although the e¡ectiveness of morphine in acute and in£ammatory pain states is well established, its e⁄cacy in neuropathic pain states still remains largely debatable and there appears to be little consensus from previous animal/clinical studies. W-Opioid receptors are located in the GN, as well as in laminae III^IV of the spinal cord (Ding et al., 1996). W-Opioid receptors in the spinal cord are known to undergo some degree of plasticity following peripheral nerve injury (Stevens et al., 1991; Go¡ et al., 1998; Porreca et al., 1998), however, little is known on the alterations that take place in the GN following neuropathy. Recent studies have demonstrated that ascending dorsal column projections play an important role in mediating tactile allodynia (Houghton et al., 1999; Sun et al., 2001). In the latter study it was shown that lesions or temporary block of ipsilateral dorsal columns in SNL rats abolished tactile allodynia. This strongly suggests that this modality is relayed through dorsal column pathways. Behavioural evidence to date indicates that spinal morphine is generally ine¡ective in relieving tactile (static) allodynia in neuropathic pain models (Lee et al., 1995; Bian et al., 1995). However, the neuronal counterpart of this activity, as recorded in these dorsal column cells, is sensitive to spinal morphine administration. One reason for this phenomenon may be due to the fact that some forms of allodynia (i.e. dynamic) is largely mediated via large-diameter, myelinated AL-¢bres, which lack opioid receptors (Zhang et al., 1998). A recent study has suggested that changes in supraspinal processing contribute to the observed poor e⁄cacy of opioids in some behavioural models of neuro-

NSC 5452 8-4-02

226

R. Suzuki and A. H. Dickenson

pathic pain, and activation of descending nociceptive facilitatory pathways is important in the maintenance of neuropathy (Kovelowski et al., 2000). We have recently discussed some of the discrepancies in animal/ clinical literature regarding the e¡ects of morphine (Suzuki et al., 1999; Dickenson and Suzuki, 1999). The route, the type of allodynia (static/dynamic), the force of the stimulus employed and the measure (threshold/suprathreshold) will all be factors in determining the e¡ectiveness of morphine. There is su⁄cient clinical evidence to support the idea that opioids can alleviate neuropathic pain (Portenoy et al., 1990; Rowbotham et al., 1991; Jadad et al., 1992). Implications for mechanisms of allodynia Overall, behavioural evidence to date suggests that in the SNL model of neuropathy, morphine administered via the intrathecal route is ine¡ective against allodynia (Bian et al., 1995; Lee et al., 1995; Wegert et al., 1997), although it produces antinociceptive and antihyperalgesic e¡ects (Wegert et al., 1997). Morphine administered via the systemic or supraspinal route, however was e¡ective against this measure (Lee et al., 1995). We have previously demonstrated that intrathecally administered morphine is e¡ective in reducing both the electrical (ANand C-¢bre-evoked response) and natural evoked responses (low- and high-intensity mechanical punctate/ heat) of spinal neurones following SNL (Suzuki et al., 1999). Furthermore, the e¡ectiveness of spinal morphine appeared to be enhanced after nerve injury at both 1 and 2 postoperative weeks. Overall, there appears to be a degree of discrepancy between the results of some previous behavioural studies with our current ¢ndings, whereby the e¡ectiveness of spinal morphine in inhibiting responses to low-threshold mechanical punctate stimuli was signi¢cantly enhanced in GN neurones after nerve injury. One obvious di¡erence is that the behavioural responses measure thresholds whereas drug e¡ects on suprathreshold responses are studied in our electrophysiological model. It is important to note that the low- and high-threshold mechanical punctate stimuli (von Frey ¢laments) employed in this study, may not directly relate to the behavioural measures of allodynia or hyperalgesia. Our results suggest that following SNL, there are alterations in the receptor system located in low-threshold sensory pathways, which project to the dorsal column. It has been proposed that there are two distinct types of mechanical allodynia in patients with neuropathic pain: static and dynamic (Ochoa and Yarnitsky, 1993), which appear to be mediated by di¡erent a¡erent ¢bres. Recent studies have shown that both types of allodynia exist in animal models of neuropathy (Field et al., 1999). Morphine, which is generally believed to be ine¡ective against mechanical allodynia, was shown to attenuate static, but

not dynamic, components of allodynia, in the SNL model of neuropathy (Field et al., 1999). This suggests that brush-evoked dynamic allodynia is mediated by AL¢bres which are insensitive to morphine, whilst AN-¢bres are largely responsible for static allodynia. Thus it is possible that the e¡ects of morphine we observed on von Frey 2 g and 9 g involves a morphine-sensitive AN-¢bre component following nerve injury. This is further supported by our present ¢ndings where we demonstrate that electrically AN-¢bre-evoked responses of SNL rats are more sensitive to the administration of spinal morphine (maximal inhibition: 52%) compared to AL¢bre-evoked responses ( 6 25%). Similar e¡ects have been demonstrated in spinal dorsal horn neurones in this animal model (Suzuki et al., 1999). In addition, we have previously shown that spinal bicuculline, which markedly facilitates AN-¢bre-evoked responses of spinal neurones, produces a similar facilitation of the static prod-evoked response, whilst exerting little e¡ect on the dynamic brush-evoked response (Reeve et al., 1998). These results strongly support the idea that AN-¢bres are responsible for mediating activity evoked by static mechanical (von Frey, prod) stimuli. Leung et al. (2001) have recently demonstrated that intravenous infusion of alfentanil (a W opiate) signi¢cantly attenuates spontaneous and von Frey pain scores in neuropathic pain patients. Most behavioural studies to date have assessed tactile allodynia through the use of von Frey ¢laments (static allodynia), hence the ¢nding in some studies, that spinal morphine is ine¡ective against this behavioural measure in animal models (Lee et al., 1995; Bian et al., 1995) yet e¡ective against the electrophysiological counterpart of the stimulus, is intriguing. Whilst opioid receptor plasticity in the spinal cord following peripheral nerve injury has been studied, little is known on the supraspinal changes accompanying altered opioid sensitivities in neuropathic pain states. We extend our observations on the e¡ectiveness of intrathecal morphine on spinal neurones in this model (Suzuki et al., 1999), by demonstrating that spinally administered morphine also attenuates the responses of GN neurones to selected peripheral natural and electrical stimuli. The ¢nding that morphine can e¡ectively attenuate the responses of GN neurones to static low-threshold mechanical stimuli may have important clinical implications for neuropathic pain patients with allodynia since its use may prove to be bene¢cial in at least some patients. Future studies on the e¡ects of systemic morphine and on the neural pathways that underlie the neuronal responses we report should provide further clues as to the relative roles of spinal and dorsal column neurones in the symptoms of neuropathy.

AcknowledgementsöThis work was supported by the Wellcome Trust.

NSC 5452 8-4-02

Plasticity in opioid modulation after nerve injury

227

REFERENCES

Al-Chaer, E., Feng, Y., Willis, W., 1998. A role for the dorsal column in nociceptive visceral input into the thalamus of primates. J. Neurophysiol. 79, 3143^3150. Angaut-Petit, D., 1975. The dorsal column system : II. Functional properties and bulbar relay of the postsynaptic ¢bres of the cat’s fasciculus gracilis. Exp. Brain Res. 22, 471^493. Arvidsson, U., Riedl, M., Chakrabarti, S., Lee, J., Nakano, A., Dado, R., Loh, H., Law, P., Wessendorf, M., Elde, R., 1995. Distribution and targeting of a mu-opioid receptor (MOR1) in brain and spinal cord. J. Neurosci. 15, 3328^3341. Beland, B., Fitzgerald, M., 2001. Mu- and delta-opioid receptors are downregulated in the largest diameter primary sensory neurons during postnatal development in rats. Pain 90, 143^150. Bian, D., Nichols, M.L., Ossipov, M.H., Lai, J., Porreca, F., 1995. Characterization of the antiallodynic e⁄cacy of morphine in a model of neuropathic pain in rats. NeuroReport 6, 1981^1984. Chapman, V., Suzuki, R., Dickenson, A.H., 1998. Electrophysiological characterization of spinal neuronal response properties in anaesthetized rats after ligation of spinal nerves L5-L6. J. Physiol. 507, 881^894. Cho, H.J., Kim, J.K., Park, H.C., Kim, D.S., Ha, S.O., Hong, H.S., 1998. Changes in brain-derived neurotrophic factor immunoreactivity in rat dorsal root ganglia, spinal cord, and gracile nuclei following cut or crush injuries. Exp. Neurol. 154, 224^230. Dickenson, A., Suzuki, R., 1999. Function and dysfunction of opioid receptors in the spinal cord. In: Kalso, E., McQuay, H., Wiesenfeld-Hallin, Z. (Eds.), Opioid sensitivity of chronic noncancer pain. Progress in pain research and management. Vol. 14, IASP, Seattle, WA, pp. 17^44. Ding, Y., Kaneko, T., Nomura, S., Mizuno, N., 1996. Immunohistochemical localization of mu-opioid receptors in the central nervous system of the rat. J. Comp. Neurol. 367, 375^402. Ferrington, D., Downie, J., Willis, W., Jr., 1988. Primate nucleus gracilis neurons: responses to innocuous and noxious stimuli. J. Neurophysiol. 59, 886^907. Field, M., Bramwell, S., Hughes, J., Singh, L., 1999. Detection of static and dynamic components of mechanical allodynia in rat models of neuropathic pain: are they signalled by distinct primary sensory neurones ? Pain 83, 303^311. Giesler, G., Cli¡er, K., 1985. Postsynaptic dorsal column pathway of the rat. II. Evidence against an important role in nociception. Brain Res. 326, 347^356. Go¡, J.R., Burkey, A.R., Go¡, D.J., Jasmin, L., 1998. Reorganization of the spinal dorsal horn in models of chronic pain: correlation with behaviour. Neuroscience 82, 559^574. Houghton, A., Hewitt, E., Westlund, K., 1999. Dorsal column lesion prevents mechanical hyperalgesia sand allodynia in osteotomy model. Pain 82, 73^80. Jadad, A.R., Carroll, D., Glynn, C.J., Moore, R.A., McQuay, H.J., 1992. Morphine responsiveness of chronic pain: double-blind randomised crossover study with patient-controlled analgesia. Lancet 339, 1367^1371. Kim, S.H., Chung, J.M., 1992. An experimental model for peripheral neuropathy produced by segmental spinal nerve ligation in the rat. Pain 50, 355^363. Koltzenburg, M., Torebjork, H.E., Wahren, L.K., 1994. Nociceptor modulated central sensitization causes mechanical hyperalgesia in acute chemogenic and chronic neuropathic pain. Brain 117, 579^591. Kovelowski, C., Ossipov, M., Sun, H., Lai, J., Malan, T., Porreca, F., 2000. Supraspinal cholecystokinin may drive tonic descending facilitation mechanisms to maintain neuropathic pain in the rat. Pain 87, 265^273. Laird, J.M., Bennett, G.J., 1993. An electrophysiological study of dorsal horn neurons in the spinal cord of rats with an experimental peripheral neuropathy. J. Neurophysiol. 69, 2072^2085. Lee, Y.W., Chaplan, S.R., Yaksh, T.L., 1995. Systemic and supraspinal, but not spinal, opiates suppress allodynia in a rat neuropathic pain model. Neurosci. Lett. 199, 111^114. Leung, A., Wallace, M., Ridgeway, B., Yaksh, T., 2001. Concentration-e¡ect relationship of intravenous alfentanil and ketamine on peripheral neurosensory thresholds, allodynia and hyperalgesia of neuropathic pain. Pain 91, 177^187. Ma, W., Bisby, M.A., 1998a. Increase of calcitonin gene-related peptide immunoreactivity in the axonal ¢bers of the gracile nuclei of adult and aged rats after complete and partial sciatic nerve injuries. Exp. Neurol. 152, 137^149. Ma, W., Bisby, M.A., 1998b. Partial and complete sciatic nerve injuries induce similar increases of neuropeptide Y and vasoactive intestinal peptide immunoreactivities in primary sensory neurons and their central projections. Neuroscience 86, 1217^1234. Ma, W., Ramer, M.S., Bisby, M.A., 1999. Increased calcitonin gene related peptide immunoreactivity in gracile nucleus after partial sciatic nerve injury: age dependent and originating from spared sensory neurons. Exp. Neurol. 159, 459^473. Miki, K., Iwata, K., Tsuboi, Y., Morimoto, T., Kondo, E., Dai, Y., Ren, K., Noguchi, K., 2000. Dorsal column thalamic pathway is involved in thalamic hyperexcitability following peripheral nerve injury : a lesion study in rats with experimental mononeuropathy. Pain 85, 263^271. Miki, K., Iwata, K., Tsuboi, Y., Sumino, R., Fukuoka, T., Tachibana, T., Tokunaga, A., Noguchi, K., 1998. Responses of dorsal column nuclei neurons in rats with experimental mononeuropathy. Pain 76, 407^415. Molander, C., Hongpaisan, J., Grant, G., 1992. Changing pattern of c-fos expression in spinal cord neurons after electrical stimulation of the chronically injured sciatic nerve in the rat. Neuroscience 50, 223^236. Noguchi, K., Kawai, Y., Fukuoka, T., Senba, E., Miki, K., 1995. Substance P induced by peripheral nerve injury in primary a¡erent sensory neurons and its e¡ect on dorsal column nucleus neurons. J. Neurosci. 15, 7633^7643. Ochoa, J., Yarnitsky, D., 1993. Mechanical hyperalgesias in neuropathic pain patients: dynamic and static subtypes. Ann. Neurol. 33, 465^472. Ossipov, M.H., Bian, D., Malan, T.P., Jr., Lai, J., Porreca, F., 1999. Lack of involvement of capsaicin-sensitive primary a¡erents in nerve-ligation injury induced tactile allodynia in rats. Pain 79, 127^133. Palecek, J., Dougherty, P.M., Kim, S.H., Paleckova, V., Lekan, H., Chung, J.M., Carlton, S.M., Willis, W.D., 1992. Responses of spinothalamic tract neurons to mechanical and thermal stimuli in an experimental model of peripheral neuropathy in primates. J. Neurophysiol. 68, 1951^ 1966. Paxinos, G, Watson, C., 1986. The rat brain in stereotaxic coordinates. Academic Press, New York. Porreca, F., Tang, Q.B., Bian, D., Riedl, M., Elde, R., Lai, J., 1998. Spinal opioid mu receptor expression in lumbar spinal cord of rats following nerve injury. Brain Res. 795, 197^203. Portenoy, R.K., Foley, K.M., Inturrisi, C.E., 1990. The nature of opioid responsiveness and its implications for neuropathic pain: new hypotheses derived from studies of opioid infusions. Pain 43, 273^286. Reeve, A.J., Dickenson, A.H., Kerr, N.C., 1998. Spinal e¡ects of bicuculline : modulation of an allodynia-like state by an A1-receptor agonist, morphine, and an NMDA-receptor antagonist. J. Neurophysiol. 79, 1494^1507. Rowbotham, M.C., Reisner-Keller, L.A., Fields, H.L., 1991. Both intravenous lidocaine and morphine reduce the pain of postherpetic neuralgia. Neurology 41, 1024^1028.

NSC 5452 8-4-02

228

R. Suzuki and A. H. Dickenson

Shortland, P., Wang, H.F., Molander, C., 1999. Distribution of transganglionically labelled soybean agglutinin primary a¡erent ¢bres after nerve injury. Brain Res. 815, 206^212. Sorkin, L., Carlton, S., 1997. Spinal anatomy and pharmacology of a¡erent processing. In: Yaksh, T., Lynch, C., Zapol, W., Maze, M., Biebuyck, J., Saidman, L. (Eds.), Anesthesia. Biologic Foundations. Lippincott-Raven, Philadelphia, PA, pp. 577^610. Stevens, C.W., Kajander, K.C., Bennett, G.J., Seybold, V.S., 1991. Bilateral and di¡erential changes in spinal mu, delta and kappa opioid binding in rats with a painful, unilateral neuropathy. Pain 46, 315^326. Sun, H., Ren, K., Zhong, C., Ossipov, M., Malan, T., Jr, Lai, J., Porreca, F., 2001. Nerve injury-induced tactila allodynia is mediated via ascending spinal dorsal column projections. Pain 90, 105^111. Suzuki, R., Chapman, V., Dickenson, A., 1999. The e¡ectiveness of spinal and systemic morphine on rat dorsal horn neuronal responses in the spinal nerve ligation model of neuropathic pain. Pain 80, 215^228. Suzuki, R., Gale, A., Dickenson, A., 2000a. The inhibitory e¡ects of the A1 adenosine receptor agonist, N6-cyclopentyladenosine on the electrical and natural evoked responses of spinal dorsal horn neurones in a rat model of mononeuropathy. J. Pain 1, 99^110. Suzuki, R., Kontinen, V., Matthews, E., Dickenson, A., 2000b. Enlargement of receptive ¢eld size to low intensity mechanical stimulation in the rat spinal nerve ligation model of neuropathy. Exp. Neurol. 163, 408^413. Tokunaga, A., E, K., Fukuoka, T., Miki, K., Dai, Y., Tsujino, H., Noguchi, K., 1999. Excitability of spinal cord and gracile nucleus neurons in rats with chronically injured sciatic nerve examined by c-fos expression. Brain Res. 847, 321^331. Wegert, S., Ossipov, M.H., Nichols, M.L., Bian, D., Vanderah, T.W., Malan, T.P., Jr., Porreca, F., 1997. Di¡erential activities of intrathecal MK801 or morphine to alter responses to thermal and mechanical stimuli in normal or nerve-injured rats. Pain 71, 57^64. Willis, W., Coggeshall, R., 1991. Sensory patways in the dorsal funiculus. In: Sensory mechanisms of the spinal cord. Plenum, New York, pp. 245^ 306. Zhang, X., Bao, L., Shi, T.J., Ju, G., Elde, R., Hokfelt, T., 1998. Down-regulation of mu-opioid receptors in rat and monkey dorsal root ganglion neurons and spinal cord after peripheral axotomy. Neuroscience 82, 223^240. Zimmermann, M., 1983. Ethical guidelines for investigations of experimental pain in conscious animals. Pain 16, 109^110. (Accepted 23 November 2001)

NSC 5452 8-4-02