A STUDY OF INTRATHECAL STRYCHNINE-INDUCED ALLODYNIA IN THE LIGHTLY ANESTHETIZED RAT

by

Hemal Khandwala

A thesis subrnitted to the School of Graduate Studies in

partial fulfilrnent of the requirements for the degree of

Master of Science

Specialization: Phannacology

Schwl of Phamacy

Mernorial University of Newfoundland

St. John's, Newfoundland

&dl1997 National Library Bibliothèque nationale du Canada Acquisitions and Acquisitions et Bibliographie Services services bibliographiques 395 Wellington Street 395, nie Wellington OttawaON KIAON4 ôttawa ON K1A ON4 Canada canada

The author has granted a non- L'auteur a accorde une licence non exclusive licence allowing the exclusive permettant à la National Library of Canada to Bibliothèque nationale du Canada de reproduce, loan, distribute or seU reproduire, prêter, distribuer ou copies of this thesis in microform, vendre des copies de cette thèse sous paper or electronic formats. la forme de microfiche/nlm, de reproduction sur papier ou sur format électronique.

The author retains ownership of the L'auteur conserve la propriete du copyright in this thesis. Neither the droit d'auteur qui protège cette thèse. thesis nor substantial extracts fiom it Ni la thèse ni des extraits substantiels may be printed or otherwise de celle-ci ne doivent être imprimes reproduced without the author's ou autrement reproduits sans son permission. autorisation. ABSTRACT

Alloâynia, an outcame of neural injury, is a condition in which non-painful stimuli (eg. light brushing or Iight touch of ciothing) evoke pain. Unlike normal

(nociceptive) pain, allodynia is triggered by input from low-threshold (AB) afferent fibers. Nomially, the synaptic transmission of Wnput appean to be modulated by glycine released from spinal intemeurons. Thus, the blockade of spinal glycine receptors with intrathecal (i.t.) strychnine (STR) produces reversible, segmentally localized allodynia in the rat; a condition blodred by the spinal administration of excitatory amino acid (EAA) receptor antagonists. These results suggest mat, in the absence of spinal glycinergic modulation, input from Ap-fibers aaquire access to spinal pain-signalling pathways. If this hypothesis is correct, then i.t STR- inâuced allodynia should be blodced by antagonists of spinal NMDAieceptors, and by neuromodulatory systems that modulate EAA-mediated synaptic transmission in the spinal cotd. Therefore, the specific objectives of this study were to: 1) determine the effed of i.t. AP-7 (NMDA-antagonist), cyciopentyladenosine (A,- agonie CPA), CGS-21680 (4-agonist), muscimol (GABAA-agonist)and baclofen

(GABA,-agonist) on STR-allodynia; 2) campare the effed of intravenous mexiletine on STRallodynia and calibrated bilateral pawpinch (without STR); 3) detemine the Mect of intravenous milacemide (a glycine prodnig and substrate for MAO-0) on STR-allodynia; and 4) test the sensitivity of this milacemide effect to

ii pretreatment with dorgyline (MAOA inhibitor) or 1-deprenyl (MAO-0 inhibitor).

Male, Sprasue- DAeyrats, fitted with i.t catheten, were lightly anesthetized with urethane. Stimulus-8voked changes in blood pressure and heart rate, and cortical elecboencephalogaphic (EEG) acüvity were continuously reoorded. After i.t STR

(40 pg), repetitive brushing d the hair with a mon tipped applicator (a nomially ineffective stimulus) evokd a marked and progressive inuease in mean arterial pressure (W)and heart rate (HR), an abrupt motor wïthdrawal(WD) response and desynchronkation of the EEG; effeds comparable to those elicited by intradermal mustard oil without STR. Pretreatment with i.t. AP-7, CPA, CGS and muscimol, but not baclofen, dosedependently inhibited STR-allodynia without affeding EEG syndirony. The ED* for i. t. AP-7 were: 1-12 pg for the evokeâ change MAP, 1.71 pg for the evoked change in HR, and 0.36 pg for the evoked change in withdrawal duration. The corresponding EDd ranged ftom 0.02-0.07 pg for CPA; 2.69-3.1 1 pg for CGS and 0.444.51 ug for muscimol. lntravenous mexiletine inhibited the responses ev&d by mxious paw-pinch (without STR) and hair deflection in STR-treated rats with equal potency (€O& = 9-1 7 mgkg). There was no conamcit change in EEG synchrony and no blockade of motor efferent pathways. lntravenous milacemide inhibited STR-allodynia (ED* = 398-404 mgkg) at doses having no effed on normal nociception. This selective anti- allodynic effect was blocked by pretreatment with Ideprenyl, but not with dorgyline.

iii The results daisstudy support the following condusions. 1) Glycine is a selective modulator of non-noxious somatosensory input in the spinal cord of the rat. 2) in the presenae d i.t STR, low oueshou afferent input aocesses an NMDA-mediated spinai sensitisation medianism normally adivated by high threshold (nociceptive) fibers. 3) Systemic mexiletine inhibits both abmal(STR-allodynia) and nomial

(noxious pinch) nociception at a cornmon central site, most likely the dorsal hom cells receiving convergent Ak and C-fiber input 4) The recniitment of neuromoâulatory systems utilizing spinal A,- and GAB4 -receptors blocks Vie abnormal responses to hair clefledion during STR Mockade of spinal glycine receptors. 5) lncreasing central glycine concentration by means of a systemic glycine prodrug (milacemide) nomalizes the spinal processing of low threshold afferent input in STR-allodynia. 6) The spinal phamacology of STR-allodynia is similar to that of dinical allodynia, and clearly distinct from that of nomal nociception. 7) STR-allodynia is a rapid, selective and valid noni'njury model for investigating this abnormal pain state. ACKNOWLEDGEMENTS

I wish to express my sincere appreciation to my researeh supervisor, Dr.

Christopher Loomis for his expert guidance Viroughout my research and thorough

review of this thesis. I am also grateful to Dr. Detlef Bieger for his advice.

I would also like to extend my gratitude to Mr. John Cruz Bautista, Ms. Elizabeth

Hodge and Ms. Janet Robinson for their skilful technical assistance and their friendship.

I would like to thank the faculty mernbers and staff of the Schwl of Phamacy for th& encouragement and kind help.

I am grateful to the School of Phamacy, School of Graduate Studies, the

Phamaceutical Manufacturer's Association Canada - Health Researcb Foundation and the Medical Research Coumil of Canada for providing financial support during

my degree program.

I would also like to thank Searle Inc. and Boehringer lngelheim for providing

milacemide and mexiletine for use in these experirnents.

v TABLEOFCONTENTS

ABTRACT ...... ii

ACKNOWLEDGEMENTS ...... v

TABLE OF CONTENTS ...... vi

UST OF FIGURES ...... x

LIST OF TABLES ...... xii

UST OF ABBREVlATiONS ...... xiii i Introduction ...... ,.,...... 1 1.1 Statement of the Research Problem ...... 1

1.2 General Organisation of the Somatosensory System ...... 5

12 1 Primary Afferent Neurons ...... 5

1.2.2 Second Order Neurons ...... 8

1.2.3 Excitatory Neurotransmitters and Neuromodulators in Somatosensory Processing in the Spinal Cord ...... 9

1.2.4 Ascending Neural Trads Conveying Sençory Information. 1O

1. 3 Neuropathic Pain ...... 12

1.3.1 General Features of Neuropathic Pain ...... 12

1.3.2 Pathophysiology of Neuropathic Pain ...... 14

1.4 Allodynia ...... 16

1.4.1 Role of &fibres in Allodynia ...... 16 1.4.2 Rob of Spinal Glycinergic lntemeurons in Allodynia .... 18

1-4.4 Spinal Phamacology of lntrathecal Strychnine-induced Alloûynia ...... 21

1.5 Role of Excitatory Amino Acid Receptors in Sensory Processing . 24

1.6 Effects of Local Anaesthetics on Normal Nociception and Newopathic Pain ...... 27

1.7 lnhibitory Modulation in Sensory Processing ...... 29

1. 7.1 Rde of Spinal Adenosine ...... 29

1.7.2 Role of Spinal GABA ...... 32

1.7.3 Role of Spinal Glycine ...... 35

1.8 Rationale and Specifc Research Objectives ...... 36

2 MethOds ...... 40

2.1 Anirnals ...... 40

2.2 Implantation of Intra-thecal Catheters ...... 40

2.3 Awte Anaesthetized Animal Preparation ...... 41

2.4 Application of Stimuli ...... 42

2.5 Dmg Administration ...... 44

2.6 Expimental Protocols ...... 45

2.6.1 Control Experiment ...... 45

2.6.2 Quantification of Strychnineinduced Ailodynia ...... 45

vii 2.6.3 Doseiesponse analysis for intrathecally administered

dmgs ...... 47

2.6.4 DPCPX Experiments ...... 48

2.6.5 Mexiletine Expen'ments ...... 49

2.7 Protocols For Milammide ...... 49

2.7.2 Cornparison of the Effed of Milaœmide on Strychnine-allodynia and Mechanical Nociception ...... 50

2.7.3 Clorgyline and L-Deprenyl Experiments ...... 53

2.8 Data Analysis ...... 54

3.1 General Observations Following Intrathecal Strychnine ...... 56

3.2 Effect of Af-7 on Strychnine-allodynia ...... 59

3.3 Effect of Mexiletine on Strychnine-allodynia and Normal Nociception ...... 61

3.4 Effects of Adenosine Agonists on Strychnine-allodynia ...... 66

3.5 Effects of GABA Agonists on Strychnine-allodynia ...... 71

3.6 Effect of Milacemide on Strychnineallodynia and Normal Nociception ...... 74

3.6.1 Effect of Clorgyline and L-Deprenyl on the Action of Milacemide ...... 78

viii 4 Discussion ...... 81

4.2 Effed of AP-7 on Strychnine-allodynia: role of NMDA receptors . . 81

4.3 Effect of Mexiletine on Strychnine-allodynia and Nomal Nociception ...... 84

4.4 ER& of Adenosine Agonists on Strychnine-allodynia: role of A, receptors ...... 89

4.5 Effed of GABA Agonists on Strychnine-allodynia: role of GA64 receptors ...... 93

4.6 Effect of Milacemide on Strychnine-allodynia: role of glycine receptors ...... 96

4.7 Future Directions ...... 100

4.8 Summary...... IO1

5 References...... -104 LIST OF FIGURES

Figure 1 Possible neurotransmitters/neuromodulators and their receptor systems implicated in the spinal processing of low-threshdd afferent input...... 4

Figure 2 Cross section of spinal cord showing the termination of primary afferent fiben in dorsal grey matter ...... 7

Figure 3 Systemic milacemide readily crosses blood-brain-barrier where it is converted to glycine by the adion of monoamine oxidase-8...... -37

Figure 4 Summary of the strychnine control protocol and dose-response protocols...... 46

Figure 5 Summary of protocols for the milacernide experiments...... 52

Figure 6 Sample polygraph tracings of arterial blood pressure, heart rate and cortical EEG illustrating the effect of non-noxious hair deflection in intrathecal saline- and strychnine-treated rats. ... 58

Figure 7 lntrathecal AP-7 dose dependently suppresses intrathecal strychnine-allodynia...... 60

Figure 8 Sample polygraph tracings of arterial blood pressure, heart rate, and cortical EEG illustrating the effect of intrathecal AP-7 on strychnineallodynia and noxious mechanical hind paw-pinch in the rat ...... 61

Figure 9 Sample polygraph tracings of arterial blood pressure, heart rate, and cortical EEG illustrating the effect of intravenous mexiletine on i.t. strychnine-allodynia and noxious mechanical hind paw pinch in the rat...... 63

Figure 10 Intravenous mexiletine dosedependently suppresses intrathecal strychnine-allodynia...... 64

Figure 1I Intravenous mexiletine dosedependently suppresses paw-pinch induced responses in absence of strychnine...... 65 Figure 12 Intrathecal CPA dosedependent1 y suppresses intrathecal strychnine9llodynia...... 67

Figure 13 Inhibition of anti-allodynic effed of CPA or CGS by pretreatment with intrathecal DPCPX ...... 69

Figure 14 lntrattiecal CGS-21680 dose-dependently suppresses intrathecal strychnine-allodynia...... 70

Figure 15 lntrathecal muscimol dose-dependently suppresses intrathecal strychnine-allodynia...... 72

Figure 16 lntravenous milacemide exerts its maximum effed on inkathecal strychnine-allodynia two hours after injedion without affecting EEG syndirony...... -75

Figure 17 lntravenous milacemide suppresses intrathecal strychnine- allodynia in a dose- and time-dependent manner, without M8ding EEG synchrony...... 76

Figure 18 Milaœmide (600mgkg, i.v.) significantly inhibits intrathecal strychnine-allodynia but has no effed on normal mechanical -ception ...... 77

Figure 19 The suppression of strychnine-allodynia by intravenous milacemide is blocked by Ideprenyl, but not clorgyline...... 79

Figure 20 Schematic diagram illustrating the postulated role of excitatory amino acidreœptor systems and neuromodulatory systems in the spinal processing of low-threshold afferent input...... 85 UST OF TABLES

Table I Classification of fiben in peripheral nerves ...... 6

Table II Cornparison d the spinal phamacology of i.t strychnine- ailodynia and normal nociception (wiümut strychnine) ...... 22

Table III ED& and 95% C.I. of AP-7 ...... 59

Table iV ED* and G.I. of i.v. mexiletine ...... 66

Table V EDdand 95% C.I. of CPA or CGS ...... -68

Table VI EDdand 95% C.I. of muscimol ...... 71

Table VI1 Effed of baciden on HD-evoked responses in strychnine-treated rats ...... 73

Table Vlll ED=s and 95% C.I. of milacemide ...... 78

xii LIST OF ABBREVlAVONS

percent degree(s) Celsius alpha, a Greek letter beta, a Greek letter gamma, a Greek letter delta, a Greek letter CI mu, a Greek letter, used to indicate micro in metric units Ci9 micmgram(s), 1Od gmms, unit of mas PI miaditer(s), 1 litres, units of volume 5441 5-hydroxytrVptamim AP A-beta, a class of primacy afferent neurwis A6 Adelta, a class of primary afferent murons 4 adenosine receptor subtype 4 adenosine receptor subtype AMPA a-amino-3-hydroxy-S-methyl4isoxazolepropioic acid MOVA analysis of variance AP-5 D-aminophosphonovalerate AP-7 (1)-2-arnino-7-phosph~eptanoicacid A~P aspartate ATP adenosine triphosphate BBB blood brain barrier BP blood pressure bpm bats per minute C a dass of pn'rnary afferent neurons CGS-21680 2p-(2carboxyethyl)phenylamino-5'-N- ethylcarbxamidoadenosine hydrochloride CI confidence inteival CNQX wano-7-nitroquinoxaIine-2, 3-dione CNS central nervous system C P-96345 (2S,3S)us-2-(diphenyimethyl)-N-((2-methoxyphenyI)niethy 11-1 -azabicyclo[2.2.2]odan-3-amine CPA N6qcJopentyl adenosine CPP 3-(2carboxypiperazin4yl)propyl-1 -phosphonoic acid CSF oerebrospinal fîuid DGG y-Dglutam y lglycine DMSO dimethyl sulfoxide DPCPX 8cyclopentyl-1 ,3-dipropylxanthine DRG Dorsal root ganglion

xiii EAA excitatory arnino acid(s) EDSo Medive dose yieldirig a 50 percent resQonse EEG 0-g- EPSP GABA y-cacid Glu a-e Gs bümuiotoiy Ggrot8Hi Gi inhibitoryGpmtein HCI hydochkricrad HD bir deflediori Ha elementalsymbdformcwawy HR heart rate Hz herb, SI, r8ci-mlseconds, unit dfrequency lASP kitemaümal Association for tfte study of Pain i.e. Id est, thet is H,m Ki glasswaro hm0 in the living body i.p. intrapefitoneai(ly) IPSP inhibitory ptsynaptic potential(s) i.t. intrathecal(ly) i.v. intravenous(1y) I Roman numeral one, used to denote spinal lamina II Roman numeral two, useâ to denote spinal lamina 111 Roman numeral three, wed to denote spinal lamina IV Roman numeral fow, used to denote spinal lamina )(a KA hinate Imonoamine oxidase A type enzyme MAO-B morioamine oxidase 6 type enzyme MAP mean artenal pressure MK-801 (+)SmethyCl O, 11 dihydro-5H-dibenzo (a,û) cyclo heptene5.1Wmine maleate nwnber of determinations

xiv NECA NK neumnin NMDA N-methyl-Paspartate nmol nanomole(s), 10' moles, nunber of molecules P probability a mor PE-1 O polyethyiene tubing (diameter -0.61 mm) PP pawginch RPlA ~(-)-N'<2-pbnyl isopropyl)adenosine S.C. =b-neoust~v) SEM standard enor af the mean SP substance P STR strychnine SYN synchrony f time WDR wide dynamic range 1 IiWoduction

1.1 Statemnt of the Reseatch Pmblem

Injury to the peripheral or central nervous system can effed an abnomal sensory *te in which innaaious stimuli acquire the ability to evoke intense pain; a conditkm 19wimi as aliodynia. Clinically, allodynic pain is triggered by input from low threshdd mechanoreceptive pn'mary afferent neurons (Meyer et al., 1985;

Campbell et al., 1988; Priœ et al., 1989; Treede et al., 1991); Viat is, by peripheral (w)SubSfrates whidi mllyhave no roi8 in pain transduction and transmission.

Akhough the pathophysiological mechanisms of allodynia are unknown, the striking change in modality subserved by Akfibers after nerve injury, and the apparently limited role of abnomially sensitizeâ C-œptors in this condition (Cline et al.,

1989) are highly suggestive of a central dysfundion.

Under romial cociditions, the synaptic transmission of low threshold afferent input appears to be modulated by glycine and GABA (yaminobutyric acid) releafed from intemeurwis in the dorsal hom of the spinal cord (Curîis et al., 1968, 1971;

Game and Lodge, 1975; Wilcockson et al., 1984; Todd, 1990; Todd and Sullivan,

1990). The loss of such inhibitory modulation, as might ocurr after nerve injury, could enhance the synaptic efkacy between Ap-fibers and pain-signalling pathways, thereby permitting the miscoding of low-threshold input as pain. 2

Consistent with this central hypothesis of allodynia are reports that the intrathecal

(i.t)injedion of strychnine (Sm;a cornpetitive glycine reœptor antagonist) into the

spinal subarachnoid space of rodents induces a pure allodynic state. Thus

imoaxxis hair defiedion (HD) in the presence (but not absence) of i.t. STR evokes

behevioral. cardiovasailar, neurochemical and electrophysiological responses

comparable to those elicited by noxious thermal, mechanical or chernical

stimulation without STR (Beyer et al., 1985, 1988; Sosnowski and Yaksh, 1989;

Yak&, 1989; Sherman and Loomis, 1994,1995,1996; Miine et al., 1996; Sherman

et al., 1996; Onaka et al., 1996). Irnportantly, these tactileevoked responses are

segmentally localized (Sherman and Loomis, 1995) and exhibit a spinal

pharmawlogy that is distinct from that of conventional nociception (Yaksh, 1989;

Sosnowski and Yaksh, 1989; Sherman and Loomis, 1994, 1995,1996). Indeed,

glycine appears to be a selective modulator of non-nociceptive input in the rat

spinal cord as i-t. STR neither enhances nor inhibits responses to high threshold

cutaneous stimulation (Sherman and Loornis, 1996). Collectively these data

indicate that the acute blodtade of spinal glycine receptors with i.t. STR yields an abnormal sensory condition having many of the features of clinical allodynia.

Previous phamacological studies have shown that STR-allodynia is dose- dependently inhibited by i.t excitatory amino acid antagonists including NBQX (2,3- dihydroxy4-nitro-7-sulpharnoyl-benzo(f) quinoxaline) and y-DGG (y-0- glutamylglycine) and by i.t glycine or betaine (N,N,N-trimethyl glycine) (Sherman 3 and Loomis, 1995. 1996) (see Table II. page 22). In mtrast. STR-allodynia D insensitive to i.t morphine, CP-96345 [neurokinin(NK), -recaptor antagonist] w neonatal treatment with capsaiun, at doses which am Maive against normpl nociception (Shemian aird Loomis, 1994, 1996). These rewlts indicate that the spinal phamiacdogy of STR4odynia b distinct from the nociception nomially conveyed by pmall diameter (A6 and C) fibers and evaked by higMhreshdd thermal, chernical or mechanical stimuli (for review see Yaksh end Malmberg,

1994).

The spinal dorsal hom contains the first synaptic contacts between peripheral afferent fibers and Vie central neurons conveying somatosensory information to the critical sites in the kain. lnsight into the spinal processing of somatosensory input couid reveal useful information about the mechanism(s) of allodynia as weil as identm modulatory systems that could be exploitecl phamiaadogically to offset the abnomial sensations triggered by light tactile input in aiiodynia. The neural systems inpiicated in the spinal proc~ssingof *input are illustrateci in Figure 1. To fwther investigate the spinal phannacology of STR- allodynia, and to explore the neufornodulatory systems afbcting low threshoid

SOmafosmsory input in the spinal cord, the present study sought to systematically detemine the efFed of GABA, adenosine and glycine in the STR-model. Also, the effed of mexiletine, Af-7 [D(-)-2-ami~?-phosphonohepb1~)icacid; an NMDA- receptor antagonist], and milaoemide (a glycine prodnig) on STR-ailodynia and Fbum L Possible neuiotnrnsrnitfeIS/neummoduIaton and their receptor systems implkatd in the spinal pmcessïng of Ibwüimshold (A4 aibmnt - input Glutamate (Glu), aspartate (Asp) and ATP are thought to be released from the cantral tminals ofefibers. Nmally, extraœllular glutamate and aspartate ad on AMPA- and kainatweceptors. NMDA-I~c~P~w,although present on second order neurons, are not nomially reauited by lw-threshold input (but may be important in abnomal Iow-ttirBShdd inprrt). Local inhibitory systems rnodulating the response ta Apinput utÏlue GAM, glyàne and adenosine. The latter, including that den'veâ from the extracellular hydrolysis of AT?, cm ad on A,- andlor 4 - recepZm. The postsynapüc Weds of these neuotransmitten are well established. The presynaptic eflects of glycine and adenosin8 are less dear and have been omitted for clarity. AIViough, in the present diagram, adenosine is shown to be released from intemeurons, its neuronal origin is unciear. 5 malnocic8ptim (without STR) was compared.

f.2 Geneni Otganisafh of UN Somafasentory

Smsuy systm cmsist of serial neual patmvays linking the pwiphery with the spinal cord, minstem, thalamus and cerebral cortex. These pathways an, comprised of prirnary afferent fibers, second order nemsand the Viird order neuroris.

1.2.1 Pn'mary Affemnt Neunont

Primary Mwent wonscarry SOmafminformafiori fmthe periphery to the spinal cord where they synapse with second or* neufans located in Ihe dorsal grey matter. Thus, the first synapse in this neural circuitry is located centrally in the outer laminae d the dorsal hom. Primary afFerent fibers found in mixed pen'pheral mesare -vided into themain types;A, B and C. A-fibers are further divided into four subgroups (a,P, y and 6)based on their diameter, degree of myelination and correspondhg condudion velocity (Table 1). These fibers seledively respond to dîfFerent modalities in accordance with the types of sensory receptors lacateâ on their pefipheral teminals. Thus, A6-fiben are activated primrily by intense medianical pressure but can also be activated by noxious thermal or chernical stimuli (Kandel and Schwae 1991). In contrast, C- Table I. CfassHication of fiben in periphemI nemes

Fiber Innervation Mean Diarneter Mean Conductian Velocity Ty~e Pm (range) mlsec (range)

Aa Primary Muscle Spindle 12 (12-20) 100 (70-120) To Skeletal Muscles

p Cutaneous Touch and 8 (5-15) 5û (30-70) Pressure Anerents

y Motor to Muscle Spindle 6 (68) 20 (15-30)

6 Mechanoreceptors, Nociceptors 3 (1-4) 15 (1 2-30)

B Sympathetic Preganglionic 3 (13) 7 (3-1 5)

C Mechanoreceptors, Nociceptors 1 (0.5-1 -5) 1 (0.5-2) Sympathetic Postganglionic

fibers are primarily adivated by noxious chernical, mechanical and thermal stimuli.

Psychophysical studies in humans demonstrate that electrical stimulation of A& fibers evoke a brief pricking or stinging sensation mile that evoked by C-fiber stimulation is described as a prolonged buming and unpleasant (Torebjork and

Odioa, 1980; Konietmy et al., 1981). Stimulation of Ag-fibers normally evokes light tactile sensations including vibration and mild pressure. These primary afferent fibers enter the spinal cord through the dorsal root entry zone. Large diameter fibers bifurcate centrally with one branch entering the dorsal columns and the second branch penetrating deep into the ne& of the dorsal hom where they reverse course and teminate in the nucleus proprius (Figure 2). Small diameter Dorsal Column

Afferent axons for touh, position and vibrations

Afferent axons for pain

Flgurs 2. CmssecUon of spinal cord showhg the tedation of ptfmaty .fikrant fi- In domai gmy matter. Small diameter fiben teminate diredly in the outer laminae of the dorsal hom where they make synaptic contact with the dmlrites a( seamd order neu#is (1). Large diameter fibers bifurcate centrally with one kandi perietrating deep into the ne& d the dond hom (II) -8 they reverse couse and tenninate in the nucieus proprius, and aie second branch entering the dorsal colmns (III). 8 fibers terminate directly in the outer laminae of the dorsal horn where they make synaptic contact with the dendrites of second order spimthalamic tract neurons

(Figure 2) (Kandel and Schwartz, 1991).

1.2.2 Second Order Neumns

The primary afferent fibers fom the synaptic contacts with second order neurons in the dorsal hom. These are either relay cells, with axons projecting to the brain stem or thalamus, or intemewons that transfer information locally to other interneurons or relay cells in the adjacent segments. Second order neurons are dassified into three general types (Besson and Chaouch, 1987; Henry, 1989): 1) noniiociceptive (receiving input from large diameter myelinated fibers), 2) nociceptive sp-c (those receiving input from C and Ai5 fibers), and 3) nociceptive non-specific (those receiving convergent input from both nociceptive and non- nociceptive primary afferent fibers). The latter are also known as wide dynamic range (WDR) neurons. WDR neurons show a brisk transient response to low- threshold mechanical stimuli and a delayed after-discharge in response to noxious stimuli. the latter representing a slow depolarization of the neuron (Henry, 1989).

Virtually al1 sensory information arising from the somatic segment of Vie body enters the spinal cord Virough the dorsal rwt. 1.2.3 ExcHatwy Neum~nsmiftemand NeummoduIators in Somafosensoty

Pmcassing In the Spin./ Cod

A variety of substances pfesent in primafy sensory neurons have been proposed as possible neurotransmitters or neufornadulators. Putative Merent neurotransmitten incluâe: substance P (SP), neurotensin. cholecystokinin, vaSOactiv8 intestinal pepüde, calcitonin gene related peptide (CGRP), somatostatin, glutamate, amteand adenosine triphosphate (ATP) (Besson and Chaouch,

1987; Henry, 1989; Yaksh and Malmberg, 1994). The precise role of many of these substances has not ben fully elucidated. Currently, glutamate and aspartate are believed to be the main excitatory neurotransmitters released from al1 ptirnary

Met-ent fibers (Kangrga and Randic, 1991 ; Dougherty et al., 1992). In nockeptive fibers, these excitatory arnino acids (EAAs) co-exist with neuropeptides including

CGRP and substance P (Battaglia and Rustioni, 1988). ATP has been implicated in the transmission of information from large diameter primary afferent fiben. Its involvement in nociception is also suggested by the depressant effeds of adenosine (the produd of its hydrolysis) on nociceptive neurons (Henry, 1989).

Synapses of primary afferent and second order neurons are subject to extensive modulation by a number of endogenous systems (Yaksh and Malmberg,

1994). Local inhibition is rnediated by glycine, adenosine and GABA (Nagy and

Daddona, 1985; Sawynok and Sweeney, 1989; Yaksh and Malmberg, 1994) (see sections 1.4.2, 1.7.1 and 1.7.2). Endogenous opioids, such as enkephalin, are 10 present in spinal intemeurons and mediate the local inhibition of nociceptive input

(Yaksh and Malmberg, 1994). Nociceptive transmission is also subjed to descending inhibitory modulation originating from sites in the pons (e.g. locus coeruleus) and medulla (i.0. nucleus raphe magnus).

1.2.4 Ascending Neurel Tmcts Convepfg Sensory In fornation

There are two main pathways conveying sensory information from the dorsal hom to the bain: 1) the dorsal column-leminiscal system; and 2) the anterolateral system. The dorsal column-leminiscal system conveys touch and vibration sense and is generally diaracterizeâ by a well defined localkation of the stimulus source.

The anterolateral system is responsible for wnveying pain, thermal sensations including cold and wamth, crude touch and pressure sensations, tickle and itch

(Kandel and Schwartz, 1991).

Dorsal miumn medial leminiscal system: The dorsal column pathway reiays somatosensory information directly to the medulla. It is compriseci of ascending axons of the large diameter pfimary afferent fibers and axons of second order neurons originating in the dorsal hom. These axons ascend ipsilateraly to the caudal medulla where they synapse onto the cells in the dorsal column nuclei

(aineate and gracile nuclei). Fiben leaving the dorsal column nuclei arch across the midline where they collect into a discrete bundle (the medial lemniscus) and ascend to the thalamus. Rom the thalamus, the somatosensory infomation is 11 conveyed to the anterior parietal lobe by thalamocortical neurons.

Anterohteral system: This systern is adually composed of two major pathways distinguished by their sites of temination: the spinothalamic tract and the spinoreticular tract. The spinothalamic tract originates principally from cells in lamina 1. This trad mediates fast pain, relayeâ from the periphery to the spinal cord by Mben. In contrast, the spinoreticular tract pn'ncipally originates from cells in lamina V. It transmits both nociceptive and nondœptive (AP-, A6- and C-fibers) infmtionand hence cells of this tract are referred to as WDR nociceptors. This trad is important in slow pain. Axons of spinoreticular trad cells end on neurons in the retiwlar formation of the brain stem, which then relay information rostrally to the thalamus and other structures in the diencephalon. Rie spinoreticular tract also relays information to the mesencephalic periaquedudal gray, Vie region sumnding the cerebral aqueduct.

It is evident fTM this bhef description that the processing of somatosensory information is cornplex, involving several parallel but discrete systems which extend throughout the neuraxis. These systems utilize multiple neurotransmitters and several levels of regulation (spinal, bulbar, thalarnic and cortical). Thus, under normal conditions, nociceptive and non-nociceptive sensory information is transmitted through distinct neural circuits, thereby ensuring appropriate interpretation of the sensory stimulus. 12 f.3 Neuropathic Pain

1.3.1 General Features of Neuropathic P.in

Aaite pain arising from tissue injury has an immediate onset, a short duration, temporally coupled to stimulus expsure and is intimately related to changes induœd by perïpheral inflammation. Physiologically, it is thought to serve an important protedive function. In contrast, chronic pain usually has delayed onset and persists long after the time mllyrequired for tissue repair. Intuitively, chronic pain involves more complex events than those of normal nociception and has no apparent biological wefulness. Along with tissue &mage and inflammation, chronic pain is &en adated wïth neu~~paWogy.This indudes changes in the

CNS leading to an increased sensitivity of central neurons; a condition known as central sensitization (Dray et al., 1994).

Neuropathic pain is a chronic condition resulting from traumatic injury to or pathology of the central or peripheral mous system. Phantom limb pain, multiple sclerosis, central post stroke pain, neuropathies such as those of diabetic, alcoholic, nutritional or traumatic origin, post herpetic neuralgia, trigeminal neuralgia, and reflex sympathetic dystrophy are some examples of neuropathic pain

(Bovie et al., 1989; Tasker, 1990; Portenoy et al., 1990; Triggs and Beric, 1992;

Woolf, 1993). it is charaderized by allodynia, hyperalgesia and spontaneous pain

(Price et al., 1989, 1992; Bonica, 1990; Coderre et al., 1993; Galer, 1995).

Spontaneous pain is pain felt in the absence of any apparent physiological 13 stimulus. Hyperalgesia is an exaggerated response to a nomaliy painful stimulus and allodynia is a condition in which pain is evoked by a normally non-painful stimuli (e.g. touch, vibration). Additional syrnptoms may include a partial loss of sensitivity, and the spread of pain to uninjured tissue (referred pain) (Price et al., l989,lQ92;Codem et al., 1993; Wooif, 1993; Rowbtham, 1995).

Unlike regular pain, neuropathic pain &en takes weeks or months to develop following the musative event (Tasker, 1990). For instance, in a group of

127 patients with spinal cord lesions, the onset of pain ranged from a month to mare Vian a year in 79% of the patients (Tasker et al., 1992). Neuropathic pain is commonly described by sufferers as a buming, Msting, ripping, tearing or sharp lancinating type of pain (Price et al., 1992; Wooif, 1993; Gonzales, 1995) and is often unbearably intense. It is diff~cultto treat and syrnptoms may last indefinitely

(Merskey, 1986). Conventional analgesics like opioids and adjuvant dnig therapy such as tricyclic antidepressants, barbiturates or anticonvulsants are only partially or temporarily effedive (Bowsher, 1991 ; Triggs and WC,1992; Baron and Saguer,

1993).

The use of opioid analgesics in neuropathic pain is highly controversial

(Anier and Mayerson, 1988; Portenoy et al., 1990; Rowbotham et al., 1991; Kupers et ai., 1991; Jadad et al., 1992). For example, neuropathic pain following amputation or nerve injury has been shown to be unresponsive to opioids (Amer and Mayerson, 1988). In contrast, patients with malignant neuropathic pain have 14 &mm a shift to the right in the opioid dose-response arrve such that an increase

in the dose will provide effective pain relief (Portenoy et al., 1990). For patients with a demonstrable sympathetic mponent (Le. syrnpathetically maintained pain),

interruption of the sympathetic moussystem with local anaesthetics (Price et al.,

lm),sympatholytic dmgs or by surgery is the major treatment approach (Woolf,

1993).

Surgical intervention in the periphery is also oontroversial. Pain relief is

usually temporary, often ineffective and may, in some cases worsen the situation

(White and Sweet, 1969; Wirth and Rutherford, 1970; Noordenbos and Wall, 1981).

Neurolytic procedures or electrical stimulation of the dorsal colurnn or the brain,

produce partial relief for intractable pain of spinal cord origin (Tasker et al., 1992).

However, these procedures are appropriate only for patients with a severe disabling pain that is refractory to al1 other treatments. Pain of this type is often associated matha terminal disease and lirnited Iife expedancy (Woolf, 1993).

f.3.2 Pathophysiology of Neumpafhic Pain

Peripheral nerve injury often leads to pathological changes including neural degeneration, neuroma formation, and spontaneous activity in damaged nerve fibers (Coderre et al., 1993). Structural changes within spinal cord, such as the loss of dorsal hom intemeurons (Kapadia and La Motte, 1987; Sugimoto et al.,

1989), some of which contain GABA and glycine (Todd and Sullivan, 1990), and the 15 ekiomial rearrangement of affmtnenre processes have also been report& The latter indude central -ng of large afferent fiben fmlamina III to lamina II of the dorsal hom, and sprouüng of sympathetic nerves into the dorsal root ganglia

(Woolf et al., 1992; Mclachlan et al., 1993). Functionally, these changes could result in either a redudion in the threshold of activation of nocicepton andlor an imsein the excitability of the central neunxis cornprising the pain-signalling pathways (e.g. 'central sensitriafbf). Central sensitization is charaderized by increased spontaneous adivity, decreased threshold of activation or inueased reçponsiveness to afFerent input, prolonged afterdischarge to repeated stimulation, and an expansion of the receptive field of dorsal hom neurons. These changes would explain, at kast in part, the phenornena of hyperalgesia, allodynia and spontaneous pain (Dubner, 1991 ; Coderre et al., 1993; Woolf, 1993).

The mechanisms underlying the development and maintenance of neuropathic pain are unclear. However, the pefipheral substrates triggering allodynia and hyperalgesia are known to be different. Hyperalgesia arises from a decrease in the threshold of activation of nociceptive C-fibers (Shir and Seltzer,

1990; Cline et al.. 1989; Lee et al., 1994). In contrast, allodynia is triggered by input from low threshold APfibers (Meyer et al., 1985; Campbell et al., 1388; Price et al., 1989; Treede et al., 1991); that is, by peripheral substrates that nomally have no role in pain transduction and transmission. The striking changes in the modality subserved by APfibers after nerve injwy are highly suggestive of a central dysfundion.

1.4 Aklodynia t4.1 Role of Apfibers in AIIodynie

Allodynia, a characteristic feature of neuropathic pain, is an abnormal amdition in which light tactile stimuli (i.e. touch or the brush of clothing against the skin) aquire the ability to evoke intense pain. As noted above, allodynic pain is triggered by input from AP-fibers, rather than abnormally sensitized C-fibers

(Campbell et al., 1988). Thus selective newe block using ischemia or nerve compression eliminated both aflodynia in the area of nerve injury along with light touch sensation in adjacent normal skin without affecting the temperature discrimination in patients with neuropathic pain suggesting that both sensations were mediated by AP-fibers (Campbell et al., 1988; Koltzenburg et al., 1992).

Conversely, nerve block with local anesthetics, that affeded A& and C-fibers (as demonstrated by absence of temperature sensibility in those patients), but not AB- fibers did not blodc allodynia (Campbell et al., 1988). In addition, response latendes for taich evoked-pain on the affeded side were nearly identical to those for the same stimulus applied to the homologous area on the normal side. 60th latendes were much faster than the latendes for the detection of pain in the normal limb (Lindblom and Verrillo, 1979; Campbell et al., 1988). These results indicate 17 that allodynic pain is evoked by ~~ input axveyed by low threshold &fibers. In support of this coriclusion, high-frequemy, low intensity electrical nenie stimulation worsens allodynia, unlike the analgesic Med observed with regular nocioeptive pain (Price et al.. 1992). Finally the selective destruction of C-fibers in rats using neonatal treatment with capsaicin did not prevent the indudion of -. alkdynia fdlowing chrwc constndKwr of the sciatic me(Shir and Seltzer, 1990).

Rather, this treatment blocked the development of thermal hyperalgesia. Overall, these data indicate Viat &fibers mediate alldoynia whereas, C- and A5-fibers mediate thermal hyperalgesia.

Since under normal conditions, the adivity in AB-fibers is perceived exdusively as touch, ligM pressure and vibration, the fundamental question arising from the phenornenon of allodynia is how this same neural activity acquires ability to evoke severe pain. One possibility is that inhibitory intemeurons nomally modulating the response of spinal neurons to low-threshold afferent input may become dysfunctional Mer injury. Indeed, intemeurons containing GABA and glycine are known to be vulnerable to ischemic and excitotoxic damage (Tureen,

1936; Davidoff et al.. 1967; Sugimoto et al., 1989, 1990). The functional loss of sudi inhibition could result in an excessive activation of second order neurons by low-threshold afferent input thereby strengthening the synaptic contact between non-ndceptive fibers and pain-signalling pathways. Thus, Ag-input could be miscoded as pain. 18 f.4.2 Role of Spinal Glycinergic lnterneurons in Allodynia

GABA and glycine are important in the spinal processing of innocuous tadile information (Yaksh, 1989, 1993). Glycine-like immunoreadivity (-LI), as demonstrated using glycine antisera, is found throughout the gray matter in the spinal cord of the rat (Van den Pol and Grocs, 1988; Todd and Sullivan, 1990). In the dorsal hom, glycine-LI is present in laminae I-N, but is most densely localized in lamina III and N (lodd and Sullivan, 1990). These results are consistent with the presence of glycine receptw-ll at axodendritic and axo-c synapses in laminae

1-111 and at dendrodendritic synapses in lamina II (Mitchell et al., 1993). Thus, binding and irnrnunocyîochemical data indicate that glycine and its corresponding receptors are localized in regions of the dorsal hom containing the central terminations of low-threshold afferent fibers (laminae Ill-IV). Furthemore, glycinergic neurons in laminae II and III of the rat spinal wrd have been shown to receive a major monosynaptic input from myelinated low threshold wtaneous primary afferent fibers (Todd, 1990). Colledively, aiment anatornical evidence supports the hypothesis that &fibers activate local glycinergic intemeurons in the spinal dorsal hom.

Pharmacological data also substantiate the relevance of glycine to the regulatiwi of the behaviow generated by low-threshold afferent transmission. Light tadile stimuli, typically eliciting no more than an orientation response in rats, evoke overt pain-like behaviours following the blockade of spinal glycine receptors with 19 i.t. STR (Yaksh, 1989). STR exhibits very high affinity ( Ka = 3-1 0 nM) for the glycine receptor (Young and Synder, 1974; Zarbin et al., 1981). This is distinct hwn the glycine modulatory site located cm the NMDA-receptor cornplex. The latter is insensitive to STR blockade. The data from STR-treated rats are in agreement with studies of single unit adivity in the trigeminal nucleus of the monkey.

lntravenous (i-v.) STR induœd a rnarked facilitation of the response of WDR msto dedical stnnulation of low-threshold medianoreceptive afferent fibers

(Yokota et al., 1979). lontophoretic delivery of STR in the lumbar spinal cord of cats diminished the inhibition phase otherwise observed following mtural and electrical stimulation of low-threshold afferent fibers (Game and Lodge, 1975).

Conversely, the iontophoretic delivery of glycine in the spinal dorsal hom diminished the responsiveness of dorsal hom neurons to light tactile stimulation

(light toucNlight pressure), decreased the size of their cutaneous receptive fields

in cats (Zieglgansberger and Hertz. 1971), and reduced the responses of s~inothalamictract neurons, induding low-threshold and WDR neurons, in monkeys

(VVilcockson et al.. 1984).

These observations suggest that the encoding of low-threshold mechanical stimuli as an innocuous event depends upon the presence of intnnsic glycinergic inhibition in the spinal dorsal hom. Indeed, genetic variants, such as the Poll

Hereford calf (Gundlach et al., 1988) and the spastic mouse (White and Heller,

1982) which exhibit up to a IO-fold decrease in STR binding in the spinal cord, 20 display exaggerated sensitivity to even rnodest cutaneous stimulation. Consistent with these animal data are dinical reports of pronounced hypersensitivity to light touch in humans during STR intoxication (Arena, 1979). Thus, glycinergic dysfunction in the spinal dorsal hom wuld be an important central abnomality in the development of allodynia.

1.4.3 Strychnineinduceci Allodynk

Further support for the role of spinal glycine in modulating low-threshold input is provided by the obsenration that phamiacological Hockade of spinal glycine receptors with i.t STR induces a pure allodynia like state in the rat (Beyer et al.,

1985, 1988; Yaksh, 1989). In the presence of STR. innocuous mechanical stimulation such as HD evoked high pitched vocalisation, biting and scratching of the stimulated dematomes and vigorous organised escape behaviour (Yaksh,

1989). These evoked responses resemble the hyperaesthesia reported by hurnans during the preconvulsive stages of STR intoxication (Arena, 1979). A striking feature of these responses is that they wuld be reliably driven by HD applied to circumsaibed areas of the body surface. These corresponded to the dematomes innervated by spinal segments near the site of i.t STR injection. Similar results were obtained using lightly anaesthetized rats pretreated with i.t. STR.

Exaggerated autonomie and motor withdrawal responses to repeated HD, comparable to those evoked by noxious stimuli in the absence of STR, were 21 reporteci (Yaksh, 1989; Sherman and Loomis, 1994). Ail HDevoked responses were obsaweâ without convulsions or seizure-like Mects and thus did not refled a general sensitization of spinal murons by i.t STR Subsequent studies in our laboratary have shown that i.t STR neither enhances nor inhibits responses to noxbus bilateml hind pawginbr or tail immersion in water at 5S°C. These resub indiieth* a) glycine is a seledive modulator of non-œptive input in the rat spinal cord; b) the blockade of spinal glycinergic receptors with il. STR yields an abnomal sensory condition resembling allodynia; and c) i.t. STR has neither a hyperalgesic nor an antinociœptive effect on normal nociception.

1.4.4 Spinal Phamacology of lntrethecaI Sîrychnineinduced Alfodynia

Clinical studies have shawn that allodynic pain is triggered by input from low threshold (nonilociceptive) fibers. Therefore, any valid experimental model of allodynia should exhiba a pharmacology that is distinct from that of normal (C-fiber mediatecl) nociception. Previais studies of the STR mode1 support this distinction

(Table II). For example, dironic treatment of neonatal rats with subcutaneous

(s.c.) capsaian at doses that seledively destroyed C-fiben by adulthood and yielded significant antinociception in th8 paw-pinch (mechanical), tail immersion

(themal) and topical xylene (chernical) tests, had no effed on STRallodynia

(Shemianand Lmis, 1996). Similarly, pain-like responses evoked by HD in STR- treated rats were insensitive to antim'œptive doses of i.t morphine and CP-96345 Table II. Cornpanson of the spinal phatmacology of if. strychninwModynia end nonnal nocicepfion (without strychnine)

Strychnine- Noxious allodynia stimulation

Agents affeding C-fibers Morphine Shemian & Loomk, 1994 Capsaicin manb Loomis, 1996 CP-96345 ~dobservati~ns (Nh-antagonist)

Glycine or analogue Betaine Sherman & Loomis, 1995 Glycine Sherman & Loomî, 1995

Fucitafoty amino acids y-DGG Sherman b Loomis, 19% Nlsstr6m et al., 1992 NBQX Sherman & Loomis, 19% Nassb6m et al., 1992

t All dwgs were administered as a single i.t. injection except capsaicin whicb was given as repeated s.c. injections during the neonatal period.

a The non-NMDA receptor antagonist CNQX was used instead of NBQX by Nasstrom et al., 1992.

(a NK,-receptor antagonist) (Shemian and Loomis, 1994; unpublished observation).

In contrast, i.t. glycine and betaine selectively inhibited STRallodynia, although their effed on normal nociception was not systematically detemined (Beyer et al.,

1985, Sherman and Loomis, 1995). These results demonstrate that agents which 23 selectively interfere wÏth the fundion of C-fibers do not attenuate STR-alloâynia in

the rat.

Glutamate and aspartate are the phcipal excitatory neurotransmitters in aie

CNS (Mayer and Westbrook, 1987; Evans, 1989; Perrchak and Cuenod, WW),

and are released from central terminais of both low- and high-threshold afferent

fiben of rats and monkeys (Skilling et al., 1988; Liu et al., 1989; Kangrga and

Randic, 1991; Sorkin et al., 1992). Indeed, glutamate-LI and aspartate-LI are

localized in the dorsal root ganglion (DRG) -11s of both large and smalldiarneter

primary afferent fiben (Wanaka et al., 1987; Battalgia and Rustioni, 1988;

Westlund et al., 1989, 1990). as well as in intemeurons of the dorsal horn of cats,

rats and monkeys (Riuoli, 1968; Fagg and Foster, 1983; Cottman et al., 1987;

StmMathisen and Ottersen, 1987). The intraspinal boutons of cat afferent hair

follide fibers, labelled intra-axonally with horseradish peroxidase, are also enndied with Lglutamate-LI (Maxwell et al., 1993). Consistent with the localization and

release of EAAs from both low- and high-thr8ShoId derent fibers, and their putative

role in somatosensory transmission, EAA receptw antagonists have been show to block the excitation of spinothalamic tract neurons by noxious thermal, mechanical, chernical and electncal stimuli in monkeys (Dougherty et al., 1992), and to vntlyinhibit STRallodynia in lightly anesthetùed and çonscious rats (Table II; Yaksh, 1989, Sherman and Loomis, 1994, 1996). Considering the selective effed of i.t. STR on low-threshold afferent input in the spinal cord 24 (Sherman and Loornis, 1996), and the marked sensitivity of this input to EAA

antagonists (Table II), it is possible that the blod

modulation by i.t STR pmits the exaggerated depolarkation of second order

neuons in pain-SigMlling pathways by bwaweshold (M)input, and the miscoding

of this input as pain.

f.5 Rok of Fycitatoty Amlno Acid Recepfom in Sensory Pmcessing

The biological actions of EAAs are mediated by at least three subtypes of

ionotropic receptm. l'bse receptor subtypes were identified on the basis of their

~8lecüveresponsiveness to the follwing agonists; a-amino-Mydroxy-S-methyl4

isoxazoI8-p~opimicacid (AMPA), kainate (KA) and N-methyl-D-aspartate (NMDA)

(Watkings et al., IWO). Colledively, the AMPA and kainate receptor subtypes are

referred to as non-NMDA receptors. While glutamate adivates both NMDA and

nonHMDA ~~OB~~OTS,aspartate is thought to ad preferentially at NMDAieceptors

(Mayer and Westkook, 1987; Dickenson, 1990). Non-NMDA receptoraupled ion channels mediate the fast component of the excitatory post synaptic potential

(EPSP). Under physiological conditions, NMDA receptorcoupled ion channels are normally blod

(Headly and Griller, 1990; Watkins et al., 1990).

Intracellular saidies using a hemiseded rat spinal cord-hind limb preparation demorisbated that the amplitude and duraüon d the EPSPs elicited by low-intensity

(touch; Ag) or high-intensity (pinch; A6 and C) mechanical stimuli are blockeâ by non-NMDA receptor antagonist, CNQX and the cell fiting of Vie dorsal hom eells was completely abolished (King and LopezGarcia, 1993). In contrast, the NMDA receptor-antagonist, AP-5 (O-aminophosphonovalerate), blocked only the longer latency spikes elicited by high-intensity cutaneous mechanical stimuli. It never abolished the œll firing produced by such stimuli (King and Lopez-Garcia, 1993) and was ineffective against responses evoked by light touch (King and Lopez-

Garcia, 1993). Thus, EAAs released fmAgfiben seledively activate non-NMDA receptors under physiological conditions. In this regard i.t. y-DGG (non-seledive

EAA-antagonist), and NBQX (a seledive AMPA-antagonist) have been shown to blo& STRallodynia in 1ightly anaesthetized rats (Sherman and Loornis, 1994,

1996); observations consistent with the mode1 of synaptic transmission illustrated in Figure 1.

The role of spinal NMDA receptors in normal nociceptive transmission has been extensively investigated. If a brief noxious stimulus, svnicient to activate C- fibers. is repeateâ at frequency 20.3 Hz, the long latency discharge recorded in second order neurons ~ro~fessivelyincreases in magnitude; a phenornenon called 26 'Wnd up" (MendeIl, 1966; Price, 1978). Wnd up' is a characteristic feature of C- fiber evoked nociception and is meâiated by €AAs acting at post-synaptic NMDA- reCBptocs (Didlevorphanol have been shown tu block the slow temporal summation of EPSPs evoked by elecWical stimulation of C- mfibers (%nd up') without reducing the response of these same neurons to

A-fi ber stimulation (Davies and Lodge, 1987; Dickenson and Sullivan, 1987;

Did

This phemenon is believed to underlie the development of central sensitization during sustained nociceptive input (Le. inflammation). Thus, Wnd up' is a physiological process mediated by NMDA receptors and is nonally restricted to noxious (C-fiber) afferent input.

it has recently ben suggested that uder pathophysiological conditions, AB afferent impulses may acquire access to the same 'wind up' rnechanism, thereby allowing low-threshold fibers to Wgger pain (Price el al., 1989; 1994a). Unlike nomial subjects, patients with nerve injury exhibit temporal summation to repeated adivaüon of AB afferents at frequencies 20.3 Hz. using either natural stimuli such as stroking the skin with a gauze pad or low-threshold elearical stimulation (Price et al., 1989). Subjects reported that the pain evoked by these normally innocuous 27 stimuli had the same buming quality and tendency to radiate as noxious heat evoked pain. Moreover, subanesthetic doses of ketarnine (an NMDAieceptor antagonist) attenuated the pain evoked by low-threshold mechanical stimulation in patients with traumatic spinal card injury, post-herpetic neuralgia, post-stroke pain and peripheral nerve injury (Backonja et al., 1994; Eide et al., 1994, 1995).

Collectively, these observations suggest that a process similar to Wnd up' is activated in allodynia and that this process is sensitive to NMDAieceptor blockade.

If, in the presence of i-tSTR, low threçhold (y)input acquires access to the same sensitization (Wnd up') mechanism nomally activated by C-fiber input, Vien STR- induced allodynia should be blocked by NMDA-receptor antagonists. While NMDA receptor antagonists are known to attenuate tadile-evoked agitation in conscious

STR-treated rats (Yaksh, 1989). their dose response effects in the anesthetized

STR model have not been detennined.

1.6 EHecfs of Local Anaesthetics on Normal Nocicepfion and Neuropafhic

Pain

Local anaesthetics such as mexiletine, lidocaine and tocainide are known to be effective in the treatment of dinical and experimental neuropathic pain, including allodynia. This includes the treatment of pain arising from peripheral nerve injury, chronic painful diabetic neuropathy, trigeminal neuralgia, causalgia and other 28 chronic pain syndromes which are resistant to conventional dmg therapy (Lindstrom and Lindblorn, 1987; Dejgard et al., 1988; Chabal et al., 1992; Colclough et al.,

1993). Indeed, a sustained analgesic response has been reported weeks after the temination of infusion in humans (Edward et al., 1985, Steinhaus and Howland,

1958; Kastnip et al., 1986). Systemic mexiletine has also been shown to relieve dironic allodynia-like symptoms in rats with ischemic spinal cord injury (Xu et al.,

1992) or spinal nerve ligation (Chaplan et al., 1995).

The phannacological effect of local maesthetics is nomally associated with frequency- and voltage-dependent blockade of Na' channels, resulting in the nerve amduaion block. However, the analgesic effect in patients with neuropathic pain is unlikely to result from simple peripheral mduction blodc Local anaesthetics do not Modc newe condudion when given systemically in clinically effective analgesic doses (Dejgard et al., 1988; Chabal et al., 1989; Rowbotham et al., 1991).

Electrophysiological studies have shmthat subanesthetic doses of lidocaine and tocainide significantly suppress C-fiber evoked polysynaptic responses of spinal murons, incompletely reduce reflex adivity in harnstnng flexor amotoneurons, and axnpletely blocked reflexes to noxious chernical and thermal stimuli in normal rats

(Woolf and Wiesenfeld-Hallin, 1985; DeJong et al., 1969; Dohi et al., 1979). C- fiber evoked Wind up' is also blocked by i.t lidocaine in the rat (Fraser et al., 1992).

In &O, subanesthetic concentrations of lidocaine inhibited direct and synaptically driven NMDA- and neurokinin-receptor mediated post-synaptic depolarkations in 29 rat spinal neurons (Nagy and Woolf, 1996). These results indicate that under physiological conditions subanesthetic doses of local anesthetics seledive block

C fiberevoked activity in the spinal axd.

Mexiletine, an orally active congener of [idocaine, is the most commonly used local anesthetic for clinical neuropathic pain (Hegarty and Portenoy, 1994) and should be effective in any valid experimental mode1 of allodynia. Moreover, if

AB-input acquires access to pain signalling-pathways in the spinal cord, Vien rnexiletine should block both the AB-evoked responses (in presence of STR) and

C-fiber evoked responses (without STR) with cumparable potency.

1.7 Inhibitory Modulation in Sensory Processing

1.7.1 Role of SpinalAdenosine

There is a growing body of evidence implicating the role of purines in the spinal modulation of afferent neurotransrnission (Sawynok et al., 1986, Delander and Hopkins, 1986). Adenosine , a probable purinergic agonist, is thought to exert its effed at specific purine receptor sites. These receptors are Gprotein coupled and are ciassifieci on the basis of their ability to inhibit (A,-receptor) or stimulate (4- receptor) adenylate cyclase through Gi and Gs proteins, respectively (Van Calker et al., 1979; Marala and Mustafa 1993; Jocùers el al., 1994). A,-agonists such as

LPlA [N% (L-2-phenylisopropyl)-adenosine] or cyclohexyladenosine in the dose 30 range of 0.5 - 10 pg, and non-seleetive (A,IAJ agonists like NECA [5'-(N-ethyl carboxamido)-adenosine] or Zchloroadenosine in the dose range of 0.5 - 3.6 pg, produced potent analgesia in the tail flidc and hot plate tests following i.t. administration (Sawynok et al., 1986; Delander and Hopkins, 1986). Conversely, i.t. adenosine receptor antagonists such as theophylline (11.8 pg) and 8- phenyltheophylline (0.12 pg) attenuated adenosine-induced antinociception

(Sawynok et al., 1986; Delander and Hopkins. 1986). Such pharmacological studies have consistently implicated A,-, rather than 4 -receptors in the spinal modulation of nociceptive input (Fastbom et al.. 1990; Karlsten et al., 1991; Poon and Sawynok, 1995; Lee and Yaksh, 1996).

That adenosine affects ABevoked responses was first dernonstrated by

Salter and Henry (1987). They observed that the depressant effed of vibration, applied to the ipsilateral hind limb, on nociceptive (*de dynamic range) neurons in the cat dorsal hom was potentiated by an adenosine uptake inhibitor

(dipyridamole, 1.2 mgkg, ix). This inhibitory effect was reversed by caffeine (20

ôO mgkg, i.v.), and unaffected by naloxone (0.1 - 0.4 mgkg, i-v.). They proposed that the analgesic effed of vibration in the spinal cord is rnediated primarily by purines and may underlie the mechanisrn of adion of transcutaneous electncal nerve stimulation (TENS) in clinical pain management, as predicted by the gate control theory (Melzadc and Wall, 1965). ENSactivates large diameter primary afferent neurons and produces analgesia in the area affected by such stimulation. 31

Convenely, in the inherited nervous system disorder known as the Lesh-Nyhan syndrome, affected diildren show involuntary movements and seif-injurious behaviour (e-g. biting of lips and fi-). Patients with mis syndrome are known to la& the enzyme, hypoxanthineguanine phosphoribosyl tramferase, thereby reduu'ng the plasma concentration d adenosine. Whether Viere is a conespondhg decrease in the CNS concentration of adenosine is unknown but caffeine does worsen this syndrome (Stone, 1981).

Based on these experimental and dinical observations, adenosine has been used in the management of neuropattiic pain. The effed of systemic adenosine on pain symptoms in seven patients with peripheral neuropathy was evaluated in double-blind, placeboantrolled crossover study (Baifrage et al., 1995).

Adenosine (50 pglkglmin, i.v.) reduced spontaneous pain, allodynia and hyperalgesia whereas placebo had no effect. In other case reports, adenosine (70 pglkghin, Ï.v.) or R-PiA (1 9 pg, i.t)alleviated spontaneous pain and touch-evoked allodynia in patients with peripheral neuropathy (Karlsten and Gordh, 1995; Sollevi et al., 1995). Low dose adenosine (70 ~glkglmin,Lv.) has also been show to effed analgesia in awake healthy volunteen subjeded to experimentally-induced ischemic pain (Segerdhal et al., 1994).

Similar results have been obtained frorn studies using animal models of neuropathic pain. Yamamoto and Yaksh (1 991) reporteci antinociception in thermal hyperalgesia with i.t. NECA (0.01 - 0.3 nmol) in rats subjected to dironic nerve 32

-dion injury (Bennet and Xie, 1988). lntrathecal R-PIA (10 ng) and NECA (5 ng) reduced the pain-like behaviaurs induced by i.t prostaglandin F, in mice

(Minami et al., 1992). LPlA or NECA, in the range of 0.31 mol i.t., dose- dependently inhibited touch-evoked agitation in conscious STR-treated rats

(Sosnowski and Yaksh. 1969). Both dmgs were equigotent in STR model and were significantly blocked by i-p. caffeine. The potent inhibitory effed of these adenosine analogues ocaimed at doses that were only mildly antinociceptive in the hot plate test (Sosnowski and Yaksh, 1989). These results demonstrate the remarkable sensitivity of abnorrnal pain states to adenosine or related agonists, in both injury and non-injury mode1 of neuropathic pain. However, the receptor subtypes (A, versus 4) mediating the anti-allodynic effect of adenosine have not been systematically investigated.

f.7.2 Role of Spinal GABA

GABA is densely locaiized in the spinal cord (Sivilotti and Nistri, 1991;

Bohlhalter et al., 1994) where it appean to modulate both nociceptive and non- nociceptive neurotransrnission. In case of nociception, GABA ,-agonists such as muscimol and GABb-agonist like bacîofen given i.t. in dose range of 0.1-1 pg, significantly increased the response latency in the tail flick and hot plate tests, and vocalkation threshold to tail shock in the rat (Roberts et al., 1986; Hammond and

Drawer, 1984; Wilson and Yaksh, 1978). lntrathecal baclofen and muscimol (0.3 - 33

flinch response to fmline injedian (SC) in a &&pendent manner (Diring and

Yaksh, 1995). Although these studies did not report any motor dysfundion at

antinociceptive doses(s lpg, il), comparable doses of muscimol anâ baclofen

(21W, i.t) produce muscle Raccidity in rats (Hammond and Drower, 1984, Diring

and Yeksh, 1995). Convemly, -+eoeptor antapists such os picrotoxin

(0.25 pg) and biaiailline (1 pg) decreased mciœptive threshold in the tail-shodc

and taiI4Wc tests (Roberts et al., 19û6). The i.t administration of CGP 35348 (30

pg) or phadophen (100 pg) (GAB&bntagonists) also blocked the antinociceptive

effects of glutamate micro-injecteci into the ventromdial medulla of monkeys

(MdIawan and Hamm,1993). These phamiacological data are supporteci by

electrophysiobgiml studies demonstrating the inhibition of glutamate- and pinch-

evoked by GABq ionophoretically applied ont0 spinothalamic tract cells in

monkeys (Willeockson et al., 1984). In vitro studies using rat spinal card

preparation have show that Ahfferent fibers activate GABAergic intemeurons which inhibited nearby dorsal hom newons (Yoshimura and Nishi, 1995). Bath application of musiml and baclofen also depressed excitatory synaptic transmission (Kangrga et al., 1991; kaki, 1994). Thus, both GA84 and GABq receptors in the opinal cord appear to mediate the antinociceptive effeds of exogenowly administered GABA agents.

GABA has also been implicated in the spinal modulation of low-threshold afferent input Thus, i.t. bicuarlline (30 pg) (GABAAAantagonist)induced an 34 allodynia-like state in the rat similar to that prodoced by i.t STR (Yaksh, 1989).

Systemic administration of badofen (O.? mgkg, i-p.), but not muscimol (Imglkg, i.p.), inhibited allodynia-like behavioun an'sing 14 âays after focal spinal cord ischemia in the rat (Hao et al., 1992). Thus, spinal GAB& specifically GAB&- receptors appean to modulate A@transmission in a mamer shilar to spinal glycine in nonnal conditions (absence d neural injury); however, in aie presence of neural injury, spinal GABq-receptors seem to play a major role in the modulation.

Consistent with this hypothesis is the observation that GABA and glycine are co-localized in int-ns in lamina I-IIIof the dorsal hom (Mitchell et al., 1993).

In fact glycine immunoreadivity in lamina 1-111 is virtually restfided to neurons that also exhibit GABA immunoreactivity (Todd anâ Sullivan, 1990) suggesting that the

M8ds of GABA and glycine on synaptic transmission may be cornplementary. The receptors rnediaüng the distinct eff8ds of glycine and GABA are co-localized on the same post-synaptic membrane (Mitchell et al., 1993; Todd et al., 1996). Finally, these GAB&- and glycine-reœptors are separately axipled to chloride ion channels leading to increased chloride condudance and neuronal

Recent eledrophysiological studies in the rat spinal cord slices have demoristrated that evoked inhibiiory post synaptic potentials (IPSP1s)consist of two distinct components; a shortduration and longduration IPSP. Bath application of

SIR and biaiailline blod

Hile the long-IPSP's are mediated by GABA acting at GAB&-recepton

(Yoshirnura and Nishi. 1995). The time course of GA6A- and glycinemediatecl

If SP's in this study are similar to those previously recorded in motmurons of cats

(Rudomin et al., 1990). if glycine and GABA are released together into the synapse upon stimulation, and if GABA and glycine ened distinct but complementary inhibition of Wnput in the spinal cd,then the alkdynia arising from Vie blockade of spinal glycine receptors should be attenuated by the concurrent activation of

Wh-,but not GAB&-rewptors.

1.7.3 Role of Spinal Glycine

The importance of glycine in the processing of AB input has been e>densively disarssed (see section 1.4.2). Thus, the disinhibition of spinal glycine by i.t. STR

(a aompetitive glycine antagonist), produces an allodynia-like state in rats, similar to that men in patients with neuropathic pain. Consistent with the role of STR- sensitive glycine receptors in the dysesthetic state is the observation that glycine and betaine (a glycine analogue) inhibited STR-induced allody nia in lightly anesthtized rats (Sherman and Loomis, 1994). However, the charged nature of these molearles at physiological pH required that they be injeded diredly into the subarachnoid space.

Milacemide (Znpentylaminoacetamide) is an orally active glycine prodnig. 36

Senwn milacemide rapidly enters and equilibrates with the CNS cornpartment where

it is metaboliseci priman'ly by MAO-B to glycinamide (Christophe et al., 1983, De

Varebeke et al., 1989) and finally to glycine (see Figure 3). Milacemide (100, 200.

400 mgkg i-p.) dosedependently increased the concentrations of glycine and glycinamide (intemediary metabolite) in the cerebrqinal fluid (CSF) of the rat

(Semba et al., 1993) and remained significantly elevated 7h after administration

(400 mgkg, i-p.) (Semba and Patsalos, 1993; Semba et al., 1993). Serum glycine concentration were unaffeded. Clinical studies have confimecl the antiamvulsant, mood elevating, cognitive and mernory enhancing properties of milaœmide (Van

Dorsser et al., 1983; Houtkooper et al., 19û6; Saletu et al., 1986; Handelman et al.,

1989; Schwartz et al., 1992). In view of the ability of systemic milacemide to increase central glycine concentration (Semba et al., 1993), and given the locus and mechanism of action of STR-allodynia, milacemide is an obvious candidate for testing in the STR model.

1.8 Rationale and Specific Research Objectives

Allodynia, a symptom of clinical neuropathic pain, is a condition in which normally non-nociceptive tactile stimuli acquire the ability to evoke exuuciating pain. Althwgh, the neuropathology is unclear, changes in the spinal processing of low-threshold afferent input are strongly suggested. Under normal conditions, WB MllACEMlDE GLYClNAMlOE ---) GLYCINE

MAO-6 MILACEMIDE -GLYCINAMIDE -GLYCINE

BLOOD

Figum 3. Systemic mllacemide mdiy cmsses bkm%bmin-banier (BBB) whmil k converted to gwlne by the Kdkn of mono.mine oxidase-8 (MAO- 8). Milaœmide rapidly equilibrates between blooâ and brain. As a substrate for MAOS, milacemide is first rnetaboiised to giya'namide and subsequently converted to glycine. Milacemide is rnetabolised primarily in Vie central rather than the peripheral cornpartment. While some glycinamide produced in the periphery can cross the blood brain barnier, its contribution to the central concentration is srnall. (Adapted from Semba et al., 1993) 38 the synaptic transmission of rnechanareceptive &fibers is modulated by glycine- mtaining intemeurwrs in the substantia gelatinosa. Consistent with the concept of a central dysfunction in allodynia is the observation that i.t STR. (cornpetitive glycine antagonist) produces reversible allodynia in both conscious and lightly anesthetized rats.

Phamacological studies have shown that STR-allodynia is insensitive to conventional analgesics (morphine, capsaicin and NK,-antagonist), but is dose- dependently reversed by i.t. glycine and betaine. The latter have no effect on mlmkeption at anti-allodynic doses. Thus, STR-allodynia exhibits a spinal pharmacology that is disünd from nmlnoa'ception, is qualitatively sirnilar to that of clinical allodynia, and may represent a usehil noninjury model for the investigation of allodynia. To further investigate the spinal phamacology of STR- allodynia and to explore the neuromodulatory systems affecting the spinal processing of AP transmission (Figure l),the aim of the present study was to investigate the role of spinal NMDA-, adenosine-, GABA-, and gl ycine-receptors in lightly anesthetized STR-treated rats.

The specific objectives of this study were:

1. To determine the effed of i.t AP-7 (an NMDAieoeptor antagonist) on STR-

induced allodynia. 39

2. To compare the effed of i.v. mexiletine on STR-induced allodynia and

normal nociception (without STR).

3. To compare the effect of i.t. cyclopentyladenasine (A,-agonist) and CGS-

21680 (h-agonist) un STR-induced allodynia and to detenine the

sensitivity of their respective effeds to pretreatrnent with i.t. DPCPX (4-

antagonist).

4. To compare the effect of i.t. muscimol (GABA,-agonist) and baclofen

(GABhagonist) on STR-induced allodynia.

5. To compare the effect of i.v. milacemide (a glycine prodnig) on STR-

induced allodynia and normal nociception (without STR) and to determine

the sensitivity of this effect to pretreatment with clorgyline or Ideprenyl. 2 Methods

2.f Animals

Male. Sprague-Dawley rats (260 - 490 g at the time of the experiments) were

obtained from the Vivarium, Animal Care Services, Mernorial University of

Newfwndland (St. John's, Canada). Animals were housed in the Animal Care

FacilÏty, with a 1247 daMigM cyde (IigMs on 07:ûû h), a temperature of 22"C,

and free access to rat chow and tap water. All experiments were condudeci in

accordance with the guidelines of the Canadian Council on Animal Care and were

approved by the Mernorial University Animal Care Committee.

2.2 Implantation of Infrathecal Catheters

Anesthesia was induced with 4% halothane (Halocarbon Laboratories, River

Edge, USA) in oxygen and maintained at a concentration of 2.0 - 2.5 %. Rats were

fitted with i.t catheters prepared from stretched polyethylene tubing (PE-10) pulled

to 4.5 x the original length (Sherman and Loomis, 1994). The catheters were filled

with sterile saline (Astra Pharma Inc., Mississauga, Canada), inserted through the

cisterna magna into the spinal subaradinoid space, and guided 8.5 cm caudally terminating near the LM2 segment. A fixed loop near the rostral end of the

catheter was sutured to the overlying muscle to seaxe the catheter and the incision

dosed. The rostral tip was extemalised on the top of the head and sealed with a 41

stainless steel plug. Animais with i.t cathetm were Vien housed individually and

permitted to recover for at least 4 days. Only those animals exhibiting nomial

feeding, drinking and grooming behavior, and having nomial $ait were used for

2.3 Acute AneesthetkedAniml Prep.l;iltkn

On the dey d the expwiment, animals were anestheüred with halothane and

the left jugular vein was cannulated. Thereafter, halothane was gradually

discontinued and anesthesia was maintained using i-v. urethane (10% whin

saline, Sigma Chernical Inc-, St Louis, USA). The initial dose of 1.1 mg was

administered slowly over 10-1 5 min as the anesthetic efF8d of halothane dedined.

Body temperahre was maintained at 37'C with a themiostatically-regulated blanket

(Harvard Instnwnents, St Laurent, CaMda). The left carotid artery was cannulated

for continuous monitoring of blood pressure and kart rate using a pressure

transducer (P23XL) and polygraph (Mode1 79E. Grass Instruments, Quincy USA).

The trachea was also cannulated and the animal was allowed to breathe

sponta'bously. The incision was then closed and the animal was positioned in a

stwedaxic apparatus (Narishige, Tokyo, Japan) with the head fimly secured using ear bars. The inci'sion and the mtad points with the ear bars were coated with 2%

lidocaine gel (Astm Phamia, k.)to reduce the basal sensory input. Cortical EEG 42 was monitored using 2 subwtaneous needle electrodes (E2, Grass instruments)

pbced 2 mm left of midline, me extending rostrally entering the skin near bregma,

the other extending caudally and entering the skin aban 2 mm caudal to the first.

The animal was then permitted to stabilize for one Mur. During this time,

anaesthesia was adjusted with i.v. urethane as required.

To ensure reproducibIe toucbevoked responses in STR-treated rats, a light

level of anaesthesia is required (Sherman and Loomis, 1994). A reliable correlation

has been obseweâ between the proportion of time that the EEG is "synchronized"

and aie depth of urethane anaesthesia (Angel et al., 1976; Lincoln et al., 1980).

For our purposes, light anaesthesia was defined as the presence of an EEG pattern which fluctuates between high amplitude ("synchronized') and low amplitude

("desynchronized" ) activity, with high amplitude adivity present for not more than

60% of the time. The basis for this cut off in the STR mode1 has been reported

previously (Sheman and Loomis, 1994). The level of anaesthesia was also assessed by observing the animal's reflex responses to stimulation.

2.4 Application of Stimuli

The innocuous tactile stimulus was applied in the fom of hair deflection

(HO). The hair on the legs, flanks and lower back was sequentially brushed with a cotton-tipped applicator using an oscillating motion (rate of 7-2 per sec; 2min 43 stimulus train applied at Smin intervals). Oscillating stimuli elicit cardiovascular, motor and newachemical responses in this model more effectively than stationary ones (Yaksh, 1989; Sherman and Loomis, 1994; Milne et al., 1996). Brushing was done witb m more force than required to move the applicator through the hair (haif defledion) çuch that mly the pelage was disturbed. In the absence of spinal STR, this stimulus produced no change in hart rate (HR) or blood pressure (BP) and was rot associatecl with any type of motor withdrawal response. After i.t STR (40 pg), the same stimulus applied to oie affeded dermatomes elicited a 1520 mm Hg increase in mean arterial pressure (MAP), a 15.20 beatslmin (bpm) increase in HR, and an abmpt motor withdrawal response (see results).

Noxious mechanical stimulation was perfonned using bilateral hind paw pinch. 60th paws were gnpped with hemostats at the pst axial border (on the mediai surface, distal to the tard joint) such that the region of skin covered &y each hemostat was 220 mm2. The force of the pinch was produced by 500 g wights attached to the handles of each hemostat This resulted in a final pressure on each paw of approw'mately 0.7 kPa. Wthdrawal of the hind paws was prevented until the end of the 10-second stimulus period. In some experiments, noxious chernical stimulation was applied in the form of a S.C. injedion (0.05 ml) of 50% mustard oil in each of the hind paws. 44 2.5 Dmg Administration

AP-7 [(+/-)-2-amino-7-phosphonoheptanoic acid), Wqclopentyladenosine

(CPA), CGS-21680 hydrochloride [2p-(2carboxyethyl)phenylethylami1~)-S-N-

eVIylcarboxamidoadenosine HCI; CGS]. DPCPX (1.3-dipropylk cyclopentylxanüiine), muscimol hydrobmmide, R(+)-badcSen hydrochloride, and dorgyline hydrochloride were obtained Rom Research Biachemicals International

(Natick, USA). Mexiletine hydrochloride [l-methyl-2~2,6-xylylo~y)~thyfamine hydrochloride; Boehringer lngelheim Inc., lngelheim am Rhein, Germany] and milacemide (2-N-pentylamino acetamide; Searl Inc., Skoki, USA) were generously provided by their manufadurers. Tyramine hydrodiloride and I-deprenyl hydrodiloride were obtained frorn Aldrich Chemical Company Inc. (Milwaukee,

USA). Mustard oil was obtained from Madison, Colman and Bell (East Rutherford,

USA).

All drugs were dissolved in 0.9% saline (Astra Pharma, Inc.) except CGS-

21680, which was dissolved in a 0.02 M solution of sodium hydroxide (Fisher

Scientific Company, Ottawa, Canada), and DPCPX which was dissolved in dimethyl suffoxide (DMSO; British Dnig House, Toronto, Canada). Mustard oil was diluted with an equal volume of 90% alcohol (Fisher Scientific Inc.). Strychnine hemisulphate (Sigma Chemical Company) was adrninistered i.t. in a volume of 4 pl to reduce the rostrocaudal spread in the CSF. All other i.t injections were made in a volume of 5 pl. lntravenous injections were made in a volume not exceeding 45 0.5 ml, and i.p. injections were made in volume of 1 ml. Dnig solutions were

flusheâ through i.t and i.v catkter using 10 pl and 1 ml stetile saline, respectively.

2.6 Erpen'mental ProfocoIs

2.6. f Contrer Expenment (l. t. vehicfe At. STR)

To establish the control responses to STR + HD (in absenoe of any dnrg

treatrnent), each rat received i.t vehicle followed 10 min later by i-t. STR (40 pg)

(Figure 4A). This dose of STR elicits optimal allodynia without convulsive adivity

in IighUy aneslhetized rats (Sherman and Lwmis, 1994). The HD stimulus (2-min

duration) was applied at 5min intervals after vehicle and strychnine injection

(Figure 4A). These stimulus trains were repeated until no HDevoked

cardiovascular or motor withdrawal responses were observed (-30 min after STR).

2.6.2 Quanfification of Strychnineinduced Aliodynia

The HD-EVOKED MAXIMUM increase in blood pressure and heart rate within 30 min of i.t STR was used for graphieal and statistical analysis (see section

2.8). The total period of time following STR injedion, during which a motor withdrawal responses could be evoked by HD was also detemined. In order to verify that each animal was maintained at an appropriate level of anaesthesia (less than 6056 synchrony in the EEG) throughout the experiment, the % synchrony was A. 1.t. vehicle / strychnine control (STR control)

O 5- 10 15 20 25 30 ...... min

B. General protocol for dose-response experiments

-- - O 60 75 80 8s 90 ,...... min

200 205 21 O 21 5 220 ...... min

Figure 4. Summary of the Strychnine (STR) control pmtocol (A) and dose- response profocols (B). All time sales are in minutes. Each hair defledion stimuius (HD) was applied for 2-min using a cotton tip applicatw. Following i.t. STR (40 pg), the HO stimulus was repeated every 5 min until the STR effeds were no longer detected (- 30 min) (A, B). One dose of i.t. AP-7, CPA, CGS, muscimol or baclofen was adrninistered 15 min before i.t STR (B; upper panel). The second dose was administered approximately two hours after the fia dose (8; lower panel). Preliminary timecourse experiments indicated mat a 15-min pretreatrnent period yielded optimal phamaoological activity for i.t. administration (A). 47 calculatecl by detemining the number of l-min intervals with a synchronous EEG

(determineâ visually) and expressing this as percentage of the experimental time.

For cbwqxme studies of i.t AP-7, CPA, CGS, muscimol and baclofen,

control responses to STR + HD were first determined in the absence of any dnig

treatment (see Figure 4A). Approximately one hour later, one dose of the above

mentioned drugs was administered i.t This was follwed 15 min later by i.t. STR

(40 pg). Mirdeftedion was then applied ab Minintewals as descn'bed previously

(Figure 48). All anirnals received multiple drug doses. The order of dosing was

from low to high to reduœ the possibility of a carry over effed and thus, the

recovery time. No more than three doses were administered to the same animal,

and no animal received any further dnig tfeatment unce the highest dose had ben

given. Preliminary tirnecourse expiments indicated that a t Minpretreatment

time was optimal for i.t drug administration and that 1.5 hr between doses was

suffident for complete recovery ftorn even the highest doses. Therefore,

subsequent doses wefe generally given approxiimately hno hours after the previous

dose (lower panel of Figure 48). For AP-7, mexiletine and CPA doseiesponse

studies, the i.t vehide/STR oontrol was repeated &er each dose to venfy complete

reeovefy from the pvious dose, and to e~urethe reproducibility of STRallodynia throughout the experiment. For the remaining doseiesponse experiments, il. 48 vehiclelSTR control responses were detemined on- at the onset of the

expriment All animals wre killed with an overdose of urethane at the end of the

expefiment.

2.6.4 DPCPX Fuperiments

The seledivity of i.t adenosine agonists for A,- and 4 -receptors were

detmined in a separate group of animals. Control responses to STR + HD were

initially established (see Figure 4A). Approxirnately one hour later, rats were

pretreated with i.t DPCPX (IO pg) followed 10 min later by i.t. CPA (0.2pg) or CGS

(5 pg). STR was adrninistered 15 min Mer CPA or CGS, and STR-allodynia was

quantifieci as describeci in section 2.6.2. The effed of CPA (0.2 pg) or CGS (5 pg)

alone (no DPCPX) on STR alladynia was also established. Thus, each rat was in

al! of the following protocols:

1.t. DMSO (15 pl) I i.t STR (40 pg) (STR control)

1.t. DPCPX (IO pg) 1 Lt. CPA (0.2 pg) l i.t. STR

1.t. DPCPX (10 pg) i1.t. CGS (5 pg) / i.t. STR

i.t. CPA (0.2 pg) / i.1. STR

i.t. CGS (5 pg) I i.t STR

A i.5hour recovery pend was allowed between protocols (sufficient for the STR- responses to retum to the control values). The dose of DPCPX (10 pg) used in mis study was selected on the basis of preliminary experirnents. 2.6.5 Mexileüne Ewpen'ments

Control responses to HD were determined 5 min after the i.t. injection of

saline (15 pl, Figure 4A). Ten minutes later, STR (40 pg) was injeded i.t. and the

HD stimulus was applied at Smin intervals (i.t vehide 1 STR control, Figure 4A)

until al1 sensory-evoked responses were no longer detectable (-30 min after the

STR injedion). Approximately 1 h after the first STR injedion, control responses

to calibrated paw pinch were detennined. Animals wre Vien injected with

mexiletine (5 or 15 mglkg i.v., infused slawly over 10 min) and calibrated paw pinch was repeated 5 min after temination of the infosion. When the responses to

noxious pinch had nomalized (about 1-2 min), i.t STR (40 pg) was injected and the

HD was applied at %in intervals for 30 min. Most animais received 2 doses of i.v.

mexiletine. Animals were allowed to recover for 2 h Mer the first mexiletine-STR treatment Before a second dose of mexiletine (15 or 30 mgkg i.v.) was given, the

STR control was repeated to verify recuvery from the initial mexiletine effed, as evidenced by a cornpiete retum of the STR control responses.

2.7 Pmtocols for Milacemide

To establish the time course of activity for each dose of milacemide in STR- mode( milacemide (100,400 or 600 mgikg) was injeded i.v. and the effect on STR- induced allodynia was followed for four hours. These doses and the 441time period 50 were chosen on the basis of preliminary expefiments.

2.7.f hmwwponse Study

The general praocd for the dose-resporise e>cpen*mentsir shwn in Figure

XA). At the onset of the expefiment, i.t vehiddSTR ooritrol reopocises were o#aW as desaibed prsvioody (sectiorrr 2.6.1-2; Figue 4).One hour later, one dose of milaoemide (100,400,600 mgkg) was slowly infuseci i.v. over perids of

1O (for 100 mglkg) -30 min (for 400 and 600 mglkg). Evwy hour thereaffer, i.t. STR

(40 W)was hjeded and the responses b HD wem detmined as describeci above

(sections 2.6.1-2). Due to the short duration of allodynia (40min), STR was administered every hour for four hours. Animals were killed wiVi an overdose d wethane at the end of the expriment.

2.7.2 Cornparison of the Effect of Milacemide on Sftychninwtllodynia and

To compare the effect of milacemide (600 mgkg) on STR-allodynia and mechanical nociception (without STR), HD-evoked responses in presence of i.t

STR (40 pg), and calibrated bilateral hind pawpindi in absence of STR were detemined every how for four houn following milacemide administration (Figure

58). Control responses to i.t. vehicielSTR were initially detemiid (Figure 4A).

Approximtely one hour later (amthe first STR Med had completely abated)

53 control respmses to calibrated winchwre detemined. Animals were Vien ihjeded with milacemide (600mgn

2.7.3 CEorgyline and L-Deprenyf Ekpdmnts

The generd proiocd fw assessing the effeds of ciorgyline or 1-deprenyl on the adion d i.v. milacmide is sham in Figue SC. lntrathecal vehidelSTR control responses weinitially detmined (Figue 4).Approximately one hour later, rats were pretreated with dorgyiine (1O mg,i.p.; Semba el al., 1993; de Varebeke et ai., 1988) or Ideprenyi (3 mgkg, Lp.; de Varebeke et al., 1988; Semba et al., 1993)

(Figure SC). Two hours later, milacemide (600 mglkg, i.v.) was infuseci over a penod d 30 min. Thereafter, HDevoked responses in the presence of i.t STR (40

W) were detemiined every hour fos far hara In a separate group of animals, the effed d dorgyline or Ideprenyi on SIR-allodynia (no milacemide) was determined every houfor four houm. To assess the selectivity of inhibition of clorgyline and

Idqxenyi for MAOd and MAO-B, respectivdy, i.v. tyrmine was irijecteâ at the end of the experiment (Figure SC). Doses ranged from 6.7 to 20 mgkg and cardiovasailar responses were recorded. 2.8 Dafa Anelysis

All cardiovaswlar data are presented as the maximum change in MAP (= systolic blood pressure + If3 pulse pressure) and HR evoked by the different sensory stimuli. The change in MAP or HR was detmined relative to the immediate (1 min) pre-stimulus control period (not relative to T=O). The m&mum of each of these evoked responses, obsenred within 30 min of i.t. STR, was used for graphical and staüsücal analysis. For the paw-pinch stimulus, the baseline MA? and HR averaged over l-min interval preœding stimulus application, was subtraded from the maximum response evoked by the stimulus, and this âiierence was reported.

With the exception of milacemide, all doseiesponse data were analysed using regression NOVA A modified t-test was used to detennine if the regression lines had significant dopes (=1). Variability associated with single measurements is indicated by the standard emof the mean (S.E.M.), mile variability associated with blocks of data is indicated by the pooled 9556 confidence interval (CI). ED, and 95% CI values were detemined for al1 dose-response ames (Figure 7, 10-1 3,

15, 17). The milacernide tirneaune data and doseiesponse data were analysed using one-way ANOVA; significant differences were identïïied using the Neuman-

Keuls test (figure16, 17, 19). Statistically significant differences among multiple independent treatment groups (i.e. DPCPX experiments) were also detemined using completely randomiseci one-way ANOVA followed by the Neuman-Keuls test 55

(Figure 14). The unpaired, two-tailed students t-test was used to th8 compare the diffw~between hivo groups (Figure 18). Methods of data analysis were based on a general text (Tallarida and Murray, 1987). 56 3 Results

3.1 Geneml Obsewaflons FolIowing Intreth8~8ISttychnine

Stroking the hair on the legs, flanks, and lower badc of the lightly

anesthetid animal with a cottm tip applicator was without effd in i.t. saline-

treated animals (Figure 64. An identical stimulus applied to rats pretreated with

i.t STR (40 pg) evoked a progressive rise in HR and BP that persisted beyonâ the

durath of the HO stimulus (Figure 68). These responses were accornpanied by

an akupt Scratchhg response near Vie site of stimulation andlor motor withdrawal,

as wall as desynchrony of the EEG (Figure 68). Normalization of the EEG (return

to a synchronous pattern), which occumd 23min after the discontinuation of the

HO stimulus, coincided with recovery of baseline BP and HR These cardiovaswlar

and motor responses were only evoked by HD at discrete sites correspondhg to

dermatomes imatedby spinal segments near the site of STR injedion. The time

course of these tadile-evoked cardiovasailar responses were comparable to those

elicited by the intradernial injection of mustard oil into the hind paw (without STR) and were greater in magnitude (Figure 6C). Average control responses to HD in

STR-treated rats were 10-15 mm Hg for MAP, 20-25 beatslmin for HR and a duration of 15-20 min during which motor withdrawal responses which could be evoked by HD. These appear as the solid horkontal line in each of the dose- response airves (e.g. Figure 7A-C). The STR effects iasted for approximately 30 min wiih the peak evoked cardiovaswlar responses appearing 5 min after i.t. STR and corikal EEG iltustmtlng the eiYi13~fof nonnoxfous hair detlecUon in introahd (Kt) saIine and Wychninofmted mfs. Hair defiedion (HD) to the legs, fiank and lmbad< 5 min affer the i.t. injedion of saline had little effed on blood pressure, hart rate or EEG (A). The same stimulus applied 5 min afler i.t. strychnine (40 pg) evoked a macked and sustained inaease in blood pressure and heart rate, and desynchrony of the EEG (6). These evoked responses were localized to dermatomes affeded by the i-t. injedion of strychnine, and obsefved without convulsions or spontaneous motor movements. Each large deflection on aie time scale (upper trace) indicates 1 min and the horizontal bar below Vie tirne trace indiCates the applicaüm of the HO stimulus. Far comparative pwposes, the responses evoked by the nitrademial injection of mustard oil are shown (C). The heart rate bachg is enlarged for darity at the bottom of the figure. Arrows indicate the i.t. injedion of saline (A), the i.t. injection of strychnine (B) or the intradermal injection of mustard oïl into the hind paws (C). (data not shown).

3.2 EChct of AP-7 on Sttychnlne-allodynk

Prebeabnent mth i.t AP-7 ( 0.2. 1.5 pg) produced a dompendent reduciion d al1 HWvokeâ cerdiovasailar and motor withdrawal responses in STR-treated rats

(Figure 7AC). The corresponding EDdand 95% Cls of AP-7 are summarised in

Table III. The EDdranged from 0.36 - 1.71 pg. The ?h synchrony of the EEG remained unchangeci afler Af-7 injection, regardless of the dose (Figure 7D).

Baseline HR and arterial blood pressure were also unaffeded by AP-7 (data not shown). To compare the spinal effect of AP-7 on STR-allodynia and normal nociception, rats pretreated with i.t. AP-7 received calibrated bilateral hind paw pinch (Without STR). A dose of i.t AP-7 (5 pg) effecting almost complete blockade of STR-ailodynia had no effect on the responses evoked by noxious pawpinch

(Figure 8).

Table Ili. ED& and 95% C.I. of AP-7

A?-7 i.t.* MAP HR WD

* AP-7 was injected 15 min before STR. Figure 7. lntratl>ecal A P-7 dosMependent& suppmsses infrathecal strychnine(STR)-allodynia. Following a 15-min pretreatment with i-t. AP-7, responses to hair deflection (HD) were determined in the presence of i.t STR (40 pg). The HDamkeâ increase in mean arterial pressure (M-AP.; A), heart rate (B), the duration for which withdrawal response cwld be evoked (C) and the % synchrony in the EEG (D) are shown. Each point represents mean i S.E.M. of 4-6 animals. Least squares regression lines and the mesponding 95% CI (dotted Iines) are shawn. The horizontal solid lines and adjacent dotted lines indicate the mean I. 95% CI of al1 STR treatments (N = 14-16) before AP-7. Sûydnim (40 )rg i-t) toilouuirig AP-7 (S a it) A HD

Sample p>aph fracings of cvtedal bloodpmssum, beatt rate and corocal €EG ijluîbiating the et%ctof intmtlrecal AP-7 on sfrychnineallodynia (A)and noxbus mechanka! hind pow-pinch (B) In fhe rot. AP-7 (5 pg, i.t ) was injected 15 min before the i.t injedion of strychnine (40 pg). The anw indicates the injection d strychnine (A). Hair deflection (HD) was applied to the legs, fianks andlawer back af the rat 5 min after strychnine. Note the complete blockade of al1 HD-evoked responses by AP-7 in the strychnine-treated rat. However, identical dose of AP-7 failed to blodc the noxious paw-pinch evoked responses (0). Each large defiedion on the tirne scale (upper tram) indicates 1 min and the horizontal bar below the üme trace indicates the application of stimulus. 62

3.3 EWtof Mexiletine on StrychnineaIIodynla and Nonnal Nockepthn

lntfavmous mexiletine (5, 15 and 30 mgkg), administered 5 min before i.t

STR, dosedependently inhibited al1 the indices of STRallodynia with ED& ranging f'13.3 - 16.6 mglkg (Table N,upper panel of Figure 9, Figure 1OA-C).

At 30 mgkg, there was a srnall but non-signifiant increase in % syndirony compared to STR.co(~trd(Figwe 100). Nevertheless, this increase was well below the previaisly established cutoff of 60% syndirony (Sherman and Loomis, 1994).

The i.v. injection of mexiletine produced an immediate dose-dependent deaease in baseline BP and HR (Le. before STRallodynia). Baseline blood pressure decreasd by 10 mm Hg (5 mgkg) to 20 mm Hg (30 mg/kg); an effed which lasted for approximately 1-2 min. In al1 cases, blood pressure retumed to normal before i.t STR was injedeci (data not shown). Baseline HR, which decreased by 30 bpm

(5 mgkg) to 60 bprn (30 mgkg), was fully recovered 5 min after the injection of mexiletine (5 mglkg). At higher doses, baseline HR remained partially depresseâ

(-4% of pre-injection values for 15 mgkg and -856 for 30 mgkg) throughout the

STR treatment.

The efktof i.v. mexiletine was also detemined against noxious paw-pinch using the same experimental animals. Calibrated bilateral hind paw pinch was applied 3 min More the injedion of i.t STR Wleüne dosdependently inhibited pinch-evoked cardiovascular responses (Figure 1 1A-B), and prevented desyndKony of the EEG (Figure 9, lower panel). Al1 rats receiving 15 mgkg dose 3

Figum 9. Sample polygraph tacings of etteriel blood pressure, heart ate, and cortical EEG illustrsting the effect of infravenous mexiletine on s~chnineallodynia(A) and noxious mechanical hhd pew pinch (8) in the rat. Mexiietine (30 mgkg, Lv.) was injecteci 5 min before the i.t injection of strychnine (40 pg). The arrow indicates the injection of strychnine (A). Hair ddedion (HD) was applied to the legs, fianks and lower back of the rat 5 min after strychnine. Note the complete blockade of al1 hair defiedion (HD)+voked responses by mexiletine in the strychnine-treated fat (A) and its inhibitory effed against calibrated bilateral hind paw-pinch without strychnine in Vie same animal (B). Each large deflection on Vie time sale (upper trace) indicates 1 min and the horizontal bar below the time trace indicates the application of stimulus. --. n 09 5 10 2030 SO l.V. Mexiletine (mgkg)

-5 1 4 1 3 5 10 20 30 50 O3 5 10 2030 50 I.V. Mexiletine (mg) 1.V. Mexiletine (mgkg)

FQum f0. lntrevenous maxilefine dosedependenfly suppresses intmthecal s&ychnine(STR)-al'odynia. Follow-ng a 5min pretreatment wi't h i.v. mexiletine, responses to hair defledion (HD) were detennined in the presence of i.t. STR (40 pg). The HD-8voked imease in mean arterial pressure (M.A.P.; A), heart rate (B), the duration for which withdrawal response wuld be evoked (C) and the % synchrmy in the EEG (O) are çhown. Each point represents mean ISEM. of 4-7 animals. Least squares regression lines and Vie corresponding 95% CI (dotted lines) are stmm. The horizontal solid lines and adjacent dotted lines indicate the mean i 95% CI of al1 ST R treatments (N=16) before mexiletine. 4 O3 s 10 2030 50 I.V. Mexiletine (mg(kg)

Fm11. lntmvenous mexiletine dosedependentiy suppresses paw-pinch induced responses in absence of swchnine (STR). Following a Win pretreatment with i.v. rnexiletine, responses to calibrated bilateral hind paw-pinch were detemined (without STR). The stimulus evoked inmeas8 in MAP.(A) and HR (B) are shown. Each point represents the mean î SEM. of 4-5 animals. Least square regression lines and the corresponding 95% CI (dotted Iines) are shown. The horizontal solid Iines indicate the mean &5% CI of al1 pawpinch evoked responses (N=16) before mexiletine. 66

attempted to mthdraw their hind paws during noxious stimulation but were

prevented from doing so to ensure a consistent noxious mechanical stimulation.

No withdrawal attempts were recorded in any of the rats receiving aie 30 mgkg

dose. The ED,'s of i.v. mexiletine against noxious pinch (without STR) and STR-

allodynia were very comparable and exhibited narraw 95% Cl's (Table N).

Table W. ED4and 95% C.I. of i.v. MavieUne

MAP Heart Rate Duration of Withdrawal

MAP Heart Rate

* STR was injected 5 min after temination of the mexiletine infusion.

3.4 Effécts of Adenosine Agonists on SttychnineaIIodynia

Pretreatment with i.t. CPA (0.001, 0.05 and 0.1 pg), 15 min before STR, dose-dependently inhibited the HDevoked rise in BP (Figure 124, HR (Figure

128) and mto~withdrawal(Figure 12C). Overall, i.t CPA had no significant effect 1.1. CPA (ng) 1.7. CPA (ng)

I.T. CPA (ng)

Figure f2. Infrathecal CPA dose-dependenüy suppresses infrathecal strychnine(STR)-allodynla. Following a 15-min pretreatment with i.t. CPA, responses to hair deflection (HO) were detemined in the presence of i.t STR (40 pg). The HD+voked increase in rnean arten'al pressure (MAP.; A), heart rate (B), the duration for which withdrawal response could be evoked (C) and the % syndirony in the EEG (O) are shown. Each point represents mean k S.E.M. of 54 animals. Least squares regression lines and the corresponding 95% Cl (dotted lines) are shmm. The horizontal solid lines and adjacent dotted lines indicate the mean î 95% CI of al1 STR treatments (N=16)before CPA injection. 68 on % synchrony of the EEG (Figure 120). However, the data from two animals receiving 0.05 w and one animal receiving 0.1 were exduded from the final resultsasthe%synchronyexceededthe60JCcutoff. TheED,ofi.t CPAraqpd fbm 0.û2 - 0.07 pg (TaMe V). The inhibitary elled of i.t CPA (0.2 pg) was bbked by prebeatment with Ihe A,- DPCPX (IO a il;Figue 13). Intwestingly, i.t WCPX alone triggered miid qmdmmus SQBfching and HD+voked uhthdrawal responses for approPOmately 5 - 10 minutes Mer injection in two animals (data not -1- 1.t CGS also imbited STR-induced allodynia in a dosedependent manner

(Figwe 14). However, doses of 2 - 5 pg were requireâ to achieve inhibition cmpaath to that d i.t CPA at nearly 1Eû the dose. The EDdof i.t. CGS ranged from 2.7 - 3.1 pg (Table V). This anti4odynic effed was signifmntly blod

Table V. EDsand 95% CJ. of CPA or COS

CPA i.t * W 0.07 0.02 - 0.20 HR 0.06 0.01 - 0.30 \ND 0.02 0.01 - 0.09

- -- * CPA and CGS were injecta 15 min before STR. DPCPX+CPA+STR CPA+STR STR

M.A.P. Heart Rate WD %SYN (mm of Hg) (Be-min) (min)

Figure 13. Inhibition of the entiaIIllroynic e-t of CPA or CGS by pmtmatment wifh intrethecal DM.DPCCP (1O pg, i.t)was administered 10 min More i-tCPA (0.2 pg) or CGS (5 pg). STR (40 pg, i.tJ was given 15 min later and responses to hair ddedion (HD) were determined at 5min intervals thereafter. The effeds of CPA (A) or CGS (B) (in absence and presence of DPCPX) on the HD-evoked incfease in mean arterial pressure (M.A.P.), heart rate, duration for which motor wi*thdrawal respanses could be evoked (WD) and % synchrony (% SYN) are shown. The solid bar represent the STRcontrol responses without any dnig treatment Each bar represents mean ISEM of 6 - 8 animals. (*p<0.05, **p

I.T. CGS-21- (pg) I.T. CGS-21680 (Ccg)

Figure 74. lntrafhecal CGS doseâependentty suppresses intathecal sWchnine(STR)-allodynia. Following a 1%in pretreatment with i. t. CGS, responses to hair defledion (HD) were determined in the presence of i.t. STR (40 pg). The HPevoked imsein manarterial pressure (MAP.; A), heart rate (B), the duration for which withdrawal response could be evoked (C) and the % synchrony in Vie EEG (D) are shown. Each point represents mean * S.E.M. of 5-6 animals. Least squares regression lines and Vie corresponding 95% CI (dotted lines) are shawn. The horizontal solid lines and adjacent dotted lines indicate the mean I95% CI of al1 STR treatments (N=18) before CGS injedion. 71 the % synchrony remaining wichanged at al1 doses. At a dose of 5 pg, i.t CGS had

no antinociceptive effed against noxious bilateral hind paw pinch (n=2, data not shown).

3.5 E#èck of GABA Agonkts on Sttychninesllodynk

1.t muscibnol (0.1, 0.5, 1 and 2.5 pg), given 15 min before STR (40 pg) injection, dose dependently inhibited al1 indices of STR-allodynia (Figure 15). This was accomplished without a change in the % synchrony of the EEG, which remained below 60% cut off. The baseline artetial blood pressure and HR were also unaffeded by i.1. muscimol (data not shown). The ED, and 95% CI for i.t. muscimol are sumrnarised in Table VI.

Table W. EDgand 95% C.I. of Muscimor

Treatment Response EDSo(pg) 95% CI

* Muscimol was injeded 15 min before STR

ki contrast, -en (up to 1O pg, i.t) failed to inhibit STR-induced allodynia

(Table VII). Rather there was a trend towards increased cardiovascular responses Figum 15. Infrathecal muscimol dose-dependenUy suppmsses infmthecal strychnine(S1R)-allodynls. Following a i5-min pretreatrnent with i.t. muscimol, responses to hair deflection (HO) were detemined in the presence of i.t. STR (40 pg). The HD-evoked imease in mean arterial pressure (MAP.; A), heart rate (B), the duration for which withdrawal response could be evoked (C) and the % synchrony in the EEG (D) are shown. Each point represents mean IS.E.M. of 5-7 animals. Least squares regression lines and the cwresponding 95% CI (dotted lines) are shown. The horizontal solid lines and adjacent do= lines indicate the mean i 95% CI of al1 STR treatments (N=23-24)before muscimol injection. Mer i.t baclofen. However, HD*voked increase in MAP and HR following

baclofen (Table VII) was not significantly different from vehiciecontrol in STR- treated rats. The withdrawal responses to HD were significantly inhibited by i.t. badOfm (m.01).Lndeed, one rat receiving the 1 w dose, and al1 rats receiving

25 I~Qexhibiteci rwe&bie hinâ limb paralysis. Thus, HD-evoked motor withdrawal responses were completely abated by high doses of baclofen at a time when cardiovasarlar responses remained unafleded. The baseline arterial blood presswe and HR were also unaffecteci by i.t badofen.

Table W. Effèct of baclohn on HBevoked responses in STR-treated rets

Treatrnent MAP HR WD (mm Hg) (beatdmin) (min)

Baclofen* 1 PS (3) 22.8 i4.9 33.3 * 13.0 6.7 * 3.3* 5 lrg (4) 25.4 i 5.0 40 i 8.4 Absent 10 Hg (5) 22.3 + 4.8 29 & 9.4 Absent

*Bad- was injected 1 minbefore STR. Data are expressed as the mean * SEM of 3-14 animals (N values show in the brackets) (Spe0.01). Mean arterial pressure (W);heart rate (HR); withdrawal duration (MîD). 74 3.6 E-t of Milacemide on StsychnLe-aIldynia and Nonnal Nociceptlon

Pretfeatment with i.v milacemide (400 and 600 mgkg) significantly inhibited

STRaliodynia (Figure 16 and 17). Maximal inhibition was observed 2-3 h after dnig administration (Figure 16). The duration of adion ranged from 2 h for 400 mgkg dose to 4 h for 600 mglkg, and the % synchrony of the EEG was unchanged at al1 doses ttuwghout the experiment (Figure 16D and 17D). The EDdat the tirne of maximum inhibition (Le. 2 h after milacemide infusion) were approximately 400 mglkg i.v. Fable VIII). Baseline BP and HR were dosedependently depressed during milacemide infiision (i.e. before STR-allodynia). Baseline BP decreased by

15 (400 rnglkg) to 25 mm Hg (600 mgkg) and HR decreased by 25 (400 mgkg) to

50 bpm (600 mgkg). For the 400 mgkg dose, BP and HR retum to normal before i.t STR was injectecl (Le. 1 h after milacemide infusion). At 600 mgkg, BP and HR retumed to nomial 3 h after inhision, except in two anirnals where baseline HR and

BP remained depressed by approxirnately 10% and 8% of pre-injection values throughout the experirnent.

[Note: Dosdependent depression of respiration was observed (visual examination) during the first 15-20 min following Vie temination of milacemide infusion].

In a separate group of animals, responses to noxious hind paw pinch

(wiMSTR) and STR-allodynia were recorded every hour for 4 h after milacernide infusion. At the time of maximum inhibition of STR-allodynia (2 h pst-infusion), Tirne (hours) lime (hours)

Figure 16. lntravenous milacemide exerts ifsmaximum effect on Intratliecal stmnine(STR)-allodynia two houn affer injection withwt affecting EEG synchmny. Hair defletSion (HD)evoked responses in the presenœ of i.t. STR (40 pg) were determineci every hour for four hours after i.v. milacemide (400 mgkg; approximate ED, value). Because of the short duration of allodynia (~Wrnin),i.t. STR was repeated every hour. The maximum evoked inaease in mean arterial pressure (MAP.; A), heart rate (B), the duration for which a motor withdrawal response could be evoked (C) and the % synchrony in the EEG (D) are shown. Each point represents mean i SEM of 68 anirnals. The horizontal dotted line is the maximum evoked response to STR (mean ISEM) in absence of any fumer treatment (N=24).(+p<0.05, *+p

Figum 77. lntravenous milacemide suppresses intratheca/ sfvchnine(STR)- allodpia in a dose and time-dependent mamer, wttnout affecthg EEG synchmny. Hair defiedion (HDbvoked responses in the presence of i.t. SIR (40 pg) were measured every hour for four houn after i.v. milacernide (100, 400 and 600 rnglkg). Dose-rasponse curves for the change in mean arterial pressure (MAP; A), hart rate (B), the duration for which wiüidrawal responses could be evoked (C) and the % synchrony in the EEG (O) were determined 2h (@ time of maximum effect) and 4h (A) after injection. Each point represents the man î SEM of 68 animals. The horizontal dotted lin8 is the maximum evoked response in STR- treated rats (rnean * SEM) before milacernide (N=24).(*p<0.05, *Sp

responses evoked by noxious pinch were unM8Cfed by milacemide (Figure 18).

Tabh VW% ED, and 95% C.I. ofmihcemide

milacemide administration

3.6.f Effbcf of CbrgyIine and L-DepmnyI on the Action of Milacemide

The efF8Cf c# i.v. milacemide (600 mgkg) on STR-allodynia was detennined

in a separate grwp of rats pretreated with either 1-deprenyl (specific MAO6

inhibitor) or ciorgyline (MAO-A inhibitor). Clorgyline had no effect on milacemide

inhibition of STRalbdynia at either 2h or 4h baseci on HD-evoked changes in MAP

(Fil8A), HR and withdrawal dumon (data not shown). In contrast, Ideprenyl

significantly MOdced the anti-allodynic Mect milacemide for up to 4 hours (Figure

lm). EEG synchrony was unaffected by any treatrnent, and clorgyline or deprenyl

by themselves had no eiWt on STR-allodynia (data not shown). Tyramine (6.7 mgkg, i.v.) injeded at the end of the experiment, induced an irnmediate and susbinecl elevaüon in arterial blood pressure in dorgyline-treated rats (Figure 186).

Tyramine, up to 20 mgkg i-v., had no M8d on Mdpressure in Ideprenyl-treated rats. A Rats wem pm-tmated with dorgyline (1O rngkg, i.p.) or Ideprenyl(3 mgkg, Lp.) two hows before milacemide (600 rnglkg, i.v.). Hair clefledion (HD)-evoked responseg in the presence of i.t. STR (40 pg) were detemined every hour for four hous aRer milacemide. Because of the short duration of allodynia (~30min), STR was repeated every W. The -mm evded increase in mean artenal pressure (MAP),detmined 2h end 4h Mer milacemide, is presented. The horizontal dotted line represmts the maximum response (mean & SEM) in SIR-treated rats in absence of any other treatment (N=l4). Each bar represents the mean i SEM of 3-5 rats.(Sp< 0.05, *Irp

B. Tyramine (6.7 mgkg, i.v.) triggered imrnediate and sustained increase in arterial blood pressure in dorgyline-treateâ rats. Even at three times the dose (20 mgkg, i.v.) tyramine had no effect in rats pre-treated with Ideprenyl.

4 Discussion

4.1 Sttychnineallodynia

Hair defledion, applied to the denatomes of rats pretreated with i.t. STR, evoked physiological responses greater than those elicited by the chernical nociceptive agent, mustard oil (see sedion 4.2). All STRdependent responses were: a) evoked (not spontaneous); b) observed without convulsions (no generalized disinhibition of spinal neurons); c) only elicited by light brushing of the hair at cirwmscribed sites corresponding to the spinal segments affected by i.t.

STR and d) cmpletely reversible. These data indicate that robust allodynia can be seledively induced with i.1. STR in animals whose somatosensory systems are othenMse ml.They are also consistent with previous studies of STR-allodynia

(Beyer et al., 1985, 1988; Yaksh, 1989; Sherman and Loomis, 1994, 1995), many of which used phasic noxious stimuli for comparative purposes, demonstrating that nonnally innocuous mechanical stimulation acquires nociceptive characteristics dwing the phamacological blockade of spinal glycine receptors. Indeed, there is growing evidence that low threshold afferent input activates central nociceptive pathways in the presence of i.t. STR (Milne et al., 1996; Sherman et al., 1996).

4.2 Effbct of AP-7 on S~chnine-allodynia:mle of NMDA mepfors

STRallodynia was dosedependently inhibited by pretreatment with the 82 selective NMDA receptor antagonist, AP-7. This effect was not due to a general depression of cardiovascular reflexes as i.t. AP-7 did not alter baseline cardiovasailar adivity. nor did it inhibit the responses evoked by a phasic noxious pindi. The lack of an effed on cortical EEG synchrony, the kt. route of administration and the axrespondingly low ED, values of AP-7 (0.4 - 1.7 pg) are al1 consistent with a spinal site of action in this stuây. Normally, spinal NMDA receptors play little or no role in the synaptic transmission of low-threshold (A-fiber) input (Urban et al., 1994). Ralher, they are known to rnediate the phenornenon of

Wnd up' (e-g. the temporal summation of long latency discharges from afferent C- fibers and a charadefistic feature of pain) in spinal neurons during repeated high- threshold (C-fiber) stimulation (Dickenson and Sullivan, t 987). Indeed, NMDA receptor antagonists have been show to block 'wind up' in vivo without affecting the neuronal responses to peripheral A-fiber stimulation (Price et al., 1994). The observation that STR-allodynia is highly sensitive to spinal NMDA receptor blockade suggests that, in the presence of STR, low-threshold afferent input acquires access to a central sensitization mechanism normall y restricted to repeated noxious stimulation.

That this sensitization process might be 'wind up' is supported by the temporal pattern of BP and Hf3 responses evoked by tonic brushing of the hair in

STR-treated rats. This consisted of a progressive increase in cardiovascular responses during the stimulus train which persisted Mer the stimulus was 83 di-&. A sirnilar pattern of HD-evoked neurochemical responses has been observed in the locus coehileus d STR-treated rats using differential normal pulse voltammetry (Milne et al., 1996). This pattern of evokeâ responses was rnatched by the s.c. irijection of mwtard oil without STR. Mustard oil is a seledive adivator d Ccioacepto~s(Wodf and Wall, 1986) that tnggers 'wind up' in Vie rat; an eff8d

bkcked by pretreatment with the NMDA antagonist, MKaO1 (Woolf and Thornpson.

1991). These results, and the recent report that rat dorsal hmnewons exhibit

'wind up' to repeated A-fiber activation fol lowing the iontophoretic application of

STR (Wkox et al.. 1W6), indicate that: a) nomally innoaious mechanical stimuli

undergo nociceptive-like processing in the spinal wrd of the rat during glycine

receptor blod

The present dose-response data for i.t. AP-7 are comparable to those of other NMDA antagonists (MUSOI and AP-5) that inhibited tactile-evoked agitation

in conscious STR-treated rats (- ED&= 3 pg) (Yaksh, 1989), and in conscious mice pretreated with i.t. prostaglandin E, or F, (ED, = 1.6 pg for APS, and 0.52 pg for MKsOl; Minami et al., 1994). The similarity of these data, determined in

Iightly anesüwüzd and conscious rats, indicate that: a) the spinal phamacdogy of STR-allodynia is unaffect4 by controlled urethane anesthesia; and b) experirnental conditions rnitigating many of the ethical issues arising frorn studies of neuropathic pain in animals cm be used reliably in the STR-model. 84 Previous studies wnduded in our laboratory showed that STR-allodynia is inhibited by i.t. y-DGG and NBQX at doses that do not affectcardiovascular or motor reflexes, or the plane of anesthesia (Shemian and Loornis, 1994, 1996). The sensitivity of STRallodynia to spinal NMDA and non-NMDA antagonists is in agreement with ouf current understanding of sensory transmission at the level of the spinal cord (Figure 20), and the hypothesis that allodynia anses from the exaggerated depolarization of spinal newons in pain paürways by low-oireshold afferent input. Hence, agents that negatively modulate such excessive excitatory drive should blodc STR-allodynia.

4.3 Effect of Mexiletine on Strychnineallodynia and No& Nociception

Mexiletine, an orally adive congener of Iidocaine, is the most commonly used local anesthetic for dinical allodynia (Hegarty and Portenoy, 1994) and should be effective in any valid model of allodynia. In the present study, mexiletine dose- dependently inhibited STR-allodynia and paw-pinch induced normal nociception

(without STR) as detemiined in the same expeflmental animals. Importantly, mexiletine was equipotent in inhibiting the responses to normal and abnomial nociceptive input (ED,'s = 9.1-1 7 mgkg, i-v.). Below 30 mgkg, this effed was achieved without a change in EEG synchrony (cortical adivity reflecting the level of anesthesia) and without affeding motor Mwent pathways; the motor withdrawal Figum 20. Schemaflc diagram illustrptlng the postulated role of excifatory amino acid (EU)-recepfor systems and nwmmodulrtory systems In the spinal pmcessing of low-thmshold (AB) uffemnt input Glutamate, released from the central terminals of @fibers, combine with AMPAlkainate-receptors (closed avals) on both second order neurons and intemeurons in the dorsal hom. The latter include excitatory (EAA) and inhibitory [GABA and glycine (Gly)] substrates. Current evidence suggests that pst-synaptic NMDA recepton (open square) do not lie adjacent to primary afTerent terminals but rather to those of excitatory intemewons (Davies and Watkins, 1983; Evans and Long, 1989; Schouenborg and Sjolund, 1986; Moms, 1989). Thur, glutamate from EAA intemeurons is thight to interad with NMOA receptors on the seccnd order neuron (Davies and Watkins, 1983; Evans and Long, 1989; Headley and Grillner, 1990; Yoshimura and Nishi, 1995). The negaüve madulation of aderiosine, GABA and Gly on EAA-rnediated transmission is also indicated. For clarity, only the adenosine (A,) receptor (open circfe) is shown. Probable sources of adenosine include low- threshold pfimary affwents and intemeurons (Nagy and Daddona, 1985; Sawynok and Sweeney, 1989) (not shown). 86 responses to calibratd pawpinch mepresen,ed in mexiletine-treated rats. At 30 mglkg or above, motor withdrawal responses were almost absent and EEG syndwony was sligitly but msignificantly inaeased, consistent m'th th8 report of sedation, reduced loamiotor activity and a slowed righting reflex in consdous rats given a similar dose of mexiletine (Xu et al., 1992).

The mexiletine doses used in the present study are similar to those reported to inhibii allodynia-like symptoms Mer ischemic spinal cord injury in the rat (Xu et al., 1992). Spinal ischemia was inducd phot~icallyby laser irradiation at mid- oxxacic region immediately after the i.~.administration of Erythrocin B. Allodynia- like behaviours developed 68 months after irradiation and persisted for at least 4 months. Mexiletine (15 m@kg and 30 mgîkg, i.p.) significantly inhibited allodynia in these rats for 45-1 20 min; a duraüon of adion sirnilar to that observed in the STR model. While a direct cornparison of mexiletiine in injury vs nokinjuw (e.g. i.t. STR) models of neuropathic pain will require more extensive and detailed experiments, the present results suggest that dironic spinal cord injury yielding localized allodynia does not display enhanced sensitivity to the antiallodynic affect of mexiletine when compared to the present pharmacological model. Strychnine- allodynia may be useful for estimating the dose-response relationships of local anesthetics in pathological central pain states, and for investigating their central mechanism(s) of action.

Systemic local anesthetics, in doses ranging from 1-25 mgkg, do not block 87 amhdim in injured or non-injwed Wpheral nerve fiben (Woolf and Wiesenfelb

Hallin, 1985; Chabal et al., 1989; Tanelian and Maclver, 1991). In the case of mexiletine, i.v. doses of 10-1 5 rnglkg yield peak senimiplasma concentrations of

1-5 pglpl in the rat (Igwernezie et al., 1991, 1992). These are within the range reported for clinical analgesia and well below the concentration required to blodc electrical conduction of peri'pheral Ab and C-fibers in vitro (,250 pglml for lidocaine)

(Tanelian and Madver, 1991). Pain in peripheral neuropathic conditions has ben attnbuted to an abnomal afferent impulse barrage into the CNS from the trapped end of the injured peripheral nerves (neuromas) and from the axotomised sensory cell bodies in dorsal root ganglia (DRG). In tum, peripheral neuromas and their assa5ated DRG's are known to be sensitive targets for local anesthetics in cases of experimental or dinial nerve injury (Wall and Devor, 1983; Chabal et al., 1989;

Devor et al., 1992). For example, systemic local anesthetics suppress ectopic neuroma and DRG diçcharge withwt blocking initiation or propagation of impulses by electrical stimulation of sciatic nerve in rats with chronic sciatic nerve ligation

(Wall and Devor, 1983; Chabal et al., 1989; Devor et al., 1992). However, these

Wgets are not relevant to the STR mode1 as no peripheral injury is induced. Thus, there is no obvious peripheral site(s) of action of mexiletine in STR-treated rats.

In contrast, subanesthetic doses of local anesthetics are known ta: a) block conventional nociceptive transmission in the spinal cord of cats (DeJong et al.,

1969; Dohi et al., 1979), b) depress a C-fiber mediated polysynaptic reflex in Vie 88 spinal cord of normal rats without blocking condudion in primary afferent fibers

(Woolf and Wiesenfeld-hallin, 1985), c) block C-fiber mediated Wnd up' in rats

(Fraser et al., 1992) and d) inhibit both direct and synaptically driven NMDA- and namkiniMeceptor mediated postsynaptic depolarkation in the rat spinal neurons

(Nagy and Woolf, 1996). In the pmsent study, i.v. mexiletine bWedthe responses of both noXiaus pinch and STRallodynia with equal potency. These results suggest that mexiletine inhibits both normal and abnomal nociception at a cornmon central site in the pain-signalling pathway. At the very least, a spinal site of action (e-g., spinal nociceptive neurons receiving convergent low- and high-threshold afferent fiber input) is likely given the lasof adion of i.t. STR in this model. Su& a site would explain the identical inhibitory potency and parallel dose-response curves of i.v. mexiletine against noxious paw-pinch and STR-allodynia. However, supra- spinal sites of adion cannot be exduded. Electrophysiological experiments will be reguired to test this hypothesis.

In wmmary, i.v. mexiletine inhibits STR-allodynia in the anesthetized rat at doses that also block normal nociception. Sub-anesthetic doses of mexiletine can suppress centrally induced allodynia in experimental anirnals that are devoid of peripheral or central nerve injury. Mexiletine appean to have an important spinal site of action in abnormal pain states. 89 4.4 €#kt of Adenosine Agonïsts on S~hnine-allodynk:rok of A,

rpceptorr

The i.v. infusion of adenosine or the i.t injedion of aôenosine analogue.

LPlA (h-agonist) has been shomi to alleviate spontaneous pain, allodynia and pin-

prid< hyperalgesia in cases d peiipheral nerve injury (Sollevi et al., 1995; Karlsten and Gordh, 1995; Beîftage et al., 1995). Consistent with this effed, CPA (an A,-

selective agonist) and CGS (an A&ect ive agonist) dosedependently inhibited

STR-allodynia. However, the ED, of i.t. CGS ( 2.7-3.1 pg) was approximately 50 tirnes greater than that of CPA (0.02 - 0.07 pg) raising concems about the seledivity af CGS at spinal A,-receptors in the present study. This was confimed

by the ability of a fucd dose of DPCPX to bled< the anü-allodynic effect of both CPA and CGS. These data are in agreement with previous reports that the spinal analgesic adion of adenosine is selectively mediated by A,-recepton whereas the motor dysfunction induced by spinal adenosine is mediated by &-ceceptors

(Fastbom et al., 1990; Karlesten et al., 1991; Pmand Sawynok, 1995; Lee and

Yaksh, 1996). In this regard, it is important to note that motor withdrawal responses evoked by noxiow paw-pinch (without STR) persisted after the il. injection of CGS

(3 5pg) (data not shoum). These resuits indicate that CGS did not adversely affect the motor fundion in the STR-model.

Rie ED& of CPA in the present study are comparable in order of magnitude to those reporteci for RPlA and NECA in conscious STR-treated rats (EDd = 0.1 90 pg RPlA and NECA; Sosnowski and Yaksh, 1989) and for RPiA (€0, = 0.2 pg) in rats with peripheral nerve injury (chnxiic ligation of LS and L6 spinal nerves distal to DRGs; Kim and Chung mode[) (Lee and Yaksh, 1996). A sirnilar ED, of CGS (8 pg) was also reporteci in the spinal nerve ligation mode1 (Lee and Yaksh, 1996).

Unlike the present study however, the anti-allodynic efi& of CGS in the Kim and

Chung model, was not blocked by pretreatment with i.t. CPT (a seledive A,- receptor antagonist). There are several possible explanations for this difference.

First, the dose of CGS used in the injury mode1 was ramer high (21 pg). Second,

10 pg of DPCPX was used in the present study compared to 3 pg of i.t. CPT.

Altematively, nerve injury may trigger the expression of &mceptorsthrough which adenosine agonists exert their effects. This apparent difference in the pharmacology of the STR- versus the spinal nerve ligation mode1 could have important implications for the investigation of adenosine agonist and needs to be clarified in future experiments.

The absence of a significant effed on cortical EEG synchrony and the low

ED, values of CPA and CGS, support a spinal site of action in the present study. lndeed the analgesic adion of i.v. adenosine in patients with peripheral nerve injury appears to be mediated by central adenosine receptors as the activation of pen-pheral adenosinereceptors is known to result in painhil symptoms (Belfrage et al., 1995). LPlA and NECA, given i.t. in doses ranging from 0.3 - 1 nmol (doses inhibiting STR-allodynia in conscious rats), also produced dosedependent 91 antinociception in rats using the hot-plate and tail-flick tests (Sosmwski et al.,

1989). Similar doses d i.t. adenosine analogs also inhibited the scratching and biiing behaviw evoked by i.t SP and NMDA; an M8Cf antagonüed by Uieophylline

(Delander and Wahl, 1988). These resuh indicate that the inhibitory Mect of aclenosine is not restrided ta abnmlpain states.

The mechanism(s) by which aderiosine Wulates sensory processing in the

@na1 cord is unclear. Haurevw, the sensitivity of experimental models of allodynia

(both injury and noninjury) to i.t. EAA antagonists and adenosine A,agonists suggests a fundional interadion between dorsal hmglutamate- and adenosine- systerns. One possibility is that A,-receptors, present on glutamatecontaining intemeurons, inhibit glutamate release and attenuate EAA-mediated synaptic transmission (see Figure 20). Indeed, recent eledrophysiological recordings in spinal cord slices of the hamster have shown that a principal efFect of adenosine is to inhibit interneuronal communication in the substantia gelatinosa (Li and Perl,

1994). Superfusion of adenosine (20-100 pg) resulted in the inhibition of al1 polysynaptic excitatory post-synaptic airrents (EPSPs) evoked by dorsal root stimulation. Moreover, binding studies have shmthat A,- and 4-receptws are located primarily on intrinsic neurons in the rat spinal cord (Choca et al., 1968).

Although the identity of these neurons was not detemined, they could indude glutamateantaining intemeurons, second-order neurons, or both. A seoond possibility is that A,+ecqHon are present on seoond order neurons which suppress 92 glutamat~vokeâdepolarization following glutamate release from the central terminais of pfimfy affefent fibers anaof excitatory intemeurons. In either case, thse postsynapoc medianims w#ild countefact the exaggerated excitatory input from low threshdd afferent fibers during glycinergic disinhibition.

Rie pre-synaptic M8dS d Pdenosine are suggested by the reparts demonstrating the inhibition of glutamate release in vitm using bain slice preparations from the hippocampus and dentate gyn~sd the rat (Fastbown and

Freedlom, 1985; Dolphine and PrBSfWich, 1985). This effect was abolished by & phenyltheophylline (10 prn) suggesting that A,-receptors were involved. While a presynaptic effed on primary afferent tminals has been proposeci (Li and Perl

1994, immunocytochemical studies failed to deted either A, or 4 receptorlike immunoreadivity on these spinal substrates (Choca et al., 1988). Thus, curent evidenoe supports a postsynapüc Kihiôitory Med d adenosine (Delander and Wahl

1988; Li and Perl, 1994), presumably on spinal substrates receiving low- andlor high4hreshoId afFerent input Regardles of the exad mechanism(s), A,ieceptors appear to be strategically located in the spinal cord to modulate sornatosensory input. 93

4.5 Effect of GABA Agonists on Strychnineallodynia: deof GABA,

receptors

The results of the present study indicate STR-allodynia is seledively

madulatecl by spinal GAI34 receptors. Thus, the i.t muscimol dosedependently

Mibited STRallodynia in lightly anestheüzed rats. Such an effet2 on STR-allodynia

has not been shown previously but is consistent with reports of allodynia in rats and

mice following i.t. biwculline, a selective GABAAantagonist (Yaksh, 1989; Onaka

et al., 1996; unpublished results). That this represented a selective spinal site of action is indicated by the la& of the effed on cortical EEG synchrony and the low

ED, values (0.4 - 0.5 pg). In contra&, i-tbaclofen failed to suppress the abnomal resporises to HO in presence of STR, even at doses which produced pronounced motor dysfunction. These data are in agreement with a previous report demonstrating the lack of Med of baclofen against STRallodynia in conscious rats

(Yaksh, 1989) and eledrophysiological data demonstrating the lad< of effect of intra-arterial CGP 35348 (60 mgkg) on the spontaneous activity of hamstring flexor amotoneurons, touch evoked-respanses or touch threshold in the spinal cord of the rat (Sivilotti and Woolf, 1994).

The present results are not surprising given the fact that a) GABA, and glycine receptors independently open chloride channels to hyperpolarize neurons

(Bomiann et al., 1987; Cohen et al., 1989); b) GABAAand glycine receptors are CO- containeci on the same postsynaptic membrane in the spinal dorsal hom (Todd et 94 al., 1996) consistent with the cdocalization of glycine and GABA in the same

presynaptic axons (Todd et al., 1996); and c) glycine immunoreactivity is virtually

reSffided to murons that also exhibit GABA immunoreadivity in the laminae I-III of

the dorsal hm(Todd, 1990; Mitchell et al., 1993). Glycine intemeurons receive

a major monosynaptic input from large diameter myelinated primary afferents in

laminae II and III. These anatomical studies provide evidence for the neuronal

ciraiitry by which glycine and GA8A might nomally modulate noniiociceptive

transmission. fhey also suggest aiat GABA and glycine, cweleased from the same

intemeurons might effed discrete but cornplementary inhibition of relevant spinal

neurons. Indeed, eledrophysiological studies using spinal cord slices from the rat

and cat have shown that GABAA- and glycine-mediated inhibition have a distinct time-course (Game and Lodge, 1975; Baba et al., 1994; Yoshimura and Nishi,

1995). The long and short IPSP, generated in substantia gelatinosa by stimulation of A6-fibers was blocked by bicuculline and STR, respectively (Yoshimura and

Nishi, 1995). The bicuculline-sensitive IPSP was up to 3 times slower than that of

STR-sensitive inhibition reflecting the difierent kinetics of GABAA- and glycine- receptors, or a different time course of the neurotransmitter uptake. It has been suggeçted that the SIR sensitive IPSP may be important in limiting the firing of the neurons and curtailing the amplitude and the ability of the EPSP to generate an adion potential. In contrast, Vie biwailline sensitive IPSP which develops more slowly and has a longer duration may be effective in preventing longer lasting 95 repetitive activation (Yoshimura and Nishi, 1995). While the exad role and signifieam of glycine- and GABA-inhibition remains to be detemineâ, the apparently complementary Meds af glycine and GABA on somatosensory input are

-stent with the inhibition of STR-allodynia by i.t GABA, agonists. Indeed, one would expect glycine or GABA (acting at GAû~-receptors). to attenuate the exaggerated firing evoked by A@-inputin presence of STR.

The role of GAS&-feceptors in the modulation of low-threshold afferent transmission is controversial. The failure of i.t. baclofen to inhibit STR-allodynia in the present expriment suggests that, in absence of nerve injury, spinal GABA,- receptors do not contribute significantly to such modulation. However, i.p. baclofen

(but not muscimol) inhibited touch-evoked allodynia following focal spinal cord ischemia in the rat (Hao et al., 1992); an effect attributed to the activation of

GAB/s,-receptors present on the spinal terminais of AP-fibers. Baciofen also reduced the hypenensitivity of spinal WDR neurons to lowintensity mechanical and electrical stimulation in rats with spinal cord ischemia (Hao et al., 1992). The ability of baclafan to effed such changes in this ischemic mode1 rnay represent an up-regulation of GABb-receptors in response to central injury andlor altered synaptic connections. Although GABA, binding sites have been reported on the tminals of large diameter primary afbrent fibers in lamina IV of normal rats (Price et al., 1987), GABA,-reoeptors may not become fundionally active until after central nerve injury. These data also suggest that the STR mode1 is relevant for 96 investigating the somatosenscxy processing but may not refied upon the altered phamiacology induced by nerve injury.

4.6 €#kt of Mikcetnide on Strychnine4I~ia:rid. of glycine rrc.pfom

Systemic administration of the glycine probug, milacemide dose- dependently inhibited STRallodynia for up to 4 hows. The time course of this effect corresponds to the increase in glycine concentration in the CSF of rats treated with milacemide (400 mglkg, i.p.) (Sernba et ai., 1993). The peak inhibition of STR- allodynia and the maximum increase in [glycine], were achieved 2-3 h after mialcemide adminishüon. That the blockade of STR-allodynia was due to glycine and net milacemide itself was confirmecl by the dependency of this effed on MAO-

B. Thus, pre-treatment with Ideprenyl, but not clorgyline. significantly blocked the anti-allodynic effed of milacemide.

There is strong phamacokinetic and biochemical evidence that senirn milacemide rapidly enters and equilibrates in the CNS cornpartment where it is metabdized by MAO-B (Semba and Patsalos, 1993; Sernba et al., 1993). In fad, over 90% of the milacernide dose is metabolized by MAO-B to glycinamide and finally glycine (Figure 3). The remaining 10% is metabolized by different routes depending on the species. ln support of this metabolic rwte, and consistent with the results of the present study, pretreatment ~ÏthIdeprenyl (2 mgkg, i.p.) 97 inhibited: a) the conversion of milacemide to glycine in vivo; and b) the ability of milacemide to increas8 the sewte threshold induced by hyperbaric oxygen in rats

(Yoâium et al., 1988). Indeed, pretreatment with I-deprenyl(2 mgkg, i.p.) almost completely blod

The ability of glycine, derived from milacemide through the action of MAO-B, to inhibit STR-allodynia is supported by previous work in our laboratory. Glycine or Maine (N,N,N-trimethyl glycine) injected directly into the spinal subaradinoid space of the rat, dose&pndently inhibited STR-allodynia (Sherman and Loomis,

1995). Moreover, i.v. milacemide had no significant effect on nomal nociception

(phasic noxious paw-pinch without STR) in the present study. The persistence of these noxiwsevoked responses indicates that cardiovascular efferent fibers were uM8ded by i.v. milacemide, and that the suppresion of cardiovascular responses ta HO in the STR mode1 represented a ûue antiallodynic effed. These results also support the hypothesis that glycine is a selective modulator of noniioxious somatosensory input in the spinal cord of the rat

In addition to the STR-sensitive glycine receptor, glycine is also known to bind to a STR-insensitive co-agonist site on the NMDA-receptor cornplex. In the case of the latter, glycine augments the excitatory effeds of NMDA (Rosse et al., 98

1991). Considering the semitivity d STRiillodynia to NMDA antagonists (see sedion 4.2), the potdntm0. d NMDAreceptor rnediated depiarization by glycine wouldbewedtoworsentheaikdynia. Weobseniednoevidencedthisinthe present study, but suWe changes in NMDA newotransmission are wilikely to be deteded using cardiavasarlar and maOr endpoints. Such an outane will require more sensitive electrophysidogical techiquerr. Nevertheless, the wmary eRed d milacemide in STRalkdynia was inhibitory. h this regard, 1 has been suggested that the accumulation of unmetabdizeâ milacemide follawing high4ose administration may antagonize the possiMe enhancement of NMDAinediated transmission by glycine (Rosse et al., 1991).

Besides glycine, sysîemic milacemide (400 mgkg, i-p.) is known to increase the concentration of serine and taurine in rat CSF by 20-25% and to decrease alanine concentration by an equal percentage (Semba and Patsalos, 1993). GABA and glutamine concentrations in the CSF were unchanged. The relevance of the decrease in alanine concentration, if any, is rot yet known, but a significant inaease Ki taurine and concentration could be of importance to Uie effect of milacemide in the STR model. Taurine, alanine, and serine are agonists at the

STR-semitive glycine receptor having the following order of potency: glycine >*b alanine > taurine ,> Lalanine > L-serine (Pan and Slaughter, 1995). Taurine is aiso shaimi to act on GA84receptors in the œrebral cortex ( Pan and Slaughter,

1995) but such 84f8d has not yet been shoimi in the spinal axd. The mparatively 99 small increase in serine and taurine concentration (20-25% versus 7580% for glycine), th& lower affinity for the glycine receptor compared to glycine, and the la& of evidence for distinct taurine and serine receptors wggest that the reported changes in &ne and taWin8 amcmtWons are mlikely to play a major role in the anti-allodynic effect of milacemide. Nevertheless, their contribution cannot be excluded at the present time.

Like Ideprenyl, milacemide is also a seledive inhibitor of MAO-8, both in vivo and in vitro (de Varebeke et al., 1989; O'Brien et al., 1994). Ten hours after a single injedion of milacemide (500 mgkg, i.p.) to the rat, the MAO-8 activity in the brain was reduced to approximately 40% of control. This adivity was significantly but incompletely recovered (approxirnately 75% of control) 40 hours after milacemide administration. As a result of this effed on MAO-BI milacemide is self-

Iimiting during long terni therapy. In the present study, only one dose of milacernide was injected per animal to avoid the complications of residwl enzyme inhibition and thus a decrease in glycine production with subsequent doses. The eventual inhibition of MAO-B by milaœmide, and thus the possible progressive decrease in the yield of glycine, may explain the use of such high doses of the prodrug in vivo to achieve a phamacological effect. Clearly, glycine proânigs that lack such enzyme inhibitory adkity would be preferable for chronic use. While the present study investigated the acute effeds of milacemide in th8 STR rnodel, these data indicate the effectiveness of a systemically administered glycine prodwg in 100 att81wafing abnomial sensiüvity to low-threshold afferent input arising in the spinal cord. 10 the extent that the loss of spinal glycinergic modulation might underlie dinical allodynia, this phamacological approach may be worth investigating, particularly in patients who respond poorly to aiment therapy.

4.7 Futum Direcfions

From an experimental point of view, the STR mode1 affords some unique advantages. First, STR-allodynia is reversible, making it possible to test dmgs under normal and abnormal (i.1. strychnine) nociceptive conditions in the same animal. Second, it is highly reproducible (with repeated STR injection) allowing for quantitative dose-response analysis (as indicated by the narrow 95% confidence intervals in Tables Ill-VI, VIII). Finally, 1 is centrally induced, of known cause, and achieved without injury to the peripheral or central nervous system. Viereby ailowing certain conclusions to be made about the site, and to a lesser extent, the mechanisms of dnig However, the experÏments that can be undertaken are limited by the short duration of allodynia (peak effed is approximately 5 min after

STR injection; duration of 20-30 min). It is not possible to test the dironic effeds of glycinergic dysfundion using this mode1 or the long term effediveness of phannacdogical interventions known to inhibit STR-allodynia. Although the phamacology of STR-allodynia mimics, qualitatively, that observed in hurnans, it 101

is not yet kmwhether the loss of glycinergic modulation underlies dinical

aikdynia. MOT^, fwther investigation is required to test mis hypothesis. One

-ble way to induce a more probged yet reversibk glycinergic dysfunction is

to 'km& out' glycine receptocs using antisense digonudeoüdes. Such a

model might prove weM to mppkmmt and extend the infocmation detemiined in

the STR model.

The interadion between glyane and GAûA in modulating low-thrBShdd

alGerent input, and Ihe consequences d simuitaneous dysfunctiori of both inhibitory

kptsin the develapment a allodynia have not yet ben investigated. Given their

kmco-localization in spinal intemeurons, a more realistic mode1 of spinal

Mibitionand allodynia must take into account, the complementary role played

by lhese inhibitory amho acids. This couid ôe adiieved iniüaily by ooadministwing

i.t. GABA and glycimantagonists. H-vw, longer term spinal glycinergic and

GABAergic dysfundion mwld require a different approach, possibly using appropriate antisense digoneuclotides.

4.7 Summary

lnaeased spinal glutamatergic tone has been shomi to yield a state of faciiitated transmission of both low- and high-intensity stimuli (Dougherty et al.,

1992). ûf aie glutamate receptor family, spinal NMDA-receptors are especially 102 important as these receptors mediate long lasting depolarisation induding the phenornenon of 'm'nd up'. Such augmentation appears to be an important spinal mechanism underlying the condition of allodynia and hyperalgesia Wshand

Malmberg, 1994). Moreover, interventions that counterad such exaggerated neurotransmission in the spinal cord should alleviate these sensory events and provide pain relief. The present research dernonstrates that the abnomal responses evoked by input from low-threshold mechanoreceptive afferents in presence of spinal STR are indeed sensitive to NMDA-receptor blockade, to i.t. glycine, GABA- and adenosineagonists and to sub-anesthetic doses of mexiletine.

The evidence obtained in this study supports the following conclusions:

Glycine is a seledive modulator of non-noxious somatosensory input in the

spinal cord of the rat.

In the presence of i.1. STR, low threshold afferent input accesses a spinal

sensitisation mechanism nomally activated by nociceptive fibers.

Mexiletine inhibits both abnomal (STRallodynia) and normal nociception at

a wmmon site, most likely dorsal hom cells receiving convergent Af3- and

C-fiber input.

The reauitment of neuromodulatory systems involving spinal adenosine A,

and GABA, receptors bldsthe abnormal sensory responses of allodynia

following glycinoceptor blockade in the spinal cord.

lncreasing central glycine concentration by means of a systemic glycine 1O3

prodnig (milacemide) is effective in blocking STR-allodynia.

6 The spinal phamacology of STR-allodynia is distinct from that of normal

nocicepion, and is similar to that currently understood in dinical allodynia.

7 STRallodynia is a useM, acute, nmnjucy model for the investigation of this

abnmal pain state. 5 References

Angel A, ûodâ J. and Gray J.D. Fluduating anesthetic state in the rat anesthetized with urethane. J Physiol. (Lond.) 259 (1976) 11-1 2P

Araki 1. lnhibitwy postsynaptic currents and the effeds of GABA on visually identifid sacral parasympathetic preganglionic neurons in neonatal rats. J Neurophysiol. 72 (1994) 2903-1 0

Arena J.M. Poisoning: Toxi~l~gy,symptoms, treatments, 4th Edn. Charles C. Thomas pub. Springfield IL (1979) pp 177-79

Amer S. and Meyerson BA La& of analgesic Meet of opioids on neuropaaiic and idiopathic fonn of pain. Pain 33 (1988) 1 1-23

Ascher P. and Nwak L Electrophysiological studies of NMDA receptors. TINS 10 (1 987) 284-88

Baba H., Yoshimura M.. Nishi S. and Shimoji K Synaptic responses of substantia gelatinosa newons to dorsal column stimulaüon in rat spinal cord in vitro. J Physiol. 478 (1994) 87-99

Backonja M., Arndt G.. Gombar K, Check B. and Zimmermann M. Response of chronic neuropathic pain syndromes to ketarnine: a preliminary study. Pain 56 (1994) 51-57.

Baron R and Saguer M. Postherpetic neuralgia. Are C-nociceptors involved in signalling and maintenance of tactile allodynia? Brain 1 16 (1993) 1477-96

Battagli G. and Rustioni A Coexistance of glutamate and substance P in dorsal root ganglion neurons of the rat and monkey. J Comp New01277 (1988) 302-12

Belfrage M., Sollevi A, Segerdahl M., Sjolund K and Hansson P. Systemic adenosine infusion alleviates spontaneous and stimulus evoked pain in patierts with penpheral neuropathic pain. Anesth. Analg. 81 (1995) 713-7

Bennett G.J. and Xie Y-K A peripheral mononeuropathy in rats that produced disorders of pain sensation like those seen in man. Pain 33 (1988) 87-1 07

Besson J.M. and Chaomb A Peripheral and spinal rnechanism of nociception. Phannacol. Rev. 67 (t 987) 67-1 86 Beyer C., Roberts 1.A and Komisnik B.R. Hyperalgesia induceâ by altered glycinergic adivity at the spinal curd. Life SU. 37 (1985) 87582.

Beyer C., Banas C., Gomora P. and Kornisnik B.R. Prevention of the oonvulsant and hyperalgesic action of strychnine by intrathecal glycine and related amino acids. Pharmacol. Biochern. Behav. 29 (1988) 73-78

Bohlhalter S., Mohler H. and Fritsdiy J-M. lnhibitory neurotransmission in the rat spinal cord: CO-(OCâIizatio11 of glyane and GABAA+w~sat GABAergic synaptic contacts demonstrated by triple immun~uorescencestaining. Brain Res. 642 (1994) 5949

Boivie J., Leijon G. and Johansson I., Central pst stroke pain: Study of the mechanisms through analysis of the sensory abnomialiües. Pain 37 (1989) i 73-85.

MicaJ.J. Causalgia and other sympathetic dystrophies. In J.J. Bonica (Ed.) The Management of Pain. Lea and Febiger Philadelphia PA (1990) 22û-43

Bormann J., Hamill O.?. and Sbvnann B. Mechanisrn of anion pemeation through chnnels gated by glycine and y-aminbutyric -d in muse cultured spinal neurons J Physiol. Lond. 276 (1987) 24346

Bawsher D. Neurogenic pain syndoms and their management. Brit. Med. Bull. 47 (1991) ô4ess

Campbell N. J., Raja N.S., Meyer AR. and Makinnon E.S. Myelinated afferents signal hyperalgesia associateci with nerve injury. Pain 32 (1988) 89-94

Chabal C., Russell L. and Burchiel KJ. The effect of intravenous lidocaine, tocainide and mexiletine on spontaneously active fibers originating in rat sciatic neurornas. Pain 38 (1989) 333-38

Chabal C., Jacobson L., Mariano A, Chaney E. and Britell W.C. The use of oral mexiletine for the treatment of pain after peripheral nerve injury. Anesthesiology 76 (1992) SI3-1 7.

Chaplan S.R., Bach F.W., Shafer S.L. and Yaksh TL. Prolonged alleviation of tactile allodynia by intravenous lidocaine in neuropathic rats. Anesthesiology 83 (1995) 77585

Chocoa J.I., Green RD. and Proudfit H.K Adenosine A, and 4 receptors of the wbstantia gelatinosa are lacated predominantiy on the intrinsic neurons: an autoradiograph study. J Phannacol. Gp.Ther. 247 (1986) 75744

Chrislophe, Kidmw J.R., Nouyen-Bui W.D., Damien C., Chatelain P. and Gillet L glycine in the rat kain. Life Science (1 983) 533

Clkm MA, Ochoa J and Torebjdrk H.E. Chroriie hyperalgesia and Wn warming caused by sensitizd C naciceptors. Brain 112 (1 989) 62147

Coderre T.J., Kah J., Vaccarino LA and Melzadc R Contribution of central newoplasticity to pathoîogical pain: review d dinical ond expimental evidence. Pain 52 (1993) 259-85

Cohen RN., Fain G.L. and Fah M.J. GABA and glycine channels Hi isdated ganglion œlls from the goldlish retina. J Physid. 417 (1989) 5342

Colclough G., Carter A and Adcemian III W.€. Mexiletine for chronic pain* The Lancet 342 (1993) 1-

CobMn C.W., Mo

Curtis D.R., HWi L. and Johnston G.AR. A pharmacological study of the depression d spinal wonsby glycine and related arnino acids. €xp. Brain Res. 6 (1968) 1-18

Curtis D.R., Duggan A W. and Johnston G.AR ïhe specifcity of strychnine as glycKie antagonist in the mammalian spinal cord Exp. Brain Res. 12 (1 971 ) 54745

DavMRA, Graham LT. Jr., ShankRP., Werman R. and AQrison M.H. Changes in mino acid concentrations amated with loss of spinal intemeurms. J Newochem. 14 (1967) 1025-31

Davies J and Watkins J-C. Rde of excitatory amino acid receptm in mono and pdysynaptic excitation in Vie cat spinal cord. Exp. Brain Res. 49 (1983) 280-90

ûavies S.N. md Lodge O. Evidence for involvement d NMDA receptors in wind up af dass II neuonr in the dorsal hom of the rat. Brain Res. 424 (1 987) 402- De Jong R.H., RoMes R and Corbin RW. Central actions of lidocaine: Synaptic transmission. Anesttwsiology 30(1969) 19-23

OejOard A, Pd- P. and KaStNp J. Mexiletine foc treatment of chronic painful diebetic newopathy. The Lamet 29 (1Wû) 9-1 1.

hlander G.E. and Hopkins C.J. Spinal adenosine modulates descending -VU pothways sümulatd by morphine. J Phamiacd. W.Ther. 239 (1986) 80-93

Delander GE. and Wahl J.J. Behwiour Hiduced by putati*ve nociceptive rmwdmmitters is inhibited by adenosine or adenosine analogue ooadministered intrathecaliy. J Pharmacd. Gp.Ther. 246 (1988) 565-70 de VwBbeke P.J., Cavalier R., Oavid43amacle M., De Mey J.ad Yodium M.B.H. Formation of Vie neurotransmitter glycine fmm Vie anticonvulsant milacemide is meâiated by kein monoamine o>a;da#, B. J Newochem. 50(19ûû) 101 1-1 6 de Varebeke P.J., Pauwels G., Buyse C., David-Ramade M., De Mey J., Roba J and Yodium M.B.H. The me(psychobapic agent milacemide is a specific enryme edivated Kihibitor d kah monoamine oxidase. Br. J Neurochem. 53 (1989) 1109- 16

Dewr M., Wall P.D. and Catalan N. Systemic lidoçaine silences edopic neuroma and DRG diçcharge without blcxking nerve amdudion. Pain 48 (1 992) 261-268

Dicicenson AH. and Sullivan AF. Evidence for a role of the NMDA receptor in the frequency dependent potentioation of deep rat dorsal hom nociceptive newons foilamng C-fiber stimulation. Neurophamiacology 26 (1 987) 1235-38

Dickenxwi AH. A cure for wind up: NMDA receptor antagonists as potential analgesics. Trends Pharmaool. Sci. 1l(199û) 307-09

Dickenson AH., Sullivan AF., Stanfa LC. and Mcqwy H.J. DextrmeViorphan and levorphand on dorsal hom nocicaptive newons in the rat. Neuropharmacol. 30 (lssi) 13036

OKing D.M. and 1.L kitrathecal badaen and rnusdmol, but not midazolarn, are mtinociceptive using the rat-fmline model. J Phamiacol. Exp. Ther. 275 (1995) 21 9-27 Dohi S., Kitahata LM., Toyoka H.. Ohtani M., Namiki Aand Taub A An analgesic adian a *ntraverouslyadministered lidocairn on dorsal hom newons responding to norious theimal stimulation. Anesthesidogy 51 (1979) 123-26 minAC. and PrestWiCtt S.A Pminto>M reverses odenosine inhibition of newonal glUBfmat8 releose. Nahm 316 (1985) 14650

Dougherty P.M., Pakœk J., Serkin L.S. snd Willis W.D. The rok dNMDAand non-NMDA PYTjtatory acid ncepton Hi the excitPüori d primate ~6ud~by~,chemicd,Ihemialondeleclricalstimuli. J NeUrosci. 12(1992) 3û2-1

Dukier R NeMMl pbkityand pain follawing peripheral tissue inflammation or meinjury. in M.R. Bond, J.E. Charlaton and C. J. Woolf (Eds) Proc. of the VI th World Congresson Pain Elsevier, Amsterdam (1 991 ) 263-276

Edwards W-T-, Habib F., Burney R.G. and 8egin G. lntravenous lidocairn in the management d variws chronic pain States. Reg. Anesth. 10 (1 985) 14

Eide P.K, Jwum E*, Stuôhaug A, Bremnes J. And Breivik H. Relief d post- herpeoc neuralgk with Ihe Nmelhyl-D-aspartic @d receptor antagonist ketamine: a douMe MM, wssover -son with morphine and placebo. Pain 58 (1994) 347-54

Eide P.K, Stubhaug A and Stenehjern AE. Central dysesthesia pain after traumatic spinal cord injury is dependent on N-rnethyl-Daspartate reœptor adivation. Newoswgery 37 (1995) 108067

Evans R.H. Phamiacdogy d segmental transmission in Vie spinal cord. Progr. NeUrObid. 33 (1 989) 255-79

Evans RH. and Long S.K Primary afferent depolarisation in the rat spinal cord is mediated by pathways utilising NMDA and norrNMDA receptors. Newosci Lett. 100 (1 989) 231 -36

Fagg G.E. md Foster AC. ANn, aad mmtrmmitters and th& pathways in the mammalian central mous system. Newoscience 9 (1983) 701 19 Fastbom J. and Fredhdm B.B. Inhibition of m] glutamate releafe from rat hippocampal sliœs by Lphenylisapropylademine. Ada. Physiol. Scand. 125 (1985) 1213

Fastbom J.. Poot C. Md Fredholm B.B. Antinociceptive effects and spinal distribution d hrro adenosine niceptor agonists aAer intrathecal administration. Ph9mracd. Torocol. 66 (1990) 6912

FM HM., Chepmen V. and Dickemon AH. Spinal local anesthefi: actions on afferentevoked respcmses and winbup of nocicepüve newons in th rat spinal OOCd: awnbinaüon with morphine pfaduœs marked potentiatim of antinociceptiori. PeWi (1992) 3341

Galer S.B. Newopathic pain of pwipheral Win: Advances in phamiacologic treatment Newology 45 supp 9 (1995) S17S25

Game C.J.A anci andodg8 DO.The pharmacology d inhibition of dorsal hom neurons by impulses in myelinated cutaneous affermts in the cat. Exp. Brain Res. 23 (1975) 7584

Goruales R.G. Central pain: diagnosis and treatrnent strategies. Newology 45 #ipp 9 (l99S)Sll-Sl6

Gundlach AL., Dodd PoR*, Grabara C.S.G., Watson W.E.J., Johnston G.A.R., Haper PAW., Demis JAand Healy P.J. Dekit of spinal cdglycinelstrychnine TeCBP(m in Kihertted myodonus of Poil i-bwbrd calves. Science 241 (1988) 1807- 9

Hammond D.L. and Orower E.J. Effects of intrathecally administered THIP, baclofen and muscimd on nociceptive threshdd. Eur. J Pharmacd. 103 (1984) 121-25

Handeimam GE., Nevins ME., Muelfer L.L., Amolde kA and Cordi AA Milacemide, a glycine produg, enhances performance of leaming tasks in mal Pnd anmestic roderrts. PMmaaA. Biochem. Behav. 34 (1989) 82S28

Hao J.X, Xu XJ., Yu Y.X,Seiger A and Wiesenfeld-Hallin Z. Transient spinal cdisaiemia induces ternporary hyprsensitivity of dorsal hom wide dynamic range neurons to myelinated, kR nol unmyelinated. fiber input J Neurophysiology 68 (1992) 384-391 Headley P.M. Md Griliner S. Excitatory mino acids and synaptic transmission: Tbevividence fw physidogicial fiwrdion. TiPS 11 (1 990) 205-1 1

HeOarty A cnd POTfenqr RK Phemirica(henipy of wopathic pain. Seminare in Neuoiogy 14 (1 994) 21 324

Henry J.L. ConcepCI d pain seiisatïcm Md its modulation. J Rheumatology 16 .uppi 19 (1989) 1û4-12

Houtkooper MA. Van Oorshcot CAE-H., Rentmeester T.W., Hopperier P.J.E.A and ûniceiira C- Double #Yd dudy d milacemide in hospitalised therapy resistant patients with epi*lepsy. Epilepsia 27 (19û6) 25562

Iginremeàe LN., Beatch G.N.,Walker M.JA and McErlane KM. Tissue distribution of mexileüne errantiomers in rats. Xenobiotica 21 (1 991 ) 1153-58

IgwemeUe LN.. BeatCti G.N.. McErlane KM. and Walker M.J.A Mexiletine's antifibrillatory Pdions are lirnited by the ocairrmœ of mvulsions in consciow animais. Ew. J. Phamracol. 210 (1992) 271-77

Jadad A.R., Candi O., Glynn C.J., Mooce R.A and McQuay H.J. Morphine responsiveness d chronk pain-Double blinâ randomized cross over study with patientumtroiled anaigesia. The Lancet 339 (1992) 1367-71

Jockers R., Linder M.€., Hohmgger M.. Nanoff C., Bertin B., Strosberg AD., Manilio S. and Freissmth M. SpecieJ difFerence in the G-protein seledivity of Vie human and bovine A, adenosine receptor. J Bioi. Chem. 269 (1994) 3207744

Kandel E. and Schwarb: J.H. Princi*plesof neural science 3 rd Eds. Elsevier pp 301-343

KangrOa 1. and-Randic M. OutRaw of endogenous aspartate and glutamate from the rat spinal dorsal hom in vitm by activation of low- and high- threshold primary afFerent fibres. Modulation by wioids. Brain Res. 553 (1991 ) 347-52

Kangrga I., Jiang M. and Randic M. Adions of (-)-baclofen on rat dorsal hom neurons. Brain Res. 562 (1 991) 265-75

Kapadia S.E. Md LaMotte C.C. DeMerentation4nduced alterations in the rat dorsal hom. 1. Cmm d peripheral nerve iniwy vs hizotomy effects ori postsynapüc a7d glid pm.J Cm. Mi. 266 (1987) 183-97 Karlsten R., Claes P., Hide 1. and Daly J.W. The antinociceptive effet3 of Kkathecally administered adenosine analogs in miœ correlates with the affin@ for the A,9denosine reœptw. Neufosci. Lett 121 (1991) 267-70 KPrtstm R snd Gordh T. Jr. An ~~sdenosine agaWst abolishes allodynia eküed by vikotion and touch after inttathecat *njedian. Anesth. Analg. 80 (1995) 844-7

KBstrclp J., Angelo H.R. PetmP. ûejgard A And Hilsted J. TreatmenQ of dumic diabetic neuropathy with intravenowr lidocaine Kifusion. Br. Med. J 292 (1986) 173

King A€. anci Lopez-Garcia J.A Excitatory amino adreceptor-meâiated neurotransmission from cutaneous afFwmts in rat dorsal hom in vitm. J Physiol 472 (l993) 44-57

Ibltrenkrg M., Lundberg LE. and Torebjdrk HE. ûynamic and sWic components d mechnical hyperalgesia in hurnan hairy skin. Pain 51 (1 992) 207-1 9 hietzny F., Perl ER., Trevino T., Light A and Hensel H. Sensory expriemes in man evoked by intraneural elearical stimulation of intact cutaneous afferent fWs. Exp. main Res. 42 (1981 ) 21 9-22

Kupers R.C., Konings H., Adriaensen H. and Gybels J. Morphine differentially aff8dS the seosory and affective pain ratings in neurogenic and idiopathic fmsof pain. Pain 47 (1991 ) 5-12 lm S.H., KaywV.. Desmeules J. and Guilbaud G. Differential adion of morphine and various opioid agonists on themial allodynia and hyperalgesia in mononewopathic rats. Pain 57 (1994) 2334û

Lee Y-W and Yaicsh T.L Phamiaoology of the spinal adenosine receptm which mediaf8s the anüallodynic adion of intrathecal adenosine agonists. J Phannacol. Exp. Ther. 277 (1996) 16424

Li J. and Perl ER Adenosine inhibition of synaptic transmission in the substantia gelatinosa. J Neuropharmacol. 72 (1 994) 161 1-21

Linodn D.W., Hentzen KI Hin T., Van der Schoot P., Clarke G. and Summerlee A.J.S. Sleep, a prerequisite for reflex rnilk ejedion in rats. W.Brain Res. 38 (1980) 15142 Lindsbom P. and Lindblom U. Anaigesic adion tocainide in trigeminal neuralgia. Pain 28 (1 987) 4550

Lindblom U. and VerriIl0 R.T. Sensory fundions in chronic newalgia. J Newd. Neurosur. Psychiat. 42 (1979) 422-35

Marala R.B. and Mustsfe S.J. DMevidence for the cwpling d A,sderiosine reeeptor to stimiatory guanine -ndiwein in bovine bain striatum. J Phannacol. Gp.Ther. 266 (1993) 294300

Maxwell D.J., Chrisie W.M., BrmAG., Ott-n 0.P and StmMathisen J. #enofiedhairfollideafferentbol\itonsinthespKialcordafthecat aree~ichedwith Lgkdamate-like immunoreacüvity. Brain Res. 606 (1993) 15641

Mayer ML. and Westbrook G.L. The physiology of excitatory arnino acids in the vertehate central moussystem. Progr. Neuroôiol. 28 (1987) 197-276

McGowan M.K and HammO.L. lntrathecal GABA, antagonists attenuate the anündception producecl by microinjeaion of lglutamate into the ventromedial medulla of Vie rat. Brain Res. 607 (1 993) 3946

McLachlan E.M.. Janig W., Drevor M. and Michaelis M. Peripheral meinjuy trïgps noradrmergic sprouting within the dorsal root ganglia. Nature 363 (1993) 54346

Melzack R. And Wall P.D. Pain mechanims: a new theory. Science 150 (1965) 971-80

Mendel1 LM. Physidogical properdies of myelinated fibre projection tu the spinal cord. Exp. Neurol. 16 (1966) 316-332

Merskey H. Classification of dironic pain, Pain suppi. 3 (1986) S2629, S216-21

Meyer, R.A, Campbell, J.N. and Raja, S.N.. Peripheral neural mechanisms of cutaneous hyperalgesia. ln: H.L. Fields. R Dubner and P. Cervero (Eds.), Advances in Pain Research and ilmrapyavol. 9, Raven Press, New York, 1985, pp. 53-7 1. Mikie, B., Duggan, S., Jhamandas, K and Loomis, C.W., Innoarous hair defiedian 8VOkes e noaoeptivdike advation dcatBChd oxidatim in the rat loau coeruleus fdMng intnthecel strychnine: a bidwmical index of allodynia uring in uh vdtmmetry. kain Res. 718 (lm)196-2û2.

Mimi T., W. R, Horiguchi S., Ito S., Hyado M. and Hayaishi O. Allodynia mddby inbathecal administration of proshgakidin F, ta corisciaw mice. Pain Sû (1992) 223-9

Minami T., Nishihare 1.. Uda R, Ito S., Hyodo M. and Hayaishi 0. Jnvdvement d glusamate mœ@m in alkdynia induced by prostaglandins E, and Fa imedinto anscious miœ. Pain 57 (1 994) 22!%31

Mitchell K, Spike R.C. and Todd AJ. An immmocytochemical study of glycine receptom and GABA Yi lamina MI1 of rat spinal dorsal hom. J Neurosci. 13 (1993) 237141

Monis R Reçponses d spinal dorsal hmneurons evoked by myelinated primary afferent stimulation are blod

Nagy J.I. end DaMcma P.€. Anatomid and cytoctiemicai relationship of aderiosine -inin9 primary Merent newons in the rat. Neurod. 15 (1985) 799813

Nagy I. and Wodf C.J. Lignocaine selectively reduces C fibre-evoked neuronal Pdivity Ki rat spinal cdin vilrr, by &masihg N.rnettiyWepartate and neuroldnin recepZordiated post-symptk depolarkation; implication for the development of novel centrally acting analgesics. Pain 64 (1 996) 59-70

Ngsstrdm J., Karlsson U. and Post C. Antinaciœptive adions of difkrent ciasses of excitatory mino acid reœptar antagonists in miœ. Ew. J. Pharmacd. 21 2 (1992) 21 -9

Noordenbos W. and Wall P.D. Implications of the failure of nerve reSBCfion and graf't to cwe chronic pain produced by nerve lesions. J Newolo. Neurosurg. Psychiat. 44 (1981 ) 106873

O'Brien E.M., Tipton KF., McCrodden J.M. and Youdim M.B.H. The interactions d milacemide with monoamine oxidase. Biochem. Pharmacol. 47 (1994) 617-23 OMko M., Minami T., Nishihara 1. and ho S. Involvement of glutamete recepton in strychnine- and biawxrllineinduoed allodynia in amci~s miœ. Anesthesidogy 84 (1 996) 121 5-22

Pan Z H. and SlaugMer M.M. Comparison of the adim of glycine and related amho acids on the isdated third order neumns from the tiger salamander retins. Newosci. 64 (1995) 153-64

PerschBk H. Md Cuenod M. In vNo fekmm d enc@mcm gkitamate eopertate in the rat striatum dwrig stimulation d cortex Neuosciericr, 35 (1990) 28347

Poon A end Sawyndc J. AntYioQoe9om by adenosine analogs and an adenosine kinase *nhiior: dependence on fomaline concentration. Eur. J. Phannacol. 286 (1 995) 17784

Porteroy RK, Fdey KM., and lmsi C.E. The nature of opioid responsiveness md its Hnplicaücm fw neuropathic pain: New hypotheses derived from studies of opioid infusions. Pain 43 (1990) 273-86.

Price D.D., Hayes R.L., Ruda M.A ad Dubner R Spatial and temporal onn nations c# input to q&dhhüc tract murons and their relation to somatic sensation. J Newophysiol. 41 (1978) 93347

Price G.W., Kelly J.S., Bowery N.G. The location of GABh receptor binding sites Hi mammalian spinal cord. Synapse l(1987) 530%

Pria D.D., Bemett J.G. Pnd Mi A Psychophysiobgicalobs8wations on patients with neuropathic pain relieved by a sympathetic block. Pain 36 (1989) 273-288.

Priœ D.D., Long S. and Huitt C. SeMary testing of pathophysiological medianism of pain in patients with reflex sympathetic dystrophy. Pain 49 (7 992) 163-1 73

Price D.D., Mao J. and Mayer 01. Central neural mechanisms of maland ekwrmal pain states; In: H.L. Fields and J-C. Liebeskind (Eds.) Pharmacological spproaches to the treatmnt of chronic pain: New concepts and critical issues Raven Press, New York (1994a) 61-84

Priœ D.D.. Mao J., Frenk H. and Mayer D.J. The NMDA receptor antagmist ~~selectively reduces temporal mmatim of second pain in man Pain 59 (1 994) 16574 Rizzdi AA Distribution of glutamic acid, aspartic y-aminokrtync add and glycine in six areas of cat spinal cord before ancl aAer transedion. main Res. 11 (1 968) 11 -1 8

Roberts L.A Bayer C. and Komisaruk 6.R Nociœpüve remto aftered GABAmgic -vity at the spinal tord. Life Sci. 39 (19û6) 1667-74

-RB., SchuuabzB-L, LeigMon M.P.. DavisRE anci OuetschS.I. AnNMDA intervenüon mtegy in schuophnnia with low dosa milacmide. Clin. m.14(1991)268-72 RGwbotham M.C.. Reismc-keller LA and Fields H.A Both Mtravemus lidocaine andmorphinereducepaindposoierpeoc~.Neurology41 (1991) 1024-28

R- R- C.M. Chronic pain: from theory to pradical managmmt Neurdogy 45 (1995) SM1O

Rudomin P., JWneciez I., Quevedo J. and Solodkin M. Pharmacological analysis of inhibition produced by lastorder intemiediate nucleus intemm rnediating nonreciprocal inhibition of motoneurons in cat spinal cd. J Neurophysiol. 63 (1990) 147-

Saletu B., GNnberger J. and Llmyer1. Acute and subaane cenbal mous system Meds d milacemide in elderly people. Oouble Mind placebo controlled quantitrdive -lography and psychometric investigation. Arc. Gerentol. Geriat 5 (1986) 165-81

Wer M.W. and Henry J.L. Evidence mat adenosine mediates üie depression of spinal dorsal hom newons vduced by mpheral vikation h the cat Neuroscience 22 (1987) 631-50

Sawyndc J., Sweeney M. 1. and White T.D. Classification of adenosine receptors mediaüng in the rat spinal cord. Br. J. Phamcol. 88 (1986) 923- 30

Sawymk J. and Sweeney MI. The decd purines in nociception. Neuroscienoe 32(1989) 55749

Schouenborg J. and Sjolund B.H. First order nociceptive synapses in rat dorsal hmare blocked by an amino antagonist. Brain Res. 376 (1 986) 394-98 Schwartr B.L., Hashtroudi S., Herüng R.1 and ûeutsch S.I. The effect of milacemide on item and source memory. Clin. Newophannacol. 15 (1992) 1144 9

Segerdahl M., EkMnn A And SdW A The influence of adenosine, ketamine and morphine on expfimeriblly induced isdiemic pain in healthy volunteers. heth. Anaia. 79 (1994) 787-91

Swnbe J: Pnd PmP.N. Milecemide effects on the temporal iriterlrelatimship of amino acids snd monoamine metabdites in rat cerekospinal fluid. Eur. J. PhPTmacol. 230 (1993) 321-26

Sembe J., CUR~G. and Patsaloé P.N. Antiepileptic chg phanmakinetics and neuropahamtics in individual rats by repetitive withdrawal of Mcmâ and cerebrospinal fluid: milacemide. Br. J Phannacdogy 1O8 (1 993) 11 1 7-24

Sherman S.E. and Loomis C.W. Morphine insensitive allodynia is pro- by intrathecal strychnine in the ligMly anesthetized rat Pain 56 (1994) 17-29.

Shemran,S.E. Pnd Loomis, C.W., Strychninedependent allodynia in the urethane- anesthetized rat is segmeritalfy disbikffed and prevented by intrathecal glycine and betaine, Can. J. Physiol. Pharmacol. 73 (1995) 16-1 705.

Sherman S.E. Md Loomis C.W. Strychnine-sensitive modulatiori is seledive for non-noxiow somatosensory input in the spinal cord. Pain 66 (1 996) 321 -30

Shemian S.E.. LW L and Dosbrovsicy J.O. Spinal strychnine alters receptive fields and sensory modalities of rat thalamic wons. Proceedings of Vie 8th World Congress on P.in Vancwver Canada lASP Press (1 996) p 245

Shir Y. and Seltrer Z A-fiber mediates meehanical hyperBSfheSia and allodynia and C-fibero mediate hyperalgesia in a new model a causalgifm pain disorder in rats. NeUrosci. Lett. 11 5 (1990) 6267

Silvilotü L and Ni* A GABA receptor mechania in the central nervous system. Pmg NewobioL 36 (1 991 ) 35-92

Sivilotti L. and WOOUC.J. Tb8 contribution of GABA,, and glycine receptors to central sensitizatim: Disinhibition and tOUCh-8voked allodynia in the spinal corci. J Neuophysid. 72 (1 ) (1994) 16979

Skilling S.R., Srnullin D.H., Beitz A J. and Larson AA Extracellular mino add mncéntrations in the dorsal spinal axd of freely moving rats follcwing veratrKline and noQceptive sümulation. J N~uT~.51 (1988) 127-32

Sollevi Al M.. Lvideberg T-, M., end Hanssori P. Clinicai note: Systmic adenoeine infusion: a new beatment modality to alleviate neuropathic pain. Pain 61 (1995) 1558

Sosnowski M. and Yaksh T.L. Rde of adenosine receptors in modulating the hypereslhesiP produoed by SPiiipI mmTeCBF)tw antagonim. Anesth. Analg. 69 (1989) 587-92

Steinhaus JE. and Howland O.€. lntravenously administered lidocaine as a supplement to nitrous oxide-thiobarbiturate anesmia. Anesth. alg. 37 (1958) 40-46

Stane T.W. Physiokg-l far adenosine and adenosine S-triphosphate in the necvous system Newoscienoe 6 (1981 ) 52-

mimoto T., BmWG.J. and Kajander KC. Strychnine enhanced trans-synaptic degeneration ddonal hom neurons h rats with an e>cperimental painful peripheral neuropathy Neurosü. Lett. 98 (1989) 13943

Taliarida R.J. Md MW R.B. 1987. Manual of Phamacological Calculations. Springer-Verlag New York pp 10.30.

Tanelian D.L. and Maclver M.8. Analgesic on cent rations of lidocairn suppress Tsrksr R Pain resu#ing from cenbal moussystm pathology (central pain). ln: J. 6oniua (Ed-), The Management d Pain. Vd 1, Lea and Feôiger, Philadelphia. PA (1990) 26486.

Tasker RR, DeCarvaho G.T.C. and Ddan E.J. lntradable pain of spinal cord aigin: cimai dimplm f~r W~~UV. J NBUTOSU~~. 77 (1 992) 373-78 Todd A J. An electton microocope study of glycine-like imrnuncweactivity in lamina Mil of the spiml dorsal hom of the rat Neurosci811~839 (1990) 387394

Todd A J. and Sullivan AC. LigM rnicroscopic study of the coexistence of GABA- like and glycine-lke imrtuw8adivities in the spinal cord of the rat. J Comp. Newd. 296 (1990) 496-505

Todd AJ., Watt C., Spke R.C. and Sieghart W. Colocalization of GABA, glycine amdtheüreceptaroatsy~psesintherstspinaluxd.JNeumsd. 16(1996)97442

Tw&jddc H.E. and Ochoa J.L. Specific sensations evoked by activity in single identifai sensory wiits in man. Ada. Physiol. Scand. 110 (1980) 445-7

Treede RD., Raja S.N., Davis KD., Meyer R.A and Campbell J.N. Evidence that alphagdrenergic reœptors mediate sympathetically maintained pain. In: M.R. Bond, J.E. CMmand C.J. WH(Eds.), Pmceedings of the Vlth World Congress on Pain, Elsevier, Amsterdam, 1991, pp. 3T7-382.

Triggs W.J. and Beric A Sensory abnomialities and dysesthesia in the anterior spiMi artery syndrome. Brain 11 5 (1992) 189-1 98.

Tureen LL Eff'd expwkmW temporay vascular ocdusion on the spinal cord. Ardr. Neuro. Psychiat. 35 (1936) 789807

Man, L., Thompson, S. W.N., and Dfay, A, Modulation of spinal excitability: CO- operation between neurokinin and excitatory amino acid neurotransmitters. TINS 17 (1994) 43238.

Ven Calkec P., Muller M. and Hamprecht B. Adenosine regulates via hvo different typer d mœptom the acamulation d C-AhRP in wbed kain cella J Newochem. 33 (1979) 999-1 005 Van den Pol AN. and Grocs T. Glycine and glycine receptor immunoreadivity in bain and spinal eord. J Newosci. 8 (198%) 472-92

Van Dors#K W., Barris D., Cordi A Md Roba J. Antiooriwlsant adivity of milecemide. kch Phprmooodyn. 266 (1963) 234

Wall P.D. and Devar M. Sensory Mwmt nipulses originate frm dwsol root -lia as dlas the pefiphery in norme1 ond nerve injured rats. Pain 17 (1983) 321-39

Wonaka A, Shiotani Y, KiymHl Matsuyama TI Kamada T, Shoisaka S. end Tohyama M. Gkrtamatdike imnvKweadive strwhms in primary sensory murons in the rat detected by a specific antisenim against glutamate. Exp. Brain Res. 65 (1987) 691-94

Watkings J.C., P. and Ha#re 1. StmcUe adivity relationships in the development d excitatory arnim acid receptor agonists and cornpetitive antagonists. Trerids Pharmaool. Sa. 11 (1990) 2533

Westlund KN. Weill D.L., Paîterson J.T. and Coggshall RE Aspartate immunoreadimmunoreadive axons in namal rat L4 dorsd roots. Brain Res. 489 (1989) 347-51

Wedknd KN. MGNeill DL, and Coggshall RE. Glutamate irnrnunoreadivity in rat dorsal root axons. Neurosci. Lett 96 (1990) 13-1 7

White J.C. and Sweet W.H. Pain and the neurosurgeon. A forty year experienoe. Thomas Sprhgfïild 111 (1 969)

White W.F. and Heller AH. Glycine receptor aiteration in the mutant mouse spastic. Nature 298 (1 982) 655-657

Wlcockson W.S., Chung J.M., Hori Y., Lee KH. and Willis W.D. Effects of iontophoretically reieased peptides on primate spinothalamic tract cells. J. Newosci. 4 (1984) 74150

Wilcax G.L., Kitto KF. and Loomis C.W. Spinal neural responses to innoarouo stimulation Wnd rrp" in the presence of i.t strychnine. Society For Neumd. 208.10 (1996) 511

Wilson P.R. and Yakh T.L Badobm is antinociœptive in spinal intrathecal space of animals. Eur. J Phannacol. 51 (1978) 323-30 Wwlh FORand Rutherfd RB. A Qvilian expwienœ with causalgia. Arc Surg 100 (1 970) 633-38

Wodf C.J. and Thompson S.W.N. The indudion and maintenance of central rensitiration is dependent on NMDA receptor adivaüon: Implication for the treatment of post-injury pain hyperse~iüvitydates. Pain 44 (1991) 293-99.

Woolf C.J.snd Wi8S8nfeld-Hallin Z me sys&mic administration of local rieothelicr proaices a seledive depression d C-alferent fibers evoked edivity in the spinal cord. Pain 23 (1 985) 361-74

VboW C.J. and Wall P.D. The relative eff8diveness of C#mery Mwmt fiben d Mmt OrÏgins in evoking a proknged facilitation d the fiexor reflex in the rat J NeUrosci. 6 (1986) 1433-42

Wodf C-JO,Shortland P. and Coggeshall R.E. Peripheral meinjwy tnggers central sprouting of myelinated afferents. Nature 355 (1992) 7578

Woo# C. The pathophysiology d penpheral neuropathic pairwbnomial peripheral input and abnomial central proœssing. Acta. Newochir. 58 (1993) 12530

Xu, X-J., Hao J.-X, Seiger, A,Amer, S., Lindblom, U. and WiesenfeldHallin, Z Syot8miC mxiletine relieves chronic allodynia like symptoms in rats with ischemic spiripl cord injury. Anesth. Analg. 74 (1992) 64-52.

Yaksh T.L. Behavioral and autonomie correlates of tadile evoked alidynia pIoduced by spinal glycine inhibition: Effeds of modulatory receptor systems and excitatory mino acid antagonists. Pain 37 (1 989) 11 1-23.

Yokoh T.L. The central phamracdogy of pain States. IASP refresher cowse syllabus (1993) 13-30

Y- T.1and Malmberg AB. Central phamiacology of nociceptive transmission, Teidbook d Pain. 3rd Ed. Edited by Wall P.D., Edinbwgh M.R. and Livingstone C. 1994, 165-200

Yamamoto T. and Yaksh T.1 Spinal pharmacology of thermal hyperesthesia induced by *mpleteligation of sciatic nerve 1. Opioid and nonopioid receptors. kiesthesidogy 75 (1991 ) 817-26

Yodiurn M.B.H., Kerem O. and Duvdevani Y. The glycine prdrug, milaoemiôe, increases the seizure threshold due to hyperbaric oxygen: Prevention by 1- deprenyl. Eur. J Phramacol. 150 (1988) 38184

Yokota T., Nishikawa N. and Nishikawa Y. Effects of strychnine upon different cISLP+Ln Of trigeminal suknideus caudalis neurons. Brain Res. 168 (1 979) 430.34

Young AB. and Synder S.H. Strychnine binding in spinal md membranes assoc@ed with the synapüc glycine feœptas: Cooperaüvity of glycine interactions. Mol. Pharmacol. 10 (1974) 790609

Yoshimura M. and Nishi S. Pn'rnary afferent evoked glycine and GABA-mediated lPSPs in substantia gelatirosa newons in the rat spinal c~din vivo. J Physiol. 482 (1 995) 29-38

Zarbin MA, Warnsley J.K and Kuhar M.J. Glycine receptors: Light microsapic autoradiograhic localization with ['HI strychnine. J Neurosci.13 (1981) 53247

ZieglgBnsberger W. and Hertr A Changes of cutaneous receptive fields of spino- cervical tract neurons and other dorsal horn neurons by microeledrophoreticaIly administered amino acids Exp. Brain Res. 13 (1971 ) 11 1 -26 TEST TARGET (QA-3)

APPLIED il IWGE. lnc -fi 1653 East Main Street ---2 Rochester, NY 14609 USA ------Phone: 71 6/48~-03OO ---- Fax: 71 6/2â8-5989