Local nitric oxide synthase activity in a model of neuropathic pain

European Journal of Neuroscience, Vol. 10, pp. 1846–1855, 1998 © European Neuroscience Association

Local nitric oxide synthase activity in a model ofneuropathic pain

Dan Levy and Douglas. W. ZochodneDepartment of Clinical Neurosciences and the Neurosciences Research Group, 182A Heritage Medical Research Building,University of Calgary, 3330 Hospital Drive NW, Calgary, Alberta T2N 4N1, Canada

Keywords: allodynia, chronic constriction injury, hyperalgesia, nerve blood flow, NO, NOS, peripheral nerve

Abstract

A local inflammatory reaction may play an important role in the development of neuropathic pain followingperipheral nerve injury. One important participant in the inflammatory response of injured peripheral nervemay be nitric oxide (NO). In this work, we examined physiological and morphological evidence for nitricoxide synthase (NOS) activation in the chronic constriction injury model of neuropathic pain in rats.Physiological evidence of local NO action was provided by studying NO-mediated changes in local bloodflow associated with the injury site. Immunohistochemistry was used to localize isoforms of NOS that mightgenerate NO.

Sciatic nerve injury associated with behavioural evidence of neuropathic pain had substantial rises in localblood flow. The NOS inhibitor NG-nitro-L-arginine methyl ester (L-NAME), but not NG-nitro-D-arginine methylester (D-NAME), reversed the hyperaemia in a dose-dependent fashion proximal to the constriction at 48 hand distally at 14 days post-operation when applied systemically or topically. Aminoguanidine, a NOSinhibitor with relatively greater selectivity for the inducible NOS (iNOS) isoform, reversed nerve hyperaemiadistal to the constriction only at 14 days. NOS-like immunoreactivity of the neuronal and endothelial isoformswas identified just proximal to the constriction at 48 h. iNOS-like immunoreactivity was observed at 7 and14 days at the constriction and distal sites, respectively.

This work provides evidence for local NOS expression and NO action in the chronic constriction injurymodel of neuropathic pain. NO has local physiological actions that include vasodilatation of microvesselsand that may be important in the development of pain sensitivity.

Introduction

Nitric oxide (NO) is a free radical toxic gas that has beenidentified as an active messenger in many biological systems(Bredt & Snyder, 1994). NO is biosynthesized from the essentialamino acid L-arginine through an oxidation reaction by a familyof nitric oxide synthases (NOS). Three distinct NOS isoformsinclude the constitutive Ca21-dependent neuronal (nNOS) andendothelial NOS (ecNOS), and the Ca21-independent inducible orimmunological NOS (iNOS) [Moncadaet al., 1991]. Followingtissue injury and inflammation, increased activity of ecNOS inendothelial cells and induction of iNOS in macrophages generatelocal increases in NO levels. Augmentation in local NO levels inturn results in vasodilatation with increased blood flow, inhibitionof platelet aggregation, plasma extravasation, cytotoxicity andmediation of cytokine-dependent processes (Moilanen & Vapaat-alo, 1995).

NO has been linked to the development of nociception. Recently,a role for NO in spinal nociceptive processing has been postulatedto occur following prolonged chemical nociception using theformalin test (Rocheet al., 1996) or peripheral injection ofcarrageenan (Melleret al., 1994). NO may mediate some of the

Correspondence: D. W. Zochodne, as above. E-mail: [email protected]

Received 22 August 1997, revised 21 December 1997, accepted 14 January 1998

neuropathic pain syndromes following peripheral nerve injuries(Meller et al., 1992). In addition to its spinal role, NO was alsoproposed to act locally and enhance nociception following localformalin injection (Kawabataet al., 1994). Peripheral nerve injurieshave also been shown to induce an increase in NOS expressionin DRG cells (Vergeet al., 1992), and the subsequent release ofNO was suggested to mediate the ongoing ectopic discharges seenin the DRG following peripheral axotomy (Wiesenfeld-Hallinet al.,1993). Lastly, NO applied locally in humans can elicit pain(Holthusen & Arndt, 1995).

The development of an animal model of partial nerve injurywith behavioural evidence for neuropathic pain (Bennett & Xie,1988) has provided a reliable and reproducible tool for investigatingthe mechanisms underlying neuropathic pain in humans. Thismodel is produced by chronic constriction of the sciatic nervewith four chromic gut sutures (43 4-0) rendering the nerveischemic (Myerset al., 1993) and inducing particular degenerationof myelinated fibres with a variable level of unmyelinated C fibreloss depending on the tightness of the constrictions (Basbaumet al., 1991 ; Mungeret al., 1992). Local periaxonal inflammation,

Local nitride oxide synthase activity 1847

in reaction to the suture material, and the formation of a localgranuloma may be critical features of this model (Clatworthyet al.,1995). The resulting nociceptive changes probably involve acomplex relationship between neuropeptides, e.g. CGRP, themicrocirculation and NO-NOS (Zochodne & Ho, 1992 ; Zochodneet al., 1995 ; Thomaset al., 1996).

In this study, we investigated local NOS-NO action in the CCImodel of neuropathic pain in the rat. By taking advantage of thepotent vasodilator characteristics of NO, we examined physiologicalevidence for NO presence by studying the temporal and spatialchanges in local nerve blood flow and their reversal by inhibitionof NOS. In addition, we studied immunoreactivity of the NOSisoforms responsible for local elaboration of NO: ecNOS, nNOSand iNOS.

Preliminary results of this work have been published in anabstract form (Levy & Zochodne, 1996).

Materials and methods

The studies used Sprague–Dawley male rats of initial weight 250–300 g (n 5 136), from the University of Calgary Animal Care Facility.Prior to surgery, rats were housed in colony cages (two–three in acage) with food and waterad libitum. Following surgery, ratswere housed individually on sawdust-covered plastic cages. Theexperimental protocols were reviewed and approved by the Universityof Calgary animal care committee following the Canadian Councilof Animal Care guidelines.

Chronic constriction injury

The chronic constriction injury (CCI) was created as previouslydescribed (Bennett & Xie, 1988). Briefly, animals were anaesthetizedwith sodium pentobarbital (65 mg/kg i.p.) and the right commonsciatic nerve was exposed at the mid-thigh level. Under a dissectingmicroscope (3 40), four ligatures (4-0 chromic gut, Ethicon, Peterbor-ough, ONT, Canada), presoaked in saline, were loosely tied in 0.5–1 mm intervals around the nerve just proximal to the trifurcation.The wound was then closed in layers using silk sutures (4-0 silk,Ethicon, Peterborough, ONT). For the behavioural tests, laser Dopplerblood flow measurements and immunohistochemical studies, thecontralateral sciatic nerve underwent sham surgery with exposure ofthe sciatic nerve without ligation or injury. For the hydrogen clearanceblood flow studies, separate animals were used as sham-exposedcontrols.

Nociceptive behavioural tests

Fourteen rats were assessed for development of chronic neuropathicpain at 48 h, 7 days and 14 days following the CCI operation. Thedevelopment of thermal hyperalgesia was assessed by measuringchanges in the latency and duration of a paw-withdrawal response tobrief radiant heat aimed to the ventral aspect of the paw using a methodpreviously described (Hargreaveset al., 1988). The development ofmechanical (tactile) allodynia was assessed using a set of calibratedvon Frey hairs (Stoelting Co., Chicago, IL, USA) with increasingstiffness (0.4–29 g/force) that exert different forces when pushedagainst the tested paw.

NO-mediated changes in local blood flow

LDF studies

A laser Doppler flowmetry (LDF) probe was used to assess localNOS-NO action in the epineurial vascular plexus of CCI nerves. This

© 1998 European Neuroscience Association,European Journal of Neuroscience, 10, 1846–1855

method detects changes in local red blood cell (RBC) flux byanalysing the frequency shift produced by the interaction betweenphotons and moving red blood cells (Shephard, 1990). Alterations inRBC flux are detected by a fibre-optic probe placed over the tissuethat detects changes in the extrinsic circulation directly beneath it.

For LDF studies, rats were anaesthetized by an initial injection ofpentobarbital (65 mg/kg i.p.) with additional doses (20 mg/kg i.p.supplemented every 2 h) as required. The left carotid artery wascannulated (PE50) for injections of fluids (1% heparin/saline) andgauging mean arterial pressure (MAP) with a transducer (P23 ID,Gould, Osnard, CA, USA). MAP was monitored and recorded on apolygraph (Model 7D Polygraph Grass instruments, Quincy, MA,USA) with animals having MAP, 90 mmHg not included. Woundson both sides were reopened, retracted and the sciatic nerves exposed.The retracted skin and muscle edges served to form a pool filled witheither mineral oil or saline (for studies with topical application ofagents) maintained at 376 1 °C with a temperature probe and controlfeedback unit (TCAT-1, Bailey, Saddle Brook, NJ, USA). TheLDF probe (1.0-mm-diameter tip) connected to a perfusion monitor(Perimed PF3, Sweden), was placed over the area of interest using amagnetic-based micromanipulator attached to a metal plate to preventmovement. The fibre-optic tip was placed in the liquid pool, perpendic-ular to the nerve, in order to generate the maximum signal from thedetection of RBC flux. Following a complete shut-off of all lightsources, including the heating lamp, baseline RBC flux was measured.For both sides studied, eight–10 different measurements were taken,starting on the constricted side, in an area of 0.5–5 mm distal andproximal to the constriction site, without removing the sutures. In sixadditional rats, the middle sutures were removed to assess RBC fluxin the granuloma area. In order to compare the constricted with thesham-operated side, measurements were taken from related sites onthe sham-operated side. For each site, a mean perfusion readingwas calculated. To evaluate local NO action on the sciatic nervemicrocirculation, baseline readings from the proximal and distal siteswere compared to a series of post-treatment readings obtainedfollowing systemic injections of NG-nitro-L-arginine methyl ester (L-NAME, 0.1, 1, 10 mg/kg i.p.,n 5 21), a broad spectrum NOSinhibitor, NG-nitro-D-arginine methyl ester (D-NAME), the inactiveenantiomer of L-NAME (10 mg/kg i.p.,n 5 7), or aminoguanidine(AG), an inhibitor of NOS with relatively greater selectivity for iNOS(10 mg/kg i.p.,n 5 7). To further assess the local action of NO oninjured nerve microcirculation, additional 48 h (n 5 7) and 14 day(n 5 7) groups of rats were used. A series of pre- and post-treatmentreadings was obtained following topical application of L-NAME orD-NAME (0.2 mL, 40 mM). For the systemic application, LDFreadings were obtained 15–20 min following injection, while for thetopical application, post-treatment readings were obtained 20–25 minfollowing application. The doses for both L-NAME and AG werechosen based on work from previous studies (Congeret al., 1995 ;Kihara & Low, 1995).

L-NAME, D-NAME and AG were purchased from Sigma Chem-icals (St Louis MO, USA) and were freshly dissolved in 0.9% salineon the day of the experiment.

Hydrogen clearance studies

Hydrogen clearance (HC) microelectrode polarography was usedto study NOS-NO action within the endoneurial plexus of theinjured nerve. Detailed description of the method and the analysisof HC curves are previously described (Zochodne & Ho, 1991,1992). Briefly, rats were anaesthetized (pentobarbital 65 mg/kgi.p.), paralysed (tubocurarine 1.5 mg/kg intra-arterial) and ventilated(model 683, Harvard rodent respirator, South Natick, MA) with a

1848 D. Levy and D. W. Zochodne

mixture of oxygen (O2) and nitrogen (N2). A 3- to 5-µm-tippedglass–platinum microelectrode, connected to a chemical microsensor(Diamond General, Ann Arbor, MI, USA) and linearly sensitiveto hydrogen, was then inserted, through an epineurial window,into the endoneurium, proximal (48 h post-operation) or distal(14 days post-operation) to the constriction site or to similar sitesin sham-operated nerves.

Following the insertion of the electrode, hydrogen (H2) wasadded to the breathing mixture until saturation and then discontinuedto allow recording of the HC curve. The first 30 s of each curvewas discarded, and the next 5 min of each curve was fitted to amono- or bioexponential curve using least squares regression:Y 5A*EXP (B*X) 1 C*EXP (D*X) for the bioexponential curve, whereY represents the hydrogen polarography current (arbitrary units),X represents time,B and D are fast and slow components of thecurve, respectively, andA and C are weighting constants.Monoexponential curves have no fast component of the washoutand were represented only asY 5 A*EXP (D*X). Endoneurialblood flow (NBF) was calculated from the slow washout componentas (–D)*100. Theoretical and experimental justification for theanalysis has been published (Dayet al., 1989; Zochodneet al.,1995). MAP was recorded through a left carotid catheter, andstudies with MAP, 90 mmHg were discarded. Arterial bloodsamples were removed from the carotid catheter, and measuredhourly (Radiometer ABL 330, Copenhagen, Denmark) with adjust-ments made to ensure a physiological level of blood gasesand pH (PO2 , . 90 Torr ; PCO2, 35–45 Torr ; pH, 7.35–7.45).Supplemental doses of pentobarbital were given everyµ 2 h tomaintain a relatively constant level of anaesthesia, as judged bythe level and stability of MAP.

Three consecutive wash-out curves were obtained for each ratin four groups of animals (n 5 6). Fifteen minutes before the firstclearance curve, a single dose of saline (0.9% 0.3 mL i.p.) wasadministered. A single dose of either L-NAME (10 mg/kg i.p.) orD-NAME (10 mg/kg i.p.) was administered 15 min prior to thesecond and third clearance curves. As L-NAME has shown toexert its effect, as judged by its effect on MAP, forµ 45 minfollowing systemic administration (unpublished observations), thetwo last clearance curves were obtained at least 45 min apart fromeach other. NBF was calculated as the mean of the last twoclearance curves and was compared to NBF calculated from thefirst curve.

NOS immunohistochemistry

Injured (CCI) and sham-operated sciatic nerves were harvested at48 h (n 5 4), 7 days (n 5 4) and 14 days (n 5 4) post-operation,placed in Zamboni’s fixative [2% paraformaldehyde, 0.5% picricacid and 0.2M phosphate-buffered saline (PBS)] overnight, washedin PBS and dimethyl sulphoxide (DMSO) followed by immersionin a cryoprotective solution (30% sucrose/PBS) for 24 h. Nervespecimens were then divided into proximal, middle and distalsections, embedded individually in O.C.T. TissueTek medium(Miles, Elkhart, IN, USA), frozen and sectioned at 16 microns ina cryostat and mounted on poly-L-lysine-coated slides. All specimens(CCI and sham) were separately incubated for 2 h with 0.2%Triton X-100, pH 7.5 (BDH, Toronto, ONT, Canada) and immunos-tained with either rabbit antimouse iNOS, rabbit antirat nNOS orecNOS polyclonal serum (Transduction Lab, Lexington, KY, USA)diluted at 1 : 500 for 48 h in moist conditions at 4 °C. Slides werethen washed with PBS and incubated with fluorescein isothiocyanate-conjugated goat immunoglobulin G (FITC, Incstar, Stillwater, MN,USA) diluted at 1 : 50 for 1 h at room temperature. Following


further PBS washing, coverslips were mounted with bicarbonate-buffered glycerol (pH 8.6) and viewed with a fluorescent microscope(Zeiss, Axioplan).

Data analysis

Statistical differences in the thermal nociception tests were analysedusing one-way analysis of variance (ANOVA) with post-ANOVA pairedcomparisons made using Student’st-test. The Mann–Whitney testwas used to analyse differences in mechanical thresholds (tactileallodynia) between sham and operated sides. To assess the influenceof NOS antagonism (L-NAME, AG and D-NAME) on NBF, anunpaired Student’st-test with posthoc Bonferroni corrections wasused to compare results before and after administration of drugs.Results are means6 SEM.

Results

Nociceptive behavioural tests

Ipsilateral thermal hyperalgesia was present in all CCI-operatedanimals tested (n 5 14). Significant hyperalgesia, represented asdecreased paw withdrawal latency (Fig. 1A), and increased withdrawaland flinching duration (Fig. 1B), was evident as early as 48 h post-operation. A gradual decrease in withdrawal latency (Fig. 1A) and anincrease in the duration of the withdrawal response (Fig. 1B) weremaximal by 14 days and significantly higher than at 48 h. Mechanicalallodynia (Fig. 2), represented as a leftward shift in the stimulus–

FIG. 1. Development of thermal hyperalgesia following CCI to the right sciaticnerve, as measured by the noxious heat-evoked hind paw withdrawal. (A)Negative difference scores between withdrawal latencies indicate hyperalgesiaon the side of sciatic nerve ligation. (B) Measurements of average pawwithdrawal and flinching duration following the paw withdrawal reflex.*P , 0.01 one-way ANOVA between postligated and preligated groups.#P , 0.01 Student’st-test between 48 h and 14 day post-operated rats.


FIG. 2. Development of mechano-allodynia following CCI, as tested with a mechanical stimulus induced by stimulating the ventral aspect of the paw with aseries of graded calibrated nylon monofilaments (von Frey hairs) with increasing stiffness. The data are presented as cumulative percentage of rats with thresholdresponse to each of the von Frey hairs in the series, with data of 0% omitted. The responses of the opposite sham-operated hind paws were identical to thecontrol group (data not shown). *P , 0.05 Mann–Whitney tests between CCI and sham paws within the same group.

FIG. 3. RBC flux, as measured by the LDF method proximal and distal to theconstriction site, and at the contralateral sham-operated side. *P , 0.001, one-way ANOVA with post-tests and Bonferroni corrections.

response curve to calibrated von Frey hairs was evident ipsilateral tothe constriction, with onset at 48 h post-operation and a gradualdecrease in threshold at 14 days post-operation.

Local blood flow

LDF studies

As illustrated in Fig. 3, baseline RBC flux was significantly higherproximal and distal to the constriction site at 48 h post-operationwhen compared to sham-operated nerves. Seven days post-operation,a slight increase in RBC flux was evident, but was not significant.RBC flux was significantly higher at the distal site at 14 dayspost-operation. Very low RBC flux at the constriction site [–45%


as compared to contralateral sham (n 5 6), data not shown] itself,indicated ischemia within the constricted zone.

Systemic administration of L-NAME, but not D-NAME—theinactive enantiomer, or AG increased MAP (data not shown) by19 6 4% (n 5 21), 10 min following injection. MAP returned tobaseline (1026 4 mmHg) 40–45 min following injection. Figure 4illustrates the changes in RBC flux following systemic administrationof L-NAME, D-NAME or AG to different groups of animals.Forty-eight hours post-operation, L-NAME, but not D-NAME orAG, reduced RBC flux proximal and distal to the constriction siteby 526 8% and 316 8%, respectively. No significant change inRBC flux was observed at 7 days following injection with eitherL-NAME, D-NAME or AG on the CCI-operated site, althoughthere was a trend toward decreased RBC flux with L-NAME.Fourteen days post-operation, both L-NAME and AG, but not D-NAME, reduced RBC flux exclusively at the distal site by42 6 11% and 306 4%, respectively. Conversely, L-NAME, butnot D-NAME or AG, increased RBC flux in the contralateralsham-operated nerves by 326 4% in all rats at all timestested. To further assess NOS-NO action on the injured nervemicrocirculation, separate groups of animals at 48 h and 14 dayswere tested for changes in nerve RBC flux following systemicadministration of increasing doses of L-NAME. Figure 5 givesthree dose–response curves for RBC flux at the proximal anddistal sites of injured nerves, and the contralateral sham-operatednerves. L-NAME was associated with dose-dependent decreases inRBC flux at the proximal site at 48 h (Fig. 5, top) and distally at14 days (Fig. 5, bottom).

Topical application of L-NAME and D-NAME to injured nervesand sham-operated nerves resulted in a similar effect of RBC flux,as seen following systemic administration without changing MAP. L-NAME, but not D-NAME, reduced RBC flux exclusively at theproximal site at 48 h (Fig. 6, top) and distally at 14 days (Fig. 6,bottom). No significant changes were observed at any other sites


FIG. 4. Effect of systemic administration of L-NAME, D-NAME and AG (10 mg/kg i.p.) on RBC flux as measured by LDF at 48 h, 7 and 14 days post-operation.All data represent the percentage change of RBC flux 15–20 min following drug administration compared to baseline readings. *P , 0.001, **P , 0.01 one-way ANOVA with post-tests and Bonferroni corrections.

FIG. 5. Changes in sciatic RBC flux produced by graded doses of L-NAME at48 days and 14 days post-operation. *P , 0.001, **P , 0.01 one-wayANOVA

with post-tests and Bonferroni corrections.


tested including the contralateral sham-operated nerve, at any ofthe times tested.

Hydrogen clearance studies

Figure 7 represents the effect of systemic application of vehicle(saline), L-NAME and D-NAME on NBF in 48 h and 14 day CCInerves at the proximal and distal sites, respectively. On the rightside are examples of hydrogen clearance curves from injurednerves following administration of L-NAME and D-NAME.Flattening of the exponential clearance curve indicates a markedreduction of NBF. At 48 h (top left) and 14 days (bottom left),NBF was elevated proximal (48 h) and distal (14 days) to theconstriction site, compared to sham-exposed NBF, in rats treatedwith either vehicle or D-NAME. Treatment with L-NAME, however,reversed this rise in NBF both proximal (48 h) and distal (14 days),but did not change NBF in the sham-exposed control nerve.

NOS immunohistochemistry

Increased immunoreactivity of the ecNOS isoform was intense anddiscrete at 48 h following CCI (Fig. 8B,C), unlike sham-operatednerves (Fig. 8A). Increased ecNOS staining was evident just proximalto the constriction site and was present as discrete deposits withinthe endoneurium. At the same time, increased nNOS immunoreactivitywas evident within the endoneurium proximal to the constriction site(Fig. 8E,F). No nNOS immunostaining was evident in sham-operatednerves (Fig. 8D). At 7 days post-operation, increased immunoreactiv-ity of the iNOS isoform could be observed within the endoneuriumat the constriction site (Fig. 9B). At 14 days, iNOS was also observedwithin the endoneurium (Fig. 9C) and epineurium (Fig. 9D) distal tothe constriction site. No iNOS immunoreactivity was evident at thecontralateral sham-operated side at 7 days (not shown) or 14 days(Fig. 9A). Both ecNOS and nNOS were not observed at 7 and 14 days.


Discussion

The major findings from this work were: (i) Chronic constrictioninjury to the sciatic nerve in rats resulted in neuropathic pain confirmed

FIG. 6. Influence of topical administration of L-NAME and D-NAME on RBCflux in 48 h and 14 days post-operated nerves. **P , 0.01 one-wayANOVA

with post-tests and Bonferroni corrections.

FIG. 7. The effect of systemic administration of L-NAME, D-NAME and vehicle (saline) on CCI and sham-exposed NBF as measured by the hydrogen clearancepolarography method. NBF was measured at the proximal site at 48 h and distal at 14 days post-operation. *P , 0.001 (compared to D-NAME or saline),###P , 0.05 (compared to sham-operated nerves) one-wayANOVA with post-tests and Bonferroni corrections.


by nociceptive behavioural testing, and was associated with localhyperaemia (rises in NBF) proximal and distal to the constriction siteat 48 h, and distal to the constriction site 14 days post-operation. (ii)Hyperaemia was reversed in a dose-dependent manner by systemicor local NOS inhibition providing physiological evidence for localNOS-NO action on nerve microvessels. (iii) Local NBF studiesindicated evidence for NOS-NO activity within both the endoneuriumand epineurium. (iv) There was NOS expression in CCI nerves withthe constitutive ecNOS and nNOS isoforms at 48 h, and iNOS atlater (7 and 14 days) timepoints.

Our findings suggest that NO is locally elaborated within theinjured sciatic nerve following CCI. Local NO may contribute to thedevelopment of hyperalgesia and allodynia directly or indirectly byinfluencing the local inflammatory and repair process of a partiallyinjured peripheral nerve. Changes in NBF provide a useful physiologi-cal index of NOS-NO action because NO is a potent vasodilator,whereas direct studies of NO presence are problematic because of itsshort half-life. Identifying changes in NBF following administrationof NOS inhibitors permitted assessment of the topographical actionof NOS-NO, and the LDF technique provided side to side differencesbetween the constricted and sham-operated nerves. Our use of boththe HC and LDF techniques allowed us to gather complementaryinformation about NO-mediated changes in NBF both in the endoneur-ium and epineurium.

Systemic application of L-NAME, but not D-NAME or AG,increased RBC flux on the sham-exposed sides and this was accompan-ied by a significant elevation in MAP in two of the three doses tested(1 and 10 mg/kg i.p.). The systemic inhibition of NO productioninduced a generalized pressure response with an increase in MAP.This increase in MAP resulted in a secondary increase in NBF inuninjured nerves due to lack of autoregulation (Low & Tuck, 1984).However, the opposite reductions in NBF by L-NAME and AG onipsilateral injured nerves indicated that there was activity of NOS


FIG. 8. Immunoreactivity of ecNOS and nNOS is selectively increased in the proximal edge of 48 h constricted sciatic nerves. Sham-operated nerves stainedwith ecNOS (A) or nNOS (D) bar5 200 mm. (B) Low magnification of constricted sciatic nerve stained with ecNOS antibody. Bar5 200 mm. (C) Highmagnification of epineurial ecNOS deposits. Bar5 50 mm. (E) and (F) Similar magnifications of constricted sciatic nerve stained with nNOS antibody. Sectionswere taken just proximal to the first suture.

and NO acting on local microvessels following CCI. Other vasoactiveagents may also contribute to the hyperaemia seen following the CCI.For example, CGRP contributes to local rises in NBF followingcomplete peripheral nerve transection and the formation of a neuroma(Zochodne & Ho, 1992 ; Zochodneet al., 1995), but its action maybe mediated, at least in part, by NO. The dose–response studies inthe present work carried out at 48 h and 14 days suggest that NOSmay be particularly active at sites proximal to the constriction at 48 hand distal to it at 14 days.

To further assess the local action of NOS-NO on the injured nervemicrocirculation and to eliminate possible confounding effects ofMAP on NBF, we used an equimolar concentration of L-NAME,applied topically on the injured nerve, as used in previous studies ofNBF (Zochodneet al., 1995 ; Kihara & Low, 1995). The localstereospecific action of L-NAME, following topical application,was also associated with a decrease in RBC flux and reversal ofhyperaemia.

L-NAME is a known potent inhibitor of ecNOS, as indicated bythe rise in MAP observed following systemic administration (Congeret al., 1995). It reversed hyperaemia at 48 h, especially proximal to


the constriction site. Conversely, AG has been suggested to serve asa more selective inhibitor of the iNOS isoform (Griffithset al., 1993)and does not influence MAP (Grishamet al., 1994). The ability ofAG to reduce RBC flux exclusively distal to the constriction site at14 days suggests that the iNOS isoform may be the active isoform atthis site. AG may also interfere with non-enzymatic glycosylation(Edelstein & Brownlee, 1992) and may inhibit diamine oxidase(Rokkaset al., 1990), an enzyme that helps to inactivate histamine(Szareket al., 1992). These actions, however, are unlikely to berelevant to our model, since increases in local histamine from AGshould have increased, rather than decreased, NBF.

The immunoreactivity of the constitutive ecNOS and nNOS, andthe inducible or immunological NOS isoforms at the proximal site at48 h and the distal site at 14 days is in agreement with our physiologi-cal evidence for NO elaboration at the same sites. At early timepoints,inflammation (Clatworthyet al., 1995) might promote the release ofNO through BK and substance P. These pro-inflammatory neuropep-tides have been shown to stimulate the vascular endothelium cells torelease nitric oxide probably through the ecNOS isoform (Bergueret al., 1993 ; Kuroiwaet al., 1995). We did not observe either ecNOS


FIG. 9. iNOS immunoreactivity in 7 and 14 days chronic constricted sciatic nerves. (A) Sham-operated nerves did not show any iNOS immunoreactivity. (B) At7 days post-operation, iNOS immunoreactivity was mainly observed in the endoneurium at the constriction site. Fourteen days post-operation iNOSimmunoreactivity was also observed in the endoneurium at the constriction site. (D) At 14 days, iNOS immunoreactivity was also observed distal to theconstriction site (bar5 50 mm).

staining of microvessels, in either sham-operated or non-operatednerves, or nNOS in nerve fibres, perhaps due to their normally lowconcentration. However, at 48 h in CCI nerves, a substantial increasein immunoreactivity of both constitutive NOS enzymes (ecNOS andnNOS) was observed in the endoneurium proximal to the constrictionsite. An increase in nNOS expression in the cell body of axotomizedprimary afferents has already been observed (Vergeet al., 1992). ThenNOS immunoreactivity is likely to be due to an expression inprimary afferents, possibly in end-bulbs formed after axotomy (Zoch-odneet al., 1997). Whether anterograde transport or local expressionunderlies this increase in immunoreactivity still needs to be evaluated.

The exact origin of ecNOS in the endoneurium proximal to the


constriction remains uncertain. It may have arisen from Schwanncells since glial cells have been shown to express both ecNOS (Maet al., 1994) and iNOS (Galeaet al., 1992). Motor neurons are alsoa conceivable source by synthesizing ecNOS following an insult(Watanabeet al., 1996). Regulation of the ecNOS transcript hasalready been suggested to occur in response to certain stimuli, e.g.shear stress (Busse & Fleming, 1995), downregulation or upregulationby hypoxia (Wang & Masden, 1995), or downregulation by tumournecrosis factor (TNF-α) [Yoshizumi et al., 1993]. It is conceivablethat local constriction of the nerve may promote upregulation of theecNOS gene and a compensatory release of NO in response to localischemia and hypoxia. Later, downregulation of the constitutive


ecNOS and nNOS to undetectable levels, at 7 and 14 days, may havebeen the result of increases in cytokine production, especially TNF-α, presumably from Schwann cells (Wagner & Myers, 1996).

The appearance of iNOS immunoreactivity in the middle sectionat 7 days and distal section at 14 days parallelled the appearance ofmacrophages that participate in axonal degeneration following partialnerve injury (Sommeret al., 1995). Macrophages are a likely sourceof NO, probably through activation of the iNOS isoform (Szabo &Thiemermann, 1995). Endothelial and Schwann cells may provideadditional sources of NO, since both have been shown to expressiNOS under various pathological conditions (Koprowskiet al., 1993 ;Gold et al., 1996). The exclusive immunolocalization of iNOS in themiddle constricted sites of the nerve at 7 days might explain the lackof response of RBC flux to NOS inhibition in the proximal and distalsites. Local rises in NO production at the middle constricted sectionmay not have influenced microvessels of the distal and proximal sitesbecause of rapid conversion of NO to nitrite (Crow & Beckman, 1995).

Local elaboration in NOS activity following peripheral nerve injuryassociated with local inflammation may directly or indirectly enhancenociception. NO may activate and sensitize primary afferents(Wiesenfeld-Hallin et al., 1993 ; Holthusen & Arndt, 1995), thusdirectly promoting hyperalgesia and allodynia. Indirectly, NO mayserve as a secondary mediator of bradykinin, a pro-inflammatoryhyperalgesic agent (Holthusen, 1997), or promote prostaglandinformation through activation of the cyclooxygenases I and II(Salveminiet al., 1993), thus contributing to C-fibre excitation (Devoret al., 1992). Recently, Tal & Eliav (1996) have shown an interestingtopographical pattern of mechanosensitivity along the CCI nervetrunk. The mechanosensitivity starts proximal to the constriction at48 h and progresses distally at later time points, resembling thepattern of NOS-NO activity we demonstrated in this work. Themechanism by which NO might generate mechanosensitivity is stillunknown, but may be related to changes in second messenger cascadesthat influence primary afferent excitability. The recent work of Thomaset al. (1996), demonstrating suppression of behavioural indices ofneuropathic pain with local L-NAME infusion, supports the hypothesisthat NO-NOS may be involved in the local generation of pain. Ourwork indicates that NOS is present and active within the injuredperipheral nerve trunk.

Conclusions

The present study demonstrates that partial peripheral nerve injuryassociated with the CCI model of neuropathic pain is associated withlocal NO-mediated changes in microvessels. These microvascularchanges provide a signature of local NO presence that has specifictemporal and spatial features in the nerve trunk. Immunohistochemicalstaining has shown that ecNOS and nNOS are expressed in theproximal section in the endoneurium 48 h post-operation. iNOS ismainly expressed within the endoneurium 14 days post-operation.The results of this work may promote our understanding on thecorrelation between the local inflammatory response that occursfollowing peripheral nerve injury and possible peripheral mechanismsunderlying neuropathic pain.

AcknowledgementsBrenda Boake provided expert secretarial assistance. Dr Ken Hargreaves,University of Minnesota kindly helped us set up the thermal hyperalgesiatesting apparatus. Drs Ray Turner and Sheldon Roth made helpful suggestions.D.W.Z. is a Medical Scholar of the Alberta Heritage Foundation for MedicalResearch. The work was supported by the Muscular Dystrophy Associationof Canada and the Medical Research Council of Canada.


Abbreviations

AG aminoguanidineCCI chronic constriction injuryDMSO dimethyl sulphoxideD-NAME NG-nitro-D-arginine methyl esterHC hydrogen clearanceLDF laser Doppler flowmetryL-NAME NG-nitro-L-arginine methyl esterMAP mean arterial pressureNBF nerve blood flowNO nitric oxideNOS nitric oxide synthaseecNOS endothelial constitutive NOSiNOS inducible NOSnNOS neuronal NOSPBS phosphate-buffered salineRBC red blood cellTNF tumour necrosis factor

References

Basbaum, A.I., Gautron, M., Jazat, F., Mayes, M. & Guilbaud, G. (1991) Thespectrum of fiber loss in a model of neuropathic pain in the rat: an electronmicroscopic study.Pain, 47, 359–367.

Bennett, G.J. & Xie, Y.K. (1988) A peripheral mononeuropathy in rat thatproduces disorders of pain sensation like those seen in man.Pain, 33, 87–107.

Berguer, R., Hottenstein, O.D., Palen, T.E., Stewart, J.M. & Jacobson, E.D.(1993) Bradykinin-induced mesenteric vasodilation is mediated by B2-subtype receptors and nitric oxide.Am. J. Physiol., 264, G492–G496.

Bredt, D.S. & Snyder, S.H. (1994) Nitric oxide: a physiologic messengermolecule.Annu. Rev. Biochem., 63, 175–195.

Busse, R. & Fleming, I. (1995) Regulation and functional consequences ofendothelial nitric oxide formation.Ann. Med., 27, 331–340.

Clatworthy, A.L., Illich, P.A., Castro, G.A. & Walters, E.T. (1995) Role ofperi-axonal inflammation in the development of thermal hyperalgesia andguarding behavior in a rat model of neuropathic pain.Neurosci. Lett., 184,5–8.

Conger, J., Robinette, J., Villar, A., Raij, L. & Shultz, P. (1995) Increasednitric oxide synthase activity despite lack of response to endothelium-dependent vasodilators in postischemic acute renal failure in rats.J. Clin.Invest., 96, 631–638.

Crow, J.P. & Beckman, J.S. (1995) Reactions between nitric oxide, superoxide,and peroxynitrite: footprints of peroxynitritein vivo. In Ignaroo, L. &Murad, F. (eds),Nitric Oxide: Biochemistry, Molecular Biology, andTherapeutic Implications. Academic Press, Toronto, pp. 17–44.

Day, T.J., Lagerlund, TD, Low, P.A. (1989) Analysis of H2 clearance curvesto measure blood flow in rat sciatic nerve.J. Physiol. (London), 414, 35–54.

Devor, M., White, D.M., Goetzl, E.J. & Levine, J.D. (1992) Eicosanoids, butnot tachykinins, excite C-fiber endings in rat sciatic nerve-end neuromas.NeuroReport, 3, 21–24.

Edelstein, D. & Brownlee, M. (1992) Mechanistic studies of advancedglycosylation end product inhibition by aminoguanidine.Diabetes, 41,552–556.

Galea, E., Feinstein, D.L. & Reis, D.J. (1992) Induction of calcium-independentnitric oxide synthase activity in primary rat glial cultures.Proc. Natl. Acad.Sci. USA, 89, 10 945–10 949.

Gold, R., Zielasek, J., Kiefer, R., Toyka, K.V. & Hartung, H.P. (1996) Secretionof nitrite by Schwann cells and its effect on T-cell activationin vitro. CellImmunol., 168, 69–77.

Griffiths, M.J., Messent, M., MacAllister, R.J. & Evans, T.W. (1993)Aminoguanidine selectively inhibits inducible nitric oxide synthase.Br. J.Pharmacol., 110, 963–968.

Grisham, M.B., Specian, R.D. & Zimmerman, T.E. (1994) Effects of nitricoxide synthase inhibition on the pathophysiology observed in a model ofchronic granulomatous colitis.J. Pharmacol. Exp. Ther., 271, 1114–1121.

Hargreaves, K., Dubner, R., Brown, F., Flores, C. & Joris, J. (1988) A newand sensitive method for measuring thermal nociception in cutaneoushyperalgesia.Pain, 32, 77–88.

Holthusen, H. (1997) Involvement of the NO/cyclic GMP pathway inbradykinin-evoked pain from veins in humans.Pain, 69, 87–92.

Holthusen, H. & Arndt, J.O. (1995) Nitric oxide evokes pain at nociceptorsof the paravascular tissue and veins in humans.J. Physiol., 487, 253–258.

Kawabata, A., Manabe, S., Manabe, Y. & Takagi, H. (1994) Effect of topicaladministration of L-arginine on formalin-induced nociception in the mouse:


a dual role of peripherally formed NO in pain modulation.Br. J. Pharmacol.,112, 547–550.

Kihara, M. & Low, P.A. (1995) Impaired vasoreactivity to nitric oxide inexperimental diabetic neuropathy.Exp. Neurol., 132, 180–185.

Koprowski, H., Zheng, Y.M., Heber-Katz, E., Fraser, N., Rorke, L., Fu, Z.F.,Hanlon, C. & Dietzschold, B. (1993) In vivo expression of inducible nitricoxide synthase in experimentally induced neurologic diseases.Proc. Natl.Acad. Sci. USA, 90, 3024–3027.

Kuroiwa, M., Aoki, H., Kobayashi, S., Nishimura, J. & Kanaide, H. (1995)Mechanism of endothelium-dependent relaxation induced by substance Pin the coronary artery of the pig.Br. J. Pharmacol., 116, 2040–2047.

Levy, D. & Zochodne, D.W. (1996) Evidence for NO action in an animalmodel of chronic neuropathic pain.Soc. Neurosci. Abst., 22, 866 (Abstract).

Low, P.A. & Tuck, R.R. (1984) Effects of changes of blood pressure,respiratory acidosis and hypoxia on blood flow in the sciatic nerve of therat. J. Physiol.(Lond.), 347, 513–524.

Ma, L., Morita, I. & Murota, S. (1994) Presence of constitutive type nitricoxide synthase in cultured astrocytes isolated from rat cerebra.Neurosci.Lett., 174, 123–126.

Meller, S.T., Cummings, C.P., Traub, R.J. & Gebhart, G.F. (1994) The roleof nitric oxide in the development and maintenance of the hyperalgesiaproduced by intraplantar injection of carrageenan in the rat.Neuroscience,60, 367–374.

Meller, S.T., Pechman, P.S., Gebhart, G.F. & Maves, T.J. (1992) Nitric oxidemediates the thermal hyperalgesia produced in a model of neuropathic painin the rat.Neuroscience, 50, 7–10.

Moilanen, E. & Vapaatalo, H. (1995) Nitric oxide in inflammation and immuneresponse.Ann. Med., 27, 359–367.

Moncada, S., Palmer, R.M. & Higgs, E.A. (1991) Nitric oxide: physiology,pathophysiology, and pharmacology.Pharmacol. Rev., 43, 109–142.

Munger, B.L., Bennett, G.J. & Kajander, K.C. (1992) An experimental painfulperipheral neuropathy due to nerve constriction. I. Axonal pathology in thesciatic nerve.Exp. Neurol., 118, 204–214.

Myers, R.R., Yamamoto, T., Yaksh, T.L. & Powell, H.C. (1993) The role offocal nerve ischemia and Wallerian degeneration in peripheral nerve injuryproducing hyperesthesia.Anesthesiol., 78, 308–316.

Roche, A.K., Cook, M., Wilcox, G.L. & Kajander, K.C. (1996) A nitric oxidesynthesis inhibitor (L-NAME) reduces licking behavior and Fos-labeling inthe spinal cord of rats during formalin-induced inflammation.Pain, 66,331–341.

Rokkas, T., Vaja, S., Murphy, G.M. & Dowling, R.H. (1990) Aminoguanidineblocks intestinal diamine oxidase (DAO) activity and enhances the intestinaladaptive response to resection in the rat.Digestion, 46 (Suppl. 2), 447–457.

Salvemini, D., Misko, T.P., Masferrer, J.L., Seibert, K., Currie, M.G. &Needleman, P. (1993) Nitric oxide activates cyclooxygenase enzymes.Proc.Natl. Acad. Sci. USA, 90, 7240–7244.

Shephard, A.P. (1990) The history of laser doppler blood flowmetry. InShepard, A.P. & Oberg, P.A. (eds),Laser Doppler Blood Flowmetry. KluwerAcademic, Massachusetts, pp. 1–16.


Sommer, C., Lalonde, A., Heckman, H.M., Rodriguez, M. & Myers, R.R.(1995) Quantitative neuropathology of a focal nerve injury causinghyperalgesia.J. Neuropathol. Exp. Neurol., 54, 635–643.

Szabo, C. & Thiemermann, C. (1995) Regulation of the expression of theinducible isoform of nitric oxide synthase. In Ignaro, L. & Murad, F.(eds), Nitric Oxide: Biochemistry, Molecular Biology and TherapeuticImplications. Academic Press, Toronto, pp. 113–153.

Szarek, J.L., Bailly, D.A., Stewart, N.L. & Gruetter, C.A. (1992) HistamineH1-receptors mediate endothelium-dependent relaxation of rat isolatedpulmonary arteries.Pulm. Pharmacol., 5, 67–74.

Tal, M. & Eliav, E. (1996) Abnormal discharge originates at the site of nerveinjury in experimental constriction neuropathy (CCI) in the rat.Pain, 64,511–518.

Thomas, D.A., Ren, K., Besse, D., Ruda, M.A. & Dubner, R. (1996)Application of nitric oxide synthase inhibitor, N omega-nitro-L-argininemethyl ester, on injured nerve attenuates neuropathy-induced thermalhyperalgesia in rats.Neurosci. Lett., 210, 124–126.

Verge, V.M., Xu, Z., Xu, X.J., Wiesenfeld-Hallin, Z. & Hokfelt, T. (1992)Marked increase in nitric oxide synthase mRNA in rat dorsal root gangliaafter peripheral axotomy:in situ hybridization and functional studies.Proc.Natl. Acad. Sci. USA, 89, 11 617–11 621.

Wagner, R. & Myers, R.R. (1996) Schwann cells produce tumor necrosisfactor alpha: expression in injured and non-injured nerves.Neuroscience,73, 625–629.

Wang, Y. & Masden, P.A. (1995) Nitric oxide synthases: gene structure andregulation. In Ignaro, L. & Murad, F. (eds),Nitric Oxide Biochemistry,Molecular Biology and Therapeutic Implications. Academic Press, Toronto,pp. 71–90.

Watanabe, M., Sakurai, M., Abe, K., Aoki, M., Sadahiro, M., Tabayashi, K.,Okamoto, K., Shoji, M. & Itoyama, Y. (1996) Inductions of Cu/Znsuperoxide dismutase- and nitric oxide synthase-like immunoreactivities inrabbit spinal cord after transient ischemia.Brain Res., 732, 69–74.

Wiesenfeld-Hallin, Z., Hao, J.X., Xu, X.J. & Hokfelt, T. (1993) Nitric oxidemediates ongoing discharges in dorsal root ganglion cells after peripheralnerve injury.J. Neurophysiol., 70, 2350–2353.

Yoshizumi, M., Perrella, M.A., Burnett, J.C. Jr & Lee, M.E. (1993) Tumornecrosis factor downregulates an endothelial nitric oxide synthase mRNAby shortening its half-life.Circ. Res., 73, 205–209.

Zochodne, D.W., Allison, J.A., Ho, W., Ho, L.T., Hargreaves, K. & Sharkey,K.A. (1995) Evidence for CGRP accumulation and activity in experimentalneuromas.Am. J. Physiol., 268, H584–H590.

Zochodne, D.W. & Ho, L.T. (1991) Influence of perivascular peptides onendoneurial blood flow and microvascular resistance in the sciatic nerve ofthe rat.J. Physiol., 444, 615–630.

Zochodne, D.W. & Ho, L.T. (1992) Hyperemia of injured peripheral nerve:sensitivity to CGRP antagonism.Brain Res., 598, 59–66.

Zochodne, D.W., Levy, D., Sun, H., Cheng, C., Lauritzen, M. & Rubin, I.(1997) Nitric oxide and nitric oxide synthetase in injured peripheral nervetrunks.Can. J. Neurol. Sci., 24, S57 (Abstract).

Local nitric oxide synthase activity in a model of neuropathic pain

Documents

Transcript of Local nitric oxide synthase activity in a model of neuropathic pain