painful mononeuropathy in c57bl/wld mice with delayed wallerian degeneration: differential effects...

Ž .Brain Research 784 1998 154–162

Research report

Painful mononeuropathy in C57BLrWld mice with delayed Walleriandegeneration: differential effects of cytokine production and nerve

regeneration on thermal and mechanical hypersensitivity 1

Claudia Sommer a,), Maria Schafers b¨a Neurologische Klinik der UniÕersitat, Wurzburg, Germany¨ ¨

b Neurologische Klinik der RWTH, Aachen, Germany

Accepted 21 October 1997

Abstract

Wallerian degeneration with macrophage influx and production of proinflammatory cytokines is a critical factor in the development ofhyperalgesia in animal models of neuropathic pain. We hypothesized that in the mouse strain with delayed Wallerian degeneration, theC57BLrWld mouse, the temporal course of mechanical allodynia and thermal hyperalgesia as well as the temporal profile of cytokineexpression after nerve injury would differ from normal mice. Here we used the model of chronic constriction injury of the sciatic nerveŽ .CCI to study the correlation of pain related behavior with peripheral nerve de- and regeneration and concomitant cytokine production.Indeed, after CCI, C57BLrWld mice showed markedly reduced thermal hyperalgesia compared to normal C57BLr6 mice, temporally

Ž .related to the delayed recruitment of hematogeneous macrophages to the injured nerve. Endoneurial tumor necrosis factor-a TNF -likeimmunoreactivity increased rapidly in normal mice but did so with a delayed time course in C57BLrWld mice. In addition, the durationof mechanical allodynia was significantly prolonged in C57BLrWld mice as compared to C57BLr6 mice, in accordance with the delayin regeneration of sensory nerve fibers in these mice. These results suggest that macrophage invasion and production of TNF mayinfluence the development of thermal hyperalgesia and that regenerative activity is linked to mechanical allodynia in peripheralmononeuropathy. q 1998 Elsevier Science B.V.

Keywords: Allodynia; Hyperalgesia; Chronic constriction injury; Tumor-necrosis-factor; Interleukin-1; Wallerian degeneration

1. Introduction

Wallerian degeneration is a critical factor in the induc-tion of hyperalgesia in animal models of neuropathic pain.Incomplete nerve injuries which induce Wallerian degener-ation lead to reproducible hyperalgesia in experimentalanimals, whereas injuries resulting in demyelination onlyhave no measurable effect on pain-related behaviorsw x17,19 . Here we used the model of chronic constrictive

Ž . w xinjury CCI 2 of one sciatic nerve to further study therelationship between the events associated with Walleriandegeneration and the time course of hyperalgesia afternerve injury. Specifically, we compared nerve pathology,

) Corresponding author. Neurologische Klinik der Universitat, Josef-¨Schneider-Str. 11, 97080 Wurzburg, Germany. Fax: q49-931-201-2697;¨E-mail: [email protected]

1 Presented in part at the 25th Annual Meeting of the Society forNeuroscience, San Diego, USA, 11r1995.

cytokine expression and pain related behavior in C57BLr6mice and in C57BLrWld mice. The C57BLrWld mouseis a strain in which Wallerian degeneration is extremely

w xslow due to a genetic defect on chromosome 4 14,15 .Accordingly, regeneration is delayed in lesioned nerves of

w xthese animals 3 . These changes are more pronounced inw xsensory than in motor nerves 10 . It has previously been

shown that the onset of thermal hyperalgesia following aŽ .chronic constrictive nerve injury CCI of the sciatic nerve

is delayed in C57BLrWld mice temporally related to theirw xdelay in Wallerian degeneration 16,18 . Since the delayed

Wallerian degeneration in these mice is associated with adiminished influx of hematogenous macrophages into theinjured nerve, we speculated that cytokines produced bythese macrophages might be a critical factor in the initia-tion of thermal hyperalgesia.

Ž .Cytokines like tumor necrosis factor-a TNF or inter-Ž .leukin-1b IL-1b have repeatedly been implied to be

Ž w x.mediators of inflammatory pain for review see Ref. 9 .

0006-8993r98r$19.00 q 1998 Elsevier Science B.V. All rights reserved.Ž .PII S0006-8993 97 01327-9

( )C. Sommer, M. SchafersrBrain Research 784 1998 154–162¨ 155

Their role in other chronic pain states or in neuropathicw xpain has only recently come to attention 5,32 . In particu-

lar, TNF given by endoneurial injection could elicit neuro-pathic pain behaviors as otherwise seen in animals with

w xpartial nerve injuries 29 . Also, TNF applied topically tosciatic nerve in vivo elicits dose-dependent firing in indi-

w xvidual afferent nerve fibers 33,36 . In contrast, administra-tion of agents which reduce the cytokine content in injuredperipheral nerve resulted in reduced hyperalgesia in the

w xanimals 26 .To test the hypothesis that cytokines produced in en-

doneurial macrophages might be a critical factor in theinitiation of hyperalgesia after nerve injury, we investi-gated the numbers of macrophages, tumor-necrosis factor-aŽ . ŽTNF -like immunoreactivity, as well as interleukin-1 IL-.1 a and b-like immunoreactivity in C57BLr6 mice and

in C57BLrWld mice. Morphometric analysis of immuno-cytochemically stained sections at various time points afterCCI was carried out. Mice were monitored for the devel-opment of thermal hyperalgesia and of mechanical allody-nia. Surprisingly, we found that the onset of mechanicalallodynia, as opposed to that of thermal hyperalgesia, wasnot delayed in C57BLrWld mice. However, the durationof mechanical allodynia was significantly prolonged com-pared to the C57BLr6-strain.

2. Materials and methods

2.1. Animals, surgery and behaÕioral testing

All experiments were approved by the local AnimalStudies Committees. Twenty-five C57BLr6 mice and 25

ŽC57BLrWld mice female, 18–21 g, supplied by Harlan.Winkelmann, Borchen, Germany were anesthetized by

intraperitoneal injection of a barbiturate solution and thesciatic nerve was exposed unilaterally at the mid-thighlevel. Three ligatures with 10-0 prolene were placed aroundthe nerve with 1 mm spacing. The ligatures were tied untilthey elicited a brief twitch in the hindlimb. This corre-sponded to blanching of the nerve and a significant reduc-tion in nerve blood flow secondary to occlusion of the

w xepineurial vasculature 19 . The wound was closed inlayers. The thermal nociceptive threshold was measured ineach hindpaw on two consecutive days before surgeryŽ .baseline and every second day from day 3 to day 9 after

w xsurgery as described before 24 . Briefly, the animals wereplaced in a plastic cage on a glass plate. A heat source waspositioned under one hind paw, and the withdrawal latencyto this thermal stimulus was recorded automatically. Pawwithdrawal latencies on the control side were subtractedfrom those on the experimental side, the resulting ‘dif-ference score’ giving a measure of thermal hyperalgesia.Withdrawal thresholds to innocuous mechanical stimuliwere assessed with von Frey hairs using the up-and-down

w xmethod of Dixon 7 at regular intervals from day 3 to 95

after surgery. Mice were placed in a plastic cage with awire mesh bottom. The plantar surface was touched per-pendicularly with a von Frey hair until slight buckling waspresent. Testing was initiated with the 0.6 mN hair. If themouse responded to touching the footpad with this hair for3 s by brisk withdrawal of the respective hindpaw, theresponse was considered positive. In case of a positiveresponse, the next weaker stimulus was presented, in caseof a negative response, the next stronger stimulus wasapplied. This procedure was repeated until six responsesŽ .either positive or negative were recorded. A 50% thresh-old, which indicates the force of von Frey hair at which ananimal reacts in 50% of the presentations, was calculatedusing the formula:

50% threshold mN s 10w X fqkd x r10,000=9.8Ž . Ž .indicating the force at which this individual rat would

Ž Žwithdraw the hindpaw in 50% of trials Xfsvalue log.units of the final von Frey hair used, ks tabular value

Ž .according to Dixon 1965; Esmean difference log units.between stimuli . A reduction in 50% thresholds was

Žinterpreted as mechanical allodynia. Animals ns5 of.each strain were sacrificed on day 4, 8, 12, 28 and 95

after surgery and the sciatic nerves were harvested. Fiveunoperated animals from both strains were also behav-iorally tested until day 95 and were then sacrificed.

2.2. Histology and immunohistochemistry

Sciatic nerve segments 4 mm of length immediatelydistal to the ligatures and the corresponding segments fromcontrol animals were processed for histology. Frozen sec-tions were stained with H&E and Kluver–Barrera myelin¨stain. Immunohistochemical staining for TNF, IL-1a andb , and macrophages was performed on frozen sections

Ž .using an ABC-system Vector . Primary antibodies werepolyclonal rabbit anti mouse TNF-a , 1:1000, polyclonalrabbit anti mouse IL-1a , 1:1000, monoclonal hamster anti

Ž .mouse IL-1b , 1:50, Genzyme , monoclonal rat anti mouseŽ .macrophages 1:100 Mac-1, CD 11b, Serotec , and poly-

Ž .clonal rabbit anti S-100 Dako , 1:5000. Controls includedomission of primary antibodies and preabsorption with the

Ž .respective recombinant cytokine Genzyme . Comparabil-ity between slides was maintained by staining of largesample sizes at a time, so that staining conditions wereexactly the same for all sections to be compared. Morpho-metric analysis of macrophages and cytokines was per-formed by a blinded investigator with a computer-assisted

Ž .morphometry system NIH-Image 6.0 by densitometri-cally identifying the immunostained cells and cell-processes. Images covering the whole nerve cross sectionwere digitized with set lighting and filter conditions at amagnification of =400. The number of immunostainedmacrophages was counted, and the number of endoneurialmacrophages per mm2 was calculated. For analysis ofcytokines, complete endoneurial area and the area of im-

( )C. Sommer, M. SchafersrBrain Research 784 1998 154–162¨156

munostained cells and cell-processes were measured andthe percentage of immunostained endoneurial area wascalculated. In Kluver–Barrera-stained sections, the area of¨stained myelin was recorded and the percentage of stainedendoneurial area was calculated.

For identification of cytokine-immunoreactive cells, tis-sue distal to the section used for morphometry was pro-cessed for immunohistochemistry using the same antibod-ies as above on serial 1 mm cryosections according to

w xmethods published previously 23 .

2.3. Statistical analysis

Comparison of behavioral data for thermal hyperalgesiaand of densitometric data was performed by two-way-ANOVA for comparison of groups and consecutive 2-tailedt-tests for comparison of mice at each time point. Mann–Whitney U-tests were used for statistical analysis of 50%thresholds to mechanical stimuli. Significance was as-sumed below p-values of 0.05.

3. Results

3.1. BehaÕior

No overt pain was observed in the animals post-oper-atively, they fed normally and gained weight like theunoperated controls. Motor function was affected duringthe first postoperative week with the animals dragging theaffected limb. The behavioral response to heating of theaffected paw in normal C57BLr6 mice was qualitatively

w xsimilar to our previous observations in rats 24 . With-drawal latencies on the sham operated side were un-

Žchanged from the preoperative values of 6.5"0.6 s data.not shown . Difference scores were y2.42"0.31 on day

3, y2.58"0.45 on day 5, y2.42"0.31 on day 7 andy1.81"0.25 on day 9 for CCI-animals. Thus, thermalhyperalgesia developed within three days and peaked dur-

Ž .ing the first week Fig. 1 .The C57BLrWld mice displayed a weak thermal hy-

peralgesic response to the injury from day 3 post surgerywith difference scores of y1.28"0.13. Difference scoreswere y1.36"0.06 on day 5, y1.28"0.13 on day 7 andy1.31"0.17 on day 9. These values differed significantlyŽ .p-0.001 from the C57BLr6 group between day 3 andŽ .7 Fig. 1 . On day 9, there was no longer a significant

difference in behavioral scores between the groups, bothdisplayed moderate hyperalgesia.

The 50% thresholds to von Frey hairs in unoperatedmice were 2 and 1.8 mN, respectively in both strains.There was no difference in thresholds between the groupsinitially. Although there was a trend to reduction of thewithdrawal thresholds in C57BLr6 mice from day 4 onthe operated side, this was not significant. A reduction inthe 50% threshold to von Frey stimuli was first unequivo-cally present on day 10 post surgery in both strains. InC57BLr6 mice, 50%-thresholds were significantly re-

Ž .duced with a minimum of 0.1 mN p-0.05 from day 10to 26 and were no longer different from the sham operatedside by day 32. In C57BLrWld mice, threshold weresignificantly lowered to values between 0.3 and 0.1 mNŽ . Ž .p-0.05 from day 10 to 65 Fig. 2 . Thus, the timeperiod with lowered mechanical withdrawal thresholds,interpreted as the duration of allodynia, was markedlyprolonged in C57BLrWld mice compared to the controlstrain.

Ž .Fig. 1. Hyperalgesia to thermal stimuli following unilateral sciatic nerve injury is plotted for two groups of animals each: C57BLr6 mice diamonds andŽ .C57BLrWld mice squares . A negative difference score is an indicator of hyperalgesia in the experimental limb. Preliminary tests immediately prior to

Ž .surgery day 0 indicate no difference in the average latencies between limbs in either group of mice. Thermal hyperalgesia is present throughout thetesting period from day 3 to 9 in C57BLr6 mice. On postoperative days 3–7, hyperalgesia in the C57BLrWld group is significantly reduced with respectto the control group. ) ) p-0.001.


Fig. 2. Withdrawal thresholds to mechanical stimulation with von Frey hairs following unilateral sciatic nerve injury are plotted for two groups of animalsŽ . Ž .each: C57BLr6 mice closed diamonds: OP side, open diamonds: sham side and C57BLrWld mice closed squares: OP side, open square: sham side .

The period of reduced withdrawal thresholds is prolonged in C57BLrWld mice.

3.2. Histology and immunohistochemistry

The myelin stain revealed a significant reduction inŽmyelin density from day 8 in C57BLr6 mice 85"5% of

.the initial density and from day 28 in C57BLrWld miceŽ .87"4% of the initial density . Differences in myelindensity between strains were significant on day 8 and 12 inaccordance with a delay in degeneration in the

Ž .C57BLrWld mice Fig. 3a .Endoneurial macrophage numbers per mm2 increased

rapidly in C57BLr6 mice with a maximum on day 8 to 12Ž .750"71 and 740"33, respectively . In C57BLrWldmice, macrophages were only moderately increased on dayŽ .8 420"59 and reached the same level as the control

Ž .strain 760"63 on day 12 post surgery.In serial 1 mm cryosections, TNF-immunoreactivity

was identified in macrophages, Schwann cells and en-Ž .doneurial fibroblasts Fig. 4a,b . IL-1a and b-immuno-

staining delineated some Schwann cells of normal nerveŽ .Fig. 4c . Soon after nerve injury, more Schwann cells

Ž .were immunoreactive for both types of IL-1 Fig. 4d ,additionally fibroblast-like structures and macrophageswere stained.

In normal controls and on the sham side of operatedanimals, the percentage of endoneurial area immunoreac-tive for TNF varied between 1.3 and 2% for the differentgroups of animals with no significant difference betweenstrains. The percentage of endoneurial area immunoreac-tive for TNF was significantly increased from day 4 to 12

Ž .in C57BLr6 mice p-0.01 with a maximum on day 12of 5.13"1.3%. Nerves from C57BLrWld mice had sig-nificantly less TNF-immunoreactivity with an increase over

Ž . Ž .controls only on day 12 4.2"0.3%, p-0.05 Fig. 3c .In normal controls and on the sham side of operated

animals, the percentage of endoneurial area immunoreac-

tive for IL-1a and b was approximately 1%. This per-centage was increased in both strains from day 12 on witha first trend already on day 4. The maximum occurred on

Žday 28 IL-1b : 1.7"0,3% in C57BLr6 mice, 1.6"0.1%in C57BLrWld mice, Fig. 3d; IL-1a 1.7 " 0.2 in

.C57BLr6 mice, 1.7"0.1 in C57BLrWld mice .

4. Discussion

By comparison of two mouse strains with different ratesof Wallerian degeneration, we could show that thermalhyperalgesia in mice with CCI is temporally related to theonset of Wallerian degeneration, of macrophage influx andto the presence of endoneurial TNF. In C57BLrWld mice,the peak increase in endoneurial TNF after nerve injurywas delayed and diminished as was the peak in thermalhyperalgesia, whereas IL-1a and b-immunoreactivity wereincreased to the same degree as in C57BLr6 mice. Fur-thermore, reduced thresholds to mechanical stimuli werepresent for a prolonged period in C57BLrWld mice.

4.1. BehaÕior

We could show that the model of painful mononeuropa-thy created by constriction injury of the sciatic nerve canbe applied to mice, and that withdrawal thresholds tothermal and mechanical stimuli can be reliably measured.Thermal hyperalgesia and mechanical allodynia ofC57BLr6 mice with CCI followed a similar temporal

w xcourse as described for rats 24 . Like in previous experi-w xments 16,18 , in C57BLrWld mice, thermal hyperalgesia

was significantly reduced during the first postoperative


week compared to the control strain. During this period,Wallerian degeneration was already advanced in C57BLr6mice, indicated by increased macrophage numbers andendoneurial TNF-immunoreactivity, but hardly present yetin the C57BLrWld mice. These findings suggest a possi-ble causal relationship between thermal hyperalgesia andcytokine action. The onset of mechanical allodynia, incontrast, did not differ between the strains. Both strainsdeveloped reduced withdrawal thresholds to von Frey hairsby day 10 after surgery. However, mechanical allodyniawas present for a much longer period in C57BLrWld micethan in C57BLr6 mice. In the fifth postoperative week,when withdrawal thresholds to mechanical stimuli wereback to normal in C57BLr6 mice but still reduced inC57BLrWld mice, the latter have reached an early stageof regeneration that C57BLr6 mice reach already by week

w xthree 18 . Thus, the period of prolonged mechanical allo-

dynia corresponds best to the delay in regeneration ofw xsensory nerve fibers in C57BLrWld mice 3,10 . In a

w xrecent study 21 , both thermal and mechanical sensitivitywere reduced in C57BLrWld mice with CCI during theobservation period of 20 days. The variability in theCCI-model in different laboratories due to type of sutureand variable tightness of the ligatures may be one explana-tion for this discrepancy.

In this, as well as in other models of neuropathic pain,the onset of mechanical allodynia was delayed comparedto other signs of neuropathic pain. There are severalpossible explanations for this observation. Allodynia mayrequire ectopic signaling from C-fiber sprouts, which ismaximal at later time points after CCI and almost absent

w xduring the first postoperative days 11,35 . Second, allody-nia may require central reorganisation so that second orderneurons usually responsive to input from nociceptive fibers

Ž . Ž .Fig. 3. a Morphometric determination of myelin density as a measure of Wallerian degeneration. b Morphometric quantification of macrophage content) Ž . Ž .after immunohistochemical staining with a Mac-1 antibody. : Significant difference p-0.05 between strains. c TNF-immunoreactivity in the

Ž . Ž . ) ) )endoneurium. d IL-1b immunoreactivity in the endoneurium. Almost identical results were obtained for IL-1a data not shown . and : Significantdifferences to controls, ) p-0.05, ) ) p-0.01. ns5 mice of each strain at each time point.


Ž .Fig. 3 continued .

only become responsive to low threshold mechanical inputw x Ž .6 . Changes in presumably inhibitory interneurons in the

w xdorsal horn have been observed in the CCI-model 20 andchanges in the expression of opioid receptors and neuro-

w xtransmitters occur 4,25,27 . However, the observation thatallodynia in the CCI-model is reversible with nerve regen-eration favors a peripheral influence on mechanical hyper-sensitivity. The hypothesis that increased ectopic activityfrom regenerating sprouts is responsible for allodynia is inaccordance with the present results, where a delay inregeneration and thus prolonged presence of sprouts isaccompanied by prolonged allodynia.

4.2. Cytokine studies

The increase in TNF immunoreactivity in C57BLr6mice followed the same time course as myelin loss andmacrophage influx. In C57BLrWld mice, the increase in

TNF-immunoreactivity was delayed and diminished in am-plitude. The macrophage peak of C57BLrWld mice wasalso delayed but reached the same magnitude as inC57BLr6 mice. Assuming that macrophages are the mainproducers of TNF, a delayed but equally high peak ofTNF-immunoreactivity would have been expected. Wemay not have picked the optimal time points to detect theTNF peak in C57BLrWld, which might occur betweenday 12 and day 28. Alternatively, due to the defectivesignaling of the damaged axons in C57BLrWld mice,their macrophages may be less activated and thus produce

w xsmaller amounts of TNF. According to Stoll et al. 28 ,only pre-phagocytotic macrophages in nerves with Walle-rian degeneration are immunoreactive for TNF. Also,Schwann cells are major producers of TNF after nerve

Žw x .injury 12,30 , present study , and lack of Schwann cellactivation in C57BLrWld mice may cause diminishedupregulation of endoneurial TNF.


Ž .Fig. 4. a Serial 1 mm cryosections of sciatic nerve from a C57BLr6 mouse with CCI, 4 days after CCI. TNF-immunoreactivity is localized to SchwannŽ . Ž . Ž .cells arrows , and resident macrophages or fibroblasts arrowheads . b Serial 1 mm cryosections of sciatic nerve from a C57BLr6 mouse with CCI, 4

Ž . Ž . Ž .days after CCI. S-100-immunoreactivity arrows corresponds to the TNF-immunoreactive Schwann-cell profiles in a . c 12 mm cryostat-sections,C57BLrWld mouse day 12 after CCI, immunostained for IL-1b. Sham operated side. Normal Schwann cells are strongly immunoreactive for IL-1b

Ž . Ž .arrows . d 12 mm cryostat-sections, C57BLrWld mouse day 12 after CCI, immunostained for IL-1b Operated side. IL-1b-immunoreactivity can beŽ . Ž . Ž .seen in activated Schwann cells arrows , presumable macrophages open arrows and fibroblast-processes arrowhead . Bars10 mm.

IL-1 immunostaining revealed a similar increase afterCCI in both mouse strains. The peak at day 28 occurredlate compared with data from the literature for nerve crushw x22,33 . However, again we cannot be sure whether we didnot miss an earlier peak between the time points studiedhere. Also, IL-1 immunostaining delineates very well theSchwann cells of normal nerve. Soon after nerve injury,this distinct immunoreactivity in the Schwann cells is lost,instead, more fibroblast-like structures and macrophages

stain positively for IL-1. Whether the loss of cytokine-im-munoreactivity in early denervated Schwann cells in factreflects loss of IL-1 expression, is not known. DenervatedSchwann cells may not retain their antigenicity as well asintact Schwann cells during tissue preparation. Studieswith ELISA and in situ hybridization are underway tofurther address this question. Another shortcoming of den-sitometry of immunostained sections, which is useful as arapid semiquantitative screening method, is that compara-


bility between slides has to be maintained by staining oflarge sample sizes at a time. Thus the semiquantitativevalues obtained by this method cannot be compared toresults from earlier or later studies, with separately stainedtissue.

There are as yet few reports on endoneurial cytokine-protein after nerve injury. Systemic levels of TNF, IL-1b

and IL-6 are increased immediately after nerve crush andsham operation and selectively between day 12 and 14

w xafter nerve crush 33 . These systemic levels as determinedin serum probably do not adequately reflect changes in

w x Žlocal cytokine concentration in the endoneurium 31 Geo-.rge and Sommer, unpublished observations . Rotshenker et

w xal. 22 measured IL-1b in conditioned medium frominjured nerve and found an increase within hours afterlesion, and 1 week thereafter. The same group recentlydescribed an increase of IL-6 as early as 2 h after injury,

w xpersisting for 21 days 36 . Of the nonneuronal cells,macrophages and fibroblasts were the major contributorsto IL-6 production. Injured nerves of C57BLrWLD miceproduced significantly lower amounts of IL-6 than didrapidly degenerating nerves of C57BLr6 mice. TNF-mRNA has recently been shown to be increased up to6-fold in the distal nerve segment after crush with an earlypeak at one day and a later increase from four days after

w xthe injury 12 .

4.3. Cytokines and neuropathic pain

Cytokines have long been implied to play a role inw xinflammatory pain 9 , and clinical trials with neutralizing

antibodies to TNF in rheumatoid arthritis have confirmedw xthis association 13 . Patients treated with neutralizing

antibodies had significantly less pain as well as decreasedarthricular inflammatory activity. Similar results were ob-

w xtained by reducing the levels of IL-1b 8 . The role ofcytokines in neuropathic pain is less clear, although thereis increasing evidence that cytokines are involved in thecausation of pain after nerve injury or inflammation atleast at certain stages. One example of inflammatory nervepain is lepromatous leprosy, which is associated with amassive increase in the production of TNF and where paincan be dramatically reduced by inhibitors of TNF-produc-

w xtion like thalidomide 1 . Exogenous TNF, applied to thesciatic nerve, induces thermal hyperalgesia in rats andincreases spontaneous activity in afferent nerve fibersw x29,33 . Also, in neuropathic pain after a constrictive nerveinjury, treatment with thalidomide reduces pain related

w xbehavior in experimental animals 26 .In the present study, we have extended the knowledge

on the role of cytokines in neuropathic pain. We couldŽ .show that 1 thermal hyperalgesia in mice with CCI is

temporally related to the onset of Wallerian degeneration,to macrophage influx and the presence of endoneurial

Ž .TNF, and 2 that mechanical allodynia in mice with CCIis temporally related to regeneration of nerve fibers. Weconclude that the various events associated with Wallerian

degeneration and regeneration, including cytokine expres-sion, differentially influence parameters of neuropathicpain.

Acknowledgements

We thank Barbara Dekant, Lydia Biko and ChristineSchmidt for their assistance in these studies. Critical read-ing of the manuscript by K.V. Toyka is gratefully ac-knowledged. Supported by Deutsche Forschungsgemein-schaft So 328r2-1.

References

w x1 R. Barnhill, A. McDougall, Thalidomide: use and possible mode ofaction in reactional lepromatous leprosy and in various other condi-

Ž .tions, J. Am. Acad. Dermatol. 7 1982 317–323.w x2 G.J. Bennett, Y.K. Xie, A peripheral mononeuropathy in rat that

produces disorders of pain sensation like those seen in man, Pain 33Ž .1988 87–107.

w x3 M.A. Bisby, S. Chen, Delayed wallerian degeneration in sciaticnerves of C57BLrOla mice is associated with impaired regeneration

Ž .of sensory axons, Brain Res. 530 1990 117–120.w x4 A. Cameron, K. Cliffer, P. Dougherty, W. Willis, S. Carlton,

Changes in lectin, GAP-43 and neuropeptide staining in the ratsuperficial dorsal horn following experimental peripheral neuropa-

Ž .thy, Neurosci. Lett. 131 1991 249–252.w x5 J.A. DeLeo, R.W. Colburn, A.J. Rickman, Cytokine and growth

factor immunohistochemical spinal profiles in two animal models ofŽ .mononeuropathy, Brain Res. 759 1997 50–57.

w x6 M. Devor, The pathophysiology of damaged peripheral nerves, in: P.Ž .Wall, R. Melzack Eds. , Textbook of pain, Churchill Livingston,

Edinburgh, 1994, pp. 79–100.w x7 W. Dixon, The up-and-down method for small samples, J. Am. Stat.

Ž .Assoc. 60 1965 967–978.w x8 M.J. Elliott, R.N. Maini, Anti-cytokine therapy in rheumatoid arthri-

Ž .tis, Baillieres Clin. Rheumatol. 9 1995 633–652.w x9 S. Ferreira, Inflammatory pain: the role of cytokines and its control

by drugs which release nitric oxide, Ann. Ist. Super. Sanita 29´Ž .1993 367–373.

w x10 M. Fruttiger, M. Schachner, R. Martini, Tenascin-C expressionduring wallerian degeneration in C57BLrWlds mice: possible impli-

Ž .cations for axonal regeneration, J. Neurocytol. 24 1995 1–14.w x11 K.C. Kajander, G.J. Bennett, Onset of a painful peripheral neuropa-

thy in rat: a partial and differential deafferentiation and spontaneousdischarge in A b and AE primary afferents, J. Neurophysiol. 68Ž .1992 734–744.

w x12 M. La Fleur, J.L. Underwood, D.A. Rappolee, Z. Werb, Basementmembrane and repair of injury to peripheral nerve: defining apotential role for macrophages, matrix metalloproteinases, and tissue

Ž .inhibitor of metalloproteinases-1, J. Exp. Med. 184 1996 2311–2326.

w x13 H.-M. Lorenz, C. Antoni, T. Valerius, R. Repp, M. Grunke, N.¨Schwerdtner, H. Nublein, J. Woody, J. Kalden, B. Manger, In vivo¨blockade of TNF-alpha by intravenous infusion of a chimeric mono-clonal TNF-alpha antibody in patients with rheumatoid arthritis, J.

Ž .Immunol. 156 1996 1646–1653.w x14 E. Lunn, V. Perry, M. Brown, H. Rosen, S. Gordon, Absence of

Wallerian degeneration does not hinder regeneration in peripheralŽ .nerve, Eur. J. Neurosci. 1 1989 27–33.


w x15 M. Lyon, B. Ogunkolade, M. Brown, D. Atherton, V. Perry, A geneaffecting Wallerian nerve degeneration maps distally on mouse

Ž .chromosome 4, Proc. Natl. Acad. Sci. USA 90 1993 9717–9720.w x16 R. Myers, C. Sommer, Role of Wallerian degeneration in pathogene-

Ž .sis of neuropathic pain, J. Neuropathol. Exp. Neurol. 52 1993 310.w x17 R.R. Myers, H.M. Heckman, H.C. Powell, Axonal viability and the

persistence of thermal hyperalgesia after partial freeze lesions ofŽ .nerve, J. Neurol. Sci. 139 1996 28–38.

w x18 R.R. Myers, H.M. Heckman, M. Rodriguez, Reduced hyperalgesiain nerve-injured WLD mice: relationship to nerve fiber phago-cytosis, axonal degeneration, and regeneration in normal mice, Exp.

Ž .Neurol. 141 1996 94–101.w x19 R.R. Myers, T.Y. Yamamoto, T.L. Yaksh, H.C. Powell, The role of

focal nerve ischemia and Wallerian degeneration in peripheral nerveŽ .injury producing hyperesthesia, Anesthesiology 78 1993 308–316.

w x20 A.K. Nachemson, G.J. Bennett, Does pain damage spinal cordneurons? Transsynaptic degeneration in rat following a surgical

Ž .incision, Neurosci. Lett. 162 1993 78–80.w x21 M.S. Ramer, G.D. French, M.A. Bisby, Wallerian degeneration is

required for both neuropathic pain and sympathetic sprouting intoŽ .the DRG, Pain 72 1997 71–78.

w x22 S. Rotshenker, S. Aamar, V. Barak, Interleukin-1 activity in lesionedŽ .peripheral nerve, J. Neuroimmunol. 39 1992 75–80.

w x23 B. Schmidt, G. Stoll, H.P. Hartung, K. Heininger, B. Schafer, K.V.Toyka, Macrophages but not Schwann cells express Ia antigen in

Ž .experimental autoimmune neuritis, Ann. Neurol. 28 1990 70–77.w x24 C. Sommer, J.A. Galbraith, H.M. Heckman, R.R. Myers, Pathology

of experimental compression neuropathy producing hyperesthesia, J.Ž .Neuropathol. Exp. Neurol. 52 1993 223–233.

w x25 C. Sommer, R.R. Myers, Neurotransmitters in the spinal cord in amodel of painful neuropathy and in nerve crush, Acta Neuropathol.

Ž .90 1995 478–485.

w x26 C. Sommer, M. Marziniak, R.R. Myers, The effect of thalidomidetreatment on vascular pathology and hyperalgesia caused by chronic

Ž .constriction injury of rat nerve, Pain 1998 , in press.w x27 C.W. Stevens, K.C. Kajander, G.J. Bennett, V.S. Seybold, Bilateral

and differential changes in spinal mu, delta and kappa opioidŽ .binding in rats with a painful, unilateral neuropathy, Pain 46 1991

315–326.w x28 G. Stoll, S. Jung, S. Jander, P.v.d. Meide, H.-P. Hartung, Tumor

necrosis factor-alpha in immune-mediated demyelination and Walle-rian degeneration of the rat peripheral nervous system, J. Neuroim-

Ž .munol. 45 1993 175–182.w x29 R. Wagner, R.R. Myers, Endoneurial injection of TNF-alpha pro-

Ž .duces neuropathic pain behaviors, Neuroreport 7 1996 2897–2901.w x30 R. Wagner, R.R. Myers, Schwann cells produce tumor necrosis

factor alpha: expression in injured and non-injured nerves, Neuro-Ž .science 73 1996 625–629.

w x31 C.X. Wang, J.A. Olschowka, J.R. Wrathhall, Increase of interleukin-1b mRNA and protein in the spinal cord following experimental

Ž .traumatic injury in the rat, Brain Res. 759 1997 190–196.w x32 L.R. Watkins, S.F. Maier, L.E. Goehler, Immune activation: the role

of pro-inflammatory cytokines in inflammation, illness responsesŽ .and pathological pain states, Pain 63 1995 289–302.

w x33 M. Wells, S.J. Racis, U. Vaidya, Changes in plasma cytokinesŽ .associated with peripheral nerve injury, J. Neuroimmunol. 39 1992

261–268.w x34 Y. Xie, W. Xiao, Electrophysiological evidence for hyperalgesia in

Ž .the peripheral neuropathy, Sci. China B 33 1990 663–672.w x35 F. Reichert, R. Levitzky, S. Rotshenker, Interleukin 6 in intact and

Ž .injured mouse peripheral nerves, Eur. J. Neurosci. 8 1996 530–535.w x36 L. Sorkin, W. Xiao, R. Wagner, R. Myers, Tumour necrosis factor-a

induces ectopic activity in nociceptive primary afferent fibers, Neu-Ž .roscience 81 1997 255–262.

painful mononeuropathy in c57bl/wld mice with delayed wallerian degeneration: differential effects...

Documents