Reduced Hyperalgesia in Nerve-Injured WLD Mice: Relationshipto Nerve Fiber Phagocytosis, Axonal Degeneration,

and Regeneration in Normal Mice

ROBERT R. MYERS, HEIDI M. HECKMAN, AND MARIO RODRIGUEZ

Department of Veterans Affairs and the Departments of Anesthesiology and Pathology (Neuropathology),University of California, San Diego, La Jolla, California 92093-0629

The pathogenesis of neuropathic pain states is influ-enced by inflammatory factors associated with nerveinjuries and may be mediated in part by the macro-phage-dependent process of Wallerian degeneration.Macrophages play a dominant role in the Wallerian(axonal) degeneration that characterizes the painfulchronic constriction injury model of neuropathy byliberating proinflammatory cytokines at the site ofnerve injury. These cytokines directly affect the struc-tural integrity of neural systems and have been impli-cated in the development of hyperalgesic states. Wehypothesized that interference with the pathologicprocess of Wallerian degeneration would alter thedevelopment of the neuropathic pain state. To test thishypothesis, we studied the development of thermalhyperalgesia in the chronic constriction injury modelof neuropathy using normal mice and mice of the WLDstrain in which recruitment of macrophages to the siteof nerve injury and Wallerian degeneration are de-layed. We compared the onset and magnitude of thehyperalgesia with quantitative measures of nerve in-jury including the phagocytic cellular activity associ-ated with Wallerian degeneration. In C57BL/6J (6J)mice, hyperalgesia peaked 3–10 days after placementof the ligatures, qualitatively matching the responsepreviously reported for rats. In C57BL/WLD (WLD)mice, there was reduced hyperalgesia temporally asso-ciated with reduced numbers of phagocytic cells in theinjured nerve. In injured WLD nerves there was areduced rate of axonal degeneration compared to simi-larly injured 6J nerves. Regeneration was correspond-ingly delayed in the WLD animals. The results suggestthat the process of Wallerian degeneration is a keyfactor in the pathogenesis of hyperalgesia. Continueddevelopment of mousemodels of neuropathic pain willbe important in exploring themolecular basis of neuro-pathic pain. Interference with the cellular mediatorsof Wallerian degeneration may be a useful therapeuticstrategy that might modulate the onset and magni-tudeofhyperalgesia followingnerve injury. r 1996Academic

Press, Inc.

INTRODUCTION

Insights into the mechanisms of neuropathic painhave increased rapidly since the development of anonneuroma model of pain (1). In this model, a focalchronic constriction injury (CCI) to nerve is created byloosely tying four ligatures around one sciatic nerve inthe rat and the resultant thermal hyperalgesia isfollowed temporally by sensitive behavioral testing ofthe affected footpad (4). The CCI model is an importantsupplement to neuroma models of pain in which thenerve is completely transected. Unlike neuroma mod-els, some fibers are spared from injury in the CCImodeland hyperalgesia develops within several days, duringthe period of nerve fiber degeneration, rather thanduring the later period of axonal regeneration as is thecase in neuromamodels (9, 17).An important character-istic of the CCI model is the intense inflammationassociated with the nerve injury and the axonal degen-eration of nearly all myelinated fibers occurring duringthe period immediately preceeding and encompassinghyperalgesia. This Wallerian degeneration of the axonis a macrophage-dependent process which we havelinked to the development of sustained hyperalgesia inexperimental animals (12). Macrophages are the domi-nant cell in the process of Wallerian degeneration andtheir associated cytokines provide stimuli for localmitotic activity and altered function of support cellsand neurons in the afferent pathway. We speculatedthat macrophages or the effects of their cytokinesprovided a chemical signal that influenced the develop-ment of hyperalgesia and that interference with theseactivities may influence the magnitude and duration ofneuropathic pain states. In support of this hypothesisare preliminary experiments which show that oraltreatment with thalidomide to reduce macrophage pro-duction of tumor necrosis factor reduces the magnitudeof hyperalgesia in rats with CCI neuropathy (18).In the current experiments, we explored the patho-

genic link between the cellular response to axonalinjury and the development of neuropathic pain bydeveloping the CCImodel inmurine species and testing

EXPERIMENTAL NEUROLOGY 141, 94–101 (1996)ARTICLE NO. 0142

940014-4886/96 $18.00Copyright r 1996 by Academic Press, Inc.All rights of reproduction in any form reserved.

the behavioral response to thermal stimuli in C57BL/6J(6J) mice and in mice of the C57BL/WLD (WLD) straininwhich there is a abnormal rate ofWallerian degenera-tion due to a delay in recruitment of hematogeneousmacrophages to the site of nerve injury. We hypoth-esized that this delay would be reflected in a delayedonset and reducedmagnitude of the thermal hyperalge-sic behavior that is characteristic of the CCI model ofneuropathic pain.

MATERIALS AND METHODS

Experiments were approved by the local AnimalStudies Committee and animals were cared for in anAAALAC-approved animal care facility. Eight 6J miceand 8 WLD mice (female, 16–21 g) were anesthetizedby IP injection of 0.2 ml/100 g body weight of a solutioncontaining sodium pentobarbital (50 mg/ml), diazepam(5 mg/ml), and saline in volume proportions of 1:1:4,respectively, and one sciatic nerve was exposed unilat-erally at the mid-thigh level. Three ligatures (10-Oprolene) were placed around either the right or the leftnerve (randomly assigned) with 1 mm spacing. Theligatures were tied until they elicited a brief twitch inthe hindlimb. In the rat, this corresponded to blanchingof the nerve and a significant reduction in nerve bloodflow secondary to occlusion of the epineurial vascula-ture (12). The wound was closed with surgical staples.The thermal nociceptive threshold was measured ineach hindpaw before surgery and at regular intervalsfrom Day 3 after surgery using a sensitive thermaltesting device and protocol that allows each animal toserve as its own control (4). This technique uses a pointsource to heat selectively only one foot at a time ratherthan heating both feet simultaneously. The method isideally suited to investigations of neuropathic paininvolving unilateral lesions to peripheral nerves. Byperforming the test separately on each foot and thensubtracting the latency to removal of the control footfrom the latency to removal of the experimental foot, a‘‘difference score’’ between the control and experimen-tal feet that reflects the degree of hyperalgesia in theexperimental foot can be generated. In control animalswith no nerve lesions, the difference score in rats has aGaussian distribution with an average of 0.04 6 0.66 s(SD) (2). In mice, baseline difference scores averaged20.03 6 0.20 (SD; data not shown). In our study, eachmouse was tested six times for 3 days precedingsurgery and on each experimental day by an operatorunaware of the experimental grouping. The averagedifference score was determined for each group ofanimals and the mean value and standard deviationswere plotted versus time. The significance of differ-ences between groups was assessed with two-factorANOVA for repeated measures and by one-factor

ANOVA for differences between individual days oftesting.These 16 animals were sacrificed 28 days after

surgery and the sciatic nerves were processed forhistology. Seven additional groups of 8 6J and 8 WLDanimals were treated in the same way and used formorphometric analysis of the nerve injury. These sci-atic nerves were removed on Day 3, 5, 7, 14, 21, 28, or35 after surgery. The nerve segments were fixed byimmersion in phosphate-buffered 2.5% glutaraldehydefor 48 h, after which the ligatures were removed by finedissection and the tissue cut into 2-mm blocks forprocessing. Following dehydration, osmium infiltra-tion, and embedding in araldite, 1-µm-thick sectionswere cut for lightmicroscopy and stained with toluidineblue or paraphenylenediamine.Morphometric analysis was performed on 1-µm-thick

sections of tissue viewed in an Olympus BH-2 lightmicroscope with an attached Cohu 5000 series televi-sion camera interfaced with a Macintosh Quadra 900computer. One or more images of each nerve section(depending on the fascicular area) was electronicallycaptured at 2003. An electronic grid with 10 3 10squares was overlaid and at each grid intersectionpoint the underlying structure was identified andcounted. The identification categories were intact my-elinated fiber, structureless space (edema), phagocyticcell, regenerating fiber, and ‘‘other’’ which includedstructures not peretinent to the experimental question.Since macrophages and Schwann cells react to injuriesin many of the same ways with phagocytosis andcytokine expression, we categorized these cells as asingle morphologic identity (phagocytic cells) in lightmicrographic quantification of the injury. Morphometryresults are expressed as a percentage of all data points.This is a standard stereological technique for quantita-tive analysis of injured nerve (19). The program wasdeveloped in our laboratory and adds functionality toimages captured in the NIH Image 1.55 softwarepackage running on a Macintosh platform. The signifi-cance of the morphological results was tested withFisher PLSD post hoc analysis following one-factorANOVA between experimental groups at each timepoint.

RESULTS

Motor function was affected by placement of the CCIligatures and the animals occasionally dragged thelimb and tended to avoid its use, but otherwise did notappear to give it extra attention. The behavioral re-sponse to heating of the injured paw in 6J animals wasqualitatively similar to those reported in rats (2, 17) inthat hyperalgesia developedwithin several days, peakedduring the next week, and then diminished linearly

95HYPERALGESIA IN WLD MICE

until a normal response was obtained several weekslater (Fig. 1).The WLD mice displayed an initial, modest hyperal-

gesic response at Day 3. This was the peak of theirhyperalgesia, however. From Day 5 and after, the WLDgroup was not hyperalgesic and differed significantly(0.004 , P , 0.04) from the 6J group whose hyperalge-sia peaked betweenDays 3 and 17 (Fig. 1). FromDay 17to Day 28 there was no significant difference in behav-ioral scores between the groups as both had moderatehyperalgesia that resolved by Day 28.Histologically, at 3 days postsurgery, edema was the

predominate pathological finding and there were veryfew degenerating axons in tissue from the WLD mice(Fig. 2a). In 6J mice, at this time point, there wasalready considerable breakdown of axons and myelin(Fig. 2b). At 7 days, many large myelinated fibers werestill well preserved in the WLD group, whereas in the6J mice, most large myelinated fibers were undergoingWallerian degeneration (Figs. 2c, 2d). Four weeks aftersurgery, there were still more large myelinated fiberspresent in the WLD mice then in the 6J mice, but thedifference was less marked (Figs. 2e, 2f).

Morphometric analysis of the tissue reinforcedthese findings. There was substantial edema in theinjured 6J nerves that peaked 14 days after nerveinjury, whereas there was less edema in WLD ani-mals which peaked later, at 21 days postinjury (Fig. 3).Three days after nerve injury only 20% of the fasciclewas occupied by intact nerve fibers in the injured 6Jgroup, while nearly twice that number remained inthe WLD group (Fig. 4). By 14 days, the number ofintact fibers in the WLD group had fallen to the le-vel seen in 6J animals at 3 days postinjury, but the levelin 6J animals at the 14-day time point was less than5%. By 21 days postinjury the WLD group reachedthis value. Regeneration was delayed in WLD animals(Fig. 5).The percentage of endoneurial space occupied by

phagocytic cells rose rapidly in injured 6J animalsbeginning 3–5 days after nerve injury and then pla-teaued between weeks 2 and 3 before peaking at 28days (Fig. 6). In WLD nerves, there was only a modestrise in phagocytic cell numbers until Day 21 when therewas a rapid rise in cell numbers that paralled the riseseen in 6J animals 2–3 weeks earlier.

FIG. 1. Hyperalgesia to thermal stimuli following unilateral sciatic nerve injury is plotted for two groups of 10 animals each: C57BL/6Jmice and C57BL/WLD mice. A negative difference score is an indicator of hyperalgesia in the experimental limb. Preliminary testsimmediately prior to surgery (Day 0) indicate no difference in the average latencies between limbs in either group of mice. By Day 3 postinjury,both groups are hyperalgesic. On subsequent testing days, however, the hyperalgesia in theWLD group is significantly reduced with respect tothe control group whose hyperalgesia peaks between postoperative Day 3 and postoperative Day 17. From Day 17 to Day 28 there is nosignificant difference between groups in their behavioral response to thermal stimuli.

96 MYERS, HECKMAN, AND RODRIGUEZ

DISCUSSION

It has been shown that the model of painful mono-neuropathy created by chronic constriction injury ofthe sciatic nerve can be applied to mice and that theirthermal hyperalgesia follows the same temporal courseas in rats. This demonstration has allowed us to be-gin hyperalgesia studies in mice with genetic defects

in nervous system function. The experimental use ofsuch mice is a proven and powerful tool in neuropa-thy research since the availability of monoclonal anti-bodies to neuroimmunologically important moleculescan support many interesting hypotheses. The useof mice in pain research will help dissect the impor-tant pathogenic factors leading to states of chronicpain.

FIG. 2. Histological sections of peripheral nerve were compared from control and WLD animals at Days 3, 7, and 28 postinjury to verifythat Wallerian degeneration had been delayed in the WLD group. Photomicrographs of sciatic nerve distal to the loose ligatures from WLDmice (a,c,e) and control mice (b,d,f) are shown here. At 3 days postinjury, there are very few degenerating axons in tissue from the WLD mice,whereas 6J mice already show a considerable amount of axon and myelin breakdown. At 7 days, many large myelinated fibers are still wellpreserved in theWLDmice (c). In the control mice, most largemyelinated fibers are undergoingWallerian degeneration (d). Twenty-eight daysafter surgery, there are still more large myelinated nerve fibers present in the WLD mice (e) than in the controls (f), but the difference is lessmarked.

97HYPERALGESIA IN WLD MICE

FIG. 3. Percentage of fascicle occupied by structureless space. Structureless space is a measure of edema and increases as a function ofWallerian degeneration. Edema peaks in 6J animals 14 days after nerve injury while it peaks 1 week later in WLD animals with delayedWallerian degeneration. Significant difference (P , 0.05) at Days 3, 7, 14, and 28.

FIG. 4. Percentage of fascicle occupied by intact myelinated nerve fibers. Nerve fiber degeneration occurs rapidly in nerve-injured 6J micebut is delayed in WLD mice. The minimum number of intact fibers occurs between postoperative week 2 and postoperative week 3 in 6J miceand between 3 and 4 weeks in WLDmice. Significant difference (P , 0.05) at Days 3, 7, and 14.

98 MYERS, HECKMAN, AND RODRIGUEZ

In WLD mice, in which there is a genetic defect thatdelays the process of Wallerian degeneration, we haveobserved that the hyperalgesic response was signifi-cantly reduced between Day 5 and Day 17 whencompared to normal animals with the same lesion. Thistemporal period corresponds to the delay in Walleriandegeneration which was advanced in the 6J mice, buthardly present in the WLD mice. Between Day 21 andDay 30, WLD and 6J mice did not differ in theirbehavioral response to heat stimuli when the resolu-tion ofWallerian degeneration in 6Jmice was similar tothe degree of Wallerian degeneration seen in WLDmice.The WLD mice had an initial painful response on

Day 3 which was unexpected. This initial hyperalgesiamay be due to the large amplitude electrophysiologic‘‘injury discharge’’ and its effect on dorsal root gangliaand central neurons (8), a process that is independentof macrophage activity.Alternatively, the initial painfulresponse could be triggered by resident macrophages orSchwann cells in the endoneurium at the site of injurywhich are activated in both 6J and WLD animals.These cells produce tumor necrosis factor alpha inresponse to nerve injury (22). The initial, local cytokineand phagocytic functions of resident macrophages andSchwann cells are reinforced and extended distally bymacrophages recruited from the systemic circulation(hematogeneous macrophages). It is this second waveof macrophage activity which normally begins 3–5 days

after injury that is delayed and reduced in WLDanimals. The reduced hyperalgesia observed betweenDay 3 and Day 9 in WLD animals is supported byevidence of diminishedWallerian degeneration inWLDanimals during this time (Fig. 2) and by reducednumbers of endoneurial macrophages at the injury siteand distally (20).That macrophages are the primary effector cell in

Wallerian degeneration has been established since1984 when it was shown that axons did not degeneratewhen they were isolated in a millipore chamber thatexcluded the entry of nonresident macrophages (3).Wallerian degeneration can also be inhibited by prevent-ing the recruitment of macrophages with a monoclonalantibody against their complement type 3 receptor (15,16). In WLD animals there appears to be an absence inthe required chemotactic signal from injured axons orSchwann cells (13). Macrophage function, per se, isnormal in WLD animals since macrophages respondappropriately to injuries outside the nervous system.One factor that may be related to this delay is thereduced level of TNF protein in WLD animals (20). Wehave identified TNF protein immunohistochemically inSchwann cell cytoplasm of normal rats and have seenthat its production is increased following nerve injury(22). While the role of sequestered TNF protein inSchwann cells is not yet clear, it may have an earlyeffect on the cascade of pathological events associatedwith Wallerian degeneration. Such effects may include

FIG. 5. The percentage of fascicle occupied by regenerating nerve fibers observed with light microscopy as small, thinly myelinatedstructures. Regeneration is seen to begin during the third postoperative week in 6J mice, but is delayed until the fifth week in WLD mice.Significant difference (P , 0.05) at day 28.

99HYPERALGESIA IN WLD MICE

altered Schwann cell communication and activation,demyelination, Schwann cell phagocytosis, and TNF-induced recruitment of macrophages to the site ofinjury. Attenuation of any of these processes couldconceivably affect temporal rates of degeneration andthe severity of the pathological processes known to beactivated in part by cytokines.It has been shown before that there are changes in

plasma cytokines associated with peripheral nerveinjury (23). Macrophage-derived interleukin-1 stimu-lates Schwann cells to express nerve growth factor(NGF) on their surfaces (11), and retrograde transportof NGF from the injury site is an important signalaffecting the function of dorsal root ganglia neurons(5–7). Recently some authors have suggested that NGFor an octapeptide of NGF is painful when injectedperipherally, but it is not clear by what mechanism thismay be mediated (10, 21). Tumor necrosis factor-alpha(TNF) liberated from activated macrophages may alsohave a role in the pathogenesis of neuropathic painsince TNF has been shown to induce demyelination andaxonal degeneration when injected into rat nerves (14).Our recent work shows that this produces abnormalpain behaviors as well as nerve fiber injury (Wagnerand Myers, unpublished). TNF is a mediator of leuko-cyte recruitment (1) and we believe that it is associatedwith the pathological alterations in several endoneu-

rial cell lines in the rat model of this neuropathy (12,17, 22).Macrophages produce many cytokines and other

secretory factors, and it is not yet clear to what degreeeach is involved in the structural and behavioral mani-festations of neuropathic pain states. We have shownthat the lack of hyperalgesia in WLD mice 5–17 daysafter focal nerve injury is temporally related to delayedrecruitment of nonresident macrophages and the devel-opment ofWallerian degeneration, the identifying char-acteristic in this strain of mice, and suggest that thecorresponding reduction in macrophage cytokine activ-ity at the site of nerve injury is involved in thepathogenesis of hyperalgesia. While the specific causalrelationship between nerve injury and hyperalgesia isnot known, these experiments suggest that the courseof neuropathic pain is influenced by activated hemato-geneous macrophages responding to peripheral nerveinjury and that altering macrophage function may be atool for modulating the course of neuropathic pain.

ACKNOWLEDGMENTS

We thank Dr. Claudia Sommer, Dr. Rochelle Wagner, and DavidMoore for their help in these studies. Supported by the Department ofVeteransAffairs and USPHS NIH Grant NS 18715.

FIG. 6. Percentage of fascicle occupied by phagocytic cells identified by light microscopy. Phagocytic cells include Schwann cells andmacrophages. In nerve-injured 6J mice, phagocytic activity associated with Wallerian degeneration begins aggressively at Day 5 andcontinues for the next 3 weeks. Cellular phagocytic activity and Wallerian degeneration are moderated and delayed in WLDmice. Significantdifference (P , 0.05) at Days 14, 21, and 35.

100 MYERS, HECKMAN, AND RODRIGUEZ

REFERENCES

1. ANDERSSON, P. B., V. H. PERRY, AND S. GORDON. 1992. Intracere-bral injection of proinflammatory cytokines or leukocyte chemo-taxins induces minimal myelomonocytic cell recruitment to theparenchyma of the central nervous system. J. Exp. Med. 176:255–259.

2. BENNETT, G. J., AND Y.-K. XIE. 1988. A peripheral mononeuropa-thy in rat that produces disorders of pain sensations like thoseseen in man. Pain 33: 87–107.

3. BEUCHE, W., AND R. L. FRIEDE. 1984. The role of non-residentcells in Wallerian degeneration. J. Neurocytol. 13: 767–796.

4. HARGREAVES, K., R. DUBNER, F. BROWN, C. FLORES, AND J. JORIS.1988. A new and sensitive method for measuring thermalnociception in cutaneous hyperalgesia. Pain 32: 77–88.

5. HENDRY, I. A., K. STOCKEL, H. THOENEN, AND L. L. IVERSEN. 1974.The retrograde axonal transport of nerve growth factor. BrainRes. 68: 103–121.

6. HEUMANN, R., D. LINDHOLM, C. BANDTLOW, M. MEYER, M. J.RADEKE, T. P. MISKO, E. SHOOTER, AND H. THOENEN. 1987.Differential regulation of mRNA encoding nerve growth factorand its receptor in rat sciatic nerve during development, degen-eration, and regeneration: Role of macrophages. Proc. Natl.Acad. Sci. USA 84: 8735–8739.

7. JOHNSON, E. M., JR, M. TANIUCHI, H. B. CLARK, J. E. SPRINGER, S.KOH, M. W. TAYRIEN, AND R. LOY. 1987. Demonstration of theretrograde transport of nerve growth factor receptor in theperipheral and central nervous system. J. Neurosci. 7: 923–929.

8. KAJANDER, K. C., S. WAKISAKA, AND G. J. BENNETT. 1992. Sponta-neous discharge originates in the dorsal root ganglion at theonset of a painful peripheral neuropathy in the rat. Neurosci.Lett. 138: 225–228.

9. KUPERS, R. C., D. NUYTTEN, M. DE CASTRO-COSTA, AND J. M.GYBELS. 1992.A time course analysis of the changes in spontane-ous and evoked behaviour in a rat model of neuropathic pain.Pain 50: 101–111.

10. LEWIN, G. R., AND L. M. MENDELL. 1992. NGF-induced heat- andmechanical hyperalgesia in adult rats. Soc. Neurosci. Abstr. 130.

11. LINDHOLM, D., R. HEUMANN, M. MEYER, AND H. THOENEN. 1987.Interleukin-1 regulates synthesis of nerve growth factor innon-neuronal cells of rat sciatic nerve.Nature 330: 658–659.

12. MYERS, R. R., T. YAMAMOTO, T. L. YAKSH, AND H. C. POWELL. 1993.

The role of focal nerve ischemia and Wallerian degeneration inperipheral nerve injury producing hyperesthesia. Anesthesiol-ogy 78: 308–316.

13. PERRY, V. H., M. C. BROWN, E. R. LUNN, AND S. GORDON. 1990.Evidence that very slow Wallerian degeneration in C57BL/Olamice is an intrinsic property of the peripheral nerve. Eur. J.Neurosci. 2: 802–812.

14. REDFORD, E. J., S. M. HALL, AND K. J. SMITH. 1995. Vascularchanges and demyelination induced by the intraneural injectionof tumour necrosis factor. Brain. 118: 869–878.

15. ROSEN, H., AND S. GORDON. 1987. Monoclonal antibody to themurine type 3 complement receptor inhibits adhesion of myelo-monocytic cells in vitro and inflammatory cell recruitment invivo. J. Exp. Med. 166: 1685–1702.

16. SAID, G., AND M. HONTEBEYRIE-JOSKOWICZ. 1992. Nerve lesionsinduced by macrophage activation. Res. Immunol. 143: 589–599.

17. SOMMER, C., J. A. GALBRAITH, H. M. HECKMAN, AND R. R. MYERS.1993. Pathology of experimental compression neuropathy pro-ducing hyperesthesia. J. Neuropathol. Exp. Neurol. 52: 223–233.

18. SOMMER, C., AND R. R. MYERS. 1994. Thalidomide inhibition ofTNF reduces hyperalgesia in neuropathic rats. Reg. Anesth. 19(2S): 1.

19. SOMMER, C., A. LALONDE, H. M. HECKMAN, M. RODRIGUEZ, ANDR. R. MYERS. 1995. Quantitative neuropathology of a focal nerveinjury causing hyperalgesia. J. Neuropathol. Exp. Neurol. 54:635–643.

20. SOMMER, C., AND M. SCHAFERS. 1995. Experimental painfulmononeuropathy: Cytokines in C57B1/6J mice and in mice withdelayed Wallerian degeneration. Soc. Neurosci. Abstr. 21: 897.

21. TAIWO, Y. O., J. D. LEVINE, R. M. BURCH, J. E. WOO, AND W. C.MOBLEY. 1991. Hyperalgesia induced in the rat by the amino-terminal octapeptide of nerve growth factor. Proc. Natl. Acad.Sci. USA 88: 5144–5148.

22. WAGNER, R., AND R. R. MYERS. 1996. Schwann cells producetumor necrosis factor alpha: Expression in injured and non-injured nerves.Neuroscience 73: 625–629.

23. WELLS, M. R., S. P. RACIS, JR., AND U. VAIDYA. 1992. Changes inplasma cytokines associated with peripheral nerve injury. J.Neuroimmunol. 39: 261–268.

101HYPERALGESIA IN WLD MICE