Rapid Calpain I Activation and Cytoskeletal Protein Degradation Following Traumatic Spinal Cord Injury: Attenuation with Riluzole Pretreatment

Journal ofNeurochemistryLippincott—Raven Publishers, Philadelphia© 1997 International Society for Neurochemistry

Rapid Calpain I Activation and Cytoskeletal ProteinDegradation Following Traumatic Spinal Cord Injury:

Attenuation with Riluzole Pretreatment

*Joe E. Springer, *Robert D. Azbill, *Sarah E. Kennedy, *Jones George,

and *tJames W. Geddes

*Department of Anatomy and Neurobiology, and tSanders-Brown Research Center on Aging,

Universily of Kentucky Medical Center, Lexington, Kentucky, U.S.A.

Abstract: Immunocytochemical and immunoblottingtechniques were used to investigate calpain I activationand the stability of the calpain-sensitive cytoskeletal pro-teins microtubule-associated protein 2 (MAP2) and spec-trin at 1, 4, and 24 h after contusion injury to the spinalcord. Spinal cord injury resulted in the activation of cal-pain I at all time points examined, with the highest levelof activation occurring at 1 h. At the same early timepoint, therewasa loss of dendritic MAP2 staining in spinalcord sections, accompanied by pronounced perikaryalaccumulation. The loss in MAP2 staining in the injuredspinal cord progressed over the 24-h survival period toaffect regions 3 mm distant to the site of injury. The pres-ence of calpain I-specific spectrin degradation was ap-parent in neuronal cell bodies and fibers as early as 1 hafter injury, with the most intense staining occurringwithin and juxtaposed to the injury site. Spectrin break-down products in neuronal cell bodies declined rapidlyat 4 h and were nearly undetectable at 24 h after injury.Immunoblot studies confirmed the immunocytochemicalresults by demonstrating a significant increase in calpainI activation, a significant decrease in MAP2 levels, and asignificant increase in spectrin breakdown. Finally, treat-ment of animals with riluzole, an inhibitor of glutamaterelease, before surgery reduced significantly the loss ofMAP2 levels observed at 24 h after injury. These resultsdemonstrate that Ca2~-dependentprotease activationand degradation of critical cytoskeletal proteins are earlyevents after spinal cord injury and that treatments thatminimize the actions of glutamate may limit their break-down. Key Words: Microtubule-associated protein 2—Spectrin — Cytoskeleton — Glutamate— Calcium — Cal-pain l—RiluzoIe—CNS injury.J. Neurochem. 69, 1592—1600 (1997).

secondary pathophysiological events result in a hostof biochemical changes, one of which is the excessiverelease of excitotoxic levels of glutamate, and the sus-tained activation of glutamate receptors leading to theaccumulation of high levels of intracellular Ca2~(Choi, 1988; Faden et al., 1989; Panter et al., 1990;Hayes et al., 1992; Young, 1992; Moriya et al., 1994;Mattson et al., 1995). One consequence of increasingintracellular Ca2~levels is the sustained activation ofmany Ca2~-dependent enzymes and proteins, includ-ing members of the calpain family (calpain I and cal-pain II) of neutral proteases (Zimmerman andSchlaepfer, 1984; Suzuki et al., 1987; Saido et al.,1994). The substrates of the calpains include the cy-toskeletal proteins microtubule-associated protein 2(MAP2), spectrin, and neurofilaments (Schlaepfer andZimmerman, 1985; Siman and Noszek, 1988; Johnsonet al., 1991). These cytoskeletal proteins are involvedin maintaining neuronal structural integrity, which isessential for normal cellular function and survival.Therefore, activation of the calpains after traumaticinjury to the CNS may play an important role in de-termining neuronal dysfunction and cell death.

The activation of one of the calpains (calpain I) inthe CNS has been linked directly to the excitotoxicconsequences of excessive glutamate receptor stimula-tion. Intraventricular administration of certain gluta-mate receptor agonists results in the rapid activationof calpain I and calpain-mediated spectrin breakdowninhippocampal neurons that eventually degenerate (Si-

It has been well documented that the considerabledegree of neuronal cell damage and loss that occursafter traumatic injury to the spinal cord is due to anumber of events that are secondary to the primaryinsult (Tator and Fehlings, 1991; Anderson and Hall,1993; Young, 1993; Lynch and Dawson, 1994). These

Received May 7, 1997; revised manuscript received June 3, 1997;accepted June 3, 1997.

Address correspondence and reprint requests to Dr. J. E. Springerat Department of Anatomy and Neurobiology, University of Ken-tucky Medical Center, 800 Rose Street, Lexington, KY 40536-0084,U.S.A.

Abbreviations used: BP, breakdown product; ECL, enhancedchemiluminescence; IR, immunoreactivity; MAP2, microtubule-as-sociated protein 2; RT, room temperature; T, thoracic; TBS, Tris-buffered saline; TTBS, Tris-buffered saline plus Tween 20.

1592

CYTOSKELETON DISRUPTION IN SPINAL CORD INJURY 1593

man and Noszek, 1988; Siman et al., 1989). Moreover,treatment with calpain inhibitors canprovide neuropro-tection from excitotoxic events associated with exces-sive glutamate receptor stimulation (Caner et al., 1993;Kampfl et al., 1995). These studies point to the role ofcalpains in glutamate-associated neuronal cell damageand death and indicate the need for identifying thecritical mechanisms by which calpain activation maycontribute to neuronal vulnerability after traumaticCNS injury.

To investigate the role of calpain I in spinal cordinjury, we examined the integrity and staining patternof the cytoskeletal proteins MAP2 and spectrin aftercontusion injury to rat spinal cord. Given that the intra-cellular accumulation of Ca2~,and, thus, the activationof Ca2~-dependent proteases, are thought to occurearly after traumatic injury to the CNS, the potentialdegradation of these cytoskeletal proteins was analyzedat early time points after injury. In addition, as calpainI undergoes autoproteolysis once activated (Cong etal., 1989; Zimmerman and Schlaepfer, 1991; Saido etal., 1994), we examined the levels of activated calpainI at the same time points. Finally, we examined therole of glutamate release after spinal cord injury as oneof the initial events leading to cytoskeletal degradation.The findings of this study demonstrate that calpain Iactivation and cytoskeletal degradation occur within 1h after traumatic spinal cord injury and that theseevents are mediated, in part, by the release of gluta-mate.

MATERIALS AND METHODS

Spinal cord injuryAll procedures used followed the guidelines established

in the U.S. Public Health Service Policy on Humane Careand Use of Laboratory Animals and the National Institutesof Health Guidefor the Care and Use ofLaboratory Animalsand were approved by the University of Kentucky Institu-tional Animal Care and Use Committee. Spinal cord contu-sion injuries were performed using theNew York Universityimpactor device, which provides an accurate and reproduc-ible method for producing contusion injury to the rat spinalcord (Constantini and Young, 1994; Basso et al., 1996).Animals receiving injury (n = 42) were anesthetized withpentobarbital (40 mg/kg) and a dorsal laminectomy per-formed to expose the spinal cord at thoracic (T) level Tb.Thevertebral column was stabilized by clamping thecolumnat vertebrae T8 and Til, and the impactor probe droppedfrom a distance of 25 mm, inflicting a contusion injury thatresults in hindlimb locomotor deficits (Constantini andYoung, 1994; Basso et al., 1996). Control animals (n = 42)were exposed to the same surgical procedures but only re-ceived a dorsal laminectomy. The body temperatures of allanimals were maintained at 37°C by placing the animals inan Isolette infant incubator (Narco Scientific), and bodytemperatures were monitoredhourly with arectal thermome-ter. The animals were then killed at 1, 4, or 24 h after injuryand their spinal cords processed for immunocytochemicalstaining (n = 8 per time point) or immunoblotting (n = 6per time point) examination of MAP2 and spectrin. Immu-

noblotting techniques were also used to examine calpain Iactivation.

Spinal cord processingFor the immunocytochemical studies, rats (eight control

and eight injured animals per time point) were anesthetizeddeeply with pentobarbital at the different times after injuryand were perfused with 100 ml of 0.1 M phosphate buffer(pH 7.4) followed by 250 ml of 4% paraformaldehyde in0.1 M phosphate buffer. A 10-mm segment of the spinalcord containing the impact site was removed, embedded in10% gelatin, and cryopreserved in 30% sucrose. Each spinalcord segment was mounted onto a microtome chuck withthe dorsal side up, sectioned on a longitudinal plane at athickness of 40 jim, and every section collected.

For the immunoblotting experiments, rats (n = 6 pergroup) were anesthetized with pentobarbital and killed bydecapitation. The spinal cords were removed rapidly and a10-mm segment containing the impact site obtained, imme-diately frozen on dry ice, and stored at —70°C.The spinalcord segments were homogenized in a 1-ml solution of 0.1M phosphate-buffered saline containing 10 jil each of pep-statin (1 mg/ml), leupeptin (2.5 mg/ml), aprotinin (2 mg/ml), phenylmethylsulfonyl fluoride (0.2 M), andEDTA (0.5M, pH 8.0). A 200-jil aliquot of a 1% solution of Triton X-100 was added, the samples homogenized again, and allowedto incubate on ice for 1 h. The samples were centrifuged at14,000 rpm for 30 min in an Eppendorf microcentrifuge, thesupernatant collected, protein levels determined by using thebicinchoninic acid assay (Pierce), and the samples aliquotedand frozen at —20°C.

Antibody characterizationThe primary antibodies used in this study included mono-

clonal antibodies against the 80-kDa subunit of calpain I(gift of G. Johnson) and MAP2 (AP-14, gift of Dr. L. I.Binder), and a polyclonal antibody against a short peptidesequence (CQQQEVY) corresponding to calpain I-mediatedproteolysis of brain spectrin (Ab 38, gift of Dr. R. Siman).Theproduction andcharacterization of each of these antibod-ies have been described previously. The calpain antibodyrecognizes the full-length 80-kDa subunit of calpain I, aswell as the autolytic 76-kDa form that is generated afteractivation (Samis et al., 1987; Guttmann et al., 1997). AP-14 recognizes MAP2a and 2b, but not 2c (Binder et al.,1984, 1986), and Ab 38 is specific for spectrin breakdownproducts (BPs) resulting from calpain I proteolysis (Rob-erts-Lewis and Siman, 1993; Roberts-Lewis et al., 1994). Itdoes not recognize full-length a-spectrin, or spectrin BPsproduced by other proteases.

ImmunocytochemistryAt the time of sectioning, every three adjacent sections

were placed into an individual well of a 48-well plate con-taining a cryoprotectant solution. This procedure was fol-lowed to maintain the dorsal—ventral spatial sequence ofeach spinal cord. One section from each well was used forthe immunocytochemical staining with one of the three anti-bodies. Using the central canal as a landmark, every spinalcord section up to 0.5 mm dorsal and ventral to the centralcanal was stained and analyzed. After removal of endoge-nous peroxidase, sections were blocked in Tris-buffered sa-line (TBS) containing 0.1% Triton X-100 and either 5%normal horse serum for MAP2, or 5% normal goat serumfor the spectrin antibody. Sections were then incubated over-

J. Neurochem., Vol. 69, No. 4, 1997

1594 J. E. SPRINGER ET AL.

night at room temperature (RT) in each antibody diluted inTBS containing 1% normal sera and 0.1% Triton X-100.The dilutions of the primary antibodies were as follows:MAP2, 1:10,000; and spectrin, 1:20,000. Sections wererinsed 5 X 5 min in TBS, and then incubated in biotinylatedhorse anti-mouse IgG (Vector), preadsorbed to rat serumproteins, for MAP2, or biotinylated goat anti-rabbit IgG(Jackson ImmunoResearch) for spectrin at dilutions of 1:200for 1 h in TBS and 1% normal sera. After 5 X 5 min rinsesin TBS, sections were incubated in streptavidin peroxidase(Jackson ImmunoResearch) for 1 h in TB S and 1% normalsera, rinsed 5 X 5 min in TBS, andreacted in TBS containing3,3 ‘-diaminobenzidine andhydrogen peroxide. The sectionswere rinsedthree times in TBS, mounted onto charged slides,dehydrated, cleared, and coverslipped. Negative controlsconsisted of spinal cord sections incubated in the absenceof primary antibody. Therewas acomplete loss of the immu-nocytochemical signal for all of the cytoskeletal proteins inthese control sections (data not shown).

ImmunoblottingSamples of spinal cord protein (MAP2, 40 jig; spectrin,

50 jig; and calpain I, 35 jig) from control and injured spinalcords (n = 6 per group) were diluted in 4x sodium dodecylsulfate—polyacrylamide gel electrophoresis loading buffer,heated to 95°Cfor 3 min. and loaded on Tris-glycine Bio-Rad Ready gels (4—15% gradient gels for MAP2 and spec-trin, and 7.5% for calpain I), and run for 90 min at 100V. The proteins were then transferred at 4°Cto enhancedchemiluminescence (ECL) nitrocellulose membranes (Am-ersham) for 3 h at 100 mA in Tris-glycine/methanol buffer,and the blots rinsed for 15 min in TBS containing 0.1%Tween 20 (TTBS). The blots were then blocked for 1 h inBlotto (5% nonfat dry milk in TTBS) at RT, incubatedovernight at RT in primary antibody (MAP2 at 1:10,000;spectrin at 1:5,000; and calpain I at 1:1,000) made up inBlotto, and rinsed the following day 3x 10 min in TTBS.Blots were then incubated for 1 h at RT in a peroxidase-conjugated goat anti-mouse antibody (Sigma; 1:50,000; ad-sorbed against rat serum) for MAP2 and calpain I, or aperoxidase-conjugated goat anti-rabbit antibody (KPI;1:50,000) for spectrin. After incubation in the secondaryantibody, the blots were rinsed 4X 10 mm, incubated for 3min in theECL detection solution (Amersham), andexposedto Hyperfilm-ECL (Amersham) for 5—20 min. Control blotswere incubated in the same fashion except the primary anti-body was omitted. There was a complete loss of specificsignal for all of the cytoskeletal proteins in these controlblots (data not shown).

Riluzole treatmentThe anticonvulsant agent, riluzole, is a potent inhibitor of

glutamate release and has been shown to be neuroprotectivein models of glutamate-mediated excitotoxicity, andrecentlywas shown to promote recovery of function after spinal cordinjury (Doble, 1996; Stutzmann et al., 1996). In the presentstudy, animals were treated with riluzole to test whetherglutamate release after spinal cord injury is involved in thedownstream events leading to cytoskeletal protein degrada-tion. Animals (n = 6) received an intraperitoneal injectionof either vehicle (saline) or 8 mg/mg riluzole (RBI) at 15min before, and 2 h after, spinal cord injury. This dose ofriluzole has been shown to be neuroprotective in several invivo models of excitotoxicity (Pratt et al., 1992; Wahl et al.,1993; Stutzmann et al., 1996). Animals were then killed at

24 h after surgery and the levels of MAP2 were quantifiedusing immunoblotting techniques as described above.

Data analysisChanges in the immunocytochemical staining pattern for

the cytoskeletal proteins over the three survival periodswere examined using a Zeiss Axioplan microscope. Themost rostral and caudal extent to which these changes oc-curred was measured (in mm) relative to the center of theimpact site. Quantitation of changes in calpain I activationand cytoskeletalprotein levels in the immunoblots was per-formed by using densitometric analysis. A total of six im-munoblots for each protein of interest were analyzed, witheach blot containing protein samples from six animals(three controls and three injured from each of the threepostsurgical time points). Each blot was scanned, theimagecaptured to a Macintosh PowerPC computer, and the rela-tive density of each individual band determined by usingNIH IMAGE 1.60 software.

The data obtained from the immunoblot experiments werecombined to obtain agroup mean for each time point, whichwasthen expressed as mean relative density units (±SEM).The analysis of the calpain I bands was expressed as therelative ratio of the 80- to 76-kDa isoforms, which is indica-tive of calpain I autolysis after activation (Zimmerman andSchlaepfer, 1991; Saido et al., 1994). As there wasno differ-ence between the three control groups, the data from thisgroup were combined andanalyzed as a single group. Statis-tical analysis of the immunoblot data was performed usinga one-way ANOVA for overall significance, followed byStudent‘s t test for group differences when justified.

FIG. 1. Immunoblot analysis and quantification of calpain I acti-vation in soluble protein fractions from control and spinal cord-injured animals. Immunoblots were quantified using densitomet-nc analysis from six independent experiments (see Materials andMethods) and expressed as the ratio of the activated calpain Iform (76 kDa) relative to the precursor form (80 kDa). The76180-kDa ratio was unchanged in control spinal cord samplesover the three time points, and the data from these groups werecombined. A significant increase in the 76/80-kDa ratio wasfound to occur in samples from injured spinal cords relative tocontrols at all time points examined (°p< 0.01; **p < 0.001).



RESULTS

Spinal cord injuryThe spinal cord impact model used results in a cen-

tral core cavitation lesion that spares the most superfi-cial layers of the surrounding white matter tracts (Con-stantini and Young, 1994; Basso et al., 1996). Allanimals receiving spinal cord injury exhibited hind-limb paralysis and a loss of hindpaw reflex behaviorthroughout the entire survival period. There was norecovery of hindlimb movement in any of the injuredanimals over the 24-h postinjury survival period. Allof the laminectomy control animals had normal loco-motion behavior and use of hindlimbs at all postsurgi-cal time points.

Calpain I activationCalpain I undergoes autolysis after activation that

results in the cleavage of the 80-kDa subunit to a form

that migrates at 76 kDa. Therefore, the ratio of the 76-kDa form relative to the 80-kDa form serves as anindicator of calpain I activation (Zimmerman andSchlaepfer, 1991; Saido et al., 1994). Both forms ofcalpain I were present in cytosolic fractions from con-trol spinal cords at all time points analyzed, with the80-kDa subunit being more prominent than the 76-kDaform (Fig. 1). This pattern of the 76/80-kDa ratio wasno different from that observed in unoperated animals(data not shown). The 76-kDa form was found toincrease dramatically by 1 h after injury to levels thatwere comparable with that of the 80-kDa form (Fig.1). Statistical analysis revealed that the 76/80-kDaratio increased significantly (4.8-fold; p < 0.001) atthis early time point in the injured spinal cord samplesrelative to controls. The 76/80-kDa ratio was still sig-nificantly elevated by 2.8-fold (p <0.001) at 4 h andby 1.4-fold (p < 0.01) at 24 h in the injured samplescompared with controls (Fig. 1).

FIG. 2. Photomicrographs of longitudinal control and injured spinal cord sections stained for MAP2. Control spinal cord sections (A)showed intense MAP2 R in dendrites and weak perikaryal staining. At 1 h after injury (B), MAP2 R decreased in dendrites aroundthe impact site and increased in cell bodies. By 4 h after injury, MAP2 IA had decreased dramatically in dendrites and cell bodies nearthe impact site (C). At 24 h after injury, MAP2 R was nearly absent in the impact site and extended 3—4 mm rostral and 1—2 mmcaudal to the impact site (D). There were no changes in MAP2 IA in white matter at any of the postinjury time points examined.Magnification bar equals 100 tim.

I Neurochem,, Vol. 69, No. 4, 1997


FIG. 3. Immunoblot analysis and quantification of MAP2 proteinlevels in soluble protein fractions from control and spinal cord-injured animals. MAP2 levels are expressed as mean relativedensity units (±SEM) from each time point after injury. Thelevels of MAP2 were unchanged in control spinal cord samplesover the three time points, and the data from those groups werecombined. A significant decrease in MAP2 levels was observedin samples from injured spinal cords relative to controls at alltime points examined (°p< 0.01; °°p< 0.001).

MAP2 immunocytochemistry and levelsSpinal cord neurons exhibited intense dendritic

MAP2 immunoreactivity (IR) with weak perikaryalstaining in all of the laminectomy control animals (Fig.2A). By 1 h after spinal cord injury, there was a loss ofdentritic MAP2 staining accompanied by the neuronalperikaryal accumulation rostral, and in close proximityto, the impact site (Fig. 2B). By 4 h after injury,perikaryal MAP2 JR was virtually absent, and the lossof dendritic staining extended 1—2 mm rostral to theimpact site (Fig. 2C). At the longest time point exam-ined (24 h), MAP2 IR was found to be absent indendrites and cell bodies up to 3 mm rostral to theimpact site (Fig. 2D).

The antibody used for the MAP2 immunocytochemi-cal studies was also used for quantification of MAP2levels. Immunoblots of cytosolic spinal cord fractions forMAP2 revealed that this cytoskeletal protein migrated toa position that was ‘=270 kDa in laminectomy controlsat all time points analyzed (Fig. 3). The size of MAP2in these control animals is consistent with that describedfor intact MAP2 (Johnson et al., 1991; Matesic andLin, 1994). The levels of MAP2 were unchanged in thecontrol animals over the three time points. In the injuredspinal cord samples, MAP2 levels were found to de-crease, relative to controls, by 48% (p < 0.01) at I hafter injury, by 81% (p < 0.001) at 4 h, and by 88%(p < 0.001) at 24 h after injury (Fig. 3).

Spectrin BPsThe spectrin antibody (Ab 38) stains for spectrin

BPs that are preferentially cleaved by calpain I (Rob-

erts-Lewis and Siman, 1993; Roberts-Lewis et al.,1994). Spectrin-BP IR was weak in laminectomy con-trol animals at all time points analyzed and was charac-terized by diffuse staining of cell bodies (Fig. 4A).At 1 h after injury, spectrin-BP JR was confined to thegray matter and was clearly evident in large cell bodies(25—30 jim in diameter) and fibers immediately rostraland caudal to the impact site (Fig. 4B). At 4 h afterinjury, spectrin-BP JR was still present in a few cellbodies, although these cells were smaller in size (<15jim in diameter) than those observed at 1 h after injury(Fig. 4C). By 24 h after injury, spectrin-BP JR wasabsent in cell bodies, but the gray matter continuedto exhibit diffuse positive staining in close proximity(<0.5 mm) to the impact site (Fig. 4D).

The antibody that recognizes spectrin BPs was usedto examine this cytoskeletal protein in immunoblotsafter spinal cord injury. The presence of calpain Icleaved spectrin in cytosolic spinal cord fractions wasvery low in all of the laminectomy control animals(Fig. 5) and was similar to levels obtained from unop-crated control animals (data not shown). At 1 h afterinjury, spectrin-BP levels were found to be elevatedsignificantly by 3.6-fold (p < 0.001) in the injuredspinal cord samples compared with controls. In addi-tion, spectrin-BP levels were still elevated significantlyat 4 h (3.2-fold; p < 0.001) and 24 h (1.7-fold; p<0.01) after injury (Fig. 5).

Riluzole treatmentAs demonstrated above, the levels of MAP2 protein

in spinal cord-injured animals receiving vehicle (n= 6) were decreased significantly (p < 0.00 1) by 84%at 24 h compared with control (n = 6) levels (Fig.6). However, in spinal cord-injured animals treatedwith riluzole (n = 6), MAP2 levels were decreasedby only 42% (p <0.01), compared with control levels,and were significantly higher (p < 0.005) than MAP2levels from injured animals treated with vehicle. Treat-ment of control animals with riluzole had no effect onMAP2 levels (data not shown). In addition, riluzoletreatment had no effect on body temperature in eithercontrol or injured animals (data not shown).

DISCUSSION

The results of this study provide evidence that trau-matic spinal cord injury results in the increased activa-tion of calpain I and the degradation of two calpain I-sensitive cytoskeletal proteins within 1 h after injury.Cytoskeletal degradation was first encountered in andaround the impact site at the earliest time point exam-ined (1 h) and progressed over time to affect areas 2—3 mm distant to the site of injury. As MAP2 and spec-trin are important for maintaining the structural integ-rity of developing and adult neurons, degradation ofthese and other major cytoskeletal elements may con-tribute to neuronal dysfunction after CNS injury(Banik et al., 1982, 1987; Siman and Noszek, 1988;

.1. Neurochem., Vol. 69, No. 4, 1997


FIG. 4. Photomicrographs of longitudinal control and injured spinal cord sections stained for calpain I-mediated spectrin BPs. Controlspinal cords exhibited weak cell body staining for spectrin BPs at all time points analyzed (A). At 1 h after injury, there was diffusestaining for spectrin BPs in the gray matter near the injury site (B). Most noteworthy at this time point was the intense staining forspectrin BPs in large neuronal cell bodies and numerous fibers that were in close proximity to the impact site (B). By 4 h after injury(C), the gray matter was still diffusely stained for spectrin BPs and a few immunoreactive neurons were present. By 24 h after injury,no cell body staining for spectrin BPs was observed, although the gray matter was still immunoreactive (D). There were no changesin spectrin BPs in white matter at any of the postinjury time points examined. Magnification bar equals 100 tim.

Taft et al., 1992; Roberts-Lewis and Siman, 1993;Matesic and Lin, 1994; Roberts-Lewis et al., 1994;Kampfl et al., 1996a,b; Posmantur et al., 1996). Sev-eral lines of evidence support the hypothesis that cy-toskeletal degradation is extremely sensitive to gluta-mate- and Ca2~-mediated excitotoxic events. MAP2,spectrin, and neurofilament proteins have been shownto undergo proteolysis after CNS insults involving ex-citotoxic mechanisms leading to calpain I activation(Banik et al., 1987; Siman and Noszek, 1988; Johnsonet al., 1989; Siman et al., 1989; Roberts-Lewis andSiman, 1993; Matesic and Lin, 1994; Roberts-Lewiset al., 1994; Posmanturet al., 1996). In addition, spinalcord injury results in the degradation of neurofilamentproteins, which is mediated, in part, by activation ofCa2~-dependent proteases (Banik et al., 1982; Iwasakiet aI., 1985, 1987).

This study is the first to demonstrate that spinal cordinjury results in the rapid degradation of MAP2 and

spectrin. MAP2 may be especially vulnerable, as it isfound in the somatodendritic compartments of neurons,regions in which high levels of excitatory amino acidreceptors and calpain J are found (Siman et al., 1985;Cotman et al., 1987; Perlmutter et al., 1988). Theinvolvement of excessive glutamate and receptor acti-vation in MAP2 degradation after spinal cord injury issupported by our finding that the loss of MAP2 ispartially attenuated in animals treated with riluzole, aninhibitor of glutamate release. Treatment with excit-atory amino acid receptor antagonists has also beenshown to attenuate the degradation of MAP2 after lat-eral fluid percussion injury (Hicks et al., 1995). Exces-sive glutamate receptor activation has also been hy-pothesized to contribute to spectrin breakdown. Gluta-mate receptor agonists, as well as ischemic andtraumatic brain injury, result in the activation of cal-pain I and the breakdown of spectrin into calpain J-cleaved fragments (Siman and Noszek, 1988; Siman



et al., 1989; Roberts-Lewis et al., 1994; Kampfl et al.,1 996a,b). Finally, several studies have demonstratedthat treatments with protease inhibitors that block cal-pain activity will inhibit MAP2 and spectrmn degrada-tion and will promote neuron survival after excitatoryor traumatic CNS injury (Arai et al., 1990; Arlinghauset al., 1991; Lee et al., 1991; Caner et al., 1993; Ramiand Kriegistein, 1993; Bartus et al., 1994). Given thesefindings, it is possible that the beneficial effects seenwith glutamate receptor antagonist treatment in trau-matic brain and spinal cord injury may be related tothe inhibition of calpain J activation and subsequentcytoskeletal degradation (Faden et al., 1988; Gomez-Pinella et al., 1989; Panter and Faden, 1992; Gentileand McJntosh, 1993; Wrathall et al., 1994).

Jn summary, the findings of this study demonstratethat calpain J is activated soon after traumatic spinalcord injury, and it occurs at times when MAP2 andspectrin degradation is evident. It should be noted,however, that calpain J autolysis is not a necessarycondition for expression of proteolytic activity (Gutt-mann et al., 1997). Therefore, it is possible that calpainJ proteolytic activity after spinal cord injury may begreater than that indicated by the presence of the auto-lyzed 76-kDa subunit. Jn addition, it cannot be ruledout that other isoforms of calpain are not involved, asa recent report has demonstrated that spinal cord injuryresults in the activation and content of the calpain IIisoform (Banik et al., 1997). Finally, it is possible that

FIG. 5. Immunoblot analysis and quantification of calpain I-cleaved spectrin in soluble protein fractions from control andspinal cord-injured animals. The levels of spectrin breakdownare expressed as mean relative density units (±SEM) from eachtime point after injury. Spectrin breakdown was unchanged incontrol spinal cord samples over the three time points, and thedata from these groups were combined. Spinal cord injury re-sulted in a significant increase in spectrin breakdown in samplesfrom injured spinal cords relative to controls at all time pointsexamined (*p < 0.01; ~*p < 0.001).

FIG. 6. lmmunoblot analysis and quantification of MAP2 levelsin soluble protein fractions from control (lanes 1 and 2) andspinal cord-injured animals treated with saline vehicle (lanes 3and 4) or riluzole (8 mg/kg; lanes 5 and 6), a potent inhibitor ofglutamate release. As shown previously, spinal cord injury re-sulted in a significant loss of MAP2 levels in saline-treated ani-mals at 24 h after injury. However, the loss of MAP2 levels wassignificantly attenuated in spinal cord-injured animals treatedwith riluzole (*p <0.001, relative to controls; **p < 0.01, relativeto controls, and p < 0.005, relative to spinal cord-injured animalstreated with vehicle).

the activation ofcalpain I and the degradation of MAP2and spectrin may simply be a consequence of the struc-tural disintegration of neurons, and not a causal factor.However, the rapid increase in intracellular Ca2~andthe rapid activation of calpain J have been shown tooccur before any gross morphological changes indica-tive of degeneration (Siman et al., 1989; Kampfl et al.,1996a,b). In addition, treatment withcalpain inhibitorscan provide neuroprotection from excitotoxic celldeath (Caner et al., 1993; Kampfl et al., 1996a,b; Pos-mantur et al., 1997). Regardless, given the findings ofthis and other studies, therapeutic strategies aimed atblocking the actions of glutamate or inhibiting the acti-vation of calpain J may limit injury-induced cytoskele-tal degradation and may prove beneficial in providingneuroprotection and functional recovery after trau-matic spinal cord injury.

Acknowledgment: We thank the following for the gift ofthe antibodies used in this study: Dr. G. Johnson (calpainI), Dr. L. Binder (AP-l4), and Dr. R. Siman (Ab 38),and also Ms. Natalie Soultanian for her excellent technicalassistance. This study was supported by NIH grants NS-30248 (J.E.S.) andAG-08974 (J.W.G.) andKentucky SpinalCord and Head Injury Research Trust grants SA-9502-K3(J.E.S.) and GA-9601-K (J.W.G.).

REFERENCESAnderson D. and Hall E. (1993) Pathophysiology of spinal cord

trauma. Ann. Emerg. Med. 22, 987—992.



Arai A., Kessler M., Lee K., and Lynch G. (1990) Calpain inhibitorsimprove the recovery of synaptic transmission from hypoxia inhippocampal slices. Brain Res. 532, 63—68.

Arlinghaus L., Mehdi S., and Lee K. S. (1991) Improved posthyp-oxic recovery with a membrane-permeable calpain inhibitor.Eur. J. Pharmacol. 209, 123—125.

Banik N., Hogan E., Powers J., and Whetstine L. (1982) Degradationofcytoskeletal proteins in experimental spinal cord injury. Neu-rochem. Res. 7, 1465—1475.

Banik N., Happel R., Sostek M., Chiu F., and Hogan E. (1987)Ca2+-mediated degradation of central nervous system (CNS)proteins: topographic and species variation. Metab. Brain Dis.2, 117—126.

Banik N., Matzelle D., Gantt-Wilford G., Osborne A., and HoganE. (1997) Increased calpain content and progressive degrada-tion of neurofilament protein in spinal cord injury. Brain Res.752, 301—306.

Bartus R. T., Baker K. L., Heiser A. D., Sawyer S. D., Dean R. L.,Elliott P. J., and Straub J. A. (1994) Postischemic administra-tion of AK275, a calpain inhibitor, provides substantial protec-tion against focal ischemic brain damage. J. Cereb. Blood FlowMetab. 14, 537—544.

Basso D., Beattie M., and Breshnahan J. (1996) Graded histologicaland locomotor outcomes after spinal cord contusion using theNYU weight-drop device versus transection. Exp. Neurol. 139,244—256.

Binder L. I., Frankfurter A., Kim H., Caceres A., Payne M. R., andRebhun L. I. (1984) Heterogeneity of microtubule-associatedprotein 2 during rat brain development. Proc. Natl. Acad. Sci.USA 81, 5613—5617.

Binder L. I., Frankfurter A., and Rebhun L. I. (1986) Differentiallocalization of MAP2 and tau in mammalian neurons in situ.Ann. NYAcad. Sci. 466, 145—166.

Caner H., Collins J. L., Harris S. M., Kassell N. F., and Lee K. S.(1993) Attenuation of AMPA-induced neurotoxicity by a cal-pain inhibitor. Brain Res. 607, 354—356.

Choi D. (1988) Glutamate neurotoxicity and diseases ofthe nervoussystem. Neuron 1, 623—634.

Cong J., Goll D., Peterson A., and Kapprell H. (1989) The role ofautolysis in activity of the Ca2+-dependent proteinases (ii-

calpain and m-calpain). J. Biol. Chem. 264, 10096—10103.Constantini S. and Young W. (1994) The effects of methyipredniso-

lone and the ganglioside GM1 on acute spinal cord injury inrats. J. Neurosurg. 80, 97—111.

Cotman C., Monaghan D., Ottersen O., and Storm-Mathisen J.(1987) Anatomical organizationof excitatory amino acid recep-tors and their pathways. Trends Neurosci. 10, 273—280.

Dohle A. (1996) The pharmacology and mechanism of action ofriluzole. Neurology 47, S233—S241.

Faden A., Lemke M., Simon R., and Noble L. (1988) N-Methyl-D-aspartate antagonist MK-80l improves outcome followingtraumatic spinal cord injury in rats: behavioral, anatomical, andneurochemical studies. J. Neurotrauma 5, 33—45.

Faden A. I., Demediuk P., Panter S. S., and Vink R. (1989) Therole of excitatory amino acids and NMDA receptors in traumaticbrain injury. Science 244, 798—800.

Gentile N. and McIntosh T. (1993) Antagonists of excitatory aminoacids and endogenous opioid peptides in the treatment of experi-mental central nervous system injury. Ann. Emerg. Med. 22,1028—1034.

Gomez-Pinella F., Tram H., Cotman C., and Nieto-Sampedro M.(1989) Neuroprotective effect of MK-80l and U-50488H aftercontusive spinal cord injury. Exp. Neurol. 104, 118—124.

Guttmann R., Elce J., Bell P., Isbell J., and Johnson G. (1997)Oxidation inhibits substrate proteolysis by calpain I but notautolysis. J. Biol. Chem. 272, 2005—2012.

Hayes R., Jenkin L., and Lyeth B. (1992) Neurotransmitter mediatedmechanisms of traumatic brain injury: acetylcholine and excit-atory amino acids. J. Neurotrauma 9, 173—187.

Hicks R. R., Smith D. H., and Mcintosh T. K. (1995) Temporalresponse and effects of excitatory amino acid antagonism on

microtubule-associated protein 2 immunoreactivity followingexperimental brain injury in rats. Brain Res. 678, 151—160.

Iwasaki Y., lizuka H., Yamamoto T., Konno H., and Kadoya S.(1985) Alleviation of axonal damage in acute spinal cord injuryby a protease inhibitor: automated morphometric analysis ofdrug effects. Brain Res. 347, 124—126.

Iwasaki Y., Yamamoto H., lizuka H., Yamamoto T., and Konno H.(1987) Suppression of neurofilament degradation by proteaseinhibitors in experimental spinal cord injury. Brain Res. 406,99—104.

Johnson G., Jope R., and Binder L. (1989) Proteolysis of tau bycalpain. Biochem. Biophys. Res. Commun. 163, 1505—1511.

Johnson G. V. W., Litersky J. M., and Jope R. 5. (1991) Degrada-tion of microtubule-associated protein 2 and brain spectrin bycalpain: a comparative study. J. Neurochem. 56, 1630—1638.

Kampfl A., Whitson J. S., Zhao X., Posmantur R., Clifton G. L., andHayes R. L. (1995) Calpain inhibitors reduce depolarizationinduced loss of tau protein in primary septohippocampal cul-tures. Neurosci. Lett. 194, 149—152.

Kampfl A., Posmantur R., Nixon R., Grynspan F., Zhao X., LiuS., Newcomb J., Clifton G., and Hayes R. (1996a) ~s-Calpainactivation and calpain-mediated cytoskeletal proteolysis follow-ing traumatic brain injury. J. Neurochem. 67, 1575—1583.

Kampfl A., Zhao X., Whitson J. S., Posmantur R., Dixon C. E., YangK., Clifton G. L., and Hayes R. L. (1996b) Calpain inhibitorsprotect against depolarization-induced neurofilament proteinloss of septo-hippocampal neurons in culture. Eur. J. Neurosci.8, 344—352.

Lee K. S., Frank S., Vanderklish P., Arai A., and Lynch G. (1991)Inhibition of proteolysis protects hippocampal neurons fromischemia. Proc. Natl. Acad. Sci. USA 88, 7233—7237.

Lynch D. R. and Dawson T. M. (1994) Secondary mechanisms inneuronal trauma. Curr. Opin. Neurol. 7, 510—516.

Matesic D. F. and Lin R. C. S. (1994) Microtubule-associated pro-tein 2 as an early indicator of ischemia-induced neurodegenera-tion in the gerbil forebrain. J. Neurochem. 63, 1012—1020.

Mattson M., Barger S., Begley J., and Mark R. (1995) Calcium,free radicals, and excitotoxic neuronal death in primary cellculture. Methods Cell Biol. 46, 187—2 16.

Moriya T., Hassan A. Z., Young W., and Chesler M. (1994) Dynam-ics of extracellular calcium activity following contusion of therat spinal cord. J. Neurotrauma 11, 255—263.

Panter S. S. and Faden A. I. (1992) Pretreatment with NMDA antag-onists limits release of excitatory amino acids following trau-matic brain injury. Neurosci. Lett. 136, 165—168.

Panier S. S., Yum S. W., and Faden A. I. (1990) Alteration in extra-cellular amino acids after traumatic spinal cord injury. Ann.Neurol. 27, 96—99.

Perlmutter L. S., Siman R., Gall C., Seubert P., Baudry M., andLynch G. (1988) The ultrastructural localization of calcium-activated protease “calpain“ in rat brain. Synapse 2, 79—88.

Posmantur R., Kampfl A., and Hayes R. (1997) A calpain inhibitorattenuates cortical cytoskeletal protein loss after experimentaltraumatic brain injury in the rat. Neuroscience 77, 875—888.

Posmantur R. M., Kampfl A., Liu S. J., Heck K., Taft W. C., CliftonG. L., and Hayes R. L. (1996) Cytoskeletal derangements ofcortical neuronal processes three hours after traumatic braininjury in rats: an immunofluorescence study. J. Neuropathol.Exp. Neurol. 55, 68—80.

Pratt J., Rataud J., Bardot F., Roux M., Blanchard J. C., LaduronP. M., and Stutzmann J. M. (1992) Neuroprotective actions ofriluzole in rodent models ofglobal and focal cerebral ischaemia.Neurosci. Lett. 140, 225—230.

Rami A. and Krieglstein J. (1993) Protective effects of calpaininhibitors against neuronal damage caused by cytotoxic hypoxiain vitro and ischemia in vivo. Brain Res. 609, 67—70.

Roberts-Lewis J. M. and Siman R. (1993) Spectrin proteolysis inthe hippocampus: a biochemical marker for neuronal injury andneuroprotection. Ann. NY Acad. Sci. 679, 78—86.

Roberts-Lewis J. M., Savage M. J., Marcy V. R., Pinsker L. R., andSiman R. (1994) Immunolocalization of calpain I-mediated



spectrin degradation to vulnerable neurons in the isehemic ger-bil brain. J. Neurosci. 14, 3934—3944.

Saido T., Sorimachi H., and Suzuki K. (1994) Calpain: new perspec-tives in molecular diversity and physiological-pathologicalinvolvement. FASEB J. 8, 814—822.

Samis J., Zboril G., and Elce J. (1987) Calpain I remains intactand intracellular during platelet activation. Immunochemicalmeasurements with monoclonal and polyclonal antibodies. Bio-chem. J. 246, 481—488.

Schlaepfer W. and Zimmerman U. (1985) Calcium-activated prote-olysis in intermediate filaments. Ann. NY Acad. Sci. 455, 552—562.

Siman R. and Noszek J. (1988) Excitatory amino acids activatecalpain I and induce structural protein breakdown in vivo. Neu-ron 1, 279—287.

Siman R., Gall C., Perlmutter L. S., Christian C., Baudry M., andLynch G. (1985) Distribution ofcalpain I, an enzyme associatedwith degenerative activity, in rat brain. Brain Res. 347, 399—403.

Siman R., Noszek J., and Kegerise C. (1989) Calpain I activationis specifically related to excitatory amino acid induction ofhippocampal damage. J. Neurosci. 9, 1579—1590.

Stutzmann J. M., Pratt J., Boraud T., and Gross C. (1996) Theeffect of riluzole on post-traumatic spinal cord injury in the rat.Neuroreport 7, 387—392.

Suzuki K., Imajoh S., Emori Y., Kawasaki H., Minami Y., and Ohno

5. (1987) Calcium-activated neutral protease and its endoge-nous inhibitor: activation at the cell membrane and biologicalfunction. FEBS Lett. 220, 271—277.

Taft W. C., Yang K., Dixon C. E., and HayesR. L. (1992) Microtu-bule-associated protein 2 levels decrease in hippocampus fol-lowing traumatic brain injury. J. Neurotrauma 9, 281—290.

Tator C. and Fehlings M. (1991) Review of the secondary injurytheory of acute spinal cord trauma with emphasis on vascularmechanisms. J. Neurosurg. 75, 15—26.

Wahl F., Allix M., Plotkine M., and Boulu R. G. (1993) Effect ofriluzole on focal cerebral isehemia in rats. Fur. J. Pharmacol.230, 209—214.

Wrathall J. R., Choiniere D., and Teng Y. D. (1994) Dose-dependentreduction of tissue loss and functional impairment after spinalcord trauma with the AMPA/kainate antagonist NBQX. J Neu-

rosci. 14, 6598—6607.Young W. (1992) Role of calcium in central nervous system injuries.

J. Neurotrauma 9 (Suppl. 1), S9—S25.Young W. (1993) Secondary injury mechanisms in acute spinal cord

injury. .1 Emerg. Med. 11 (Suppl. 1), 13—22.Zimmerman U. and Schlaepfer W. (1984) Calcium-activated neutral

protease (CANP) in brain and other tissues. Prog. Neurobiol.23, 63—78.

Zimmerman U. and Schlaepfer W. (1991) Two stage autolysis ofthe catalytic subunit initiates activation of g-calpain. Biochim.Biophys. Acta 1078, 192—198.


s

s

a

a

a

a

Documents

Rapid Calpain I Activation and Cytoskeletal Protein Degradation Following Traumatic Spinal Cord Injury: Attenuation with Riluzole Pretreatment