Opioid and Nonopioid Mechanisms May Contribute to Dynorphn’s Pathophysiological Actions in 3pinal Lord lnjury Alan I. Faden, MD

It has been suggested that the opioid dynorphin, an endogenous agonist for K-opiate receptors, contributes to secondtissue damage after spinal cord injury. To evaluate this hypothesis further, effects of intrathecally administered dynorphin (Dyn) A-(1-17), dynorphin antiserum, or the K-selective opiate antagonist nor-binaltorphimine (nor-BNI) were studied in rats subjected to standardized impact trauma to the thoracic spinal cord. Effects of intrathecal Dyn A(1-17) were also compared to those of Dyn A-(2-17), which is inactive at opiate receptors, in uninjured and injured animals. Both Dyn A-(1-17) and Dyn A-(2-17) produced motor dysfunction in uninjured rats, but Dyn A-(1-17) was approximately 2.5 times more potent. At lower doses of Dyn A-(1-17), paraparesis was markedly attenuated by norBNI; nor-BNI was less effective at higher doses of Dyn A-( 1- 17) and did not modify the motor dysfunction produced by Dyn A-(2-17). Treatment with dynorphin antiserum significantly improved outcome after trauma as compared to control treatment with normal rabbit serum or leucine-enkephalin antiserum. Dyn A-(1-17), but not Dyn A-(2-17) at similar doses, exacerbated neurological dysfunction after spinal cord injury. Pretreatment with nor-BNI attenuated neurological dysfunction after traumatic spinal cord injury to a similar degree in rats administered saline or Dyn A-(117). These observations support the hypothesis that dynorphin contributes to certain pathophysiological changes after traumatic spinal cord injury through both opiate-receptor (K-receptor)-mediated and nonopioid mechanisms. ary

Faden AI. Opioid and nonopioid mechanisms may contribute to dynorphin’s pathophysiological actions in spinal cord injury. Ann Neurol 1990;27:67-74 Tissue damage after spinal cord trauma may result, in part, from delayed or secondary injury caused by endogenous (autodestructive) factors [ 11. Use of opiatereceptor antagonists to treat spinal cord injury (SCI) was based on the hypothesis that endogenous opioids are released following trauma and contribute to secondary injury by reducing blood flow within the microcirculation [21. Subsequent studies demonstrated that opiate-receptor antagonists reverse posttraumatic spinal cord ischemia [2, 31, improve somatosensory evoked responses across the injury zone [ 3 , 41, limit histopathological changes [5], and enhance chronic neurological recovery f2, 5-11) after traumatic or ischemic SCI. In all reports, best effects were obtained with doses of opiate-receptor antagonists substantially higher than those necessary to reverse effects at popiate receptors, suggesting that the beneficial effects occur either through actions at other opiate receptors or through actions other than those mediated by opiate

receptors. In support of an opiate-receptor-mediated mechanism, it has been shown that the beneficial effects of various opiate-receptor antagonists are stereospecific 17, 121, with only the levoisomer showing activity, and that structurally different opiate-receptor antagonists produce similar therapeutic effects [2, 5, 7, 91. A large number of endogenous opioids and fragments have been found since the identification of the enkephalins in 1975 [131. Most of these opioid substances are derived from 3 prohormone precursors: proenkephalin, pro-opiomelanocortin, and prodynorphin (proenkephalin B) [14]. We have suggested that the prodynorphin-derived opioid dynorphin (Dyn) A( 1- 17), a proposed endogenous ligand for the K-opiate receptor 1141, is a secondary injury factor after spinal trauma, with its pathophysiological actions mediated predominantly through K-opiate receptors [9, 151. A number of observations from a variety of experimental

From the Center for Neural Injury, Department of Neurology, University of California at San Francisco and Veterans Administration Medical Center, San Francisco, CA.

Address correspondence to Dr Faden, Neurology Service (127), Veterans Administration Medical Center, 4150 Clement S t , San Francisco, CA 94121.

Received for publication Dec 6, 1988, and in revised form Mar 15 and Jun 7 , 1989. Accepted for publication Jun 7 , 1989.

67

studies appear to support this hypothesis. Dynorphin, but not enkephalins or P-endorphin, accumulates at the injury site following traumatic SCI in rats in concentrations that are significantly correlated with severity of injury 1151. After intrathecal administration, dynorphin causes hindlimb paralysis that clinically simulates posttraumatic paraparesis 116, 171; dynorphin is far more potent than other opioids or dynorphin fragments in producing this effect 1181. K-Opiate receptors have been found within the spinal cord in a variety of species 119-231 and are upregulated after traumatic SCI 1241. Opiate antagonists that have enhanced activity at K-opiate receptors are more effective or more potent, or both, than naloxone in treating SCI; these antagonists include the relatively nonselective WIN44,441-3 171 and nalmefene {51, as well as the highly selective K-receptor antagonist norbinaltorphimine (nor-BNI) 191, whereas the 6-selective opiatereceptor antagonist ICI154,129 is ineffective 181. Other observations appear to be incompatible with the dynorphin-K-receptor hypothesis of SCI. Hall and associates (1987) 1251 have shown that U50488h, a selective K-opiate receptor agonist, improves spinal cord blood flow (SCBF) after traumatic SCI. It has also been found that the dynorphin fragments Dyn A-(217) 1183 and Dyn A-(3-13) 126, 271, which have little or no activity at opiate receptors 1281, also produce hindlimb paralysis after intrathecal administration, whereas U50488h does not 118, 261. Furthermore, some groups have reported that dynorphin-induced hindlimb paralysis in rats is neither reversed nor prevented by even very large doses of naloxone 117, 18, 26, 271. In contrast, others have shown that opiatereceptor antagonists can attenuate or block the paralytic effects of dynorphin 129, 30). It also appears that Dyn A-( 1- 17) and Dyn A-( 1- 13), which are active at opiate receptors, are more potent in producing motor dysfunction than Dyn A42- 17) and Dyn A-(3- 13), respectively Il8, 261. Taken together, these observations suggest that dynorphin-induced paralysis includes both opiate-receptor-mediated and nonopioid components, and that the ability of opiate-receptor antagonists to inhibit motor dysfunction caused by dynorphin may therefore be critically dependent on the dynorphin dose. To test this concept and to assess further the role of dynorphin and K-opiate receptors in SCI, we conducted a series of parallel, independent experiments that (1)compared the potency of Dyn A-(1-17) and Dyn A-(2-17) in causing hindlimb paralysis, as well as the ability of the K-selective opiate antagonist nor-BNI to block this effect; (2) evaluated the ability of antibodies to Dyn A-( 1-17) to limit posttraumatic motor dysfunction; (3) compared the abihty of Dyn A( 1- 17) and Dyn A-(2- 17) to exacerbate posttraumatic paralysis; and (4) determined the effect of nor-BNI, administered intrathecally, in limiting posttraumatic 68 Annals of Neurology Vol 27 N o 1 January 1990

motor dysfunction both with and without superimposed dynorphin administration.

Methods Spinal Cord Znjuy Model After anesthesia with ketamine hydrochloride (15 mg/kg intramuscularly) and sodium pentobarbital (40 mg/kg intraperitoneally), male Sprague-Dawley rats (300 t 25 gm) were subjected to traumatic thoracic SCI using the weightdrop method, as previously described in detail [ 5 } . In brief, after laminectomy to expose the T9 spinal segment, the spinal cord was traumatized using the Allen method, in which a 10-gm weight is dropped a distance 5 cm through a fiberglass guide tube onto a plastic impounder placed on the exposed dura mater. This method produces a consistent degree of incomplete injury, as reflected by both chronic neurological recovery and histopathological examination; control animals demonstrate persistent, moderately severe, spastic paraparesis 4 weeks after trauma [51. The model was developed as an alternative to similar weight-drop models in cats and rabbits and appears to exhibit a similar profile of biochemical or pharmacological responses to that found in the other species. For chronic injury studies, animals are followed and evaluated over a 4-week postinjury period, since we and another investigator [111have found that there is little additional recovery between 4 and 10 weeks after injury.

Dynorphin-induced Paralysis Studies Twenty-four hours before drug administration, a PE-10 catheter was threaded into the intrathecal space to the T8 level through a suboccipital incision, using a method modified from Yaksh and Rudy 131) and previously described in detail [16}. All drugs were administered in an equal volume (10 111) of artificial cerebrospinal fluid: 20 nmol Dyn A-(117) (n = 8), 20 nmol Dyn A-(2-17) (n = 7), 20 nmol Dyn A-(1-17) plus 100 pdkg nor-BNI (n = 7), 30 nmol Dyn A(1-17) (n = lo), 30 nmol Dyn A-(1-17) plus 100 p,g/kg nor-BNI (n = 7), 30 nmol Dyn A-(2-17) (n = ?), 40 nmol Dyn A-(2-17) (n = 5), 50 nmol Dyn A-(2-17) (n = 6), or 50 nmol Dyn A-(2-17) plus 100 pdkg nor-BNI (n = 6). Choice of doses for both dynorphin and nor-BNI was based on earlier dose-response studies with these compounds {9, 181.

Antibody Studies Two independent studies were performed. In study 1, at 24 hours after placement of an intrathecal line at T9, animals received 10 p1 of neat antiserum to Dyn A-(1-17) (Peninsula Laboratories, Belmont, CA) or equal volumes of normal rabbit serum equivalent with regard to protein content (each n = 9).This antiserum shows 42% cross-reactivity with Dyn A-(1-13), 4% with Dyn A-(1-lo), and 0.02% with Dyn A(1-9); no cross-reactivity is seen with Dyn A-(1-8), Dyn A-( 1-7), Dyn A-( 1-6), leucine-enkephalin, methionineenkephalin, or a-neo-endorphin. Fifteen minutes after antisera administration, animals were subjected to traumatic SCI. Four hours after injury, animals were administered a second dose of either Dyn A antiserum or normal rabbit serum (each = 10 pl). Animals were subsequently evaluated for 4 weeks and neurological outcome was determined.

In study 2, animals were randomly assigned to treatment with Dyn A-( 1-17) antiserum, antiserum to leucineenkephalin (Llugh), or normal rabbit serum, all equivalent with regard to protein content (each group n = 10). Animals were treated with 10 pl of neat antiserum at 4 and 24 hours after traumatic SCI (as in study 1)and evaluated for 4 weeks for neurological outcome. The leucine-enkephalin antiserum shows less than 0.06% cross-reactivity with methionineenkephalin and 0.12% to Dyn A-(1-17).

.. a

al

5

b

4

al

3 2

5: z

, ..... ....,

1

0 20 m o l e

Neurological Scoring Animals were blindly scored on a daily basis using an 8-point ordinal scale based on motor function as follows: 0 = no hindlimb movement; 1 = minimum, spontaneous hindlimb movement; 2 = spontaneous hindlimb movement but unable to support weight; 3 = supports weight but unable to walk; 4 = walks with severe spasticity and ataxia; 5 = walks with moderate ataxia and spasticity; 6 = walks with mild ataxia and spasticity; 7 = normal motor function. Rats were also graded as either walkers (score = 4-7) or nonwalkers (score = 0-3). In addition, animals were evaluated on their ability to maintain themselves on an inclined plane for 5 seconds and the maximum angle was noted; this method was originally developed by Rivlin and Tator 1321and was previously described by us in detail 151.

Data Analysis Differences in angle board scores (parametric variable) across groups were analyzed using either paired t tests or analysis of variance (ANOVA) followed by Dunnett’s tests. Neurological ordinal scores (nonparametric variable) were compared using Kruskal-Wallis ANOVA or Mann-Whitney U tests, or both. Neurological scores were also compared in terms of walkers or nonwalkers using the Fisher’s Exact Probability Test, a useful way to distinguish functional categories and previously shown by us to be a good discriminator of treatment effects IS, 7, 91. A p value of less than 0.05 was considered statistically significant.

Results Dy norphin-induced Paralysis Studies At a dose of 20 nmol, Dyn A-(1-17) produced complete paralysis in 7 of 8 animals at 15 minutes after injection, when motor effects are maximal (Fig 1). After a similar dose, Dyn A-(2-17) caused no paralysis in 6 of 7 animals and only minimal weakness in another ( p < 0.01, Mann-Whitney U test) (Fig 2). Pretreatment with nor-BNI at a dose of 100 pglkg significantly attenuated ( p < 0.05, Mann-Whitney U test) paralysis

20 nmole

+ nor-BNI

Dynorphin Plus Trauma Studies In these experiments the intrathecal line was placed at the time of laminectomy just before injury. Dyn A-(1-17) (30 nmol, n = 23), Dyn A-(2-17) (30 nmol, n = 8),or an equal volume of physiological saline (n = 16) was administered 5 minutes prior to traumatic SCI. Animals were subsequently followed for 4 weeks for neurological evaluation. Other animals received Dyn A-(1-17) (30 nmol) plus 100 pgkg norBNI (n = 8) or nor-BNI alone (100 (*g/kg, n = 8) using the same paradigm. Animals were followed for 4 weeks after injury and evaluated for neurological recovery.

100

2

. ... 30 m o l e

a 30 nmole

+ nor-BNI

1

80

a,

Y

4

60

c

C

al

$2

40

a,

a

20

n

20 m o l e

20 nmole

+ nor-BNI

30 m o l e

30 m o l e

+ nor-BNI

Fig 1 . Effects of 20 nmol and 30 nmol dynorphin (Dyn) A-(I17) on neurological function of rats 15 minutes afer intrathecal (IT) administration, a time comsponding t o worst function. At 20 nmol, D y n A-(1-1 7 ) produced complete paralysis in 7 of 8 animals and 30 nmol caused complete paralysis in 9 of 10 animals. The paralytic e&s of D y n A-(1-1 7 ) at 20 nmol, but not at 30 nmol, were signijicantly (p < 0.05) attenuated 6y IT pretreatment with the u-selective opiate-receptorantagonist norbinaltorphimine (nor-BNI, 100 pglkg). Upper panel shows zndiuidual animal neuroscores (dots), as well as the median scores (histograms). Lower panelsimplijies the data in terms of functional (i.e., walking) ability.

induced by Dyn A-( 1-17) at a dose of 20 nmol; only 1 of 7 animals showed complete paralysis and 6 of 7 animals retained walking ability (Fig 1). At a higher dose (30 nmol) Dyn A-(1-17) caused significantly greater motor dysfunction (90% versus 15% in nonwalkers) than Dyn A-(2-17) ( p < 0.01, MannWhitney U test), but nor-BNI was less effective in preventing Dyn A-( l-l7)-induced paralysis, with only 3 of 7 animals retaining walking ability. Comparison of Figures 1 and 2 reveals that Dyn A-(1-17) is approximately 2.5 times more potent than Dyn A-(217) in producing hindlimb paralysis after intrathecal administration, similar to what we have previously reported in other dose-response studies [ 181. Moreover, in contrast to Dyn A-(1-17), paralytic effects of Dyn A-(2-17) at doses equieffective to those of Dyn A-(117) were not affected by treatment with nor-BNI (Fig 2).

Antibody Studies Treatment with antiserum to Dyn A-(l-l7) 15 minutes before and 4 hours after traumatic SCI signifiFaden: Dynorphin’s Actions in Spinal Cord Injury 69

0

Normal Rabbit Serum Dynorphin Antiserum

3 2

: 1

0

20 nmole

30 nmole

40 nmole

50 nmole

50 nrnole

+ nor-BNI 0

1

2

3 Neuroscore

4

5

6

7

'"1 80 60 40 20

20 nmole

30 nmole 40 m o l e

50 nmole

50 m o l e

+ nor-BNI

Fig 2.Comparison of the ej$cts of dynorphin (Dyn)A-(2-17) at 20 nmol, 30 nmol, 40 nmol, and 50 nmol on motor dysfunction at 15 minutes afer intrathecal administration. Comparison of Figures 1 and 2 shows that Dyn A-(1-1 7 ) is approximately 2.5 times more potent than Dyn A-(2-1 7 ) in causing motor weakness. In contrast to its ability to modzfji paralysis caused by Dyn A-(1-1 7), nor-binaltorphimine (nor-BNl) has no e&kt on motor dysfunction caused by Dyn A-(2-1 7 ) at equieffective doses. Data are presented as in Figure I .

cantly improved chronic neurological recovery as compared to that of control animals treated with normal rabbit serum ( p < 0.05, Mann-Whitney U test). Nearly 4 times the number of animals treated with dynorphin antiserum gained walking ability as compared to control subjects (78% versus 22%), although angle differences on the inclined plane test did not differ significantly between the groups (Fig 3). Treatment with Dyn A-(1-17) antiserum at 4 and 24 hours after trauma resulted in a higher percentage of walkers (30%) than in leucine-enkephalin antibody-treated (0%) or normal rabbit serum-treated (10%) animals, although these differences did not reach statistical significance (Fig 4). However, animals treated with Dyn A-( 1- 17) antiserum showed significantly better angle scores than either of the other groups (each p < 0.05 by ANOVA and Dunnett's tests) (Fig 4).

Dynorphin Exacerbation Studies Treatment with dynorphin (30 nmol) before traumatic injury significantly worsened neurological outcome as compared to control animals pretreated with saline (Fig 5) or animals pretreated with equal doses of Dyn A-(2- 17). Four-week motor scores were significantly worse in Dyn A-(l-l7)-treated animals than in control subjects (Kruskal-Wallis ANOVA, p < 0.01;

70 Annals of Neurology

Vol 27 No 1 January 1990

A

Normal Rabbit Serum

Normal Rabbit Serum

Dynorphin Antiserum

Dynorphin Antiserum

Fig 3. Treatment with antiserum t o dynorphin A-(I -1 7) before and 4 hours afer traumatic spinal cord injury signzjkantly (p < 0.05) improved chronic neurological recovery (at 4 week afer trauma) as compared to control animals treated with normal rabbit serum (upper panel). Seventy-eight percent of dynorphin antiserum-treated animals regained walking ability versus 22% of control subjects (middle panel). Altbough a sfight trend was seen toward a smaller angle change from baseline in dynorphinantiserum-treated animals as compared to control animals, these dtfferencesdid not reach significance (lower panel).

Mann-Whitney U test, p < 0.05), as were angle scores (ANOVA, p < 0.05). In contrast, animals treated with Dyn A-(2-17) did not differ significantly from control subjects (data not shown). Pretreatment with nor-BNI significantly improved neurological recovery to a similar degree in animals administered Dyn A-(1-17) or an equal volume of saline ( p < 0.05 and p < 0.01, respectively, Mann-Whitney U test) (Fig 5). Thus, traumatized animals administered either nor-BNI alone or nor-BNI plus Dyn A-( 1- 17) (30 nmol) had a similar recovery rate (Fig 5): In each group 75% of animals regained walking ability versus 12.5% of control subjects. Angle scores also recovered to a similar degree in the 2 nor-BNI-treated groups (Fig 5); each

8o

0

1

0

1

2

3 4 Neuroscore

Normal Rabbit Serum Leu-Enkephalin Antiserum Dynorphin Antiserum

5

6

Saline Dyn 1-17 Dyn 1-17 + nor-BNI

7

2

3

4

5

6

7

Neuroscore

Y ~

5... g

30 20

a al 10 n

Normal Rabbit Serum

Leu-Enkephalin Antiserum

Dynorphin Antiserum

"

30

Saline

Dyn 1-17

Dyn 1-17

+ nor-BNI

nor-BNI

-

(I)

?? m $

20 -

1-

v

m a,

2

Normal Rabbit Serum

Leu-Enkephalin Antiserum

Dynorphin Antiserum

10

-

q-

0

Dyn 1-17

Fig 4. Treatment with antiserum to dynorphin A-(I-1 7) at 4 and 24 hours after spinal cord injury enhances 4-week neurological recovery as compared to control animals treated with normal rabbit serum or antiserum to leucine-enkephalin, Although neumcore dzfferences did not quite reach statistical significance, inclined phne scores were significantly better in animals treated with dynorphin antiserum than in either control group (each p < 0.05).

showed significantly lower angle changes (ANOVA, p < 0.001; Dunnett's test, eachp < 0.01) than the saline control animals or animals pretreated with Dyn A-(117) alone, respectively.

Discussion The present studies are consistent with the hypothesis that dynorphin contributes to secondary SCI after trauma and that such effects are, at least in part, mediated by opiate receptors. First, paralysis caused by Dyn A-( 1-17) administration is significantly attenuated by pretreatment with the opiate-receptor antagonist nor-BNI (which is highly K-selective); in contrast, the paralytic effects of Dyn A-(2-17), which is inactive at opiate receptors and requires substantially higher doses for this action, are unaffected by nor-BNI pretreatment. Second, administration of dynorphin anti-

Dyn 1-17 + nor-BNI

nor-BNI

Fig 5 . Effectsof pretreatment with intrathecally administered saline, dynorphin (Dyn)A-(l-17) (30 nmol), nor-binaltorphimine (nor-BNI, 100 mglkg), or Dyn A-(1-1 7)plus norBNI on motor function and inclinedplane (angle)scores after traumutic spinal cord injury. Dyn A-(I-l7) pretreatment significantly worsened outcome as compared to that of salinetreated control subjects (p < 0.05) on each measure. Nor-binaltorphiminepretreatment improved outcome significantly (p < 0.01)and to the same degree in animals pretreated with saline orDyn A-(1-17).

serum significantly limits the neurological deficits produced by traumatic SCI. Third, administration of exogenous dynorphin exacerbates posttraumatic SCI; pretreatment with a K-selective opiate-receptor antagonist attenuates posttraumatic neurological deficits to a similar degree in animals pretreated with Dyn A-(117) and in saline-treated control subjects. These results may explain much of the controversy in the experimental literature regarding dynorphininduced paralysis. Przewlocki and associates (291 found that dynorphin, at a dose of 4.7 nmol, caused paralysis that was blocked by high doses (10 mg/kg) but not lower doses (1 mg/kg) of naloxone. In contrast, 3 other groups have found that dynorphin-induced

Faden: Dynorphin's Actions in Spinal Cord Injury 7 1

paralysis (with higher dynorphin doses) is not prevented by high-dose naloxone administration 118, 26, 333. Differences in dynorphin dose or potency in the various studies may explain the disparate observations. Consistent with this interpretation, whereas motor dysfunction caused by lower doses of Dyn A-(1-17) (20 nmol) is significantly attenuated by pretreatment with the selective K-receptor antagonist nor-BNI, paralysis produced by higher doses of dynorphin (30 nmol) is only modestly and insignificantly attenuated by nor-BNI pretreatment. From present and previous studies, it becomes evident that dynorphin induces paralysis through both opiate-receptor-mediated and nonopioid mechanisms. Although Dyn A-(2- 17), which is largely inactive at opiate receptors, produces paralysis, it requires 2.5 times the dose of Dyn A-(117) to cause these effects, similar to what has been found by us previously 1183. Moreover, Dyn A-(217)-induced motor dysfunction is not prevented by pretreatment with nor-BNI, whereas paralysis caused by Dyn A-(1-17) is reduced. The fact that the opiatereceptor-mediated paralytic effects occur at lower doses may have substantial relevance, both physiologically and in terms of treatment of SCI. The ability of dynorphin antiserum to limit the effects of traumatic SCI is perhaps the strongest evidence implicating dynorphin in secondary SCI. This effect is specific, since it is not replicated by normal rabbit serum or antiserum to leucine-enkephalin. Although the observation that dynorphin pretreatment exacerbates posttraumatic paraparesis may not be unexpected, it is noteworthy that nor-BNI pretreatment produces similar degrees of recovery in animals pretreated with either Dyn A-(1-17) or an equal volume of vehicle. Thus, a K-selective opiate-receptor antagonist not only improves posttraumatic neurological recovery, presumably by preventing secondary injury caused by endogenous dynorphin release, but also blocks the exacerbation produced by exogenous dynorphin administration. Other data also support a potential pathophysiological role for dynorphin in secondary injury. Dyn A-(117)-immunoreactive material accumulates at the injury zone after impact trauma in direct proportion to the severity of injury 1151. In addition, intrathecal administration of Dyn A-( 1-1 3) produces severe histological damage to the spinal cord 1333. That pathophysiological actions of dynorphin are mediated, in part, through opiate receptors is suggested by several lines of evidence. Both the paralytic effects of Dyn A-(1-17) and the neurological deficit in traumatized animals that is exacerbated by Dyn A-( 117) are attenuated by treatment with nor-BNI; norBNI shows a high degree of selectivity for K-opiate receptors both in vitro and in vivo, the latter at doses similar to those used in the present experiments [343. 72 Annals of Neurology Vol 27 No 1 Januaxy 1990

Similarly, motor dysfunction caused by intrathecal Dyn A-( 1-17) is prevented by pretreatment with MR1452, a purported K-receptor antagonist 1301, whereas the non-opiate-receptor-mediated paralytic effects of Dyn A-(2-17) are unaffected by treatment with nor-BNI. In general, opiate-receptor antagonists that show enhanced activity at K-opiate receptors have been found to be more effective or more potent, or both, than naloxone in improving outcome after traumatic SCI. These antagonists include WIN44,441-3 E71, nalmefene 151, and nor-BNI 193. In contrast, the high doses of naloxone required to improve outcome after traumatic SCI as well as to attenuate paralysis caused by intrathecal dynorphin are inconsistent with a p-receptor mechanism. Furthermore, Spampinato and Candeletti 1303 have reported that dynorphin-induced motor dysfunction occurs in rats made tolerant to morphine. That 6-receptors are not involved is indicated by the failure of the &receptor antagonist ICIl54,129 to alter neurological dysfunction after SCI

181. Apparently inconsistent with the dynorphin-weceptor hypothesis is the observation that U50488h, a K-opioid agonist, improves SCBF after traumatic injury 1253. However, U50488h has a number of physiological actions in addition to its K-agonist activity that may affect SCBF after injury, including an ability to antagonize certain actions of excitatory amino acids and calcium 1253. Also, considerable evidence supports the existence of separate populations of K-receptors (isoreceptors) within the spinal cord [19,21, 301; thus, dynorphin and U50488h may well be acting, in part, through different isoreceptors. In support of this concept, Zukin and associates L353 have recently demonstrated separate populations of high-affinity and lowaffinity K-opiate receptors in rat brain homogenates; whereas U50488h and dynorphin show similar potencies at the high-affinity site, dynorphin is nearly 300 times more active than U50488h at the low-affinity site. The striking differences between U50488h and dynorphin with regard to paralytic actions 1181 and central cardiovascular activity 1361 may indicate that these physiological actions are mediated through the lower-affinity K-site. Possibly consistent with this hypothesis is our recent observation that U50488h and the related K-agonist U69593 worsen outcome in dose-dependent fashion after traumatic fluid-percussion brain injury in rats; however, these effects occur only with relatively high doses (McIntosh and colleagues, unpublished data, 1988). The mechanism(s) by which dynorphin may contribute to secondary tissue injury remains speculative. Dynorphin has previously been reported to reduce SCBF after intrathecal administration through what was postulated to be nonopioid actions, because the effects were not blocked by the nonselective opiate antagonist

naloxone 127); however, the authors did not evaluate opiate-receptor antagonists at lower dynorphin doses, and they did not study K-active or K-selective antagonists. McIntosh and associates [37] showed a significant correlation between accumulation of dynorphin in local tissue and regional declines in cerebral blood flow (CBF) after traumatic brain injury; CBF changes after trauma were reversed by the K-active opiate-receptor antagonist WIN44,441-3 but not by its dextrostereoisomer, indicating an opiate-receptor mechanism. Although opiate antagonists also restore SCBF after traumatic SCI {2, 31, these compounds have other actions that may contribute to their protective action, including ability to attenuate changes in calcium flux 1381 and to limit decreases in intracellular free magnesium after trauma 139). In addition, Caudle and Isaac [40) have recently reported that dynorphin may produce effects at high doses, including paralysis after intrathecal administration, that are partially mediated through excitatory amino acid receptors. Preliminary studies from our laboratory have confirmed the latter observation and demonstrated that paralysis following intrathecal administration of high doses of Dyn A-(117), as well as Dyn A-(2-17), are significantly attenuated by preueatment with noncompetitive Nmethyl-D-aspartate (NMDA) antagonists. NMDA antagonists have also been shown to limit tissue damage and improve outcome after traumatic SCI C41). In summary, the present studies indicate that dynorphin may contribute to SCI through both opiate-receptor-mediated and non-opiate-receptor-mediated actions. The relative role of each of these effects, as well as the specific mechanism(s) by which they are mediated, remain to be established.

This work was supported by a grant from the National Institutes of Health, NINCDS (NS23422-01A1). We adhered to the principles enumerated in the Guidefor the Caw and Use of kboratoty Animals, prepared by the Committee on Care and Use of Laboratory Animals of the Institute of Laboratory Resources, National Research Council (DHEW publication no. [NIH] 85-23, 1985). Dynorphin A-(1-17), dynorphin A-(2-17), and dynorphin antiserum were obtained from Peninsula Laboratories. I thank Drs P. S. Portoghese and A. E. Takemori for providing nor-binaltorphimine, and Drs Brian Cox and Gregory Mueller for the leucine-enkephalin antiserum. I also thank Gretchen Avery, Peter Halt, Brendan Chan, and Rohit Bakshi for technical assistance, and Eleanor Leveau for preparation of this manuscript.

References Faden AI. Role of TRH and opiate receptor antagonists in limiting central nervous system injury. In: Waxman S, ed. Physiological Basis for Functional Recovery in Neurological Disease, vol47. New York: Raven Press, 1988:531-546 Faden AI, Jacobs TP, Holaday JW.Endorphins in experimental spinal injury: therapeutic effects of naloxone. Ann Neurol 1981; 10:326-332

3. Young W, Flamm ES, Demopoulos HB, et al. Naloxone ameliorates postuaumatic ischemia in experimental spinal contusion. J Neurosurg 1981;55:209-219 4. Inoue Y. Evoked spinal potentials in the Wistar rat: effect of cord compression and drugs. Nippon Seikeigeka Gakkai Zasshi 1986;60:777-785 5. Faden AI, Sacksen I, Noble LJ. Opiate-receptor antagonist nalmefene improves neurological recovery after traumatic spinal cord injury in rats through a central mechanism. J Pharmacol Exp Ther 1988;245:742-748 6. Faden AI, Jacobs TP, Holaday JW.Opiate antagonists improve neurologic recovery after spinal injury. Science 1981;211:493494 7. Faden AI, Jacobs TP. Opiate antagonist WIN44,441-3 stereospecifically improved neurological recovery after ischemic spinal injury. Neurology 1985;35:1311- 13 15 8. Faden AI, Jacobs TP, Zivin JA. Comparison of naloxone and a delta-selective antagonist in experimental spinal “stroke.” Life Sci 1983;33(supplI):707-710 9. Faden AI, Takemori AE, Portoghese TS. K-Selective opiate antagonist nor-binaltorphimine improves outcome after traumatic spinal cord injury in rats. CNS Trauma 1987;4:227-237 10. Flamm ES, Young W, Demopoulos HB, et al. Experimental spinal cord injury: treatment with naloxone. Neurosurgery 1982;10227-231 11. Arias MJ. Effect of naloxone on functional recovery after experimental spinal cord injury in the rat. Surg Neurol 1985; 23:440-442 Faden AI. Naloxone acts at central opiate recep12. Holaday JW, tors to reverse hypotension, hypothermia, and hypoventilation in spinal shock. Brain Res 1980;189:295-299 13. Bloom FE. The endorphins. A growing family of pharmacologically pertinent peptides. Annu Rev Pharmacol Toxic01 1982; 23:151-170 14. Cox BM. Endogenous opioid peptides: aguide to structures and terminology. Life Sci 1982;31:1645-1658 15. Faden AI, Molineaux CJ, Rosenberger JC, et al. Endogenous opioid immunoreactivity in rat spinal cord following traumatic injury. Ann Neurol 1985;17:386-390 16. Faden AI, Jacobs TP. Dynorphin induces partially reversible paraplegia in the rat. Eur J Pharmacol 1983;91(2/3):321-324 17. Herman BH, Goldstein A. Antinociception and paralysis induced by intrathecal dynorphin A. J Pharmacol Exp Ther 1985;232:27-32 18. Faden AI, Jacobs TP. Dynorphin-related peptides cause motor dysfunction in the rat through a non-opiate action. Br J Pharmacol 1984;81:2 7 1-2 7 6 19. Arrali B, Gouarderes C, Mazargd H, et al. Evidence for multiple “kappa” binding sites by use of opioid peptides in the guinea-pig lumbo-sacral spinal cord. Neuropeptides 1982;3:5364 20. Traynor JR, Kelley PD, Rance MJ. Multiple opiate binding sites in rat spinal cord. Life Sci 1982;31:1377-1380 21. Gouarderes C, Attali B, Audigier Y, Cros J. Interaction of selective mu and delta llgands with the kappa, subtype of opiate binding sites. Life Sci 1983;33(suppl I):175-178 22. Ctlonkowski A, Costa T, Przewlocki R, et al. Opiate receptor binding sites in human spinal cord. Brain Res 1983;267:392396 23. Mack KJ, KiUian A, Weyhenmeyer JA. Comparison of mu, delta, and kappa opiate binding sites in rat brain and spinal cord. Life Sci 1984;34:281-285 24. b i n s SA, Faden AI. Traumatic injury alters opiate receptor binding in spinal cord. Ann Neurol 1986;19:498-501 25. Hall ED, Wolf DL, Althaus JS, and Von Voigtlander PF. Beneficial effects of the kappa opioid receptor agonist U50488h in acute CNS trauma models. Brain Res 1987;435:174-180

Faden: Dynorphin’s Actions in Spinal Cord Injury 73

26. Stevens CW, Yaksh TL. Dynorphin A and related peptides administered intrathecally in the rat: a search for putative kappa opiate receptor activity. J Pharmacol Exp Ther 1986;238:833838 27. Long JB, Kinney RC, Malcolm DS, et al. Intrathecal dynorphin A (1- 13) and dynorphin A (3-1 3) reduce rat spinal cord blood flow by non-opioid mechanisms. Brain Res 1987;436:374-379 28. Walker JM, Moises HC, Coy DH, et al. Non-opiate effects of dynorphinanddes-tyr-dynorphin.Science 1982;218:1136-1138 29. Przewlocki R, Shearman GT, Herz A. Mixed opioidnon-opioid effects of dynorphin and dynorphin-related peptides after their intrathecal injection in rats. Neuropeptides 1983;3:233-240 30. Spampinaro S, Candeletti S. Characterization of dynorphin Ainduced antinociception at spinal level. Eur J Pharmacol 1985;110:21-30 31. Yaksh TL, Rudy TA. Chronic catheterization of the spinal subarachnoid space. Physiol Behav 1976;17:1031-1036 32. Rivlin AS, Tator CH. Objective clinical assessment of motor function after experimental spinal cord injury in the rat. J Neurosurg 1977;47:5 7 7-58 1 33. Long JB, Petras JM, Mobley WC, Holaday JW. Neurological dysfunction following intrathecal injection of Dyn A-(1-13) in the rat: 11. Non-opioid mechanisms mediate loss of motor function, sensory and autonomic function. J Pharmacol Exp Ther 1988;246:1167-1174 34. Portoghese PS, Lipkowski AW, Takemori AE. Binaltorphimine

74 Annals of Neurology

Vol 2 7

No 1 January 1990

35.

36.

37.

38.

39.

40.

41.

and nor-binaltorphimine, potent and selective K-opioid receptor antagonists. Life Sci 1987;40:1287-1292 Zukin RS, Eghbali M, Olive D, et al. Characterization and visualization of rat and guinea pig brain K-opiate receptors: evidence for K~ and K? opioid receptors. Proc Natl Acad Sci USA 1988;85:4061-4065 Hassen AH, Feuerstein G, Faden AI. Kappa opioid receptors modulate cardiorespiratory function in hindbrain of nuclei of rat. J Neurosci 1984;4:2213-2221 McIntosh TK, Hayes RL, DeWitt DS, et al. Endogenous opioids may mediate secondary damage after experimental brain injury. Am J Physiol 1987;253:E565-E574 Stokes BT, Hollinden G, Fox P. Improvement in injuryinduced hypocalcia by high-dose naloxone intervention. Brain Res 1984;290:187-190 Vink R, McIntosh TK, Faden AI. Treatment with the thyrouopin-releasing hormone analog CG3703 restores magnesium homeostasis following traumatic brain injury in rats. Brain Res 1988;460:184-188 Caudle RM, Isaac L. A novel interaction between dynorphin (1-13) and an N-methyl-Baspanate site. Brain Res 1988;443: 329-332 Faden AI, Lemke M, Simon RP, Noble LJ. N-methyl-B aspartate antagonist MK80 1 improves outcome following traumatic spinal cord injury in rats: behavioral, anatomic and neurochemical studies. J Neurotrauma 1988;5:33-45