Detrimental effects of exogenous glutamate on spinal cord neurons during brief ischemia in vivo

Combined use of an epidural cooling catheter and systemic moderate hypothermia enhances spinal cord protection against ischemic injury in rabbits

BACKGROUND

Epidural placement of a cooling catheter can protect against ischemic spinal cord injury. With the use of rabbits, we investigated whether this epidural cooling technique, when combined with systemic moderate hypothermia, can protect the spinal cord against ischemic metabolic stress.

METHODS

New Zealand white rabbits (n = 28) were assigned to 1 of 4 different groups. Animals underwent abdominal aortic occlusion for 30 minutes using a 3F balloon catheter. Group 1 (n = 7) underwent epidural cooling by the catheter and systemic moderate hypothermia (35 °C) induced with a cooling blanket. Group 2 (n = 7) underwent epidural cooling under systemic normothermia (38.5 °C). Group 3 (n = 7) underwent systemic moderate hypothermia (35 °C) without epidural cooling. Group 4 (n = 7) underwent neither epidural nor blanket cooling as a negative control. Neurologic status of their hind limbs was graded according to the modified Tarlov scale at 1, 2, and 7 days after surgery.

RESULTS

During infrarenal aortic ischemia, epidural temperature was significantly lower in group 1 (18.5 °C ± 0.8 °C) than in group 2 (28.6 °C ± 1.0 °C; P = .0001), group 3 (34.2 °C ± 0.06 °C; P = .0001), or group 4 (38.5 °C ± 0.2 °C; P = .0001). Hind limb function recovery was greater in group 1 (mean Tarlov score, 4.9 ± 0.057) than in group 2 (2.6 ± 0.3; P = .0028), group 3 (2.1 ± 0.34; P = .0088), or group 4 (0.0 ± 0.0; P = .0003).

CONCLUSIONS

Epidural cooling catheter combined with systemic moderate hypothermia produced additive cooling ability and protected the spinal cord against ischemia in rabbits more effectively than either intervention alone.

Protective use of N-methyl-D-aspartate receptor antagonists as a spinoplegia against excitatory amino acid neurotoxicity

OBJECTIVE

Paraplegia remains a serious complication of thoracic and thoracoabdominal aortic operations. To avoid this dreadful complication, N-methyl-D-aspartate (NMDA) receptor antagonists have been examined in the ischemic or excitotoxic neuronal injury model. In the present study, we evaluated the protective efficacy of NMDA receptor antagonists that were infused segmentally after aortic clamping, as a spinoplegia, to reduce aspartate neurotoxicity in the spinal cord.

METHODS

Infrarenal aortic isolation was performed in New Zealand white rabbits. Group A animals (n = 7) were pretreated with the segmental infusion of MK-801, a noncompetitive NMDA receptor antagonist, followed by segmental aspartate (50 mmol) infusion for 10 minutes. Group B animals (n = 6) received pretreatment with CGS19755, a competitive NMDA receptor antagonist, followed by the same aspartate infusion as group A. Group C animals (n = 7) received vehicle only, followed by aspartate infusion as a control group. In addition, group D animals (n = 6) were pretreated with MK-801 that was administrated intravenously 1 hour before aspartate infusion. Neurologic status was assessed at 12, 24, and 48 hours after operation by using the Tarlov score. The spinal cords were procured at 48 hours for histopathologic analysis to determine the extent of excitotoxic neuronal injury.

RESULTS

Most of the animals in groups A and D revealed full recovery or mild motor disturbance. Group B and C animals exhibited paraplegia or paraparesis with marked neuronal necrosis. In the Tarlov score at 48 hours, group A animals represented better neurologic function than group C (P < .01) and similar motor function to group D animals. Severe histopathologic change was not observed in groups A and D. Animals in groups A and D showed a greater number of motor neurons than animals in groups B and C (P < .01). The difference could be due to chance between group A and D animals (P = .08).

CONCLUSIONS

These results showed that the segmental infusion of noncompetitive NMDA receptor antagonist as an intraoperative spinoplegia could have a protective effect on the spinal cord neurons against excitotoxic neuronal injury in vivo. On the other hand, efficacy of the use of competitive antagonist was suggested to be limited in this model, probably because of the insurmountable obstacle of the blood-brain barrier.

CLINICAL RELEVANCE

Paraplegia is a devastating complication during surgical repair of the thoracic and thoracoabdominal aortas. Excitatory amino acids neurotoxicity through the N-methyl-D-aspartate (NMDA) receptor is no doubt the pathologic hallmark of ischemic and postischemic spinal cord injury. Systemic administration of either a competitive or noncompetitive NMDA antagonist has been reported to have neuroprotective effect, in terms of preoperative treatment, with dose-related central sympathomimetic and sedative effects. Local administration, particularly of a noncompetitive NMDA antagonist, infused segmentally after aortic clamping could therefore be a potent intraoperative pharmacologic strategy to minimize the effective dose that retains NMDA antagonism without undesirable adverse effects. Our ability to reproduce this model could facilitate pharmacologic prevention or provide a new surgical technique as a spinoplegia for NMDA receptor-mediated neuronal injury.

Effect of the free radical scavenger MCI-186 on spinal cord reperfusion after transient ischemia in the rabbit

OBJECTIVE

Paraplegia remains a serious complication of aortic operations. The production of free radicals during reperfusion after transient ischemia is believed to induce secondary spinal neuronal injury, resulting in paraplegia. The aim of the present study was to clarify the protective effect and method of administration of antioxidants on the neurological and histological outcome in the animal model for reperfusion injury after transient spinal cord ischemia.

METHODS

New Zealand white rabbits underwent surgical exposure of the abdominal aorta that was clamped for 15 minutes to achieve spinal cord ischemia. Group A animals received two 10 mg/kg doses of 3-methyl-1-phenyl-2-pyrazolin-5-one (MCI-186) at the time of release of the aortic clamp and 30 minutes later. In group B, MCI-186, 5 mg/kg, was given three times, at the time of aorta clamp release, 30 minutes and 12 hours later. In group C (control group), one dose of vehicle was administered. Neurological status was assessed using modified Tarlov's score until 168 hours after operation. Spinal cord sections were examined microscopically to determine the extent of ischemic neuronal damage.

RESULTS

Groups A and B animals had better neurological function than group C (p < 0.001). In contrast, group C animals exhibited paraplegia or paraparesis with marked neuronal necrosis. The number of surviving neurons within examined sections of the spinal cord was significantly greater in group B than in group C (p < 0.001).

CONCLUSION

In a 15-minute ischemia-reperfusion model using rabbits, systemic repetitious administration of MCI-186, a free radical scavenger, was found to have a protective effect on the spinal cord neurons both neurologically and histologically. We postulate that the drug minimizes the delayed neuronal cell death for reperfusion injury after transient ischemia by reducing the free radical molecules. Moreover, it was thought that we could protect delayed neuronal cell death more effectively by administering MCI-186 12 hours later.

Neuroprotection of adult rat dorsal root ganglion neurons by combined hypothermia and alkalinization against prolonged ischemia

Ischemia and ischemia-induced secondary events, such as acidosis and excessive activation of receptors by amino acids, trigger neuron death. The isolation and dissociation of dorsal root ganglion (DRG) involves time during which the neurons are ischemic due to being densely packed within the intact DRG and surrounded by a connective tissue coat. Thus, the longer the time between killing the host animal and when the DRG are dissociated, the longer the neurons are ischemic and exposed to ischemia-induced secondary causes of neuron death. It is well established that hypothermia and alkalinization each separately protect neurons from ischemia and ischemia-induced secondary causes of neuron death, but there are no data on the neuroprotection provided by simultaneous hypothermia and alkalinization. The present experiments were designed to determine the combination of hypothermic and alkaline conditions that yield the largest number of viable neurons dissociated from intact DRG maintained ischemic for up to 4 h. Hypothermia (20 degrees C>15 degrees C>37 degrees C) and alkalinization (pH 9.3>pH 8.3>pH 7.4) increased the yield of viable neurons compared with the yield from DRG maintained under physiological conditions. Hypothermia and alkalinization combined (20 degrees C/pH 9.3) provided the greatest neuroprotection with a yield of viable neurons after 1 h of ischemia 2.5-fold larger than that from DRG maintained under physiological conditions (37 degrees C/pH 7.6). Over 4 h of ischemia, the yield of viable neurons from DRG maintained under both hypothermic/alkaline and physiological conditions decreased in a linear manner, but those at 20 degrees C/pH 9.3 had a 4.5-fold greater yield of viable neurons than those at 37 degrees C/pH 7.6. Thus, combined hypothermia and alkalinization provide significantly greater protection against ischemia and ischemia-induced secondary causes of neuron death than either alone.

Neuroprotective effects of N-methyl-d-aspartate receptor antagonist on aspartate induced neurotoxicity in the spinal cord in vivo

OBJECTIVE

Much evidence has been gathered to show that neurotoxicity of excitatory amino acids is mainly activated through an N-methyl-D-aspartate (NMDA) receptor cascade. We evaluated the protective effects of NMDA receptor antagonists, MK-801 and CGS19755 on spinal cord neurons using the NMDA receptor mediated neurotoxicity model in vivo.

METHODS

New Zealand white rabbits underwent an infrarenal aortic isolation. Group A animals (n = 7) received segmental aspartate (50 mM) infusion for 10 minutes. Group B animals (n = 6) were pretreated with MK-801 (6mg/kg), a noncompetitive NMDA receptor antagonist, that was administrated intravenously for 3 hours beginning 1 hour before the segmental infusion of aspartate (50 mM) of 10 minutes. Group C animals (n = 6) received pretreatment with CGS19755 (30mg/kg), a competitive NMDA receptor antagonist, that was administrated in the same fashion as group B, followed by the segmental infusion of aspartate (50 mM). Neurologic status was scored at 12, 24, and 48 hours after operation using the Tarlov score. All the animals were sacrificed for histologic assessment at 48 hours.

RESULTS

Group A animals exhibited paraplegia or paraparesis with marked neuronal necrosis. Group B and C animals showed significantly better neurologic function compared with group A (p = 0.0013, A vs. B) (p = 0.0011, A vs. C). Pathohistological change was not observed in group B and C animals.

CONCLUSIONS

NMDA receptor antagonists can have protective effects on spinal cord neurons against aspartate induced neurotoxicity. This model may be useful in assaying protective agents in the spinal cord against neuronal injury mediated by NMDA receptors in vivo.

Exogenous aspartate neurotoxicity in the spinal cord under metabolic stress in vivo

BACKGROUND

Considerable evidence exists that neurotoxicity of excitatory amino acids is related to the neuronal injury, including paraplegia. However, little is known about aspartate neurotoxicity in the spinal cord in vivo. We evaluated the detrimental effects of exogenous aspartate on spinal cord neurons under metabolic stress.

METHODS

New Zealand white rabbits underwent an infrarenal aortic isolation. Group A animals (n = 7) received segmental aspartate 50 mmol/L) infusion for 10 minutes. Group B animals (n = 7) received saline as a negative control. Group C animals (n = 5) received segmental aspartate 100 mmol/L) infusion for 5 minutes. Group D animals (n = 7) were pretreated with segmental infusion of (+)-5-methyl-10,11-dihydro-5H-dibenzo[a,d]cycloheptan-5,10-imine (MK-801) (6 mg/kg), a noncompetitive N-methyl-D-aspartate (NMDA) receptor antagonist for 1 minute, followed by segmental infusion of aspartate (50 mmol/L) for 9 minutes. Group E animals (n = 7) received vehicle only, followed by aspartate (50 mmol/L) infusion as a control of group D. Neurologic status was assessed at 12, 24, and 48 hours after operation using the Tarlov score.

RESULTS

Group A animals exhibited paraplegia or paraparesis with marked neuronal necrosis. Group B and C animals recovered fully. Group D animals showed significantly better neurologic function (p = 0.0007) compared with group E animals that exhibited paraplegia or paraparesis.

CONCLUSIONS

Exogenous aspartate can have detrimental effects on spinal cord neurons under metabolic stress. This model may be useful in assaying neuronal injury mediated by NMDA receptor in vivo.

White matter injury in spinal cord ischemia: protection by AMPA/kainate glutamate receptor antagonism

BACKGROUND AND PURPOSE

Spinal cord ischemia is a serious complication of surgery of the aorta. NMDA receptor activation secondary to ischemia-induced release of glutamate is a major mechanism of neuronal death in gray matter. White matter injury after ischemia results in long-tract dysfunction and disability. The AMPA/kainate receptor mechanism has recently been implicated in white matter injury.

METHODS

We studied the effects of AMPA/kainate receptor blockade on ischemic white matter injury in a rat model of spinal cord ischemia.

RESULTS

Intrathecal administration of an AMPA/kainate antagonist, 6-nitro-7-sulfamoyl-(f)-quinoxaline-2, 3-dione (NBQX), 1 hour before ischemia reduced locomotor deficit, based on the Basso-Beattie-Bresnahan scale (0=total paralysis; 21=normal) (sham: 21+/-0, n=3; saline: 3.7+/-4.5, n=7; NBQX: 12. 7+/-7.0, n=7, P<0.05) 6 weeks after ischemia. Gray matter damage and neuronal loss in the ventral horn were evident after ischemia, but no difference was noted between the saline and NBQX groups. The extent of white matter injury was quantitatively assessed, based on axonal counts, and was significantly less in the NBQX as compared with the saline group in the ventral (sham: 1063+/-44/200x200 microm, n=3; saline: 556+/-104, n=7; NBQX: 883+/-103, n=7), ventrolateral (sham: 1060+/-135, n=3; saline: 411+/-66, n=7; NBQX: 676+/-122, n=7), and corticospinal tract (sham: 3391+/-219, n=3; saline: 318+/-23, n=7; NBQX: 588+/-103, n=7) in the white matter on day 42.

CONCLUSIONS

Results indicate severe white matter injury in the spinal cord after transient ischemia. NBQX, an AMPA/kainate receptor antagonist, reduced ischemia-induced white matter injury and improved locomotor function.

Prevention of ischemic spinal cord injury: comparative effects of magnesium sulfate and riluzole

PURPOSE

Excitotoxic mechanisms have been implicated in the pathophysiology of spinal cord ischemic injury induced by aortic cross-clamping. We investigated the effects of the anti-excitotoxic drugs magnesium sulfate (MgSO(4)) and riluzole in a rabbit model of spinal cord ischemia.

METHOD

The infrarenal aorta of New Zealand albino white rabbits (n = 68) was occluded for 40 minutes. Experimental groups included: a control group, which received only vehicle (n = 17); group A (n = 17), which received riluzole (8 mg/kg) before clamping; group B (n = 17), which received MgSO(4) (100 mg/kg) before clamping; and group C (n = 17), which received riluzole (8 mg/kg) and MgSO(4) (100 mg/kg) before clamping. Five additional rabbits had the same operation, but did not undergo aortic clamping (sham operation). The neurological status of the rabbits was assessed at 24 hours, 48 hours, and then daily for as long as 120 hours by using a modified Tarlov scale. The rabbits were killed at 24 hours (n = 3 per group), 48 hours (n = 4 per group), and 120 hours (n = 10 per group) postoperatively. Spinal cords were harvested for histopathologic and immunohistochemistry examinations for microtubule-associated protein-2 (MAP-2), a cytoskeletal protein specific from neurons.

RESULTS

No major adverse effect was observed with either riluzole or MgSO(4). All control rabbits became severely paraplegic. All riluzole-treated and MgSO(4)-treated animals had a better neurological status than control animals. Typical morphological changes characteristic of neuronal necrosis in the gray matter of control animals was demonstrated by means of the histopathological examination, whereas riluzole or magnesium prevented or attenuated necrotic phenomenons. Moreover, MAP-2 immunoreactivity was completely lost in control rabbits, whereas it was preserved, either completely or partially, in rabbits treated with riluzole or magnesium. Riluzole was more effective than MgSO(4) in preventing paraplegia caused by motor neuron injury (P <.01 ). Riluzole and MgSO(4) had no additive neuroprotective effect.

CONCLUSION

These results demonstrate that riluzole and, to a lesser extent, MgSO(4) may afford significant spinal cord protection in a setting of severe ischemia and may, therefore, be considered for clinical use during "high-risk" operations on the thoracic and thoracoabdominal aorta.

Selective perfusion of segmental arteries in patients undergoing thoracoabdominal aortic surgery

BACKGROUND

Reattachment of segmental arteries is one method used to prevent paraplegia associated with thoracoabdominal aortic repair. Nevertheless, even when important segmental arteries are reattached, ischemia causing spinal injury may occur during anastomosis.

METHODS

In 27 patients undergoing thoracoabdominal aortic repair, we attempted to perfuse the segmental arteries to be reattached with catheters connected to the distal bypass circuit. To identify perioperative risk factors for spinal ischemia, we examined changes in spinal somatosensory evoked potentials.

RESULTS

A median value of four segmental arteries were perfused in 20 (74%) of the 27 patients. Changes in somatosensory evoked potential indicative of spinal ischemia were observed in 13 patients (48%). The only risk factor associated with changes in evoked potentials revealed by a multivariate analysis was prolonged aortic cross-clamp time (> 120 minutes). Of the 2 patients who suffered paraplegia, one had the longest clamp time and the other showed spinal cord necrosis due to embolic shower.

CONCLUSIONS

Despite selective perfusion of segmental arteries, spinal ischemia associated with aortic cross-clamping may occur when clamping is prolonged over 120 minutes. Most of the changes appear to be reversible, however.

Retrograde perfusion with a sodium channel antagonist provides ischemic spinal cord protection

BACKGROUND

Neuronal voltage-dependent sodium channel antagonists have been shown to provide neuroprotection in focal and global cerebral ischemic models. We hypothesized that retrograde spinal cord venous perfusion with phenytoin, a neuronal voltage-dependent sodium channel antagonist, would provide protection during prolonged spinal cord ischemia.

METHODS

In a rabbit model, spinal cord ischemia was induced for 45 minutes. Six groups of animals were studied. Controls (group I, n = 8) received no intervention during aortic cross-clamping. Group II (n = 8) received systemic phenytoin (100 mg). Group III (n = 4) received systemic phenytoin (200 mg). Group IV (n = 8) received retrograde infusion of room temperature saline (22 degrees C) only. Group V (n = 8) and group VI (n = 9) received retrograde infusion of 50 mg and 100 mg of phenytoin, respectively, (infusion rate: 0.8 mL x kg(-1) x min(-1) during the ischemic period). Mean arterial blood pressure was monitored continuously. Animals were allowed to recover for 24 hours before assessment of neurologic function using the Tarlov scale.

RESULTS

Tarlov scores (0 = complete paraplegia, 1 = slight lower limb movement, 2 = sits with assistance, 3 = sits alone, 4 = weak hop, 5 = normal hop) were as follows (mean +/- SEM): group I, 0.50 +/- 0.50; group II, 0.25 +/- 0.46; group IV, 1.63 +/- 0.56; group V, 4.13 +/- 0.23; and group VI, 4.22 +/- 0.22 (p < 0.0001 V, VI versus I, II, IV by analysis of variance). No differences in mean arterial blood pressure were observed. All animals in group III became profoundly hypotensive and died before the conclusion of the 45-minute ischemic time.

CONCLUSIONS

Retrograde venous perfusion of the spinal cord with phenytoin, a voltage-sensitive sodium channel blocker, is safe and provides significant protection during prolonged spinal cord ischemia.

Riluzole prevents ischemic spinal cord injury caused by aortic crossclamping

BACKGROUND

Recent studies support the involvement of glutamate neurotoxicity in the pathophysiology of spinal cord injury induced by aortic crossclamping. We investigated the effects of riluzole, a neuroprotective drug that blocks glutamatergic neurotransmission, in a rabbit model of spinal cord ischemia.

METHODS

The infrarenal aortas of New Zealand White albino rabbits (n = 40) were occluded for 40 minutes. Experimental groups were as follows: sham operation group (n = 5), control group undergoing occlusion but receiving no pharmacologic intervention (n = 10), experimental group A (n = 10) receiving 8 mg/kg riluzole intravenously 30 minutes before ischemia, experimental group B (n = 10) receiving 4 mg/kg riluzole intravenously 30 minutes before ischemia and at the onset of reperfusion, and experimental group C (n = 10) receiving 8 mg/kg riluzole intravenously at the onset of reperfusion. Neurologic status was assessed at 6, 24, and 48 hours after the operation and then daily until the fifth day. All animals were killed at 24, 48, or 120 hours after the operation. Spinal cords were harvested for histopathologic studies, immunohistochemical studies for microtubule-associated protein 2, and search for morphologic features of apoptosis by the terminal deoxynucleotidyltransferase-mediated deoxyuridine triphosphate-biotin nick-end labeling staining method.

RESULTS

All animals in the control group became paraplegic. Except for 1 rabbit in group C, all riluzole-treated animals had better neurologic function. Luxol fast blue and terminal deoxynucleotidyltransferase-mediated deoxyuridine triphosphate-biotin nick-end labeling staining methods demonstrated typical morphologic changes characteristic of necrosis and apoptosis in control animals. Riluzole prevented or attenuated ischemia-induced necrosis, apoptosis, and cytoskeletal proteolysis, depending on the dose and the timing of administration.

CONCLUSION

Riluzole may have therapeutic utility during high-risk operations on the thoracoabdominal aorta.

Selective protection of gray and white matter during spinal cord ischemic injury

BACKGROUND

Ischemic injury in the gray matter is associated with excitatory amino acid neurotransmitters (EAA) release, and in the white matter is associated with intracellular sodium accumulation. We investigated the protective effect during spinal ischemia of the EAA antagonist, 2-carboxypiperazinyl-propylphosphonic acid (CPP), and the sodium channel blocker (2,6-dimethylphenylcarbamoylmethyl) triethylammonium bromide (QX).

METHODS

Sprague-Dawley rats were randomized in four groups, received intrathecally 10 microL of saline, CPP, QX, or QX/CPP, and underwent balloon occlusion of the aorta. Proximal pressure was lowered by exsanguination. In the acute protocol, 28 rats were used to calculate the length of occlusion, resulting in paraplegia in 50% of animals (P50). In the chronic study, 60 rats underwent 11' occlusion. The chronic animals were scored daily for 28 days and submitted to cord histology.

RESULTS

The P50 of QX (11'22") and QX/CPP (11'54") were longer than saline (10'39"), suggesting a beneficial effect. Neurologic scores of all treatment groups (p = 0.0001) and histologic scores of CPP (p = 0.003) and QX/CPP (p = 0.002) were better than saline.

CONCLUSIONS

Protection of spinal cord during ischemia can be achieved with intrathecal administration of selective agents directed to the gray and white matter.

[Glutamate neurotoxicity during spinal cord ischemia--neuroprotective effects of glutamate receptor antagonists]

Evidence is accumulating that glutamate, a major neurotransmitter, exerts potent neurotoxic activity during ischemia. In our laboratory, a delayed-onset paraplegia model using rabbits has been developed and described. The severity of the ischemic event in this model, i.e., extracellular glutamate overload, is believed to influence the etiology of this borderline lesion. We hypothesized that glutamate receptor antagonists (MK-801, NBQX) would attenuate the delayed neuronal dysfunction that follows spinal cord ischemia. Infrarenal aortic segments from 18 New Zealand white rabbits were isolated for 5 minutes and infused at a rate of 2 ml/min. Group I (n = 6) received normothermic L-glutamate (20 mM). Group II (n = 6) received 3 mg of MK-801 and normothermic L-glutamate (20 mM). Group III (n = 6) received 3 mg of NBQX and normothermic L-glutamate (20 mM). Neurologic function was assessed at 6, 24, and 48 hours after surgery according to the modified Tarlov scale. After 48 hours, the rabbits were euthanized and spinal cords were harvested for histologic examination. The neurologic function of three rabbits in group I showed acure paraplegia and the other three showed delayed-onset paraplegia, whereas all group II animals had nearly intact neurologic function and all group III animals showed mild neurologic disturbance. Histologic examination of spinal cords from rabbits in group I showed evidence of moderate spinal cord injury with necrosis of central gray matter and adjacent white matter and axonal swelling, whereas spinal cords from group II showed small and localized spinal cord injuries and those from group III revealed no evidence of cord injury. These results indicate that MK-801 and NBQX exert different neuroprotective effects related to different mechanisms of glutamate neurotoxicity mediated by the NMDA receptor and non-NMDA receptor, which initiate a deleterious cascade of biochemical events that ultimately results in delayed-onset paraplegia.

[Glutamate neurotoxicity during spinal cord ischemia--the neuroprotective effects of adenosine]

Evidence is accumulating that glutamate, a major neurotransmitter, exerts potent neurotoxic activity during ischemia. In our laboratory, a delayed-onset paraplegia model using rabbits has been developed and described. The severity of the ischemic event in this model, i.e., extracellular glutamate overload, is believed to influence the etiology of this delayed neuronal dysfunction. Adenosine, an endogenous neuromodulator, is released after acute ischemic insult and provides neuroprotection by actions on neuronal and glial cells in the still viable border zone of the ischemic focus. We hypothesized that the neuroprotective action of adenosine is associated with inhibition of glutamate neurotoxicity following ischemia. Infrarenal aortic segments from 11 New Zealand white rabbits were isolated for 5 minutes and infused at a rate of 2 ml/min. Group I (n = 6) received normothermic L-glutamate (20 mM). Group II (n = 5) received 75 mg of adenosine and normothermic L-glutamate (20 mM). Neurologic function was assessed at 6, 24, and 48 hours after surgery according to the modified Tarlov scale, After 48 hours, the rabbits were euthanized and their spinal cords were harvested for histologic examination. The neurologic function of three rabbits in group I showed acute paraplegia and the other three showed delayed-onset paraplegia, whereas all group II animals and nearly intact neurologic function. Histologic examination of spinal cords from rabbits in group I showed evidence of moderate spinal cord injury with central gray matter and adjacent white matter necrosis and axonal swelling, whereas spinal cords from group II revealed no evidence of cord injury. Adenosine A1-receptor activation is suspected to reduce excitatory amino acids by controlling the activation of the voltage-dependent NMDA receptor. These results indicate that the neuroprotective effect of adenosine is associated with inhibition of glutamate neurotoxicity, which initiates a deleterious cascade of biochemical events that ultimately result in delayed-onset paraplegia.