Ann Thorac Surg 1997;63:1057-1062
© 1997 The Society of Thoracic Surgeons
Original Article: Cardiovascular
Detrimental Effects of Exogenous Glutamate on Spinal Cord Neurons During Brief Ischemia In Vivo
Atsuo Mori, MD,
Toshihiko Ueda, MD,
Tsukasa Nakamichi, MD,
Mikito Yasudo, MD,
Ryo Aeba, MD,
Hiroshi Odaguchi, MD,
Atsuhiro Mitsumaru, MD,
Tsutomu Ito, MD,
Ryohei Yozu, MD,
Atsuo Koto, MD,
Shiaki Kawada, MD
Departments of Cardiovascular Surgery and Neurology, Keio University, Tokyo, Japan
Accepted for publication October 30, 1996.
 |
Abstract
|
|---|
Background. Paraplegia remains a serious complication of thoracoabdominal aortic operations. However, despite growing in vitro evidence, it has been difficult to demonstrate glutamate neurotoxicity in vivo because of the reuptake activity that occurs. We hypothesized that glutamate can be toxic to the spinal cord under metabolic stress.
Methods. Infrarenal aortic isolation was performed in New Zealand white rabbits. Group A animals (n = 7) then received a segmental infusion of glutamate (50 mmol/L) for 5 minutes. Group B animals (n = 7) received saline as a negative control. Group C animals (n = 6) were pretreated with a segmental infusion of 2,3- dihydroxy-6-nitro-7-sulfamoyl-benzo(f)-quinoxaline (4 mg/kg), a competitive 
-amino-3-hydroxy-5-methylisoazole-4-propionic acid/kainate antagonist, followedby the segmental infusion of glutamate (30 mmol/L) for 4 minutes. Group D animals (n = 6) received the vehicle agents only, followed by the same glutamate infusion (30 mmol/L) as in group C as a control for group C. Neurologic status was assessed at 12, 24, and 48 hours after operation and scored using the Tarlov system.
Results. Group A animals exhibited paraplegia or paraparesis with marked neuronal necrosis. Group B animals recovered fully. Group C animals had better neurologic function than group D animals (p = 0.0039).
Conclusions. Exogenous glutamate can have detrimental effects on spinal cord neurons during a brief period of ischemia. This model may be useful for the purpose of assaying a glutamate receptor antagonist in vivo.
|
Introduction
|
|---|
Paraplegia remains a serious surgical complication associated with the repair of the thoracic and thoracoabdominal aorta. The prevalence of postoperative paraplegia has been reported to range from 5% to 20%, depending on the site and extent of the aortic aneurysm [1]. Numerous surgical and pharmacologic interventions have been proposed for preventing paraplegia; however, no method yet completely avoids this dreadful complication.
Glutamate is an excitatory neurotransmitter that is abundant in the central nervous system, including the spinal cord. Recent evidence has shown that glutamate is potentially neurotoxic to brain neurons and is involved in the pathophysiologic mechanism of a variety of neuronal diseases. Glutamate has been reported to destroy neurons through its actions on N-methyl-D-aspartate (NMDA) and non-NMDA receptors, by inducing calcium and sodium ion influxes that ultimately lead to neuronal death in cortical cell cultures [2].
Glutamate is not toxic to neuronal tissue when it is applied exogenously under physiologic conditions, probably because the extracellular glutamate concentration is maintained within the nontoxic level by presynaptic reuptake [3]. We undertook this study to evaluate the toxicity of exogenously administered glutamate in vivo during a short period of ischemia. We also examined the protective effects of 2,3-dihydroxy-6-nitro-7-sulfamoyl-benzo(f)-quinoxaline (NBQX), a highly selective antagonist of the -amino-3-hydroxy-5-methylisoazole-4-propionic acid (AMPA)/kainate glutamate receptor, against the spinal cord injury in our model.
|
Material and Methods
|
|---|
Experimental Model
All New Zealand white rabbits weighing 3.5 to 4.0 kg were anesthetized with 1.5% halothane and 98.5% oxygen. Animals were then placed supine and allowed to breathe spontaneously without endotracheal intubation or mechanical ventilation. A venous catheter (24 gauge) was placed in a marginal ear vein, and cefazolin (10 mg/kg) was administered as a single dose. Central ear arterial pressure and rectal temperature were monitored continuously. A heating pad was placed underneath each animal to maintain normal body temperature. After sterile preparation, a median laparotomy was made. The abdominal aorta was dissected just inferior to the left renal vein and above the bifurcation. After systemic anticoagulation with heparin sulfate (60 U/kg intravenous bolus), a 20-gauge catheter was introduced over a guide wire from the right femoral artery to the abdominal aorta. The tip of this catheter, which could be seen through the aorta, was placed 5 mm above the bifurcation of the aorta.
Group A rabbits (n = 7) received a segmental infusion of 50-mmol/L glutamate solution, which was adjusted to physiologic osmolarity with NaCl and distilled water, through the femoral arterial catheter. At the same time, the abdominal aorta was clamped with a vascular clip immediately inferior to the left renal vein and the femoral arterial catheter was snared around the bifurcation of the aorta. The posterior mesenteric artery was also clamped with a vascular clip. The infused glutamate solution was warmed to 39°C in an incubator and infused into the isolated segment at a rate of 2 mL/min for 5 minutes (Fig 1
). After 5 minutes the clamps were released and the snare around the bifurcation loosened. The infusion catheter was removed, and the abdomen was closed in two layers. Saline solution alone was infused at the same rate and in the same manner as in group B animals (n = 7). The rabbits were then returned to their cages and allowed free access to water and food.
View larger version (34K):
[in this window]
[in a new window]
|
Fig 1. . Experimental model. (PMA = posterior mesenteric artery; NBQX = 2,3-dihydroxy-6-nitro-7-sulfamoyl-benzo[f]-quinoxaline.)
|
|
|
Pharmacologic Study
|
|---|
The NBQX was purchased from Tocris Cockson Ltd (Bristol, UK) and dissolved in 0.9% NaCl to reach a concentration of 4 mg/mL. Group C rabbits (n = 6) were pretreated with the segmental infusion of NBQX (4.0 mg/kg) administered through the femoral catheter for 2 minutes, followed by the segmental infusion of 30-mmol/L glutamate delivered at a rate of 2 mL/min for 4 minutes. Group C animals received a continuous intravenous infusion of NBQX (0.1 mg/kg per hour) for 6 hours after the operation. This tapering protocol was employed to exclude the paradoxical deleterious effects of NBQX resulting from glutamate receptor up-regulation [4]. Group D rabbits (n = 6) received only the vehicles used for NBQX treatment, followed by a segmental infusion of 30-mmol/L glutamate administered in the same manner as was done in group C.
|
Neurologic and Histopathologic Evaluation
|
|---|
Hindlimb function was scored at 12, 24, and 48 hours after operation using the modified Tarlov system (0 = no movement, 1 = slight movement, 2 = sits with assistance, 3 = sits alone, 4 = weak hop, 5 = normal hop). After 48 hours animals were reanesthetized and sacrificed with an overdose of intravenously administered pentobarbiturate. The spinal cord was fixed with a 10% formaldehyde solution perfused through a femoral arterial catheter in a fashion identical to that used at the first operation. This was followed by immersion fixation for 2 weeks. Cross-sections from the lower thoracic through the sacral cord were stained with hematoxylin-eosin, Luxol fast blue, and Nissl, and the neuropathologist, who was blinded to the experimental group, performed the histologic assessment using light microscopy.
|
Statistical Analysis
|
|---|
We used the Mann-Whitney U test to compare the postoperative neurologic status of animals between the groups. Other parameters between the two groups were compared by analysis of variance for repeated measures or Student's t test, where appropriate. A p value of 0.05 was taken to indicate significance.
|
Animal Care
|
|---|
All animals received humane care and treatment in accordance with the "Guide for the Care and Use of Laboratory Animals" published by the National Institutes of Health (NIH publication 85-23, revised 1985).
|
Results
|
|---|
Five of 7 rabbits in group A exhibited paraplegia (Tarlov score of 0) at 12, 24, and 48 hours after the procedure. Two rabbits showed paraparesis (Tarlov score of 1) at all the time points (Table 1). Group B animals showed intact neurologic function (Tarlov score of 5) at all the time points. Group A rabbits had a significantly worse neurologic score than the animals in group B (p < 0.001).
Three of 6 rabbits in group C had recovered fully (Tarlov score of 5) at 48 hours. The other 3 rabbits exhibited a mild to moderate disturbance in motor function (Tarlov score of 3 or 4) at 48 hours. Three rabbits in group D exhibited paraplegia (Tarlov score of 0), and 1 rabbit showed paraparesis (Tarlov score of 1) at 48 hours. Group C rabbits had a significantly better neurologic score than the animals in group D (p = 0.0039).
Histologic studies showed neuronal injury in all spinal cords of the animals in group A. Sections from group A showed severe and extensive gray matter necrosis with vascular necrosis in the anterior and dorsal horns of the lumbosacral cord. Axonal degeneration was noted in the white matter in group A (Fig 2). In contrast, group B showed a normal histologic appearance (Fig 3). Group C showed a mild neuronal change with hyperchromatic nuclei and eosinophilic changes localized in the anterior horns in some sections of the sacral cord. There was no white matter degeneration in group C (Fig 4). Group D exhibited moderate gray matter degeneration with a limited lesion in the white matter (Fig 5).
View larger version (126K):
[in this window]
[in a new window]
|
Fig 2. . (A, B) Photomicrographs of sections of spinal cord from a rabbit in group A. (Hematoxylin-eosin and Luxol-fast blue: A, x25; B, x125, both before 53% reduction.)
|
|
View larger version (122K):
[in this window]
[in a new window]
|
Fig 3. . (A, B) Photomicrographs of histologic sections of spinal cord from a rabbit in group B. (Hematoxylin-eosin and Luxol-fast blue: A, x25; B, x125, both before 53% reduction.)
|
|
View larger version (127K):
[in this window]
[in a new window]
|
Fig 4. . (A, B) Photomicrographs of histologic sections of spinal cord from an animal in group C. (Hematoxylin-eosin and Luxol-fast blue: A, x25; B, x125, both before 53% reduction.)
|
|
View larger version (142K):
[in this window]
[in a new window]
|
Fig 5. . (A, B) Photomicrographs of histologic sections of spinal cord from an animal in group D. (Hematoxylin-eosin and Luxol-fast blue: A, x25; B, x125, both before 53% reduction.)
|
|
Rectal temperatures were similar in group A and B animals and in group C and D animals immediately before the infrarenal aorta was clamped (Table 2). There were no significant differences between the groups in the systemic mean blood pressure and pulse rate during aortic clamping (Table 2).
|
Comment
|
|---|
This study showed that the regional administration of glutamate during 5 minutes of ischemia causes spinal cord injury in rabbits in vivo. Furthermore, the protective effects conferred by NBQX indicate that the neuronal injury induced in our experimental setup may be due to glutamate neurotoxicity.
There is growing evidence that glutamate neurotoxicity contributes to the spinal cord injury that occurs during ischemia. For example, Regan and Choi [5] have shown that a 5-minute exposure to glutamate induces neuronal degeneration in rat spinal cord cells in culture. However, although numerous in vitro studies have been performed that show this, it has been difficult to demonstrate glutamate neurotoxicity in vivo. For example, Mangano and Schwarcz [3] infused a 300-mmol/L glutamate solution into the hippocampus of a rat for 2 weeks at a rate of 0.5 µL/h but found that this caused no neuronal injury. It was assumed that this was due to the activity of a high-affinity reuptake system present in presynaptic nerve terminals and glia that removes extracellular glutamate from the synaptic cleft and masks the neurotoxicity of exogenous glutamate in vivo. This is further supported by the fact that the infusion of glutamate analogues, including NMDA, quisqualate, and kainate, which are not taken up by the high-affinity uptake system, induces limbic paraplegia and histologic neuronal degeneration in rat in vivo [6, 7].
Two factors are probably responsible for causing the glutamate neurotoxicity seen in the current study. The first is that infrarenal aortic clamping resulted in spinal cord ischemia and reduced the efficiency of the high-affinity reuptake system in removing glutamate. Because this reuptake system is highly dependent on the adenosine triphosphate energy store, the depletion of adenosine triphosphate brought about by ischemia may cause the extracellular accumulation of glutamate. The second factor is that the regional infusion of glutamate into the isolated aortic segment may have prevented the exogenous glutamate from being washed out by blood flow and caused the extracellular glutamate concentration to reach toxic levels. Because regional infusion has been reported to be a useful technique for administering an agent to the spinal cord in high concentrations [8], we therefore used this technique to administer the glutamate in our model.
In previous studies infrarenal aortic clamping for 15 to 20 minutes was found to induce paraplegia and neuronal degeneration in rabbit spinal cord [9]. However, in the present study 5 minutes of infrarenal aortic clamping was not observed to be sufficient to produce neuronal injury, though the combination of 5 minutes of ischemia and regional glutamate infusion produced paraplegia or paraparesis and neuronal degeneration. This indicates that 5 minutes of ischemia may be long enough to compromise the ability of the high-affinity reuptake system to remove glutamate. These results agree with those from previous studies, which showed that about 3 minutes of ischemia causes the depletion of the intracellular adenosine triphosphate store [10]. The glutamate that accumulated during 5 minutes of ischemia may have then excessively activated the NMDA and non-NMDA receptors in postsynaptic neuronal membranes that act as cationic ionophores, permitting sodium and calcium ion influx, and ultimately leading to neuronal death [11].
After removal of the aortic clamp and discontinuation of the glutamate infusion, exogenous glutamate in the synaptic cleft may have been washed out by reperfusion. However, the intracellular elevation of calcium ions during the 5 minutes of ischemia may have activated several calcium-dependent enzymes, including protein kinase C, phospholipase A2, nitric oxide synthase, and other proteases, leading to late neuronal degeneration. It is also possible that endogenous glutamate leaking from injured neurons sustained the increase in the extracellular level of glutamate and produced further neuronal damage in a self-propagating manner. In fact, a sustained elevation of glutamate level has been detected, not only during ischemia, but also after reperfusion in spinal cord ischemia models using stereotactic microdialysis [12, 13].
The histologic findings in group A (the glutamate infusion group) were analogous to those seen in the setting of spinal cord infarction. In this group axonal injury was also detected in the severely injured segment of the spinal cord. Degeneration of the white matter was observed only in that white matter surrounding the lesion in the severely injured gray matter. There was no instance of isolated white matter degeneration. Therefore degeneration of the white matter was probably due to toxic agents, such as nitric oxide, that diffused out from the injured gray matter. However, central gray matter necrosis with relative sparing of the white matter is typical in the setting of glutamate neurotoxicity [4, 7]. This discrepancy may be due to the differences in the concentration of the glutamate infused. The lower concentration of glutamate (30 mmol/L) was noted to induce histologic degeneration with relative axonal sparing identical to that seen in previous studies. For this reason, we chose this lower glutamate concentration to examine the effects of the glutamate receptor antagonist. The NBQX-treated rabbits showed no white matter injury. It is possible therefore that secondary reactions triggered by the deregulation of calcium homeostasis lead to the white matter degeneration.
Faden and Simon [14] have reported that MK-801, a noncompetitive NMDA receptor antagonist, attenuates the spinal cord injury in rat in vivo. In addition, a non-NMDA antagonist has recently been reported to reduce the neuronal degeneration caused by brain and spinal cord ischemia [4, 15]. Moreover, it has been demonstrated that the non-NMDA receptors rather than the NMDA receptors, play an important role in spinal motor neuron toxicity [16]. For this reason we chose to administer NBQX, the selective AMPA/kainate receptor antagonist, as a blocker in our pharmacologic study. The NBQX was found to partially protect the spinal cord against the injury produced by the segmental infusion of glutamate. These results support the notion that the neuronal degeneration occurring in this model may be due to glutamate excitotoxicity. There are several possible explanations for the glutamate neurotoxicity mediated by the AMPA/kainate receptor. For example, excessive AMPA/kainate receptor activation is assumed to cause sodium and passive chloride influx, which leads to neuronal swelling. Subsequent depolarization could cause the intracellular calcium ion concentration to be increased through the voltage-dependent calcium channel and sodium-calcium ion exchanger. In addition, previous studies have demonstrated that activation of the AMPA/kainate receptor can stimulate the release of xanthine oxidase, leading to the formation of free radicals [17]. We speculate that the glutamate excitotoxicity mediated by the NMDA receptor might also participate in this process. This excitotoxicity might contribute to the formation of the residual degenerative lesions, which were not completely prevented by NBQX treatment in our study.
The rabbit spinal cord ischemia model caused by occluding the infrarenal aorta is a highly established and useful means of producing local ischemia. However, the mechanism of the resulting paraplegia is multifactorial. Hypoxia, hypoperfusion, anaerobic metabolism, ischemic-reperfusion injury, platelet adhesion, neutrophil activation, and glutamate neurotoxicity may all contribute to its development. However, considering that a clamp time as short as 5 minutes is all that is necessary in our model to produce spinal cord injury, factors besides glutamate neurotoxicity may be minimal in our model as compared with the rabbit spinal cord ischemia model. We therefore believe that our model may provide a refined approach to the evaluation of the ability of protective agents to reduce glutamate neurotoxicity in ischemic spinal cord injury.
|
Acknowledgments
|
|---|
We gratefully acknowledge the help of Mr Takashi Kimura with the pathologic analysis. We are also grateful to Etsuko Doi for her technical assistance.
|
Footnotes
|
|---|
Address reprint requests to Dr Mori, Department of Cardiovascular Surgery, Keio University, 35 Shinanomachi, Shinjuku, Tokyo 160, Japan.
|
References
|
|---|
- Svensson LG, Crawford ES, Hess KR, Coselli JS, Safi HJ. Experience with 1509 patients undergoing thoracoabdominal aortic operations. J Vasc Surg 1993;17:357–70.[Medline]
- Choi DW, Maulucci-Gedde M, Kriegstein AR. Glutamate neurotoxicity in cortical cell culture. J Neurosci 1987;7:357–68.[Abstract]
- Mangano RM, Schwarcz R. Chronic infusion of endogenous excitatory amino acids into rat striatum and hippocampus. Brain Res Bull 1983;10:47–51.[Medline]
- Redmond JM, Zehr KJ, Blue MY, et al. AMPA glutamate receptor antagonism reduces neurologic injury after hypothermic circulatory arrest. Ann Thorac Surg 1995;59:579–84.[Abstract/Free Full Text]
- Regan RF, Choi DW. Glutamate neurotoxicity in spinal cord cell culture. Neuroscience 1991;43:585–91.[Medline]
- Hugon J, Vallat JM, Spencer PS, Leboutet MJ, Barthe D. Kainic acid induces early and delayed degenerative neuronal changes in rat spinal cord. Neurosci Lett 1989;104:258–62.[Medline]
- Pisharodi M, Nauta H. An animal model for neuron-specific spinal cord lesions by the microinjection of N-methylaspartate, kainic acid, and quisqualic acid. Appl Neurophysiol 1985;48:226–33.[Medline]
- Seibel PS, Theodore P, Kron IL, Tribble CG. Regional adenosine attenuates postischemic spinal cord injury. J Vasc Surg 1993;18:153–60.[Medline]
- Martinez-Arizala A, Long JB, Riggamonti DD, et al. Deteriorating stroke model of spinal ischemia in the rabbit is associated with a marked hyperemia. Neurology 1989;39:371–2.
- Mauney MC, Blackbourne LH, Langenburg SE, Buchanan SA, Kron IL, Tribble CG. Prevention of spinal cord injury after repair of the thoracic or thoracoabdominal aorta. Ann Thorac Surg 1995;59:245–52.[Abstract/Free Full Text]
- McDonald JW, Silverstein FS, Cardona D, Hudson C, Chen R, Johnston MV. Systemic administration of MK-801 protects against N-methyl-D-aspartate and quisqualate-mediated neurotoxicity in perinatal rats. Neuroscience 1990;36:589–99.[Medline]
- Simpson RK, Robertson CS, Goodman JC. Spinal cord ischemia–induced elevation of amino acids: extracellular measurement with microdialysis. Neurochem Res 1990;15:635–9.[Medline]
- Rokkas CK, Helfrich LR, Lobner DC, Choi DW, Kouchoukos NT. Dextrorphan inhibits the release of excitatory amino acids during spinal cord ischemia. Ann Thorac Surg 1994;58:312–20.[Abstract]
- Faden AI, Simon RP. A potential role for excitotoxins in the pathophysiology of spinal cord injury. Ann Neurol 1988;23:623–6.[Medline]
- Euler M, Seiger A, Holmberg L, Sundstrom E. NBQX, a competitive non-NMDA receptor antagonist, reduces degeneration due to focal spinal cord ischemia. Exp Neurol 1994;129:163–8.[Medline]
- Urca G, Urca R. Neurotoxic effects of excitatory amino acids in the mouse spinal cord: quisqualate and kainate but not N-methyl-D-aspartate induce permanent neural damage. Brain Res 1990;529:7–15.[Medline]
- Dykens JA. Isolated cerebral and cerebellar mitochondria produce free radicals when exposed to elevated Ca2+ and Na+: implications for neurodegeneration. J Neurochem 1994;63:584–91.[Medline]
This article has been cited by other articles:
|
|
|
|
 
Y. Cho, T. Ueda, A. Mori, T. Nakamichi, H. Shimizu, Y. Inoue, and S. Kawada
Exogenous aspartate neurotoxicity in the spinal cord under metabolic stress in vivo
Ann. Thorac. Surg.,
November 1, 2000;
70(5):
1496 - 1500.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
G. K. Kanellopoulos, X. M. Xu, C. Y. Hsu, X. Lu, T. M. Sundt, N. T. Kouchoukos, and P. H. Chan
White Matter Injury in Spinal Cord Ischemia : Protection by AMPA/Kainate Glutamate Receptor Antagonism Editorial Comment: Protection by AMPA/Kainate Glutamate Receptor Antagonism
Stroke,
August 1, 2000;
31(8):
1945 - 1952.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
T. Ueda, H. Shimizu, A. Mori, I. Kashima, K. Moro, and S. Kawada
Selective perfusion of segmental arteries in patients undergoing thoracoabdominal aortic surgery
Ann. Thorac. Surg.,
July 1, 2000;
70(1):
38 - 43.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
J. J. Gangemi, J. A. Kern, S. D. Ross, K. S. Shockey, I. L. Kron, and C. G. Tribble
Retrograde perfusion with a sodium channel antagonist provides ischemic spinal cord protection
Ann. Thorac. Surg.,
June 1, 2000;
69(6):
1744 - 1748.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
L. Lang-Lazdunski, C. Heurteaux, N. Vaillant, C. Widmann, and M. Lazdunski
RILUZOLE PREVENTS ISCHEMIC SPINAL CORD INJURY CAUSED BY AORTIC CROSSCLAMPING
J. Thorac. Cardiovasc. Surg.,
May 1, 1999;
117(5):
881 - 889.
[Abstract]
[Full Text]
[PDF]
|
|
|
|
|
|
|
F. M. Follis, K. S. Blisard, P. S. Varvitsiotis, S. B. Pett Jr, R. T. Temes, and J. A. Wernly
Selective protection of gray and white matter during spinal cord ischemic injury
Ann. Thorac. Surg.,
May 1, 1999;
67(5):
1362 - 1369.
[Abstract]
[Full Text]
[PDF]
|
|
|
Top
Abstract
Introduction
