| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2002;74:838-845
© 2002 The Society of Thoracic Surgeons
a Division of Cardiac Surgery, The Johns Hopkins Medical Institutions, Baltimore, Maryland, USA
b Division of Cardiovascular Medicine, The Johns Hopkins Medical Institutions, Baltimore, Maryland, USA
c Division of Neuropathology, The Johns Hopkins Medical Institutions, Baltimore, Maryland, USA
d Division of Neurology, The Johns Hopkins Medical Institutions, Baltimore, Maryland, USA
* Address reprint requests to Dr Gott, Division of Cardiac Surgery, Johns Hopkins Medical Institutions, Blalock 618, 600 North Wolfe St, Baltimore, MD 21287, USA
e-mail: vgott{at}csurg.jhmi.jhu.edu
Presented at the Forty-eighth Annual Meeting of the Southern Thoracic Surgical Association, San Antonio, TX, Nov 8–10, 2001.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and methodsResultsCommentAcknowledgmentsDiscussionReferences |
|---|
Methods. The infra-renal aorta was cross-clamped in 45 male New Zealand white rabbits for 20 minutes. Control animals received no pretreatment. Diazoxide-treated animals were dosed (5 mg/kg) 15 minutes before cross-clamp. A third group underwent 5 minutes of IPC 30 minutes before cross-clamp. Two groups received KATP antagonists, 5-hydroxydecanoic acid (5-HD, 20 mg/kg) or glibenclamide (1.0 mg/kg), before diazoxide administration. Systemic hypotension was induced in a final group with excess isoflurane. Tarlov Scoring was used to assess neurologic function at 24 and 48 hours, after which, the spinal cords were procured for histopathological analysis.
Results. Tarlov scoring demonstrated marked improvement in the Diazoxide group compared with control at 24 hours (p < 0.02) and 48 hours (p < 0.009). Moreover, no further neurologic injury occurred in this group at 7 days. IPC-treated animals showed neurologic improvement but were not significantly different from controls. Further, administration of glibenclamide was effective in antagonizing diazoxide’s protective effect.
Conclusions. Administration of diazoxide resulted in significant improvement in neurologic outcome in this model. This protective effect improved outcome at both early and late time points. Further, the antagonistic effect of glibenclamide implicates diazoxide’s ATP-dependent potassium channel agonism as the mechanism of protection. Overall, this study suggests that diazoxide may be useful in the prevention of neurologic injury after thoracic aneurysm surgery.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsDiscussionReferences |
|---|
Even less well understood is the phenomenon of delayed-onset paraplegia seen in a small subset of patients who will awaken with no paralysis only to manifest a neurologic deficit 1 to 5 days postoperatively. Several theories have been put forth in an attempt to explain this delayed-onset paraplegia, including postoperative hypotension, thrombosis of or embolization to the anterior spinal artery, and occlusion of reimplanted intercostal arteries [2]. None of these, however, adequately explains this phenomenon, suggesting the presence of additional mechanisms of spinal cord injury after extensive thoracoabdominal aneurysm repairs. Although ischemic neuronal cell death is often attributed to necrosis [3], recent studies have demonstrated that ischemia and reperfusion can evoke DNA fragmentation and chromatin condensation (indicative of apoptosis), leading to cell death in 24 to 48 hours [4]. This, in part, may help to explain the development of neurologic deficits in initially normal postoperative patients.
Ischemic preconditioning (IPC) was first described in 1986 as a paradoxical form of cardio-protection, whereby brief ischemic insults can protect the myocardium from subsequent lethal ischemia [5]. This effect, initially thought to be mediated through the activation of cell surface ATP-sensitive potassium channels (surfaceKATP) [6], can be abolished by the KATP channel antagonist, glibenclamide [7]. Subsequent studies have, however, demonstrated that the mechanism of this preconditioning is likely to be, at least in part, modulated through activation of mitochondrial ATP-sensitive potassium (mitoKATP) channels [8–10]. In these studies, diazoxide, a potent and specific mitoKATP channel opener, mimics the protection seen with short periods of noninjurious ischemia. Further, 5-hydroxydecanoic acid (5-HD), a specific mitoKATP antagonist, has been shown to abolish both ischemic preconditioning and diazoxide-induced pharmacological preconditioning [11]. More recently, ischemic preconditioning (IPC) has been shown to reduce neurologic injury in models of spinal cord ischemia [12, 13, 14].
Therefore, we undertook this study to determine if pharmacological induction of preconditioning by the mitochondrial KATP channel agonist, diazoxide, could reduce both immediate and delayed-onset paraplegia in a well-characterized rabbit model of the spinal cord ischemia.
|
Material and methods |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsDiscussionReferences |
|---|
Forty-five male New Zealand white rabbits weighing 2.0 to 4.0 kg were used in these experiments. The rabbits were allowed access to standard rabbit food and water ad libitum. All rabbits were neurologically intact before anesthesia and instrumentation.
Surgical procedure
The rabbits were initially anesthetized with sodium pentobarbital (20 mg/kg) administered through a 24-gauge angiocath placed in an ear vein. A second 22-gauge catheter was then placed in an artery in the ear for blood pressure monitoring. Animals were then shaved, endotracheally intubated, and placed in a right side down position. Anesthesia was maintained with isoflurane (1 to 2%) administered via a veterinary Narkomed 2 anesthesia system. A probe was inserted 8.0 cm into the rectum, and body temperature was measured and maintained between 37.0° and 38.5°C with a circulating warm water underbody heating pad. All animals were prepared and draped in a sterile fashion consistent with the guidelines set forth by the institutional Animal Care and Use Committee for chronic, sterile surgical preparations.
A 5.0-cm left flank incision was then made, and dissection was carried out to expose the retroperitoneal, infra-renal aorta. After the dissection, all animals were given 300 U of heparin sodium intravenously. During the experimental protocol, an atraumatic vascular clamp was applied within 1.0 cm of the renal artery takeoff to occlude the aorta for 20 minutes. Upon release of the cross-clamp, animals were resuscitated with Lactated Ringers and the wounds closed in two layers with 2-0 Vicryl suture. The rabbits received a combination of buprenorphine (0.1 mg IM q 12 hours) and acepromazine (0.6 mg IM q 12 to 24 hours) for the duration of the experiments (48 hours to 7 days) to treat pain and anxiety.
Experimental protocol
After intubation and baseline hemodynamic measurements, the rabbits were placed in one of six groups (Fig 1):
control (n = 7), diazoxide (DZ, n = 13), ischemic preconditioning group (IPC, n = 6), 5-hydroxydecanoic acid plus diazoxide (5-HD+DZ, n = 6), glibenclamide plus diazoxide (Glib+DZ, n = 6), and the isoflurane group (ISO, n = 7). The control group received normal saline vehicle alone. The DZ group received a dose of 5 mg/kg IV, 15 minutes before the 20-minute cross-clamp. The IPC group underwent a 5-minute ischemic insult followed by 30 minutes of reperfusion before the cross-clamp. The 5-HD+DZ group received 5-HD (20 mg/kg IV) 5 minutes before DZ administration, whereas the Glib+DZ group received glibenclamide (1 mg/kg) 30 minutes before DZ administration. In the ISO group, we manipulated the systemic blood pressure by varying the amount of inhalational anesthetic agent (isoflurane) to mimic the transient systemic hypotensive effect caused by diazoxide during the 15-minute pre-cross-clamp period.
|
Histologic study
All rabbits were sacrificed using sodium pentobarbital (150 mg/kg) administered intravenously through an ear vein followed by antegrade, aortic perfusion of 2 L cold (0.9%) NaCl. The thoracic (normal control) and lumbar (experimental region of ischemia) were immediately extracted and flash fixed in 10% buffered formalin. They were embedded in paraffin and serial transverse sections were cut (8 µm) for hematoxylin and eosin staining and specific staining of DNA fragmentation (deoxynucleotidyltransferase-mediated dUTP-biotin nick end labeling [TUNEL]) using a biological-histochemical system (Apop Tag kit; Intergen Company, Purchase, NY). For TUNEL staining, sections were deparaffinized in Citrosolve (Fisher Scientific) and rehydrated in baths of descending ethanol concentration. To expose the DNA, slides were then gently boiled in a microwave for three cycles of 5 minutes in 10 mmol/L sodium citrate, pH 6.0. All subsequent steps were performed as outlined in the Apoptag manual. The color was developed in a solution of DAB and H2O2 in phosphate-buffered saline. Sections were counterstained in methyl green.
Histologic damage was scored using a system developed in our laboratory for this study. A neuropathologist, blinded to the treatment, was responsible for all the scoring. A score of 0 to 4 was assigned to each section as follows: 0, frank necrosis; 1, severe cellular damage; 2, moderate cellular damage; 3, mild cellular damage; and 4, normal histologic appearance.
Chemicals
Diazoxide (Hyperstat [15 mg/mL], Schering, Inc., Kenilworth, NJ) was obtained from the hospital pharmacy. 5-Hydroxydecanoic acid (5-HD), and glibenclamide (Glib) were purchased from Sigma Chemical Co. (St. Louis, MO). 5-HD was dissolved in 0.9% NaCl to a concentration of 10 mg/mL. Glibenclamide was dissolved in polyethylene gycol 400, 0.1 N NaOH, 95% EtOH, and 0.9% NaCl in a 1:1:1:2 cocktail to a concentration of 1.0 mg/mL.
Statistical analysis
Data are presented as means ±SEM. Statistical comparison was evaluated by the two-tailed unpaired Student’s t test, with p less than 0.05 considered significant. The Spearman rank order correlation test was used to evaluate the association between neurologic function and histopathological score, with p less than 0.05 considered significant.
|
Results |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsDiscussionReferences |
|---|
Twenty minutes of infra-renal aortic cross-clamp resulted in severe lower-extremity neurologic deficits in the control group, with average Tarlov scores at 24 and 48 hours of 1.43 ± 0.81 and 1.0 ± 0.68, respectively (Fig 2a). Diazoxide (DZ)-treated animals had significantly better Tarlov scores at both 24 hours (4.0 ± 0.36, p < 0.02) and 48 hours (3.83 ± 0.21, p < 0.009). Further, when assessed for 7 days, diazoxide-treated animals demonstrated no delayed-onset paraplegia (Tarlov: 4.25 ± 0.25, p = NS compared with DZ at 48 hours). Although the group that underwent ischemic preconditioning (IPC) also appeared to have a better outcome compared with control, this difference did not reach statistical significance.
|
|
|
To objectively quantify the amount of neuronal damage seen in these histologic sections, we devised a scoring system described where a score of 0 to 4 was assigned to each section as follows: 0, frank necrosis; 1, severe cellular damage; 2, moderate cellular damage; 3, mild cellular damage; and 4, normal histologic appearance. A neuropathologist, blinded to the treatment, scored all samples. Although there appeared to be a difference between groups, no statistical significance was noted (Table 2). Overall, however, histopathological score did correlate with clinical neurologic function (48-hour Tarlov score), with a Spearman rank order correlation coefficient of 0.467 (p < 0.013).
|
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsDiscussionReferences |
|---|
The histology of the spinal cords confirms the clinical observations. Whereas all of the thoracic sections were normal, control rabbits with severe neurologic deficits exhibited pathologic processes within the lumbar regions. Qualitatively, evidence of both necrosis and apoptosis (TUNEL) was apparent in these sections. Diazoxide-treated animals, on the other hand, not only exhibited preserved neurologic function but also demonstrated less histologic damage. Although an attempt to quantify the differences seen on histologic exam revealed no statistical significance, the histology does provide anatomic correlation with neurologic function.
Murry and associates first described ischemic preconditioning (IPC) in the myocardium in 1986 [5]. Therefore, most of the studies investigating the mechanism of IPC protection have employed models of myocardial ischemia. ATP-sensitive potassium channels were first implicated in IPC when it was discovered that administration of pharmacological agonists of these channels could mimic and antagonists could abolish the protection seen with brief periods of ischemia [15]. Initially, cell surface channels were thought to be the end effectors, with cardio-protection resulting from action potential shortening [6]. More recently, it has been shown that the degree of protection is independent of action potential shortening [16] and that protection is also seen in unstimulated myocytes, where alteration of the action potential cannot be a factor [17, 18]. Further studies using diazoxide (DZ), a more than 1,000-fold more potent agonist of mitochondrial ATP-sensitive potassium channels (mitoKATP) [19], and the mitoKATP blocker 5-hydroxydecanoic acid (5-HD) have provided evidence that these mitochondrial channels are the end effectors of both ischemic [20, 21, 22] and pharmacological [23, 24] preconditioning.
Although first studied in the myocardium, there has also been a great deal of interest in the study of ischemic preconditioning in the brain [24, 25] and spinal cord [12–14]. Similarly, the effect of diazoxide on the brain has been extensively investigated. To date, however, there have been only a few studies that have addressed the mechanism of IPC in the spinal cord and no research has been published on the effect of diazoxide in a model of spinal cord ischemia. In this study, the use of glibenclamide to abolish the protection seen with diazoxide suggests that the activation of neuronal ATP-sensitive potassium channels plays a central role in diazoxide-induced neuro-protection. However, glibenclamide is not specific for mitochondrial KATP channels and 5-hydroxydecanoic acid (5-HD), the most widely used mitoKATP-specific antagonist, did not inhibit the action of diazoxide in this model. Based on these data, it is impossible to determine if this protective effect is mediated at the level of the cell surface channels (surfaceKATP) or at the level of the mitochondrion (mitoKATP). We believe that the inability of 5-HD to prevent this protection is secondary to its hydrophilic nature. Whereas both diazoxide and glibenclamide are lipophilic agents and are, therefore, more likely to cross the blood-brain barrier, 5-HD is extremely hydrophilic. This may explain its inability to block the effects of diazoxide or ischemic preconditioning in the model, but this certainly requires further investigation.
In summary, administration of diazoxide resulted in significant improvement in neurologic outcome in this model of spinal cord ischemia. This protective effect improved function at both early (24 to 48 hours) and late (7 days) time points. Moreover, the mechanism of protection appears to be related to diazoxide’s ATP-sensitive potassium channel activity, although further research is necessary to determine whether mitochondrial ATP-sensitive potassium channels are the ultimate end effectors of this response. Overall, this study suggests that diazoxide may be useful in the prevention of neurologic injury after thoracic aneurysm surgery.
Diazoxide is currently used clinically, in its parenteral form, for the treatment of hypertensive emergencies. Although not widely used for this indication, secondary to its side effects (sodium and water retention, hyperglycemia), we have submitted a clinical protocol to study the effects of diazoxide in patients undergoing hypothermic circulatory arrest. This trial is based on previously published work from our laboratory demonstrating significant improvement in neurologic function using diazoxide in our canine model of hypothermic circulatory arrest-induced brain injury [26]. Given our findings in this current study, a clinical trial examining diazoxide’s efficacy in the prevention neurologic injury after thoracoabdominal aortic aneurysm repair is also warranted.
|
Acknowledgments |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsDiscussionReferences |
|---|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsDiscussionReferences |
|---|
There are also three kinds of neuronal injury, and you have explained these reasonably well. Obviously, there is the ischemic injury that results in the death of a cell. We all are intimately familiar with that. Also, there is the injury that can occur from reperfusion. And then there is the ischemic injury that is programmed cell death called apoptosis. This process can be understood as a cell having a trigger that will cause it to die later, and we think this contributes to the delayed neurological deterioration seen in these patients on occasion.
There is research ongoing to address all three of these types of injury. Much of our recent work at UVA and your work as well has been looking more and more at ameliorating reperfusion injury and apoptosis. You and your team have addressed the pharmacological treatment thought to increase the tolerance of the neuronal tissue of the cord during the ischemic period, specifically attempting to precondition the tissue, a concept that is familiar to us as cardiac surgeons.
You gave one dose of your agent and one dose of two different agents thought to block the action of the original agent, diazoxide. You also tried to use actual ischemic preconditioning in an attempt to mimic the effects of the diazoxide.
As I saw it, there was some good news and some bad news in this work. The bad news was that the ischemic preconditioning effect did not parallel the diazoxide, only one of your blockers actually blocked the effects of the diazoxide, and the histological differences did not achieve definable significance. I do not think any of this is terribly important to your conclusions. I think the good news is that the drug actually did work, and I think it may prove to be a very useful addition to our armamentarium. This led me to have several questions for you that might allow you to elucidate and cast some light on the work that you did.
I very well understand the time and cost constraints with developing every imaginable drug and time dose-response curve, but I thought you might tell us just briefly why you chose just one time to give the drug and one dose of both the drug and the blocking agents. Furthermore, with regard to the pharmacokinetics of diazoxide itself, I certainly thought that there might have been some effect of the diazoxide not just on the ischemic period itself but during the reperfusion and possibly even on preventing or ameliorating delayed cell death. Do you have evidence or knowledge of an ability of this agent to last that long pharmacologically?
And finally, with regard to the apoptosis or programmed or delayed cell death, did you think you ended up with enough evidence to speculate about diazoxide’s role in preventing that injury? There is obviously some controversy about the TUNEL technique, and in your work, only your diazoxide animals were really followed for the full week after the injury was induced. Did you have other evidence such as elevated caspase levels or DNA repair protein levels that you saw in those groups that might shed some light on that?
In summary, I think we are just beginning to elucidate all the mechanisms and the possible pharmacological preventive measures to deal with all the different aspects of spinal cord ischemia. I certainly think that your work represents an important step in the right direction. Thanks very much for asking me to discuss your work.
DR CAPARRELLI: Thank you, Dr Tribble, for those kind comments and questions. To first address the issue of dose-response, we did not choose to pursue a dose-response curve in this study. We based our dosing of 5 mg/kg IV of diazoxide 15 minutes before cross-clamp and the dosing of the blocking agents on previously published studies in the myocardium, as well as on our experience using these drugs in our canine model of hypothermic circulatory arrest. Further, we have only reported on the successful antagonistic effect of glibenclamide in this presentation. In the manuscript, we also report on a group that received 5-hydroxydecanoic acid (5-HD), a specific mitochondrial ATP-sensitive potassium channel antagonist, before diazoxide administration. Although this is the most widely used antagonist of preconditioning in the myocardium, it was ineffective in our preparation. We believed that this was due to the fact that unlike diazoxide and glibenclamide, which are lipophilic, 5-HD is hydrophilic and may not cross the blood-brain barrier.
It is true that the failure of the ischemic preconditioning regimen to affect a significant protective response is somewhat perplexing. However, examination of the raw data does shed some light on this issue. Of the 6 animals that underwent the ischemic preconditioning protocol, there were 3 animals that demonstrated normal neurological function and 3 animals that were completely paralyzed. This variability seems to suggest that ischemic preconditioning might be effective in this model but is inconsistent. This is in contrast to the diazoxide-treated group, where preservation of neurological function was seen almost uniformly.
Moving to the question of the pharmacokinetics of diazoxide. Again, in the myocardium, there have been several studies investigating the role of ATP-sensitive potassium channels in both early and late-phase preconditioning. These studies demonstrate that pharmacological triggers of ischemic preconditioning, such as adenosine, norepinephrine, and opioids, are able to induce a delayed or late-phase protective effect. Although diazoxide has a long half-life (48 ± 12 hours), preconditioning induced by diazoxide is thought to be due the activation of complex cellular protective mechanisms and not merely because of persistence of the drug in the system. Therefore, there may be a role for the use of this drug both as a bolus before cross-clamp as well as a continuous infusion during reperfusion. This is something that we have experience with, again in our circulatory arrest model, where we administer a bolus of diazoxide before arrest and a continuous infusion during rewarming. Conversely, the existence of this late-phase protective effect suggests that there may be a role for administration of diazoxide several hours or days before a planned aortic surgery. Because diazoxide’s ability to drastically reduce systemic pressure is short-lived, administration (in a monitored setting) before surgery could eliminate this as a factor in the intraoperative setting.
To address the question of apoptosis, there have been several studies demonstrating that apoptosis plays an important role in neuronal death after ischemia. Our histology did not, however, bear that out to the degree that we analyzed it. In our histopathological sections, there appeared to be much more necrotic cell death. Finally, I agree that TUNEL assay is not specific for apoptosis, and we are currently in the process of examining other markers of apoptosis, including caspase activity. This work, however, is still in its preliminary stages. Thank you again for the privilege of presenting our findings today.
|
References |
|---|
|
TopAbstractIntroductionMaterial and methodsResultsCommentAcknowledgmentsDiscussionReferences |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Sirlak, S. Eryilmaz, M. Bahadir Inan, Y. S. Sirin, O. Besalti, L. Yazicioglu, E. Ozcinar, E. Erdemli, R. Tasoz, A. H. Elhan, et al. Effects of carbamazepine on spinal cord ischemia. J. Thorac. Cardiovasc. Surg., October 1, 2008; 136(4): 1038 - 1043.e4. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Roseborough, D. Gao, L. Chen, M. A. Trush, S. Zhou, G. M. Williams, and C. Wei The Mitochondrial K-ATP Channel Opener, Diazoxide, Prevents Ischemia-Reperfusion Injury in the Rabbit Spinal Cord Am. J. Pathol., May 1, 2006; 168(5): 1443 - 1451. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. Shi, T. Kazui, X. Jiang, N. Washiyama, K. Suzuki, K. Yamashita, and H. Terada NS-7, a novel Na+/Ca2+ channel blocker, prevents neurologic injury after spinal cord ischemia in rabbits J. Thorac. Cardiovasc. Surg., February 1, 2005; 129(2): 364 - 371. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I. K. Toumpoulis The role of early ischemic preconditioning in spinal cord protection after transient aortic occlusion Eur. J. Cardiothorac. Surg., November 1, 2003; 24(5): 852 - 852. [Full Text] [PDF]
|
|
|
|
|
|
C. S. Isbir and K. Ak Reply to Toumpoulis Eur. J. Cardiothorac. Surg., November 1, 2003; 24(5): 853 - 853. [Full Text] [PDF]
|
|
|
|
 |
|
 I. K. Toumpoulis and C. E. Anagnostopoulos Early ischemic preconditioning for spinal cord protection Ann. Thorac. Surg., October 1, 2003; 76(4): 1340 - 1341. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ANN THORAC SURG | ASIAN CARDIOVASC THORAC ANN | EUR J CARDIOTHORAC SURG |
| J THORAC CARDIOVASC SURG | ICVTS | ALL CTSNet JOURNALS |
