| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ann Thorac Surg 2005;79:1307-1314
© 2005 The Society of Thoracic Surgeons
a Department of Cardiothoracic Surgery, New York, New York
b Department of Neurosurgery,New York, New York
c Department of Biomathematics, Mount Sinai School of Medicine, New York, New York
Accepted for publication May 12, 2004.
* Address reprint requests to Dr R. B. Griepp, Mount Sinai School of Medicine, Department of Cardiothoracic Surgery, One Gustave L. Levy Place, Box 1028, New York, NY 10029 (E-mail: hagl{at}exch.thg.mh-hannover.de).
 |
Abstract |
|---|
 Top Abstract IntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
METHODS: Twenty-seven pigs were studied, divided between two tasks. 13 underwent 90 minutes HCA at 20°C and were trained from postoperative day (POD) 7; 14 were unoperated controls. The maze includes a holding area, 8 rooms, and a center hallway. One piece of apple is placed in each baited room on each of 10 days of learning evaluation. After a pig enters a room, doors to all other rooms close, and the pig must return to the holding area. In task 1, 6 of 8 rooms were baited, and each day's session ended when each baited room had been entered, or after 20 trials. In task 2, initially only the right- or left-sided rooms were baited. Pigs were evaluated each day until they entered 4 baited rooms, or for 15 trials; the process was then repeated, baiting the other side.
RESULTS: Intraoperative physiology and postoperative recovery showed no differences between task 1 or 2 pigs. Task 1 did not distinguish between control and HCA groups (p = 0.5), but task 2 revealed significantly (p = 0.04) better learning in controls.
CONCLUSIONS: The significantly poorer performance of pigs after HCA suggests that the reversal of baited rooms in task 2 provides the sensitivity to detect cognitive dysfunction. The maze is a promising tool to investigate in pigs the mild cerebral damage often seen after HCA.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
With increasing experience, however, the incidence of devastating neurologic sequelae of HCA has diminished substantially, and the major issue with regard to cerebral protection has become the prevention of subtle cerebral injury. Ergin and coworkers [4] established a generally accepted distinction between permanent neurologic damage–usually due to embolic events or gross malperfusion– and the more subtle type of injury resulting from suboptimal cerebral protection, which he termed "temporary neurologic dysfunction" (TND). This syndrome, which consists of delayed awakening and prolonged confusion, has been demonstrated to correlate with increasing duration of HCA. The manifestations of TND are unassociated with any changes in imaging studies and usually improve within a few days, and were therefore initially assumed to be benign. But further study has shown that TND is associated with long-term deficits in cognitive function, including decreased learning ability, impaired fine motor control and deficiencies in memory [5]. Routine neurologic examinations are not sensitive enough to reveal these relatively subtle changes, but they can be detected using a complex neuropsychological test battery.
In order to study cerebral protection strategies systematically, an animal model large enough to undergo cardiopulmonary bypass (CPB) is required. In recent years, the pig has become the animal model of choice, but there is still very limited experience in evaluating neurologic recovery in the pig. It has gradually become apparent that our initial methods of behavioral evaluation may not be sensitive enough to detect potential neuroprotective effects with regard to the kinds of subtle cognitive dysfunction that is increasingly the focus of clinical concern. To address this problem, we have increased the sensitivity of our histopathological assessments by incorporating immunocytochemical staining for the presence of various kinds of injury predisposing to cell death (to detect apoptosis, for example) [6, 7]. But before considering potential therapies for clinical trials, we think there should be evidence that a particular intervention not only reduces brain damage as measured histologically in an animal model, but also improves functional outcome as assessed neurophysiologically or behaviorally [8]. Tests of cognitive function are well established for small animals [9] and primates, but have been used only rarely in the porcine model.
In developing a sensitive behavioral measure for evaluation of global cerebral injury in the pig model, we have chosen to focus on tasks that require acquisition of new information, since it has been reported [10, 11] that extended periods of ischemia can lead to significant deficits in learning and memory both in animals and in humans. These deficits are in large part due to the high sensitivity of the hippocampus to ischemia [12]. Successful learning of a new task requires that the subject store and then retrieve information with regard to location or other features of specific places or objects.
In this article we report our results in establishing and validating various tasks in a multi-room maze as a means of evaluating cognitive function in pigs. We found that subtle impairment in the ability of pigs to learn can be detected following HCA. This underscores the usefulness of the pig as a clinically relevant model for the study of ways of improving cerebral protection during aortic arch surgery.
|
Material and Methods |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Overall, 27 animals were included in the study. To evaluate the sensitivity of each maze task, healthy unoperated control animals were compared with postoperative pigs that had undergone 90 minutes of HCA at a brain temperature of 20°C. For task 1, there were 7 operated pigs and 7 unoperated controls; for task 2, there were 6 operated pigs and 7 controls. Orientation sessions in operated pigs were not started until POD 7 to assure that the animals had recovered well enough to fulfill the physical requirements for learning in the maze. The evaluation was then performed according to the protocol for task 1 or 2.
All animals received humane care in compliance with the "Principles of Laboratory Animal Care" formulated by the National Society for Medical Research and the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health (National Institutes of Health Publication No. 88 to 23, revised 1996). The protocol for this study was approved by the Mount Sinai Institutional Animal Care and Use Committee.
Maze Apparatus
The testing apparatus was an 8-room maze (Fig 1) housed in a 5.7x3.0 m lighted room on the same floor as the animal colony. The entire apparatus was made of sheets of extruded hollow polypropylene, steel braces, and angle iron.
|
A food tray was mounted on the far wall of each room, and an additional food tray was mounted in the holding area. A fan mounted above the maze at the end opposite the entry/exit door operated continuously. The ambient noise level of the room was approximately 70 dB, and the ambient temperature was approximately 23°C. A custom-made electronic circuit controlled the opening and closing of each door.
Four video cameras were mounted on the ceiling. Each camera was capable of monitoring two adjacent rooms of the maze. Images from all eight rooms were monitored continuously by use of a video observation system (ULTRAK, Lewisville, TX) that mixed the signals from the four video cameras. The output of the video system was saved to videotape by a Daewoo VCR (Model DV-K84N). Figure 2 is a picture of a pig in the maze, taken looking down from the ceiling.
|
Behavioral Evaluation in the Maze
ORIENTATION
Before evaluation of learning of each task, animals were given two orientation sessions, which were not scored, on each of two consecutive days. During these sessions, the animals wandered freely through all rooms of the maze. The lengths of the four sessions were 10 minutes and 5 minutes on the first day, and 5 minutes and 2.5 minutes on the second day. Each animal was returned to its transport cage for 10 minutes between the two sessions on the same day. At the beginning of each orientation session, a piece of apple was placed in each of the food trays in 6 of 8 rooms. The 6 baited rooms included 3 rooms on the right side of the maze and 3 on the left (randomly chosen for each animal). In addition, a piece of apple was placed in a food tray outside the entry/exit door of the maze.
TASK 1: EVALUATION OF ABILITY TO LEARN
Following orientation, a piece of apple was placed in each food tray of an animal's 6 baited rooms (the same as those baited during the orientation sessions). During the learning evaluation phase, animals were permitted to enter only one room on each trial. A room entry was defined as the placement of the animal's front feet beyond a room's passageway (represented in Fig 1 by the dashed lines). When an animal entered a room, all of the doors to the other rooms closed immediately. After the animal exited a room, the door to that room was closed as well. The entry/exit door to the maze was then opened, and the animal exited the maze. Whenever an animal obtained a slice of apple for a correct room entry, it also received an additional piece of apple when it exited the maze. The additional slice of apple was placed in the tray mounted on the outside wall of the holding area near the entry/exit door. Animals remained in the holding area for 30 seconds between trials.
A room entry was scored as correct if it was the first entry of the session into a baited room. An error was defined as an entry into a room that had never been baited, or an entry into a previously baited room from which the food had already been obtained. The number of correct entries and the total number of errors were determined. A percent correct was calculated for each animal on each day of learning evaluation by dividing the number of correct entries by the total number of room entries.
On each day, animals were evaluated until each had entered all of the 6 baited rooms or until 20 trials were completed. To improve their scores, the animals had to learn which rooms were never baited, and remember which rooms they had previously entered. Each animal was evaluated for a total of ten sessions over a period of 12 days.
TASK 2: EVALUATION OF ABILITY TO LEARN
For this task, the 8 rooms were divided into two sets: the 4 rooms on either the left or the right side of the maze (2x4 maze). Animals were assigned randomly to either the left or the right set for the first phase of the evaluation on each day, and then switched to the opposite side for the second phase of each day's evaluation. Each of the 4 rooms in the appropriate set was baited with a slice of apple before the first or second phase of each day's learning evaluation.
On each trial, animals were permitted to enter the maze and select one room. Following entry into a room, all of the doors to the other seven rooms closed. Animals exited the maze after each trial. As in task 1, if an animal entered a baited room for the first time, it obtained a slice of apple in the room and then was given an additional piece of apple in the tray in the holding area. If the pig entered an unbaited room, this was considered an error, and no reinforcement was given. Learning evaluation on each set was terminated if a pig entered all four baited rooms, or if it made 15 room entries, even if one or more of the baited rooms had not been entered.
Following the first phase of a day's learning evaluation, each animal was returned to its transport cage for 5 minutes. During the 5-minute break, apple slices were placed in the trays in the second set of rooms. To improve a score during the second phase of each day's trial, therefore, each pig had to learn that all four apple slices would be on the same side of the maze; to remember which rooms had previously been entered, and to realize that the slices would be on the side opposite from the rooms which had been baited earlier that day. Animals had 10 learning evaluation sessions over a 12-day period.
Experimental Protocol
HCA PROTOCOL
The surgical protocol for HCA has been described in detail previously [13]. Briefly, the animals were connected to CPB, cooled for 45 minutes (using 
-stat principles) to reach the target brain temperature, and the pump was then stopped for 90 minutes. Methylprednisolone was given before HCA, and mannitol was administered during reperfusion. After HCA, the animals were rewarmed over a period of 60 minutes. Hemodynamic and neurophysiological data were recorded throughout cooling, HCA and rewarming. After the chest was closed (3 hours after the start of rewarming), anesthesia was stopped to allow spontaneous breathing through the endotracheal tube. If breathing was adequate and the animals regained consciousness, they were extubated and brought into the recovery area.
EARLY BEHAVIORAL RECOVERY
From POD 1 to POD 5, all animals were scored on a gross behavioral scale that evaluates mental status, appetite, and gait by a physician blinded to the experimental protocol, as described in earlier reports [13]. Daily scores were determined after several time-standardized inspections in the morning, at noon and in the evening. A score of 9 is normal, and 0 indicates coma or death.
Statistics
Analysis of variance with repeated measures (SAS) and paired t-tests were performed to compare the groups, and to test the effect of repeated exposure to the maze task on the percent of trials during which the animals entered a baited room (percent correct).
|
Results |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
|
|
TASK 1: 6 OF 8 MAZE
In Figure 4 the mean values of percent of correct entries for each of the 10 learning evaluation sessions are depicted. The horizontal line at 30.1% represents the percent of correct entries that would be expected if animals had made their room choices at random. As can be seen, the percent correct on days 1 and 2 were near the random level in both groups. From day 3 to 7, however, the percent correct increased steadily. Over days 7 to 10, the mean percent correct was 64.7% in the controls, and 63% in the operated group: this difference is not significant.
|
TASK 2: 2x4 MAZE
As in task 1, the percent correct was calculated for each animal for each day of learning evaluation. Figure 5 illustrates the percent of correct choices for the animals on each of the 10 days scored. The horizontal line at 24.1% shows the percent correct if animals had selected rooms at random. Performance on day 1 was close to this random level in both groups.
|
|
Comment |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Despite these and other gains in understanding the pathophysiology of cerebral injury following HCA, we felt the need to find another tool to detect the kind of subtle cognitive dysfunction associated with TND [4], which occurs in a disturbingly high percentage of patients who require prolonged cerebral protection during repair of complex aortic arch pathology. Since the incidence of TND reveals a clear correlation clinically with the duration of HCA [3, 16], TND serves as a sensitive marker of imperfect cerebral protection. Clinically, systematic testing of neuropsychological function can most sensitively identify the individuals with symptoms associated with TND. For this reason, we felt that an approximation of the clinical situation in an animal model could be achieved by testing learning ability in pigs. Since experience in the porcine model is limited, we had to begin by showing that pigs are able to acquire new information [17]. The next step was to show that our learning task was sensitive enough to distinguish between different levels of cognitive function. We present here our first attempts to demonstrate – by comparing healthy unoperated animals with a group which underwent HCA – that a subtle impairment in learning occurs after an HCA protocol similar to that used clinically, and that it can be detected by a maze task.
The rapid improvement in the animals' scores on both the 6 of 8 and 2x4 maze tasks indicates that pigs, like many other foraging animals, can quickly acquire a high level of accuracy on spatial tasks with food reinforcement. Within a few sessions, healthy animals increased their successful choices in both tasks from a random level (24% to 30% correct) to a level greater than 65%. The rapid learning of these tasks is consistent with the results of a recent report by Laughlin and Mendl [18], who trained pigs using an 8-arm radial arm maze using food for reinforcement.
Several factors influenced the development of the maze tasks described here, including the physical constraints placed on the task by the size of the pig and the dimensions of our testing room. The cost of housing the pigs and their rapid weight gain forced us to develop a task that could be completed within 3 weeks, including a minimum of 7 days for recovery from surgery. Our task also needed to have a level of difficulty that was sufficiently high to detect subtle differences in learning following HCA. Tasks can be made more sensitive to cognitive dysfunction either by increasing the amount of information that must be retained, or by lengthening the retention interval [19].
Our first attempt to train pigs involved a T-maze because it requires a simple apparatus and has been used to demonstrate ischemia-induced deficits in other animal models [20, 21]. In pilot experiments, we required animals to retain a memory for the arm in which they received a piece of apple. On the next trial, the animals received an apple if they entered the opposite arm. As expected, animals learned this task very quickly, and a delay was needed to increase the difficulty of the task. However, when we attempted to interpose a lengthy delay between trials, problems were encountered because the animals would not wait passively. We therefore decided to make the task more difficult by increasing the number of items that needed to be stored in memory, which also had the effect of lengthening the retention interval. Designing a multi-choice maze seemed the optimal solution.
In the present study, we were able to demonstrate that a multi-room maze can be used as a tool to detect cognitive dysfunction in pigs. Surgery itself appeared to produce no lasting adverse impact on the ability of the animals to move and run the maze. On POD 7, all animals had recovered fully, as indicated by a neurobehavioral score of 9 (normal). Moreover, there were no differences in the number of room entries during the orientation sessions between unoperated and postoperative animals. The fact that task 1 could not differentiate between the healthy and ischemic animals is also a clear indicator that possible physical limitations in the postoperative animals had no significant impact on the outcome of the maze tasks.
Why task 1 was not sensitive enough to detect ischemic injury after 90 minutes of HCA at 20°C remains an unanswered question. It is our assumption that animals used their working memories in choosing which rooms to enter and which to avoid. However, there are alternative explanations, including the use of odor cues to determine where they had been in that session, which might represent hard-wired innate responses rather than learned behavior. Various strategies, including the use of a fan and the placement of apple slices strategically around the perimeter of the maze, were used to minimize any odor cues.
The 6 of 8 maze task and the 2x4 maze task are similar in that both require working memory to store the locations of previously entered rooms. The tasks differ in their reliance on what has been termed reference memory: the memory for the location of rooms or arms that are never baited in multi-choice spatial tasks, which includes two of the rooms in the 6/8 maze. Since the few animals trained when all eight rooms were baited learned as quickly as those with only six rooms baited (data not shown), reference memory does not seem likely to be very important in this model.
In the 2x4 maze, all rooms were baited each day, but in subsets of 4 rooms. The need to remember that all rooms on the same side were either baited or unbaited, followed by the reversal of baited and unbaited rooms during the second phase of each trial, seems to have placed a greater burden on working memory than did the simpler task of remembering which of the 8 rooms had previously been entered. In addition, animals performing the 2x4 maze task had to wait 5 minutes between the first and the second phase of training, increasing the duration of the task's requirement for working memory. Thus, the second task included several aspects which enhanced its demands on working memory, and therefore succeeded in detecting impaired cognitive function in postoperative animals that were able to learn the first task without any difficulty.
Because HCA in the present study was carried out using cooling and rewarming on CPB in anesthetized animals, it is not possible to rule out the possibility that anesthesia or CPB contributed to the cognitive dysfunction demonstrated: an experiment with anesthesia controls and anesthesia + CPB controls would be required. Since HCA as used clinically in aortic surgery (and surgery for congenital heart disease) cannot be carried out in the absence of anesthesia and CPB, such trials do not seem to justify the time and expense which would have been required to clarify this issue. A significant role for CPB and anesthesia in contributing to cognitive dysfunction after HCA is also not consistent with repeated clinical observations that TND correlates with the duration of HCA.
Conclusions
In conclusion, the two maze tasks described herein have the potential for use as tools for the behavioral assessment of pigs, and presumably of other animals of similar size. It seems that a complex learning task is required in order to detect mild HCA-induced cerebral damage in pigs. The significantly poorer performance of pigs after HCA in task II suggests that the reversal of baited rooms provided the necessary sensitivity to detect subtle cognitive dysfunction. The maze looks promising as a sophisticated tool to explore the pathophysiology of the mild cerebral injury often seen clinically after HCA. Although the sensitivity of such tasks in discriminating between small differences in the degree of cerebral damage remains to be determined in future studies, the current study establishes the pig as a promising model in which to study how to improve cerebral protection and preserve intellectual capacity during aortic surgery.
|
Acknowledgments |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
|
References |
|---|
|
TopAbstractIntroductionMaterial and MethodsResultsCommentAcknowledgmentsReferences |
|---|
Related Article
This article has been cited by other articles:
|
|
 |
|
  N. Khaladj, S. Peterss, P. Oetjen, R. von Wasielewski, G. Hauschild, M. Karck, A. Haverich, and C. Hagl Hypothermic circulatory arrest with moderate, deep or profound hypothermic selective antegrade cerebral perfusion: which temperature provides best brain protection? Eur J Cardiothorac Surg, September 1, 2006; 30(3): 492 - 498. [Abstract] [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 |
= task 1; 
= task 2. (POD = postoperative day.)
