151
BehaviouralBrainResearch, 41 (1990) 151-160 Elsevier BBR 01119
The effects of forebrain ischaemia on spatial learning J.J. Hagan and M. Beaughard* CNS Pharmacology Labs, Organon International B. V., Oss (The Netherlands) (Received 19 December 1989) (Revised version received 26 June 1990) (Accepted 28 August 1990)
Key words: Forebrain; Ischaemia; Place navigation; T-maze; Hippocampus; Rat
Rats were subjected to 15 rain of forebrain ischaemia using the 4-vessel occlusion method. Following recovery they were trained using place navigation learning in a Morris water maze and forced choice rewarded alternation in a T-maze. Ischaemic rats were impaired in place navigation learning but the deficit was transient and there was no impairment of subsequent transfer test performance. Food-rewarded forced-choice alternation in a T-maze revealed a persistent impairment in ischaemic rats. The behavioural deficits were associated with neuropathological damage in the CA 1cell layer of the dorsal hippocampus with varying degrees of damage in layers CA2, CA3 and C A 4. Granule cells in the dentate gyrus were not visibly affected. Variable amounts of lesion damage were found in the dorsolateral striatum. Ischaemic rats are therefore impaired on both place navigation and forced choice rewarded alternation, suggesting that ischaemic brain damage affects reference and working memory processes to different extents. Forced choice alternation may be the more sensitive method of assessing cognitive changes caused by forebrain ischaemia.
INTRODUCTION
Survivors of cardiac arrest, or cerebral ischaemia due to other causes, often suffer anterograde amnesia whilst other cognitive and linguistic functions remain unimpairedS,25-27,30,31. Autopsy studies reveal that the CA~ cell layer of the hippocampus is strikingly vulnerable to ischaemia and in some cases CA~ cell death has been associated with memory loss in the near absence of other neurological or neuropathological signs 5A3, suggesting that it may be sufficient to cause the memory loss. Other species show the same regional vulnerability to cell loss following brief periods of ischaemia and reperf u s i o n 19-21'23. Cell death is thought to be caused by excitatory amino acids which trigger
neurotoxicity by inducing calcium overload. In support of this argument extracellular levels of glutamate and aspartate have been shown to rise dramatically during ischaemia 2 and disruption of the major excitatory input to the hippocampus, either by entorhinal cortex lesions 14 or granule cell destruction ~3 protects the CA1 cells from ischaemic damage. The short-term neurological and behavioural consequences of ischaemia subside rapidly4"9 to reveal persistent learning and memory deficits. Preoperatively trained rats are impaired in the performance of radial maze 6 and T-maze t a s k s 19"29. Rats trained after the ischaemic event are impaired in both radial maze 7 and a learning set place navigation task but not simple place navigation learning 1. Several authors suggest that
*Present address: RL-CERM, Route de Marsat, 63203 Riom Cedex, France. Correspondence: J.J. Hagan, Present address: SmithKline Beecham, Coldharbour Road, The Pinnacles, Harlow, Essex CM19 5AD, U.K. 0166-4328/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
152 following ischaemia reference, or trial-independent memory, is better preserved than working or t r i a l - d e p e n d e n t m e m o r y 6,7,28,29. In s o m e studies 29 this claim is well founded, although in others the suggestion is based on statistical trends 7'28. Furthermore, the conclusion is largely derived from experiments in which effects on different memory systems are inferred from different aspects of a single task. Finally, comparisons between experiments are confounded by marked variations in the extent of lesion damage. In the present experiments we examined the effects of transient forebrain ischaemia on two independent tasks within the same group of ischaemic rats. These were acquisition of place navigation learning in a watermaze 16"~7 and forced choice alternation in a T-maze s'=. Proprioceptive or kinaesthetic cues partly contribute to successful solution of the T-maze but both tasks depend largely for their solution on the use of spatial or extra-maze cues and both are thought to require intact septo-hippocampal systems s'1°'17'22. In the water maze the invariant location of the escape platform characterises it as a reference memory task. In the T-maze correct solution on the choice trial is dependent upon the information given on the immediately preceding forced choice run and it therefore constitutes a working memory task. MATERIALS AND METHODS
Naive male Wistar rats (Charles River) weighing 300-350 g, were housed in a temperature(22°C) and humidity-controlled environment under a 12-h light/dark cycle (lights on at 07.00 h). Rats were group-housed, 4 or 5 to a cage, during place navigation testing and had ad libitum access to food and water. During forced choice alternation training they were singly housed with ad libitum access to water. An individually adjusted amount of normal lab food (approximately 15 g) was provided at the same time each day (16.30 h) about one hour after the end of training for the last block of animals. Cerebral ischaemia was produced by the 4-vessel occlusion technique 2°'21. The rats were anaesthetised with halothane ( 2 ~ ) in a 3:1,
N 2 0 : O 2 gas mixture and the vertebral arteries were electrocauterised at the first cervical vertebra. Four days later both common carotid arteries were isolated under halothane anaesthesia and atraumatic clips were placed around each artery. Twenty-four hours later the carotid clips were tightened for 15 rain to produce a transient cerebral ischaemia. The loss of righting reflex during this period was used as a selection criterion and animals which convulsed during the ischaemia or the reperfusion period were excluded from the study. During and after ischaemia (3 h) all animals were placed under the warming light to prevent hypothermia due to their comatose state. Sham-operated controls underwent the two periods of anaesthesia without cauterization of the cervical vertebra or ligation of the common carotid arteries. Other experiments have reported no difference between non-operated controls and rats given vertebral artery cauterization alone 28. Thirteen to twenty days were allowed for recovery between surgery and the start of place navigation testing. Place navigation training was conducted in a Morris water maze. Briefly, rats were trained over 4 sessions to find the spatial location of a hidden platform (11 cm diameter) submerged 1 cm below the water surface in a circular black pool (2.1 m diameter) filled to a depth of 24.5 cm with water. Temperature was maintained at 26 _+ 1 °CI6"17. Both the pool and the platform were constructed of grey/black polyvinyl. This, plus the fact that the platform stood 1 cm below the surface, rendered the platform invisible at water level. The pool was housed in a room which was painted black and illumination was provided by several lamps placed around the perimeter. Conspicuous cues, (wall plates, room door, ceiling frame, camera etc.) were provided around the pool. Behaviour was monitored via an overhead video camera and the rat's position was computed from the video image and stored, when required, on disc. Each rat was given 120 s of adaptation to the pool prior to training. During training a trial began when the rat, held facing the side wall, was immersed in the water. Latency to escape onto the hidden platform was recorded with a stopwatch.
153 If a rat failed to locate the platform within 120 s it was placed on, or guided to it. The rat remained on the platform for 30 s, was removed and returned to a holding cage to await the next trial. Four trials were given on day one followed by 8, 8 and 4 trials on subsequent days. Rats were trained in squads of 4 or 5 to one of four pool locations (NE, NW, SE, SW). The platform remained in its fixed location throughout training but the starting position on each trial was randomly selected from one of 4 points - North, South, East or West. These points defined four 90 ° quadrants on the pool surface (NE, NW, SE, SW). The platform occupied a position in the quadrant which was approximately midway between the pool centre and the side. Immediately after the last training trial the platform was removed and swim paths were recorded during a 60-s transfer test. Two measures of spatial bias were calculated from the data (1) time spent in the training quadrant and (2) the number of occasions on which each rat crossed the previous location of the platform (annulus entries). T-maze training started at the end of testing in the water maze. Prior to the start of training rats were reduced to 85 ~o of their free feeding weight and stabilised at this level for one week. Forced choice alternation training was carried out using a T-maze constructed of dark grey polyvinyl. The start area (21 cm x 13 cm) was separated from the run arm (1 m x 13 cm) by a vertically sliding door. At the end of the run arm sliding doors opened into the choice arms (58 x 13 cm), at the end of which food pellets (Noyes 37 rag) were provided in a 2-cm deep food well. The maze was surrounded by a 2.5-cm lip, stood 20 cm above table height and was open to a variety of room cues. Low level room lighting provided illumination. On the first day of pretraining rats were habituated to the apparatus for 5 min with all doors open and food pellets (Noyes 37 mg) scattered throughout the maze. In two sessions, which were run on the following day, pellets were restricted to the choice arms and then finally to the food cups. This was followed by a forced choice trial, in which one arm was closed and the open arm was baited. The rat was placed in the start area with
the door open and it was allowed to traverse the maze and eat the pellet in the open arm. In the choice trial, which followed immediately, both arms were open and the previously closed arm was baited. A correct choice was recorded if the rat chose the baited arm. The time from the start of the trial to reaching the food well was recorded as the choice latency. Animals were trained in groups of 3 or 4. Each forced choice trial was followed immediately by a choice trial and the animal was then returned to a holding cage where it remained whilst the other animals in the group completed their trials. Forced choices were allocated left or right on a quasi-random basis. For the first 4 training days 5 pairs of forced-choice and free-choice trials were given, this was then increased to 10 pairs of trials/day. Training continued for a total of 70 trials. Two blocks of delay trials were then given on consecutive days. Each rat was placed in a separate holding cage for either 30 s or 60 s between each forced-choice trial and choice trial. Both place navigation and forcedchoice alternation training were carried out with the experimenter blind to the treatment conditions. At the end of behavioural testing (10 weeks after surgery) rats were deeply anaesthetised with an overdose of Nembutal (60mg/rat). The ascending aorta was cannulated and perfused with phosphate-buffered saline for 60 s followed by FAM fixative [formaldehyde, 40~o glacial % Observations ~00STRIATUM
0
CA 1
,. I ' - - I
CA2/CA3/CA 4
r-
% Neuronal d a m a g e
Fig. 1. The frequency with which different grades of neuropathological damage were found in the CA~, CA2, CA3, C A 4 and dorsolateral striatum ofischaemic rats. Data are based on separate evaluation of left and right hemipheres. In control rats (not shown) 100% of animals fall in category 0 for all regions.
154
Fig. 2. Coronal section through a control (A) and ischaemic (B) dorsal hippocampus showing the extensive loss of CA~ cells. CA1 cell loss in this section is evaluated as grade 5 and the lesser damage evident in CA2, CA3 and C A 4 is scored as grade 1. In (C) grade 5 damage is found in all cell fields but note that both upper and lower blades of the dentate gyrus appear normal ( x 20). (D) shows the dorsolateral striatum from a control rat and contrasts with (E) in which ischaemia caused extensive damage with a high density of small acidophillic remnants. Corpus callosum is visible in the top left corner of both sections ( × 43).
155
156
acetic acid and methanol (99~o) in the ratio 1/1/8]. After fixation the brains were removed and stored in fixative until they were embedded in paraffin for sectioning. Brains sections (10/~m) were stained with Luxol fast blue and Cresyl violet and examined under a light microscope. Previous studies reported a preponderance of damage to the dorsolateral striatum, the CA 1 region and the CA 2, C A 3 and CA4 cells fields of the hippocampus. Assessment was therefore restricted to these areas and performed according to a method described previously 2. Morphological damage in the dorsolateral part of the striatum was graded using a scale of 0 to 3, with 0 representing normal tissue; 1, a few neurons damaged (
BehaviouralBrainResearch, 41 (1990) 151-160 Elsevier BBR 01119
The effects of forebrain ischaemia on spatial learning J.J. Hagan and M. Beaughard* CNS Pharmacology Labs, Organon International B. V., Oss (The Netherlands) (Received 19 December 1989) (Revised version received 26 June 1990) (Accepted 28 August 1990)
Key words: Forebrain; Ischaemia; Place navigation; T-maze; Hippocampus; Rat
Rats were subjected to 15 rain of forebrain ischaemia using the 4-vessel occlusion method. Following recovery they were trained using place navigation learning in a Morris water maze and forced choice rewarded alternation in a T-maze. Ischaemic rats were impaired in place navigation learning but the deficit was transient and there was no impairment of subsequent transfer test performance. Food-rewarded forced-choice alternation in a T-maze revealed a persistent impairment in ischaemic rats. The behavioural deficits were associated with neuropathological damage in the CA 1cell layer of the dorsal hippocampus with varying degrees of damage in layers CA2, CA3 and C A 4. Granule cells in the dentate gyrus were not visibly affected. Variable amounts of lesion damage were found in the dorsolateral striatum. Ischaemic rats are therefore impaired on both place navigation and forced choice rewarded alternation, suggesting that ischaemic brain damage affects reference and working memory processes to different extents. Forced choice alternation may be the more sensitive method of assessing cognitive changes caused by forebrain ischaemia.
INTRODUCTION
Survivors of cardiac arrest, or cerebral ischaemia due to other causes, often suffer anterograde amnesia whilst other cognitive and linguistic functions remain unimpairedS,25-27,30,31. Autopsy studies reveal that the CA~ cell layer of the hippocampus is strikingly vulnerable to ischaemia and in some cases CA~ cell death has been associated with memory loss in the near absence of other neurological or neuropathological signs 5A3, suggesting that it may be sufficient to cause the memory loss. Other species show the same regional vulnerability to cell loss following brief periods of ischaemia and reperf u s i o n 19-21'23. Cell death is thought to be caused by excitatory amino acids which trigger
neurotoxicity by inducing calcium overload. In support of this argument extracellular levels of glutamate and aspartate have been shown to rise dramatically during ischaemia 2 and disruption of the major excitatory input to the hippocampus, either by entorhinal cortex lesions 14 or granule cell destruction ~3 protects the CA1 cells from ischaemic damage. The short-term neurological and behavioural consequences of ischaemia subside rapidly4"9 to reveal persistent learning and memory deficits. Preoperatively trained rats are impaired in the performance of radial maze 6 and T-maze t a s k s 19"29. Rats trained after the ischaemic event are impaired in both radial maze 7 and a learning set place navigation task but not simple place navigation learning 1. Several authors suggest that
*Present address: RL-CERM, Route de Marsat, 63203 Riom Cedex, France. Correspondence: J.J. Hagan, Present address: SmithKline Beecham, Coldharbour Road, The Pinnacles, Harlow, Essex CM19 5AD, U.K. 0166-4328/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
152 following ischaemia reference, or trial-independent memory, is better preserved than working or t r i a l - d e p e n d e n t m e m o r y 6,7,28,29. In s o m e studies 29 this claim is well founded, although in others the suggestion is based on statistical trends 7'28. Furthermore, the conclusion is largely derived from experiments in which effects on different memory systems are inferred from different aspects of a single task. Finally, comparisons between experiments are confounded by marked variations in the extent of lesion damage. In the present experiments we examined the effects of transient forebrain ischaemia on two independent tasks within the same group of ischaemic rats. These were acquisition of place navigation learning in a watermaze 16"~7 and forced choice alternation in a T-maze s'=. Proprioceptive or kinaesthetic cues partly contribute to successful solution of the T-maze but both tasks depend largely for their solution on the use of spatial or extra-maze cues and both are thought to require intact septo-hippocampal systems s'1°'17'22. In the water maze the invariant location of the escape platform characterises it as a reference memory task. In the T-maze correct solution on the choice trial is dependent upon the information given on the immediately preceding forced choice run and it therefore constitutes a working memory task. MATERIALS AND METHODS
Naive male Wistar rats (Charles River) weighing 300-350 g, were housed in a temperature(22°C) and humidity-controlled environment under a 12-h light/dark cycle (lights on at 07.00 h). Rats were group-housed, 4 or 5 to a cage, during place navigation testing and had ad libitum access to food and water. During forced choice alternation training they were singly housed with ad libitum access to water. An individually adjusted amount of normal lab food (approximately 15 g) was provided at the same time each day (16.30 h) about one hour after the end of training for the last block of animals. Cerebral ischaemia was produced by the 4-vessel occlusion technique 2°'21. The rats were anaesthetised with halothane ( 2 ~ ) in a 3:1,
N 2 0 : O 2 gas mixture and the vertebral arteries were electrocauterised at the first cervical vertebra. Four days later both common carotid arteries were isolated under halothane anaesthesia and atraumatic clips were placed around each artery. Twenty-four hours later the carotid clips were tightened for 15 rain to produce a transient cerebral ischaemia. The loss of righting reflex during this period was used as a selection criterion and animals which convulsed during the ischaemia or the reperfusion period were excluded from the study. During and after ischaemia (3 h) all animals were placed under the warming light to prevent hypothermia due to their comatose state. Sham-operated controls underwent the two periods of anaesthesia without cauterization of the cervical vertebra or ligation of the common carotid arteries. Other experiments have reported no difference between non-operated controls and rats given vertebral artery cauterization alone 28. Thirteen to twenty days were allowed for recovery between surgery and the start of place navigation testing. Place navigation training was conducted in a Morris water maze. Briefly, rats were trained over 4 sessions to find the spatial location of a hidden platform (11 cm diameter) submerged 1 cm below the water surface in a circular black pool (2.1 m diameter) filled to a depth of 24.5 cm with water. Temperature was maintained at 26 _+ 1 °CI6"17. Both the pool and the platform were constructed of grey/black polyvinyl. This, plus the fact that the platform stood 1 cm below the surface, rendered the platform invisible at water level. The pool was housed in a room which was painted black and illumination was provided by several lamps placed around the perimeter. Conspicuous cues, (wall plates, room door, ceiling frame, camera etc.) were provided around the pool. Behaviour was monitored via an overhead video camera and the rat's position was computed from the video image and stored, when required, on disc. Each rat was given 120 s of adaptation to the pool prior to training. During training a trial began when the rat, held facing the side wall, was immersed in the water. Latency to escape onto the hidden platform was recorded with a stopwatch.
153 If a rat failed to locate the platform within 120 s it was placed on, or guided to it. The rat remained on the platform for 30 s, was removed and returned to a holding cage to await the next trial. Four trials were given on day one followed by 8, 8 and 4 trials on subsequent days. Rats were trained in squads of 4 or 5 to one of four pool locations (NE, NW, SE, SW). The platform remained in its fixed location throughout training but the starting position on each trial was randomly selected from one of 4 points - North, South, East or West. These points defined four 90 ° quadrants on the pool surface (NE, NW, SE, SW). The platform occupied a position in the quadrant which was approximately midway between the pool centre and the side. Immediately after the last training trial the platform was removed and swim paths were recorded during a 60-s transfer test. Two measures of spatial bias were calculated from the data (1) time spent in the training quadrant and (2) the number of occasions on which each rat crossed the previous location of the platform (annulus entries). T-maze training started at the end of testing in the water maze. Prior to the start of training rats were reduced to 85 ~o of their free feeding weight and stabilised at this level for one week. Forced choice alternation training was carried out using a T-maze constructed of dark grey polyvinyl. The start area (21 cm x 13 cm) was separated from the run arm (1 m x 13 cm) by a vertically sliding door. At the end of the run arm sliding doors opened into the choice arms (58 x 13 cm), at the end of which food pellets (Noyes 37 rag) were provided in a 2-cm deep food well. The maze was surrounded by a 2.5-cm lip, stood 20 cm above table height and was open to a variety of room cues. Low level room lighting provided illumination. On the first day of pretraining rats were habituated to the apparatus for 5 min with all doors open and food pellets (Noyes 37 mg) scattered throughout the maze. In two sessions, which were run on the following day, pellets were restricted to the choice arms and then finally to the food cups. This was followed by a forced choice trial, in which one arm was closed and the open arm was baited. The rat was placed in the start area with
the door open and it was allowed to traverse the maze and eat the pellet in the open arm. In the choice trial, which followed immediately, both arms were open and the previously closed arm was baited. A correct choice was recorded if the rat chose the baited arm. The time from the start of the trial to reaching the food well was recorded as the choice latency. Animals were trained in groups of 3 or 4. Each forced choice trial was followed immediately by a choice trial and the animal was then returned to a holding cage where it remained whilst the other animals in the group completed their trials. Forced choices were allocated left or right on a quasi-random basis. For the first 4 training days 5 pairs of forced-choice and free-choice trials were given, this was then increased to 10 pairs of trials/day. Training continued for a total of 70 trials. Two blocks of delay trials were then given on consecutive days. Each rat was placed in a separate holding cage for either 30 s or 60 s between each forced-choice trial and choice trial. Both place navigation and forcedchoice alternation training were carried out with the experimenter blind to the treatment conditions. At the end of behavioural testing (10 weeks after surgery) rats were deeply anaesthetised with an overdose of Nembutal (60mg/rat). The ascending aorta was cannulated and perfused with phosphate-buffered saline for 60 s followed by FAM fixative [formaldehyde, 40~o glacial % Observations ~00STRIATUM
0
CA 1
,. I ' - - I
CA2/CA3/CA 4
r-
% Neuronal d a m a g e
Fig. 1. The frequency with which different grades of neuropathological damage were found in the CA~, CA2, CA3, C A 4 and dorsolateral striatum ofischaemic rats. Data are based on separate evaluation of left and right hemipheres. In control rats (not shown) 100% of animals fall in category 0 for all regions.
154
Fig. 2. Coronal section through a control (A) and ischaemic (B) dorsal hippocampus showing the extensive loss of CA~ cells. CA1 cell loss in this section is evaluated as grade 5 and the lesser damage evident in CA2, CA3 and C A 4 is scored as grade 1. In (C) grade 5 damage is found in all cell fields but note that both upper and lower blades of the dentate gyrus appear normal ( x 20). (D) shows the dorsolateral striatum from a control rat and contrasts with (E) in which ischaemia caused extensive damage with a high density of small acidophillic remnants. Corpus callosum is visible in the top left corner of both sections ( × 43).
155
156
acetic acid and methanol (99~o) in the ratio 1/1/8]. After fixation the brains were removed and stored in fixative until they were embedded in paraffin for sectioning. Brains sections (10/~m) were stained with Luxol fast blue and Cresyl violet and examined under a light microscope. Previous studies reported a preponderance of damage to the dorsolateral striatum, the CA 1 region and the CA 2, C A 3 and CA4 cells fields of the hippocampus. Assessment was therefore restricted to these areas and performed according to a method described previously 2. Morphological damage in the dorsolateral part of the striatum was graded using a scale of 0 to 3, with 0 representing normal tissue; 1, a few neurons damaged (
