- Email: [email protected]
NMDA systems in the amygdala and piriform cortex and nicotinic effects on memory function
Cognitive Brain Research 17 (2003) 475–483 www.elsevier.com / locate / cogbrainres
Research report
NMDA systems in the amygdala and piriform cortex and nicotinic effects on memory function Helen May-Simera, Edward D. Levin* Neurobehavioral Research Laboratory, Department of Psychiatry, Box [3412, Duke University Medical Center, Durham, NC 27710, USA Accepted 29 April 2003
Abstract Both nicotinic cholinergic and NMDA glutaminergic systems are important for memory function. Nicotine has been found repeatedly to significantly improve working memory performance in the radial-arm maze. The NMDA antagonist dizocilpine has been found to impair working memory performance. There is neuropharmacological evidence that these two systems are functionally related. Nicotine is potent at releasing many transmitters including glutamate. The current study was conducted to examine the interaction of nicotinic and NMDA systems within the amygdala with regard to working and reference memory. Rats were trained on a working / reference procedure on a 16-arm radial maze. After acquisition, local infusion cannulae were implanted bilaterally into the amygdala and piriform cortex using stereotaxic techniques. Then 20 min prior to running the rats on the radial-arm maze, they were injected subcutaneously with (2) nicotine ditartrate at doses of 0 and 0.4 mg / kg. Following this, the rats received local infusions of (1) dizocilpine maleate (MK-801) at doses of 0, 2, 6 and 18 mg per side into the lateral amygdala or piriform cortex 10 min prior to running on the radial-arm maze. Each of the eight nicotine and dizocilpine combinations was administered to each rat in a counterbalanced order. After completion of the drug sessions the rats were sacrificed, and using histological methods the cannulae placements were verified. Acute amygdalar infusions of the NMDA glutamate receptor antagonist dizocilpine induced dose-related working and reference memory deficits in the radial-arm maze. Systemic nicotine was not seen to reverse these effects. Dizocilpine infusions into the adjacent piriform cortex did not impair memory function, supporting the specificity of dizocilpine effects in the amygdala. Latency effects were seen with both drugs in both areas. Latencies were decreased with both systemic nicotine and dizocilpine in both the lateral amygdala and the piriform cortex. This study demonstrated the importance of NMDA glutamate systems in the amygdala for appetitively-motivated spatial memory performance. 2003 Elsevier B.V. All rights reserved. Theme: Neural basis of behavior Topic: Learning and memory: systems and functions–animals Keywords: Nicotine; Dizocilpine; MK-801; Memory; Amygdala; Piriform cortex; Radial-arm maze
1. Introduction Numerous studies have shown that both N-methyl-Daspartate (NMDA) glutaminergic and nicotinic cholinergic receptor systems are important for memory function (for reviews see Refs. [7,10,11]). We have found that nicotinic– NMDA interactions are important inasmuch as systemic nicotine administration can substantially attenuate the memory impairment caused by systemic administration of the NMDA antagonist dizocilpine (MK-801) [16]. This *Corresponding author. Tel.: 11-919-681-6273; fax: 11-919-6813416. E-mail address: [email protected] (E.D. Levin). 0926-6410 / 03 / $ – see front matter 2003 Elsevier B.V. All rights reserved. doi:10.1016 / S0926-6410(03)00163-0
nicotine-induced reversal of the memory impairment caused by dizocilpine may be related to the fact that nicotine stimulates glutamate release [23,41]. The anatomic substrate of this interaction is not yet known. Determining the anatomic sites for nicotinic–NMDA interactions and memory function will help advance our understanding of the complex neural basis of cognitive function and will help in the development of novel avenues of treatment for cognitive dysfunction. The NMDA antagonist dizocilpine causes substantial impairment of spatial learning and memory [36,37]. Cole et al. [2] found that both 0.1 and 0.2 mg / kg of dizocilpine caused significant deficits in an operant delayed matching to position task in a delay-dependent manner. Ward et al.
476
H. May-Simera, E.D. Levin / Cognitive Brain Research 17 (2003) 475–483
[39] showed that dizocilpine can produce a dose-dependent disruption of radial-arm maze memory performance. Acquisition of working and reference memory performance on an eight-arm radial maze was impaired by 0.0625 mg / kg of dizocilpine [37], but retention of performance was not affected. We have found working and reference memory impairments in the 16-arm radial maze caused by dizocilpine at doses as low as 25 mg / kg [16]. NMDA systems in the amygdala are important components for glutamate involvement in cognitive function. Pre-training intra-amygdala infusion of AP5 or AP7, other NMDA receptor antagonists, hinders retention of a condition startle response [26]. The immediate post-training intra-amygdala infusion of AP5 hinders retention of stepthrough inhibitory avoidance [19], of a mixture of active and inhibitory avoidance called continuous avoidance [9], and of a step-down inhibitory avoidance and a water finding task. Nicotinic agonists have been found in rodent and nonhuman primate studies to improve performance on a variety of memory tasks. The 16-arm radial maze has been used to show the relative specificity of nicotine-induced improvement in working, but not reference, memory [14,15]. In a complementary fashion, nicotinic antagonists such as mecamylamine have been shown to impair working memory function. A variety of studies have shown that acute treatment with nicotine or nicotinic agonists can improve working memory function in the radial-arm maze in rats [3,13,15]. An acute dose of 0.2 mg / kg nicotine injected subcutaneously 20 min before testing significantly improved working memory with the typical inverted Ushaped function for memory-enhancing drugs. This effect was specific to working memory (changing memory specific to each session). Reference memory (unchanging memory constant across sessions) errors were not significantly affected. It has also been found that acute treatment with other selective nicotinic agonists significantly improves memory performance [12]. Nicotine has been shown to attenuate the amnestic effects of dizocilpine. Dizocilpine (0.1 mg / kg) caused substantial deficits in both working and reference memory on the 16-arm radial maze [16]. This dose of dizocilpine eliminated the memory improvement caused by 0.2 mg / kg nicotine. However a higher dose of nicotine, 0.4 mg / kg, which is usually too high to cause a significant improvement in working memory performance, did cause a significant reduction in dizocilpine-induced deficits in both working and reference memory. One likely point of nicotinic–NMDA interaction regarding memory is the hippocampus. There is a significant nicotinic interaction with hippocampal NMDA systems. However, this interaction is quite different from what we have seen on a systemic level. We found that local ventral hippocampal dizocilpine infusions reversed the systemic effect of nicotine on memory such that systemic nicotine increased working memory error rates on the radial-arm
maze [18]. Interestingly, this reversal of the systemic nicotine effect was seen with ventral hippocampal dizocilpine doses that were not seen by themselves to impair memory function. The amygdala is another area likely to be important for nicotinic–NMDA interactions and memory. The amygdala is known to play an important role in memory function [22]. However, little is known about the role of nicotinic systems in the amygdala regarding memory. There have been several studies that suggest cholinergic activity as the mechanism by which the amygdala modulates memory [24,28,34]. Both muscarinic cholinergic and nicotinic cholinergic systems have been shown to be involved [24,28]. While the effects of cholinergic drugs on aversive tasks appear to involve the amygdala [34], cholinergic drug effects on spatial tasks do not seem to depend on the amygdala [34]. Several different studies have shown the amygdala to be responsive to nicotine exposure. Fos activity was increased upon nicotinic stimulation [35,40]. The amygdala has been generally thought to be primarily involved in the processing of emotional memories [21,29,33], however, there is evidence that the amygdala plays a broader role in learning and memory function. The amygdala is also involved in non-aversively motivated memory. While most tests of amygdalar function in learning and memory have involved training with aversive stimuli, there is also evidence that the amygdala is important for learning tasks that use appetitive motivation when the reinforcement is of high affective value [6,8,31]. In some studies, such as delayed non-matching-to-sample and the three panel runway appetitive task, basolateral amygdala lesions in rats led to working memory impairments [28,30]. Ohno et al. have shown working but not reference memory to be impaired by basolateral amygdala lesions [28]. Systemic treatment with indirect cholinergic agonists or direct muscarinic agonists improved the memory deficit. Aversive tasks have a different neural basis in the amygdala. Riekkinen and colleagues reported that the passive avoidance deficits caused by amygdaloid lesions were not reversible by nicotine or muscarinic treatment [34]. Not all studies have found impairments due to amygdala lesions. In some studies, such as spatial alternation, a lesion did not lead to any significant changes in performance [38]. The basolateral amygdala has high concentrations of NMDA glutamate receptors, and receives afferent glutaminergic projections from the cortex and thalamus [27]. Recent findings suggest that the NMDA receptor system in the amygdala is involved in aversively motivated learning. Microinfusion of an NMDA receptor antagonist, D,L-2-amino-5-phosphate ( D,L-AP5), into the basolateral amygdala impairs one-trial inhibitory avoidance learning [19], Pavlovian conditioning and extinction of fear-potentiated startle [4,26], blocking both acquisition and expression of contextual fear conditioning [5,20]. Most importantly pre- and post-training intra-amygdala administra-
H. May-Simera, E.D. Levin / Cognitive Brain Research 17 (2003) 475–483
tion of NMDA receptor antagonists has been shown to impair memory storage [9,19,26]. The purpose of this study is to investigate the interactions between nicotinic and glutaminergic systems with respect to memory function. The amygdala may be a critical site for the significant interactions found between nicotine and the NMDA antagonist dizocilpine. The effect of dizocilpine infusions into the lateral amygdala or piriform cortex combined with systemic nicotine on the working and reference memory in rats was assessed using a 16-arm radial maze. The object of this study is to shed light on the question of how nicotinic systems interact with amygdalar glutamate systems in the neural foundation of memory function and to help in the development of new treatments for cognitive dysfunction.
2. Methods
2.1. General design The purpose of this study was to determine the role of amygdalar systems in the expression of systemic nicotine effects. In particular this study investigated the interactions between systemic nicotine actions and amygdalar glutaminergic systems with respect to memory function. The effect of dizocilpine infusions into the basolateral amygdala combined with systemic nicotine on the working and reference memory in rats was assessed using a 16-arm radial maze. The rats were trained until asymptotic levels of choice accuracy had been reached, a minimum of 18 sessions on the radial-arm maze. Local infusion cannulae were then implanted bilaterally into the basolateral amygdala using stereotaxic techniques. After a short period of post-surgery training on the maze to test for adverse effects due to surgery, the drug trials were carried out. At 20 min prior to running the rats, they were injected subcutaneously with (2) nicotine ditartrate at doses of 0 and 0.4 mg / kg. Following this, they were infused with (1) dizocilpine maleate at doses of 0, 2, 6 and 18 mg per side, 10 min prior to running in every combination possible, making eight separate drug trials. The animals were sacrificed after completion of the drug sessions, and using histological methods the cannulae placements were verified.
2.2. Subjects A total of 24 young adult female Sprague–Dawley rats (Taconic Farms, Germantown, NY, USA) weighing between 165 and 220 g prior to training were used. The rats were housed in a rat colony room with a reverse 12-h light / 12-h dark cycle. All behavioral testing was carried out during the dark phase, their most active period. Testing
477
was done in a lit room to facilitate the animals use of visual cues to orient themselves around the maze. The rats were housed three per cage during the initial training phase and then one per cage after cannulation. They had ad libitum access to water, with a restricted diet administered daily after testing, to maintain body weights at |85% of free feeding levels.
2.3. Radial-arm maze training In order to test both working and reference memory, a radial-arm maze (RAM) with 16 arms was used. The maze, made out of wood and painted black, is constructed with a 50-cm central platform elevated 30 cm off the floor. A total of 16 arms, 10 cm360 cm, project radially from the central platform, each with a food cup 2 cm from the distal end. The maze was positioned in a quiet room with many extra-maze visual cues for the rat to orientate itself. The rats were handled for at least two 5-min periods on consecutive days in order to tame them. To acquaint the rats with the food reinforcement, Froot Loops (a sugarcoated cereal) were given as part of their food rations. The shaping phase consisted of placing the rats on the central platform inside a plastic cylinder, which also contained eight Froot Loop halves. The rats were given 5 min to eat the cereal. Once the rats ate over six pieces in one trial they were ready to be trained. The rats were trained 5 days a week for a minimum of 18 sessions. For each rat there was a unique pattern of 12 baited and four unbaited arms that was used for each session. The patterns did not include three sequentially unbaited arms. At the beginning of the training session, the maze was baited accordingly. The rat was then placed in the plastic cylinder positioned on the central platform for 10 s. This allowed for orientation and to prevent biased arm entry. Timing of the session began once the cylinder was removed and the rat was able to roam freely around the maze. The rat was allowed to run the maze for a maximum of 10 min or until all 12 baited arms had been entered. Arm choice was recorded when all four paws of the rat crossed the threshold of an arm. It was noted whether the cereal piece was eaten or not. As the positioning of the unbaited arms did not change over the sessions, entrance into unbaited arms tested reference memory and all entrances into unbaited arms were marked as reference memory errors. Baited arms were not re-baited once the rat had eaten the cereal, so repeated entries into baited arms were marked as working memory errors. Latency (seconds per entry) was calculated by dividing the total time of the session (seconds) by the number of arms entered. Working and reference memory errors as well as the latency were recorded after each session. After 18 sessions the rats were cannulated; those not undergoing cannulation continued to be run on the maze.
478
H. May-Simera, E.D. Levin / Cognitive Brain Research 17 (2003) 475–483
2.4. Cannulation For the acute local drug infusions into the basolateral amygdala, the rats were bilaterally implanted with stainless steel guide cannulae (22 gauge, Plastics One, Roanoke, VA, USA) via stereotaxic surgery. An intra-peritoneal injection of 0.75 mg / kg ketamine and 0.15 mg / kg medetomidine was given to anesthetize the animal prior to surgery. The rat’s head was shaved from between the eyes to behind the ears and then secured in the stereotaxic apparatus (David Kopf Instruments, Tajunga, CA, USA) with ear bars and bite bar elevating the head to 15 mm above the intra-aural line. The rat was placed on a stereotaxic heating pad to maintain its body temperature. An incision was made from behind the eyes to just in front of the ears. The bregma was then located. The bregma was used as a point of reference for the coordinates of the basolateral amygdala obtained from the Pellegrino atlas [32]. The guide cannulae were lowered through holes drilled in the skull to the following coordinates from bregma: AP 21.4 mm, ML 15.5 mm, DV 28.5 mm. Four screws were anchored in the skull between the cannulae; wire was wrapped around these screws before applying the cranioplastic cement to secure the cannulae. Dummy cannulae were placed into the guide cannulae to prevent infection or blockage. After surgery was complete the rat was injected with 0.15 mg / kg antisedan to counteract the effects of the medetomidine. The rats were given 1 week to recover from surgery and were checked and weighed every 2 days to ensure they were recovering well and showing no adverse effects due to surgery.
followed by an infusion 10 min later of dizocilpine at a rate of 0.126 ml / min using a Harvard Instruments infusion pump. The rats were run on the maze 10 min after the infusion. The drug treatments were spaced at least 48 h apart, to allow the drugs to clear the rat’s system. Each animal received all eight drug treatments in a counterbalanced design.
2.6. Histology Following the completion of the drug trials, each rat was anesthetized with sodium pentobarbital (50 mg / kg). To mark the cannulae placements, 0.5 ml of a solution of saline and Chicago sky-blue dye was infused into each cannulae. Perfusion of each rat was carried out using a 9% phosphate-buffered saline solution followed by a 0.1 M 4% paraformaldehyde solution. The rat brains were removed and stored in 4% paraformaldehyde until slicing. Before slicing, the cerebellum was cut off and the brains were washed in distilled H 2 O. The brains were frozen on dry ice before being sliced on a cryostat (Cryocut 1800, Reichert-Jung) to prepare histological slides. The slides were studied under a microscope to verify and record the cannulae placements.
2.7. Data analysis The choice accuracy and response latency measures were assessed by a within subjects design analysis of variance. A P-value of less than 0.05 was considered significant for main effects. Comparisons were made between the effect of dizocilpine in the presence and absence of nicotine.
2.5. Drug treatment The rats were trained until asymptotic levels of choice accuracy had been reached, a minimum of 18 sessions, after which they were cannulated. Following the cannulation the animals began the drug studies. The drug treatments consisted of a subcutaneous injection of (2) nicotine ditartrate (saline vehicle) at doses of 0.0 (saline) or 0.4 mg / kg, and infusions of (1) dizocilpine maleate (saline vehicle) into the basolateral amygdala, via the cannulae, at doses of 0, 2, 6 and 18 mg per side. There being eight different combinations of these drugs, the animals were subjected to eight separate drug trials. Then 1 week after surgery the rats were given a trial run on the maze to determine any loss in performance due to the adverse effects of surgery. Following the post-surgery run, the rats were injected and infused with two test doses of artificial cerebrospinal fluid (ACSF) to accustom the animals to the procedure. Once these trial runs had been completed the drug treatments began. The nicotine injection was delivered subcutaneously,
3. Results
3.1. Histology Cannula placements were verified histologically. Many of the placements were found to have missed the specific target area of the basolateral amygdala. However they still fell into the range of the lateral amygdala, thus the data were still applicable for this study. Out of the 24 rats that underwent surgery, nine succeeded in having both (left and right sides) cannulae on target and had sufficient data to be analyzed. A further six rats had bilateral placements in the piriform cortex. These placements served as a control to the amygdaloid placements. Animals with cannulae in two different brain locations on opposite sides of the brain were discarded from the study. Fig. 1 shows two coronal diagrams of the brain at an anterior–posterior coordinate of 21.4 with the cannulae placements marked. Fig. 1a shows
H. May-Simera, E.D. Levin / Cognitive Brain Research 17 (2003) 475–483
479
mean working errors were higher among the animals with placements in the amygdala (9.5268.49) than the piriform cortex (7.1265.51). The 0.4-mg / kg dose of nicotine did not have a significant effect on working memory errors and did not reverse the dizocilpine induced impairment. The significant effect of dizocilpine was only seen with the high doses of 18 mg per side and only within the amygdala (P,0.005 without nicotine; P.0.05 with nicotine). Neither drug appeared to have any effect on working errors in the rats with piriform cortex cannulae. The reference memory errors showed a similar pattern of results to the working memory errors (Fig. 3). There appears to be an overall significant effect of dizocilpine on reference errors (P,0.005). There is a nearly significant effect of dizocilpine and the placement of the cannulae (P,0.08) although this is not as great an effect as for the working memory errors. The dizocilpine effect was only observed in the amygdalar cannulae (amygdala: P,0.005). The significant effect of dizocilpine with the amygdalar infusions was only seen among the highest dose of dizocilpine, 18 mg per side (P,0.0005 without nicotine; P,0.0005 with nicotine). Nicotine did not seem to have a significant effect in rats with either cannulae locations or on any of the dizocilpine doses. There was no significant interaction between the two drugs. As in the working memory errors, the animals with amygdalar placements had overall slightly more errors (5.4264.24) compared with rats with the piriform cortex placements (4.0262.07). The animals with cannulae placements in the piriform cortex showed virtually no effect from either drug.
3.3. Latency Fig. 1. Placement of cannulae for amygdala (a) and piriform cortex (b) [32].
the placements located within the amygdala and Fig. 1b the placements within the piriform cortex.
3.2. Choice accuracy Working memory was significantly affected by dizocilpine. The analysis of variance showed that overall doses of dizocilpine had a significant effect on working errors (P,0.005). There was a significant interaction of dizocilpine and the placement area (P,0.05). The dizocilpine effect was mainly due to the placements in the amygdala as there was a significant working memory impairment caused by dizocilpine infusion into the amygdala (P,0.005), but not into the piriform cortex (Fig. 2). There was no significant effect of nicotine administration in either the amygdala or the piriform cortex dizocilpine infusion studies. There were also no significant interactions of dizocilpine and nicotine for either brain placement. Regardless of the treatment the animals received, the
Response latency showed a significant main effect of dizocilpine (P,0.0001), the latencies decreasing as the dizocilpine dose increased (Fig. 4). Nicotine also had a significant effect on response latency (P,0.001), lowering the latencies at the 2 and 6 mg per side doses of dizocilpine. There was also a significant nicotine3 dizocilpine interaction (P,0.05). The latency reducing effects of nicotine and dizocilpine were less than additive. Tests of the simple main effects of nicotine at each dose of dizocilpine showed quite significant nicotine-induced reductions in latency with dizocilpine vehicle infusions (P, 0.0001) and low (2 mg per side) dizocilpine infusions (P,0.0005), but not quite significant nicotine effects on latency with 6 mg per side dose of dizocilpine (P50.08) and not nearly significant nicotine effects with 18 mg per side (P50.27). No interactions of drug effects on latency with brain infusion area were seen. Latency scores appeared to be generally higher with the placements within the piriform cortex than the amygdala. However, there were no significant latency differences between the rats with cannulae in the piriform cortex and those with cannulae in the amygdala with control infusions. The rats with the piriform cortex cannulae did have
480
H. May-Simera, E.D. Levin / Cognitive Brain Research 17 (2003) 475–483
Fig. 2. Dizocilpine infusion into the amygdala (n59) and piriform cortex (n56) and systemic nicotine: effects on working memory (mean6S.E.M.).
slightly higher mean latency scores (27.363.2 s per entry) than those with the amygdala cannulae (22.864.1 s per entry), but these were not nearly significantly different (P50.44).
4. Discussion Acute amygdalar infusions of the NMDA glutamate receptor antagonist dizocilpine induced dose-related work-
Fig. 3. Dizocilpine infusion into the amygdala (n59) and piriform cortex (n56) and systemic nicotine: effects on reference memory (mean6S.E.M.).
H. May-Simera, E.D. Levin / Cognitive Brain Research 17 (2003) 475–483
481
Fig. 4. Dizocilpine infusion into the amygdala (n59) and piriform cortex (n56) and systemic nicotine: effects on response latency (mean6S.E.M.).
ing and reference memory deficits in the radial-arm maze. Systemic nicotine was not seen to reverse these effects. It was expected that nicotine would reverse the memory deficit caused by dizocilpine infused into the amygdala. The 0.4-mg / kg nicotine ditartrate dose was sufficient to block the amnestic effect of systemic dizocilpine [16], which is why that dose was used in the current study. A lower dose of 0.2 mg / kg of nicotine ditartrate was not seen in a previous study to be effective in reversing the amnestic effects of systemic dizocilpine. Higher doses of nicotine have adverse effects that limit their usefulness in improving memory function. Dizocilpine infusions into the adjacent piriform cortex did not impair memory function, supporting the specificity of dizocilpine effects in the amygdala. Latency effects were seen with both drugs in both areas. Latencies were decreased with increasing doses of dizocilpine and with administration of nicotine. This study demonstrated the importance of NMDA glutamate systems in the amygdala for appetitively-motivated spatial memory performance. The currently observed effects of dizocilpine in the amygdala contrast with previously observed effects in the ventral hippocampus. In a previous study in our laboratory, the same experiments were carried out on animals with cannulae into the ventral hippocampus [18]. In that study none of the three doses of dizocilpine caused significant working memory impairments when administered alone, but all three caused significant working memory deficits when administered in conjunction with systemic nicotine.
No effect of either treatment, alone or together, was seen with reference memory or response latency. The variation in results indicates the different roles the brain regions play in cognitive function. The finding that ventral hippocampal dizocilpine caused systemic nicotine to actually impair memory points to the importance of ventral hippocampal NMDA receptors for the positive effects of nicotine on memory. Blocking these receptors unveiled the expression of other nicotinic actions that impair working memory function. In the previous study by Levin et al. [16], it was proposed that nicotine may have counteracted the dizocilpine induced memory deficits through its action in stimulating glutamate release [23]. Because nicotine and NMDA receptors are similar in structure, cross-reactivity of ligands for these receptors may also have been involved in the observed interactions. Our findings in the present study complicate the situation, as nicotine was shown not to have an effect on choice accuracy. This could have been due to the method of administration. The direct infusion of dizocilpine into the brain might have been too intense and localized, so that the systemic nicotine could not exert its effects. Infusion of dizocilpine also enables the delivery of the drug to a precise location within the brain targeting the NMDA receptors in a specific region of the brain, not just all receptor sites. In the current study the animals were pretrained so that the drug effects on memory performance would be assessed during the stable asymptotic post acquisition phase
482
H. May-Simera, E.D. Levin / Cognitive Brain Research 17 (2003) 475–483
of performance. The within-subjects counterbalanced design was used so that possible carryover drug effects would not be confounded with order of treatment dose. At least 2 days were interposed between drug doses to minimize carryover effects. The piriform cortex infusions served as a useful control. Inasmuch as the memory effects were specific to the amygdala and not the piriform cortex, widespread diffusion of dizocilpine from the infusion site in the amygdala did not appear to be the cause of the observed effects on memory function. Also the larger volume of the dye infused for verification (0.5 ml per side) did not appear to invade nuclei outside the target areas. Tissue damage from the infusions was not seen to be a problem. Even subtle tissue damage with repeated drug infusion would not have been confounded with the observed drug effects because the drugs were administered in a repeated measures counterbalanced design so that the order of drug infusion was not confounded with dose. Previously it had been found that systemic nicotine administration improved working memory in the radial arm maze [11]. The present study showed that infusion of the NMDA antagonist dizocilpine had a significant effect in the amygdala but not the piriform cortex on both working and reference memory errors. Nicotine (0.4 mg / kg) did not reverse memory impairment caused by dizocilpine. Latency was significantly decreased with the rats displaying an increased level of activity with increasing dose of dizocilpine infused into both brain regions. Nicotine effected latency. As previously demonstrated by McLamb et al. [25] and Levin et al. [16], the current study found that dizocilpine impaired memory. However, in contrast to our earlier study [16], nicotine attenuated the dizocilpine induced memory deficit. The difference in findings could be due to the type of drug delivery involved in each study. In the present study, dizocilpine was infused directly into the amygdala rather than being injected subcutaneously [16]. Systemic dizocilpine has a global effect on all parts of the central and peripheral nervous system, whereas locally infused dizocilpine is specific to one area in the brain. This suggests that extra-amygdala NMDA glutamate receptors may have different effects to those in the amygdala. It is also important to note the dose of nicotine administered. A variety of studies have shown that acute treatment with nicotine or nicotinic agonists can improve working memory function in the radial-arm maze and other tasks [3,12,13]. In six separate studies, it was found that a single dose of 0.2 mg / kg nicotine injected subcutaneously significantly improved working memory [15]. Improvements could also be seen at lower and higher doses (0.1 and 0.4 mg / kg), but with the typical inverted U-shaped function for memory-enhancing drugs, it was found that 0.2 mg / kg nicotine was the most effective dose. For this particular study 0.4 mg / kg was used as the previous study by Levin et al. [16] had found that particular dose to be most
effective at reversing the effects of dizocilpine. It could be that different doses of nicotine are needed to counteract the effects of locally infused dizocilpine. The current study found that blockade of the lateral and basolateral amygdalar NMDA glutamate receptors with dizocilpine significantly impaired working and reference memory. This effect was regionally specific with the same doses infused into the adjacent piriform cortex having no discernible effect. Previously, we have found that these doses given into the ventral hippocampus did not by themselves impair memory performance on the radial-arm maze [18]. Nicotine co-administration did not significantly interact with the effects of dizocilpine infused into the amygdala. This also was regionally selective with a substantial nicotine-induced memory impairment seen with the same dose range of dizocilpine infused into the ventral hippocampus. Nicotinic receptors in the amygdala are important for memory function as shown by our work with local nicotinic antagonist infusions [1]. This is similar to what we have found in the hippocampus [17]. However, the current study shows that nicotinic interactions with NMDA glutamate systems in the hippocampus and amygdala differ substantially.
Acknowledgements The authors thank Channelle Christopher for her invaluable assistance in this study. This research was supported in part by a grant from the Philip Morris External Research Foundation.
References [1] N. Addy, A. Nakajama, E. Levin, Nicotinic mechanisms of memory: effects of acute local DHbE and MLA infusions in the basolateral amygdala, Cogn. Brain Res. 16 (2003) 51–57. [2] B.J. Cole, M. Klewer, G.H. Jones, D.N. Stephens, Contrasting effects of the competitive NMDA antagonist CPP and the noncompetitive NMDA antagonist MK-801 on performance of an operant delayed matching to position task in rats, Psychopharmacology 111 (1993) 465–471. [3] M.W. Decker, J.D. Brioni, A.W. Bannon, S.P. Arneric, Diversity of neuronal nicotinic acetylcholine receptors: lessons from behavior and implications for CNS therapeutics—minireview, Life Sci. 56 (1995) 545–570. [4] W. Falls, M. Miserendino, M. Davis, Extinction of fear-potentiated startle: blockade by infusion of an NMDA antagonist into the amygdala, J. Neurosci. 12 (1992) 854–863. [5] M. Fanselow, J. Kim, Acquisition of contextual Pavlovian fear conditioning is blocked by application of an NMDA receptor antagonist D,L-2-amino-5-phosphonovaleric acid to the basolateral amygdala, Behav. Neurosci. 108 (1994) 210–212. [6] N. Hiroi, N. White, The lateral nucleus of the amygdala mediates expression of the amphetamine-produced conditioned place preference, J. Neurosci. 11 (1991) 2107–2116. [7] I. Izquierdo, Role of NMDA receptors in memory, Trends Pharmacol. Sci. 12 (1991) 128–129. [8] R.P. Kesner, J.M. Williams, Memory for magnitude of rein-
H. May-Simera, E.D. Levin / Cognitive Brain Research 17 (2003) 475–483
[9]
[10] [11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
forcement: dissociation between the amygdala and hippocampus, Neurobiol. Learn. Mem. 64 (1995) 237–244. M. Kim, J.L. McGaugh, Microinfusion of an N-methyl-D-aspartate antagonist impairs avoidance learning in rats, Soc. Neurosci. Abstr. 16 (1990) 767. E.D. Levin, Nicotinic systems and cognitive function, Psychopharmacology 108 (1992) 417–431. E.D. Levin, The role of nicotinic acetylcholine receptors in cognitive function, in: F. Clementi, C. Gotti, D. Fornasari (Eds.), Handbook of Experimental Pharmacology: Neuronal Nicotinic Receptors, Springer, New York, 2000, pp. 587–602. E.D. Levin, A.H. Rezvani, The promise of nicotinic treatment for cognitive dysfunction, Curr. Drug Targets—CNS Neurol. Disord. 1 (2002) 423–431. E.D. Levin, D. Torry, Nicotine effects on memory performance, in: P.B.S. Clarke, M. Quik, K. Thurau, F. Adlkofer (Eds.), International Symposium on Nicotine: The Effects of Nicotine on Biological ¨ Systems II, Birkhauser, Boston, 1995, pp. 329–336. E.D. Levin, P. Kim, R. Meray, Chronic nicotine effects on working and reference memory in the 16-arm radial maze: interactions with D1 agonist and antagonist drugs, Psychopharmacology 127 (1996) 25–30. E.D. Levin, S. Kaplan, A. Boardman, Acute nicotine interactions with nicotinic and muscarinic antagonists: working and reference memory effects in the 16-arm radial maze, Behav. Pharmacol. 8 (1997) 236–242. E.D. Levin, C. Bettegowda, T. Weaver, N.C. Christopher, Nicotine– dizocilpine interactions and working and reference memory performance of rats in the radial-arm maze, Pharmacol. Biochem. Behav. 61 (1998) 335–340. E.D. Levin, A. Bradley, N. Addy, N. Sigurani, Hippocampal alpha 7 and alpha 4 beta 2 nicotinic receptors and working memory, Neuroscience 109 (2002) 757–765. E.D. Levin, D. Sledge, A. Baruah, N. Addy, Hippocampal NMDA blockade and nicotinic effects on memory function, Brain Res. Bull. (2003) (in press). K.C. Liang, M. Davis, Intra-amygdala injection of N-methyl-Daspartate receptor antagonists impairs memory in an inhibitory avoidance task, Soc. Neurosci. Abstr. 10 (1990) 767. S. Maren, G. Aharonov, D.L. Stote, M.S. Fanselow, N-methyl-Daspartate receptors in the basolateral amygdala are required for both acquisition and expression of conditional fear in rats, Behav. Neurosci. 110 (1996) 1365–1374. J.L. Martinez, K.C. Liang, A. Oscos, Amnesia induced by stimulation of the amygdala is attenuated by hexamethonium, Psychopharmacology 81 (1983) 310–314. J. McGaugh, I. Introini-Collison, A. Nagahara, L. Cahill, Involvement of the amygdaloid complex in neuromodulatory influences on memory storage, Neurosci. Biochem. Rev. 14 (1990) 425–431. D.S. McGehee, M.J.S. Heath, S. Gelber, P. Devay, L.W. Role, Nicotine enhancement of fast excitatory synaptic transmission in CNS by presynaptic receptors, Science 269 (1995) 1692–1696. C.K. McIntyre, M.E. Ragozzino, P.E. Gold, Intra-amygdala infusions of scopolamine impair performance on a conditioned place preference task but not a spatial radial maze task, Behav. Brain Res. 95 (1998) 219–226. R. McLamb, L. Williams, K. Nanry, W. Wilson, H. Tilson, MK-801
[26]
[27]
[28]
[29]
[30]
[31] [32] [33]
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
483
impedes the acquisition of a spatial memory task in rats, Pharmacol. Biochem. Behav. 37 (1990) 41–45. M.J. Miserendino, C.B. Sananes, K.R. Melia, M. Davis, Blocking of acquisition but not expression of conditioned fear-potentiated startle by NMDA antagonists in the amygdala, Nature 345 (1990) 716– 718. D.T. Monaghan, C.W. Cotman, Distribution of N-methyl-D-aspartatesensitive L-[ 3 H]glutamate-binding sites in rat brain, J. Neurosci. 5 (1985) 2909–2919. M. Ohno, T. Yamamoto, S. Watanabe, Involvement of cholinergic mechanisms in impairment of working memory in rats following basolateral amygdaloid lesions, Neuropharmacology 31 (1992) 915– 922. M.B. Parent, J.L. McGaugh, Posttraining infusion of lidocaine into the amygdala basolateral complex impairs retention of inhibitory avoidance training, Brain Res. 661 (1994) 97–103. M.A. Peinado-Manzano, The role of the amygdala and the hippocampus in working memory for spatial and non-spatial information, Behav. Brain Res. 38 (1990) 117–134. M.A. Peinado-Manzano, Amygdala, hippocampus and associative memory in rats, Behav. Brain Res. 61 (1994) 175–190. L.J. Pellegrino, A.S. Pellegrino, A.J. Cushman, A Stereotaxic Atlas of the Rat Brain, 2nd Edition, Plenum, New York, 1979. A.E. Power, B. Roozendaal, J.L. McGaugh, Glucocorticoid enhancement of memory consolidation in the rat is blocked by muscarinic receptor antagonism in the basolateral amygdala, Eur. J. Neurosci. 12 (2000) 3481–3487. P. Riekkinen Jr., M. Riekkinen, J. Sirvio, Cholinergic drugs regulate passive avoidance performance via the amygdala, J. Pharmacol. Exp. Ther. 267 (1993) 1484–1492. O. Salminen, T. Seppa, H. Gaddnas, L. Ahtee, The effects of acute nicotine on the metabolism of dopamine and the expression of Fos protein in striatal and limbic brain areas of rats during chronic nicotine infusion and its withdrawal, J. Neurosci. 19 (1999) 8145– 8151. M.L. Shapiro, C. O’Connor, N-methyl-D-aspartate receptor antagonist MK-801 and spatial memory representation: working memory is impaired in an unfamiliar environment but not in a familiar environment, Behav. Neurosci. 106 (1992) 604–612. M.L. Shapiro, Z. Caramanos, NMDA antagonist MK-801 impairs acquisition but not performance of spatial working and reference memory, Psychobiology 18 (1990) 231–243. R. Wan, K. Pang, D. Olton, Hippocampal and amygdaloid involvement in non-spatial and spatial working memory in rats: effects of delay and interference, Behav. Neurosci. 108 (1994) 866–882. L. Ward, S.E. Mason, W.C. Abraham, Effects of the NMDA antagonists CPP and MK-801 on radial-arm maze performance in rats, Pharmacol. Biochem. Behav. 35 (1990) 785–790. K. Watanabe, K. Hashimoto, T. Nishimura, K. Tsunashima, Y. Minabe, Expression of Fos protein in rat brain following administration of a nicotinic acetylcholine receptor agonist epibatidine, Brain Res. 797 (1998) 135–142. S. Wonnacott, J. Irons, C. Rapier, B. Thorne, G.G. Lunt, Presynaptic modulation of transmitter release by nicotinic receptors, in: A. Nordberg, K. Fuxe, B. Holmstedt, A. Sundwall (Eds.), Progress in Brain Research, Vol. 79, Elsevier, New York, NY, 1989, pp. 157– 163.
Research report
NMDA systems in the amygdala and piriform cortex and nicotinic effects on memory function Helen May-Simera, Edward D. Levin* Neurobehavioral Research Laboratory, Department of Psychiatry, Box [3412, Duke University Medical Center, Durham, NC 27710, USA Accepted 29 April 2003
Abstract Both nicotinic cholinergic and NMDA glutaminergic systems are important for memory function. Nicotine has been found repeatedly to significantly improve working memory performance in the radial-arm maze. The NMDA antagonist dizocilpine has been found to impair working memory performance. There is neuropharmacological evidence that these two systems are functionally related. Nicotine is potent at releasing many transmitters including glutamate. The current study was conducted to examine the interaction of nicotinic and NMDA systems within the amygdala with regard to working and reference memory. Rats were trained on a working / reference procedure on a 16-arm radial maze. After acquisition, local infusion cannulae were implanted bilaterally into the amygdala and piriform cortex using stereotaxic techniques. Then 20 min prior to running the rats on the radial-arm maze, they were injected subcutaneously with (2) nicotine ditartrate at doses of 0 and 0.4 mg / kg. Following this, the rats received local infusions of (1) dizocilpine maleate (MK-801) at doses of 0, 2, 6 and 18 mg per side into the lateral amygdala or piriform cortex 10 min prior to running on the radial-arm maze. Each of the eight nicotine and dizocilpine combinations was administered to each rat in a counterbalanced order. After completion of the drug sessions the rats were sacrificed, and using histological methods the cannulae placements were verified. Acute amygdalar infusions of the NMDA glutamate receptor antagonist dizocilpine induced dose-related working and reference memory deficits in the radial-arm maze. Systemic nicotine was not seen to reverse these effects. Dizocilpine infusions into the adjacent piriform cortex did not impair memory function, supporting the specificity of dizocilpine effects in the amygdala. Latency effects were seen with both drugs in both areas. Latencies were decreased with both systemic nicotine and dizocilpine in both the lateral amygdala and the piriform cortex. This study demonstrated the importance of NMDA glutamate systems in the amygdala for appetitively-motivated spatial memory performance. 2003 Elsevier B.V. All rights reserved. Theme: Neural basis of behavior Topic: Learning and memory: systems and functions–animals Keywords: Nicotine; Dizocilpine; MK-801; Memory; Amygdala; Piriform cortex; Radial-arm maze
1. Introduction Numerous studies have shown that both N-methyl-Daspartate (NMDA) glutaminergic and nicotinic cholinergic receptor systems are important for memory function (for reviews see Refs. [7,10,11]). We have found that nicotinic– NMDA interactions are important inasmuch as systemic nicotine administration can substantially attenuate the memory impairment caused by systemic administration of the NMDA antagonist dizocilpine (MK-801) [16]. This *Corresponding author. Tel.: 11-919-681-6273; fax: 11-919-6813416. E-mail address: [email protected] (E.D. Levin). 0926-6410 / 03 / $ – see front matter 2003 Elsevier B.V. All rights reserved. doi:10.1016 / S0926-6410(03)00163-0
nicotine-induced reversal of the memory impairment caused by dizocilpine may be related to the fact that nicotine stimulates glutamate release [23,41]. The anatomic substrate of this interaction is not yet known. Determining the anatomic sites for nicotinic–NMDA interactions and memory function will help advance our understanding of the complex neural basis of cognitive function and will help in the development of novel avenues of treatment for cognitive dysfunction. The NMDA antagonist dizocilpine causes substantial impairment of spatial learning and memory [36,37]. Cole et al. [2] found that both 0.1 and 0.2 mg / kg of dizocilpine caused significant deficits in an operant delayed matching to position task in a delay-dependent manner. Ward et al.
476
H. May-Simera, E.D. Levin / Cognitive Brain Research 17 (2003) 475–483
[39] showed that dizocilpine can produce a dose-dependent disruption of radial-arm maze memory performance. Acquisition of working and reference memory performance on an eight-arm radial maze was impaired by 0.0625 mg / kg of dizocilpine [37], but retention of performance was not affected. We have found working and reference memory impairments in the 16-arm radial maze caused by dizocilpine at doses as low as 25 mg / kg [16]. NMDA systems in the amygdala are important components for glutamate involvement in cognitive function. Pre-training intra-amygdala infusion of AP5 or AP7, other NMDA receptor antagonists, hinders retention of a condition startle response [26]. The immediate post-training intra-amygdala infusion of AP5 hinders retention of stepthrough inhibitory avoidance [19], of a mixture of active and inhibitory avoidance called continuous avoidance [9], and of a step-down inhibitory avoidance and a water finding task. Nicotinic agonists have been found in rodent and nonhuman primate studies to improve performance on a variety of memory tasks. The 16-arm radial maze has been used to show the relative specificity of nicotine-induced improvement in working, but not reference, memory [14,15]. In a complementary fashion, nicotinic antagonists such as mecamylamine have been shown to impair working memory function. A variety of studies have shown that acute treatment with nicotine or nicotinic agonists can improve working memory function in the radial-arm maze in rats [3,13,15]. An acute dose of 0.2 mg / kg nicotine injected subcutaneously 20 min before testing significantly improved working memory with the typical inverted Ushaped function for memory-enhancing drugs. This effect was specific to working memory (changing memory specific to each session). Reference memory (unchanging memory constant across sessions) errors were not significantly affected. It has also been found that acute treatment with other selective nicotinic agonists significantly improves memory performance [12]. Nicotine has been shown to attenuate the amnestic effects of dizocilpine. Dizocilpine (0.1 mg / kg) caused substantial deficits in both working and reference memory on the 16-arm radial maze [16]. This dose of dizocilpine eliminated the memory improvement caused by 0.2 mg / kg nicotine. However a higher dose of nicotine, 0.4 mg / kg, which is usually too high to cause a significant improvement in working memory performance, did cause a significant reduction in dizocilpine-induced deficits in both working and reference memory. One likely point of nicotinic–NMDA interaction regarding memory is the hippocampus. There is a significant nicotinic interaction with hippocampal NMDA systems. However, this interaction is quite different from what we have seen on a systemic level. We found that local ventral hippocampal dizocilpine infusions reversed the systemic effect of nicotine on memory such that systemic nicotine increased working memory error rates on the radial-arm
maze [18]. Interestingly, this reversal of the systemic nicotine effect was seen with ventral hippocampal dizocilpine doses that were not seen by themselves to impair memory function. The amygdala is another area likely to be important for nicotinic–NMDA interactions and memory. The amygdala is known to play an important role in memory function [22]. However, little is known about the role of nicotinic systems in the amygdala regarding memory. There have been several studies that suggest cholinergic activity as the mechanism by which the amygdala modulates memory [24,28,34]. Both muscarinic cholinergic and nicotinic cholinergic systems have been shown to be involved [24,28]. While the effects of cholinergic drugs on aversive tasks appear to involve the amygdala [34], cholinergic drug effects on spatial tasks do not seem to depend on the amygdala [34]. Several different studies have shown the amygdala to be responsive to nicotine exposure. Fos activity was increased upon nicotinic stimulation [35,40]. The amygdala has been generally thought to be primarily involved in the processing of emotional memories [21,29,33], however, there is evidence that the amygdala plays a broader role in learning and memory function. The amygdala is also involved in non-aversively motivated memory. While most tests of amygdalar function in learning and memory have involved training with aversive stimuli, there is also evidence that the amygdala is important for learning tasks that use appetitive motivation when the reinforcement is of high affective value [6,8,31]. In some studies, such as delayed non-matching-to-sample and the three panel runway appetitive task, basolateral amygdala lesions in rats led to working memory impairments [28,30]. Ohno et al. have shown working but not reference memory to be impaired by basolateral amygdala lesions [28]. Systemic treatment with indirect cholinergic agonists or direct muscarinic agonists improved the memory deficit. Aversive tasks have a different neural basis in the amygdala. Riekkinen and colleagues reported that the passive avoidance deficits caused by amygdaloid lesions were not reversible by nicotine or muscarinic treatment [34]. Not all studies have found impairments due to amygdala lesions. In some studies, such as spatial alternation, a lesion did not lead to any significant changes in performance [38]. The basolateral amygdala has high concentrations of NMDA glutamate receptors, and receives afferent glutaminergic projections from the cortex and thalamus [27]. Recent findings suggest that the NMDA receptor system in the amygdala is involved in aversively motivated learning. Microinfusion of an NMDA receptor antagonist, D,L-2-amino-5-phosphate ( D,L-AP5), into the basolateral amygdala impairs one-trial inhibitory avoidance learning [19], Pavlovian conditioning and extinction of fear-potentiated startle [4,26], blocking both acquisition and expression of contextual fear conditioning [5,20]. Most importantly pre- and post-training intra-amygdala administra-
H. May-Simera, E.D. Levin / Cognitive Brain Research 17 (2003) 475–483
tion of NMDA receptor antagonists has been shown to impair memory storage [9,19,26]. The purpose of this study is to investigate the interactions between nicotinic and glutaminergic systems with respect to memory function. The amygdala may be a critical site for the significant interactions found between nicotine and the NMDA antagonist dizocilpine. The effect of dizocilpine infusions into the lateral amygdala or piriform cortex combined with systemic nicotine on the working and reference memory in rats was assessed using a 16-arm radial maze. The object of this study is to shed light on the question of how nicotinic systems interact with amygdalar glutamate systems in the neural foundation of memory function and to help in the development of new treatments for cognitive dysfunction.
2. Methods
2.1. General design The purpose of this study was to determine the role of amygdalar systems in the expression of systemic nicotine effects. In particular this study investigated the interactions between systemic nicotine actions and amygdalar glutaminergic systems with respect to memory function. The effect of dizocilpine infusions into the basolateral amygdala combined with systemic nicotine on the working and reference memory in rats was assessed using a 16-arm radial maze. The rats were trained until asymptotic levels of choice accuracy had been reached, a minimum of 18 sessions on the radial-arm maze. Local infusion cannulae were then implanted bilaterally into the basolateral amygdala using stereotaxic techniques. After a short period of post-surgery training on the maze to test for adverse effects due to surgery, the drug trials were carried out. At 20 min prior to running the rats, they were injected subcutaneously with (2) nicotine ditartrate at doses of 0 and 0.4 mg / kg. Following this, they were infused with (1) dizocilpine maleate at doses of 0, 2, 6 and 18 mg per side, 10 min prior to running in every combination possible, making eight separate drug trials. The animals were sacrificed after completion of the drug sessions, and using histological methods the cannulae placements were verified.
2.2. Subjects A total of 24 young adult female Sprague–Dawley rats (Taconic Farms, Germantown, NY, USA) weighing between 165 and 220 g prior to training were used. The rats were housed in a rat colony room with a reverse 12-h light / 12-h dark cycle. All behavioral testing was carried out during the dark phase, their most active period. Testing
477
was done in a lit room to facilitate the animals use of visual cues to orient themselves around the maze. The rats were housed three per cage during the initial training phase and then one per cage after cannulation. They had ad libitum access to water, with a restricted diet administered daily after testing, to maintain body weights at |85% of free feeding levels.
2.3. Radial-arm maze training In order to test both working and reference memory, a radial-arm maze (RAM) with 16 arms was used. The maze, made out of wood and painted black, is constructed with a 50-cm central platform elevated 30 cm off the floor. A total of 16 arms, 10 cm360 cm, project radially from the central platform, each with a food cup 2 cm from the distal end. The maze was positioned in a quiet room with many extra-maze visual cues for the rat to orientate itself. The rats were handled for at least two 5-min periods on consecutive days in order to tame them. To acquaint the rats with the food reinforcement, Froot Loops (a sugarcoated cereal) were given as part of their food rations. The shaping phase consisted of placing the rats on the central platform inside a plastic cylinder, which also contained eight Froot Loop halves. The rats were given 5 min to eat the cereal. Once the rats ate over six pieces in one trial they were ready to be trained. The rats were trained 5 days a week for a minimum of 18 sessions. For each rat there was a unique pattern of 12 baited and four unbaited arms that was used for each session. The patterns did not include three sequentially unbaited arms. At the beginning of the training session, the maze was baited accordingly. The rat was then placed in the plastic cylinder positioned on the central platform for 10 s. This allowed for orientation and to prevent biased arm entry. Timing of the session began once the cylinder was removed and the rat was able to roam freely around the maze. The rat was allowed to run the maze for a maximum of 10 min or until all 12 baited arms had been entered. Arm choice was recorded when all four paws of the rat crossed the threshold of an arm. It was noted whether the cereal piece was eaten or not. As the positioning of the unbaited arms did not change over the sessions, entrance into unbaited arms tested reference memory and all entrances into unbaited arms were marked as reference memory errors. Baited arms were not re-baited once the rat had eaten the cereal, so repeated entries into baited arms were marked as working memory errors. Latency (seconds per entry) was calculated by dividing the total time of the session (seconds) by the number of arms entered. Working and reference memory errors as well as the latency were recorded after each session. After 18 sessions the rats were cannulated; those not undergoing cannulation continued to be run on the maze.
478
H. May-Simera, E.D. Levin / Cognitive Brain Research 17 (2003) 475–483
2.4. Cannulation For the acute local drug infusions into the basolateral amygdala, the rats were bilaterally implanted with stainless steel guide cannulae (22 gauge, Plastics One, Roanoke, VA, USA) via stereotaxic surgery. An intra-peritoneal injection of 0.75 mg / kg ketamine and 0.15 mg / kg medetomidine was given to anesthetize the animal prior to surgery. The rat’s head was shaved from between the eyes to behind the ears and then secured in the stereotaxic apparatus (David Kopf Instruments, Tajunga, CA, USA) with ear bars and bite bar elevating the head to 15 mm above the intra-aural line. The rat was placed on a stereotaxic heating pad to maintain its body temperature. An incision was made from behind the eyes to just in front of the ears. The bregma was then located. The bregma was used as a point of reference for the coordinates of the basolateral amygdala obtained from the Pellegrino atlas [32]. The guide cannulae were lowered through holes drilled in the skull to the following coordinates from bregma: AP 21.4 mm, ML 15.5 mm, DV 28.5 mm. Four screws were anchored in the skull between the cannulae; wire was wrapped around these screws before applying the cranioplastic cement to secure the cannulae. Dummy cannulae were placed into the guide cannulae to prevent infection or blockage. After surgery was complete the rat was injected with 0.15 mg / kg antisedan to counteract the effects of the medetomidine. The rats were given 1 week to recover from surgery and were checked and weighed every 2 days to ensure they were recovering well and showing no adverse effects due to surgery.
followed by an infusion 10 min later of dizocilpine at a rate of 0.126 ml / min using a Harvard Instruments infusion pump. The rats were run on the maze 10 min after the infusion. The drug treatments were spaced at least 48 h apart, to allow the drugs to clear the rat’s system. Each animal received all eight drug treatments in a counterbalanced design.
2.6. Histology Following the completion of the drug trials, each rat was anesthetized with sodium pentobarbital (50 mg / kg). To mark the cannulae placements, 0.5 ml of a solution of saline and Chicago sky-blue dye was infused into each cannulae. Perfusion of each rat was carried out using a 9% phosphate-buffered saline solution followed by a 0.1 M 4% paraformaldehyde solution. The rat brains were removed and stored in 4% paraformaldehyde until slicing. Before slicing, the cerebellum was cut off and the brains were washed in distilled H 2 O. The brains were frozen on dry ice before being sliced on a cryostat (Cryocut 1800, Reichert-Jung) to prepare histological slides. The slides were studied under a microscope to verify and record the cannulae placements.
2.7. Data analysis The choice accuracy and response latency measures were assessed by a within subjects design analysis of variance. A P-value of less than 0.05 was considered significant for main effects. Comparisons were made between the effect of dizocilpine in the presence and absence of nicotine.
2.5. Drug treatment The rats were trained until asymptotic levels of choice accuracy had been reached, a minimum of 18 sessions, after which they were cannulated. Following the cannulation the animals began the drug studies. The drug treatments consisted of a subcutaneous injection of (2) nicotine ditartrate (saline vehicle) at doses of 0.0 (saline) or 0.4 mg / kg, and infusions of (1) dizocilpine maleate (saline vehicle) into the basolateral amygdala, via the cannulae, at doses of 0, 2, 6 and 18 mg per side. There being eight different combinations of these drugs, the animals were subjected to eight separate drug trials. Then 1 week after surgery the rats were given a trial run on the maze to determine any loss in performance due to the adverse effects of surgery. Following the post-surgery run, the rats were injected and infused with two test doses of artificial cerebrospinal fluid (ACSF) to accustom the animals to the procedure. Once these trial runs had been completed the drug treatments began. The nicotine injection was delivered subcutaneously,
3. Results
3.1. Histology Cannula placements were verified histologically. Many of the placements were found to have missed the specific target area of the basolateral amygdala. However they still fell into the range of the lateral amygdala, thus the data were still applicable for this study. Out of the 24 rats that underwent surgery, nine succeeded in having both (left and right sides) cannulae on target and had sufficient data to be analyzed. A further six rats had bilateral placements in the piriform cortex. These placements served as a control to the amygdaloid placements. Animals with cannulae in two different brain locations on opposite sides of the brain were discarded from the study. Fig. 1 shows two coronal diagrams of the brain at an anterior–posterior coordinate of 21.4 with the cannulae placements marked. Fig. 1a shows
H. May-Simera, E.D. Levin / Cognitive Brain Research 17 (2003) 475–483
479
mean working errors were higher among the animals with placements in the amygdala (9.5268.49) than the piriform cortex (7.1265.51). The 0.4-mg / kg dose of nicotine did not have a significant effect on working memory errors and did not reverse the dizocilpine induced impairment. The significant effect of dizocilpine was only seen with the high doses of 18 mg per side and only within the amygdala (P,0.005 without nicotine; P.0.05 with nicotine). Neither drug appeared to have any effect on working errors in the rats with piriform cortex cannulae. The reference memory errors showed a similar pattern of results to the working memory errors (Fig. 3). There appears to be an overall significant effect of dizocilpine on reference errors (P,0.005). There is a nearly significant effect of dizocilpine and the placement of the cannulae (P,0.08) although this is not as great an effect as for the working memory errors. The dizocilpine effect was only observed in the amygdalar cannulae (amygdala: P,0.005). The significant effect of dizocilpine with the amygdalar infusions was only seen among the highest dose of dizocilpine, 18 mg per side (P,0.0005 without nicotine; P,0.0005 with nicotine). Nicotine did not seem to have a significant effect in rats with either cannulae locations or on any of the dizocilpine doses. There was no significant interaction between the two drugs. As in the working memory errors, the animals with amygdalar placements had overall slightly more errors (5.4264.24) compared with rats with the piriform cortex placements (4.0262.07). The animals with cannulae placements in the piriform cortex showed virtually no effect from either drug.
3.3. Latency Fig. 1. Placement of cannulae for amygdala (a) and piriform cortex (b) [32].
the placements located within the amygdala and Fig. 1b the placements within the piriform cortex.
3.2. Choice accuracy Working memory was significantly affected by dizocilpine. The analysis of variance showed that overall doses of dizocilpine had a significant effect on working errors (P,0.005). There was a significant interaction of dizocilpine and the placement area (P,0.05). The dizocilpine effect was mainly due to the placements in the amygdala as there was a significant working memory impairment caused by dizocilpine infusion into the amygdala (P,0.005), but not into the piriform cortex (Fig. 2). There was no significant effect of nicotine administration in either the amygdala or the piriform cortex dizocilpine infusion studies. There were also no significant interactions of dizocilpine and nicotine for either brain placement. Regardless of the treatment the animals received, the
Response latency showed a significant main effect of dizocilpine (P,0.0001), the latencies decreasing as the dizocilpine dose increased (Fig. 4). Nicotine also had a significant effect on response latency (P,0.001), lowering the latencies at the 2 and 6 mg per side doses of dizocilpine. There was also a significant nicotine3 dizocilpine interaction (P,0.05). The latency reducing effects of nicotine and dizocilpine were less than additive. Tests of the simple main effects of nicotine at each dose of dizocilpine showed quite significant nicotine-induced reductions in latency with dizocilpine vehicle infusions (P, 0.0001) and low (2 mg per side) dizocilpine infusions (P,0.0005), but not quite significant nicotine effects on latency with 6 mg per side dose of dizocilpine (P50.08) and not nearly significant nicotine effects with 18 mg per side (P50.27). No interactions of drug effects on latency with brain infusion area were seen. Latency scores appeared to be generally higher with the placements within the piriform cortex than the amygdala. However, there were no significant latency differences between the rats with cannulae in the piriform cortex and those with cannulae in the amygdala with control infusions. The rats with the piriform cortex cannulae did have
480
H. May-Simera, E.D. Levin / Cognitive Brain Research 17 (2003) 475–483
Fig. 2. Dizocilpine infusion into the amygdala (n59) and piriform cortex (n56) and systemic nicotine: effects on working memory (mean6S.E.M.).
slightly higher mean latency scores (27.363.2 s per entry) than those with the amygdala cannulae (22.864.1 s per entry), but these were not nearly significantly different (P50.44).
4. Discussion Acute amygdalar infusions of the NMDA glutamate receptor antagonist dizocilpine induced dose-related work-
Fig. 3. Dizocilpine infusion into the amygdala (n59) and piriform cortex (n56) and systemic nicotine: effects on reference memory (mean6S.E.M.).
H. May-Simera, E.D. Levin / Cognitive Brain Research 17 (2003) 475–483
481
Fig. 4. Dizocilpine infusion into the amygdala (n59) and piriform cortex (n56) and systemic nicotine: effects on response latency (mean6S.E.M.).
ing and reference memory deficits in the radial-arm maze. Systemic nicotine was not seen to reverse these effects. It was expected that nicotine would reverse the memory deficit caused by dizocilpine infused into the amygdala. The 0.4-mg / kg nicotine ditartrate dose was sufficient to block the amnestic effect of systemic dizocilpine [16], which is why that dose was used in the current study. A lower dose of 0.2 mg / kg of nicotine ditartrate was not seen in a previous study to be effective in reversing the amnestic effects of systemic dizocilpine. Higher doses of nicotine have adverse effects that limit their usefulness in improving memory function. Dizocilpine infusions into the adjacent piriform cortex did not impair memory function, supporting the specificity of dizocilpine effects in the amygdala. Latency effects were seen with both drugs in both areas. Latencies were decreased with increasing doses of dizocilpine and with administration of nicotine. This study demonstrated the importance of NMDA glutamate systems in the amygdala for appetitively-motivated spatial memory performance. The currently observed effects of dizocilpine in the amygdala contrast with previously observed effects in the ventral hippocampus. In a previous study in our laboratory, the same experiments were carried out on animals with cannulae into the ventral hippocampus [18]. In that study none of the three doses of dizocilpine caused significant working memory impairments when administered alone, but all three caused significant working memory deficits when administered in conjunction with systemic nicotine.
No effect of either treatment, alone or together, was seen with reference memory or response latency. The variation in results indicates the different roles the brain regions play in cognitive function. The finding that ventral hippocampal dizocilpine caused systemic nicotine to actually impair memory points to the importance of ventral hippocampal NMDA receptors for the positive effects of nicotine on memory. Blocking these receptors unveiled the expression of other nicotinic actions that impair working memory function. In the previous study by Levin et al. [16], it was proposed that nicotine may have counteracted the dizocilpine induced memory deficits through its action in stimulating glutamate release [23]. Because nicotine and NMDA receptors are similar in structure, cross-reactivity of ligands for these receptors may also have been involved in the observed interactions. Our findings in the present study complicate the situation, as nicotine was shown not to have an effect on choice accuracy. This could have been due to the method of administration. The direct infusion of dizocilpine into the brain might have been too intense and localized, so that the systemic nicotine could not exert its effects. Infusion of dizocilpine also enables the delivery of the drug to a precise location within the brain targeting the NMDA receptors in a specific region of the brain, not just all receptor sites. In the current study the animals were pretrained so that the drug effects on memory performance would be assessed during the stable asymptotic post acquisition phase
482
H. May-Simera, E.D. Levin / Cognitive Brain Research 17 (2003) 475–483
of performance. The within-subjects counterbalanced design was used so that possible carryover drug effects would not be confounded with order of treatment dose. At least 2 days were interposed between drug doses to minimize carryover effects. The piriform cortex infusions served as a useful control. Inasmuch as the memory effects were specific to the amygdala and not the piriform cortex, widespread diffusion of dizocilpine from the infusion site in the amygdala did not appear to be the cause of the observed effects on memory function. Also the larger volume of the dye infused for verification (0.5 ml per side) did not appear to invade nuclei outside the target areas. Tissue damage from the infusions was not seen to be a problem. Even subtle tissue damage with repeated drug infusion would not have been confounded with the observed drug effects because the drugs were administered in a repeated measures counterbalanced design so that the order of drug infusion was not confounded with dose. Previously it had been found that systemic nicotine administration improved working memory in the radial arm maze [11]. The present study showed that infusion of the NMDA antagonist dizocilpine had a significant effect in the amygdala but not the piriform cortex on both working and reference memory errors. Nicotine (0.4 mg / kg) did not reverse memory impairment caused by dizocilpine. Latency was significantly decreased with the rats displaying an increased level of activity with increasing dose of dizocilpine infused into both brain regions. Nicotine effected latency. As previously demonstrated by McLamb et al. [25] and Levin et al. [16], the current study found that dizocilpine impaired memory. However, in contrast to our earlier study [16], nicotine attenuated the dizocilpine induced memory deficit. The difference in findings could be due to the type of drug delivery involved in each study. In the present study, dizocilpine was infused directly into the amygdala rather than being injected subcutaneously [16]. Systemic dizocilpine has a global effect on all parts of the central and peripheral nervous system, whereas locally infused dizocilpine is specific to one area in the brain. This suggests that extra-amygdala NMDA glutamate receptors may have different effects to those in the amygdala. It is also important to note the dose of nicotine administered. A variety of studies have shown that acute treatment with nicotine or nicotinic agonists can improve working memory function in the radial-arm maze and other tasks [3,12,13]. In six separate studies, it was found that a single dose of 0.2 mg / kg nicotine injected subcutaneously significantly improved working memory [15]. Improvements could also be seen at lower and higher doses (0.1 and 0.4 mg / kg), but with the typical inverted U-shaped function for memory-enhancing drugs, it was found that 0.2 mg / kg nicotine was the most effective dose. For this particular study 0.4 mg / kg was used as the previous study by Levin et al. [16] had found that particular dose to be most
effective at reversing the effects of dizocilpine. It could be that different doses of nicotine are needed to counteract the effects of locally infused dizocilpine. The current study found that blockade of the lateral and basolateral amygdalar NMDA glutamate receptors with dizocilpine significantly impaired working and reference memory. This effect was regionally specific with the same doses infused into the adjacent piriform cortex having no discernible effect. Previously, we have found that these doses given into the ventral hippocampus did not by themselves impair memory performance on the radial-arm maze [18]. Nicotine co-administration did not significantly interact with the effects of dizocilpine infused into the amygdala. This also was regionally selective with a substantial nicotine-induced memory impairment seen with the same dose range of dizocilpine infused into the ventral hippocampus. Nicotinic receptors in the amygdala are important for memory function as shown by our work with local nicotinic antagonist infusions [1]. This is similar to what we have found in the hippocampus [17]. However, the current study shows that nicotinic interactions with NMDA glutamate systems in the hippocampus and amygdala differ substantially.
Acknowledgements The authors thank Channelle Christopher for her invaluable assistance in this study. This research was supported in part by a grant from the Philip Morris External Research Foundation.
References [1] N. Addy, A. Nakajama, E. Levin, Nicotinic mechanisms of memory: effects of acute local DHbE and MLA infusions in the basolateral amygdala, Cogn. Brain Res. 16 (2003) 51–57. [2] B.J. Cole, M. Klewer, G.H. Jones, D.N. Stephens, Contrasting effects of the competitive NMDA antagonist CPP and the noncompetitive NMDA antagonist MK-801 on performance of an operant delayed matching to position task in rats, Psychopharmacology 111 (1993) 465–471. [3] M.W. Decker, J.D. Brioni, A.W. Bannon, S.P. Arneric, Diversity of neuronal nicotinic acetylcholine receptors: lessons from behavior and implications for CNS therapeutics—minireview, Life Sci. 56 (1995) 545–570. [4] W. Falls, M. Miserendino, M. Davis, Extinction of fear-potentiated startle: blockade by infusion of an NMDA antagonist into the amygdala, J. Neurosci. 12 (1992) 854–863. [5] M. Fanselow, J. Kim, Acquisition of contextual Pavlovian fear conditioning is blocked by application of an NMDA receptor antagonist D,L-2-amino-5-phosphonovaleric acid to the basolateral amygdala, Behav. Neurosci. 108 (1994) 210–212. [6] N. Hiroi, N. White, The lateral nucleus of the amygdala mediates expression of the amphetamine-produced conditioned place preference, J. Neurosci. 11 (1991) 2107–2116. [7] I. Izquierdo, Role of NMDA receptors in memory, Trends Pharmacol. Sci. 12 (1991) 128–129. [8] R.P. Kesner, J.M. Williams, Memory for magnitude of rein-
H. May-Simera, E.D. Levin / Cognitive Brain Research 17 (2003) 475–483
[9]
[10] [11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
forcement: dissociation between the amygdala and hippocampus, Neurobiol. Learn. Mem. 64 (1995) 237–244. M. Kim, J.L. McGaugh, Microinfusion of an N-methyl-D-aspartate antagonist impairs avoidance learning in rats, Soc. Neurosci. Abstr. 16 (1990) 767. E.D. Levin, Nicotinic systems and cognitive function, Psychopharmacology 108 (1992) 417–431. E.D. Levin, The role of nicotinic acetylcholine receptors in cognitive function, in: F. Clementi, C. Gotti, D. Fornasari (Eds.), Handbook of Experimental Pharmacology: Neuronal Nicotinic Receptors, Springer, New York, 2000, pp. 587–602. E.D. Levin, A.H. Rezvani, The promise of nicotinic treatment for cognitive dysfunction, Curr. Drug Targets—CNS Neurol. Disord. 1 (2002) 423–431. E.D. Levin, D. Torry, Nicotine effects on memory performance, in: P.B.S. Clarke, M. Quik, K. Thurau, F. Adlkofer (Eds.), International Symposium on Nicotine: The Effects of Nicotine on Biological ¨ Systems II, Birkhauser, Boston, 1995, pp. 329–336. E.D. Levin, P. Kim, R. Meray, Chronic nicotine effects on working and reference memory in the 16-arm radial maze: interactions with D1 agonist and antagonist drugs, Psychopharmacology 127 (1996) 25–30. E.D. Levin, S. Kaplan, A. Boardman, Acute nicotine interactions with nicotinic and muscarinic antagonists: working and reference memory effects in the 16-arm radial maze, Behav. Pharmacol. 8 (1997) 236–242. E.D. Levin, C. Bettegowda, T. Weaver, N.C. Christopher, Nicotine– dizocilpine interactions and working and reference memory performance of rats in the radial-arm maze, Pharmacol. Biochem. Behav. 61 (1998) 335–340. E.D. Levin, A. Bradley, N. Addy, N. Sigurani, Hippocampal alpha 7 and alpha 4 beta 2 nicotinic receptors and working memory, Neuroscience 109 (2002) 757–765. E.D. Levin, D. Sledge, A. Baruah, N. Addy, Hippocampal NMDA blockade and nicotinic effects on memory function, Brain Res. Bull. (2003) (in press). K.C. Liang, M. Davis, Intra-amygdala injection of N-methyl-Daspartate receptor antagonists impairs memory in an inhibitory avoidance task, Soc. Neurosci. Abstr. 10 (1990) 767. S. Maren, G. Aharonov, D.L. Stote, M.S. Fanselow, N-methyl-Daspartate receptors in the basolateral amygdala are required for both acquisition and expression of conditional fear in rats, Behav. Neurosci. 110 (1996) 1365–1374. J.L. Martinez, K.C. Liang, A. Oscos, Amnesia induced by stimulation of the amygdala is attenuated by hexamethonium, Psychopharmacology 81 (1983) 310–314. J. McGaugh, I. Introini-Collison, A. Nagahara, L. Cahill, Involvement of the amygdaloid complex in neuromodulatory influences on memory storage, Neurosci. Biochem. Rev. 14 (1990) 425–431. D.S. McGehee, M.J.S. Heath, S. Gelber, P. Devay, L.W. Role, Nicotine enhancement of fast excitatory synaptic transmission in CNS by presynaptic receptors, Science 269 (1995) 1692–1696. C.K. McIntyre, M.E. Ragozzino, P.E. Gold, Intra-amygdala infusions of scopolamine impair performance on a conditioned place preference task but not a spatial radial maze task, Behav. Brain Res. 95 (1998) 219–226. R. McLamb, L. Williams, K. Nanry, W. Wilson, H. Tilson, MK-801
[26]
[27]
[28]
[29]
[30]
[31] [32] [33]
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
483
impedes the acquisition of a spatial memory task in rats, Pharmacol. Biochem. Behav. 37 (1990) 41–45. M.J. Miserendino, C.B. Sananes, K.R. Melia, M. Davis, Blocking of acquisition but not expression of conditioned fear-potentiated startle by NMDA antagonists in the amygdala, Nature 345 (1990) 716– 718. D.T. Monaghan, C.W. Cotman, Distribution of N-methyl-D-aspartatesensitive L-[ 3 H]glutamate-binding sites in rat brain, J. Neurosci. 5 (1985) 2909–2919. M. Ohno, T. Yamamoto, S. Watanabe, Involvement of cholinergic mechanisms in impairment of working memory in rats following basolateral amygdaloid lesions, Neuropharmacology 31 (1992) 915– 922. M.B. Parent, J.L. McGaugh, Posttraining infusion of lidocaine into the amygdala basolateral complex impairs retention of inhibitory avoidance training, Brain Res. 661 (1994) 97–103. M.A. Peinado-Manzano, The role of the amygdala and the hippocampus in working memory for spatial and non-spatial information, Behav. Brain Res. 38 (1990) 117–134. M.A. Peinado-Manzano, Amygdala, hippocampus and associative memory in rats, Behav. Brain Res. 61 (1994) 175–190. L.J. Pellegrino, A.S. Pellegrino, A.J. Cushman, A Stereotaxic Atlas of the Rat Brain, 2nd Edition, Plenum, New York, 1979. A.E. Power, B. Roozendaal, J.L. McGaugh, Glucocorticoid enhancement of memory consolidation in the rat is blocked by muscarinic receptor antagonism in the basolateral amygdala, Eur. J. Neurosci. 12 (2000) 3481–3487. P. Riekkinen Jr., M. Riekkinen, J. Sirvio, Cholinergic drugs regulate passive avoidance performance via the amygdala, J. Pharmacol. Exp. Ther. 267 (1993) 1484–1492. O. Salminen, T. Seppa, H. Gaddnas, L. Ahtee, The effects of acute nicotine on the metabolism of dopamine and the expression of Fos protein in striatal and limbic brain areas of rats during chronic nicotine infusion and its withdrawal, J. Neurosci. 19 (1999) 8145– 8151. M.L. Shapiro, C. O’Connor, N-methyl-D-aspartate receptor antagonist MK-801 and spatial memory representation: working memory is impaired in an unfamiliar environment but not in a familiar environment, Behav. Neurosci. 106 (1992) 604–612. M.L. Shapiro, Z. Caramanos, NMDA antagonist MK-801 impairs acquisition but not performance of spatial working and reference memory, Psychobiology 18 (1990) 231–243. R. Wan, K. Pang, D. Olton, Hippocampal and amygdaloid involvement in non-spatial and spatial working memory in rats: effects of delay and interference, Behav. Neurosci. 108 (1994) 866–882. L. Ward, S.E. Mason, W.C. Abraham, Effects of the NMDA antagonists CPP and MK-801 on radial-arm maze performance in rats, Pharmacol. Biochem. Behav. 35 (1990) 785–790. K. Watanabe, K. Hashimoto, T. Nishimura, K. Tsunashima, Y. Minabe, Expression of Fos protein in rat brain following administration of a nicotinic acetylcholine receptor agonist epibatidine, Brain Res. 797 (1998) 135–142. S. Wonnacott, J. Irons, C. Rapier, B. Thorne, G.G. Lunt, Presynaptic modulation of transmitter release by nicotinic receptors, in: A. Nordberg, K. Fuxe, B. Holmstedt, A. Sundwall (Eds.), Progress in Brain Research, Vol. 79, Elsevier, New York, NY, 1989, pp. 157– 163.