Effects of memantine and donepezil on cortical and hippocampal acetylcholine levels and object recognition memory in rats

Neuropharmacology 61 (2011) 891e899

Contents lists available at ScienceDirect

Neuropharmacology journal homepage: www.elsevier.com/locate/neuropharm

Effects of memantine and donepezil on cortical and hippocampal acetylcholine levels and object recognition memory in rats Jouni Ihalainen a, *,1, Timo Sarajärvi b,1, Doug Rasmusson c,1, Susanna Kemppainen a, Pekka Keski-Rahkonen d, Marko Lehtonen d, Pradeep K. Banerjee e, Kazue Semba f, Heikki Tanila a, g a

A. I. Virtanen Institute, Department of Neurobiology, University of Eastern Finland, P.O. Box 1627, Yliopistonranta 1 C, 70211 Kuopio, Finland Department of Clinical Medicine/Neurology, University of Eastern Finland, P.O. Box 1627, Yliopistonranta 1 C, 70211 Kuopio, Finland Department of Physiology and Biophysics, Dalhousie University, Halifax, NS B3H 1X5, Canada d School of Pharmacy, University of Eastern Finland, P.O. Box 1627, Yliopistonranta 1 C, 70211 Kuopio, Finland e Forest Laboratories Inc., Harborside Financial Center, Plaza V, 19th Floor, Jersey City, NJ 07311, USA f Department of Anatomy and Neurobiology, Dalhousie University, Halifax, NS B3H 1X5, Canada g Department of Neurology, Kuopio University Hospital, P.O. Box 1627, Neulaniementie 2, 70211 Kuopio, Finland b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 25 January 2011 Received in revised form 6 June 2011 Accepted 8 June 2011

This preclinical study investigated the ability of memantine (MEM) to stimulate brain acetylcholine (ACh) release, potentially acting synergistically with donepezil (DON, an acetylcholinesterase inhibitor). Acute systemic administration of either MEM or DON to anesthetized rats caused dose-dependent increases of ACh levels in neocortex and hippocampus, and the combination of MEM (5 mg/kg) and DON (0.5 mg/kg) produced significantly greater increases than either drug alone. To determine whether ACh release correlated with cognitive improvement, rats with partial fimbria-fornix (FF) lesions were treated with acute or chronic MEM or DON. Acute MEM treatment significantly elevated baseline hippocampal ACh release but did not significantly improve task performance on a delayed non-match-to-sample (DNMS) task, whereas chronic MEM treatment significantly improved DNMS performance but only marginally elevated baseline ACh levels. Acute or chronic treatment with DON (in the presence of neostigmine to allow ACh collection) did not significantly improve DNMS performance or alter ACh release. In order to investigate the effect of adding MEM to ongoing DON therapy, lesioned rats pretreated with DON for 3 weeks were given a single intraperitoneal dose of MEM. MEM significantly elevated baseline hippocampal ACh levels, but did not significantly improve DNMS task scores compared to chronic DON-treated animals. These data indicate that MEM, in addition to acting as an NMDA receptor antagonist, can also augment ACh release; however, in this preclinical model, increased ACh levels did not directly correlate with improved cognitive performance. Ó 2011 Elsevier Ltd. All rights reserved.

Keywords: Acetylcholine Alzheimer’s disease Hippocampus In vivo microdialysis Memantine Recognition memory

1. Introduction

Abbreviations: Ab, amyloid-b; ACh, acetylcholine; AChE, acetylcholinesterase; AChEIs, AChE inhibitors; aCSF, artificial cerebrospinal fluid; AD, Alzheimer’s disease; ANOVA, analysis of variance; DNMS, delayed non-match to sample; DON, donepezil; FF, fimbria-fornix; i.p., intraperitoneal; LCeMS/MS, liquid chromatographyemass spectrometry; KW, KruskaleWallis; MEM, memantine; NMDA, N-methyl-D-aspartate. * Corresponding author. Tel.: þ358 50 3051720; fax: þ358 17 162048. E-mail addresses: Jouni.Ihalainen@uef.fi (J. Ihalainen), timo.sarajarvi@uef.fi (T. Sarajärvi), [email protected] (D. Rasmusson), susanna.kemppainen@uef.fi (S. Kemppainen), pekka.keski-rahkonen@uef.fi (P. Keski-Rahkonen), marko. lehtonen@uef.fi (M. Lehtonen), [email protected] (P.K. Banerjee), [email protected] (K. Semba), Heikki.Tanila@uef.fi (H. Tanila). 1 Jouni Ihalainen, Timo Sarajärvi and Doug Rasmusson contributed equally to this work. 0028-3908/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.neuropharm.2011.06.008

Memantine (MEM), an uncompetitive antagonist (open-channel blocker) of N-methyl-D-aspartate (NMDA) receptors, is used clinically for the treatment of Alzheimer’s disease (AD). In several randomized, placebo-controlled clinical trials involving patients with moderate to severe AD, MEM has been shown to provide global, cognitive, functional, and behavioral benefits (Reisberg et al., 2003; Tariot et al., 2004; Winblad and Poritis, 1999). Because MEM acts via a mechanism that does not involve inhibition of acetylcholinesterase (AChE) it is commonly administered in combination with AChE inhibitors (AChEIs) in patients who are in moderate to severe stages of the disease. Support for this practice comes from two large randomized, placebo-controlled studies in patients with moderate to severe AD, which demonstrated that

892

J. Ihalainen et al. / Neuropharmacology 61 (2011) 891e899

adding MEM to a stable AChEI regimen resulted in significantly better outcomes than AChEIs alone on multiple clinical measures (Grossberg et al., 2008; Tariot et al., 2004). Despite the favorable clinical outcomes of such studies, relatively a little progress has been made in addressing the mechanism behind this combined effect. In vitro measurements suggest that the combined effect does not likely occur on the level of AChE inhibition, since MEM has been shown to have no effect on the activities of the AChEIs donepezil (DON) and galantamine (Wenk et al., 2000). Although MEM and DON have independently demonstrated efficacy in improving memory in animal models, no preclinical study has thus far shown a cooperative effect from the co-administration of both drugs (Wise and Lichtman, 2007; Yamada et al., 2005). One mechanism by which MEM could clinically augment the effect of DON is through the enhancement of acetylcholine (ACh) release, an effect that would lead to increased synaptic concentrations of ACh and could therefore potentiate the effect of AChEIs. Systemic or intracerebroventricular administration of high-affinity NMDA receptor antagonists, such as MK-801 (Hasegawa et al., 1993) and CPP (Giovannini et al., 1994, 1997; Hutson and Hogg, 1996), has previously been shown to increase ACh release in several brain regions. This possible mechanism is supported by a microdialysis study in freely moving rats, which found increased ACh release in the ventral tegmental area and nucleus accumbens after acute subcutaneous injection of MEM at 20 mg/kg; however, this dose is several-fold higher than the concentration used clinically (Shearman et al., 2006). In the present study, we tested whether administration of MEM at a clinically relevant dose enhanced ACh release in the hippocampus and cerebral cortex of anesthetized or freely moving rats. To ensure sufficient detection of changes in ACh levels and to equalize degradation of ACh between treatments, we used the AChEI neostigmine in the dialysate. This prevented us from seeing the already well-established increase in extracellular ACh levels by systemically administered AChEI DON, but allowed us to see potential interactions between MEM and DON on ACh release. We also assessed whether clinically relevant doses of MEM and DON, in combination, improved cognition in fimbria-fornix (FF)-lesioned rats in a short-term object recognition task. The model and the task were chosen based on previous evidence that FF-lesion in rats impairs short-term object recognition similarly to a local microinfusion of scopolamine into the hippocampus, whereas local microinfusion of physostigmine enhances object recognition (Hunsaker et al., 2007). 2. Materials and methods 2.1. Animals A total of 89 male Wistar rats (225e350 g) were anesthetized and used in the initial microdialysis studies, while 61 male Wistar rats (32 weeks old; 400e550 g at the beginning of the experiment) were used for the behavioral study. Rats in the behavioral study were initially food-deprived until they reached 80e85% of their free-feeding weight, but were allowed access to fresh water ad libitum. Rats were housed individually in a controlled environment (illuminated from 7:00 AMe7:00 PM, temperature 22  1  C, humidity 50%e60%). The experiments were conducted according to the Declaration of Helsinki and the guidelines of the Council of Europe (Directive 86/609), the Finnish and the Canadian Councils of Animal Care and the NIH Guide for the Care and Use of Laboratory Animals, and approved by the State Provincial Office of Eastern Finland and by the Dalhousie University Committee on Laboratory Animals. 2.2. In vivo microdialysis in anesthetized rats To test the effects of MEM and DON on ACh levels in the neocortex and hippocampus, we first administered the drugs in anesthetized rats using a single intraperitoneal (i.p.) injection, and monitored ACh levels using microdialysis probes implanted in each brain region. We then tested whether the drug effects occur at or

near the cholinergic terminals by administering the drugs locally within the hippocampus, using reverse dialysis through the hippocampal microdialysis probe. 2.2.1. Surgery Rats were anesthetized with urethane (SigmaeAldrich, St. Louis, MO; 1.6 mg/kg, i.p.). This constant level of anesthesia over several hours provided a “behavioral clamp”, which enabled a direct assessment of the drugs’ effects on ACh levels. After placing the animal in a stereotaxic frame, a microdialysis probe (2 mm membrane length, 0.5 mm outer diameter, 20,000 MW cutoff; CMA, Sweden) was implanted into the left dorsal hippocampus (4.1 mm posterior to bregma, 2.5 mm lateral to midline, and 3.5 mm ventral to the cortical surface). For studies that used peripheral injections of MEM and DON, a second, identical probe was implanted into the left somatosensory cortex (1.3 posterior, 2.0 lateral, and 2.0 ventral). 2.2.2. Treatment A total of 53 rats (6 or 7 animals/group) received i.p. drug injections of MEM (Forest Research Institute, Jersey City, NJ) at 2.5, 5, or 10 mg/kg, DON (SigmaeAldrich, Saint Louis, MO) at 0.25, 0.5 or 1.0 mg/kg, or a combination of both drugs (MEM: 5 mg/kg; DON: 0.5 mg/kg). An additional 36 animals received direct hippocampal administration of the drugs dissolved in the perfusate (see below) via reverse microdialysis. Seven of these rats received a low concentration (1 mM) of one of the two drugs, which produced no change in ACh release. Consequently, the remaining 29 animals received a higher concentration (100 mM) of MEM alone, DON alone, or MEM plus DON. 2.2.3. In vivo microdialysis For both systemic and local administration experiments, the microdialysis probes were perfused continuously at a flow rate of 2 ml/min with artificial cerebrospinal fluid (aCSF; 3 mM KCl, 125 mM NaCl, 1.3 mM CaCl2, 1 mM MgSO4), to which 5 mM neostigmine (SigmaeAldrich, Saint Louis, MO), an AChEI, was added to prevent ACh hydrolysis. After 1 h of stabilization, a series of ten 15-min microdialysis samples were collected. For peripheral administration, drugs were injected i.p. at the beginning of sample 6; control animals were given an injection of the vehicle (saline). For the reverse dialysis experiments, the dialysis solution was replaced throughout sample 6 with one containing MEM alone, DON alone, or MEM plus DON. The efficiency of the microdialysis probe was tested after each experiment by placing it in a standard solution of ACh (10 pM) for 15 min and measuring the ACh content appearing in the perfusate, to ensure that the in vitro recovery of the probe was greater than 5%. The dialysis samples were analyzed for ACh content using high performance liquid chromatography with electrochemical detection (Waters; Missasauga, Ontario, Canada), according to the procedure used by Materi and Semba (2001). 2.2.4. Data analysis and statistics The mean ACh concentration measured in the first 2 samples was considered the baseline ACh level for each animal. The effect of each drug on spontaneous ACh release was determined using sample 6, immediately after peripheral drug injection, or during reverse dialysis. Statistical evaluation used the 5% level of significance. Student’s t-test was used when appropriate, but when the assumption of equal variance between groups was not met, a nonparametric KruskaleWallis (KW) test was used. 2.3. In vivo dialysis in freely moving rats with fimbria-fornix lesions 2.3.1. Overview The time course of experimental procedures is summarized in Fig. 1A. To investigate the cognitive effects associated with changes in ACh levels after drug treatment, a DNMS short-term object recognition memory task was developed (Fig. 1B; see below). The training phase continued for each animal until it reached a criterion of 80% or more correct choices on three consecutive testing days (prelesion performance), then sham or FF-lesioning was performed to impair afferent cholinergic innervation of the septal (dorsal) hippocampus. Chronic microdialysis probes were implanted to allow in vivo assessment of extracellular ACh levels during subsequent DNMS testing. The rats were divided into 5 treatment groups (Table 1), treated for three weeks with MEM, DON, or placebo, and DNMS task performance was again observed (DNMS Test 1; Fig. 1). Finally, rats that had been treated chronically with MEM were given an acute dose of DON, while DON-treated rats were given an acute dose of MEM, to test the effects of drug combinations on task performance (DNMS Test 2). After a 5-day washout of the acute treatment (while continuing chronic treatment), the microdialysis probes were connected to a collection device, and DNMS Tests 1 and 2 were repeated under the same conditions, to monitor ACh release during DNMS task performance. 2.3.2. Fimbria-fornix lesioning After DNMS training (see below), each rat was randomly assigned to sustain either a sham lesion or a partial fimbria-fornix (FF) lesion (Table 1). The rats were anaesthetized with a 3:1 mixture (2 mL/kg, i.p.) of ketamine (KetaminolÒ/KetalarÒ, 50 mg/mL, Intervet International B.V, Boxmeer, the Netherlands) and medetomidine (DomitorÒ, 1 mg/mL, Orion Pharma, Turku, Finland). Each animal was then placed in

J. Ihalainen et al. / Neuropharmacology 61 (2011) 891e899

Fig. 1. A, Timeline of the procedures in the behavioral experiments. B, Schematic representation of the DNMS object recognition task protocol. Each trial consists of two phases. Sample phase (A): the sample object covers the middle food well (A1), and the rat displaces the object to receive a food reward (A2). Choice phase (B): one of the lateral food wells (randomly either left or right side) is covered with a novel object, while the other side is covered with an identical copy of the sample object (B1). To receive a food reward, the rat must displace the novel object (B2). DNMS indicates delayed non-match to sample; DON, donepezil; MEM, memantine; PBO, placebo. a stereotaxic frame. The lesion groups received bilateral electrolytic lesions, produced by passing 500 mA anodal current for 40 s through a tungsten electrode (0.1 mm in diameter, un-insulated about 0.75 mm at the tip of the electrode). The lesion coordinates were 1.5 mm posterior to bregma; 0.7 mm and 1.6 mm lateral to midline; 4.4 mm and 4.5 mm below dura for fornix and fimbria, respectively (Fig. 2B, D). Animals in the sham-lesioned group were treated identically, except that the electrode tip was only lowered 2.0 mm below dura and no current was applied (Fig. 2A, C). A microdialysis guide cannula (MAB 4.15.IC, Microbiotech/se AB, Stockholm, Sweden) was implanted in all animals, just above the right dorsal hippocampus (mm from bregma: AP e 4.4 mm; L e 2.5 mm; V e 1.6 mm). Atipamezole hydrochloride (AntisedanÒ vet, Orion Pharma, Turku, Finland, 0.5 mg/kg, s.c.) was used as an anti-sedative agent and carbiprofen (RimadylÒ, Vericore Ltd, Dundee, UK, 5 mg/kg, s.c.) was used for post-operative pain relief. Saline (5 mL, i.p.) was injected after operation to maintain fluid balance.

Table 1 Treatment groups. Group (n)

Lesion Type

Chronic Treatment

Acute Treatment

A (15) B (12) C (15) D (9) E (10)

Sham FF FF FF FF

Placebo MEM DON Placebo Placebo

MEM DON MEM MEM DON

DON indicates donepezil; FF, fimbria-fornix lesion; MEM, memantine.

893

Fig. 2. AeB, Cresyl violet staining showing the fimbria-fornix area in sham-operated controls (A) and rats with partial fimbria fornix lesion (B). CeD, Density of acetylcholinesterase-positive neurites in the hippocampus of sham-operated (C) and partial fimbria fornix lesioned (D) rats. The dashed rectangle illustrates the location of the microdialysis cannula in the right hippocampus. Scale bar ¼ 500 mm. E, AChE staining intensity in the hippocampus of sham-operated and lesioned rats. Staining intensities were measured by densitometry and expressed as relative optic density (%) compared to the sham group. ***p < 0.001, compared to the sham-operated control group (t-test).

2.3.3. Treatment Chronic administration of MEM and DON were initiated only after lesioning, to focus on the symptomatic effects of the drugs and to minimize any potential neuroprotective effects of the two drugs (Akasofu et al., 2008; Heim and Sontag, 1995). Drug doses were selected based on previous studies (Barnes et al., 1996; Kosasa et al., 1999). The chronic MEM dose was 30 mg/kg/day, while the acute (i.p.) dose was 5 mg/kg. The chronic and acute (i.p.) doses of DON were identical (2.5 mg/kg/day). Chronic drug treatment (via drinking water) was administered for 3 weeks, beginning two days after the surgery; the sham group (A) and the two FF lesion groups randomized to placebo (D and E) received drinking water without drugs (Table 1). The bottles were changed twice per week. The daily drug intake was controlled by adjusting the concentration in the drinking water according to individual water consumption, measured over a 24 h period before the drug administration, and again after every change of water bottles. 2.3.4. In vivo microdialysis Twenty-three days post-surgery (the day before the first DNMS/microdialysis experiment; see below), each rat was individually placed into a test cage, and the microdialysis probe (active membrane length 2 mm, diameter 0.20 mm, MAB 4.15.2.Cu, Microbiotech/se AB, Stockholm, Sweden) was connected using flexible PEEK tubing (o.d. 0.51 mm; i.d. 0.13 mm; Upchurch, Oak Harbour, WA, USA) to a rotating liquid swivel (375/D/22QM, Instech Laboratories Inc, PA, USA), and inserted through the guide cannula. The hippocampus was continuously perfused with Ringer’s solution (145 mmol/l NaCl, 2.7 mmol/l KCl, 1.2 mmol/l CaCl2, and

894

J. Ihalainen et al. / Neuropharmacology 61 (2011) 891e899

1.0 mmol/l MgCl2) at a rate of 0.5 ml/min for 18 h (CMA/100 Microinjection Pump, Solna, Sweden). The next day, the perfusion speed was increased to 2.0 ml/min and neostigmine was added to the perfusion fluid (final concentration: 750 nM); perfusion continued for 2.5 h prior to sampling. The ACh sampling included 60 min of baseline activity, followed by a 10 min exploration of the DNMS hole-board without objects (Table 2). After a 20 min idle period, the DNMS task (see below) was performed multiple times over a 10 min period. After another 20 min idle period, the DNMS task was repeated over another 10 min interval. The rats completed on average 12 trials during each 10-min testing session. The total sampling time was 2 h 30 min and included 12 samples (each 10 min, 20 ml; Table 2). At the end of DNMS/microdialysis Test 1, the perfusion fluid was switched to normal Ringer’s solution and the flow was decreased to a rate of 0.5 ml/min. The next day, DNMS/microdialysis Test 2 was performed in a manner identical to Test 1, except that the animals were injected either with MEM (groups A, C and D) or DON (groups B and E), 1 h prior to the microdialysis experiment (see Table 1). 2.3.5. Determination of extracellular ACh levels The analytical method used for determining the concentration of acetylcholine in the microdialysis samples was based on liquid chromatographyemass spectrometry (LCeMS/MS) with a special ionization technique, described previously (KeskiRahkonen et al., 2007). The detection limit of this technique is 0.15 nM, with linearity maintained over the concentration range of 0.15e73 nM. In our studies, a microdialysis sample size of 15 ml was sufficient for detection, and stable-isotope-labeled acetylcholine was included in the analysis as an internal standard. The ACh levels were analyzed at the Department of Pharmaceutical Chemistry, University of Kuopio, Finland. 2.3.6. Histology At the conclusion of testing, the rats were anesthetized with pentobarbiturateechloralhydrate cocktail (60 mg/kg, i.p.), then perfused with heparinized 0.9% saline followed by 4% paraformaldehyde in 0.1 M Na-phosphate buffer (12 mL/min; each 10 min). The brains were rapidly removed, cryoprotected in 30% sucrose for 48 h, and stored at 80  C. Coronal sections (35 mm) were cut on a freezing sliding microtome. The sections of fimbria-fornix (lesion) area were mounted on gelatincoated slides and stained with iron (Prussian blue reaction) and cresyl violet (Fig. 2A, B). The second series of hippocampal sections (near the cannula location) were histochemically stained for AChE (Hedreen et al., 1985) (Fig. 2C, D). The sections were digitized using an Olympus DP50 digital camera with low magnification (2), and the images were converted to 8-bit gray scale images using Adobe Photoshop 7.0. The density of AChE staining in CA2/CA3 region of dorsal hippocampus (mm from bregma: AP -4.4 mm) was measured using the ScionImage (NIH) program. Densitometric measurements in unstained regions of corpus callosum were also performed, to provide an internal control for background staining. The accurate placement of microdialysis cannulae, the size of the FF-lesion and AChE staining density were verified in all animals. 2.4. Data analysis and statistics The staining intensity of AChE-positive fibers in the hippocampus was compared between sham and FF-lesioned group using Student’s t-test. The five treatment groups were compared on the pre-lesion DNMS task performance using one-way analysis of variance (ANOVA). Pre- and post-operative task performance was analyzed using ANOVA for repeated measures, with the task phase as the withinsubject factor and the lesion as the between-subjects factor. The drug effects on the post-lesion DNMS task performance or ACh levels were also analyzed with ANOVA for repeated measures, with the test day as within-subject factor and both chronic and acute drug treatment as between-subject factors. The group differences were analyzed using one-way ANOVA followed by Dunnett’s post-hoc test. 2.5. Delayed non-match-to-sample (DNMS) object recognition task Instead of spontaneous object exploration, we employed a food-motivated DNMS task to tax object recognition to ensure stable performance over large

Table 2 Microdialysis sampling schedule during DNMSa performance. Sample number Sampling time 1e3 4 5e6 7 8e9 10 11e12

3 1 2 1 2 1 2

      

10 10 10 10 10 10 10

minb (60 min total) min min (20 min total) min min (20 min total) min min (20 min total)

Activity Baseline (no task) Exploration of empty hole-board No task DNMS performance No task DNMS performance No task

DNMS indicates delayed non-match to sample object recognition task. a See below for description of DNMS Task procedure. b Every second sample was collected at baseline, totaling 3 samples over 60 min.

number of trial per day to allow simultaneous microdialysis. The details of the DNMS procedure have been described previously (Ihalainen et al., 2010); in brief: 2.5.1. Apparatus A rectangular box (41  27  35 cm) containing a start and goal area, separated by a sliding door; the goal area included six feeding wells in the floor board, evenly spaced in two parallel rows (Fig. 1B). 2.5.2. Pre-training During the habituation period, rats were trained to find food rewards in the goal area upon opening of the sliding door. Over the testing period, the reward was progressively concealed by one of several types of small objects (plastic, wood, steel, glass or porcelain), which the animals learned to displace to collect the reward (3e5 days). 2.5.3. Training Pre-training was followed by the DNMS protocol, consisting of a sample phase and a choice phase (20 trials per day). During the sample phase, the animals displaced a randomly selected object from a central well to receive the reward (Fig. 1BA1e2). In the choice phase (20e25 s later), the animals learned to displace a novel object from a lateral well to receive the reward, rather than displacing the familiar object from the opposite lateral well (Fig. 1B-B1e2). The training phase continued for an average of 5.5 weeks, until the animals reached a pre-lesion performance level of 80% or more correct choices on three consecutive days (Pre-Test performance). The total number of different objects during training and testing sessions was 22. The same objects were presented repeatedly between training and testing days. The same object was presented at most twice in daily session, but paired with a different object. 2.5.4. Testing Once the criterion was reached the FF-lesioning was performed (see above). The animals were allowed to recover for two weeks during the chronic drug treatment phase, and three additional days of pre-training were performed prior to the two DNMS tests (see Timeline, Fig. 1A).

2.6. DNMS Test 1 DNMS Test 1 took place 17 days after FF lesion, after 15 days of chronic treatment. Test 1 consisted of 20 trials (sample and choice phases), as in the training phase. The result of Test 1 was combined with the result attained on the last day of pre-training, to generate a two-day average score.

2.7. DNMS Test 2 DNMS Test 2, also consisting of 20 trials, took place one day after Test 1. At 1 h before testing, the chronically treated animals were additionally injected with either MEM (groups A, C and D) or DON (groups B and E; Table 1).

3. Results 3.1. Combination of systemic MEM and DON increased cortical and hippocampal ACh levels in anesthetized rats In anesthetized rats, systemic administration of MEM produced a dose-dependent increase in basal extracellular ACh levels in both the hippocampus and neocortex (Fig. 3A). At the highest dose used (10 mg/kg), there was approximately a 250% increase in release in both areas (hippocampus: mean increase 247%, KW statistic ¼ 15.2, p ¼ 0.003; cortex: mean 241%, KW ¼ 16.3; p ¼ 0.001). DON also produced a dose-dependent increase in basal ACh levels (Fig. 3B), albeit smaller than the MEM effect (mean 109% increase in hippocampus and 134% increase in neocortex; KW ¼ 16.2, p ¼ 0.001 and KW ¼ 18.0, p ¼ 0.0004, respectively). Since neostigmine (an AChEI) was added to the perfusate to enable the collection of ACh during microdialysis, the turnover of ACh should have been constant across samples; thus, the increased levels of ACh seen in this study are likely due to increased synaptic release rather than to decreased degradation by AChE inhibition, even in the DON samples. The increased ACh levels seen after DON administration returned rapidly to a near-baseline level in the second 15-min sample after injection (data not shown). The ACh release returned to basal levels more slowly after MEM injection, but was also not significantly

J. Ihalainen et al. / Neuropharmacology 61 (2011) 891e899

895

ACh release from the hippocampus, compared to a 167% increase with MEM alone and 73% increase with DON alone (differences between groups, KW ¼ 13.1, p ¼ 0.001). A one-tailed t-test comparing the combined drug effect with the expected sum (240%) of the means of the each drug alone was significant (p < 0.05), suggesting cooperativity. Similarly, in the cortex, the combination produced a much larger increase (480%) than either MEM alone (93%) or DON alone (104%; KW ¼ 10.1, p ¼ 0.007). A t-test comparing the combined drug group with the expected sum of 197% was statistically significant (p < 0.05). 3.2. Locally applied MEM and DON increased ACh levels in the hippocampus, but without a synergistic effect

Fig. 3. Effects of systemic memantine (A) and donepezil (B) on basal ACh release in the hippocampus and somatosensory cortex of anesthetized rats. ACh indicates acetylcholine; SEM, standard error of the mean *p < 0.05; **p < 0.01 compared to animals treated with 0 mg/kg of drug.

larger than baseline in the second post-injection sample (data not shown). The combination of intermediate concentrations of each drug (0.5 mg/kg DON and 5 mg/kg MEM) had a synergistic effect on basal ACh release (Fig. 4). The combination produced a 583% increase in

Fig. 4. Effect of acute, systemic memantine (MEM; 5 mg/kg), donepezil (DON; 0.5 mg/kg), and combined treatment (MEM þ DON) on basal ACh release (n ¼ 11 animals in each group). Treatment with memantine plus donepezil produced a significantly greater increase in ACh release than treatment with either drug alone; the mean increase in the combined group was greater than the sum of the increases in the two single-treatment groups (SUM). ACh indicates acetylcholine; SEM, standard error of the mean *p < 0.05; **p < 0.01.

In principle, systemic MEM and DON administration might affect ACh in any of the brain circuits leading to activation of the cholinergic CNS neurons. Delivery of the drug directly to the vicinity of the cholinergic terminals should help determine the extent to which such networks might contribute. Low concentrations (1 mM) of MEM and DON applied directly to the hippocampus by reverse dialysis had little effect on ACh release in anesthetized rats (þ0.5% and þ7.8%, respectively; data not shown). Administration of each drug at a concentration of 100 mM, on the other hand, was effective in increasing ACh release (Fig. 5). The increases seen during 100 mM MEM (þ64%, n ¼ 11) and 100 mM DON reverse dialysis (þ40%, n ¼ 7) were statistically significant compared to baseline (t10 ¼ 4.11, p < 0.01 and t6 ¼ 2.54, p < 0.05, respectively). The co-administration of both drugs also produced a significant increase (74%, t10 ¼ 4.57, p < 0.01). While this value was greater than either drug alone, statistical comparison found no significant difference between the three treatments (ANOVA: F2,26 ¼ 0.998, p ¼ 0.38). Thus, there was no statistical support for an additive effect of the two drugs when applied locally to the hippocampus. 3.3. Hippocampal ACh levels increased during DNMS task in freely moving rats Pre-lesioned rats were trained to perform a DNMS task until they reached a criterion of 80% or more correct choices on three consecutive testing days. The rats reached the criterion after approximately 5.5 weeks of training, and there were no differences in choice accuracy between the groups before the lesions (One-way ANOVA: F(4,45) ¼ 0.11; p ¼ 0.98). In sham-lesioned animals treated with placebo for 3 weeks, hippocampal ACh levels varied during different phases of the DNMS task (F(2,32) ¼ 140.1; p < 0.001). During exploration of the empty hole-board, ACh release was elevated compared to baseline

Fig. 5. Effects of intrahippocampal administration of high doses (100 mM) memantine (MEM), donepezil (DON), and both drugs combined (MEM þ DON), on basal ACh release in anesthetized rats. Although ACh release after intrahippocampal coadministration of both drugs was greater than that seen with either drug alone, the differences were not significant (ANOVA: F2,26 ¼ 0.998, p ¼ 0.38).

896

J. Ihalainen et al. / Neuropharmacology 61 (2011) 891e899

(empty cage with bedding), but a clear further increase was observed during the object recognition task performance (Fig. 6, Group A). 3.4. FF-lesion impaired DNMS task performance but did not significantly affect hippocampal ACh level The partial FF-lesions in this study were effective in decreasing the DNMS object recognition task performance by approximately 20% compared to pre-lesion performance levels (Fig. 7.: F(1,21) ¼ 15.8; p ¼ 0.001). Histology was used to confirm the size of the FF-lesions (Fig. 2A, B) and the location of microdialysis probes in dorsal hippocampus or somatosensory cortex; 15 animals were excluded from the analysis (Group A, n ¼ 3; Group B, n ¼ 4, Group C, n ¼ 3; Group D, n ¼ 3; Group E, n ¼ 2) due to ineffective or insufficient lesioning (evidenced by a lack of decreased AChE staining), incorrect microdialysis probe placement, detached microdialysis guide cannula or death after surgery. AChE staining intensity showed that the partial FF-lesion produced more than a 50% decrease in the number of AChE-positive neurites in the dorsal hippocampus, compared to sham lesioning (one-way ANOVA: F(1,38) ¼ 73.96; p < 0.001) (Fig. 2CeE). However, FF-lesioning did not reduce baseline hippocampal ACh levels compared to those with a sham lesion (Fig. 6, Group A vs. Group D þ E, white bars; F(1,21) ¼ 1.1; p ¼ 0.79). Our data is consistent with earlier studies showing no change in hippocampal ACh levels after partial FF-lesioning (Erb et al., 1997; Lapchak et al., 1991), likely due to a compensatory increase of ACh release from the remaining cholinergic terminals. 3.5. Chronic MEM improved DNMS task performance but did not increase hippocampal ACh release significantly Immediately prior to testing the effect of chronic drug treatment (DNMS Test 1, Fig. 1), the rats underwent three days of postoperative testing on the DNMS task, during which time their performance approached an asymptote. The choice accuracy on the DNMS task differed significantly between the chronic treatment groups on Test 1 (Fig. 8, left panel: F(3,40) ¼ 4.1; p ¼ 0.014). Chronic administration of MEM in drinking water (30 mg/kg/day p.o.) for 15 days significantly improved the DNMS task performance of lesioned rats, compared to that of lesioned animals treated with placebo (Fig. 8, left panel, Group B vs. Groups D þ E) but only marginally elevated hippocampal ACh levels during the task (Fig. 6, Group B vs. Group D þ E, dark bars; F(1,17) ¼ 0.14; p ¼ 0.15. Chronic DON also resulted in a moderate but statistically nonsignificant

Fig. 6. Effect of chronic drug treatments on hippocampal ACh levels during different phases of the delayed non-match to sample object recognition task: baseline (samples 1e3), empty hole-board exploration (sample 4), and object recognition task (samples 7 and 10). SEM indicates standard error of the mean.

Fig. 7. Delayed non-match to sample object recognition task performance before and after the fimbria-fornix lesion. Sham group (white columns, n ¼ 11) and lesion groups (groups D and E; black columns, n ¼ 12). SEM indicates standard error of the mean **p ¼ 0.001, compared to sham-lesioned group (t-test).

improvement in the DNMS task performance compared to the placebo lesion group (Fig. 8, left panel, Group C vs. Groups D þ E). Hippocampal ACh levels in lesioned animals were not significantly altered after chronic DON treatment in the presence of neostigmine in the dialysate (Fig. 6, Group C vs. Group D þ E). 3.6. Acute MEM injection in rats treated chronically with DON or placebo did not significantly improve DNMS task performance but increased hippocampal ACh levels Although chronic MEM did not significantly elevate ACh levels in this study, we investigated whether short-term MEM treatment could increase brain ACh levels, possibly contributing to the functional synergy observed in clinical trials in which MEM was added to a background of chronic DON treatment (Grossberg et al., 2008; Tariot et al., 2004). One day after DNMS Test 1, animals that had been chronically treated with DON or placebo for 3 weeks, were acutely injected with a dose of MEM (5 mg/kg; IP) and tested again (DNMS Test 2; Fig. 8, right panel). Similarly, animals chronically treated with MEM or placebo received an acute dose of DON (2.5 mg/kg; IP). An ANOVA-model with the test day as withinsubject and both chronic (placebo vs. drug) and acute (MEM vs. DON) drug treatment as between-subject factors revealed a highly

Fig. 8. Task performance after FF-lesioning on the DNMS Test 1 (left panel) and Test 2 (right panel). DNMS indicates delayed non-match-to-sample; DON, donepezil (2.5 mg/kg/day chronic treatment; 2.5 mg/kg acute treatment); MEM, memantine (30 mg/kg/day chronic treatment; 5 mg/kg acute treatment); PBO, placebo; SEM, standard error of the mean. *p < 0.05; **p < 0.01 compared to placebo-treated lesioned animals (Dunnett’s test).

J. Ihalainen et al. / Neuropharmacology 61 (2011) 891e899

Fig. 9. Baseline ACh levels after chronic drug treatment (white bars) and after an acute drug challenge in chronically treated animals (dark bars). The dialysates were collected 1e2 h after acute drug injection. Acute drug effects (dark vs. white bars) within individual groups were the following: Group A: **p ¼ 0.006; Group B: p ¼ 0.08; Group C: **p ¼ 0.006; Group D: *p ¼ 0.013, Group E: p ¼ 0.89 (One-way ANOVA). ACh indicates acetylcholine; DON, donepezil; MEM, memantine; PBO, placebo, SEM, standard error of the mean.

significant effect of the test day (F(1,39) ¼ 41.0, p < 0.001). This implies that all animals performed better on the second day with the acute drug administration (MEM or DON; Fig. 8). However, the day  acute drug interaction or the three-way interaction day  chronic drug  acute drug were nonsignificant (p > 0.71), indicating that there was no difference between the treatments. It should be noted that, since chronic MEM treatment restored the performance of lesioned rats to the level of sham animals, the apparent lack of an effect of acute DON treatment in chronic MEMtreated animals may have been due to a ceiling effect. A similar effect may have occurred in chronic DON-treated animals treated with acute MEM. A different picture emerged in the analysis of ACh levels. Not only was there a significant difference between the days as in the ACh levels F(1,39) ¼ 13.9, p < 0.01) but also a highly significant day  acute drug interaction (F(1,39) ¼ 34.2, p < 0.001). This derives from the fact that MEM increased ACh release while DON did not (Fig. 9). Furthermore, the three-way interaction day  chronic drug  acute drug was nonsignificant (F(1,39) ¼ 2.4, p > 0.13), indicating that MEM increased ACh release whether given after chronic placebo or after chronic DON (Fig. 9). 4. Discussion This study demonstrates that acute administration of MEM significantly increases ACh levels in the hippocampus of freely moving rats and in both the hippocampus and neocortex of anesthetized animals. A moderate increase that did not reach statistical significance was also observed after chronic MEM administration, compared to placebo, in lesioned animals. Chronic, oral administration of MEM also significantly improved short-term recognition memory of FF-lesioned rats. Whereas the memory enhancement by MEM was more pronounced after chronic administration, the acute dose was more effective in increasing ACh release, suggesting that the long-term, memory-improving effect of MEM is likely distinct from its short-term effects involving increased ACh release. The enhancement of ACh levels by MEM in this study was observed in both anesthetized and freely moving animals, although the effect was less pronounced in the awake animals, most likely due to higher baseline activity in cholinergic neurons. Systemically

897

administered MEM in this study had a strong effect on ACh release in both cortex and hippocampus of anesthetized rats, increasing release 3.5-fold at the highest dose used (10 mg/kg) compared to baseline. We also found that acute administration of MEM (5 mg/kg; IP), a dose known to produce plasma drug levels comparable to those seen in patients (Danysz et al., 1997; Parsons et al., 2007), resulted in a significant enhancement of hippocampal ACh release in freely moving rats. An earlier study demonstrated enhancement of ACh release from nucleus accumbens and ventral tegmental area (Shearman et al., 2006) after acute MEM administration (although at a higher dose than our study), but to our knowledge this is the first reported evidence that MEM at clinically relevant doses is able to enhance ACh release in the hippocampus. In contrast to the effect of AChEIs, which prolong the availability of ACh once it is released, the increased ACh levels caused by MEM appear to be due to increased synaptic release. While several mechanisms for NMDA receptor antagonist-mediated increase in ACh release have been proposed in the literature (Hanania and Johnson, 1999; Kim et al., 1999), our data obtained using reverse dialysis indicate that locally applied MEM in our system likely exerts its effects at or near cholinergic terminals in the hippocampus itself rather than on the cholinergic cell bodies in the medial septum, which are approximately 5 mm anterior to the location of the microdialysis probes. However, because the effect required as high MEM concentration as 100 mM we cannot fully exclude spread of MEM all the way to the medial septum. Because of the large number of experimental groups, we did not include a control group receiving a placebo injection after chronic placebo treatment; thus we cannot assess whether the stress of acute injection plays a role in the observed increases in ACh release, or in cognitive performance, due to increased vigilance. However, there was no increase in hippocampal ACh release in the chronic placebo group that received an acute DON injection (Fig. 9). Since ACh release is a sensitive measure of stress-induced increase in vigilance (Day et al., 1998), we consider it unlikely that the injection itself causes any significant changes in vigilance affecting the task performance. After chronic MEM administration, only a trend of increased ACh release was observed. However, this finding should be interpreted with caution for two reasons. First, the applied microdialysis method is sensitive to revealing changes in ACh levels in the same animal across different time points, but is less suitable for comparing levels between animals (which would have required a laborious no-net-flux method to calibrate the ACh levels across subjects). However, we believe that some group comparison is possible even with this method, since the baseline ACh levels were similar in the two placebo and the DON groups, and the variations were small (Fig. 6). Second, the within-group variation was much larger in the group treated chronically with MEM than in other three groups (Fig. 6). This may reflect the fact that MEM was administered in drinking water and the rats may have had very different drinking pattern during the experiment, resulting in different momentary MEM concentrations, although the concentrations on average were comparable. It is also possible that acute and chronic MEM differentially affect ACh release. Although several studies have reported that NMDA receptor antagonists increase brain ACh levels after acute dosing (Giovannini et al., 1994, 1997; Hasegawa et al., 1993; Hutson and Hogg, 1996), there is no evidence that a similar increase in brain ACh levels can be maintained after chronic administration. We did not observe a robust increase in hippocampal ACh levels following 3 weeks of treatment with MEM, and it is possible that tolerance may have developed to this effect of MEM over time. Development of tolerance to certain effects of MEM (e.g. changes in locomotor activity) following

898

J. Ihalainen et al. / Neuropharmacology 61 (2011) 891e899

chronic treatment has been reported (Hesselink et al., 1999; Kos and Popik, 2005). Interestingly, tolerance to the pro-cognitive effects of MEM does not develop over time (Minkeviciene et al., 2004, 2008). In this study, we observed a significant improvement in DNMS task performance in FF-lesioned rats following 3 weeks of treatment with MEM. After acute MEM treatment (5 mg/kg, IP), we observed a substantial, significant rise in hippocampal ACh levels, but only a trend toward improvement in DNMS task performance. Intriguingly, a comparable acute dose of MEM (2e10 mg/kg; IP) was shown to reverse scopolamine-induced learning impairment in mice in one study (Drever et al., 2007), but resulted in no change or even cognitive impairment in APP transgenic mice (Van Dam et al., 2005) or senescent rats (Creeley et al., 2006) when given alone. Therefore, it is possible that increased ACh release by MEM contributes to its cognition-improving effect only in models involving severely compromised (muscarinic) ACh neurotransmission. Despite a dramatic loss of cholinergic terminals in the hippocampus in FF-lesioned rats (as indicated by AChE staining), the extracellular ACh levels were not decreased, indicating compensatory elevation of ACh release by the preserved terminals. In contrast, an acute scopolamine injection leaves no time for any compensatory mechanisms to develop. Because of the discrepancy between MEM-induced ACh release and DNMS task performance after chronic and acute MEM administration, it is likely that MEM-induced cognitive enhancement in this experimental model also involves other, probably NMDA receptor mediated mechanisms of action, which may or may not be related to augmented ACh release. It is noteworthy that selective neurotoxin lesioning of either cholinergic or GABAergic septal projections to the hippocampus leads to only a modest memory deficit, whereas simultaneous damage to both projections, as occurs with FF-lesioning, results in memory impairment similar to lesioning of the hippocampus itself (Pang et al., 2001). Therefore, it is possible that MEM treatment in FF-lesioned rats corrects the resulting imbalance between GABAergic and glutamatergic functions, independent of its action on ACh release, and thereby reverses the memory deficit. Elucidation of these exact mechanisms calls for further studies. In freely moving rats, we did not observe an increase in ACh levels by DON, most likely due to presence of neostigmine (another AChEI, added to the dialysis buffer to prevent ACh hydrolysis during sample collection). However, in anesthetized rats, DON was able to increase ACh levels even in the presence of neostigmine, suggesting that in this experiment, either the neostigmine did not completely eliminate the baseline catabolism of ACh in the tissue, or DON directly increased the very low level of ACh release during anesthesia. One possible explanation for this apparent difference between awake and anesthetized animals is that the low levels of ACh collected in anesthetized rats are more dependent on small changes in AChE inhibition, when spontaneous ACh release is at a minimum, than in alert animals. Another factor that may contribute is that the hepatic P450 CYP3A enzyme, which metabolizes DON (Jann et al., 2002), is inhibited by urethane (Meneguz et al., 1999), meaning that the effective concentration of DON may be higher in urethane-anesthetized animals. In any event, the primary effect of DON, inhibition of the breakdown of ACh, cannot be monitored in microdialysis studies because of the necessity of using an AChEI in the perfusate to enable sample collection. Notwithstanding this limitation, our experiment on anesthetized rats demonstrated that the combination of MEM and DON was considerably more effective than either drug alone in increasing extracellular ACh concentrations, thus supporting the prediction that their combined effect would be additive (or cooperative) because they act via different mechanisms.

5. Conclusions Our data present evidence that MEM is able to enhance extracellular levels of ACh in the hippocampus, a brain structure highly relevant for declarative memory and Alzheimer’s disease, and that this mechanism involves enhanced ACh release rather than modification of its enzymatic breakdown. This effect was also observed in freely moving animals after an acute injection; however, after chronic MEM administration the increase in ACh release was no longer significant, suggesting the development of tolerance. Chronic MEM administration, on the other hand, resulted in a significant improvement in short-term recognition memory in fimbria-fornix lesioned rats. Although the combination of MEM and DON resulted in additive effect on ACh extracellular levels in anesthetized animals, a similar effect was not observed in freely moving animals. Based on our findings of augmented ACh release by MEM, it would be tempting to speculate that the cognitive benefits seen clinically with the combination of MEM and DON treatment in patients with AD (Grossberg et al., 2008; Tariot et al., 2004) results from a synergistic effect of the two drugs, with MEM increasing ACh release and DON preventing its enzymatic degradation. However, the lack of significant ACh release after chronic MEM, which is most relevant for the clinical effect, does not support this notion. It appears that the effect of MEM on ACh release is lost after chronic administration, so a more parsimonious explanation for the beneficial effect of the MEM þ DON combination is that they engage two completely independent mechanisms, both favorable to cognition.

Disclosure statement Dr. Pradeep Banerjee is an employee of Forest Research Institute, Inc., Jersey City, NJ, USA.

Role of the funding source This work was supported by a grant from Forest Laboratories, Inc. (MEM-TX-13), as well as the Canadian Institutes of Health Research (MOP-44016). Author Pradeep Banerjee, from Forest Research Institute, Inc., contributed to study conception, preparation of protocols, data review and presentation, and manuscript preparation. Acknowledgments We thank Pasi Miettinen for technical assistance in histology, Milka Määttä for assistance in animal training, Heather Allen for technical assistance, and Michael L. Miller and Vojislav Pejovic from Prescott Medical Communications Group for editorial and medical writing suggestions. References Akasofu, S., Kimura, M., Kosasa, T., Sawada, K., Ogura, H., 2008. Study of neuroprotection of donepezil, a therapy for Alzheimer’s disease. Chem. Biol. Interact. 175, 222e226. Barnes, C.A., Danysz, W., Parsons, C.G., 1996. Effects of the uncompetitive NMDA receptor antagonist memantine on hippocampal long-term potentiation, shortterm exploratory modulation and spatial memory in awake, freely moving rats. Eur. J. Neurosci. 8, 565e571. Creeley, C., Wozniak, D.F., Labruyere, J., Taylor, G.T., Olney, J.W., 2006. Low doses of memantine disrupt memory in adult rats. J. Neurosci. 26, 3923e3932. Danysz, W., Parsons, C.G., Kornhuber, J., Schmidt, W.J., Quack, G., 1997. Aminoadamantanes as NMDA receptor antagonists and antiparkinsonian agentse preclinical studies. Neurosci. Biobehav. Rev. 21, 455e468. Day, J.C., Koehl, M., Le Moal, M., Maccari, S., 1998. Corticotropin-releasing factor administered centrally, but not peripherally, stimulates hippocampal acetylcholine release. J. Neurochem. 71, 622e629.

J. Ihalainen et al. / Neuropharmacology 61 (2011) 891e899 Drever, B.D., Anderson, W.G., Johnson, H., O’Callaghan, M., Seo, S., Choi, D.Y., Riedel, G., Platt, B., 2007. Memantine acts as a cholinergic stimulant in the mouse hippocampus. J. Alzheimers Dis. 12, 319e333. Erb, C., Klein, J., Koppen, A., Loffelholz, K., Jeltsch, H., Cassel, J.C., 1997. Modulation of hippocampal acetylcholine release after fimbria-fornix lesions and septal transplantation in rats. Neurosci. Lett. 231, 5e8. Giovannini, M.G., Camilli, F., Mundula, A., Pepeu, G., 1994. Glutamatergic regulation of acetylcholine output in different brain regions: a microdialysis study in the rat. Neurochem. Int. 25, 23e26. Giovannini, M.G., Giovannelli, L., Bianchi, L., Kalfin, R., Pepeu, G., 1997. Glutamatergic modulation of cortical acetylcholine release in the rat: a combined in vivo microdialysis, retrograde tracing and immunohistochemical study. Eur. J. Neurosci. 9, 1678e1689. Grossberg, G., Manes, F., Allegri, R., Miguel, L., Robledo, L., Gloger, S., Xie, L., Jia, X., Perhach, J., 2008. A multinational, randomized, double-blind, placebocontrolled, parallel-group trial of memantine extended-release capsule (28 mg, once daily) in patients with moderate to severe Alzheimer’s disease. Alzheimers Dement. 4, T793. Hanania, T., Johnson, K.M., 1999. Regulation of NMDA-stimulated [14C]GABA and [3H]acetylcholine release by striatal glutamate and dopamine receptors. Brain Res. 844, 106e117. Hasegawa, M., Kinoshita, H., Amano, M., Hasegawa, T., Kameyama, T., Nabeshima, T., 1993. MK-801 increases endogenous acetylcholine release in the rat parietal cortex: a study using brain microdialysis. Neurosci. Lett. 150, 53e56. Hedreen, J.C., Bacon, S.J., Price, D.L., 1985. A modified histochemical technique to visualize acetylcholinesterase-containing axons. J. Histochem. Cytochem. 33, 134e140. Heim, C., Sontag, K.H., 1995. Memantine prevents progressive functional neurodegeneration in rats. J. Neural Transm. Suppl. 46, 117e130. Hesselink, M.B., Smolders, H., De Boer, A.G., Breimer, D.D., Danysz, W., 1999. Modifications of the behavioral profile of non-competitive NMDA receptor antagonists, memantine, amantadine and (þ)MK-801 after chronic administration. Behav. Pharmacol. 10, 85e98. Hunsaker, M.R., Rogers, J.L., Kesner, R.P., 2007. Behavioral characterization of a transection of dorsal CA3 subcortical efferents: comparison with scopolamine and physostigmine infusions into dorsal CA3. Neurobiol. Learn. Mem 88, 127e136. Hutson, P.H., Hogg, J.E., 1996. Effects of and interactions between antagonists for different sites on the NMDA receptor complex on hippocampal and striatal acetylcholine efflux in vivo. Eur. J. Pharmacol. 295, 45e52. Ihalainen, J., Sarajarvi, T., Kemppainen, S., Keski-Rahkonen, P., Lehtonen, M., Tanila, H., 2010. A novel delayed non-match to sample object recognition task that allows simultaneous in vivo microdialysis. J. Neurosci. Methods 189, 210e215. Jann, M.W., Shirley, K.L., Small, G.W., 2002. Clinical pharmacokinetics and pharmacodynamics of cholinesterase inhibitors. Clin. Pharmacokinet. 41, 719e739. Keski-Rahkonen, P., Lehtonen, M., Ihalainen, J., Sarajarvi, T., Auriola, S., 2007. Quantitative determination of acetylcholine in microdialysis samples using liquid chromatography/atmospheric pressure spray ionization mass spectrometry. Rapid Commun. Mass Spectrom. 21, 2933e2943. Kim, S.H., Price, M.T., Olney, J.W., Farber, N.B., 1999. Excessive cerebrocortical release of acetylcholine induced by NMDA antagonists is reduced by GABAergic and alpha2-adrenergic agonists. Mol. Psychiatry 4, 344e352.

899

Kos, T., Popik, P., 2005. A comparison of the predictive therapeutic and undesired side-effects of the NMDA receptor antagonist, memantine, in mice. Behav. Pharmacol. 16, 155e161. Kosasa, T., Kuriya, Y., Matsui, K., Yamanishi, Y., 1999. Effect of donepezil hydrochloride (E2020) on basal concentration of extracellular acetylcholine in the hippocampus of rats. Eur. J. Pharmacol. 380, 101e107. Lapchak, P.A., Jenden, D.J., Hefti, F., 1991. Compensatory elevation of acetylcholine synthesis in vivo by cholinergic neurons surviving partial lesions of the septohippocampal pathway. J. Neurosci. 11, 2821e2828. Materi, L.M., Semba, K., 2001. Inhibition of synaptically evoked cortical acetylcholine release by intracortical glutamate: involvement of GABAergic neurons. Eur. J. Neurosci. 14, 38e46. Meneguz, A., Fortuna, S., Lorenzini, P., Volpe, M.T., 1999. Influence of urethane and ketamine on rat hepatic cytochrome P450 in vivo. Exp. Toxicol. Pathol. 51, 392e396. Minkeviciene, R., Banerjee, P., Tanila, H., 2004. Memantine improves spatial learning in a transgenic mouse model of Alzheimer’s disease. J. Pharmacol. Exp. Ther. 311, 677e682. Minkeviciene, R., Banerjee, P., Tanila, H., 2008. Cognition-enhancing and anxiolytic effects of memantine. Neuropharmacology 54, 1079e1085. Pang, K.C., Nocera, R., Secor, A.J., Yoder, R.M., 2001. GABAergic septohippocampal neurons are not necessary for spatial memory. Hippocampus 11, 814e827. Parsons, C.G., Stoffler, A., Danysz, W., 2007. Memantine: a NMDA receptor antagonist that improves memory by restoration of homeostasis in the glutamatergic system e too little activation is bad, too much is even worse. Neuropharmacology 53, 699e723. Reisberg, B., Doody, R., Stoffler, A., Schmitt, F., Ferris, S., Mobius, H.J., 2003. Memantine in moderate-to-severe Alzheimer’s disease. N. Engl. J. Med. 348, 1333e1341. Shearman, E., Rossi, S., Szasz, B., Juranyi, Z., Fallon, S., Pomara, N., Sershen, H., Lajtha, A., 2006. Changes in cerebral neurotransmitters and metabolites induced by acute donepezil and memantine administrations: a microdialysis study. Brain Res. Bull. 69, 204e213. Tariot, P.N., Farlow, M.R., Grossberg, G.T., Graham, S.M., McDonald, S., Gergel, I., 2004. Memantine treatment in patients with moderate to severe Alzheimer disease already receiving donepezil: a randomized controlled trial. JAMA 291, 317e324. Van Dam, D., Abramowski, D., Staufenbiel, M., De Deyn, P.P., 2005. Symptomatic effect of donepezil, rivastigmine, galantamine and memantine on cognitive deficits in the APP23 model. Psychopharmacology (Berl.) 180, 177e190. Wenk, G.L., Quack, G., Möbius, H.J., Danysz, W., 2000. No interaction of memantine with acetylcholinesterase inhibitors approved for clinical use. Life Sci. 66, 1079e1083. Winblad, B., Poritis, N., 1999. Memantine in severe dementia: results of the M-best study (benefit and efficacy in severely demented patients during treatment with memantine). Int. J. Geriatr. Psychiatry 14, 135e146. Wise, L.E., Lichtman, A.H., 2007. The uncompetitive N-methyl-D-aspartate (NMDA) receptor antagonist memantine prolongs spatial memory in a rat delayed radial-arm maze memory task. Eur. J. Pharmacol. 575, 98e102. Yamada, K., Takayanagi, M., Kamei, H., Nagai, T., Dohniwa, M., Kobayashi, K., Yoshida, S., Ohhara, T., Takuma, K., Nabeshima, T., 2005. Effects of memantine and donepezil on amyloid beta-induced memory impairment in a delayedmatching to position task in rats. Behav. Brain Res. 162, 191e199.