INTRODUCTION

Glutamate receptor blockers with selective action on N-methyl-D-aspartate (NMDA) receptors have wide therapeutic potential in a variety of central nervous system (CNS) disorders ranging from the acute neurodegeneration (eg stroke and trauma), chronic neurodegeneration (eg Parkinson's disease (PD), Alzheimer's disease (AD), and Huntington's disease (HD) (Danysz et al, 1995; Parsons et al, 1998; Krystal et al, 1999), to symptomatic treatment (eg epilepsy, PD, drug dependence, depression, anxiety, and chronic pain) (Meldrum, 1986; Danysz et al, 1995; Krystal et al, 1999; Skolnick, 1999).

In these conditions, an excessive release of glutamate is thought to play a major role in triggering neuronal death via NMDA receptors (Choi, 1988; Danysz et al, 1995; Parsons et al, 1998). NMDA receptor channel blockers represent one group of pharmacologically active agents able to antagonize NMDA receptor activation by acting directly on the phencyclidine recognition site within the channel. Since these agents act in a ‘use-dependent’ manner, only blocking the channel in the open state (Foster and Wong, 1987; Huettner and Bean, 1988), they have been termed uncompetitive NMDA receptor antagonists.

Dizocilpine maleate (MK-801) is a prototypical NMDA-receptor channel blocker with a high affinity to the ion channel (Wong et al, 1986). However, high affinity NMDA receptor channel blockers have been observed to elicit psychotomimetic properties in humans (Krystal et al, 1999), and thus have been useful only experimentally in vitro and in vivo. In contrast, memantine (1-amino-3,5-dimethyladamantane), an uncompetitive NMDA receptor antagonist of moderate affinity to the receptor-associated ion-channel, exhibits fast blocking and unblocking kinetics, high voltage dependency (Kornhuber et al, 1991; Parsons et al, 1993), and is well tolerated at clinical doses (Parsons et al, 1999). Memantine is suggested as a neuroprotective agent for the treatment of several dementias, particularly AD (Görtelmeyer and Erbler, 1992; Muller et al, 1995; Parsons et al, 1999). Moreover, memantine and acetylcholinesterase inhibitors are currently the only agents widely used to improve cognition and daily activities in patients with AD (Kilpatrick and Tilbrook, 2002; Winblad et al, 2002). Memantine has been approved for clinical use in Europe and is currently undergoing clinical trials in the United States (Kilpatrick and Tilbrook, 2002). Thus, information concerning the molecular pharmacology of memantine is important to understand its actions and to anticipate potential adverse effects.

We sought to characterize the genes with altered expression in adult rat brain after MK-801 and memantine treatment in order to better understand the multiple actions of NMDA/glutamate antagonists. cDNA microarray-based hybridization techniques offer a high throughput platform that allows for rapid and cost-effective screens to detect effects of pharmacological agents (Schena et al, 1995; Grunblatt et al, 2001; Ueda et al, 2002). Here, we utilized a microarray containing 1090 cDNA fragments and in situ hybridization (ISH) to determine the gene expression alterations in response to treatment with uncompetitive NMDA receptor antagonists in rat limbic cortex. We were able to identify several genes with altered expression in response to MK-801 and memantine treatment. In addition to already known and identified genes involved in either pharmacological or toxicological effects, the study also uncovered a number of new gene/treatment correlations. Characterization of these genes may lead to better understanding of the molecular pharmacology of uncompetitive NMDA receptor antagonists. More importantly, some of these genes may constitute novel targets for new therapeutic use.

MATERIALS AND METHODS

Male Wistar rats (initial weight 250–280 g, National Laboratory Animal Center, University of Kuopio, Kuopio, Finland) were kept under standardized temperature, humidity, and lighting conditions (12-h light/dark cycles) with free access to water and food. All attempts were made to minimize pain and discomfort of the experimental animals. All animal treatments were approved by the Experimental Animal Committee of the University of Kuopio, and they have been carried out in accordance with the guidelines issued by the Society for Neuroscience.

Doses and duration of the treatment with uncompetitive NMDA receptor antagonists, MK-801 and memantine, used in this study were based on the previous studies. Memantine intraperitoneally (i.p.) dose 25 mg/kg (4 h) was shown to increase significantly BDNF mRNA levels in the rat brain (Marvanová et al, 2001) and produce plasma levels reaching ∼6.5 μM concentrations, which are still within the range specific to NMDA receptor affinity (Kornhuber et al, 1994). Moreover the neuroprotective dose of memantine in rats has been shown to be 10–30 mg/kg in acute indications (Wenk et al, 1994). A 5 mg/kg i.p. dose of memantine in rats produces plasma levels of ∼1 μM that can be considered therapeutically relevant (Parsons et al, 1999). It has been suggested that neuroprotective doses of MK-801 (0.3–0.8 mg/kg; i.p.) block NMDA receptors in the brain, and these doses produce plasma concentrations within the range of NMDA receptor affinity (Gill et al, 1991; Massieu et al, 1993).

Rats were injected i.p. with saline (control group of animals) or with 5, 10, 25, and 50 mg/kg of memantine HCl, and killed 1–48 h afterwards as indicated. In all, 1 mg/kg MK-801 (dizocilpine maleate) (RBI, Natick, MA, USA) was applied to rats that were killed 4 h afterwards.

The rats were narcotized with CO2 before decapitation. Brains were rapidly removed and posterior cingulate and anterior retrosplenial cortices were rapidly and carefully dissected and immediately stored on dry ice to be processed for RNA isolation. Brains for ISH histochemistry were rapidly removed and rinsed in PBS (pH 7.5), placed on dry ice, and stored at −80°C. Coronal brain sections (14 μm) were cut on a Leica CM 3000 cryostat and thaw-mounted onto SuperFrost/Plus (Menzel-Glaeser, Germany) slides. Brain sections from saline (control) and treated rats were comounted onto the same slides and fixed in 4% paraformaldehyde solution, and stored in 95% ethanol at 4°C until use.

cDNA Array

Rat 1.0 Atlas Glass Microarrays for dual color analyses (Clontech Laboratories, Palo Alto, CA, USA) were utilized in this study. The array contains 86 controls and 1090 well-characterized transcripts with assigned function and a broad functional range class. Each gene is represented by ‘long oligo’, single-stranded cDNA fragments of around 80 bases that lack repetitive elements and have minimal homology with other genes on the array. The full gene list can be found at http://www.uku.fi/aivi/supplementary_information.htm.

RNA Extraction, Probe Synthesis, and Hybridization

Total RNA was isolated from frozen tissue using Trizol reagent (Gibco BRL, MD, USA) according to the manufacturer's recommended protocol. The total RNA was then DNase I treated (Ambion, Inc., Texas, USA). The integrity of RNA was verified with a 1% agarose-formaldehyde minigel and concentrations were obtained by measuring absorbance at 260 nm. Fluorescently labeled probes were synthesized from total RNA (20 μg) by reverse transcription and indirect incorporation of fluorescent dye (Cy3 and Cy5) according to the manufacturer's protocol. Following hybridization (42°C, for 16 h) with fluorescently labeled cDNAs in hybridization chambers, all arrays were washed and then scanned (ScanArray 5000, GSI Lumonics, Inc., Billerica, MA, USA).

Gene Expression Data Analysis

Raw scanned images of Cy3 and Cy5 fluorescence were processed and the signal from each spot was then quantified using QuantArray analysis software (GSI Lumonics, Inc.). The local background was subtracted from the fluorescent values of each spot to obtain an intensity value. For each individual treatment, duplicate microarrays were hybridized with rat brain mRNA samples obtained from two rat individuals. In this hybridization study a ‘swapped dyes’ design was used. On one microarray, the treatment sample was labeled with Cy5, and control was labeled with Cy3. The dyes were then switched where the treatment sample was labeled with Cy3 and control was labeled with Cy5. In principle, the ratios of treatment (Cy-5)/control (Cy-3) and treatment (Cy-3)/control (Cy-5) should be approximately equal. In order to account for any differential incorporation of the fluorescent labels and differences in scanning parameters, the signal intensities for Cy3 and Cy5 were globally normalized and the expression ratios between experimental and control samples were determined. Genes were upregulated if they displayed a ratio ⩾1.5 and downregulated if the ratio was ⩽0.67 in both replicate dye swap experiments. The cut-off value for gene induction or suppression were chosen based on observed gene expression data, and previously reported microarray profiling studies of brain regions, which showed that the magnitude of gene expression differences is often less than two-fold in brain tissue (Mirnics et al, 2000). The analyzed expression data sets of memantine and MK-801 experiments are located on the publicaly available website (http://www.uku.fi/aivi/supplementary_information.htm).

In situ Hybridization

ISH was performed as described previously (Marvanová et al, 2001). Briefly, the antisense oligonucleotides of rat RL/IF-1 (5′-GTAGGGCAACTCATCTTCCGTGAATTCTGACTCCG-3′), rat GDI-1 (5′-GGAGATCATGCACACATAGATGTCTGGCTTCCTGT-3′), rat PAK-1 (5′-CTCTGCACACAGCTGCTATCTGGCCTTCATCCATA-3′), rat SHPS-1 (5′-CTGTTTGATTCGGAGGAGGTAGAGGGCAGCCATCA-3′), rat NVP-2 (5′-CACCAGCCCTACTTCTGCATGTCACATTGCAACAG-3′), rat GRP78 (5′-GGCCTGAGAATAGCGAGCAGCAACGTCTCGAGCTT-3′), rat neuromedin B receptor (5′-AACTGCTGTTATCCGAGCTACCGATGCGTGCTACT-3′), rat ARF-3 (5′-GAGCTCATCCTCCGCCAGCATCCTCATCAGCTCTT-3′), rat prostaglandin E2 receptor (5′-CTGCATGCGAATAAGGTTGAGGATCACACTGATG-3′), rat calcium channel β (5′-TTCACTCTGTACCTCAGCCAGGCTGGAGCGAGTGT-3′), rat HSP90β (5′-TCTTCATCAATGCCCAGGCCTAGTTTAATCATGCG3′), rat dopamine D4 receptor (3′-TCGGCAGCCGCCAGGCTCACGATGAAGTAGTTGGT-5′), rat c-fos (5′-GCAGCGGGAGGATGACGCCTCGTAGTCCGCGTTGAAACCCGAGAA-3′), and rat FGF-2 (5′-AGTGTCTAAAGAGAGTCAGCTCTTAGCAGACATTG-3′), rat K+/Na+ transporting ATPase 2β (5′-TTTAGCTGGGTCCGGTTGAACTGGCAGGCACGTT-3′), rat F-spondin (5′-ACACTTCTCCGCCTGCTCCAGATCCTCATTACAGT-3′) mRNAs were 3′ end-labeled to a specific activity of 1–2 × 107 cpm/pmol using terminal deoxynucleotidyl transferase (MBI Fermentas, Vilnus, Lithuania) and a 30 : 1 molar ratio of α-[33P]-dATP (2000 Ci/mmol, New England Nuclear, Boston, MA) to probe. Hybridization was performed for 21 h at 42°C on paraformaldehyde postfixed sections in the presence of 1–3 × 103 cpm/μl of labeled probe in buffer containing 50% formamide, 4 × standard saline citrate (SSC) (1 × SSC: 150 mM NaCl, 15 mM sodium citrate), and 10% dextran sulfate. After incubation, the sections were washed and opposed onto Hyperfilm-βmax films (Amersham Life Sciences, Inc., Buckinghamshire, England) for 5 days to 3 weeks and developed for 5 min in D-19 (Kodak, France). Nonspecific hybridization was determined by adding 100-fold excess of unlabeled oligonucleotide probe to the hybridization buffer. Nonspecific hybridization was not significantly above the background.

Immunoblotting

Tissues from posterior cingulate and anterior retrosplenial cortices were rapidly dissected and well homogenized in 1 ml of lysis buffer (50 mM Hepes, pH 7.5, 150 mM NaCl, 2.5 mM MgCl2, Protease inhibitors and 0.1 mM GDP). The homogenate was spun for 10 min at 600 g at 4°C to remove nuclear fractions and intact cells. The supernatant was transferred to SW 40 swinging bucket rotor tubes and centrifuged at 100 000 g for 30 min at 4°C. After centrifugation, the supernatant containing cytosolic fraction was removed to a separate tube and the obtained pellet containing membrane protein was resuspended in 250 μl of lysis buffer. All protein fractions were stored in −80°C until use. After measuring protein concentrations (Dc Protein Assay, BioRad, Richmond, CA, USA), samples consisting of equal amounts of protein (20 μg) were subjected to 7.5% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Proteins were transferred to nitrocellulose membranes (Protran, Schleicher & Schuell, Jeene, NH) and then blocked with 3% bovine serum albumin (BSA) in TBS containing 0.1% Tween 20 (TBS-T). Immunoblotting was carried out with the rabbit GDI-1 (GDP-dissociation inhibitor 1) (Amersham Life Sciences, Inc.) and anti-actin (c-11, goat polyclonal IgG) (Santa Cruz Biotechnology, Santa Cruz, CA, USA) primary antibodies using 1 : 1000 dilutions at 4°C overnight. Filters were rinsed in water twice and in PBS containing 0.05% Tween twice. Bound immunoglobulins were detected with anti-rabbit (Amersham) or anti-goat (Santa Cruz Biotechnology) horseradish peroxidase-conjugated secondary antibodies at 1 : 1000 dilution for 1 h at room temperature, rinsed, and followed by chemiluminescence detection using luminol (Sigma-Aldrich, Helsinki, Finland) as the substrate. Signals were exposed to films for 1–10 min.

Data Analysis

Optical densities from immunoblot and ISH films were quantified by a video-based image analysis system (MCID M4, Imaging Research, Inc., St Catharines, Ontario, Canada). The anatomical brain regions were determined according to Paxinos and Watson (1986). For ISH autoradiograms, 14C-microscales were exposed simultaneously to convert optical densities to nCi/g statistical analysis was performed with ANOVA followed by Dunnett's post hoc test.

RESULTS

Expression Array Data Analysis

To examine the molecular mechanisms contributing to the effect of MK-801 and memantine, mRNA expression was profiled using Clontech Rat 1.0 glass arrays containing 1090 genes of known function. Expression levels and profiles of genes were obtained after saline and MK-801 (1 mg/kg; 4 h), or saline and memantine (25 mg/kg; 4 h) systemic administrations in rat posterior cingulate and anterior retrosplenial cortices. Duplicate microarrays were hybridized and the ratios (treated vs control) were calculated for each individual array. Analysis of both arrays from each treatment revealed that 28 genes were differentially altered after memantine administration (Tables 1 and 3; Figures 1 and 2) while 34 genes were differentially altered after MK-801 administration (Tables 2 and 3; Figures 1 and 2). A total of 13 genes were regulated by both treatments of which 12 were upregulated and one was downregulated (Table 3).

Table 1 Profiling of Genes Altered by Memantine Administration (i.p.; 25 mg/kg; 4 h) in Rat Limbic Cortex
Full size table
Table 3 The Number of Genes Similarly and Differentially Regulated by Memantine and MK-801 in Rat Limbic Cortex by Microarray and ISH Analysis
Full size table
Figure 1
figure 1

Acute systemic administration of memantine and MK-801 regulated gene expression of candidate genes in rat brain. Representative autoradiograms from ISH showing differential gene expression of RL/IF-1 (a, d), GDI-1 (b, e), and NVP-2 (c, f) after saline (Sal), memantine (Mem), and MK-801 (MK) treatments. Retrosplenial cortex (RSC), temporal cortex (TemC), dentate gyrus (DG), hippocampal regions (CA1 and CA3).

Full size image
Figure 2
figure 2

The effect of acute memantine (25 mg/kg) and MK-801 (1 mg/kg) systemic administration in rat brain. The rats were i.p. injected with 25 mg/kg of memantine or 1 mg/kg of MK-801 and killed 4 h later. In all, 14 μm coronal brain sections were hybridized by radioactive ISH with oligonucleotide probes. Exposure time for autoradiograms was 1–3 weeks. Histograms of the region-expression levels are shown for memantine (a) and for MK-801 (b) acute administration. Data are mean±SEM obtained from five to six brains in each group. Significant differences between saline and treated rats are indicated: *p<0.05; **p<0.01, ANOVA followed by Dunnett's post hoc test. Retrosplenial cortex (RSC), temporal cortex (TemC), dentate gyrus (DG), hippocampal regions (CA1 and CA3).

Full size image
Table 2 Profiling of Genes Altered by MK-801 Administration (i.p.; 1 mg/kg; 4 h) in Rat Limbic Cortex
Full size table

ISH Analysis of Candidate Genes

ISH was used to independently verify the results from previous microarray analyses as well as to determine brain expression distribution, dose response, and time course. ISH from five to six rat brains treated with MK-801 (i.p.; 1 mg/kg; 4 h) or memantine (i.p.; 5–25 mg/kg; 4 h) was performed for several up- or downregulated candidate genes. Candidate genes selected for ISH analysis were based on groups of genes regulated by both treatments, by only memantine, or by only MK-801.

ISH Analysis of Candidate Genes after Memantine Treatment

From the microarray analysis, 28 candidate genes were identified of which five were previously confirmed by the literature, and 10 were confirmed by ISH with one false positive (Tables 1 and 3). Moreover, we confirmed the regulation of memantine selectively altered genes: PAK-1, SHPS-1, calcium channel β, and F-spondin. Additionally, one false negative was observed (c-fos). Quantification of these results can be found as supplemental material (http://www.uku.fi/aivi/supplementary_information.htm).

Regenerating liver inhibitory factor-1 (RL/IF-1), GDI-1, SHPS-1, PAK-1, and neural visinin Ca2+-binding protein 2 (NVP-2) were further subjected to study memantine-induced dose response and time course. mRNA levels of RL/IF-1 gene was significantly increased in all investigated brain regions (32±3.4%) (Figures 1a and 2a). Gene expression levels of GDI-1 were significantly increased in retrosplenial cortex with an increase of 25±4.4% and in all studied hippocampal regions with an increase of 29±4.4% (Figures 1b and 2a). NVP-2 mRNA level was decreased in retrosplenial and temporal cortices (86±1.0%), and 22±5.2% increased in DG (Figures 1c and 2a). Moreover, the gene SHPS-1 was significantly increased 14±1.0% in retrosplenial and temporal cortices, and 26±6.5% in DG (Table 4). Furthermore, gene expression of PAK-1 gene was induced with 28±4.4% increase in studied cortical regions, and 31±2.4% in hippocampal DG, CA1 and CA3 regions (Table 4).

Table 4 Memantine (M) (i.p.; 5–50 mg/kg; 4 h) Dose Response of Gene Expression in Rat Brain Regions
Full size table

A dose response to memantine (5–50 mg/kg) was performed with the RL/IF-1, GDI-1, SHPS-1, and PAK-1 genes (Table 4). The lowest dose of memantine used, 5 mg/kg, which can be considered clinically relevant, was able to increase significantly mRNA levels of GDI-1 and PAK-1 in several regions of rat hippocampus. The second lowest dose of memantine (i.p.; 10 mg/kg; 4 h) was able to increase significantly PAK-1 mRNA level in all studied cortical and hippocampal regions (Table 4). With increased dose of memantine, mRNA levels of RL/IF-1, GDI-1, and PAK-1 were also induced in other brain regions, and the effect was more pronounced. SHPS-1 mRNA levels were increased only after high doses of memantine (i.p.; 25–50 mg/kg; 4 h) (Table 4).

Peak induction of RL/IF-1 was observed in the cingulate and parietal cortices, and striatum (CPu) at 1 and 4 h after administration of memantine (i.p.; 25 mg/kg; 4 h) (Figure 3). Furthermore, peak induction of SHPS-1 mRNA was observed in retrosplenial cortex at 4 and 8 h after memantine (i.p.; 25 mg/kg; 4 h) administration (data not shown). Other brain regions may also be regulated by acute MK-801 and memantine administration but were not quantitated.

Figure 3
figure 3

The time course of memantine effect on RL/IF-1 mRNA expression. The rats were i.p. injected either with saline and killed after 4 h or with memantine (25 mg/kg) and killed 1, 4, 8, 24, or 48 h later. Optical densities were quantified by MCID/M4-analysis system from cingulate cortex (a), parietal cortex (b), and striatum (c) for RL/IF-1 gene. Values are expressed as percentage of saline (mean±SEM from three rats in each treatment group). Significant differences between saline and memantine treated rats are indicated: *p<0.05; **p<0.01, ANOVA, Dunnett's post hoc test.

Full size image

ISH Analysis of Candidate Genes of MK-801 Treatment

From the microarray analysis, 34 candidate genes were identified of which nine were previously confirmed by the literature, and seven were confirmed by ISH with two false positives (Tables 2 and 3). Moreover, we confirmed the regulation of MK-801 selectively altered genes: FGF-2, prostaglandin EP2 receptor, and dopamine D4 receptor. Additionally, one false negative was observed (HSP90β). Quantification of these results can be found as supplemental material (http://www.uku.fi/aivi/supplementary_information.htm).

Expression of RL/IF-1 was increased in retrosplenial and temporal cortices (38±6.5%) as well as in hippocampal regions (35±3.1%) (Figures 1d and 2b). mRNA expression of GDI-1 was also increased in retrosplenial cortex (33±5.9%) and CA3 hippocampal region (15±5.3%) (Figures 1e and 2b). NVP-2 mRNA level was decreased in retrosplenial and temporal cortices (85±2.5%). Moreover, an acute administration of MK-801 increased mRNA levels of the NVP-2 gene in hippocampal DG (15±5.9%) (Figures 1f and 2b).

Study of GDI-1 Protein Induction

Acute administration of memantine (i.p.; 10 and 25 mg/kg; 8 h) significantly increased protein levels of Rab/GDI-1 complex (∼84 kDa) in the cytosolic fraction (29±5.3%; 64±6.8%), and also in the membrane fraction of retrosplenial cortex (33±4.2%; 52±1.8%) (Figure 4). Cytosolic and membrane levels of free GDI-1 (∼56 kDa) were unchanged after memantine administration.

Figure 4
figure 4

Western blot analysis of GDI protein expression in retrosplenial cortex after memantine treatment. The rats were injected with memantine (i.p.; 10 or 25 mg/kg) for an 8 h time period, and protein from the dissected retrosplenial cortex was obtained from cytosolic and membrane fractions. Proteins were run on SDS/polyacrylamide gels and transferred to nitrocellulose filters. The filters were probed either with anti-GDI-1 or anti-actin antibodies. Values are expressed as percentage of saline. Data shown are mean±SEM from three rats. *p<0.05; **p<0.01, significantly different compared to saline-treated rats (ANOVA, Dunnett's post hoc test).

Full size image

DISCUSSION

Agents acting as noncompetitive antagonists of NMDA/glutamate receptor induce the expression of several genes in limbic cortical regions, such as the cingulate, retrosplenial, and entorhinal cortices (Sharp et al, 1991; Gass et al, 1993; Tomitaka et al, 1996; Marvanová et al, 2001). In this study, we have applied expression profiling of MK-801 (1 mg/kg) and memantine (5–50 mg/kg) in the posterior cingulate and anterior retrosplenial cortices to look more carefully at induced genes, and to delineate molecular pharmacologic effects of NMDA/glutamate receptor antagonists in the brain.

We screened the expression of 1090 known rat genes representing about 2% of entire rat genome by using cDNA microarrays. We identified 13 genes regulated by both NMDA receptor antagonists of which 12 genes were upregulated and one was downregulated. In addition, 28 genes were regulated by memantine and 34 were regulated by MK-801 administration. Confidence in the microarray results are supported by (1) identification of genes previously known to be altered by NMDA receptor antagonist treatment, (2) verification of novel regulated genes by ISH, and (3) verification by immunoblotting.

Antagonist of NMDA receptor subtype of glutamate receptor, protect against brain damage in neurological disorders (Danysz et al, 1995; Parsons et al, 1998; Krystal et al, 1999). However, in the previous studies it was observed that administration of NMDA receptor antagonists with high affinity or at high doses, can produce psychotomimetic effect in humans and neurotoxicity in specific neuronal subpopulations in the CNS of rodents, in particular neurons of the limbic cortex. Injured neurons in layers III and IV of the posterior cingulate and anterior retrosplenial cortex are characterized by abnormal cytoplasmic vacuoles (Olney et al, 1989; Sharp et al, 1991). At a 1 mg/kg dose of MK-801, this vacuole formation is reversible and not detected after 24 h (Olney et al, 1989). Therefore, it is likely that not only genes related to beneficial effects of MK-801 and memantine treatments, but also genes related to neurotoxic or stress-response effects of cortical cells are revealed in this study. In this study, we observed increases in mRNA levels of c-fos and HSP70 in the retrosplenial and cingulate cortex of adult rat treated with 25 mg/kg memantine and 1 mg/kg MK-801. Moreover, we observed increased transcriptional production of FGF-2 after MK-801 treatment.

Genes coding for a diverse range of proteins whose function is related to secretion, synaptic vesicle trafficking, and regulation of membrane trafficking were shown to be upregulated by different NMDA receptor antagonist treatments. GDI-1 (RabGDI, GDI-α) transcript level was increased by both NMDA receptor antagonists. The lowest dose of memantine increased GDI-1 mRNA levels in the hippocampal subfields, CA1 and CA3, and also in the cortex after higher doses. Moreover, we observed significantly increased levels of Rab/GDI-1 protein complex in membrane and cytosolic fractions of rat retrosplenial cortices. However, since GDI-1 interacts with most Rab proteins of which approximately 40 have been identified, we were unable to determine the specific Rab protein(s) involved in this process. The production of GDI-1 protein is critical for vesicular membrane transport in both the endocytic and also exocytic pathways via regulation of Ras-like GTPases of the Rab family proteins (Pfeffer et al, 1995; Martinez and Goud, 1998; Luan et al, 1999; Seabra et al, 2002). Moreover, GDI-1 has an important role in vivo to suppress hyperexcitability of the CA1 pyramidal neurons, which changes the firing properties of the neurons (Ishizaki et al, 2000).

Consistent with previous reports demonstrating that MK-801 and memantine stimulate hypothalamic–puituitary–adrenal activity by activating corticotropin-releasing factor or cortocoliberin (CRF) in the hypothalamus (Lee et al, 1994; Zhou et al, 1998), we found the increased transcript levels of CRF after both treatments. Neuropeptide Y (NPY) and neuromedin B receptor were also upregulated by both treatments, whereas somatostatin receptor 2 was downregulated after MK-801. Taken together, these results suggest effects on multiple neuropeptide systems.

Memantine and MK-801 elicited also induction of the immediate early gene RL/IF-1 in several brain regions. The RL/IF-1 gene is the rat homolog of human MAD-3, and possesses I-κB activity of broad specificity and thus it is able to inhibit the binding of p50–p60 NF-κB, c-Rel-p50, and RelB-p50, but not homodimeric NF-κB to the κB site (Tewari et al, 1992). NF-κB is a ubiquitous transcription factor that regulates transcription of multiple genes involved in immune and inflammatory responses (Baeuerle and Henkel, 1994). Previous studies have reported that inappropriate regulation of NF-κB/Rel-mediated transcription is associated with pathological conditions including acute inflammatory responses, toxic/septic shock, acute phase reactions, atherosclerosis, radiation damage, and also with several neuropathologies (Grilli et al, 1993; Baldwin, 1996; O’Neill and Kaltschmidt, 1997). In post-mortem tissue from patients with neurodegenerative disease or cultured neurons exposed to neurotoxic stimuli, increased NF-κB activity in cells has been reported (Kaltschmidt et al, 1995; Terai et al, 1996; Hunot et al, 1997). Furthermore, several studies have demonstrated that stimulation of both NMDA and non-NMDA glutamate receptors strongly stimulate NF-κB in vitro (Guerrini et al, 1995; Kaltschmidt et al, 1995). Additionally, another gene involved in immune and inflammatory responses, prostaglandin EP2 receptor (Narumiya et al, 1999), was downregulated.

This study revealed several new regulated transcripts in the brain after acute memantine or MK-801 treatment. While a 25 mg/kg dose of memantine causes more general expression changes in RL/IF-1 and GDI-1 transcripts, as shown by representative autoradiograms, regional changes were observed only at lower doses of memantine. These findings suggest caution in over extrapolation of gene expression changes, especially in cases like RL/IF-1 and GDI-1, where the ISH expression pattern is diffuse and regional differences may be subtle and revealed only by lower doses. Moreover, these changes were observed and reported only after acute administration. In a clinical setting of AD treatment, memantine is used chronically in relatively low doses. The 25 mg/kg dose results in plasma levels reaching ∼6.5 μM concentration that are still specific to NMDA receptor antagonism, although it is approximately seven times higher than the typical therapeutic values (∼1 μM) (Kornhuber et al, 1994). The current study was designed to identify early effects of memantine treatment and potential new drug targets for NMDA receptor antagonists. Thus, future studies in a rat model should use chronic administration of lower memantine doses (1–5 mg/kg) to mimic the clinical regimen, and in this way reveal gene candidates responsible for its beneficial therapeutical effect. Moreover, as the rat genome project advances toward completion, whole genome expression array analysis could eventually be possible to perform.

In summary, we have used microarrays to study differentially expressed genes in response to acute treatments of high and modest affinity NMDA receptor antagonists. This study identified new genes not previously implicated in the molecular pharmacologic actions of high affinity and medium affinity NMDA receptor channel blockers, MK-801 and memantine, in the brain. Some caution should be taken in extrapolation of our results since only a portion of the microarray results were confirmed by independent methods. Nonetheless, genes regulated by a clinically relevant dose of memantine might be important for its therapeutic effects, and could potentially constitute new targets for investigation. Further and more detailed characterization of these gene relationships to neural function may lead to a better understanding of neurodegenerative processes and ultimately their treatment.