Abstract
The neural cell adhesion molecule (NCAM) is the predominant carrier of α2,8 polysialic acid (PSA) in the mammalian brain. Abnormalities in PSA and NCAM expression are associated with schizophrenia in humans and cause deficits in hippocampal synaptic plasticity and contextual fear conditioning in mice. Here, we show that PSA inhibits opening of recombinant NMDA receptors composed of GluN1/2B (NR1/NR2B) or GluN1/2A/2B (NR1/NR2A/NR2B) but not of GluN1/2A (NR1/NR2A) subunits. Deficits in NCAM/PSA increase GluN2B-mediated transmission and Ca2+ transients in the CA1 region of the hippocampus. In line with elevation of GluN2B-mediated transmission, defects in long-term potentiation in the CA1 region and contextual fear memory in NCAM/PSA-deficient mice are abrogated by application of a GluN2B-selective antagonist. Furthermore, treatment with the glutamate scavenger glutamic-pyruvic transaminase, ablation of Ras-GRF1 (a mediator of GluN2B signaling to p38 MAPK), or direct inhibition of hyperactive p38 MAPK can restore impaired synaptic plasticity in brain slices lacking PSA/NCAM. Thus, PSA carried by NCAM regulates plasticity and learning by inhibition of the GluN2B-Ras-GRF1-p38 MAPK signaling pathway. These findings implicate carbohydrates carried by adhesion molecules in modulating NMDA receptor signaling in the brain and demonstrate reversibility of cognitive deficits associated with ablation of a schizophrenia-related adhesion molecule.
Introduction
The neural cell adhesion molecule (NCAM) is expressed by all cell types in the nervous system, where it promotes neuron-neuron and neuron-glia adhesion via homophilic and heterophilic interactions with other cell adhesion and extracellular matrix molecules. It is involved in signal transduction and critical developmental events, such as cell migration and proliferation, neurite outgrowth and fasciculation, synaptogenesis and synaptic plasticity (Dityatev et al., 2000; Maness and Schachner, 2007). Functions of NCAM depend on its glycosylation, particularly on polysialylation, i.e., attachment of long homopolymers of α2,8-linked sialic acid residues, known as polysialic acid (PSA) (Rutishauser and Landmesser, 1996; Dityatev et al., 2004; Weinhold et al., 2005). PSA expression is regulated in an age-, cell type- and activity-dependent manner, being highest during early development and in brain regions associated with neuroplasticity in the adult. A growing body of evidence has implicated NCAM and PSA as risk factors for major neuropsychiatric and neurodegenerative disorders, such as schizophrenia, bipolar disorder, depression, anxiety disorder, and Alzheimer's disease (Brennaman and Maness, 2008). This is in line with increasing appreciation of the roles played in brain malfunctions by other cell adhesion molecules, such as the NCAM-associated cell adhesion molecule L1 and neuroligins (Dahme et al., 1997; Südhof, 2008). Particularly noteworthy are the drastically decreased expression of polysialylated NCAM (PSA-NCAM) in hippocampi of patients with schizophrenia (Barbeau et al., 1995) and a link between NCAM and neurocognitive changes in schizophrenia (Sullivan et al., 2007).
These data agree with the widely recognized role of NCAM and its associated PSA as an important mediator of synaptic plasticity in the adult (Lüthi et al., 1994; Muller et al., 1996; Bukalo et al., 2004). Learning and synaptic plasticity in vitro and in vivo are altered in the CA1 and CA3 hippocampal subfields and/or in the dentate gyrus of mice deficient for NCAM (Cremer et al., 1994, 1998; Muller et al., 1996; Senkov et al., 2006; Stoenica et al., 2006). In different types of synapses, synaptic plasticity is modulated by NCAM through either PSA-dependent or PSA-independent mechanisms. Enzymatic removal of NCAM-associated PSA and genetic ablation of polysialyltransferase ST8SiaIV—required for polysialylation of NCAM in the adult hippocampus—lead to learning defects and impairment of long-term potentiation (LTP) and long-term depression (LTD) in the CA1 region of the hippocampus, but not in mossy fiber–CA3 synapses or in perforant path synapses of the dentate gyrus (Becker et al., 1996; Muller et al., 1996; Eckhardt et al., 2000; Senkov et al., 2006). Application of soluble PSA-NCAM to acute slices of NCAM-deficient mice restores normal CA1 LTP and improves contextual fear memory (Senkov et al., 2006). Since application of non-polysialylated NCAM is not effective in restoring normal CA1 LTP, the combined data indicate that PSA is both necessary and sufficient for normal induction of CA1 LTP.
However, the mechanisms by which NCAM regulates synaptic plasticity in a PSA-dependent manner have remained elusive. In light of data suggesting that PSA prevents activation of GluN2B (former NR2B; for the modern nomenclature, see Collingridge et al., 2009)-containing NMDA receptors at low concentrations of glutamate (Hammond et al., 2006), we have now investigated GluN2B-mediated neurotransmission in NCAM/PSA-deficient hippocampal slices and its impact on synaptic plasticity and contextual fear learning in NCAM/PSA-deficient mice. By restoring plasticity and learning in NCAM/PSA-deficient mice through inhibition of GluN2B and downstream signaling, we demonstrate reversibility of abnormalities associated with the deficit in a schizophrenia-related cell adhesion molecule.
Materials and Methods
Animals.
Constitutively NCAM-deficient (NCAM−/−) mice (Cremer et al., 1994) and their wild-type (NCAM+/+) littermates were inbred onto the C57BL/6J genetic background for 10 generations. Two- to 3-month-old mice of both sexes were used for electrophysiological and biochemical experiments. Three-month-old Ras-GRF1-deficient mice (Giese et al., 2001), inbred onto the C57BL/6J genetic background for at least 6 generations, were used for electrophysiological recordings.
For behavioral experiments, we used 3- to 5-month-old NCAM−/− males and their littermates. At least 1 week before starting the experiments, mice were transferred from the animal facility of the Zentrum für Molekulare Neurobiologie to a small vivarium at the Department of Neurophysiology and Pathophysiology, where they were housed individually with food and water ad libitum on a reversed 12:12 light/dark cycle (light on at 11:00 P.M., and all behavioral experiments were performed in the afternoon, during the dark phase of the cycle when mice are active) under constant temperature (22 ± 1°C) and humidity (55 ± 5%). All treatments and behavioral procedures were approved by the Food and Veterinary Office in a governmental body for Social Affairs, Family, Health and Consumer Protection in Hamburg, by the Office of Animal Welfare at the Department of Veterinary Public Health, Nutrition and Food Safety in Rome, or have been performed according to RIKEN regulations.
Single-channel recordings from recombinant NMDA receptors.
GluN1-1a with an HA tag in the pEGFP-N1 vector (Holmes et al., 2002) was a kind gift from Kevin Holmes and Gregory Dekaban (University of Western Ontario, London, Ontario, Canada). GluN2A in the pRK5 vector and GluN2B in the pRK7 vector were kind gifts of Peter Seeburg (Max-Planck Institute for Medical Research). Chinese hamster ovary (CHO) cells were transiently cotransfected with 1 μg of GluN1 and 1 μg of GluN2A, or GluN2B, or 0.5 μg of GluN2A plus 0.5 μg of GluN2B cDNAs per a well of 6-well plates using Lipofectamine (Invitrogen), as described by the manufacturer. The expression of GluNs was confirmed by Western blotting. The transfected CHO cells were harvested 48 h after transfection to prepare the plasma membrane fraction for reconstitution in lipid bilayers using the “tip-dip” method as described previously (Hammond et al., 2006), with minor modifications. The bilayer formation was initiated by successive transfer of two monolayers onto the tip of the patch pipette in an asymmetric saline condition with “outside-out” configuration. The bath solution contained pseudoextracellular fluid composed of the following (in mm): 125 NaCl, 5 KCl, 1.25 NaH2PO4, 0.001 glycine, and 5 Tris HCl. The pseudointracellular fluid, consisting of the following (in mm): 110 KCl, 4 NaCl, 2 NaHCO3, 1 MgCl2, 0.1 CaCl2, and 2 3-N-morpholinopropanesulfonic acid, was used as the pipette solution. Single-channel currents were elicited by application 0.3, 1, and 3 μm glutamate (here and elsewhere unless specified otherwise, Sigma). Increasing glutamate concentrations elevated the open probability rather than the conductance of NMDA receptors. GluN1/2A, GluN1/2B, and GluN1/2A/2B currents were recorded at +85 mV, +92 mV and +75 mV, respectively. At these voltages we observed patches containing only single channels with many events as possible and a stable base line. After recordings from channels derived from cells transfected with GluN1 plus GluN2A or GluN1 plus GluN2B subunits, 10 μm APV or 0.25 μm Ro 25-6981 was added to block activity, respectively, and, thus, to confirm the identity of recorded channels as NMDA receptors. To isolate the activity of GluN1/2A/2B receptors from that of GluN1/2A and GluN1/2B receptors during recordings from channels derived from cells transfected with GluN1 plus GluN2A and GluN2B subunits, 50 nm ZnCl2 and 0.25 μm Ro 25-6981 were present in the extracellular solution during all recordings. To determine these concentrations of Ro 25-6981 and ZnCl2 (needed to block GluN1/2B and GluN1/2A receptors, respectively), we recorded GluN1/2B single-channel activity (n = 5) elicited with 3 μm glutamate and tested concentrations of Ro 25-6981 ranging from 0.05 to 0.25 μm, with increments of 0.05 μm, until channel activity was completely abolished (Fig. 1A). Similarly, for GluN1/2A, 3 μm glutamate and increasing concentrations of ZnCl2 (10–50 nm) with increments of 10 nm were tested until channels (n = 5) were completely blocked (Fig. 1B). After successful inhibition of both GluN1/2B and GluN1/2A receptors with Ro 25-6981 and ZnCl2, respectively, the voltage was changed to ensure that inhibition was not voltage-dependent. APV was added at the end of these recordings to confirm the identity of recorded channels as NMDA receptors. PSA (colominic acid, Fluka) was delivered to the extracellular fluid in increasing concentrations ranging from 1 μg/ml to 10 μg/ml, as indicated in the text, in the presence of 3 μm glutamate. Data processing of single-channel recordings was performed as described previously (Vaithianathan et al., 2005). Briefly, the all-point-current-amplitude histograms were constructed for 3000 ms traces, 5–6 per patch. The single-channel Po was calculated as Po = Ro/(Rc + Ro), where Rc and Ro represent areas under the current-amplitude histogram corresponding to closed and open states, respectively, fitted with the sum of two Gaussian curves using the Microcal Origin program. The single-channel chord conductance was calculated as Iamp/(Vm − Vrev), where Iamp is the single-channel current amplitude at the membrane potential Vm and Vrev is the reversal potential assumed to be 0 mV.
Recordings of NMDA receptor-mediated responses and LTP in hippocampal slices.
After brief CO2 sedation, decapitation and removal of the brain, sagittal hippocampal sections were cut with a Leica VT 1000M vibratome in ice-cold artificial CSF (ACSF) containing the following (in mm): 250 sucrose, 24 NaHCO3, 25 glucose, 2.5 KCl, 1.25 NaH2PO4, 2 CaCl2, 1.5 MgCl2, pH 7.3. The slices were then kept for at least 2 h before the start of recordings at room temperature in carbogen-bubbled ACSF, containing 120 mm NaCl instead of 250 mm sucrose. For isolation of NMDA receptor-mediated field EPSPs (fEPSPs), MgCl2 concentration was lowered to 0.25 mm during recordings. Recordings were performed in the same solution, in a submerged chamber, at room temperature (22−24°C). Recordings of fEPSPs were performed in the stratum radiatum of the CA1b subfield with glass pipettes filled with ACSF and having a resistance of 1–2 MΩ. Schaffer collaterals were stimulated with a glass pipette having a resistance of 1 MΩ and placed ∼300 μm closer to the CA3 subfield than the recording electrode. Basal synaptic transmission was monitored at 0.05 Hz. The intertheta burst stimulation (TBS) interval was 20 s and four TBSs were applied to induce LTP. TBS consisted of 8 bursts delivered at 5 Hz. Each burst consisted of four pulses delivered at 100 Hz. Duration of pulses was 0.2 ms and stimulation strength was set to provide baseline fEPSPs with amplitude of ∼50% from the subthreshold maximum.
The data were recorded at sampling rate of 10 kHz, low-pass filtered at 3 kHz and analyzed using EPC10 amplifier and Pulse software (both from Heka). SigmaPlot (Systat Software Inc.) was used for data presentation and digital subtraction of responses elicited by single pulses for precise measurements of paired-pulse facilitation (defined as the ratio between the slopes of fEPSPs elicited by the second and first pulses).
Enzymatic and pharmacological treatments of hippocampal slices.
To remove PSA before electrophysiological recordings of LTP, acute hippocampal slices from C57BL/6J mice were incubated for 2 h at 35°C in 2 ml of carbogen-bubbled ACSF supplemented by recombinantly produced active (truncated by 245 amino acids) form of PSA-degrading endosialidase of bacteriophage K1F (endoNF; 9.6 μg/ml) (Stummeyer et al., 2005). The slices incubated without any enzyme or with a mutated (two substituted amino acids) inactive form of endoNF (9.6 μg/ml) (Stummeyer et al., 2005) served as controls. To verify removal of PSA by endoNF and no changes in expression of PSA by inactive endoNF, all treated slices were fixed overnight in 4% formaldehyde in phosphate buffered saline, pH 7.3, cut in 30 μm-thick subslices and immunostained for PSA using a monoclonal antibody (clone 735) (Eckhardt et al., 2000). The selected concentration of endoNF guaranteed removal of PSA in all recorded slices (supplemental Fig. 1, available at www.jneurosci.org as supplemental material).
The following drugs were used in our experiments: [(R)-[(S)-1-(4- bromo-phenyl)-ethylamino]-(2,3-dioxo-1,2,3,4-tetrahydro-quinoxalin- 5-yl)-methyl]-phosphonic acid (NVP-AAMO77, 0.25 μm, generously supplied by Dr. Yves Auberson from Novartis Pharmaceuticals), an antagonist producing preferential inhibition of GluN2A-containing receptors; R-(R*,S*)-(-(4-hydroxyphenyl)-(R)-methyl-4-(phenyl-methyl)-1-piperidinepropranolol (Ro 25-6981 hydrochloride, 0.5 or 5 μm, as indicated in the text), as a GluN2B-selective antagonist; dl-2-amino-5-phosphonovaleric acid (APV, 50 μm; Tocris Bioscience) as a subtype unselective antagonist of NMDA receptors; glutamic-pyruvic transaminase (GPT, 5 U/ml), as glutamate scavenger, dialyzed against ACSF and coapplied with 2 mm pyruvate; PSA in the form of colominic acid sodium salt from Escherichia coli (10 μm; Fluka); 4-(4-fluorophenyl)-2-(4-methylsulfinylphenyl)-5-(4-pyridyl)-1H-imidazole (SB 203580, 10 μm); bicuculline methiodide (20 μm), as antagonist of GABAA receptors; and (2S)-3-[[(1S)-1-(3,4-dichlorophenyl)ethyl]amino-2-hydroxypropyl] (phenylmethyl) phosphinic acid (CGP 55845, 10 μm; Tocris Bioscience), as antagonist of GABAB receptors. In all experiments with pharmacological rescue of LTP, the slices were perfused with a drug-containing ACSF immediately after a transfer to the recording chamber, i.e., ∼40 min before induction of LTP.
For isolation of the NMDA receptor-mediated component ([2] in Fig. 2A), the AMPA receptor-mediated fast component of the fEPSP [1], the amplitude of which was taken as 100%, was blocked by perfusion with ACSF containing 0.25 mm Mg2+, the AMPA/kainate receptor antagonist NBQX (10 μm) and the GABAA/glycine receptor antagonist picrotoxin (50 μm). Application of 0.25 μm NVP or 5 μm Ro 25-6981 isolated fEPSPs mediated predominantly by GluN2B- or GluN2A-containing receptors [3], respectively, which could be blocked to 1–1.5% level by the NMDA receptor antagonist APV (50 μm) [4]. The ratio between amplitudes of NMDA and AMPA receptor-mediated components, was computed as ([2] − [4])/([1] − [4]), and the amplitude of NVP- or Ro 25-6981-insensitive component was computed as ([3] − [4])/([2] − [4]), where numbers refer to the aforementioned components (shown for NVP-insensitive component in Fig. 2A).
For statistical comparisons of LTP levels in different groups, measurements for these groups were performed in an interleaved manner and treated using ANOVA followed by multiple comparisons of groups with the post hoc Bonferroni t test (SigmaStat, Systat Software Inc.). When raw data did not pass the normality test, ANOVA was performed on ranks.
Two-photon Ca2+ imaging.
Hippocampal slices were obtained from 6- to 7-week-old Sprague Dawley rats. Slices were incubated for 2 h prior recording and then transferred to the recording chamber and superfused using a custom micro-perfusion system with a total volume of 2 ml and a flow rate of 2–3 ml/min. During recordings the slices were superfused with carbogenated Mg2+-free ACSF containing the following (in mm): 120 NaCl, 2.5 KCl, 1 NaH2PO4, 26.2 NaHCO3, 2 CaCl2, 11 glucose, 2 Na-pyruvate, and 0.25 4-methoxy-7-nitroindolinylglutamate (MNI-Glu; Tocris Bioscience). AMPA and kainate receptors were blocked with 20 μm CNQX (Tocris Bioscience); mGluR5 were blocked with 5 μm Fenobam (Tocris Bioscience). Action potentials were blocked with 2 μm tetrodotoxin. Whole-cell current-clamp recordings were made from CA1 pyramidal cells at room temperature (23−24°C). The intracellular solution contained the following (in mm): 120 K-methanesulfonate, 10 HEPES, 4 MgCl2, 10 Na-phosphocreatine, 0.4 NaGTP, 4 Na2ATP, 0.1 Alexa 594 (Invitrogen), 0.250 Fluo4 (Invitrogen). Recordings were discarded if the baseline Fluo4 fluorescence had changed >25% or the membrane potential had changed >5 mV during the course of the experiment. Uncaging of MNI-Glu was done for 20 ms at a single spot with a 405 nm Laser (vortex mode of Olympus FV1000 Laser Scanning Microscope) at aspiny parts of dendrites to elicit an NMDA receptor-mediated response. Ca2+-dependent fluorescence of Fluo4 was measured with line-scan two-photon imaging (excitation at 810 nm) before and after uncaging. Active or inactive forms of endoNF were applied to slices at a concentration of 576 μg/ml via pipettes with a diameter of tips ∼30 μm. In the series of experiments with the active form of endoNF, 2 μg/ml colominic acid was sequentially added to the bath to inhibit responses potentiated by endoNF. At the end of each experiment, 10 μm Co101244 (Tocris Bioscience), as selective GluN1/2B antagonist (Zhou et al., 1999; Kohl and Dannhardt, 2001), was applied, which largely blocked the response. Data are presented as mean ± SEM and statistically evaluated using ANOVA with repeated measures and the post hoc Bonferroni t test.
Analysis of p38 and p42/p44 phosphorylation.
Whole hippocampi were isolated from brains of adult mutants and control littermates for homogenization in ice-cold lysis buffer containing 10 mm Tris-Cl, pH 7.5, 150 mm NaCl, 1 mm EDTA, 1 mm Na-orthovanadate and 1 mm PMSF plus 1% NP-40, 0.1% SDS and protease inhibitor cocktail (Roche Applied Science). This buffer was also used for homogenization of inactive and active endoNF-treated hippocampal slices (4 hippocampal slices from one mouse were pooled for one sample). Crude protein extracts were obtained by clearing the lysates by centrifugation at 13,000 × g for 20 min at 4°C. Proteins were subjected to SDS-PAGE and transferred onto nitrocellulose membrane (Protran, Schleicher & Schuell). Following pretreatment in 5% nonfat dry milk powder in 0.1% Tween 20 in Tris-buffered saline, pH 7.5 (TBS), membranes were incubated in 5% BSA in TBS containing anti-phospho-p38 (Thr180/Tyr182), anti-phospho-p44/42 (Thr202/Tyr204) or anti-total p38 and anti-total p44/42 antibodies (1:1000; Cell Signaling Technology). After washing in 0.1% Tween 20 in TBS, blots were incubated with horseradish peroxidase-conjugated secondary antibody (1:10,000). Blots were stripped and reblotted with GAPDH antibody as a loading control. Immunoreactivity was detected by enhanced chemiluminescence (ECL kit, GE Healthcare) on Kodak Biomax x-ray film (Sigma) according to the manufacturer's instructions. Band intensities were quantified by densitometry and normalized using intensities of NCAM+/+ probes from the same blot (for NCAM−/− hippocampi) or intensities of inactive endoNF-treated slices from the same mouse (for active endoNF-treated samples). Data are presented as mean ± SEM and compared using ANOVA with repeated measures, followed by the post hoc Bonferroni t test.
Fear conditioning.
To assess both contextual and cued fear memories, an auditory fear conditioning paradigm was used (Senkov et al., 2006). Fear conditioning was performed as follows: (1) after 2 d of handling and habituation in a neutral context (home cage) for 5 min each day, the mouse was injected intrahippocampally with Ro 25-6981, or with vehicle as control, 15 min before training (Zhao et al., 2005); (2) subsequently, the mouse was placed into a conditioning chamber bearing conditioned context (CC+) and allowed to freely explore this context for 180 s. Freezing levels during this time were taken as a baseline; (3) in a next step, the mouse was subjected to two pairings of 20 s episodes of conditioned stimulus (CS+) presentation (delivered to half of the mice as a patterned tone, i.e., a series of 50-ms-long 75 dB tones of 2.5 kHz at 1 Hz, and to the other half of the mice as a continuous tone of the same frequency) separated by 60 s and coterminated with a footshock as unconditioned stimulus (US); (4) 60 s after the last shock, the mouse was returned to its home cage (supplemental Fig. 2, available at www.jneurosci.org as supplemental material).
The following tests were performed on days 1 and 7 after induction of fear conditioning: (1) the mouse was placed into the conditioning chamber bearing context CC+ used for fear conditioning for 3 min to assess its retention of contextual memory and then returned to its home cage; (2) after ∼2–3 h the same mouse was placed in a new neutral context (CC−), a tone cage (see below), and was allowed to freely explore this context for 180 s; in this way we could measure the level of generalization, as freezing in response to CC− and discrimination between CC+ versus CC−; (3) a 60 s episode of CS+ was presented (as patterned or continuous tone previously paired with a foot shock) to assess tone memory; 120 s later a 60 s presentation of unpaired CS− tone (either patterned or continuous, but different from CS+) was given; (4) 60 s after the last tone presentation the mouse was returned to its home cage.
Two fear conditioning chambers with two distinct CC+ contexts were used for conditioning: an aluminum chamber which was cleaned before placing of a mouse with 75% ethanol, and a Plexiglas chamber which was cleaned with 1% acetic acid in water (supplemental Fig. 2, available at www.jneurosci.org as supplemental material). There were two tone cages bearing neutral CC− context: a tone cage with removable celled walls flavored with vanilla bakery oil, and another one containing removable circled walls flavored with lemon bakery oil, both cages were cleaned with water before placing of a mouse. One of two distinct CC+ and one of two distinct CC− were presented to the mouse to minimize dependency of results on a particular choice of CC+ and CC−. Patterned and continuous tones were used as CS+ and CS− because mice had been observed to show a good discrimination between these tones in a fear conditioning paradigm (Tang et al., 2003; Senkov et al., 2006).
All experiments were performed in a double-walled sound-attenuated isolation cubicle (for details, see Senkov et al., 2006). The unconditioned stimulus was a 1 s, 0.7 mA (AC) middle intensity shock delivered to the rod floor. Delivery of tones and shocks was controlled by customized software (Signal 1.88, CED, Cambridge, UK) using a Pentium-II computer through an AD/DA converter (Micro 1401 board, CED, Cambridge, UK).
Behavior of mice during conditioning and tests was recorded on a Pentium-4 computer, using a video frame-grabber (Pixelsmart, CA) which digitized the video streaming with 3 frames per second and stored it on the hard-drive as a set of 8-bit gray images. These images were then converted into “avi” format movies by the freely available Java-based program ImageJ (http://rsb.info.nih.gov/ij). All recorded movies were then analyzed using Microsoft Windows Media Player 9 by a trained observer blind to genotype and treatment of the respective mouse. Freezing time, as a measure of fear-related memory, was quantified as described previously (Tang et al., 2001). Briefly, freezing was detected if the animals remained motionless, except for respiratory movements, in a tense posture that was characterized by horizontal positioning of the head, a stretched state of the body and stiffening of the tail.
Data are expressed as mean ± SEM. To analyze the effects of the injected compounds on contextual memory, we used ANOVA (Systat, Systat Software GmbH) having genotype and Ro 25-6981 treatment as between group factors, and day and context or tone as within group factors. ANOVA followed by the post hoc Bonferroni t test if significant effects of factors were detected. Additionally, statistical comparison between freezing responses to CS+ and CS− on a given day within each group was performed using the nonparametric Wilcoxon signed-rank test. Significant differences between groups were detected by the Wilcoxon test and the t test for the same groups, and we therefore refer in the text only to p-values provided by the Wilcoxon test. Statistical significance of differences was accepted if p < 0.05.
Surgical procedures.
Chronic cannulation of mice was performed as described previously (Senkov et al., 2006). In brief, mice were anesthetized with a mixture of Ketamin and Rompun (Ketamin 100 mg/kg, Rompun 16 mg/kg, 0.01 ml/g body weight, i.p.), placed in a stereotaxic frame (Narishige), their scalps were incised and retracted and two small holes were drilled. Bilaterally, sterile stainless steel 23 gauge double guide cannulas (Plastics One) were stereotaxically implanted reaching the border of the dorsal hippocampus under following coordinates (AP: −2 mm and ML: ±1.5 mm from the bregma; DV, −1.5 mm from the skull surface) according to Paxinos (2001). For NCAM−/− mice, AP was −1.5 mm, because their brain is shorter than normal due to abnormal development of olfactory bulbs and ventricles (Cremer et al., 1994). Guide cannulas were fixed on the skull bone by fast dental kallocryl cement (CPGM rot; Speiko). After the operation lasting for ∼30 min, mice were placed back into their home cages and kept under a warm (∼37°C) lamp until awakening. Metamizol (50 mg/kg body weight, i.p.) was used as a postoperative analgesic. The following behavioral investigations were performed after the mice had fully recovered 5–7 d after surgery.
Drugs for behavioral studies.
To test whether the block of GluN2B-containing receptors in the hippocampus would improve learning in NCAM−/− mice, we used the GluN2B selective antagonist Ro 25-6981 at a concentration of 20 μg/μl (0.5 μl/site/min), which is close to the concentration previously used for injection into the hippocampus and shown to be ineffective in fear conditioning in wild-type mice (Zhao et al., 2005). The bilateral injection cannulas (#C235, Plastics One) were connected to a pump (TSE Systems) through a Tygon tubing system. The injection cannulas were gently inserted into the bilateral guide cannulas ∼10 s before injection, extending the tips of the guide cannulas by 1 mm. The site of the injection was aimed to the hilus of the left and right hippocampi. The injection cannulas were left within the guide cannulas for at least 60 s after injection.
Since fear conditioning is sensitive to stress, especially when it occurs before training, special care was taken to minimize stress during i.p. or intrahippocampal injection procedures. To this end, all mice were slightly sedated with a mixture of CO2 and room air for the period of injection, i.e., approximately for 2 min. Recovery after anesthesia was fast (2–3 min) and without any side effects in the following fear conditioning experiments (Senkov et al., 2006).
Histology.
To verify the injection sites, brains were investigated histologically after the behavioral tests. Mice were reinjected with solution containing the dye methylene blue, using the same injection procedure as used for injection of Ro 25-6981. Fifteen minutes after this reinjection the mice were deeply anesthetized with pentobarbital (100 mg/kg body weight) and transcardially perfused with 4% paraformaldehyde in PBS, pH 7.4, during 20 min. Brains were postfixed overnight in the same fixative at 4°C. The brains were then cut into 100 μm thin slices near the injecting area. Slices were then inspected microscopically and photographed.
Results
PSA inhibits GluN1/2B and GluN1/2A/2B receptors
Our previous pharmacological study revealed that PSA inhibits GluN2B-containing receptors in primary hippocampal cultures (Hammond et al., 2006). In the mature hippocampus, NMDA receptors predominantly have the following composition: GluN1/2A, GluN1/2B and GluN1/2A/2B (Köhr, 2006). To identify which of these NMDA receptor subtypes are inhibited by PSA, we transfected Chinese hamster ovary cells with these three combinations of GluN subunits, prepared membranes from transfected cells, fused these with lipid bilayer membranes formed at the tip of patch pipette and compared activities of reconstituted NMDA receptors in the absence and presence of PSA. Single-channel activity was detected by an upward transition of the current, which represented the open state (Fig. 1A–C). The mean values of open probability and conductance of recorded channels are shown in Figure 1D. GluN1/2A/2B were identified by their low sensitivity to antagonists of GluN1/2A (50 nm ZnCl2) and GluN1/2B receptors (0.25 μm Ro 25-6981). Single-channel activities elicited by glutamate in GluN1/2B (Fig. 1A) and GluN1/2A/2B (Fig. 1C) receptors were sensitive to PSA. In contrast, addition of PSA had no significant effect on GluN1/2A receptors (Fig. 1B). Statistical evaluation revealed a progressive decrease in the open channel probability of GluN1/2B and GluN1/2A/2B receptors when the PSA concentration was increased from 1 to 2 and 3 μg/ml, without changes in single-channel conductance (Fig. 1E). After addition of 1, 2, and 3 μg/ml PSA, the open probability of GluN1/2B receptors decreased to 72.7 ± 1.1, 50.6 ± 1.1, 21.6 ± 0.7% of control levels, respectively (Fig. 1E). The open probability of GluN1/2A/2B receptors decreased after addition of 1, 2, and 3 μg/ml PSA to 73.8 ± 4.9, 59.4 ± 6.8, 33.4 ± 4.1%, respectively (Fig. 1E). ANOVA revealed that the inhibition by PSA was stronger for GluN1/2B than for GluN1/2A/2B receptors. Neither the open probability nor conductance of GluN1/2A receptors was affected by 1, 2, 3 and even by 10 μg/ml PSA (Fig. 1E). These data demonstrate that PSA selectively inhibits diheteromeric GluN1/2B and triheteromeric GluN1/2A/2B receptors.
NCAM−/− mice have increased GluN2B-mediated transmission in hippocampal CA3-CA1 synapses
Data from single-channel recordings suggests that the GluN2B-mediated component of synaptic transmission might be larger in NCAM-deficient (NCAM−/−) mice compared with wild-type control littermates (NCAM+/+ mice). We tested this hypothesis in hippocampal slices. The ratio between amplitudes of NMDA and AMPA receptor-mediated responses was not different between genotypes: 9.7 ± 1.7% in NCAM−/− mice versus 9.3 ± 0.5% in NCAM+/+ mice (Fig. 2A). Also, the basal AMPA receptor-mediated synaptic transmission and its paired-pulse modulation were normal in NCAM−/− mice (supplemental Fig. 3, available at www.jneurosci.org as supplemental material). To isolate the GluN2B-mediated component of postsynaptic responses, we used the NMDA receptor antagonist NVP at a concentration of 0.25 μm that inhibits ∼80% of currents mediated by GluN1/2A receptors and only ∼20% of currents mediated by GluN1/2B receptors (Bartlett et al., 2007). The amplitude of this component was larger in NCAM−/− mice (35.7 ± 4.9%) than in NCAM+/+ mice (22.0 ± 1.3%) (p < 0.05, t test). Complementary experiments with evaluation of the component insensitive to the GluN2B-specific antagonist Ro 25-6981 (5 μm) revealed a reduction of this presumably GluN2A-mediated component in NCAM−/− mice (64.0 ± 2.3%) compared with NCAM+/+ mice (74.8 ± 1.4%) (p < 0.01, t test). A sum of NVP- and Ro 25-6981-insensitive responses was on average 97% and 100% of total NMDA receptor-mediated response for NCAM+/+ and NCAM−/− mice, respectively. These data underline the validity of the procedure used here for dissection of GluN2A- and GluN2B-mediated components.
To test whether an increase in the GluN2B-mediated component reflects a deficit in inhibition of GluN2B-containing receptors by PSA in NCAM−/− mice, we applied PSA to hippocampal slices and found that it inhibited the NVP-insensitive component more strongly in NCAM−/− than in NCAM+/+ mice (20.1 ± 2.5% vs 12.0 ± 1.7%, respectively; p < 0.05, t test) (Fig. 2C). Together, these observations suggest that despite normal total NMDA receptor-mediated transmission in NCAM−/− mice, there is an increase in transmission via GluN2B.
Acute removal of PSA upregulates GluN2B-mediated Ca2+ transients in CA1 pyramidal cells
To further investigate whether a deficit of PSA in hippocampal slices would increase GluN2B-mediated Ca2+ influx, we performed recordings of Ca2+ transients in the aspiny, presumably extrasynaptic, part of CA1 apical dendrites (Fig. 3A) before and after application of the specifically PSA-digesting enzyme endosialidase NF (endoNF) (Stummeyer et al., 2005). As control, we used slices treated with an endoNF inactivated by point mutations in the active site (Stummeyer et al., 2005). Ca2+ influx was elicited by local uncaging of glutamate and recorded using the Ca2+-sensitive dye Fluo4 (Fig. 3B,C). Treatment with endoNF increased the integral of recorded Ca2+ signals to 209 ± 31% relatively to the signals recorded before treatment. In line with findings that PSA suppresses the opening of GluN2B-containing receptors only at low micromolar glutamate concentrations (Hammond et al., 2006), the difference between Ca2+ transients recorded in control and endoNF-treated cells was observed during the late phase of transients, i.e., when concentration of uncaged glutamate was low. PSA applied after endoNF treatment largely reversed the increase of the Ca2+ signal (137 ± 9% of signal before endoNF treatment) (Fig. 3B–D). Further application of Co101244, a selective GluN2B antagonist (Zhou et al., 1999), reduced the Ca2+ signals to 23 ± 6%, showing that they were predominantly mediated by GluN2B (Fig. 3B,D). Inactive endoNF did not significantly change the integral of Ca2+ signals (113 ± 8%), whereas further application of Co101244 reduced the responses to 19 ± 5% (Fig. 3B,D). These experiments directly visualize the acute upregulation of GluN2B-mediated Ca2+ transients in CA1 pyramidal cells after removal of PSA.
Impaired LTP in NCAM/PSA-deficient hippocampus can be restored via inhibition of GluN2B-containing receptors
Analysis of synaptic plasticity induced by theta-burst stimulation of Schaffer collaterals/commissural fibers revealed an impairment of LTP in CA1 of NCAM−/− mice (123.9 ± 2.3%) compared with NCAM+/+ mice (144.4 ± 4.1%) (Fig. 4A,C). Since our data demonstrate an increase in GluN2B-mediated transmission in NCAM−/− mice (Fig. 2) and since activation of GluN2B-containing receptors might promote synaptic depression (Massey et al., 2004), we next tested whether inhibition of GluN2B would normalize LTP in NCAM−/− mice. Indeed, application of the GluN2B-selective antagonist Ro 25-6981 (0.5 μm) increased LTP in NCAM−/− mice to 137.6 ± 4.9%, a value not significantly different from the level of 138.4 ± 3.3% recorded in parallel (i.e., in an interleaved manner) in NCAM+/+ mice (Fig. 4B,C). We next recorded LTP in the presence of a glutamate scavenger GPT, a substance known to substantially reduce concentrations of extrasynaptic glutamate (Tsvetkov et al., 2004) and, thus, to reduce activity of GluN2B-containing receptors. This treatment also increased LTP in NCAM−/− mice to 136.8 ± 3.2%, being not significantly different from the level of 147.3 ± 6.2% found in parallel recordings in NCAM+/+ mice (Fig. 4D–F). Thus, LTP in NCAM−/− mice can be fully restored by treatments that decrease transmission via GluN2B or reduce extrasynaptic glutamate concentration.
Similarly to ablation of NCAM, removal of PSA with recombinant endoNF (supplemental Fig. 1, available at www.jneurosci.org as supplemental material) impaired LTP (Fig. 5A,C), as previously reported for another preparation of endosialidase N (Becker et al., 1996; Muller et al., 1996). As a negative control, we show that inactive endoNF did not inhibit LTP and was not different from sham controls (Fig. 5A,C). Blockade of GluN2B with Ro 25-6981 (Fig. 5A,C) increased levels of LTP in endoNF-treated slices to levels not different from those observed after treatment with inactive endoNF or in sham controls (Fig. 5A,C). Thus, these data replicate results from the experiments in NCAM−/− mice and demonstrate that impaired LTP in either NCAM- or PSA-deficient slices is due to improper signaling via GluN2B.
One possible explanation for our results would be that elevated signaling via GluN2B expressed by interneurons in NCAM/PSA-deficient mice facilitates activation of interneurons and hence impairs LTP via enhanced activation of GABAA/GABAB receptors. Although normal amplitudes of unitary IPSCs were reported in hippocampal cultures of NCAM−/− neurons (Dityatev et al., 2000), no information is available regarding recruitment of interneurons during TBS in NCAM/PSA-deficient slices. Here, we investigated this issue by recording LTP in the presence of GABAA and GABAB receptor inhibitors, bicuculline and CGP 55845, in endoNF-treated slices. As the difference between endoNF-treated and control slices remained ∼20% after blockade of GABAergic inhibition (Fig. 5B,C), we conclude that it is unlikely that modulation of GABAergic inhibition mediates the effects of PSA on LTP.
Inhibition of enhanced p38 MAPK activity inNCAM/PSA-deficient hippocampus normalizes impaired LTP
As GluN2B-containing receptors are involved in both potentiation and depression of synaptic transmission (Tang et al., 1999; Massey et al., 2004), a critical question pertains to the downstream signaling mechanisms of PSA-dependent GluN2B-containing receptors. Previous reports highlighted p38 mitogen-activated protein kinase (MAPK) as a key player in synaptic depression (Bolshakov et al., 2000; Hsieh et al., 2006). We therefore investigated the levels of active MAPKs in the hippocampus of NCAM−/− mice and found an almost fourfold upregulation of p38 MAPK phosphorylation (Fig. 6A,B) and a twofold increase in phosphorylated p44 (Fig. 6A,B), whereas phosphorylation of p42 (Fig. 6A,B) and protein levels of p38, p42 and p44 (supplemental Fig. 4, available at www.jneurosci.org as supplemental material) were not changed. The level of phosphorylated p38 was also upregulated in active endoNF- compared with inactive endoNF-treated hippocampal slices (Fig. 6A,C), without changes in the protein level (supplemental Fig. 4, available at www.jneurosci.org as supplemental material). Upregulation of p38 phosphorylation in endoNF-treated slices was, however, lower than in NCAM−/− mice. This can be due to an increase in phosphorylated p38 in control slices during dissection of slices and/or due to low levels of basal spontaneous neuronal activity in acute slices (compared with intact brains), which would reduce the level of GluN2B and, hence, p38 activation in endoNF-treated slices. Still, during induction of LTP, the GluN2B-p38 pathway might be activated in endoNF-treated as well as in NCAM−/− slices by theta-burst stimulation-evoked glutamate release.
Indeed, inhibition of p38 MAPK activity with SB 203580 restored LTP in NCAM−/− mice to 137.2 ± 6.1% (Fig. 7A–C). Also, LTP was restored to normal levels by addition of SB 203580 to endoNF-treated slices (to 133.5 ± 3.7%) (Fig. 7D,E,G). SB 203580 had no effect on LTP in slices of wild-type mice (Fig. 7C,G).
Enhanced GluN2B-mediated activation of p38 MAPK via Ras-GRF1 results in reduction of LTP inNCAM/PSA-deficient hippocampus
The Ca2+-dependent Ras-guanine releasing factor (GRF) 1 may mediate activation of p38 MAPK by GluN2B-containing receptors (Li et al., 2006). We therefore studied LTP in mice deficient in Ras-GRF1. In contrast to wild-type mice (Fig. 7D), enzymatic removal of PSA in this mutant did not impair LTP (135.3 ± 2.9% in sham vs 133.3 ± 2.5% in endoNF-treated mutants) (Fig. 7F,G), suggesting that enhanced GluN2B-mediated activation of p38 MAPK via Ras-GRF1 accounts for reduction of LTP in the absence of NCAM/PSA.
Impaired contextual fear conditioning in NCAM−/− mice can be rescued via suppression of intrahippocampal GluN2B-containing receptors
NCAM−/− mice are impaired in contextual and cued (tone) memories in fear conditioning paradigms (Stork et al., 2000; Senkov et al., 2006). To verify the hypothesis that this deficit in hippocampus-dependent contextual learning is related to increased activity of GluN2B-containing receptors, we attempted to rescue learning deficits by injection of Ro 25-6981 into the hippocampi of NCAM−/− and NCAM+/+ mice 15 min before training and assessed its effects on contextual and cued fear memories on the first and seventh test days (Fig. 8A). Intrahippocampal injection was necessary in these experiments to exclude inhibition of GluN2B-containing receptors that are required for fear conditioning-associated neuroplasticity outside the hippocampus, for instance in the amygdala and prefrontal cortex (Rodrigues et al., 2001; Zhao et al., 2005). NCAM−/− mice injected with vehicle showed a twofold decrease in contextual memory compared with vehicle-injected NCAM+/+ mice at every test day (Fig. 8B). This robust learning deficit was fully rescued by intrahippocampal injection of Ro 25-6981, which normalized not only short- and long-term contextual memory, but also the animals' ability to discriminate between conditioned and neutral contexts (Fig. 8B). In contrast, cued memory in NCAM−/− mice was not improved by injection of Ro 25-6981 (Fig. 8B), although discrimination of tones was significantly facilitated on the seventh test day (Fig. 8B). In NCAM+/+ mice, injection of Ro 25-6981 had no significant effect on discrimination of contexts, although it had a tendency to reduce discrimination of tones.
Discussion
The present study provides evidence that PSA-NCAM restrains signaling via a subset of GluN2B-containing receptors. Our working model linking this phenomenon to a reduction in LTP can be summarized as follows: In wild-type animals, NCAM-associated PSA inhibits activation of GluN2B-containing receptors (Hammond et al., 2006), possibly by steric hindrance of the ligand to access the glutamate binding site. Importantly, this inhibition occurs only at low micromolar concentrations of glutamate (Hammond et al., 2006). In the absence of NCAM or PSA, GluN2B-mediated activity at low concentrations of glutamate becomes more prominent and results in activation of p38 MAPK, which together with reduced signaling via GluN2A-containing receptors impairs LTP (supplemental Fig. 5, available at www.jneurosci.org as supplemental material). Thus, in addition to the well recognized anti-adhesive property of PSA, being critically important for modulation of cell interactions during development (Rutishauser, 2008), PSA regulates synaptic plasticity in adulthood via modulation of a specific population of glutamate receptors. PSA inhibits GluN2B-containing receptors at low micromolar concentrations of glutamate found in the extracellular space (Sherwin, 1999; Ueda et al., 2000), which are two orders of magnitude below glutamate concentrations in the synaptic cleft after vesicular glutamate release. Thus, it is plausible to expect that although PSA may act on both synaptic and extrasynaptic GluN2B receptors, it would effectively inhibit transmission preferentially via extrasynaptic GluN2B receptors. The influence of PSA on GluN2B at stimulated synapses would be negligible since glutamate is released at high concentrations and then PSA is not effective. As GluN2B-containing NMDA receptors can signal via multiple intracellular pathways (for review, see Ponimaskin et al., 2007), synaptically and extrasynaptically, they may influence LTP positively or negatively depending on the balance between these pathways. For instance, during early development and in mice overexpressing GluN2B (Tang et al., 1999; Köhr et al., 2003) the balance is shifted toward synaptic (rather than extrasynaptic) signaling that promotes LTP. However, manipulations leading to lateral movement of synaptic GluN2B to extrasynaptic sites (reduction of either CaMKII/GluN2B or PSD-95/GluN2B association) result in impairment of LTP (Gardoni et al., 2009). Also in hippocampal slices from adult NCAM/PSA-deficient mice, the balance between synaptic and extrasynaptic pathways seems to shift in favor of GluN2B-mediated extrasynaptic signaling and activation of p38 MAPK, as discussed below.
Although one should be cautious with interpretation of pharmacological data regarding the function of extrasynaptic NMDA receptors (Neyton and Paoletti, 2006), it is noteworthy that several lines of evidence point to an increase in extrasynaptic GluN2B-mediated signaling in NCAM/PSA-deficient slices and its importance for modulation of LTP: (1) PSA suppresses opening of native ifenprodil/Ro 25-6981-sensitive receptors by low concentrations of glutamate (Hammond et al., 2006). (2) PSA suppresses the activity of recombinantly expressed GluN1/2B and GluN1/2A/2B, but not GluN1/2A receptors. (3) In hippocampal slices of NCAM−/− mice, the NVP-insensitive component is larger and is more strongly inhibited by PSA than in slices from NCAM+/+ mice. (4) Ro 25-6981-sensitive component is enlarged in slices from NCAM−/− mice. (5) Removal of PSA increases GluN2B-mediated Ca2+ transients in the aspiny part of dendrites. (6) Inhibition of GluN2B-containing receptors with Ro 25-6981 restores levels of LTP in NCAM- or PSA-deficient slices to the levels seen in slices from wild-type mice. (7) Mice deficient in the glial glutamate transporter GLT-1 have an increased extracellular concentration of glutamate and impaired LTP (Katagiri et al., 2001). (8) Reduction of extrasynaptic glutamate concentrations by the glutamate scavenger GTP restores LTP in NCAM−/− mice. (9) A substantial fraction of GluN2B-containing receptors is located extrasynaptically in the adult hippocampus (for review, see Köhr, 2006). (10) The synaptic pool of NMDA receptors activates the extracellular signal-regulated kinases, whereas the extrasynaptic pool triggers a signaling cascade that results in the inactivation of this pathway (Ivanov et al., 2006).
Interestingly, at 22−24°C the extrasynaptic signaling in response to a single electrical pulse is exaggerated compared with 35−37°C (Asztely et al., 1997). The recording at temperatures of 22−24°C was thus instrumental in our in vitro measurements for a better detection of differences between NCAM genotypes in terms of GluN2B-mediated signal and its inhibition by PSA (Fig. 2). However, in response to short trains of stimuli or theta-burst stimulation, extrasynaptic receptors are activated at both 22−24°C and 35−37°C (Asztely et al., 1997). In line with this data, it has been shown that slices from NCAM−/− mice and endoN-treated slices at 35−37°C (Muller et al., 1996) were similarly impaired in LTP as in our present and several previous studies conducted at room temperature. Furthermore, our behavioral experiments with intrahippocampal inhibition of GluN2B support the view that the revealed mechanism is valid at physiological conditions in vivo.
The following data support the view that the mechanisms downstream of GluN2B-containing receptors in NCAM/PSA-deficient slices involve signaling via Ras-GRF1 to the Rac effector p38 MAPK: (1) The level of phosphorylated p38 MAPK is upregulated in NCAM-deficient mice and in endoNF-treated slices, while it is reduced in Ras-GRF1−/− mice (Li et al., 2006). (2) The level of phosphorylated p44 is co-upregulated although in a smaller degree than p38, as expected in response to activation of Ras-GRF1 (Li et al., 2006). (3) An inhibitor of p38 restores levels of LTP in PSA or NCAM-deficient slices to those seen in wild-type mice. (4) Ras-GRF1 is shown to mediate synaptic depression through p38 MAPK (Li et al., 2006). (5) Recordings of LTP in acute slices from NCAM−/− mice of different ages (G. Kochlamazashvili, M. Schachner, and A. Dityatev, unpublished data) revealed that impairment in LTP appears after developmental switch in NMDA receptors downstream signaling from Sos to Ras-GRFs (Tian et al., 2004). (6) PSA deficiency does not lead to impaired LTP in Ras-GRF1−/− mice.
Activation of p38 MAPK signaling has also been shown to mediate impairment in LTP by tumor necrosis factor (Butler et al., 2004) and the Aβ peptide generated from the amyloid precursor protein, which is widely believed to underlie the pathophysiology of Alzheimer's disease (Hsieh et al., 2006). p38 MAPK regulates the removal of synaptic AMPA and NMDA receptors from the postsynaptic cell surface, although the substrates for p38 that mediate these effects in response to tumor necrosis factor α, Aβ, or activation of NMDA receptors remain to be identified. Among the many molecules signaling downstream of p38 are transcription factors, protein kinases and the small GPTase Rab5, which is essential for early endocytosis (Shi and Gaestel, 2002). An attractive hypothesis is that activation of p38 MAPK may accelerate glutamate receptor endocytosis by stimulating the formation of the guanyl nucleotide dissociation inhibitor-Rab5 complex (Huang et al., 2004).
At the behavioral level, our data on the in vitro induction of LTP match the outcome of fear conditioning in NCAM−/− mice. This paradigm has been used to show impairment in cued (amygdala-dependent) and contextual (hippocampus- and amygdala-dependent) fear memories in NCAM−/− mice (Stork et al., 2000) and in mice deficient for the polysialyltransferases ST8SiaII/STX or ST8SiaIV/PST (Angata et al., 2004; Senkov et al., 2006). The importance of PSA in formation of hippocampus-dependent contextual memory is further underscored by the normalization of memory in NCAM−/− mice via application of PSA-NCAM, but not by NCAM without PSA (Senkov et al., 2006) and by disrupted contextual learning in wild-type mice after acute enzymatic removal of PSA from the dorsal hippocampus (Lopez-Fernandez et al., 2007). Here, we demonstrate that intrahippocampal injection of a GluN2B-specific antagonist normalizes contextual fear conditioning. Thus, PSA carried by NCAM regulates both synaptic plasticity and hippocampus-dependent learning via modulation of NMDA receptors.
NCAM is a strong candidate gene for neurocognitive changes associated with schizophrenia (Sullivan et al., 2007). NCAM−/− mice have enlarged lateral ventricles and deficits in prepulse inhibition of the acoustic startle (Wood et al., 1998), reminiscent of findings in schizophrenic patients. Furthermore, patients with schizophrenia have higher concentrations of the soluble cleaved extracellular domain of NCAM in prefrontal cortex, cingulate cortex, hippocampus, and CSF (Vawter et al., 2001; see their references). In addition to a deficit in PSA expression (Barbeau et al., 1995), a polymorphism in the promoter region of the polysialyltransferase ST8SiaII gene is associated with schizophrenia (Arai et al., 2006; Tao et al., 2007). Recently, cognitive dysfunction has been recognized as a primary and enduring core deficit in schizophrenia. Although new medications may improve other functions, they do not normalize neurocognitive deficits in schizophrenia (Peuskens et al., 2005). Thus, our demonstration that abnormalities in GluN2B-mediated signaling lead to cognitive abnormalities in NCAM/PSA-deficient brains may be an important contribution toward understanding the mechanisms of neurocognitive deficits in schizophrenia and their reversibility. The recent progress in dissection of the molecular mechanisms underlying single-gene mental disorders, such as tuberous sclerosis, neurofibromatosis type I, fragile X and Rett‘s syndromes (Kelleher and Bear, 2008), suggests novel treatments to restore synaptic and cognitive abnormalities in patients suffering from these disorders. Our study makes a step in a similar direction for schizophrenia and may encourage attempts to treat schizophrenia by targeting the GluN2B–p38 MAPK pathway.
Footnotes
-
M.S. was supported by the New Jersey Commission for Spinal Cord Research. M.S. and A.D. were supported by the Deutsche Forschungsgemeinschaft. R.G.-S. and K.S. received financial support from the European Community (FP 6: PROMEMORIA, Contract no. LSHM-CT-2005-512012). We thank Harold Cremer for NCAM-deficient mice, Kodeeswaran Parameshwaran for help with single-channel analysis, Achim Dahlmann for genotyping, and Eva Kronberg for animal care. The vectors for expression of the GluN subunits were kind gifts from Peter Seeburg (Max-Planck Institute for Medical Research, Heidelberg, Germany), and Kevin Holmes and Gregory Dekaban (both from University of Western Ontario, London, Ontario, Canada). NVPAAM077 was a kind gift from Dr. Yves Auberson at Novartis Institutes for BioMedical Research.
- Correspondence should be addressed to either of the following: Alexander Dityatev, Department of Neuroscience and Brain Technologies, The Italian Institute of Technology, Via Morego, 30, 16163 Genova, Italy, alexander.dityatev{at}iit.it; or Melitta Schachner, Zentrum für Molekulare Neurobiologie, Universitätsklinikum Hamburg-Eppendorf, Martinistrasse 85, D-20251 Hamburg, Germany, melitta.schachner{at}zmnh.uni-hamburg.de