Logo of molcellbPermissionsJournals.ASM.orgJournalMCB ArticleJournal InfoAuthorsReviewers
Mol Cell Biol. 2006 Oct; 26(19): 7283–7298.
PMCID: PMC1592877

NF-κB/Rel Regulates Inhibitory and Excitatory Neuronal Function and Synaptic Plasticity

Abstract

Changes in synaptic plasticity required for memory formation are dynamically regulated through opposing excitatory and inhibitory neurotransmissions. To explore the potential contribution of NF-κB/Rel to these processes, we generated transgenic mice conditionally expressing a potent NF-κB/Rel inhibitor termed IκBα superrepressor (IκBα-SR). Using the prion promoter-enhancer, IκBα-SR is robustly expressed in inhibitory GABAergic interneurons and, at lower levels, in excitatory neurons but not in glia. This neuronal pattern of IκBα-SR expression leads to decreased expression of glutamate decarboxylase 65 (GAD65), the enzyme required for synthesis of the major inhibitory neurotransmitter, γ-aminobutyric acid (GABA) in GABAergic interneurons. IκBα-SR expression also results in diminished basal GluR1 levels and impaired synaptic strength (input/output function), both of which are fully restored following activity-based task learning. Consistent with diminished GAD65-derived inhibitory tone and enhanced excitatory firing, IκBα-SR+ mice exhibit increased late-phase long-term potentiation, hyperactivity, seizures, increased exploratory activity, and enhanced spatial learning and memory. IκBα-SR+ neurons also express higher levels of the activity-regulated, cytoskeleton-associated (Arc) protein, consistent with neuronal hyperexcitability. These findings suggest that NF-κB/Rel transcription factors act as pivotal regulators of activity-dependent inhibitory and excitatory neuronal function regulating synaptic plasticity and memory.

Stimulus-coupled changes in synaptic plasticity are required for the storage, retrieval, and removal of acquired information collectively referred to as memory formation (28, 32, 39). Such changes are facilitated by both modifications of existing synaptic effectors and the de novo synthesis of new gene products regulated by various transcriptional regulators. These processes are tightly controlled by the coordinated action of both excitatory and inhibitory neurotransmitters derived from glutamatergic neurons and GABAergic (where GABA is γ-aminobutyric acid) interneurons, respectively (47, 54). While the vast majority of studies to date have focused on the cyclic AMP-responsive transcription factor (CREB) regulating excitatory neuron function (7, 32-34, 62, 72), more recently, other transcription factors, including members of the NF-κB/Rel family of transcription factors, have been implicated in experience-based synaptic adaptations (38, 45, 49, 55). However, our understanding of their precise role in regulating synaptic plasticity remains rudimentary at best.

Although NF-κB/Rel factors were originally implicated as central regulators of the immune and inflammatory responses, both basal expression and stimulus-coupled induction of NF-κB/Rel factors occur in neurons and glial cells (23, 30, 31, 45, 48, 55).Activation of NF-κB/Rel proceeds through the site-specific phosphorylation, polyubiquitylation, and proteasome-mediated degradation of the major NF-κB/Rel inhibitor protein, IκBα (41). The newly liberated NF-κB/Rel complex rapidly translocates into the nucleus, where it engages cognate κB enhancer elements in a variety of cellular target genes including the IκBα gene eliciting an auto-inhibitory feedback loop (65). Substitution of the two key serine phospho-acceptor sites in IκBα with alanines (S32A/S36A) generates a potent, nondegradable inhibitor of NF-κB/Rel activation termed the IκBα superrepressor (IκBα-SR) (6, 61, 74), which serves as a useful tool to probe NF-κB action in vivo (40, 71).

As transcriptional regulators, NF-κB/Rel proteins can potentially either positively or negatively regulate the expression of genes governing changes in synaptic plasticity and cognitive functions (73). Several reports support a positive link between the activation of NF-κB/Rel factors and the induction of long-term potentiation (LTP) or long-term depression, experimental correlates of learning and memory (1, 17, 20, 48). Observations in the crab also support a positive role for NF-κB/Rel action in emotional learning and fear responses (16, 50). Similarly, mice lacking the p50 subunit of NF-κB/Rel display impaired emotional learning and decreased anxiety-related responses but exhibit increased exploratory activity (37, 38). In contrast, other studies suggest a negative correlation between NF-κB action and synaptic function. For example, NF-κB/Rel activation has also been shown to impair the generation of synaptic currents in hippocampal neurons (20). Increased NF-κB/Rel action is also associated with the accelerated onset of cognitive deficits in an experimental model of Alzheimer's disease (2).

These apparently conflicting observations could reflect distinct roles for the various NF-κB/Rel factors in regulating different cognitive behaviors in select brain regions. Alternatively, these different outcomes may involve specific NF-κB/Rel actions in distinct neuronal subtypes and/or in glia. In support of the latter, Meffert et al. have reported that loss of the RelA/p65 subunit of NF-κB, in the context of concomitant tumor necrosis factor receptor 1 (TNFR1) deficiency, in both neuronal and glial cells results in impaired spatial learning and memory (49). Additionally, Fridmacher et al. and Kaltschmidt et al. (19, 29), using the CamKII promoter to direct expression of the NF-κB/Rel inhibitor exclusively in excitatory neurons of the forebrain, have demonstrated a clear positive requirement for NF-κB/Rel action in regulating synaptic plasticity and memory. To date the potential function of NF-κB/Rel in inhibitory GABAergic interneurons remains unexplored (47).

In our present study, we have utilized the prion promoter and enhancer (70) to direct the expression in neurons of a potent, nondegradable inhibitor of NF-κB/Rel activation, termed the IκBα superrepressor. IκBα-SR+ mice display robust IκBα-SR expression in inhibitory interneurons, somewhat lower levels of expression in excitatory neurons, and no detectable expression in glia. These IκBα-SR animals were used to explore the role of NF-κB/Rel in various electrophysiological and biochemical parameters in the brain. Our findings reveal that pan-neuronal inhibition of NF-κB action results in a marked enhancement of activity-dependent synaptic signaling and select cognitive functions including learning and memory.

MATERIALS AND METHODS

Generation and characterization of IκBα-SR bigenic mice.

A total of eight transgenic pTet-O-HA-IκBα-SR founder lines were generated by pronuclear injection of linearized DNA into DBA inbred zygotes, and the resulting mice were screened by PCR to detect the presence of the transgene. Positive transgenic mice were crossed with FVBN mice carrying a Prp/TtA transgene (Tet-off). Bigenic mice were screened by PCR for both transgenes. At 1 month of age, FVBN/DBA bigenic litters were divided into two groups and maintained on either a standard or a doxycycline-supplemented diet. At 4 to 6 months of age, mice were anesthetized with isoflurane and sacrificed by cervical dislocation. Brains from control and IκBα-SR+ mice were removed, fixed in 4% paraformaldehyde, embedded in paraffin, sectioned, and stained with Nissl stain for histological examination. Strict adherence to institutional and NIH guidelines was maintained in all procedures relating to the care and treatment of mice. As discussed in the results section, three independently derived transgenic lines, demonstrating readily detectable IκBα-SR transgene in the absence of doxycycline but near complete inhibition of IκBα-SR expression in the presence of doxycycline, were selected for further study. Use of three independently derived founder lines minimized effects related to the site of transgene integration.

Immunoprecipitation, immunoblotting, and immunofluorescence.

Mice from each group were anesthetized with isoflurane, perfused with saline followed by 4% paraformaldehyde, at 4 to 8 months of age. Brains were removed, and lysates from whole tissue or specific regions were prepared with a low-stringency buffer containing 50 mM Tris-HCl, pH 8, 120 mM NaCl, 5 mM EDTA, 0.5% (wt/vol) NP-40, supplemented with fresh 1× protease inhibitor cocktail (Calbiochem). Membrane-enriched fractions were isolated as described in Pharmingen protocols online. Lysates were either resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis directly or were first immunoprecipitated with hemagglutinin (HA)-specific monoclonal antibodies (262K at a dilution of 1:200; Cell Signaling) conjugated to protein G agarose (25 μl of packed beads). Membranes were immunoblotted with HA-specific polyclonal antibodies (Y11 at a dilution of 1:1,00; Santa Cruz Biotechnology) and the antibodies indicated in the legends overnight at 4°C and visualized by chemiluminescence. For immunohistochemical and immunofluorescence analysis, brains were removed from selected mice following anesthetization with isoflurane and perfusion with saline and 4% paraformaldehyde and embedded in cryo-matrix mounting medium (22-oxyacalcitriol [OCT]; Tissue-Tek), frozen, and cryosectioned into 10-μm sections. Additionally, primary neuronal cultures (hippocampal or cortical) were grown on treated slides and probed. Frozen sections or slides were fixed in 4% paraformaldehyde and permeabilized in 0.2% Triton X-100. Sections were either stained with Nissl or incubated with primary antibodies (indicated in the legend) overnight at 4°C and probed with specific antibodies as outlined in the legends, followed with fluorescein isothiocyanate-conjugated or Alexa-conjugated secondary antibodies (1:1,000) for 60 min at room temperature. Sections were counter-stained with DAPI (4′,6′-diamidino-2-phenylindole; 1:500) for 5 min for nuclear staining. Sections were mounted in Gel/Mount (Biomeda) and visualized under UV light on a Nikon E600 microscope connected to a SPOT advanced software camera. Several fields were compared for intensity of positive immunostaining by a technician blinded to the genotypes.

Evaluation of NF-κB/Rel signaling by electrophoretic mobility shift assay (EMSA) and RPA.

Bigenic IκBα-SR+ and IκBα-SR mice were treated with kainic acid (KA) (22 mg/kg) by intraperitoneal injection at 4 to 6 months of age (44). Mice were anesthetized, and hearts were perfused with saline at 7 to 8 h after KA injection. Brains were removed and lysed in nuclear extraction buffer (20 mM Tris-HCl, pH 7.8, 125 mM NaCl, 5 mM MgCl2, 0.2 mM EDTA, 12% [wt/vol] glycerol, and 0.1% [wt/vol] NP-40) supplemented with 10 μg/ml aprotinin, 1 mM phenylmethylsulfonyl difluoride, and 1 mM dithiothreitol.Samples with matched protein concentrations were incubated with a κB/Rel enhancer probe radiolabeled with [32γ-P]ATP. Nucleoprotein-κB/Rel complexes were separated on nondenaturing gels and visualized by autoradiography. Control and bigenic mice were treated with KA as described above. After 8 h, brains were removed, and total RNA was isolated with an RNA isolation kit (Access RT PCR system; Promega) according to the manufacturer's directions. An RNase protection assay (RPA) was performed using the Riboquant system (Pharmingen), and NF-κB-targeted RNA sequences were detected with a specific probe template set (mAPO-3).

Establishing primary neuronal cultures.

Embryonic day 15 embryos were removed from gestating mice and placed in ice-cold 1× Hanks balanced salt solution. Whole brains from these embryos were rapidly removed, and the hippocampal formation and cortex were isolated and washed in ice-cold 1× Hanks balanced salt solution. The tissue was dissociated with papain using a Worthington papain dissociation system, and isolated cells were plated at a concentration of 4 × 104 cells/ml on poly-l-lysine-coated six-well plates in neurobasal medium supplemented with B-27. Cortical cultures enriched in GABAergic neurons (>95%) were grown in the presence of valine (25 μg/ml), pyruvate (2 mM), and α-ketoglutarate (5 mM) and in the absence of glutamine. Hippocampal cultures enriched in glutamatergic neurons (>87%) were maintained in medium supplemented with 2 mM glutamine as described previously (76). Cultures were incubated at 37°C in a humidified atmosphere containing 5% CO2. On day 3 the medium was changed and supplemented with conditioned culture medium and mitotic inhibitor solution (5-fluoro-2-deoxyuridine, cytosine-d-arabinofuranoside, and uridine). Fresh medium was added every 24 h. On day 6, cultures were incubated with fresh medium without glutamic acid and maintained for up to 2 weeks. In experiments using enriched glial cultures, cultures were shifted into Dulbecco's modified Eagle's medium/F12 l-valine lacking KCl but supplemented with 10% fetal calf serum, G5 (Gibco), and penicillin/streptomycin to induce neuronal cell death.

Establishing postmitotic (adult) brain neurons and glial cells from the hippocampus.

Hippocampal regions were rapidly dissected from the brains of postnatal day 1 (PN1) mice in 2 ml of Hibernate-A, supplemented with B27 and 0.5 mM l-glutamine. Meninges and excess white matter were removed. The hippocampus was sliced perpendicularly to the long axis and transferred to a tube containing fresh Hibernate-A/B27 and 0.5 mM l-glutamine and rocked at 30°C for 30 min. The slices were transferred to fresh Hibernate-A with papain (20 U/mg) and rocked at 30°C for 45 min. Slices were transferred to fresh Hibernate-A/B27 incubated at room temperature for 5 min. Slices were pipetted 15 times, and the pieces were allowed to settle for 2 min. The supernatant (cell suspension) was removed and carefully applied to the top of a gradient (35%, 25%, 20%, and 15%) formed with OptiPrep in Hibernate-A/B27 medium and centrifuged at 800 × g for 15 min. The top 6 ml (debris) was discarded. The underlying 2 ml, designated fraction 1, was enriched for oligodendrocytes. The next layer, fraction 2, contained neurons with accessory glia cells. Fraction 3 was enriched for neurons, and the pellet, fraction 4, was enriched for microglia. Each fraction was pelleted at 200 × g for 1 min and then resuspended in Neurobasal-A/B27 with l-glutamine and 1× gentamicin. Neurons were plated at 90 to 320 cells/mm2 on poly-d-lysine-coated plates in fresh Neurobasal-A/B27 medium supplemented with 0.5 mM l-glutamine, 10 μg/ml gentamicin, and 5 ng/ml fibroblast growth factor-β.Glial cells from each representative fraction (oligodendrocytes, microglia, and mixed glial fraction), were resuspended at a concentration of 3 × 106 cells in 12 ml of MEM10 containing 1% penicillin/streptomycin, and 2 ml of each suspension was added to each well of a six-well plate precoated with poly-d-lysine. All cultures were maintained in a highly humidified atmosphere containing 5% CO2 at 37°C.

Reverse transcription-PCR analysis of GAD65 mRNA levels.

Whole hippocampus from either untrained or trained (subjected to maze testing) mice in each genotypic group was collected for RNA quantification. RNA was extracted using an RNAeasy kit from QIAGEN. The purified RNA was DNase treated (Ambion, Austin, TX) and reverse transcribed from mRNA to cDNA using a first-strand synthesis kit (Invitrogen, Carlsbad, CA). The amount of cDNA was quantified using real-time PCR and primers sets designed to amplify GAD65. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and β-actin mRNA levels were used as internal controls for normalization. Amplified mRNA transcripts were visualized on 1.5% agarose gels.

Characterization of IκBα-SR bigenic mice.

At 1 month of age, bigenic littermates were screened by PCR to confirm the presence of the transgene, divided into two groups, and maintained on a normal or doxycycline-supplemented diet. At 4 to 8 months, male mice from three independently derived founder lines were selected for further analysis. Progeny from each founder line behaved similarly in each of the various assays.

Electrophysiological analysis. (i) Slice preparation.

The animals were first anesthetized with halothane, and brains were rapidly removed and cooled in ice-cold artificial cerebrospinal fluid (ACSF) containing the following: 125 mM NaCl, 5 mM KCl, 1.25 mM NaH2PO4, 25 mM NaHCO3, 25 mM d-glucose, 2 mM CaCl2, 1 mM MgSO4, 0.01 mM glycine, and 1 mM kynurenic acid bubbled with a mixture gas of 95% O2 and 5% CO2. For slicing, 1 mM kynurenic acid was also added to ACSF. The solution pH was adjusted to 7.3 to 7.4, and osmolarity was set at 300 ± 10. The hippocampus was dissected, and 350-μm slices were cut transversely. A cut was routinely made between the CA1 and CA3 area right after slicing, and in some recordings, both CaCl2 and MgSO4 concentration was increased to 4 mM to reduce seizure-like activity and population spikes. All slices were collected in a holding chamber in the same solution maintained at 31 ± 1°C. A single slice was transferred to an interface recording chamber that was constantly perfused with the gas-saturated ACSF at 1 ml/min at 29 ± 1°C. The recording chamber was also constantly superfused with gas-saturated moist air during recording. The remaining slices were kept in the holding chamber until tested.

(ii) Extracellular field recordings.

Field recordings were used to assess synaptic transmission and plasticity in the Schaffer collateral/commissural CA1 pathway. For field excitatory postsynaptic potential (fEPSP) recordings, a glass pipette was filled with the ACSF or 1 mM NaCl (3 to 5 M resistance) and placed within the striatum radiatum. For high-frequency stimulation, two bipolar tungsten electrodes (S1 and S2) were placed on opposite sides of the recording electrode along the Schaffer collateral fibers in the striatum radiatum. The test stimulation was delivered alternately through S1 and S2 once per minute. After 30 to 60 min of recording for control tetanus stimulation, three or four trains of square pulses (100 pulses at 100 Hz) were delivered with 3- to 4-min intervals through S1, while the S2 was turned off. In some cases, theta burst stimulation was also used. These two stimulation methods were compared in the previous whole-cell recordings, and no clear difference was noted. For late-phase long-lasting LTP (L-LTP), field potentials evoked by S1 and S2 were monitored in the same way as in the control for a minimum of 180 min and, in most cases, up to 300 min after the LTP induction protocol. The peak amplitude (or 30 to 70% rising slope) of all fEPSPs recorded from an individual slice were normalized to the mean peak amplitude (or slope) during the 30 min before the theta burst or high-frequency stimulation, and these normalized values were used to compared LTPs induced in various treatment groups. Field potentials were recorded with an Axon amplifier 2B (Axon Instruments, Union City, CA). Data acquisition and analysis were done off-line using P-Clamp 9 software and Origin. All data are presented as the mean ± standard error of the mean. Significance was assessed at a P value of <0.05, using Dunnett's t test. Any recordings lasting less than 3 h were excluded from the final analysis. An average of three slices/mouse and three mice/group were used for the analysis of L-LTP. A researcher blinded to the genotypes of the mice performed all recordings.

Behavioral analysis.

Control and bigenic mice were assessed at 6 to 8 months of age in a blind-controlled series of behavioral tests, including (i) a Morris water maze test and (ii) a radial arm maze.

(i) Morris water maze test.

A pool (diameter 140 cm) was filled with opaque water (24 ± 1°C), and mice were trained to locate first a visible platform (days 1 to 2) and then a submerged hidden platform (days 3 to 5) in two daily sessions 3.5 h apart, each consisting of three 60-s trials (10-min intertrial intervals). Mice acquired spatially encoded information with visual cues outside the maze to locate the platform. Mice that failed to find the hidden platform within 60 s were placed on the platform for 15 s. For analysis of data, the pool was divided into four quadrants. During the visible platform training, the platform was moved to a different quadrant for each session. During the hidden platform training, the platform location was kept constant for each mouse (in the center of the target quadrant). The starting point at which the mouse was placed into the water was changed for each trial. Time to reach the platform (latency), path length, and swim speed were recorded with a Noldus Instruments EthoVision video tracking system set to analyze two samples per second. Since there were no significant differences in average swim speeds between the different groups of mice during the visible platform sessions, the time required to locate the platform (latency) was used as the main measure for analysis. A 60-s probe trial (platform removed) was carried out as described in the legend.

(ii) Radial arm maze.

One week before testing, animals were placed on a restricted diet of 80% to 90% of food levels so that their initial body weight decreased by a maximum of 15%. This diet was maintained throughout the testing period; however, mice were given free access to water. Mice received food reward pellets in their home cages 1 day before pretraining to become accustomed to the novel food in a familiar environment. For pretraining, each mouse was placed in the central platform of the eight-arm radial maze and allowed to explore and consume food pellets scattered throughout the entire maze for a 15-min period. During training, mice were allowed to take a pellet from each food dispenser located at the distal end of each arm. A trial was finished after the subject mouse consumed the pellet in each of the eight arms. As a mouse consumed a pellet from each arm, the next arm opened, and when the mouse entered this arm, the first arm closed. Thus, the mouse was sequentially guided through each arm of the maze. Immediately after the training, maze acquisition trials were performed with all eight arms baited with food pellets. Mice were placed on the central platform and allowed access to all eight arms for 15 min. The session was terminated immediately after all eight food pellets concealed at the end of the arms were consumed as measured by breaking a sensor beam (“head poke”) or after 15 min. An “arm visit” was defined as entering the arm and breaking both sensor light beams. Mice were confined in the center platform for 2 s after each arm choice, thus reducing behaviors such as clockwise serial searching strategies. Animals were subjected to one session per day. A MedPC software program was used for both the training and testing phases. For each trial, the following were automatically recorded: (a) latency to complete the maze and retrieve all pellets, (b) number of errors, (c) choice of arms, and (d) number of different arms chosen within the first eight choices. The operator also manually recorded the number of pellets eaten by each mouse. Each of these parameters was used to detect abnormalities in the acquisition and retention of spatially encoded information. In each case, the operator was blinded to the genotype of the mice being examined. Data are expressed as mean ± standard error of the mean. Differences among means were evaluated by analysis of variance (ANOVA) and a Tukey-Kramer test. Learning curves were compared by a repeated-measures ANOVA using contrasts to assess differences between specific groups of mice. For all analyses, the null hypothesis was rejected at the 0.05 level.

RESULTS

Generation and characterization of transgenic mice expressing an IκBα-SR inhibitor of NF-κB action in neurons.

To explore the relationship between NF-κB/Rel factor activation, synaptic signaling, and higher-order cognitive function, we generated multiple independent lines of transgenic mice expressing an HA-tagged version of the NF-κB/Rel inhibitor, IκBα-SR. The IκBα-SR transgene was cloned downstream of a tetracycline transactivator (tTA)-responsive promoter (Tet-O) facilitating regulatable expression. By breeding these mice to a second transgenic line of mice where the prion promoter (Prp) was used to direct tTA expression (a generous gift from Stanley Prusiner, University of California, San Francisco), expression of the Tet-O-IκBα-SR transgene was restricted to neurons. In the presence of the blood-brain barrier-permeable tetracycline analogue, doxycycline, transcription of the tTA-regulated gene is blocked (Fig. (Fig.1A)1A) (22). In the presence or absence of doxycycline, bigenic Prp-tTA/Tet-O-HA-IκBα-SR mice were both viable and fertile. Furthermore, these mice displayed anatomically and structurally normal gross brain morphology as indicated by Nissl and anti-calbindin immunofluorescence staining of the hippocampal regions (Fig. (Fig.1B1B).

FIG. 1.
Generation and biochemical characterization of bigenic IκBα-SR mice. (A) Transgenic mice encoding an N-terminal, HA-tagged IκBα-(SS32/36AA) superrepressor under the control of tetracycline response element (Tet-O) were ...

In each of the eight independently derived lines, expression of the IκBα-SR transgene was clearly detectable (designated IκBα-SR+ or SR+), and its expression was suppressed in the presence of doxycycline (designated IκBα-SR or SR) (Fig. (Fig.1C).1C). Three independently derived lines of FVBN/DBA bigenic mice exhibiting the tightest regulation of transgene expression by doxycycline and control mice were selected for further analysis (Fig. (Fig.1C,1C, designated by asterisks) to ensure that the observed effects were attributable to transgene expression rather than to a positional effect related to the site of transgene insertion. Among several tissues examined, expression of the IκBα-SR transgene appeared confined to brain, and within the brain, transgene expression was detected in all regions examined, including cerebellum, cortex, thalamus, and mid-brain (Fig. 1D and E). The levels of HA-tagged transgenic IκBα-SR protein in hippocampal neurons isolated from IκBα-SR+ mice proved comparable to the levels endogenous IκBα protein observed in either IκBα-SR or control mice (Fig. (Fig.1F).1F). Of note, the level of endogenous IκBα protein was lower in IκBα-SR+ neurons due to the loss of NF-κB/Rel action required for the auto-regulatory induction of the IκBα gene (65). These equivalent levels of transgene and endogenous protein mitigate against an artifactual gain-of-function phenotype sometimes observed when transgenes are overexpressed in vivo.

Immunofluorescence staining of IκBα-SR+ brain sections confirmed IκBα-SR transgenic protein expression in neurons from the hippocampus, cortex, medulla, hypothalamus, and Purkinje cells of the cerebellum (data not shown). The HA-tagged IκBα-SR protein was not detected in matched neuro-anatomical regions from control or IκBα-SR mice. In vivo expression of the HA-IκBα-SR gene product was not detected in macroglia when sections were probed with both anti-HA and anti-glial fibrillary acidic protein (GFAP) antibodies, but expression was detectable in MAP2-positive neurons (Fig. (Fig.1G1G).

To confirm functional inhibition of NF-κB in vivo following IκBα-SR expression, we assessed both DNA binding and induction of NF-κB target genes. Using EMSAs, we compared KA-induced NF-κB/Rel DNA binding activity in IκBα-SR+ mice and IκBα-SR mice. While robust NF-κB/Rel activation was detected in various brain regions from IκBα-SR mice, this response was markedly impaired in the IκBα-SR+ mice (Fig. (Fig.2A2A ). These KA-induced nucleoprotein complexes contained p50, p52, c-Rel, and RelA protein species as determined by supershifting with specific antibodies (data not shown). Primary neuronal or glial cultures from embryonic brain tissues from IκBα-SR+, IκBα-SR, and control mice were either untreated (−) or exposed to TNF-α or high concentrations of KCl (22 mM) to induce membrane depolarization. In IκBα-SR+ neurons, the activation of NF-κB/Rel in response to either stimuli was markedly inhibited under both basal and stimulated conditions. In contrast, NF-κB/Rel DNA binding was evident in both unstimulated and, to a greater extent, in stimulated in neurons from both IκBα-SR and control mice (Fig. (Fig.2B).2B). Consistent with expression of IκBα-SR in neurons, but not glia, NF-κB/Rel activation was detected in glial cultures from IκBα-SR+ mice as well as from IκBα-SR and control mice (Fig. (Fig.2C2C).

FIG. 2.
Functional inhibition of NF-κB/Rel DNA binding by IκBα-SR expression. (A) NF-κB/Rel activation induced by in vivo administration of KA in various brain regions from IκBα-SR+ and IκBα-SR ...

RPAs were used to detect KA induction of various endogenous NF-κB/Rel-regulated target genes, specifically Fas and the p55 TNF-α receptor (Fig. (Fig.2D).2D). RPAs revealed that IκBα-SR expression in neurons effectively inhibited KA induction of the Fas and p55 TNF-α receptor genes.In contrast, both genes were effectively induced by KA in both IκBα-SR and control mice (Fig. (Fig.2E).2E). Together these findings demonstrate that expression of the IκBα-SR transgene is restricted to neurons, is tightly regulated by doxycycline, can functionally inhibit KA-induced nuclear translocation of NF-κB/Rel factors in neurons but not in glial cells, and is able to effectively impair the activation of various endogenous NF-κB/Rel-inducible target genes.

IκBα-SR is expressed in both cortical inhibitory GABAergic interneurons and hippocampal excitatory glutamatergic neurons.

To assess potential differences in IκBα-SR expression in functionally distinct neuronal subtypes, we probed HA-specific immunoprecipitations of lysates prepared from primary hippocampal excitatory neurons or from cortical inhibitory GABAergic neuronal cultures. HA-IκBα-SR transgene expression was detected only in IκBα-SR+ lysates and, taking into account β-tubulin protein input levels, expression levels of IκBα-SR approximately seven times higher were detected in cortical cultures enriched for GABAergic interneurons relative to levels detected in hippocampal cultures enriched for excitatory glutamatergic neurons (Fig. (Fig.3A).3A). The relative levels of inhibitory GABAergic neurons and excitatory glutamatergic neurons present in cortical cultures and hippocampal cultures were quantitated by immunofluorescence staining using antibodies specific for vesicular GABA transporter VGAT (GABAergic) versus vesicular glutamate transporter VGLUT (glutamatergic), respectively. Cortical neuronal cultures were enriched for GABAergic interneurons (98% VGAT+ cells), whereas hippocampal cultures primarily contained glutamatergic cells (83% VGLUT+ cells) (Fig. (Fig.3B).3B). Enhanced prion-promoter driven HA-IκBα-SR expression in GABAergic interneurons is consistent with the activity of the prion promoter in vivo, where either the endogenous prion protein or a Prp-enhanced green fluorescent protein transgene was robustly expressed in these neuronal cells with little or no protein detected in glia (5, 51).

FIG. 3.
IκBα-SR expression in both excitatory glutamatergic neurons and inhibitory GABAergic interneurons. (A) Neuronal cultures were isolated from embryonic day 17 brains from IκBα-SR+ and IκBα-SR ...

Inhibition of NF-κB by IκBα-SR results in deceased GAD65 expression in inhibitory GABAergic interneurons.

Probing hippocampal lysates from IκBα-SR+ mice revealed markedly decreased levels of glutamate decarboxylase (GAD65), the rate-limiting enzyme required for the synthesis of the inhibitory neurotransmitter, GABA (Fig. (Fig.3C).3C). However, no difference in the levels of the vesicular GABA transporter protein (VGAT) was observed in lysates from mice in all three groups, arguing against a selective loss of these inhibitory GABAergic interneurons in IκBα-SR+ mice (Fig. (Fig.3C).3C). Similarly, primary neuronal cultures isolated from IκBα-SR+ mice expressed significantly lower levels of GAD65 (Fig. (Fig.3D3D).

Consistent with the immunoblotting results, GAD65 mRNA levels were decreased relative to levels seen in either IκBα-SR or control mice. Both end-point PCR and real-time PCR (Fig. (Fig.3E)3E) analysis confirmed lower GAD65 mRNA transcripts in GABAergic interneurons in IκBα-SR+ mice, suggesting that the GAD65 gene is regulated either directly or indirectly by NF-κB/Rel. Together, these findings indicate that NF-κB/Rel is required for expression of GAD65 in GABAergic interneurons and suggest that NF-κB/Rel factors may regulate inhibitory neuronal function.

Inhibition of NF-κB/Rel action by IκBα-SR expression in neurons leads to increased LTP in the hippocampus.

We next investigated the potential impact of IκBα-SR expression in both inhibitory and excitatory neurons on the induction of LTP following tetanus or high-frequency stimulation. The averages of peak amplitudes of fEPSPs (S1) from all trials collected 15 to 30 min before the induction of LTP (100%) were used to normalize peak amplitude of fEPSPs of each individual trial acquired over 300 min after tetanus stimulation (Fig. (Fig.4A-D).4A-D). Similar recordings were obtained using theta burst high-frequency stimulation (data not shown). While tetanus stimulation of the Schaffer-collateral pathway induced higher fEPSP amplitude (LTP) in all three groups, we observed significantly greater potentiation in hippocampal slices from several IκBα-SR+ mice (Fig. (Fig.4D).4D). In all cases, the enhanced S1 fEPSP lasted for a minimum of 180 min, and in most cases, responses were recorded for 300 min, consistent with enhanced L-LTP. To monitor basal synaptic signaling, a second stimulating electrode was placed at the other side of the recording electrode to evoke control fEPSPs (S2) between S1 trials. The S2 fEPSPs evoked by this unstimulated pathway did not show any significant change after the tetanus stimulation of the S1 pathway and remained steady for the duration of the recording. The LTP induced in hippocampal slices from IκBα-SR+ mice was consistently higher than that observed in either IκBα-SR or control mice, (P < 0.05, Dunnett's t test) (Fig. (Fig.4D).4D). The observed enhancement of synaptic signaling did not appear to reflect an overall increase in synapse numbers in IκBα-SR+ mice. Similar levels of synapsin, synaptophysin, and complexin as well as myelin basic protein were observed in sections or hippocampal membrane fractions isolated from either IκBα-SR+, IκBα-SR or control mice. In addition, overall neuronal and glial cell content was equivalent in mice from all three groups as evidenced by comparable levels of the neuronal cell marker MAP2 or the glial cell marker GFAP (data not shown).

FIG. 4.
IκBα-SR+ transgenic mice exhibit enhanced LTP and increased Arc expression as indicators of neuronal hyperexcitability. Representative recordings of tetanus stimulation induced LTP along the Schaffer collateral/commissural CA1 ...

The induction of an immediate early gene, arc (for activity-regulated, cytoskeleton-associated), was used as a marker of increased excitability of glutamatergic neurons potentially as a result diminished GABA-mediated inhibitory inputs (64). As expected, Arc protein was not expressed under basal conditions in hippocampal cultures isolated from either IκBα-SR+, IκBα-SR, or control mice (Fig. (Fig.4E).4E). However, using low Mg2+/high K+ supplemented medium to facilitate neuronal excitation (3), robust Arc expression was detected in cultures from IκBα-SR+ mice but not from IκBα-SR or control mice (Fig. (Fig.4F).4F). Arc induction was blocked in IκBα-SR+ cultures following pretreatment with the NMDA (N-methyl-D-aspartate) antagonist, AP5, but not with the AMPAR ((alpha-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid receptor) antagonist, CNQX (6-cyano-2,3-dihydroxy-7-nitro-quinoxaline) (Fig. (Fig.4F).4F). Of note, the induction of Arc by CNQX in all three mice lines is consistent with a recent report in which inhibiting AMPAR activity strongly potentiates activity-dependent Arc expression at the level of transcription by blocking AMPAR-coupled Gi signaling (60). Taken together, these findings support the notion that IκBα-SR expression in neurons may disturb the homeostatic balance between inhibitory interneurons and excitatory neuronal functions, leading to hyperexcitability, increased LTP, and induction of activity-dependent target genes like arc.

Selective inhibition of NF-κB/Rel action in neurons alters cognitive behaviors.

To explore in vivo the impact of IκBα-SR expression in neurons on cognitive function, four groups of FVBN/DBA genetically matched male mice at 6 to 8 months of age were evaluated in a range of behavioral tests. Each cohort included mice from each of the following groups: (i) nontransgenic or singly transgenic mice (control; n = 57), (ii) bigenic mice from three independently derived lines maintained on doxycycline from 1 month of age (IκBα-SR; n = 63), (iii) bigenic mice off doxycycline and expressing the IκBα superrepressor (IκBα-SR+; n = 91), and (iv) bigenic mice on doxycycline in utero to suppress transgene expression during development but subsequently switched to a nondoxycycline diet at 1 month of age to allow transgene expression (IκBα-SR+*; n = 29). Of note, all IκBα-SR+ mice appeared to be significantly more active in their home cages and were much more sensitive to handling-induced seizures.

Using the elevated plus maze to assess anxiety levels, no differences were detected in mice from any of the groups. This normal or unimpaired anxiety phenotype was confirmed using the zero maze, with mice in all groups spending equivalent amounts of time in the maze (Table (Table1).1). Interestingly, in the plus maze, both IκBα-SR+ and IκBα-SR+* mice consistently exhibited increased exploratory activity as indicated by the total distance moved in both the open arms (P < 0.01, Tukey-Kramer test) and in the closed arms of the plus maze (P < 0.01, Tukey-Kramer test) (Table (Table1).1). IκBα-SR+ and IκBα-SR+* mice also exhibited heightened exploratory activity with more rearing events relative to their doxycycline-treated littermates (IκBα-SR) or control mice, (P < 0.01) in the open-field test (52). This increased exploratory activity could reflect diminished inhibitory tone leading to increased excitatory neuronal activity in the IκBα-SR+ mice. However, direct effects related to IκBα-SR expression in excitatory neurons cannot be excluded.

TABLE 1.
Activity and anxiety phenotypes as assessed by plus and zero mazesa

In evaluating hippocampus- and cortex-dependent nonspatial learning and memory by novel object recognition testing (59), we observed that mice in all three groups displayed a normal response. All mice explored both objects equivalently on day 1 and spent more time exploring the novel object versus the familiar object on day 2 of testing (P < 0.01) (data not shown). Of note, IκBα-SR+ mice spent significantly more time overall exploring both objects on each day, consistent with an enhanced exploratory behavior. Examination of sensorimotor function by rotorod testing also revealed no significant differences between these various transgenic animals (data not shown).

IκBα-SR+ mice exhibit enhanced performance in spatial memory tests.

In view of a recent report describing impaired spatial learning in mice lacking p65/RelA expression and NF-κB action in all neurons and in glial cells (49), we evaluated spatial memory in the IκBα-SR+, IκBα-SR, and control mice. Using the Morris water maze, we observed that the IκBα-SR+ mice located both the visible and hidden platforms in significantly less time than their control counterparts (P < 0.01) (Fig. (Fig.5A).5A). Swim speeds for mice in all groups were similar (control, 17.9 ± 0.7 cm/s; IκBα-SR+, 14.8 ± 0.3 cm/s; IκBα-SR, 16.7 ±0.6 cm/s; IκBα-SR+*, 16.5 ± 0.5 cm/s). These data indicate an overall enhancement of spatial learning and memory in IκBα-SR+ mice. There was a significant genotype by session interaction in both phases of the test (P = 0.0433, repeated-measures ANOVA). Interestingly, IκBα-SR+* mice expressing the IκBα-SR transgene after 1 month of age exhibited even stronger acceleration of spatial learning and memory. Bigenic IκBα-SR+ mice learned much faster than IκBα-SR or control mice (Fig. (Fig.5A,5A, inset), as shown by the change in overall performance between the first and last sessions of each test phase and by the changes in latency (per mouse) between two subsequent sessions of the test (P < 0.05, Tukey-Kramer test). Conversely, bigenic IκBα-SR mice on doxycycline, which silenced transgene expression, displayed levels of spatial learning and memory similar to control mice.

FIG. 5.
IκBα-SR+ bigenic mice exhibit enhanced spatial learning and memory. (A) IκBα-SR+ mice located a visible platform significantly faster than control or IκBα-SR mice (P < 0.01). ...

The retrieval of spatially encoded information (i.e., memory) was assessed by a probe trial (57), in which the platform was removed from the water maze and the mice were examined for their ability to remember the specific target quadrant where the platform was previously located. After completion of the test, both IκBα-SR+ and IκBα-SR mice spent significantly more time in the target quadrant than in any other quadrant, indicating that all mice had learned and recalled the platform location by the end of testing (P < 0.01;, Tukey-Kramer) (data not shown). However, when probe trials were performed at earlier times during testing with a second cohort of IκBα-SR and IκBα-SR+ mice (n = 5 for each group), we observed that IκBα-SR+ mice spent significantly more time in the target quadrant than in any other quadrant even after only one day of testing (for SR+, P < 0.05; target versus any other quadrant) (Fig. (Fig.5B).5B). In contrast, the IκBα-SR mice failed to show any preference for the target quadrant on day 1. On day 2, the IκBα-SR+ mice showed an even stronger preference for the target quadrant (for SR+, P = 0.0051; P < 0.05 target versus right quadrant and P < 0.01 target versus left and opposite quadrants). While the IκBα-SR− mice exhibited a trend toward favoring the target quadrant on day 2, this result did not reach statistical significance (P = 0.198). Even on day 3, the IκBα-SR+ mice continued to show a preference for the target quadrant, (SR+, P < 0.05; target versus any other quadrant). In contrast to the accelerated retrieval of spatial memory observed in IκBα-SR+ mice, IκBα-SR mice display a preference for the target quadrant only on day 3 of testing (SR, P = 0.0025; P < 0.05 target versus right quadrant and P < 0.01 target versus left and opposite quadrants). These findings are consistent with the overall enhanced performance of IκBα-SR+ mice in the Morris water maze.

Since the Morris water maze test may involve a component of stress relayed to exposure to water, we tested a second cohort of mice using the radial arm maze, paralleling the study described by Meffert et al. (49). Using time to complete the maze (latency) as a measure of spatial learning and memory, we observed that IκBα-SR+ mice again displayed improved spatial learning and memory relative to either IκBα-SR or control mice (P < 0.01) (Fig. (Fig.5C).5C). After two pretraining sessions, faster times were recorded for IκBα-SR+ mice on day 1 of the maze test and the IκBα-SR+ mice continued to exhibit enhanced performance throughout the remainder of the test. Additionally, using the “number of errors” in the radial arm maze paradigm as reported by Meffert et al. for p65−/− TNFR−/− mice, the IκBα-SR+ mice also performed better than the IκBα-SR or control mice (Fig. (Fig.5C,5C, inset).

To determine the relationship between maze performance and synaptic signaling, we assessed potential correlations between the magnitude of LTP recorded (percent over baseline) and latency in radial arm maze-trained mice. Striking correlations between the increase in LTP and either the time recorded (r = −0.95; P = 0.0134; n = 5) or in errors made in the last trial (r = −0.98; P = 0.0048; n = 5) were observed. Correcting the maze data for potential differences in performance in trial 1 [percent improvement measure calculated as the (performance in the first trial − performance in the last trial)/performance in the first trial], the magnitude of the LTP recorded (percentage over baseline) was strongly correlated with percent improvement in time to complete the maze (r = 0.88; P = 0.0239; n = 5) and in errors made (r = 0.81; P = 0.049; n = 5). These findings reveal a strong correlation between improved LTP and enhanced performance in the radial arm maze. Thus, based on these two independent tests, we suggest that the neuronal pattern of expression of IκBα-SR in these mice results in enhanced spatial learning and memory.

Activity-dependent recovery of synaptic signaling and gluR1 levels in trained IκBα-SR+ mice.

In recording synaptic signaling along the Schaffer collateral/commissural pathway, the intensity of the electrical stimulation (input [I]) and the peak amplitude of the fEPSPs (output [O]) evoked were used to establish an I/O function as a measure of the basal synaptic transmission. Strikingly, untrained IκBα-SR+ mice exhibited significantly impaired basal synaptic transmission relative to levels in similarly untrained control mice (Fig. (Fig.6A).6A). In contrast, I/O functions in IκBα-SR mice were equivalent to levels detected in control mice (data not shown). In sharp contrast to the impaired I/O function observed in naive mice, maze-trained IκBα-SR+ mice displayed a completely restored I/O function (Fig. (Fig.6B6B).

FIG. 6.
Activity dependent recovery of synaptic transmissions and GluR1 levels in trained IκBα-SR+ mice. (A) Examples of fEPSPs evoked by stimulating Schaffer collaterals and recording in CA1 stratum radiatum and recorded from individual ...

One possible explanation for the impaired basal synaptic strength in untrained IκBα-SR+ mice relates to markedly decreased GluR1 levels detected in IκBα-SR+ mice. Immunoblot analysis showed that levels of AMPA-type GluR1 subunits in synaptosomal membrane preparations were significantly decreased in naive IκBα-SR+ mice compared to untrained IκBα-SR littermates or nontransgenic control mice (Fig. (Fig.6C).6C). In contrast, maze training resulted in a much greater activity-dependent increase in GluR1 levels in the IκBα-SR+ hippocampus than in either IκBα-SR littermates or control mice (Fig. (Fig.6D,6D, lane 1 versus lanes 3 and 5; note the lower levels of basal GluR1 in untrained SR+ lysates relative to SR or control lysates). A similar activity-dependent restoration of GAD65 levels was not observed in IκBα-SR+ mice. This activity-dependent increase in GluR1 may be a consequence of hyperexcitation of glutamatergic neurons occurring in the context of diminished inhibitory tone and/or altered excitatory circuits in IκBα-SR+ mice.

Synaptic signaling induced by behavioral training was evaluated by EMSA using CREB and NF-κB/Rel-specific radiolabeled probes. NF-κB/Rel DNA binding was increased in trained IκB-SR and control mice but, consistent with expression of IκBα-SR, was unaffected in IκBα-SR+ mice (Fig. (Fig.6E).6E). Of note, CREB DNA activity was slightly increased in IκBα-SR+ mice under both basal and maze-trained conditions (Fig. (Fig.6E),6E), suggesting that this factor may be activated in response to pan-neuronal NF-κB inhibition. While direct effects of IκBα-SR expression on excitatory neuronal activity cannot be excluded, the activity-dependent restoration of synaptic strength (I/O), improved synaptic signaling (LTP and Arc levels), and enhanced cognitive functions are consistent with heightened excitatory activity resulting from impaired GAD65-dependent inhibitory neuronal function in IκBα-SR+ mice.

DISCUSSION

NF-κB as a regulator of synaptic plasticity and memory formation.

Alterations in synaptic plasticity reflect composite changes occurring not only in excitatory neurons but also within inhibitory interneurons (32-34, 67). Our findings suggest a surprising and previously unrecognized role for the NF-κB/Rel family of transcriptional factors as critical modulators of the homeostatic interplay occurring between inhibitory and excitatory neuronal function. Further, our studies reveal that NF-κB is an important positive regulator of GAD65, an enzyme that is critical for establishment of GABAergic interneuron-mediated inhibitory tone in vivo.

Using the prion promoter-enhancer, we have generated a transgenic mouse model in which a dominantly acting inhibitor of NF-κB action is exclusively expressed in neurons. This inhibitor, termed IκBα-SR, is strongly expressed in GABAergic inhibitory interneurons and, to a lesser extent, in excitatory neurons. As noted, IκBα-SR expression results in decreased expression of GAD65 in GABAergic interneurons. IκBα-SR expression also leads to impaired basal synaptic signaling, likely due to decreased synaptosomal AMPAR-type glutamate receptor (GluR1) expression, resulting in sharply impaired I/O function in untrained or naive mice (Fig. (Fig.7).7). However, after these IκBα-SR+ mice are subjected to an experience- or task-based activity (e.g., maze training), AMPAR-type GluR1 levels are markedly increased in synaptosomal membranes, and I/O function is completely restored. We suspect that this training converts formerly “silent” dendritic spines into active ones, facilitating increased neuronal excitation (21). Enhanced activity-dependent synaptic signaling in IκBα-SR+ mice could reflect increased AMPAR-mediated neuronal excitation occurring as a consequence of diminished GAD65-derived inhibitory tone, although direct effects of IκBα-SR expression in excitatory neurons may also contribute to the observed phenotype. Consistent with this proposed model, IκBα-SR+ mice exhibit increased L-LTP and synaptic activity-dependent gene expression, enhanced physical and exploratory activity, higher incidence of seizures, and improved performance in various tests of spatial learning.

FIG. 7.
Our proposed model for altered basal and activity-dependent excitatory synaptic signaling in IκBα-SR+ versus IκBα-SR mice. Under basal conditions, signaling through AMPA-type GluR1 in IκBα-SR ...

Expression of IκBα-SR in GABAergic interneurons results in decreased GAD65 expression.

Excitatory neurons and inhibitory interneurons represent the opposing “yin-yang” of synaptic function and memory formation. Notably, during any given behavioral task, >90% of excitatory neurons remain silent, whereas almost all of the inhibitory interneurons are active (18, 27). Recently, a number of studies have implicated the NF-κB/Rel family of transcriptional regulators in excitatory neuronal function and spatial memory (19, 49). However, the role of these factors in inhibitory GABAergic interneurons, which comprise more than 30% of all neurons in the adult mammalian central nervous system, has not been explored.

IκBα-SR+ expression resulted in decreased transcription of GAD65, a rate-limiting enzyme required for GABA synthesis in GABAergic interneurons and generation of the inhibitory tone. GABA is formed from the alpha-decarboxylation of glutamate by the GAD isoforms, GAD65 and GAD67 (10). In the dentate gyrus and CA1 region of the rat hippocampus, GAD65 is localized primarily in synaptosomes and regulates the vesicular pool of GABA, allowing responses to short-term increases in demand during activity-dependent synaptic signaling (56, 68). GAD67 appears to be primarily responsible for the synthesis of the metabolic GABA pool and supports tonic levels of synaptic transmission (11, 12). Mice in which the GAD67 gene is disrupted die at birth, likely as a result of the dramatically lower production of GABA synthesis; GAD67−/− mice have 90% less GABA levels than normal mice (4, 9). In contrast, GAD65−/− mice survive, and the total GABA content is only marginally decreased (26, 35, 36). However, GAD65−/− mice are prone to seizures (35), have diminished GABA release following K+ stimulation of the visual cortex, and exhibit altered visual cortical plasticity (26). These findings suggest that GAD65 plays an important role in GABAergic synaptic transmission. Indeed, in view of the large amounts of GABA in neuronal cell bodies and the different intraneuronal distributions of GAD65 and GAD67, it has been suggested that GAD67 might be involved in the synthesis of GABA for general metabolic activity through the tricarboxylic acid cycle, whereas GAD65-derived GABA participates in regulating synaptic transmission at active spines (63).

The long-term regulation of GAD is complex, involving both transcriptional and posttranscriptional mechanisms. Studies of gad67 and gad65 gene expression as well as analysis of their significantly different regulatory regions suggest that transcriptional regulation involves different intracellular mechanisms (63). Our observation of a specific decrease in GAD65, but not GAD67, mRNA transcripts and protein levels in IκBα-SR+ mice raises the possibility that GAD65 may correspond to an NF-κB target gene in GABAergic interneurons. Alternatively, changes in GAD65 expression may involve more indirect mechanisms resulting from the pan-neuronal expression of IκBα-SR+ in these mice.

Loss of NF-κB action in neurons leads to hyperexcitability and enhanced LTP.

Experience-based neuronal activity results in progressive depolarization of the postsynaptic neuron in response to glutamate. In contrast, stimulated inhibitory GABAergic neurons synthesize and release GABA (47), triggering hyperpolarization of the postsynaptic neuron. This GABA-mediated inhibitory tone essentially acts as a neurochemical brake to inhibit the presynaptic release of other excitatory neurotransmitters and attenuate the excitatory signal (18, 27). It has previously been reported that disruption of inhibitory inputs on neurons results in unopposed excitatory firing leading to increased seizure activity (69). Consistent with this finding is the observation that GABA withdrawal or pharmacological inhibition of GABAergic function triggers neuronal hyperexcitability (8). Thus, these IκBα-SR+ animals provided a unique opportunity to explore in vivo the impact of inhibiting NF-κB/Rel action on the interplay between inhibitory interneurons and excitatory neurons regulating synaptic signaling.

As an experimental correlate of activity-dependent synaptic plasticity, LTP is typically induced in a biphasic manner following high-frequency stimulation. The early-phase LTP occurs independently of new gene expression and involves the activation of several protein kinases and the recruitment of existing AMPAR into active synapses (42, 53). In contrast, L-LTP requires new gene transcription and protein synthesis and is thus considered to be the most likely mechanism underlying the long-lasting changes required for long-term memory. Several candidate genes have been identified as molecular analogs of long-term memory. One such synaptic target is the immediate-early gene encoding the protein Arc (24). Typically induced following neuronal activation through NMDA receptor (NMDAR), synapse-specific Arc expression serves to facilitate synaptic plasticity and long-term memory consolidation and is a strong indicator of activity-dependent synapse excitability (64). Spontaneous induction of Arc expression in low Mg2+/high K+ medium (3) confirmed the hyperexcitability of IκBα-SR+ neurons likely due to reduced GAD65 expression in GABAergic interneurons leading to impaired inhibitory tone which may, in turn, explain the increased LTP and enhanced Arc expression. However, effects due to IκBα-SR expression in excitatory neurons on synaptic signaling and excitatory firing may also contribute to this phenotype and cannot be excluded.

Neuronal IκBα-SR expression alters both basal and activity-dependent synaptic signaling.

Under normal conditions, experience- or activity-based neuronal activity reflects a balance achieved primarily through glutamatergic neuronal excitation and GABAergic interneuron-mediated inhibitory tone. Our studies suggest that NF-κB may play a dual role in modulating synaptic signaling by regulating select functions in these distinct neuronal subtypes. Consistent with prior reports, our studies demonstrate that NF-κB is required during the maintenance phase of the synaptic response for regulating basal AMPAR expression and function (43, 54, 78). Conversely, during the constructive phase of activity-dependent synaptic signaling, activation of transcriptional factors including CREB, CREM (32, 33), or serum response factor (58), in addition to NF-κB, leads to increased AMPAR expression and enhanced excitatory firing (32, 77). Simultaneously, activity-dependent induction of NF-κB/Rel action in GABAergic interneurons increases GAD65 levels, resulting in enhanced GABA-mediated inhibitory tone effectively attenuating excitatory neuron firing.

Increased LTP and enhanced learning and memory have also been reported in two other studies involving overexpression of either the NMDA-type glutamate receptor, NR2B, or the KIF17 kinesin motor protein that are required for trafficking of these receptors to active spines in excitatory neurons (66, 75). Our findings suggest that IκBα-SR regulation of GABAergic neuronal function may also result in increased GluR1 trafficking to active spines as a consequence of decreased GAD65-derived inhibitory tone. Additionally, IκBα-SR expression in excitatory neurons could promote increased GluR1 expression, resulting in enhanced LTP as a consequence of altering the ratios of silent synapses in the postsynaptic density. However, this latter possibility seems less likely, given the inhibitory effects of the IκBα-SR on neuronal synaptic strength under basal conditions.

Inhibition of NF-κB action by IκBα-SR expression in neurons leads to enhanced cognitive functions.

The predominant phenotypes resulting from IκBα-SR expression include increased late long-term potentiation, neuronal hyperexcitability, increased incidence of handling seizures, hyperactivity, and heightened exploratory activity. Of note, consistent with the increased exploratory phenotype induced by IκBα-SR expression, mice lacking the gene for the p50 subunit of NF-κB/Rel also exhibit increased exploratory activity (38). Altered exploratory activity has not been reported for either the p65-deficient mice or the CamKII-promoter IκBα-SR mice (19, 49). Further, transgenic mice overexpressing TNF-α, a potent activator of NF-κB, displayed decreased exploratory behavior in open-field tests (15).

However, one of the most striking phenotypes resulting from IκBα-SR expression in vivo is an enhanced performance in two independent tests of spatial learning and memory. Indeed, we observe a strong positive correlation between synaptic signaling (LTP) and radial arm maze performance (46). This enhanced spatial memory in Prp-IκBα-SR+ mice is in marked contrast with prior reports in which mice either lacking expression of RelA subunit of NF-κB (RelA−/− TNFR−/−) in neurons and glia (49) or exclusively expressing the IκBα-SR transgene in excitatory neurons (19) exhibit impaired spatial memory. One explanation for these disparate findings relates to the cell types involved. While neurons are key mediators of synaptic signaling, there is mounting evidence that glial cells, by far the most numerous cell type in the brain, are also essential for learning and memory (14, 25). Glial cells act as key modulators of glutamate-mediated neurotransmission (13, 14) and, thus, can potently affect excitatory stimulation without impacting inhibitory tone. Consequently, loss of both TNF-α receptor and RelA/p65 expression in both neurons and glial cells in the study of Meffert et al. (49) could affect synaptic plasticity quite differently than inhibiting NF-κB/Rel action only in neurons. Another important difference between these two models relates to the fact that Meffert et al. selectively deleted only the RelA subunit of NF-κB, while we have employed a more broadly acting inhibitor that potently impairs many, if not all, members of the NF-κB/Rel family. Potential compensatory effects of other Rel family members could contribute to the differences observed.

In the other IκBα-SR transgenic model described by Kaltschmidt et al. (19), the CamKII promoter was used to specifically target IκBα-SR transgene expression to neurons in the forebrain. Such expression leads to diminished LTP and impaired spatial learning and memory (29). CamKII-SR mice express the SR transgene exclusively in excitatory neurons in the forebrain, while Prp-IκBα-SR mice express the transgene robustly in inhibitory GABAergic interneurons and, to a lesser extent, in excitatory neurons in multiple brain regions. Notwithstanding the key differences between these two model systems with respect to promoter activity, cell types, and brain regions involved as well as developmental expression, these mice do serve as interesting contrasts with respect to inhibiting NF-κB action in different neuronal cell types that regulate synaptic plasticity through fundamentally opposing forces, i.e., excitatory versus inhibitory neurotransmissions. Thus, it is not surprising that quite different spatial learning and memory phenotypes are observed in mice with altered excitatory circuits in the context of intact versus impaired inhibitory neuronal function.

Various forms of memory formation and retrieval are likely to be contingent on different transcriptional mechanisms occurring in discrete cell types in select brain regions and stored with different time constants. While a number of studies have implicated the NF-κB/Rel family of transcription factors as positive regulators of excitatory neuronal function and spatial memory, our study now also identifies NF-κB/Rel as an important regulator of inhibitory interneuron function through its effects on GAD65 expression. Overall, we suggest that NF-κB plays an important role in determining the balance between inhibitory neuronal activity and excitatory neuronal firing that importantly shapes changes in synaptic plasticity, LTP, and memory formation.

Acknowledgments

We thank Stanley Prusiner (University of California, San Francisco) for supplying the Prp/tTa mice and Paul Worley (Johns Hopkins) for generously providing Arc antibodies. We also thank Bobby Benitez, JoDee Fish, Jorge Palop-Esteban, and Vikram Rao (Gladstone Institutes) for their technical expertise and helpful comments. We also thank members of the Greene laboratory, Lennart Mucke, Steven Finkbeiner, and Michael Cowley, for helpful discussions. We are very grateful to Tom MacMahon, Kevin Deitchman, and Robert Messing (Ernest Gallo Clinic and Research Center) for their collegiality and invaluable assistance with the radial arm maze. We thank Angela Rizk-Jackson and Timothy F. Pfankuch for their assistance with the behavioral testing. We thank John Carroll, Jack Hull, Stephen Gonzales, and Chris Goodfellow for assistance in the preparation of figures, Stephen Ordway and Gary Howard for editorial assistance, and Robin Givens and Sue Cammack for administrative support.

This work was supported by the J. David Gladstone Institutes (W.C.G.) and NIH grant AG 20904 (J.R.) and the Extramural Research Facilities Improvement Program Project (C06 RR018928).

REFERENCES

1. Albensi, B. C., and M. P. Mattson. 2000. Evidence for the involvement of TNF and NF-κB in hippocampal synaptic plasticity. Synapse 35:151-159. [PubMed]
2. Arancio, O., H. P. Zhang, X. Chen, C. Lin, F. Trinchese, D. Puzzo, S. Liu, A. Hegde, S. F. Yan, A. Stern, J. S. Luddy, L. F. Lue, D. G. Walker, A. Roher, M. Buttini, L. Mucke, W. Li, A. M. Schmidt, M. Kindy, P. A. Hyslop, D. M. Stern, and S. S. Du Yan. 2004. RAGE potentiates Aβ-induced perturbation of neuronal function in transgenic mice. EMBO J. 23:4096-4105. [PMC free article] [PubMed]
3. Arias, R. L., and M. R. Bowlby. 2005. Pharmacological characterization of antiepileptic drugs and experimental analgesics on low magnesium-induced hyperexcitability in rat hippocampal slices. Brain Res. 1047:233-244. [PubMed]
4. Asada, H., Y. Kawamura, K. Maruyama, H. Kume, R. Ding, F. Y. Ji, N. Kanbara, H. Kuzume, M. Sanbo, T. Yagi, and K. Obata. 1996. Mice lacking the 65 kDa isoform of glutamic acid decarboxylase (GAD65) maintain normal levels of GAD67 and GABA in their brains but are susceptible to seizures. Biochem. Biophys. Res. Commun. 229:891-895. [PubMed]
5. Barmada, S., P. Piccardo, K. Yamaguchi, B. Ghetti, and D. A. Harris. 2004. GFP-tagged prion protein is correctly localized and functionally active in the brains of transgenic mice. Neurobiol. Dis. 16:527-537. [PubMed]
6. Brown, K., S. Gerstberger, L. Carlson, G. Franzoso, and U. Siebenlist. 1995. Control of IκBα proteolysis by site-specific, signal-induced phosphorylation. Science 267:1485-1488. [PubMed]
7. Carlezon, W. A., Jr., R. S. Duman, and E. J. Nestler. 2005. The many faces of CREB. Trends Neurosci. 28:436-445. [PubMed]
8. Casasola, C., T. Montiel, E. Calixto, and S. Brailowsky. 2004. Hyperexcitability induced by GABA withdrawal facilitates hippocampal long-term potentiation. Neuroscience 126:163-171. [PubMed]
9. Condie, B. G., G. Bain, D. I. Gottlieb, and M. R. Capecchi. 1997. Cleft palate in mice with a targeted mutation in the gamma-aminobutyric acid-producing enzyme glutamic acid decarboxylase 67. Proc. Natl. Acad. Sci. USA 94:11451-11455. [PMC free article] [PubMed]
10. Erlander, M. G., N. J. Tillakaratne, S. Feldblum, N. Patel, and A. J. Tobin. 1991. Two genes encode distinct glutamate decarboxylases. Neuron 7:91-100. [PubMed]
11. Erlander, M. G., and A. J. Tobin. 1991. The structural and functional heterogeneity of glutamic acid decarboxylase: a review. Neurochem. Res. 16:215-226. [PubMed]
12. Feldblum, S., M. G. Erlander, and A. J. Tobin. 1993. Different distributions of GAD65 and GAD67 mRNAs suggest that the two glutamate decarboxylases play distinctive functional roles. J. Neurosci. Res. 34:689-706. [PubMed]
13. Fields, R. D. 2004. The other half of the brain. Sci. Am. 290:54-61. [PubMed]
14. Fields, R. D., and B. Stevens-Graham. 2002. New insights into neuron-glia communication. Science 298:556-562. [PMC free article] [PubMed]
15. Fiore, M., L. Probert, G. Kollias, K. Akassoglou, E. Alleva, and L. Aloe. 1996. Neurobehavioral alterations in developing transgenic mice expressing TNF-α in the brain. Brain Behav. Immun. 10:126-138. [PubMed]
16. Freudenthal, R., and A. Romano. 2000. Participation of Rel/NF-κB transcription factors in long-term memory in the crab Chasmagnathus. Brain Res. 855:274-281. [PubMed]
17. Freudenthal, R., A. Romano, and A. Routtenberg. 2004. Transcription factor NF-κB activation after in vivo perforant path LTP in mouse hippocampus. Hippocampus 14:677-683. [PubMed]
18. Freund, T. F., and G. Buzsaki. 1996. Interneurons of the hippocampus. Hippocampus 6:347-470. [PubMed]
19. Fridmacher, V., B. Kaltschmidt, B. Goudeau, D. Ndiaye, F. M. Rossi, J. Pfeiffer, C. Kaltschmidt, A. Israel, and S. Memet. 2003. Forebrain-specific neuronal inhibition of nuclear factor-κB activity leads to loss of neuroprotection. J. Neurosci. 23:9403-9408. [PubMed]
20. Furukawa, K., and M. P. Mattson. 1998. The transcription factor NF-κB mediates increases in calcium currents and decreases in NMDA- and AMPA/kainate-induced currents induced by tumor necrosis factor-α in hippocampal neurons. J. Neurochem. 70:1876-1886. [PubMed]
21. Gazzaley, A., S. Kay, and D. L. Benson. 2002. Dendritic spine plasticity in hippocampus. Neuroscience 111:853-862. [PubMed]
22. Gossen, M., and H. Bujard. 1992. Tight control of gene expression in mammalian cells by tetracycline-responsive promoters. Proc. Natl. Acad. Sci. USA 89:5547-5551. [PMC free article] [PubMed]
23. Guerrini, L., F. Blasi, and S. Denis-Donini. 1995. Synaptic activation of NF-κB by glutamate in cerebellar granule neurons in vitro. Proc. Natl. Acad. Sci. USA 92:9077-9081. [PMC free article] [PubMed]
24. Guzowski, J. F., G. L. Lyford, G. D. Stevenson, F. P. Houston, J. L. McGaugh, P. F. Worley, and C. A. Barnes. 2000. Inhibition of activity-dependent Arc protein expression in the rat hippocampus impairs the maintenance of long-term potentiation and the consolidation of long-term memory. J. Neurosci. 20:3993-4001. [PubMed]
25. Haydon, P. G. 2001. GLIA: listening and talking to the synapse. Nat. Rev. Neurosci. 2:185-193. [PubMed]
26. Hensch, T. K., M. Fagiolini, N. Mataga, M. P. Stryker, S. Baekkeskov, and S. F. Kash. 1998. Local GABA circuit control of experience-dependent plasticity in developing visual cortex. Science 282:1504-1508. [PMC free article] [PubMed]
27. Henze, D. A., Z. Borhegyi, J. Csicsvari, A. Mamiya, K. D. Harris, and G. Buzsaki. 2000. Intracellular features predicted by extracellular recordings in the hippocampus in vivo. J. Neurophysiol. 84:390-400. [PubMed]
28. Hong, E. J., A. E. West, and M. E. Greenberg. 2005. Transcriptional control of cognitive development. Curr. Opin. Neurobiol. 15:21-28. [PubMed]
29. Kaltschmidt, B., D. Ndiaye, M. Korte, S. Pothion, L. Arbibe, M. Prullage, J. Pfeiffer, A. Lindecke, V. Staiger, A. Israel, C. Kaltschmidt, and S. Memet. 2006. NF-κB regulates spatial memory formation and synaptic plasticity through protein kinase A/CREB signaling. Mol. Cell. Biol. 26:2936-2946. [PMC free article] [PubMed]
30. Kaltschmidt, C., B. Kaltschmidt, and P. A. Baeuerle. 1993. Brain synapses contain inducible forms of the transcription factor NF-κB. Mech. Dev. 43:135-147. [PubMed]
31. Kaltschmidt, C., B. Kaltschmidt, and P. A. Baeuerle. 1995. Stimulation of ionotropic glutamate receptors activates transcription factor NF-κB in primary neurons. Proc. Natl. Acad. Sci. USA 92:9618-9622. [PMC free article] [PubMed]
32. Kandel, E. R. 2001. The molecular biology of memory storage: a dialog between genes and synapses. Biosci. Rep. 21:565-611. [PubMed]
33. Kandel, E. R., and C. Pittenger. 1999. The past, the future and the biology of memory storage. Philos. Trans. R. Soc. Lond. B 354:2027-2052. [PMC free article] [PubMed]
34. Kaplan, M. P., and T. Abel. 2003. Genetic approaches to the study of synaptic plasticity and memory storage. CNS Spectr. 8:597-610. [PubMed]
35. Kash, S. F., R. S. Johnson, L. H. Tecott, J. L. Noebels, R. D. Mayfield, D. Hanahan, and S. Baekkeskov. 1997. Epilepsy in mice deficient in the 65-kDa isoform of glutamic acid decarboxylase. Proc. Natl. Acad. Sci. USA 94:14060-14065. [PMC free article] [PubMed]
36. Kash, S. F., L. H. Tecott, C. Hodge, and S. Baekkeskov. 1999. Increased anxiety and altered responses to anxiolytics in mice deficient in the 65-kDa isoform of glutamic acid decarboxylase. Proc. Natl. Acad. Sci. USA 96:1698-1703. [PMC free article] [PubMed]
37. Kassed, C. A., and M. Herkenham. 2004. NF-kappaB p50-deficient mice show reduced anxiety-like behaviors in tests of exploratory drive and anxiety. Behav. Brain Res. 154:577-584. [PubMed]
38. Kassed, C. A., A. E. Willing, S. Garbuzova-Davis, P. R. Sanberg, and K. R. Pennypacker. 2002. Lack of NF-κB p50 exacerbates degeneration of hippocampal neurons after chemical exposure and impairs learning. Exp. Neurol. 176:277-288. [PubMed]
39. Kelleher, R. J., III, A. Govindarajan, and S. Tonegawa. 2004. Translational regulatory mechanisms in persistent forms of synaptic plasticity. Neuron 44:59-73. [PubMed]
40. Lavon, I., I. Goldberg, S. Amit, L. Landsman, S. Jung, B. Z. Tsuberi, I. Barshack, J. Kopolovic, E. Galun, H. Bujard, and Y. Ben-Neriah. 2000. High susceptibility to bacterial infection, but no liver dysfunction, in mice compromised for hepatocyte NF-κB activation. Nat. Med. 6:573-577. [PubMed]
41. Lilienbaum, A., and A. Israel. 2003. From calcium to NF-κB signaling pathways in neurons. Mol. Cell. Biol. 23:2680-2698. [PMC free article] [PubMed]
42. Lissin, D. V., S. N. Gomperts, R. C. Carroll, C. W. Christine, D. Kalman, M. Kitamura, S. Hardy, R. A. Nicoll, R. C. Malenka, and M. von Zastrow. 1998. Activity differentially regulates the surface expression of synaptic AMPA and NMDA glutamate receptors. Proc. Natl. Acad. Sci. USA 95:7097-7102. [PMC free article] [PubMed]
43. Marini, A. M., X. Jiang, X. Wu, F. Tian, D. Zhu, P. Okagaki, and R. H. Lipsky. 2004. Role of brain-derived neurotrophic factor and NF-κB in neuronal plasticity and survival: from genes to phenotype. Restor. Neurol. Neurosci. 22:121-130. [PubMed]
44. Matsuoka, Y., Y. Kitamura, M. Okazaki, K. Terai, and T. Taniguchi. 1999. Kainic acid-induced activation of nuclear factor-κB in rat hippocampus. Exp. Brain Res. 124:215-222. [PubMed]
45. Mattson, M. P., and S. Camandola. 2001. NF-κB in neuronal plasticity and neurodegenerative disorders. J. Clin. Investig. 107:247-254. [PMC free article] [PubMed]
46. Mazzucchelli, C., C. Vantaggiato, A. Ciamei, S. Fasano, P. Pakhotin, W. Krezel, H. Welzl, D. P. Wolfer, G. Pages, O. Valverde, A. Marowsky, A. Porrazzo, P. C. Orban, R. Maldonado, M. U. Ehrengruber, V. Cestari, H. P. Lipp, P. F. Chapman, J. Pouyssegur, and R. Brambilla. 2002. Knockout of ERK1 MAP kinase enhances synaptic plasticity in the striatum and facilitates striatal-mediated learning and memory. Neuron 34:807-820. [PubMed]
47. McBain, C. J., and A. Fisahn. 2001. Interneurons unbound. Nat. Rev. Neurosci. 2:11-23. [PubMed]
48. Meberg, P. J., W. R. Kinney, E. G. Valcourt, and A. Routtenberg. 1996. Gene expression of the transcription factor NF-κB in hippocampus: regulation by synaptic activity. Brain Res. Mol. Brain Res. 38:179-190. [PubMed]
49. Meffert, M. K., J. M. Chang, B. J. Wiltgen, M. S. Fanselow, and D. Baltimore. 2003. NF-κB functions in synaptic signaling and behavior. Nat. Neurosci. 6:1072-1078. [PubMed]
50. Merlo, E., R. Freudenthal, and A. Romano. 2002. The IκB kinase inhibitor sulfasalazine impairs long-term memory in the crab Chasmagnathus. Neuroscience 112:161-172. [PubMed]
51. Mironov, A., Jr., D. Latawiec, H. Wille, E. Bouzamondo-Bernstein, G. Legname, R. A. Williamson, D. Burton, S. J. DeArmond, S. B. Prusiner, and P. J. Peters. 2003. Cytosolic prion protein in neurons. J. Neurosci. 23:7183-7193. [PubMed]
52. Myhrer, T. 1988. Exploratory behavior and reaction to novelty in rats with hippocampal perforant path systems disrupted. Behav. Neurosci. 102:356-362. [PubMed]
53. Nicoll, R. A. 2003. Expression mechanisms underlying long-term potentiation: a postsynaptic view. Philos. Trans. R. Soc. Lond. B 358:721-726. [PMC free article] [PubMed]
54. Ohba, S., T. Ikeda, Y. Ikegaya, N. Nishiyama, N. Matsuki, and M. K. Yamada. 2005. BDNF locally potentiates GABAergic presynaptic machineries: target-selective circuit inhibition. Cereb. Cortex. 15:291-298. [PubMed]
55. O'Neill, L. A., and C. Kaltschmidt. 1997. NF-κB: a crucial transcription factor for glial and neuronal cell function. Trends Neurosci. 20:252-258. [PubMed]
56. Petroff, O. A. 2002. GABA and glutamate in the human brain. Neuroscientist 8:562-573. [PubMed]
57. Raber, J., D. Wong, G. Q. Yu, M. Buttini, R. W. Mahley, R. E. Pitas, and L. Mucke. 2000. Apolipoprotein E and cognitive performance. Nature 404:352-354. [PubMed]
58. Ramanan, N., Y. Shen, S. Sarsfield, T. Lemberger, G. Schutz, D. J. Linden, and D. D. Ginty. 2005. SRF mediates activity-induced gene expression and synaptic plasticity but not neuronal viability. Nat. Neurosci. 8:759-767. [PubMed]
59. Rampon, C., Y. P. Tang, J. Goodhouse, E. Shimizu, M. Kyin, and J. Z. Tsien. 2000. Enrichment induces structural changes and recovery from nonspatial memory deficits in CA1 NMDAR1-knockout mice. Nat. Neurosci. 3:238-244. [PubMed]
60. Rao, V. R., S. A. Pintchovski, J. Chin, C. L. Peebles, S. Mitra, and S. Finkbeiner.2006. AMPA receptors regulate transcription of the plasticity-related immediate-early gene Arc. Nat. Neurosci. 9:887-895. [PubMed]
61. Reuther, J. Y., G. W. Reuther, D. Cortez, A. M. Pendergast, and A. S. Baldwin, Jr. 1998. A requirement for NF-κB activation in Bcr-Abl-mediated transformation. Genes Dev. 12:968-981. [PMC free article] [PubMed]
62. Scott, R., R. Bourtchuladze, S. Gossweiler, J. Dubnau, and T. Tully. 2002. CREB and the discovery of cognitive enhancers. J. Mol. Neurosci. 19:171-177. [PubMed]
63. Soghomonian, J. J., and D. L. Martin. 1998. Two isoforms of glutamate decarboxylase: why? Trends Pharmacol. Sci. 19:500-505. [PubMed]
64. Steward, O., and P. Worley. 2002. Local synthesis of proteins at synaptic sites on dendrites: role in synaptic plasticity and memory consolidation? Neurobiol. Learn. Mem. 78:508-527. [PubMed]
65. Sun, S. C., P. A. Ganchi, D. W. Ballard, and W. C. Greene. 1993. NF-kappa B controls expression of inhibitor I kappa B alpha: evidence for an inducible autoregulatory pathway. Science 259:1912-1915. [PubMed]
66. Tang, Y. P., E. Shimizu, G. R. Dube, C. Rampon, G. A. Kerchner, M. Zhuo, G. Liu, and J. Z. Tsien. 1999. Genetic enhancement of learning and memory in mice. Nature 401:63-69. [PubMed]
67. Teyler, T. J., and P. DiScenna. 1987. Long-term potentiation. Annu. Rev. Neurosci. 10:131-161. [PubMed]
68. Tian, N., C. Petersen, S. Kash, S. Baekkeskov, D. Copenhagen, and R. Nicoll. 1999. The role of the synthetic enzyme GAD65 in the control of neuronal gamma-aminobutyric acid release. Proc. Natl. Acad. Sci. USA 96:12911-12916. [PMC free article] [PubMed]
69. Treiman, D. M. 2001. GABAergic mechanisms in epilepsy. Epilepsia 42(Suppl. 3):8-12. [PubMed]
70. Tremblay, P., Z. Meiner, M. Galou, C. Heinrich, C. Petromilli, T. Lisse, J. Cayetano, M. Torchia, W. Mobley, H. Bujard, S. J. DeArmond, and S. B. Prusiner. 1998. Doxycycline control of prion protein transgene expression modulates prion disease in mice. Proc. Natl. Acad. Sci. USA 95:12580-12585. [PMC free article] [PubMed]
71. van Hogerlinden, M., B. L. Rozell, L. Ahrlund-Richter, and R. Toftgard. 1999. Squamous cell carcinomas and increased apoptosis in skin with inhibited Rel/nuclear factor-kappaB signaling. Cancer Res. 59:3299-3303. [PubMed]
72. Waltereit, R., and M. Weller. 2003. Signaling from cAMP/PKA to MAPK and synaptic plasticity. Mol. Neurobiol. 27:99-106. [PubMed]
73. West, A. E., E. C. Griffith, and M. E. Greenberg. 2002. Regulation of transcription factors by neuronal activity. Nat. Rev. Neurosci. 3:921-931. [PubMed]
74. Whiteside, S. T., M. K. Ernst, O. LeBail, C. Laurent-Winter, N. Rice, and A. Israel. 1995. N- and C-terminal sequences control degradation of MAD3/IκBα alpha in response to inducers of NF-κB activity. Mol. Cell. Biol. 15:5339-5345. [PMC free article] [PubMed]
75. Wong, R. W., M. Setou, J. Teng, Y. Takei, and N. Hirokawa. 2002. Overexpression of motor protein KIF17 enhances spatial and working memory in transgenic mice. Proc. Natl. Acad. Sci. USA 99:14500-14505. [PMC free article] [PubMed]
76. Yavin, Z., and E. Yavin. 1980. Survival and maturation of cerebral neurons on poly(l-lysine) surfaces in the absence of serum. Dev. Biol. 75:454-459. [PubMed]
77. Yin, J. C., and T. Tully. 1996. CREB and the formation of long-term memory. Curr. Opin. Neurobiol. 6:264-268. [PubMed]
78. Yu, Z., G. Cheng, X. Wen, G. D. Wu, W. T. Lee, and D. Pleasure. 2002. Tumor necrosis factor alpha increases neuronal vulnerability to excitotoxic necrosis by inducing expression of the AMPA-glutamate receptor subunit GluR1 via an acid sphingomyelinase- and NF-κB-dependent mechanism. Neurobiol. Dis. 11:199-213. [PubMed]

Articles from Molecular and Cellular Biology are provided here courtesy of American Society for Microbiology (ASM)
PubReader format: click here to try

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...