Positive and negative regulation of EAAT2 by NF-κB: a role for N-myc in TNFα-controlled repression
2005; Springer Nature; Volume: 24; Issue: 3 Linguagem: Inglês
10.1038/sj.emboj.7600555
ISSN1460-2075
AutoresRaquel Sitcheran, Pankaj Gupta, Paul B. Fisher, Albert S. Baldwin,
Tópico(s)Cell Adhesion Molecules Research
ResumoArticle20 January 2005free access Positive and negative regulation of EAAT2 by NF-κB: a role for N-myc in TNFα-controlled repression Raquel Sitcheran Raquel Sitcheran Lineberger Comprehensive Cancer Center, University of North Carolina at Chapel Hill, NC, USA Search for more papers by this author Pankaj Gupta Pankaj Gupta Department of Pathology, College of Physicians and Surgeons, Columbia University Medical Center, New York, NY, USA Search for more papers by this author Paul B Fisher Paul B Fisher Department of Pathology, College of Physicians and Surgeons, Columbia University Medical Center, New York, NY, USA Departments of Neurosurgery and Urology, College of Physicians and Surgeons, Columbia University Medical Center, New York, NY, USA Search for more papers by this author Albert S Baldwin Corresponding Author Albert S Baldwin Lineberger Comprehensive Cancer Center, University of North Carolina at Chapel Hill, NC, USA Department of Biology, University of North Carolina at Chapel Hill, NC, USA Search for more papers by this author Raquel Sitcheran Raquel Sitcheran Lineberger Comprehensive Cancer Center, University of North Carolina at Chapel Hill, NC, USA Search for more papers by this author Pankaj Gupta Pankaj Gupta Department of Pathology, College of Physicians and Surgeons, Columbia University Medical Center, New York, NY, USA Search for more papers by this author Paul B Fisher Paul B Fisher Department of Pathology, College of Physicians and Surgeons, Columbia University Medical Center, New York, NY, USA Departments of Neurosurgery and Urology, College of Physicians and Surgeons, Columbia University Medical Center, New York, NY, USA Search for more papers by this author Albert S Baldwin Corresponding Author Albert S Baldwin Lineberger Comprehensive Cancer Center, University of North Carolina at Chapel Hill, NC, USA Department of Biology, University of North Carolina at Chapel Hill, NC, USA Search for more papers by this author Author Information Raquel Sitcheran1, Pankaj Gupta2, Paul B Fisher2,3 and Albert S Baldwin 1,4 1Lineberger Comprehensive Cancer Center, University of North Carolina at Chapel Hill, NC, USA 2Department of Pathology, College of Physicians and Surgeons, Columbia University Medical Center, New York, NY, USA 3Departments of Neurosurgery and Urology, College of Physicians and Surgeons, Columbia University Medical Center, New York, NY, USA 4Department of Biology, University of North Carolina at Chapel Hill, NC, USA *Corresponding author. 22-000 Lineberger Comprehensive Cancer Center, CB#7295, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599-7295, USA. Tel.: +1 919 966 3652; Fax: +1 919 966 0444; E-mail: [email protected] or [email protected] The EMBO Journal (2005)24:510-520https://doi.org/10.1038/sj.emboj.7600555 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions Figures & Info The glutamate transporter gene, EAAT2/GLT-1, is induced by epidermal growth factor (EGF) and downregulated by tumor necrosis factor α (TNFα). While TNFα is generally recognized as a positive regulator of NF-κB-dependent gene expression, its ability to control transcriptional repression is not well characterized. Additionally, the regulation of NF-κB by EGF is poorly understood. Herein, we demonstrate that both TNFα-mediated repression and EGF-mediated activation of EAAT2 expression require NF-κB. We show that EGF activates NF-κB independently of signaling to IκB. Furthermore, TNFα can abrogate IKKβ- and p65-mediated activation of EAAT2. Our results suggest that NF-κB can intrinsically activate EAAT2 and that TNFα mediates repression through a distinct pathway also requiring NF-κB. Consistently, we find that N-myc is recruited to the EAAT2 promoter with TNFα and that N-myc-binding sites are required for TNFα-mediated repression. Moreover, N-myc overexpression inhibits both basal and p65-induced activation of EAAT2. Our data highlight the remarkable specificity of NF-κB activity to regulate gene expression in response to diverse cellular signals and have implications for glutamate homeostasis and neurodegenerative disease. Introduction The excitatory amino-acid transporter, EAAT2/GLT-1, is responsible for the majority of clearance of the neurotransmitter glutamate from neuronal synapses in the central nervous system (CNS) (Tanaka et al, 1997). Impaired glutamate uptake by glial cells can result in cell death from excessive levels of glutamate and overstimulation of glutamate receptors (Choi, 1988). Indeed, glutamate toxicity is implicated in a wide variety of neurodegenerative disorders, including Alzheimer's disease, amyotrophic lateral sclerosis and multiple sclerosis (Doble, 1999). The importance of regulating glutamate transport is underscored by the observation that mice lacking EAAT2 develop progressive neurodegeneration and epilepsy as a result of aberrant glutamate homeostasis (Rothstein et al, 1996; Tanaka et al, 1997). Therefore, it is extremely important to understand the regulation of EAAT2 expression in glial cells. Interestingly, EAAT2 is positively regulated by epidermal growth factor (EGF) and negatively regulated by the proinflammatory cytokine, tumor necrosis factor α (TNFα) (Zelenaia et al, 2000; Schluter et al, 2002; Su et al, 2003). However, the requirements for growth factor- and cytokine-mediated regulation of EAAT2 expression have not been clearly elucidated. TNFα activates the NF-κB family of transcription factors, which are ubiquitously expressed and are pivotal in controlling diverse cellular processes, including immune responses, cell proliferation and differentiation (Israel, 2000; Baldwin, 2001; Silverman and Maniatis, 2001; Ghosh and Karin, 2002; Li and Verma, 2002). Increasingly, it has become evident that NF-κB also plays important roles in the CNS (O'Neill and Kaltschmidt, 1997). In most cells, the predominant induced NF-κB complex is a p65/RelA and p50 heterodimer, which is generally a positive regulator of gene-specific transcription. Activation of NF-κB typically involves regulating the stability of the inhibitory protein, IκB. Signal-induced phosphorylation, ubiquitination and degradation of IκB induce NF-κB nuclear accumulation and subsequent target-specific DNA binding. Phosphorylation of IκB is carried out by the IκB kinase (IKK) complex, which is comprised of IKKα, β and γ subunits (Israel, 2000; Ghosh and Karin, 2002; Yamamoto and Gaynor, 2004). Notably, IKKβ is the primary kinase responsible for cytokine-induced IκB phosphorylation and NF-κB activation (Ghosh and Karin, 2002). The regulation of NF-κB activity by growth factors such as EGF is less well understood. For example, EGF has been reported to stabilize IκB and block NF-κB activation (Mehta and Besner, 2003; Banan et al, 2004a, 2004b). However, in carcinoma cells that overexpress EGF receptor family members, EGF has been shown to induce IκBα degradation and NF-κB DNA binding (Sun and Carpenter, 1998; Biswas et al, 2000). Additionally, NF-κB-inducing kinase (NIK) has been reported to be complexed with the EGF receptor, which potentiates EGF activation of NF-κB (Chen et al, 2003). Consistent with EGF-controlled activation of NF-κB, Anest et al (2004) have recently reported positive regulation of the immediate-early gene c-fos by NF-κB through a mechanism involving constitutive association of p65 with the promoter. Clearly, further studies are necessary to understand the regulation of EGF-responsive genes by NF-κB. The opposing regulation of EAAT2 expression by cytokines and growth factors provides a unique opportunity to study how different physiological or pathological signals elicit distinct transcriptional responses from a single promoter. In this work, we establish that both positive and negative regulation of EAAT2 gene expression are controlled by NF-κB. We show that EGF induces NF-κB recruitment to the EAAT2 promoter in a manner that does not involve IκB degradation or enhanced p65/RelA nuclear accumulation. Furthermore, our data reveal that NF-κB is an intrinsic positive regulator of EAAT2 gene expression and that TNFα-mediated repression involves another transcription factor, N-myc. These experiments have important implications for understanding differential gene expression as well as the altered glutamate homeostasis associated with different CNS disorders. Results EGF and TNFα differentially regulate EAAT2 expression Previous reports have demonstrated that diverse signals can positively or negatively regulate the expression of EAAT2 in astrocytes and we sought to determine whether this dual regulation occurred in human H4 astroglioma cells. Using quantitative real-time RT–PCR analysis in wild-type H4 cells, we observed induction of EAAT2 mRNA in response to treatment with EGF (Figure 1A). In contrast, in response to TNFα treatment and consistent with previous reports (Su et al, 2003), EAAT2 mRNA expression was decreased. Figure 1.EAAT2 expression is differentially regulated by TNFα and EGF. (A) Quantitative real-time RT–PCR analysis of EAAT2 mRNA levels in H4 glioma cells treated with EGF or TNFα. PCR reactions were performed in triplicate and EAAT2 mRNA levels were normalized to β-actin as an endogenous control. (B) NF-κB and N-myc consensus sites in the EAAT2 promoter. The relative positions and sequences of putative NF-κB and N-myc DNA-binding sites are indicated. Arrows indicate the position of primers used for ChIPs (Figures 3 and 7). Download figure Download PowerPoint Given the strong role of NF-κB in TNFα-induced gene expression and the less well-established role of this transcription factor in EGF-induced gene expression, we analyzed the EAAT2 proximal promoter sequence for potential NF-κB-dependent regulatory elements. We identified six putative NF-κB consensus binding sites, including two sites (+133 and +265) in the region corresponding to the 5′ untranslated region (5′UTR) (Figure 1B). The −583 and −272 sites were previously reported and a reporter construct containing these NF-κB sites in the EAAT2 promoter was shown to remain responsive to both EGF- and TNFα-mediated regulation (Su et al, 2003). Results from our initial studies found that changes in EAAT2 mRNA levels were not due to alterations in NF-κB p65 or p50 protein levels resulting from EGF or TNFα treatments (data not shown), raising the potential that alteration of NF-κB transcriptional activity is associated with the differential control of EAAT2 gene expression by growth factors and cytokines. NF-κB is recruited to the EAAT2 promoter in response to TNFα and EGF To address the potential role of NF-κB in regulating EAAT2 gene expression, we asked whether TNFα and EGF induce NF-κB DNA binding to the consensus sites in the EAAT2 promoter. Electrophoretic mobility shift assays (EMSAs) were performed with nuclear extracts from H4 cells treated with TNFα or EGF. TNFα strongly induced DNA/protein complex formation at the −583 position, which consists of the consensus sequence 5′-GGGGCATCCC-3′ (Figure 2A). The binding of complex 1 was rapidly induced after 15 min of treatment and persisted for at least 4 h. Complex 2 was induced with slightly slower kinetics. We observed that EGF treatment also induced binding to this element, albeit weakly as compared to that of TNFα. EGF induction of complex 1 occurred at 1 h and complex 2 was also more weakly induced (Figure 2A). Experiments were also performed with a probe in which the −583 consensus site was mutated in order to determine whether the induced binding was specific for NF-κB. Indeed, both TNFα- and EGF-induced NF-κB complexes did not form with the mutated probe (Figure 2B). The EGF-induced increase in NF-κB DNA binding to the −583 site was further investigated using a DNA affinity purification assay (DAPA) where nuclear proteins were allowed to bind to an immobilized, biotinylated oligo containing the −583 sequence (see Materials and methods). Bound proteins were eluted and subsequently analyzed by Western blot with a p65-specific antibody. We found that NF-κB p65 DNA binding to the −583 site of the EAAT2 promoter was strongly increased in response to TNFα and weakly increased in response to EGF (Figure 2C). Results from these experiments clearly confirmed our EMSA results and also identified the NF-κB p65 subunit as part of both TNFα- and EGF-induced DNA-binding complexes. Figure 2.TNFα and EGF induce NF-κB DNA binding at the EAAT2 promoter. (A) EMSAs were performed with nuclear extracts from H4 cells treated with TNFα or EGF over a 4 h time course using a probe corresponding to the wild-type −583 site (EκB (−583)). A darker gel exposure is shown in the EGF panel to better visualize weak induction of DNA binding. (B) NF-κB binding using the same extracts from (A) was compared to a mutant EκB (−583) site. Mutated sequences are underlined. (C) DAPAs were performed by incubating H4 nuclear extracts with an immobilized, biotinylated EκB (−583) double-stranded oligo. Bound protein complexes were washed, eluted with 1 × sample buffer and subjected to PAGE and Western blot analysis with a p65-specific antibody. (D) EMSAs were performed using nuclear extracts from cells treated with TNFα (0.5 h) or EGF (4 h) in the presence or absence of p65 or p50 antibodies. Supershifted complexes are indicated with asterisks (*p65, **p50). (E) EMSA was performed with a probe corresponding to the +265 site in the EAAT2 5′UTR using H4 nuclear extracts treated with TNFα or EGF over a 2 h time course. (F) EMSA was performed using a mutant +265 probe. Mutated sequences are underlined. Download figure Download PowerPoint To further characterize the proteins in the EMSA complexes, we performed supershift assays with antibodies to specific NF-κB family members. In TNFα-treated cells, complex 1 was completely supershifted with α-p65 and α-p50 antibodies and complex 2 was partially supershifted with α-p50 (Figure 2D). In EGF-treated cells, although weak supershifted bands were formed, we observed complete loss of complex 1 binding by incubating nuclear extracts with α-p65 or α-p50, indicating that these antibodies did not generate a clear supershift but nevertheless, specifically interfered with complex 1 formation. Results from these experiments demonstrated that the TNFα- and EGF-induced NF-κB complex 1 primarily consists of the p65–p50 heterodimer (Figure 2D). The TNFα-induced complex 2 contains p50 as well as the p52 subunit, whereas c-Rel and RelB are absent (see Supplementary data). Next, we examined NF-κB binding to one of the putative consensus NF-κB sites in the region of the EAAT2 gene corresponding to the 5′UTR. Interestingly, we observed a different pattern of binding to the +265 site, which consists of the sequence 5′-GGGGCGTTCCC-3′. Results from EMSAs showed that EGF strongly induced NF-κB binding in a manner comparable to that of TNFα (Figure 2E). Additionally, EGF induced binding at 30 min that was diminished by 4 h post-treatment. In contrast, TNFα rapidly induced NF-κB binding 15 min after the addition of EGF and this binding persisted for up to 4 h, which is a pattern similar to that seen with TNFα at the −583 site (Figure 2A). We observed three specific complexes that bind to the +265 site, and mutation of this site eliminated complex formation (Figure 2F). Supershift analyses indicated that the inducible complex 2 is a heterodimer of the p65 and p50 NF-κB subunits (see Supplementary data). We are currently investigating the nature of the additional bands that form at the +265 site. Additionally, we observed binding to other NF-κB sites in the EAAT2 promoter, including −334 as well the three sites at −272, −251 and +133 (see Supplementary data). To examine the kinetics of in vivo recruitment of NF-κB proteins to the EAAT2 promoter in untreated and EGF- or TNFα-treated H4 cells, we performed chromatin immunoprecipitation (ChIP) assays using a p65-specific antibody (Figure 3A and B). The kinetics of detectable p65 recruitment to the EAAT2 promoter is greater with TNFα than EGF treatment over the course of 4 h. Moreover, the overall levels of p65 recruitment are approximately three times greater with TNFα compared with that of EGF. Although ChIP analysis does not itself measure affinity or the extent of promoter occupancy, the kinetics of p65 promoter recruitment by this approach is consistent with results from the EMSAs and real-time PCR analysis of EAAT2 expression. Overall, these data indicate that TNFα and EGF induce NF-κB recruitment to the EAAT2 promoter in vivo and that the subunit composition of NF-κB complexes is similar for both inducers. Figure 3.NF-κB p65 is present at the EAAT2 promoter in vivo in response to EGF and TNF. (A) H4 cells were treated with TNFα (20 ng/ml) or EGF (50 ng/ml) for the indicated time points. ChIP assays were performed using a p65-specific antibody. IgG precipitations are shown as a negative control. Prior to immunoprecipitation, a portion of samples was reserved for input controls. Immunoprecipitated chromatin was prepared and subjected to quantitative real-time PCR analysis using primers that amplify the −583 region (see Figure 1B and Materials and methods). Reactions were performed in triplicate and values are represented as a percentage of input DNA. (B) DNA from real-time PCR analysis was visualized by agarose gel electrophoresis with EtBr. Download figure Download PowerPoint Positive and negative regulation of EAAT2 gene expression requires NF-κB Thus far, our data demonstrate in vitro and in vivo that EGF and TNFα induce the formation of NF-κB DNA-binding complexes that are correlated with either transcriptional activation or repression. To determine whether regulation of EAAT2 expression required NF-κB, we tested the activity of an EAAT2 promoter-driven luciferase reporter and compared this to EAAT2 reporters that had mutations in NF-κB consensus sites (Figure 4A). Mutation of the −583 site significantly impaired the constitutive activation of EAAT2, indicating that NF-κB plays a role in controlling uninduced levels of gene expression. Indeed, the activity of a triple mutant reporter (−583, −272 and −251 sites) was completely abolished. Interestingly, TNFα-mediated repression of reporter activity was only partially abrogated with the −583 mutant reporter, whereas EGF-mediated activation was completely lost. The fact that TNFα-mediated repression is not completely abolished with the −583 mutant likely reflects the contribution of additional NF-κB sites that were not mutated. When we examined the activity of an EAAT2 reporter in which the NF-κB sites in the 5′UTR (+133/+265) were mutated, we again observed lower basal activity. However, in this context, we observed complete loss of TNFα-mediated repression, whereas EGF was still able to activate (Figure 4A). These results suggest that distinct NF-κB sites may contribute preferentially to activation or repression of EAAT2. Next, we monitored EAAT2 mRNA levels in H4 astroglial cells expressing either vector or IκBα super-repressor (SR), a nondegradable, dominant-negative inhibitor of all NF-κB complexes. Stable expression of IκBα-SR resulted in loss of NF-κB DNA-binding activity as assayed by EMSAs (data not shown). H4 cells stably expressing IκBα-SR were no longer sensitive to TNF-mediated inhibition of EAAT2 expression (Figure 4B) as compared with cells expressing vector alone. Additionally, induction of EAAT2 gene expression by EGF was also abrogated in H4 IκBα-SR cells. These results definitively show that NF-κB activity is required for both TNFα-mediated inhibition and EGF-mediated induction of EAAT2 expression in glial cells. Figure 4.NF-κB is required for regulation of EAAT2 expression. (A) Luciferase reporter assays were performed in H4 cells with wild-type and mutant EAAT2 promoter constructs that had mutations in three NF-κB consensus sites (see Materials and methods). Treatments were performed with 20 ng/ml TNFα or 50 ng/ml EGF. Transfections were performed in triplicate. Normalized RLUs (relative luciferase units) represent luciferase activity normalized to a Renilla reporter. (B) Wild-type and IκBα-SR-expressing H4 cells were treated with TNFα or EGF for the indicated time points. Total RNA was isolated, reverse-transcribed with MMLV-RT and quantitative real-time PCR analysis of EAAT2 cDNA was subsequently performed. EAAT2 mRNA levels were normalized to β-actin mRNA as an internal control. Download figure Download PowerPoint TNFα and EGF use different signaling pathways to regulate NF-κB Given the clear quantitative and kinetic differences in the ability of EGF and TNFα to elicit the formation of NF-κB DNA-binding complexes at the EAAT2 promoter, we set out to address whether these signals utilized the same pathways to regulate NF-κB activity. We asked whether both EGF and TNFα employed the classical mechanism of signal-induced IκBα phosphorylation and degradation to trigger NF-κB nuclear localization. In H4 cells treated with TNF, the typical kinetics of IκBα phosphorylation and degradation at 15–30 min were observed, which is consistent with induction of DNA binding at that time point (Figure 5A). However, treatment with EGF altered neither the phosphorylation nor the stability of IκBα. Similar results were also observed with IκBβ and IκBε (Figure 5B). In order to confirm activation of the EGF signaling pathway, we assessed whether downstream targets of that pathway were activated with treatment. Indeed, we observe induction of MEK1/2 phosphorylation after EGF treatment (Figure 5C). To determine whether EGF-induced stabilization of IκB affected NF-κB levels in the nucleus, we performed Western blot analysis of nuclear extracts from H4 cells. Results from these experiments showed that TNFα treatment results in greatly increased nuclear p65 levels, whereas EGF did not significantly alter the pre-existing, constitutive levels of nuclear p65 (Figure 5D). Together, these data suggest that TNFα acts through the classical IκB degradation pathway to induce NF-κB DNA binding and transcriptional repression, whereas EGF regulates the activity of constitutively nuclear NF-κB complexes independently of IκB to regulate positively EAAT2 expression. Figure 5.EGF and TNFα signal to NF-κB via different signaling pathways. Western blot analysis was performed with protein extracts from H4 cells that were treated with either TNF or EGF for the indicated time points. Cytoplasmic fractions were used unless otherwise indicated. Blots were reprobed with β-tubulin to demonstrate equal loading of protein. (A) Total and phosphorylated IκBα proteins. (B) IκBβ and IκBε proteins. (C) Phosphorylated MEK1/2 in whole cell lysates. (D) p65 protein levels in nuclear extracts from cells treated with TNF or EGF. Download figure Download PowerPoint TNFα-mediated repression overcomes intrinsic activation of EAAT2 by NF-κB We next asked whether the activating or repressive signal would be dominant with cotreatment of H4 cells with EGF and TNFα. Results from transient transfection assays with the wild-type EAAT2-luciferase reporter show that the repressive signal by TNFα prevents EGF-mediated activation of EAAT2. We sought to determine whether a signal that elicits NF-κB activation alone would activate or repress EAAT2 gene expression. Thus, we tested whether overexpression of p65 or IKKβ, which is essential for TNFα-mediated activation of NF-κB, would positively or negatively regulate EAAT2. Our results demonstrated that p65 and IKKβ expression dramatically increased wild-type EAAT2 reporter activation (Figure 6A), suggesting that NF-κB has an inherent ability to activate EAAT2. Interestingly, TNFα treatment resulted in repression of both p65- and IKKβ-mediated EAAT2 activation. Mutation of the −583, −272 and −251 NF-κB sites resulted in a significant decrease in basal EAAT2 expression and loss of the ability of IKKβ to induce gene expression (Figure 6B). Additionally, we find that TNFα activates a reporter containing NF-κB consensus sites from the MHC class I promoter, demonstrating that the repressive effects of TNFα are promoter specific (Figure 6C). Based on these data, we suggest that TNFα functions at two distinct levels to regulate NF-κB activity and repress EAAT2 gene expression: (i) inducing NF-κB DNA binding through IκB signaling and (ii) regulating different signaling pathways and/or transcriptional regulators that ultimately convert NF-κB to a transcriptional repressor (see Figure 9). Figure 6.TNFα-mediated repression overcomes activation of EAAT2 by NF-κB. (A) The effect of TNFα on EGF-, IKKβ- or p65-mediated activation of the wild-type EAAT2 reporter was tested. Normalized RLUs represent luciferase units normalized to Renilla control reporter activity. (B) The ability of IKKβ to regulate the triple mutant EAAT2 reporter was tested and compared to that of wild type. (C) TNFα-mediated repression is specific to the EAAT2 promoter. The activity of a 3x-κB-luciferase promoter containing consensus NF-κB sites from the MHC class I promoter was tested in the presence of TNFα or EGF. Download figure Download PowerPoint Figure 7.N-myc is recruited to the EAAT2 promoter in vivo. Using an N-myc antibody, ChIP assays were performed in H4 cells at various time points in the presence or absence of TNFα or EGF. Immunoprecipitated DNA was analyzed by (A) quantitative real-time PCR and (B) agarose gel electrophoresis. (C) DAPAs were performed by incubating H4 nuclear extracts with an immobilized, biotinylated double-stranded oligo that contains an N-myc consensus site (see Materials and methods). Bound protein complexes were washed, eluted with 1 × sample buffer and subjected to PAGE and Western blot analysis with an N-myc-specific antibody. Download figure Download PowerPoint A role for N-myc in TNFα-mediated repression of EAAT2 To address the mechanism by which TNFα induces repression, we investigated whether other regulatory factors were involved. The EAAT2 promoter contains consensus binding sites for several transcription factors, including N-myc (see Figure 1B), a basic helix–loop–helix transcription factor that is necessary for neurogenesis (Knoepfler et al, 2002) and is frequently amplified in neuroblastoma (Schwab, 1993). Using ChIP assays, we found that N-myc recruitment to the EAAT2 promoter is significantly increased with TNFα treatment as compared with EGF treatment (Figure 7A and B). Furthermore, using DNA purification assays, we observed increased TNFα-induced N-myc DNA binding to a probe spanning the −534 to −505 region of the EAAT2 promoter, which contains a putative CACGCG N-myc site (Figure 7C). These data suggest that N-myc may be involved in TNFα-mediated repression of EAAT2. To test this possibility, we constructed an EAAT2-luciferase reporter that had both N-myc consensus sites mutated and tested its activity in transient transfection assays (see Materials and methods). We observed that TNFα was unable to repress the activity of the N-myc mutant reporter, while EGF remained able to activate (Figure 8A). Next, we examined whether overexpression of N-myc would affect the basal activity of the EAAT2 promoter or the ability of NF-κB to activate EAAT2. Indeed, in luciferase reporter assays, we observed that N-myc repressed both basal and p65-induced activation of the wild-type EAAT2 reporter in a dose-response manner (Figure 8B). Importantly, the loss of NF-κB activation was not due to decreased p65 or p50 protein levels by N-myc (Figure 8B). Together, these data indicate that the preferential, early recruitment of N-myc to the EAAT2 promoter with TNFα treatment contributes to the repression of intrinsic NF-κB-mediated activation. Figure 8.N-myc-binding sites in the EAAT2 promoter are required for N-myc- and TNFα-mediated repression. (A) Wild-type or N-myc mutant reporters were transfected into H4 cells. Untreated and TNFα- and EGF-treated cells were assayed for luciferase activity. (B) The wild-type EAAT2-luciferase reporter in the absence or presence of p65 was cotransfected H4 cells with increasing amounts of N-myc. Western blot analysis with p65- or p50-specific antibodies was performed on whole cell lysates from transfected cells. Download figure Download PowerPoint Figure 9.A model for the differential regulation of EAAT2 expression by NF-κB. (A) TNFα signals to IKK to trigger IκB phosphorylation and degradation, thereby eliciting NF-κB nuclear translocation, DNA binding and transcriptional repression of EAAT2. We propose that TNFα functions to induce NF-κB DNA binding as well as regulate different signaling pathways and/or other regulators, such as N-myc, to convert NF-κB to a transcriptional repressor. (B) EGF signaling regulates the activity of pre-existing NF-κB nuclear complexes to induce transcriptional activation. EGF may also regulate the activity or recruitment of other regulators at the promoter that cooperate with NF-κB to regulate positively EAAT2 transcription. Download figure Download PowerPoint Discussion Here we investigate the regulation of EAAT2/GLT-1 expression by growth factors and cytokines. We show that EAAT2 gene expression is negatively regulated by TNFα and positively regulated by EGF in H4 astroglioma cells. Previous work suggesting that regulation of EAAT2 expression was mediated by NF-κB was based on inhibiting NF-κB activity using antioxidants, such as PDTC (Schreck et al, 1992; Sato et al, 1996; Zelenaia et al, 2000; Su et al, 2003). However, it has recently been demonstrated that these antioxidants have pleiotropic effects and do not
Referência(s)