| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 279, Issue 31, 32869-32881, July 30, 2004
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
B PATHWAY*
¶||
**
From the
Institute of Cell Biology and Immunology, University of Stuttgart, Allmandring 31, D-70569 Stuttgart, Germany, ¶Xantos Biomedicine AG, Fraunhoferstrasse 22, D-82152 Martinsried, Germany, ||Department of Physiology and Neurobiology, Eötvös Loránd University, Pázmány P. stny. 1/C, H-1117 Budapest, Hungary, and Department of Molecular Neurobiology, Rijksuniversiteit Groningen, Kerklaan 30, P. O. Box 14, 9750 AA Haren, The Netherlands
Received for publication, October 27, 2003 , and in revised form, May 14, 2004.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
B (NF-B) activation were apparent upon TNF treatment. Both TNFR1 and TNFR2 induced the NF-B pathway, yet with distinguishable kinetics and upstream activating components, TNFR1 only induced transient NF-B activation, whereas TNFR2 facilitated long term phosphatidylinositol 3-kinase-dependent NF-B activation strictly. Glutamate-induced triggering of the ionotropic N-methyl-D-aspartate receptor was required for the enhanced and persistent phosphatidylinositol 3-kinase-dependent NF-B activation by TNFR2, indicating a positive cooperation of TNF and neurotransmitter-induced signal pathways. TNFR2-induced persistent NF-B activity was essential for neuronal survival. Thus, the duration of NF-B activation is a critical determinant for sensitivity toward excitotoxic stress and is dependent on a differential upstream signal pathway usage of the two TNFRs. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
We have recently investigated the role of TNF and its receptors in retinal ischemia and unraveled an antagonistic function of TNFR1 and TNFR2. TNFR2 exerts neuroprotection in a phosphatidylinositol 3-kinase (PI3K) dependent manner, which is counterbalanced by the neurodegenerative action of TNFR1 (10). TNFR1 has also been associated with the cell death of hippocampal neurons responding to TNF (11), whereas inhibition of TNFR2 expression by the use of antisense oligonucleotides sensitized neuronal-like cells toward apoptosis (12). To elucidate the neuroprotective signal pathways emanating from TNFR2, we have established an in vitro model of apoptosis induction of primary cortical neurons by glutamate-triggered excitotoxicity to study mechanisms controlling apoptosis sensitivity of neurons within wild type (C57BL/6) and TNF transgenic TNFR1-/- and TNFR2-/- mice. Our data reveal, for the first time, differential activation kinetics of NF-B and differential usage of upstream signaling components by TNFR1 and TNFR2. TNFR2, but not TNFR1, induces persistent NF-B activation via a signaling pathway involving PI3K and PKB/Akt, which is strongly enhanced by N-methyl-D-aspartate receptor co-stimulation.
|
EXPERIMENTAL PROCEDURES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
-globin gene (14). Four mouse lines were investigated and showed similar TNF expression patterns differing only in the level of and duration of expression. The expression pattern was verified by in situ hybridization as described previously (15). The pBSII-SK plasmid containing a cDNA of the murine TNF gene was digested either with PstI or EcoRI for synthesis of a 35S-labeled sense or antisense probe using T3 or T7 RNA polymerase.
Primary Cortical Cell Culture—Cortical cultures were prepared as follows. Embryonic brains were extracted (embryonic day 14), the meninges were removed, and the cortices were dissected immediately. Neurons were recovered by mechanical dissociation. Cells were plated at a density of 
1 x 105 cells/cm2 (96- or 6-well plates) or 2.5 x 104 cells/cm2 (24-well plates) on poly-L-lysine (Sigma) coated plates. Neurobasal medium with B27 supplement (Invitrogen) and 0.5 mM glutamine with a final concentration of 0.1 mg/ml gentamycin (Invitrogen) and 2.5 µg/ml amphotericin B (Sigma) was used as culture medium. On the second day in vitro, cells were treated with 10 µM cytosine arabinoside for 48 h to inhibit non-neuronal cell growth. Subsequently, the medium was completely exchanged, and every third day afterward, fresh medium was added.
Cells were either stimulated with increasing concentrations of glutamate for 1 h (Figs. 2B, 3, A and C, and 8C) or for 24 h with the indicated concentrations of TNF (Figs. 3B and 8, A and B) or TNFR agonistic antibodies (Fig. 4, A-C) followed by a 1-h stimulation of 250 µM glutamate (Sigma) and the determination of the cell viability 24 h later. When indicated, cells were incubated prior to the glutamate stimulus with an inhibitor of PI3K (LY294002 (25 µM), Calbiochem, or wortmannin (100 nm), Sigma), NMDAR (MK801 (10 µM), Tocris), caspases (Z-VAD-fmk (20 µM), BACHEM), or NF-B (MG-132 (20 µM), Calbiochem; geldanamycin (0.5 µM), Calbiochem; or BAY11-7082 (20 µM), Alexis Biochemicals). To determine the kinetics of cell death induction, primary cortical neurons were incubated for 1 h with 250 and 500 µM glutamate and cell viability was assessed directly after the stimulus (=1-h) or 1 (=2-h), 3 (=4-h), 5 (=6-h), and 23 (=24-h total incubation time) h later (Fig. 2A). Agonistic antibodies (2 µg/µl; HyCult Biotechnology) were cross-linked by a mouse anti-rat antibody (2 µg/µl; Jackson ImmunoResearch, Dianova, Hamburg, Germany) to obtain optimum triggering of the TNFRs.
|
|
|
|
Western Blot—Cytosolic extracts from primary cortical neurons (2 x 106) were prepared by washing twice with ice-cold phosphate-buffered saline (PBS) followed by the addition of 0.4 ml of the cell lysis buffer (10 mM HEPES, pH 7.9, 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM DTT, and 1 mM phenylmethylsulfonyl fluoride). Lysates were incubated on ice for 20 min. 25 µl of 10% Nonidet P-40 was added for 2 min, and nuclei were collected by centrifugation. The supernatants were boiled for 5 min in Laemmli's sample buffer (2% SDS, 5% DTT), and 50 µg of protein was separated by SDS-polyacrylamide gel electrophoresis. After transfer to a nitrocellulose membrane (Hybond ECL, Amersham Biosciences), proteins were detected with a specific primary and a horse-radish peroxidase-conjugated secondary antibody using enhanced chemiluminescence according to the manufacturer's instructions (LuminGlo, New England Biolabs).
Primary antibodies included were a rabbit polyclonal antibody specific for the phospho-PKB/Akt serine 473 residue (1:1000; Cell Signaling), total PKB/Akt (1:1000; Cell Signaling), IB (1:500; Santa Cruz Biotechnology), phospho-IB (1:1000; Cell Signaling), TNF (1:1000; BIOSOURCE), and glutamate receptor 1 (1:300; Chemicon).
Immunocytochemistry and Immunohistochemistry—Cells were fixed with 4% paraformaldehyde in PBS, pH 7.4, for 20 min at room temperature, permeabilized with 0.1% Triton X-100/PBS for 5 min, and blocked with 5% goat serum in PBS. Overnight incubation with a polyclonal rabbit anti-p65 antibody (1:500; Oncogene) or rabbit anti-TNF antibody (1:1000, Santa Cruz Biotechnology) at 4 °C was followed by a 1-h incubation with a fluorescent-labeled Alexa 546 or 488 anti-rabbit antibody (1:1000; Molecular Probes) at room temperature. Neurons were labeled with a monoclonal neuronal-specific mouse anti-unique--tubulin antibody (1:2000; Babco) for 1.5 h at room temperature and a Alexa 488 or 546 anti-mouse antibody (1:1000).
Glia cells were detected using a monoclonal mouse anti-glial fibrillary acidic protein antibody (1:2000; Sigma) and a secondary Alexa 546-labeled anti-mouse antibody (1:1000). The glia cell content resulted in <0.5%.
Determination of apoptosis induction was assessed after fixation of stimulated (1 h 250 µM glutamate) and unstimulated cells by incubation with an annexin V-specific fluorescein isothiocyanate-conjugated antibody (1:20; BD Biosciences) for 1 h at room temperature followed by a 1-h incubation with DAPI (0.1 µg/ml) for nuclear staining,
For immunohistochemical staining, paraffin-embedded sections were permeabilized with 0.02% H2O2-methanol for 30 min and blocked with 10% fetal calf serum for 20 min. Sections then were incubated with specific antibodies (rabbit-anti-TNF (Genzyme) at 1:1000; rat-antimac3 (BD Biosciences) at 1:200; and rat-anti-CD45 (BD Biosciences) at 1:100 in 0.1% bovine serum albumin/Tris-buffered saline) at 4 °C overnight, washed, incubated with a biotin-labeled anti-rabbit IgG or anti-rat-IgG (Vector Laboratories at 1:200 in 2.5% bovine serum albumin/Tris-buffered saline), and stained with avidin-diaminobenzidine (Sigma) according to the manufacturer's protocol.
Electrophoretic Mobility Shift Assay (EMSA)—For preparation of nuclear extracts, 2 x 106 primary cortical neurons were washed once with cold PBS, resuspended in 0.4 ml of lysis buffer (10 mM HEPES, pH 7.9, 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM DTT, 1 mM phenylmethylsulfonyl fluoride) and then incubated on ice for 20 min. Subsequently, 25 µl of 10% Nonidet P-40 was added for 2 min, and nuclei were collected by centrifugation and resuspended in 20 mM HEPES, pH 7.9, 400 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM DTT, and 1 mM phenylmethylsulfonyl fluoride, shaken for 20 min, and then centrifuged. The supernatant containing nuclear proteins was used for EMSA. The NF-B consensus oligonucleotide (5'-ATCAGGGACTTTCCGCTGGGGACTTTCCG-3') was 5'-labeled with [
-32P]ATP using polynucleotide kinase. Binding reaction was performed by incubating 10 µg of nuclear protein extracts with 4 µg of poly(dI-dC) in a 20-µl reaction mixture containing 5 mM MgCl2,50mM KCl, 5 mM HEPES, pH 7.8, 0.2 mM EDTA, 5 mM DTT, and 10% glycerol. The double-stranded oligonucleotide probe (50,000 cpm/µl) was added for 20 min at room temperature. Binding specificity was determined by competition with excess (100 ng) unlabeled NF-B oligonucleotide. DNA/protein complexes were analyzed on non-denaturing 6% polyacrylamide gels and visualized by PhosphorImaging. For supershift assays, nuclear extracts were preincubated with p65, p50, or cRel antibodies (1 µg/ml; Santa Cruz Biotechnology) for 30 min at room temperature.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
|
Further evidence for distinct proapoptotic and anti-apoptotic pathways emanating from TNFR1 and TNFR2, respectively, was obtained by the use of TNFR subtype-specific agonistic antibodies (27), allowing selective stimulation of either receptor in wild type neurons. TNFR specificity of the antibodies was verified by analyzing reactivity on TNFR1-/- and TNFR2-/- neurons (data not shown). The TNFR1-specific agonist alone elicited neuronal cell death in wild type neurons in a dose-dependent manner and did not protect from glutamate-induced apoptosis (Fig. 4A). Conversely, selective stimulation of TNFR2 had no cytotoxic effect but instead led to an inhibition of neurotransmitter-triggered excitotoxicity (Fig. 4B). Again, the blocking of PI3K by LY294002 reverted the TNFR2-mediated protection (Fig. 4B), whereas the TNFR1-mediated cell death was not affected by the PI3K inhibitor (Fig. 4A). The critical involvement of PI3K signaling was confirmed by usage of wortmannin after selective TNFR2 stimulation (Fig. 4C) or TNF preincubation (data not shown).
NMDAR Signaling Enhances TNFR2-induced PKB/Akt Activation—PKB/Akt is considered to be a prominent and ubiquitous downstream target protein of PI3K (28). Because of the TNF- and PI3K-dependent protection of cortical neurons from glutamate-induced apoptosis, we investigated PKB/Akt activity during excitotoxic stimulation of untreated and TNF-pretreated wild type, TNFR1-/-, and TNFR2-/- mice as well as that of NR2B/TNF transgenic mice. Western blot analyses were performed with protein lysates from cortical neurons treated with TNF and/or glutamate in the presence or absence of a specific NMDAR blocker (MK801) or a selective PI3K inhibitor (LY294002). Intense and persistent PKB/Akt phosphorylation was detected in both TNFR1-/- and wild type neurons after TNF and glutamate co-stimulation (Fig. 5A, lanes 4, 8, and 12). This was partially inhibited by LY294002 in wild type cells and completely abolished in TNFR1-/- cells (Fig. 5A, lanes 6, 10, and 14). In TNFR2-/- neurons, only a weak transient PKB/Akt activation was detected upon glutamate and TNF treatment (Fig. 5A, lanes 4, 8, and 12). These data show that the TNF-induced persistent PKB/Akt phosphorylation is TNFR2-mediated and PI3K-dependent. Moreover, we reveal to date an unrecognized strong enhancement of this signal pathway by the glutamate-activated NMDAR.
|
NMDAR Signaling Increases TNFR2-induced NF-B Activation—NF-B activation in primary cortical neurons was determined by immunofluorescence microscopy of nuclear translocated p65 (Fig. 6A) as well as by the EMSA (Fig. 6, E and F). The lone TNF treatment of TNFR1-/- or TNFR2-/- neurons revealed characteristic differences in NF-B activation (Fig. 6B). Nuclear NF-B translocation after stimulation of TNFR1 reached a maximum between 20 and 40 min (Fig. 6B, white bars), whereas TNFR2-mediated NF-B translocation increased gradually, peaked after 4 h of stimulation, and was still discernable after 24 h (Fig. 6B, black bars). Wild type neurons, which express both TNFRs, exhibited intermediate translocation kinetics (Fig. 6B, gray bars). Upon glutamate stimulation of TNF-pretreated (24 h) cultures, a strong enhancement of the residual low level NF-B activity could be identified by nuclear translocation (Fig. 6C) and EMSA (Fig. 6E) in wild type and in TNFR1-/- neurons, whereas TNFR2-/- neurons showed no enhancement of NF-B activation by glutamate (Fig. 6, C, white bars, and E, lanes 2, 4, and 8). Of note, upon glutamate treatment alone, no NF-B activation was discerned (Fig. 6, C and E, lanes 3, 7, and 11). In TNF-pretreated wild type neurons, weak NF-B activity (Fig. 6, B, gray bars, and E, lane 2) strongly increased after glutamate stimulation and persisted at elevated levels for at least 4 h (Fig. 6, C, gray bars, and E, lanes 4, 8, and 12). In these cells, the enhancement was strongly reduced by MK801 and LY294002 when assessed during the first 2 h (Fig. 6, C, gray bars, and E, lanes 5, 6, 9, and 10) and NF-B activity was no longer discernable by EMSA in LY294002-treated cultures at 4 h after glutamate stimulation (Fig. 6E, lane 14). These results indicate an important contribution of a TNFR2-dependent pathway for NF-B activation in wild type neurons, in particular at later time points of glutamate exposure. TNFR1-/- neurons had comparable activation kinetics to wild type neurons with one remarkable difference. At every time point measured, NF-B activation was completely inhibited by LY294002 (Fig. 6, C, black bars, and E, lanes 6, 10, and 14), clearly suggesting that, in cortical neuronal cells, TNFR2-mediated NF-B activation is PI3K-dependent. Moreover, the virtually complete blocking of the enhancement of the TNF-induced NF-B response by MK801 discloses the essential role of NMDAR co-activation in wild type and TNFR1-/- neurons (Fig. 6E, lanes 5, 9, and 13). By contrast, PI3K dependence of neuronal TNFR1-mediated NF-B activation was apparently less stringent. For example, although the NF-B response of TNF-pretreated TNFR2-/- neurons was partially LY 294002-sensitive 1 h post-glutamate treatment, it remained completely unaffected by LY294002 at 2 h (Fig. 6, C and E, lanes 4 versus 6 and 8 versus 10). Of greater relevance, EMSA confirmed that TNFR1-mediated NF-B activation does not persist and lacks enhancement through NMDAR activation. Similar to PKB/Akt activity, NR2B/TNF transgenic neurons displayed basal NF-B activity because of the constitutive expression of TNF (Fig. 6, D and F, lane 1). Additional glutamate stimulation enhanced nuclear NF-B translocation (Fig. 6D) and its activation (Fig. 6F, lanes 2, 5, and 8), which could be partially reduced by MK801 (Fig. 6, D and F, lanes 3, 6, and 9) and fully abrogated by LY294002 2 and 4 h after glutamate stimulation (Fig. 6, D and F, lanes 7 and 10). Further, incubation of NR2B/TNF transgenic neurons with LY294002 inhibited endogenous NF-B activation as well (Fig. 6F, lane 12), indicating that chronic TNF expression in these cells apparently imposes a bias on TNF signaling toward TNFR2 pathways. Supershift analysis using antibodies against p65, p50, and cRel identified p50 and p65 as the predominant NF-B subunits (Fig. 6G) in all of the mouse strains tested. Although the latter experiment is in accordance with the view that both TNFR1 and TNFR2 use the canonical NF-B pathway (9), the differential kinetics of activation and sensitivity toward PI3K inhibitors in primary cortical neurons indicate distinct upstream signal pathway usage of the two TNFRs. TNFR1-mediated induction of NF-B is transient, whereas TNFR2 can induce a more lasting NF-B activation. The latter is strictly PI3K-dependent and enhanced by co-stimulation of the ionotropic NMDA receptor. In contrast, the TNFR1-mediated NF-B activation is less dependent on PI3K and not enhanced by NMDA receptor activation.
|
B Degradation Is Sensitive to PI3K Inhibition—To assess the molecular level of PI3K/Akt-mediated NF-B activation, IB
phosphorylation and subsequent degradation were analyzed in NR2B/TNF transgenic or TNF-pretreated wild type, TNFR1-/-, and TNFR2-/- neurons. In TNFR1-/- cells, the phosphorylation and depletion of IB occurred earlier and was stronger as compared with wild type cells (Fig. 7A, lanes 3-7). LY294002 sensitivity of IB phosphorylation and degradation was more pronounced in TNFR1-/- neurons as compared with wild type neurons (Fig. 7A, lanes 10-14). Consistent with the shorter NF-B activation kinetics in TNFR2-/- neurons, only transient IB phosphorylation and degradation were observed in these cells (Fig. 7A, lanes 3-7) that were largely LY204002-insensitive (Fig. 7A, lanes 10-14). NR2B/TNF transgenic neurons exhibited continuous (Fig. 7B, lanes 2-6) LY294002-sensitive (Fig. 7B, lanes 8-12)IB phosphorylation and depletion due to the continuous expression of TNF. This finding is consistent with the basal activation of NF-B observed in these cells (Fig. 6, C and E). LY294002 treatment significantly raised IB levels and decreased the proportion of phosphorylated IB in NR2B/TNF transgenic neurons (Fig. 7B, lane 7), indicating permanent PI3K-dependent signaling through TNFR2. The transient LY249002-insensitive reduction in IB levels observed at around 60 min (Fig. 7B, lane 11) probably reflects concomitant TNFR1 activity and/or other TNFR2-independent signals. To verify specific proteasomal degradation, IB levels were assessed in the presence of the proteasomal inhibitor, MG132, which prevented the stimulus-dependent reduction in IB protein levels in all of the neuronal cultures (Fig. 7, C and D).
|
B Activation Is a Central Mediator of Neuroprotection—To scrutinize the protective role of NF-B from the excitotoxic insult of cortical neurons, we blocked NF-B by the use of three different inhibitors known to interfere with NF-B activation at distinct levels, i.e. the assembly of the IKK complex via geldanamycin, the blocking of IB phosphorylation with BAY 11-7082, and the interference with proteasomal degradation of phosphorylated I-B via MG-132 (29-32). Although MG-132, geldanamycin, and BAY 11-7082 treatment themselves showed a slight toxic effect in the in vitro neuronal culture system, all of these inhibitors completely blocked TNF-induced protection in TNFR1-/- and wild type cortical neurons (Fig. 8, A-C). TNFR2-/- neurons served as a negative control and revealed no influence of the applied inhibitors on the excitotoxic sensitivity of these cells (Fig. 8, A-C). Likewise, the resistance of NR2B/TNF-neuronal cultures toward glutamate-induced cell death could be abrogated by pharmacological intervention with NF-B activation (Fig. 8D). These data provide strong evidence that NF-B activation and thus transcription of NF-B-dependent anti-apoptotic genes play a major role in TNF-mediated protection of neurons from excitotoxic insults. |
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
B activation and by pharmacological inhibition of signal pathway components in the TNFR knock-out mouse strains that neuroprotection is mediated via a TNFR2-PI3K-Akt-NF-B pathway in which the duration of NF-B activation is the critical determinant in mounting resistance toward excitotoxic insults.
The high concentration of soluble recombinant murine TNF required to achieve significant protection in vitro is already indicative of an involvement and critical role of TNFR2, which has at 37 °C a lower affinity for soluble TNF and a reduced signal competence as compared with TNFR1, thus necessitating higher TNF concentrations for signaling (33). Moreover, the primary neurons derived from the TNF transgenic mouse line that are protected from excitotoxicity express TNF not only in its processed form but also at the neuronal cell membrane, a prerequisite for efficient TNFR2 activation. Both an autotropic and a juxtatropic signaling of membrane-expressed TNF are conceivable. As cortical neurons co-express both TNFRs, we conclude that, under constitutive membrane TNF expression in the transgenic neurons, the neuroprotective TNFR2 pathway is dominant over TNFR1 signals. Direct support for differential roles of the two TNFRs in signaling neuronal survival comes from the use of TNFR-specific agonistic antibodies where TNFR2- but not TNFR1-specific agonists promoted protection from excitotoxicity. A differential role of TNFR1 and TNFR2 has also been reported for hippocampal neurons responding to TNF alone with TNFR1 inducing both NF-B activation and cell death, whereas TNFR2 was found to stimulate p38 mitogen-activated protein kinase (11). However, this study neither addressed the functional significance of TNFR2-mediated p38 activation nor resolved the underlying mechanism of apoptosis dominance in these cells in the presence of an apparently strong NF-B activation (11), which is considered to play an important role in neuroprotection (34). As an example, the cellular inhibitors of apoptosis proteins, well known NF-B-target gene(s) up-regulated by TNF, have been recently identified as facilitators in down-regulating apoptosis sensitivity within neurons (35, 36). The prominent role of NF-B in protection from neurodegeneration has recently received further support from a transgenic model characterized by forebrain-specific, inducible, and complete ablation of NF-B activity by I-B superrepressor (37). Hippocampal slice cultures from these NF-B-deficient animals presented with strongly increased lesions upon excitotoxic stress.
In accordance with the important role of NF-B in neuroprotection, here we show that upon in vitro glutamate exposure of TNF-treated cortical neurons, a PI3K-dependent PKB/Akt phosphorylation was ensued by NF-B activation. The kinetics of both PKB/Akt and NF-B activation support a TNFR2-initiated signal pathway to NF-B in a PI3K-dependent manner and the involvement of PKB/Akt. A potential role of other kinases located downstream of PI3K that could contribute to NF-B activation in response to TNF, e.g. PKC
, cannot be formally ruled out. However, in vivo, PKC is apparently a tissue-selective NF-B activator with its dominant function (with respect to NF-B activation) in lung tissues (39). Moreover, the main action of PKC in the context of TNF-mediated NF-B activation has now been recognized to be at the level of RelA transactivation (38). In contrast, our data indicate that PI3K operates at a level upstream of IB degradation, i.e. probable at the level of IKK activation or IB itself as shown in other cell systems (39-43). Examples with established cell lines show that PKB/Akt may serve as an IKK kinase in response to TNF (42). More recent data indicate that IKK is the prime target of PKB/Akt phosphorylation in the heterotrimeric IKK complex, resulting in IKK activation and subsequent IkB phosphorylation (47). Our data with primary neurons reveal PKB/Akt activation in a TNFR2-specific manner. Strict PI3K dependence of TNFR2-mediated IB phosphorylation and activation of p50/p65 as well as neuroprotection is in full agreement with a cell context-specific function of PKB/Akt in the canonical NF-B pathway.
Of note, NF-B activation via TNFR1-induced pathways did not protect cortical neurons from glutamate-induced apoptosis, which is in accordance with data from hippocampal neurons responding to TNF only (11). Interestingly, our data reveal that, in primary neurons, TNFR1- and TNFR2-induced pathways differ with respect to the kinetics and duration of NF-B activation with the TNFR1 inducing a rapid but transient NF-B response and TNFR2 generating a more persistent response, still discernible after several hours when stimulated with TNF alone. As an implication, the duration of NF-B activation is critical to achieve significant tissue protection. The usage of different upstream activators of NF-B by TNFR1 and TNFR2 is a probable explanation for the observed differential kinetics and duration of the NF-B response. Indeed, it becomes increasingly apparent that the cell-specific quantitative composition with signal proteins contributing to a complex and multi-regulated pathway critically determines response phenotypes. For NF-B activation, for example, a multitude of combinatorial possibilities exist, constituting canonical and non-canonical pathways, and comprise of several potential upstream IKK-activating kinases and several IB isoforms with differential sensitivity to degradation and feedback regulation (45-48). The individual composition of a cell with the relative amounts of each of these signaling molecules may not only influence the potential linkage to receptor proximal signals but also the kinetics and duration of NF-B activation. Our data provide a clear example that differences in the initial signal input (TNFR1 versus TNFR2 pathway), although activating the same intracellular key molecule (NF-B), are translated into a completely different signal output (cell death versus survival) by virtue of a differential duration of the active state of this critical molecule.
Blocking the NMDA receptor with MK801 lead to a reduction of PKB/Akt and NF-B activation, revealing an as of yet unknown cross-talk among a cytokine receptor, TNFR2, and the ionotropic NMDA receptor. It can be speculated that this positive cooperation with respect to PKB/Akt and NF-B activation may serve as a regulative feedback to limit cytotoxic signals emanating from overactivated NMDARs. In favor of this reasoning, only persistent NF-B activation is associated with protection from cell death. Our data show that this requires a cooperative action of TNFR2 with the ionotropic NMDA receptor, which meet above or at the level of PI3K-PKB/Akt. The molecular mechanisms of the NMDAR-TNFR2 cooperation are not yet resolved. It is apparent from sequential stimulation studies (data not shown) that the signals emanating from TNFR2 have to precede the excitotoxic insult. At first glance, this would be in accordance with strictly TNF-mediated acquisition of an apoptosis-resistant state of the cell via an NF-B-dependent transcriptional response without a contribution of NMDAR signals. However, pharmacological inhibition of the PI3K pathway just before glutamate stimulation not only blocked the TNFR2-dependent and NMDAR-enhanced stimulation of PKB/Akt and NF-B but also fully abrogated protection from apoptosis. This finding suggests that the acute signal cross-talk between the two receptor systems is essential for establishing a protective response. We propose that sustained PKB/Akt activity ensures prolonged NF-B activity and counterbalances the negative feedback of NF-B-dependent IB expression typical for TNFR1-mediated signals causing transient NF-B activity (9). Indeed, neurons protected from excitotoxicity showed consistently reduced/lower IkB levels as compared with sensitive ones. In support of our reasoning, a recent study has linked calcium-induced pathways via PI3K and Akt toward NF-B activation in neuronal cells (51).
A critical function of other PI3K-PKB/Akt pathways in neuronal survival has been proposed, and several downstream targets aside from NF-B, in particular Bad, potentially interfere with apoptotic signals and could be involved in neuroprotection (49). For example, calcineurin-induced Bad dephosphorylation seems to play an important role in calcium-induced apoptosis within hippocampal neurons, an effect feasible in glutamate-treated cortical neurons as well (50). However, a dominant role for other PKB/Akt substrates, including Bad, in TNF-induced protection of cortical neurons from excitotoxicity seems unlikely because selective inhibitors of NF-B activation operating downstream of PKB/Akt, such as a specific inhibitor of IB phosphorylation (31, 32) or inhibition of proteasomal IB degradation (30), prevented protection from glutamate-induced cell death.
TNF is up-regulated in a number of neurodegenerative disorders including Alzheimer's disease, stroke, and multiple sclerosis. However, its function remained to be defined because both neurotoxic and neuroprotective effects during disease pathogenesis have been described previously (50-55). The TNF transgenic mouse line described here clearly shows that restricted neuronal TNF expression is a priori not harmful. To the contrary, the response phenotype of primary neurons isolated from these mice suggests that TNF expression conditions the neurons toward resistance to excitotoxic insults. Support for the in vivo relevance of a TNFR2-mediated neuroprotective pathway was previously obtained by us in a model of retinal ischemia (10). We now have preliminary evidence in a model of cerebral ischemia that NR2B/TNF transgenic mice display reduced lesions as compared with wild type animals.3
In conclusion, here our data present that TNF mediates neuroprotection via TNFR2 and that NMDAR co-signaling provides new mechanistic insights and sheds new light into the role of TNF in the brain. We propose that TNF functions as an important regulatory cytokine in the central nervous system with differential signaling through the two distinct TNFRs determining its contribution to degenerative and regenerative processes. Thus, depending on the type and state of disease, environmental/external factors, cellular composition of the affected tissue, and intracellular availability of TNFR signaling components, TNF may function to aggravate or ameliorate disease. A complete understanding of the differential signals that are initiated by TNFR1 and TNFR2 in neurons and of their cross-talk with other pathways that are important for neuronal viability is essential in order to derive therapeutically effective strategies. From our study, we conclude that selective triggering of TNFR2 could be a novel therapeutic strategy in neurodegenerative diseases because it actively induces an anti-apoptotic state in neurons by cooperative action with neurotransmitter receptor signals. This mechanism elucidated here can now be exploited for therapeutic gain in combination with or independent of drugs interfering with apoptotic signal pathways.
|
FOOTNOTES |
|---|
A recipient of a Landesgraduiertenförderung stipend from the state of Baden-Wuerttemberg.
Supported by grants from Volkswagen Foundation, EU Commission, Hertie Foundation, and the Deutsche Forschungsgemeinschaft (SFB 495).
** To whom correspondence should be addressed. Tel.: 49-711-685-6986; Fax: 49-711-685-7484; E-mail: klaus.pfizenmaier{at}izi.unistuttgart.de.
1 The abbreviations used are: TNF, tumor necrosis factor; TNFR, TNF receptor; PI3K, phosphatidylinositol 3-kinase; PKB, protein kinase B; UTR, untranslated region; NF-B, nuclear factor-B; NMDA, N-methyl-D-aspartate; NMDAR, NMDA receptor; NR2B, NMDAR subunit 2B; Z, benzyloxycarbonyl; fmk, fluoromethyl ketone; PBS, phosphate-buffered saline; DTT, dithiothreitol; EMSA, electrophoretic mobility shift assay; IKK, IB kinase; ANOVA, analysis of variance.
2 L. Marchetti, M. Klein, K. Schlett, K. Pfizenmaier, and U. L. M. Eisel, unpublished data.
3 U. L. M. Eisel and H. Petit, unpublished data.
|
ACKNOWLEDGMENTS |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTAL PROCEDURESRESULTSDISCUSSIONREFERENCES |
|---|
) and 500 (
) µM glutamate induces neuronal death in a time-dependent manner. Cell viability was assessed at the indicated times after a 1-h glutamate stimulus. B, dose-dependent neuronal cell death in the absence (
and 
, NR2B/TNF; 
and 
, TNFR1-/- (R1-/-); 
and 
, TNFR2-/- (R2-/-)). B, TNF dose-dependent protection of primary neuronal cultures after exposure to excitotoxic concentrations of glutamate (1 h, 250 µM) in the absence (open symbols) or presence (filled symbols) of the LY294002 (
, NR2B/TNF plus geldanamycin. Cell death in A, B, C, and D was assessed by MTT assay (n = 3) as described under "Experimental Procedures."
