Address correspondence and reprint requests to Professor Dr Josef Krieglstein, Philipps-Universität Marburg, Institut für Pharmakologie und Toxikologie, Ketzerbach 63, D-35032 Marburg, Germany. E-mail: email@example.com
Preconditioning by a sublethal stimulus induces tolerance to a subsequent, otherwise lethal insult and it has been suggested that reactive oxygen species (ROS) are involved in this phenomenon. In the present study, we determined whether preconditioning activates the transcription factor nuclear factor-κB (NF-κB) and how this activation contributes to preconditioning-induced inhibition of neuronal apoptosis. Preconditioning was performed by incubating mixed cultures of neurons and astrocytes from neonatal rat hippocampus with xanthine/xanthine oxidase or FeSO4 for 15 min followed by 24 h of recovery which protected the neurons against subsequent staurosporine-induced (200 nm, 24 h) apoptosis. The cellular ROS content increased during preconditioning, but returned to basal levels after removal of xanthine/xanthine oxidase or FeSO4. We detected a transient activation of NF-κB 4 h after preconditioning as shown by immunocytochemistry, by a decrease in the protein level of IκBα as well as by electrophoretic mobility shift assay. Preconditioning-mediated neuroprotection was abolished by antioxidants, inhibitors of NF-κB activation and cycloheximide suggesting the involvement of ROS, an activation of NF-κB and de novo protein synthesis in preconditioning-mediated rescue pathways. Furthermore, preconditioning increased the protein level of Mn-superoxide dismutase which could be blocked by antioxidants, cycloheximide and κB decoy DNA. Our data suggest that inhibition of staurosporine-induced neuronal apoptosis by preconditioning with xanthine/xanthine oxidase or FeSO4 involves an activation of NF-κB and an increase in the protein level of Mn-superoxide dismutase.
The transcription factor nuclear factor-κB (NF-κB) has been initially identified as a lymphoid-specific protein that binds to the κ-light chain gene enhancer (Sen and Baltimore 1986) and has meanwhile been found in all cells including neurons and astrocytes (O'Neill and Kaltschmidt 1997). NF-κB is a heterodimer protein predominantly composed of a 50-kDa and a 65-kDa subunit (Sen and Baltimore 1986). The transcription factor resides in the cytosol in its inactive state complexed with the inhibitory protein IκBα (O'Neill and Kaltschmidt 1997). Upon activation, IκBα undergoes phosphorylation and degradation, thus releasing the p50–p65 heterodimer for translocation to the nucleus where it triggers the transcription of various genes (Brown et al. 1993; Sun et al. 1993). NF-κB belongs to the Rel family of related transcription factors that control a broad range of physiological and pathophysiological processes (O'Neill and Kaltschmidt 1997). The role of NF-κB in the regulation of apoptosis is controversially discussed. Activation of NF-κB has been demonstrated to mediate excitotoxin-induced apoptosis in rat striatum (Qin et al. 1998; Grilli and Memo 1999a) and to promote neurotoxicity of amyloid-β (Aβ) (Bales et al. 1998). The infarct volume after transient focal ischaemia was found to be reduced in p50 knock-out mice suggesting that an activation of NF-κB contributes to ischaemic cell death (Schneider et al. 1999). On the other hand, a constitutive activation of NF-κB has been shown to be required for the maintenance of neuronal growth and integrity (Kaltschmidt et al. 1994). In addition, inhibition of apoptosis by nerve growth factor (NGF) has been shown to involve an activation of NF-κB with subsequent inhibition of caspase-8 and an increase in bcl-2 protein level (de Moissac et al. 1998; Maggirwar et al. 1998; Wang et al. 1998). An inhibition of NF-κB-induced gene transcription induced NGF-resistant apoptosis in PC12 cells (Taglialatela et al. 1997) and potentiated Aβ peptide-mediated apoptotic damage in neurons (Kaltschmidt et al. 1999b). Similarly, a lack of the p50 subunit increased the vulnerability of hippocampal neurons to excitotoxic injury (Yu et al. 1999).
We recently demonstrated that ROS were involved in the neuroprotective effect mediated by preconditioning with X/XO or FeSO4 (Ravati et al. 2000). In the present study, we tried to find out whether an activation of the transcriptional factor NF-κB and an increase in the NF-κB regulated SOD-2 gene contribute to preconditioning-induced inhibition of staurosporine-induced apoptosis.
Materials and methods
Neurobasal medium, Leibovitz's L15 medium, antibiotics, B27 supplement, glutamine and papain were purchased from Life Technologies, Eggenstein, Germany. Staurosporine, Fe2SO4, xanthine, xanthine oxidase, Nonidet NP-40, Hoechst 33258, Tween 20, poly-l-lysine, trypsin inhibitor, cycloheximide, vitamin E, 2-hydroxyoestradiol, pyrollidine dithiocarbamate (PDTC), interleukin-1β (IL-1β), tumour necrosis factor-α (TNF-α), phenylmethylsulphonyl fluoride (PMSF), leupeptin, sodium dodecylsulphate (SDS), bovine serum albumin (BSA), the lactate dehydrogenase (LDH) assay kit and dimethylsulphoxide (DMSO) were obtained from Sigma-Aldrich, Deisenhofen, Germany. The bicincolinic acid protein assay kit and the Super Signal-chemiluminescence system were from Pierce, Rockford, USA. Dihydrorhodamine 123 was obtained from Molecular Probes, Eugene, USA. Polyclonal antibodies against Cu,Zn-superoxide dismutase (SOD-1) and Mn-superoxide dismutase (SOD-2) and polyclonal anti-IκBα antibodies were purchased from RDI, Flanders, USA, and Santa Cruz, Heidelberg, Germany, respectively. Monoclonal antibodies against glial fibrillary acidic protein (GFAP) and the p65 subunit of NF-κB were from Roche Diagnostics, Mannheim, Germany. Biotinylated anti-mouse IgG and fluorescein streptavidin were obtained from Vector Laboratories, Grünberg, Germany. Polyclonal antineurofilament antibodies, horseradish peroxidase-conjugated anti-rabbit IgG, double-strand oligonucleotide containing the NF-κB specific binding consensus sequence (single-stranded sequence: 5′-AGTTGAGGGGACTTTCCCAGGC-3′) and the single-base mutated double-strand oligonucleotide (single-stranded sequence: 5′-AGTTGAGCGACTTTCCCAGGC-3′) that does not bind to NF-κB were from Promega, Mannheim, Germany. [γ-32P]ATP was purchased from Amersham Pharmacia Biotech, Freiburg, Germany.
Mixed hippocampal cultures containing both neurons and astrocytes were prepared from neonatal Fischer 344 rats (P1–P2) as described by Sengpiel et al. (1998). Briefly, the isolated hippocampi were dissected, incubated at 37°C for 20 min in Leibovitz's L15 medium supplemented with 1 mg/mL papain and 0.2 mg/mL BSA, and gently triturated. Thereafter, the cell suspension was layered onto culture medium containing 1% trypsin inhibitor and 1% BSA, centrifuged at 200 g for 10 min and the pellet was resuspended and seeded at a density of 2 × 104 cells/cm2 into poly-l-lysine-coated Petri dishes. For immunocytochemistry, cells were plated onto poly-l-lysine-coated glass cover slips that were placed into Petri dishes. Cells were cultured in a humidified atmosphere of 5% CO2 and 95% air at 37°C in neurobasal medium supplemented with 0.5 mm glutamine, B27 supplement, 20 U/mL penicillin and 20 µg/mL streptomycin for 10 days. Medium was exchanged after 3 days in culture. The mixed cultures contained 60% neurons and 40% astrocytes as evaluated by immunocytochemical staining with anti-neurofilament antibodies (1 : 2000) and anti-GFAP antibodies (1 : 1000) in three independent series of experiments. This culture system is more physiological than separated cultures.
After 10 days in culture, preconditioning was performed by incubating hippocampal cells with X/XO (10 µm/0.5 mU per mL) or FeSO4 (100 µm) for 15 min followed by 24 h of recovery if not otherwise mentioned. Thereafter, apoptosis was induced by a treatment with 200 nm staurosporine for 24 h. Staurosporine is widely used as a model to induce apoptosis in neuronal cells (Prehn et al. 1997; Krohn et al. 1999; Ravati et al. 1999, 2000; Ahlemeyer et al. 2001). Double-stranded oligonucleotide with a specific NF-κB-binding consensus sequence (κB decoy DNA) and the single-base mutated double-stranded oligonucleotide (mutant DNA) both containing phosphorothioate bonds to reduce degradation in culture media at 37°C (Campbell et al. 1990) were added directly to the culture medium and were present only during recovery. Cycloheximide, PDTC and lactacystine were present during preconditioning as well as during recovery. Vitamin E and 2-hydroxyoestradiol were present only during preconditioning.
Cell viability was determined by the release of the cytosolic enzyme LDH into the culture medium. After exposure to staurosporine for 24 h, LDH activities in the culture medium of 5 culture flasks per group were measured by a commercial photometric LDH assay kit. Values are expressed as LDH activity in the culture medium. One unit LDH is defined as the amount of enzyme which catalyzes the formation of 1 µmol/L of NAD per min.
Determination of the percentage of apoptotic neurons and astrocytes
To ensure that the release of LDH corresponded to neuronal death and not to glial death, additional experiments were performed. Mixed cultures were incubated with Hoechst 33258 to visualize nuclear morphology which allowed us to detect apoptotic and viable cells followed by GFAP immunostaining to discriminate between neurons and astrocytes. Briefly, after fixing the cells with methanol, they were incubated with the DNA fluorochrome Hoechst 33258 (10 µg/mL) for 15 min, stored at − 20°C for 20 min and incubated for 30 min with phosphate-buffered saline (PBS) containing 1% BSA. Thereafter, monoclonal anti-GFAP-antibodies (1 : 1000) were added overnight at 4°C. After washing with PBS, cells were incubated with biotinylated anti-mouse IgG (1 : 100) for 1 h at room temperature. Antibody–antigen complexes were amplified by an incubation with fluorescein streptavidin (1 : 500) for 30 min at 37°C in the dark. Thereafter, the cells were washed and observed under a confocal laser scanning microscope (LSM 510, Zeiss, Germany). Negative controls were performed by omitting the primary antibody. The number of neurons and astrocytes with apoptotic features and the total number of neurons and astrocytes were counted in four randomized subfields of 4 different Petri dishes containing approximately 80 cells per subfield. Values are expressed as percent ratio of neurons and astrocytes with apoptotic features vs. the total number of neurons and astrocytes, respectively.
Measurement of ROS
The ROS were measured using the lipophilic nonfluorescent dye dihydrorhodamine 123 which accumulated in the mitochondria and was oxidized by ROS to the positively charged fluorescent derivative rhodamine 123. Determination of the ROS level in single cells was performed as previously described (Ravati et al. 1999). Briefly, cells were stained with 5 µm dihydrorhodamine 123 for 15 min. Digital video imaging of rhodamine 123 fluorescence was conducted using a fluorescence microscope (Axiovert 100, Zeiss, Germany) with attenuated ultraviolet illumination from a 75-W xenon lamp. Fluorescence intensity was measured at 490 nm excitation wavelength and 510 nm emission wavelength. An electronic shutter which opened during image acquisition only, minimized photobleaching and phototoxicity. Images were taken by a CCD camera (C 2400–87; Hamamatsu, Germany) and were digitalized as 256 × 256 pixels. Before measurement of fluorescent values, a background picture was taken that was later subtracted from the images. Data were analysed using Argus 50 software (Hamamatsu, Germany). Fluorescence intensities were given as arbitrary units (Fl.U) per cell.
After nuclear staining with Hoechst 33258, cultured hippocampal cells were fixed and permeabilized with methanol at − 20°C for 20 min and then incubated for 30 min with PBS containing 1% BSA. Thereafter, primary monoclonal antibodies against the p65 subunit of NF-κB (1 : 100) were added overnight at 4°C. After washing with PBS, cells were incubated with biotinylated anti-mouse IgG (1 : 100) for 1 h at room temperature. Antibody–antigen complexes were amplified by an incubation with fluorescein streptavidin (1 : 500) for 30 min at 37°C in the dark. Thereafter, the cells were washed and observed under a confocal laser scanning microscope (LSM 510; Zeiss, Germany). Negative controls were performed by omitting the primary antibody.
Cell cultures were washed once with PBS, collected at 4°C and then lysed on ice in a buffer containing 0.25 m sucrose, 20 mm Tris pH 7.5, 20 mm KCl, 1.5 mm MgCl2, 1 mm EGTA, 1 mm dithiothreitol (DTT), 1 mm PMSF, 0.2 µg/mL leupeptin and 10 µg/mL trypsin inhibitor. The cells were gently aspirated five times through a 25 G needle and centrifuged at 1500 × g for 15 min at 4°C. The supernatants were collected on ice and were measured for protein content using a protein assay kit. In each sample, 30 µg protein was incubated for 5 min in sample buffer at 95°C, loaded and separated on 15% (for detecting IκBα protein) or 20% (for detecting SOD proteins) SDS–polyacrylamide gels and transferred to nitrocellulose membranes. Blots were blocked for 3 h at 20°C in PBS containing 2% BSA, 5% non fat dry milk and 0.02% Tween 20 (blocking buffer). Incubations with primary polyclonal antibodies against IκBα (1 : 1000), SOD-1 (1 : 7000) and SOD-2 (1 : 7000) were performed overnight at 4°C. The next day, the blots were incubated for 1 h with horseradish peroxidase-conjugated anti-rabbit IgG (1 : 2000) at 20°C in blocking buffer. For detection of the signals, Super Signal-chemiluminescence system was used.
Nuclear extraction and electrophoretic mobility shift assay
Nuclear extracts were prepared as described by Schreiber et al. (1989). Briefly, cells were washed with PBS and resuspended at 4°C in 400 µL buffer containing 10 mm HEPES, pH 7.9, 0.1 mm EDTA, 0.1 mm EGTA, 10 mm KCl, 1 mm DTT and 0.5 mm PMSF. Cells were allowed to swell on ice for 15 min. After the addition of 25 µL of a solution of 10% Nonidet NP-40 and mixing, the samples were centrifuged at 20 000 g for 30 s and resuspended in 50 µL buffer containing 20 mm HEPES, pH 7.8, 25% (v/v) glycerol, 400 mm NaCl, 1 mm EGTA, 1 mm EDTA, 1 mm DTT and 1 mm PMSF. After vigorous shaking at 4°C for 20 min, the nuclear extract was centrifuged at 20 000 g for 4 min and the supernatants were frozen in aliquots at − 80°C after analysing the protein concentration.
Electrophoretic mobility shift assay was performed using double-stranded 22 base pair oligonucleotides (25 ng) containing the NF-κB consensus sequence. Double-stranded oligo-nucleotides were end-labelled with [γ-32P]ATP (3000 Ci/mmol) for 1 h at 37°C by the T4 polynucleotide kinase. For the binding reaction, 5 µL nuclear extract (10 µg protein) was incubated with 2 µL buffer consisting of 20 mm HEPES, pH 7.9, 20% (v/v) glycerol, 100 mm KCl, 0.5 mm EDTA, 2 mm DTT, 0.25% Nonidet NP-40 and 0.1 mm PMSF, and 2 µL 0.1% poly(dI-C), as well as 2 µL 1% BSA, 4 µL Ficoll-buffer (20% Ficoll 400, 20 mm HEPES, 100 mm KCl, 0.1 mm PMSF) and the [γ-32P]-labelled probe (10000 cpm) for 30 min at room temperature. DNA–protein complexes were loaded on a pre-electrophoresed (10 min, 200 V) 4% nondenaturing polyacrylamide gel in 11.2 mm Tris, 11.2 mm boric acid and 0.25 mm EDTA for 1.5 h at 200 V at 4°C followed by autoradiography.
Neuroprotection by preconditioning is reversed by antioxidants, inhibitors of NF-κB activation and cycloheximide
Preconditioning was performed in hippocampal cultures by an incubation with either X/XO (10 µm/0.5 mU per mL) or FeSO4 (100 µm) for 15 min followed by 24 h of recovery which protected the neurons against a subsequent staurosporine-induced (200 nm, 24 h) apoptotic damage as determined by the release of LDH into the culture medium (Fig. 1). In addition, we assessed neuronal as well as astrocytic damage in cultures treated with vehicle, staurosporine with and without preconditioning (Table 1). The experiments were performed by incubating mixed cultures with Hoechst 33258 to visualize nuclear morphology which allowed us to detect apoptotic and viable cells, followed by GFAP-immunostaining to discriminate between neurons and astrocytes. Staurosporine induced apoptosis only in a few astrocytes and neuronal damage (Table 1) corresponds with cellular viability as measured by the release of LDH (Fig. 1) suggesting that LDH release reflects predominantly neuronal death. In previous studies, we demonstrated that the ROS level increased during the 15 min of preconditioning with X/XO or FeSO4 (Ravati et al. 2000). Monitoring the ROS-formation kinetics during the 15-min period of preconditioning revealed that X/XO-induced increase in ROS could be immediately reduced by 2-hydroxyoestradiol and vitamin E. However, significant effects of both antioxidants on FeSO4-induced increase in ROS could only be measured at 15 min of incubation, but not during the first 10 min (Ravati et al. 2000). Because neuroprotection by preconditioning with X/XO was abolished by the antioxidants 2-hydroxyoestradiol (1 µm) and vitamin E (10 µm) as well as by PDTC (1 µm), the proteasome inhibitor lactacystine (0.1 µm) and the protein synthesis inhibitor cycloheximide (1 µm) we suggest that ROS, an activation of NF-κB and de novo protein synthesis were involved in the protective mechanism (Fig. 1a). Neuroprotection by preconditioning with FeSO4 could be blocked by PDTC, lactacystine and cycloheximide, but not by the antioxidants 2-hydroxyoestradiol or vitamin E (Fig. 1b). The antioxidants, lactacystine and cycloheximide alone had no effect on neuronal viability under control conditions (data not shown).
Table 1. Preconditioning with X/XO or FeSO4 prevents staurosporine-induced neuronal damage
Apoptotic neurons (%)
Apoptotic astrocytes (%)
Primary hippocampal cultures were preconditioned (pre) for 15 min with 10 µm X and 0.5 mU/mL XO (X/XO) or 100 µm FeSO4 (FeSO4) in the presence and absence of vitamin E (10 µm). After a recovery of 24 h, cells were treated with staurosporine (200 nm, Stau). Controls received vehicle only. The percentage of apoptotic neurons and astrocytes were determined 24 h after starting the exposure to staurosporine by nuclear staining with Hoechst 33258. GFAP immunostaining allowed the discrimination between neurons and astrocytes. Values are given as means ± SD of four experiments. Different from cultures treated with staurosporine alone: *p < 0.001 as evaluated by one-way analysis of variance with subsequent Scheffé test.
Preconditioning with X/XO or FeSO4 reduced staurosporine-induced increase in the cellular ROS level
Measuring the ROS level within single neurons, we found a three-fold increase 4 h after starting the exposure to staurosporine (Table 2) which is consistent with previous findings in rat hippocampal neurons (Prehn et al. 1997; Krohn et al. 1999; Ahlemeyer et al. 2001). Staurosporine-induced increase in the ROS level was reduced when the neurons were preconditioned with X/XO or FeSO4(Table 2). Suppression of staurosporine-induced increase in the ROS level by preconditioning with X/XO, but not with FeSO4, was abolished by the antioxidants vitamin E and 2-hydroxyoestradiol being present only during preconditioning (Table 2). The lack of effect of both antioxidants to block the antioxidative effect of preconditioning with FeSO4 can be explained by an initial insensitivity of FeSO4 against radical scavengers as previously described (Ravati et al. 2000).
Table 2. Preconditioning with X/XO or FeSO4 prevents staurosporine-induced increase in the ROS level
Primary hippocampal cultures were preconditioned (pre) for 15 min with 10 µm X and 0.5 mU/mL XO (X/XO) or 100 µm FeSO4 (FeSO4) in the presence and absence of vitamin E (10 µm) or 2-hydroxyoestradiol (1 µm). After a recovery of 24 h, cells were treated with staurosporine (200 nm, Stau) or vehicle (controls). The cellular ROS level was measured 4 h after starting the exposure to staurosporine using dihydrorhodamine 123. Fluorescence intensities of rhodamine 123 were determined in neuronal cells which were identified by morphological criteria. Values are given as means ± SD of 6–8 experiments. Different from cultures treated with staurosporine alone: *p < 0.01. Different from cultures preconditioned with X/XO and treated with staurosporine; †p < 0.01 as evaluated by one-way analysis of variance with subsequent Scheffé test.
Preconditioning-induced neuroprotection involves the activation of NF-κB
Because lactacystine abolished preconditioning-mediated neuroprotection (Fig. 1), we performed additional experiments to get further evidence that activation of NF-κB is involved in the mechanism of neuroprotection. We used κB decoy DNA as a specific inhibitor of NF-κB activation. A single-base mutant oligonucleotide which does not bind NF-κB (mutant DNA) was used as negative control. As shown in Fig. 2, κB decoy DNA abolished neuroprotection by preconditioning with X/XO and FeSO4, whereas mutant DNA did not. The κB decoy DNA and mutant DNA alone had no effect on neuronal viability under control conditions (Fig. 2).
Next, we evaluated the activation of NF-κB after preconditioning by three different approaches: (i) by immunocytochemistry using confocal laser scanning microscopy to detect the translocation of the p65 subunit of NF-κB from the cytosol into the nucleus; (ii) by western blot analysis of the inhibitory IκBα protein level which decreases upon activation of NF-κB; and (iii) by electrophoretic mobility shift assay. In controls, p65 immunoreactivity was found in the cytosol only (Fig. 3a). Preconditioning with X/XO (Fig. 3b) or FeSO4(Fig. 3c) resulted in a translocation of p65 from the cytosol into the nucleus in most of the cells. Translocation of p65 started at 1 h, reached its maximum at 4 h and was still increased at 24 h after preconditioning (data not shown). X/XO-mediated translocation of p65 from the cytosol to the nucleus could be blocked by vitamin E indicating the requirement of ROS for the observed effect (Fig. 3d). The FeSO4-mediated translocation of p65 was not inhibited by vitamin E. Activation of NF-κB as determined by translocation of its p65 subunit was found in neurons as well as in astrocytes as identified by concomitant neurofilament- or GFAP-staining (data not shown).
Upon activation of NF-κB, the complexed inhibitory protein IκBα is degraded by proteasomes and releases the p50-p65 heterodimer for translocation to the nucleus. Therefore, western blot analysis of the IκBα protein level can be used as a method to evaluate activation of NF-κB. As shown in Fig. 4, we detected a marked decrease in the IκBα protein level 4 h after preconditioning with X/XO or FeSO4. Activators of NF-κB, TNF-α (10 ng/mL) and IL-1β (10 ng/mL) were used as positive controls to demonstrate that a decrease in IκBα protein level followed the activation of NF-κB. The decrease in IκBα protein level by preconditioning with X/XO was prevented by vitamin E, PDTC and lactacystine (Fig. 4a), whereas the preconditioning with FeSO4 was abolished by vitamin E and lactacystine, but not by PDTC (Fig. 4b). The former results were confirmed by electrophoretic mobility shift assay showing an enhanced NF-κB activity by preconditioning with X/XO as well as by treatment with TNF-α and IL-1β(Fig. 5). The increase in NF-κB activity by preconditioning with X/XO was blocked by vitamin E, 2-hydroxyoestradiol, PDTC and lactacystine (Fig. 5).
Preconditioning with X/XO or FeSO4 increased the protein level of SOD-2 which could be blocked by 2-hydroxyoestradiol, cycloheximide and κB decoy DNA
Preconditioning with X/XO or FeSO4 increased the protein level of the NF-κB-regulated antioxidative enzyme SOD-2 starting from 4 h and up to 24 h after preconditioning, whereas the protein level of SOD-1 remained unchanged (Fig. 6). The increase in SOD-2 protein level by preconditioning with X/XO or FeSO4 was abolished by 2-hydroxyoestradiol, cycloheximide and κB decoy DNA (Fig. 7) suggesting the requirement of ROS, an activation of NF-κB and a subsequent protein synthesis of SOD-2 for the observed effect.
Although many studies have shown the protective effect of preconditioning, the mechanisms which induce and mediate tolerance are still unknown. Despite various processes and intracellular cascades initiated by the onset of preconditioning, the existence of a common downstream regulator collecting the information of upstream signalling events has been proposed (Grilli and Memo 1999b). One candidate would be the transcription factor NF-κB which is activated in the central nervous system by many agents such as TNF-α, IL-1β, phorbolesters, hydrogen peroxide, ultraviolet light, bacterial and viral proteins, glutamate, opioids, NGF or Aβ (Schutze et al. 1992; Baldwin 1996; O'Neill and Kaltschmidt 1997; Liu et al. 2000). In this study, we provide evidence that preconditioning with X/XO and FeSO4 protected cultured neurons against subsequent staurosporine-induced damage by an increase in ROS followed by an activation of NF-κB and a subsequent increase in NF-κB regulated gene expression. This suggestion is based on our findings that antioxidants inhibited preconditioning-mediated neuroprotection as well as the activation of NF-κB and that preconditioning-induced neuroprotection was abolished by inhibitors of NF-κB activation and cycloheximide. Four different approaches were used to inhibit NF-κB activation: (i) vitamin E and 2-hydroxyoestradiol scavenge ROS, thereby inhibiting the redox signalling pathway of NF-κB activation; (ii) PDTC is an antioxidant, but it has also been described to inhibit NF-κB activation due to oxidation of glutathione (Brennan and O'Neill 1996); (iii) lactacystine is a proteasome inhibitor and blocks the cleavage of the inhibitory subunit IκBα from the p65-p50 heterodimer; and (iv) the κB decoy DNA reduces the NF-κB-binding to the appropriate gene promoters by competing with nuclear DNA. Interestingly, neuroprotection by preconditioning with FeSO4 could be abolished by PDTC, but not by the radical scavengers vitamin E or 2-hydroxyoestradiol supporting the above mentioned hypothesis that PDTC-mediated NF-κB inhibition involves other mechanisms beside its antioxidant property.
Several groups have reported that NF-κB activation inhibited apoptosis (de Moissac et al. 1998; Maggirwar et al. 1998; Wang et al. 1999), whereas in other studies it promoted cell death (Bales et al. 1998; Qin et al. 1998; Grilli and Memo 1999b; Schneider et al. 1999). The explanation for the conflicting results concerning an anti-apoptotic versus a pro-apoptotic role for NF-κB activation is still not clear and has been described as ‘janus faces’ of NF-κB (Lipton 1997). Whether activation of NF-κB results in enhanced cell survival or apoptosis seems to depend on the cell type, the nature and intensity of the activating stimulus and the concomitant activation of other transcription factors (Grilli et al. 1995; Baichwal and Baeuerle 1997; Lin et al. 1999). Alternatively, the NF-κB dimer species, which is activated and translocated to the nucleus, has been suggested to play a regulatory role, e.g. that an increased p65 translocation enhanced neuronal survival, while increased p50 translocation could have a repressor role (Gu et al. 2000). Consistent with these findings, we found a translocation of p65 from the cytosol to the nucleus 4 h after preconditioning (Fig. 3) which resulted in enhanced cell survival. In addition, the kinetic of the activation has been proposed to influence the cellular outcome, e.g. that a transient activation of NF-κB is protective, whereas a long-lasting activation entails deleterious events (Schneider et al. 1999).
An activation of NF-κB has been shown to involve ROS-dependent as well as ROS-independent pathways (Okamoto et al. 1997). In our study, preconditioning was performed using X/XO which leads to the formation of superoxide radicals and hydrogen peroxide (Satoh et al. 1998) and FeSO4 which generates hydroxyl radicals via the Fenton reaction (Müller and Krieglstein 1995). In previous studies, we have already shown that the cellular ROS level increased during preconditioning with X/XO or FeSO4 and that antioxidants being present during preconditioning with X/XO abolished the increase in ROS (Ravati et al. 2000). Significant effects of the antioxidants in FeSO4-induced increase in ROS level could only be measured at 15 min of incubation, but not during the first 10 min suggesting that an immediate and continuous radical scavenging is necessary to block ROS-mediated preconditioning (Ravati et al. 2000). Consistently, we found that preconditioning-induced neuroprotection as well as the activation of NF-κB with X/XO, but not with FeSO4 were abolished by antioxidants. Thus, it seems likely that ROS mediated the activation of NF-κB as well as the protective effect of preconditioning. Different kinds of ROS may possess different NF-κB activating properties. Superoxide radicals have been reported to be more potent activators of NF-κB than hydroxyl radicals (Wang et al. 1999), but hydroxyl radicals are suggested to mediate hypoxia-induced preconditioning in the brain (Rauca et al. 2000) and in the heart (Das et al. 1999). In our previous study, we used dihydrorhodamine 123 to detect ROS during preconditioning with X/XO or FeSO4 (Ravati et al. 2000). Because dihydrorhodamine 123 detects generalized ‘oxidative stress’ rather than the production of a particular oxidizing species, it remains open which kind of ROS is responsible for activation of NF-κB by X/XO or FeSO4. In addition, not only the kind, but also the amount and duration of the increase in the ROS level are important for the activation of NF-κB and for cell survival. When cultured neurons were treated with low concentrations of hydrogen peroxide a marked increase in NF-κB activation was found, whereas high concentrations of hydrogen peroxide inhibited the translocation of the p65 subunit into the nucleus (Kaltschmidt et al. 1997).
As we suggest that preconditioning-induced neuroprotection was due to an activation of NF-κB followed by enhanced transcriptional activity of NF-κB-regulated genes, we next evaluated the protein level of one of these genes, the antioxidative enzyme SOD-2 (White et al. 2000). We did so because preconditioning not only protected the neurons, but also prevented staurosporine-induced increase in the cellular ROS level (Table 2). ROS have been shown to mediate staurosporine-induced neuronal death and radical scavengers (Prehn et al. 1997; Krohn et al. 1999; Ravati et al. 2000), an increase in the protein levels of antioxidant enzymes (Prehn et al. 1997; Ahlemeyer et al. 2001) as well as SOD mimetics (Patel 1999) have been identified as potent neuroprotective agents. Preconditioning-induced neuroprotection as well as the increase in the SOD-2 protein level could be blocked by 2-hydroxyoestradiol, cycloheximide and κB decoy DNA. This suggests that de novo synthesis and an activation of NF-κB are required for the antioxidative and neuroprotective effect of preconditioning and that an increase in the SOD-2 protein level is involved in this process. Because a pronounced induction of SOD-2 gene is suggested to result from a complex interaction between NF-κB and other transcription factors (Maehara et al. 1999; Xu et al. 1999), we cannot exclude that activation of other transcription factors in addition to NF-κB are involved in the increase in SOD-2 protein expression. Because NF-κB regulates the expression of various genes with either pro-apoptotic function such as p53, amyloid precursor protein, bax or interleukin-converting enzyme or anti-apoptotic properties such as bcl-2 or bcl-xL (Grilli and Memo 1999b), it is possible that other NF-κB regulated genes beside SOD-2 are involved in the antioxidative and neuroprotective effect of preconditioning. The exact identification of the genes and proteins which are transcriptionally regulated by NF-κB will provide further insight into its functional relevance.
Mixed hippocampal cultures, which are used in this study, consist of neurons and astrocytes. We cannot distinguish whether preconditioning-mediated rescue pathways were activated in neurons, astrocytes or in both cell types. We observed an activation of NF-κB as determined by translocation of the p65 in neurons as well as in astrocytes. In contrast to our results, exposure to Aβ has been reported to decrease constitutive NF-κB activity in cortical neurons and to increase NF-κB activity in astrocytes (Bales et al. 1998). Constitutive neuronal NF-κB activity depended on neuron–glial interaction as hippocampal neurons were devoid of activated NF-κB in co-culture with astrocytes, whereas separating the neurons from glial components resulted in an increase in neuronal NF-κB activity (Kaltschmidt and Kaltschmidt 2000). In addition, activation of NF-κB has been shown to mediate retrograde signalling from the axons, dendrites or synapses to the cell body (Gisiger 1998; Kaltschmidt et al. 1999a).
To summarize, we have shown that preconditioning with X/XO or FeSO4 protected cultured neurons against a subsequent otherwise deleterious exposure to staurosporine. Preconditioning was found to activate the transcription factor NF-κB and to increase the protein level of the antioxidant enzyme SOD-2, parameters which are suggested to contribute to the mechanism of neuroprotection.
The authors would like to thank Sandra Engel for excellent technical assistance and Emma Esser for careful reading of our manuscript. The experiments were supported by a grant from the Deutsche Forschungsgemeinschaft (Kr 354/16–3).