Tumor Necrosis Factor-α-induced Cell Killing and Activation of Transcription Factor NF-κB Are Uncoupled in L929 Cells*

The induction of transcription factor NF-κB has been shown to counteract tumor necrosis factor (TNF)-α-induced cell death in various cell types. In this study, we investigated the role of NF-κB for TNF-α-triggered cell death in the widely used mouse cell line L929 by various approaches. Inhibition of the mitochondrial permeability transition by bongkrekic acid impaired TNF-α-induced cell death without affecting the activity of NF-κB. The reduction of NF-κB-mediated gene expression by the synthetic steroid dexamethasone was associated with a decrease in TNF-α-mediated cell killing, suggesting that NF-κB does not protect L929 cells from TNF-α-induced cell death. This concept was reinforced by experiments employing L929 cell lines stably overexpressing a transdominant negative form of IκB-α. These cell lines were unable to activate NF-κB and to inducibly express the IL-6 gene, but they showed the same susceptibility toward TNF-α-mediated cell death as L929 wild-type cells.

The cytokine TNF-␣ 1 plays a pivotal role in a variety of inflammatory, immunological, and pathological processes. Two well-studied receptors for TNF-␣ are TNF-R55 and TNF-R75. These receptors belong to the rapidly growing family of TNF receptors, including the recently discovered LT-␤ receptor, DR3/WSL-1/TRAMP, RANK, and TRAIL (for reviews, see Refs. 1 and 2). Binding of the extracellular ligand leads to trimerization of the receptors and association of numerous cytoplasmic proteins to the intracellular receptor domain. This binding is induced by ligation of the TNF-R55 or clustering of the intracellular C-terminal domains, which are sufficient for the induction of cell death, which coincides with the activation of transcription factor NF-B (3).
NF-B is normally retained in its inactive state in the cytoplasm of cells by interaction with an inhibitory IB molecule (for review, see Ref. 4). Triggering of cells with a variety of inflammatory cytokines, including TNF-␣, induces phosphorylation, ubiquitinylation, and degradation of IB (for reviews, see Refs. 5 and 6). This allows the DNA-binding dimer to enter the nucleus, to bind to its cognate DNA, and to induce the tran-scription of its target genes (for a review, see Ref. 7). The target genes encode a variety of proteins involved in immune responses, cell growth, and apoptosis. Among those are the antiapoptotic proteins Mn-superoxide dismutase, the zinc finger protein A20, and c-IAP2 (8). Previous studies showed that the activation of NF-B can counteract TNF-␣-induced cell death (for a review, see Ref. 9). Fibroblasts from mice lacking the transactivating p65 subunit are more sensitive against the cytopathic effect of TNF-␣ than fibroblasts from p65 ϩ/ϩ mice (10). The inhibition of NF-B in various cell types by stable overexpression of a transdominant negative form of IB-␣ renders those cells more susceptible to cell killing by TNF-␣ (11)(12)(13). However, NF-B has also apoptosis-promoting activities. Expression of high levels of the NF-B subunit c-Rel in bone marrow cells leads to apoptosis and autophagocytic cell death (14). Accordingly, the overexpression of the viral and cellular anti-apoptotic proteins E1B 19K and Bcl-2 impairs NF-B-dependent transactivation (15,16).
In this study, we investigated the impact of NF-B activation on TNF-␣-mediated cell killing in murine L929 fibrosarcoma cells. This widely used cell line is of particular interest because TNF-␣-induced cell death occurs without co-apoptotic stimuli, such as cycloheximide or actinomycin D. Using several different strategies, we demonstrated that TNF-␣-triggered cell killing and activation of transcription factor NF-B are uncoupled in L929 cells.

EXPERIMENTAL PROCEDURES
Cell Culture and Stable Transfections-Murine L929sA fibrosarcoma cells were maintained in Dulbecco's modified Eagle's medium containing 10% FCS and 1% (v/v) penicillin/streptomycin. All cells were grown in an incubator at 37°C and 5% CO 2 . The plasmid RSV-IB-␣ S32/36A was obtained by inserting the HindIII fragment from the plasmid CMV-IB-␣ S32/36A (17) into the HindIII site of a eukaryotic RSV expression vector. The NF-B-dependent reporter gene construct p(IL6B) 3 50hu.IL6 was described previously (18). The plasmids were purified on CsCl gradients, and the L929 cells were transfected with CaCl 2 as described (16). Stably transfected cells were selected in 200 g/ml G418 for several weeks.
Western Blot Analysis and Luciferase Assays-For Western blotting, the proteins were separated by SDS-polyacrylamide gel electrophoresis and transferred onto a polyvinylidene difluoride membrane (Millipore, Bedford, MA) using a semidry blot apparatus (Bio-Rad). The detection of IB-␣ proteins was done as described (19). For the determination of luciferase activity, the cells were washed with isotonic buffer and lysed in 100 l of lysis buffer (Tropix, Bedford, MA). The luciferase assays were performed according to the manufacturer's instructions (Promega, Mannheim, Germany) and quantified in a Duo Lumat LB 9507 luminometer (Berthold, Wildbad, Germany).
Determination of Cell Viability and IL-6 Production-The cytotoxic activity of TNF-␣ was determined by the colorimetric 3-(4,5-dimethyl thiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) assay essentially as described (20). L929sA cells were grown at a density of 1 ϫ 10 4 cells per well in 96-well microtiter plates. Following the addition of TNF-␣ (2000 units/ml) and/or the indicated substances, 20 l of a MTT solution (5 mg/ml PBS) was added to all wells at the times given in the figure legends. After another 3-h incubation, supernatants were removed, followed by the addition of 100 l of a 24:1 (v/v) isopropyl alcohol-HCl solution for 15 min. The absorbance of each well was determined with an automated plate reader (Digiscan, Asys Hitech, Eugendorf, Austria) at 550 nm. The IL-6 concentrations were determined with an antihuman IL-6 enzyme-linked immunosorbent assay kit (Boehringer Mannheim) according to the manufacturer's instructions.

RESULTS
BA Reduces TNF-␣-induced Cytotoxicity but Not the Activation of NF-B-To study the linkage between TNF-␣-induced NF-B activation and cell death in L929 cells, we investigated the effect of BA, which prevents the generation of mitochondrial permeability transition pores during the apoptotic process (21). The addition of 50 M BA clearly decreased the cytotoxic effects of various concentrations of TNF-␣ on L929 cells (Fig.  1A). In contrast, the TNF-␣-induced DNA binding activity of NF-B was not influenced by BA at various concentrations (Fig. 1B). The influence of BA on TNF-␣-stimulated NF-B-de-pendent transcription was tested by adding BA, TNF-␣, or a combination of both to a pool of L929 cells stably transfected with a NF-B-dependent luciferase reporter gene. BA was unable to influence the basal or TNF-␣-induced transactivation by NF-B (Fig. 1C). These results suggest that it is possible to reduce cytotoxicity without affecting NF-B activation in L929 cells.
Dexamethasone Impairs TNF-␣-induced NF-B Transactivation and Cytotoxicity-A protective role of NF-B for TNF-␣induced cell killing would imply that the inhibition of this transcription factor should result in enhanced cell killing. Therefore, we investigated the effects of DEX on cell death and NF-B-dependent gene expression induced by TNF-␣. L929 cells stably transfected with a NF-B-dependent reporter gene were treated with different combinations of TNF-␣ and DEX as indicated ( Fig. 2A). The TNF-␣-induced gene expression was significantly impaired by DEX, a known inhibitor of NF-B gene expression (22). Previous reports attributed the inhibitory effect of glucocorticoids to the increased production of the inhibitory IB-␣ molecule, which then would dissociate NF-B from its cognate DNA (23,24). To test whether this proposed mechanism also applies to L929 cells, the influence of DEX on TNF-␣-activated NF-B was tested by EMSAs. As seen in Fig.  2B, DEX had no effect on the DNA binding activity of NF-B. The same extracts were assayed for the relative levels of IB-␣ in Western blot experiments. Short-time treatment of cells with TNF-␣ resulted in the induced degradation of IB-␣, which was resynthesized after 2 h (Fig. 2B). Treatment of cells with DEX had no impact on the relative levels of IB-␣ at any time point, suggesting that DEX interferes in L929 cells with NF-B-mediated transactivation rather than detectably influencing the IB-␣ levels. The effect of DEX on TNF-␣-induced cell death was tested by adding DEX, TNF-␣, or a combination of both to the cells. When cell viability was measured after various incubation times (Fig. 2C), it was found that DEX significantly protected the L929 cells from cell death, indicating that NF-B activation does not protect this cell line from TNF-␣-induced cell death.
Stable Overexpression of IB-␣ Abrogates TNF-␣-induced Transcription of IL-6 without Affecting Cell Killing-In order to inhibit NF-B in a highly specific way, L929 cells were stably transfected with the plasmid RSV-IB-␣ S32A/S36A, an expression vector encoding a transdominant negative human IB-␣ mutant. Because serines 32 and 36 in this mutant were changed to alanines, the inducible phosphorylation and subsequent degradation was prevented, thus conferring an increased half-life to this IB-␣ form (17). Several L929 cell clones were tested for the expression of IB-␣ protein in Western blot experiments (Fig. 3A). Three of the cell clones (2D1, 3A6, and 3B1) were found to constitutively express the transdominant negative form of human IB-␣, which migrates slightly slower than the endogenous murine IB-␣ protein from L929 cells. L929 control cells that were stably transfected with the empty expression vector did not contain this slower migrating IB-␣ band (Fig. 3A). These IB-␣-overexpressing cell clones were further characterized by testing the TNF-␣-induced DNA binding activity of NF-B by EMSAs. The induced DNA binding capacity of all IB-␣-overexpressing clones was strongly impaired when compared with L929 control cells ( Fig. 3B and data not shown). The effect of IB-␣ overexpression on the transcription of the endogenous NF-B target gene IL-6 was tested by measuring the TNF-␣-triggered production of this cytokine. In contrast to the L929 control cells, treatment of the three IB-␣-overexpressing cell clones with TNF-␣ did not result in strongly elevated IL-6 levels (Fig. 4A). In summary, these results demonstrate that the L929 cell clones 2D1, 3A6, and 3B1 are unable to significantly activate the transcriptional activity of NF-B. However, TNF-␣ killed these cell clones with the same efficiency and kinetics as the L929 control cells (Fig.  4B), indicating that NF-B activation neither activates nor represses TNF-␣-mediated cell killing, thus establishing that both events are unrelated and uncoupled in L929 cells.

DISCUSSION
The role of NF-B in the regulation of apoptosis does not yield a homogeneous picture yet. There are numerous examples showing apoptosis-promoting as well as apoptosis-antagonizing effects of NF-B activation (for a review, see Ref. 9). Some inducers of cell death, such as TNF-␣, directly activate NF-B; others may influence B-dependent gene expression by alternative pathways. In this respect, it was shown that an activated form of Caspase-3 can cleave the unphosphorylated IB-␣ protein, which leads to the generation of a constitutive inhibitor of B-dependent gene expression (25). The events leading to TNF-␣-triggered cell death and NF-B activation are initiated by the trimerization of the TNF receptor. This leads to the induced association of the intracellular TRADD protein and subsequent recruitment of further proteins, including TRAF-2 and MORT1/FADD. The molecular events leading to NF-B activation and cell killing diverge relatively early in this signal cascade. The expression of a transdominant negative form of FADD inhibits TNF-R55-mediated cell death but not B-dependent transcription (26). TRAF-2-deficient mice display a functional NF-B activation and an increased sensitivity to TNF-␣, suggesting an additional NF-B-independent pathway of cell protection that is mediated by TRAF-2 (27).
In this study, we show that the TNF-␣-triggered activation of NF-B and the induction of cell death are uncoupled events in L929 cells. A protective role of NF-B would imply that the inhibition of this transcription factor would lead to enhanced cell death. However, even the partial inhibition of NF-B transactivation by DEX was accompanied by a decrease in cell killing. In the case of L929 cells, the DEX-mediated repression of NF-B seems to be independent from up-regulation of IB-␣ and can probably be attributed to the recently proposed repression of transactivation (28). Conversely, the activation of NF-B does not promote killing of L929 cells, because BA allowed the full activation of NF-B while reducing the cytotoxic effects of TNF-␣. Evidence for the uncoupling of NF-B activation and cytotoxicity in L929 cells was obtained by the unchanged susceptibilities to cell death in cell lines stably overexpressing a transdominant negative variant of IB-␣. The role of NF-B as a promoter or attenuator of cell killing may therefore depend on the nature of the apoptosis-inducing stimulus as well as on the cell type. Accordingly, the inhibition of NF-B activation by IB-␣ overexpression did not alter the sensitivity of the human breast cancer cell line MCF7 toward TNF-␣ (29), whereas the HT1080 fibrosarcoma cells and Jurkat T-cells were more susceptible to the detrimental effects of TNF-␣ (12,13).