Resistance to Tumor Necrosis Factor-α (TNF-α)-induced Apoptosis in Rat Hepatoma Cells Expressing TNF-α Is Linked to Low Antioxidant Enzyme Expression

In order to study the mechanisms of resistance to tumor necrosis factor-α (TNF-α), we have constructed two stable transfectants producing TNF-α (Yv12–2 and Yv13–44) from the rat hepatoma H4IIE cell, which does not produce TNF-α. H4IIE cells were highly sensitive to apoptosis induced by TNF-α, whereas Yv2–12 and Yv13–44 cells were resistant. Manganous superoxide dismutase was not up-regulated in Yv2–12 and Yv13–44 cells and was unresponsive to induction by exogenous TNF-α and by H2O2 in H4IIE cells and in the transfectants. Catalase expression and activity were lower in Yv2–12 and Yv13–44 cells than in H4IIE cells; furthermore, the transfectants were more susceptible to H2O2. Treatment with exogenous TNF-α down-regulated catalase in H4IIE cells but not in Yv2–12 and Yv13–44 cells. Treatment of H4IIE cells with the catalase inhibitor 3-amino-1,2,4-triazole rendered them resistant to exogenous TNF-α. These data suggest a causal relationship between resistance to TNF-α and low catalase activity. Expression of copper and zinc containing superoxide dismutase was also decreased, whereas expression of glutathione peroxidase-1 was unchanged in Yv2–12 and Yv13–44 cells. Data from a microarray point to a down-regulation of genes in the resistant clones that code for antioxidative proteins and proteins involved in glutathione synthesis and function. We assume that a prooxidant signal linked to the down-regulation of antioxidant defense may be associated with resistance to apoptosis induced by TNF-α.

Tumor necrosis factor-␣ (TNF-␣) 1 is a pleiotropic cytokine acting in a paracrine or autocrine fashion on a wide variety of target cells via two members of the TNF-␣ receptor family, TNF receptor 1 (TNFR1) and TNF receptor 2 (TNFR2). Binding of TNF-␣ trimers to these receptors activates signal transduction pathways that lead to different and even antagonistic effects on cell function and survival (reviewed in Refs. 1 and 2): 1) TNF-␣ supports the formation of proinflammatory molecules that are constituents of the immune response. 2) TNF-␣ can exert mitogenic actions and supports proliferation processes. 3) TNF-␣ can induce apoptosis via caspase 8 activation and the mitochondrial pathway. 4) TNF-␣ supports (a) survival pathway(s), supposedly via the formation of antiapoptotic proteins. Development of resistance against TNF-␣-induced cytotoxicity has been demonstrated in a variety of experimental models, including the repeated administration of exogenous TNF-␣ to cultured cells (3) and animals (4,5) or the overexpression of endogenous TNF-␣ in cultured cells (6,7).
Among the possible mechanisms that might cause TNF-␣ resistance is the induction of protective proteins that counteract the apoptotic action of TNF-␣. Induction of the mitochondrial antioxidant enzyme manganous superoxide dismutase (MnSOD) by TNF-␣ has been reported in a wide variety of cell types and has been assumed to function as a central protective mechanism against TNF-␣-induced apoptosis (8,9). In experiments with stable transfection of the TNF-␣ gene into tumor cells, up-regulation of MnSOD paralleled TNF-␣ resistance (10,11). Overexpression of MnSOD protected cells against TNF-␣ cytotoxicity, whereas the expression of antisense MnSOD-RNA increased the susceptibility of cells toward TNF-␣ (9,12). It is known that mitochondrial generation of reactive oxygen species (ROS) is involved in the cytotoxicity of TNF-␣ (13,14) and that inhibitors of mitochondrial function prevent the induction of MnSOD (15). It may therefore be assumed that protection by MnSOD induction against TNF-␣-induced cytotoxicity is an adaptation reaction to mitochondrial superoxide production.
Up-regulation of MnSOD as a mechanism of TNF-␣ resistance suggests that TNF-␣-resistant cells might also be resistant against oxidants. It has been shown that tumor cells overexpressing TNF-␣ and thus rendered resistant against TNF-␣ were also protected against the redox-cycling cytostatic drug adriamycin (10,16,11); it may be assumed that the protection was because of the removal of adriamycin-generated superoxide anion radicals by the up-regulated MnSOD activity. In the liver, TNF-␣ has been shown to be involved in viral and immune-mediated hepatitis and drug-induced liver failure (17). Consequently, under pathophysiological conditions related to oxidative stress the liver can be subject to prolonged exposure to TNF-␣, and resistance to TNF-␣-induced liver damage has been shown in animals exposed to TNF-␣ (4,5). However, the mechanisms underlying TNF-␣ resistance in liver cells have not yet been elucidated. In the present study, two TNF-␣resistant clones were constructed from a TNF-␣-sensitive hepatoma cell line in order to investigate whether in the transfectants the antioxidant enzyme system would be induced.

EXPERIMENTAL PROCEDURES
Cell Lines and Culture Conditions-H4IIE cells (a rat hepatoma cell line), transfectants derived from H4IIE cells, and L929 cells (a murine fibrosarcoma cell line) were grown in Dulbecco's modified Eagle's me-* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Stable Transfection of TNF-␣ into H4IIE Cells-TNF-␣ cDNA was obtained by RT-PCR from total rat hepatocyte RNA isolated with Trizol (Invitrogen). Hepatocytes were prepared from male Wistar rats (250 -300 g) by a modified method according to Lindl and Bauer 1989 (19). The amplified product was purified and cloned into the HindIII site of the pSPT18 vector. Correct sequence was verified in randomly selected pSPT18-TNF-␣ clones by the use of the T7 and SP6 sequencing primers. For stable transfection into H4IIE cells, the pTRE vector (Clontech) was used. Correct sequence was verified in randomly selected pTRE-TNF-␣ clones with vector-specific primers. For selection of tetracycline-responsive transfectants, cells were transfected with 10 g of pTet-ON or pTet-OFF using calcium phosphate precipitation. Clones were selected in the presence of 800 g/ml G418 48 h after transfection for 2 weeks. For establishing TNF-␣ overexpressing cells, the selected clones were co-transfected with pTRE-TNF-␣ and pTK-Hyg (Clontech), a plasmid that encodes hygromycin resistance. Clones were selected in the presence of 500 g/ml hygromycin B 48 h after transfection, using calcium phosphate precipitation. Isolates of clones Yv2-12 and Yv13-44 were tested for TNF-␣ expression by RT-PCR.
TNF-␣ Detection-Total TNF-␣ concentration in the culture medium was determined by commercially available rat TNF-␣ ELISA according to the manufacturer's recommendations (Endogen, Boston, MA). This kit is sensitive to 5 pg/ml TNF-␣ and does not cross-react with TNF-␤, interleukin-1, or interleukin-6. Biologically active TNF-␣ was quantified by the L929 assay as described previously (20). One unit of TNF-␣ was defined as the reciprocal of the dilution at which 50% L929 cytotoxicity was observed.
Catalase Activity-Catalase activity was determined by measuring the initial rate of decay of H 2 O 2 absorbance at 240 nm, using an extinction coefficient of 43.6 M Ϫ1 cm Ϫ1 . Prior to the catalase assay, homogenates were incubated with ethanol and Triton X-100 according to Cohen et al. (21). Each assay mixture consisted of a cell sample (150 g of protein/ml) and H 2 O 2 at an initial concentration of 19 mM H 2 O 2 in 50 mM phosphate buffer (pH 7.0).
RNA Isolation and Northern Blot Analysis-Total RNA was isolated from cells using Trizol (Invitrogen). Total RNA (5-10 g) was resolved by electrophoresis in a 1.5% agarose, 2.25 M formaldehyde gel. RNA was transferred to nylon membrane (Amersham Biosciences) as described by Sambrook et al. (22). Purified cDNAs were labeled with [␣ 32 P]dCTP (111 TBq/mmol; Hartmann Analytic, Braunschweig, Germany) by random hexamer priming (Roche Diagnostics). Blots were prehybridized and hybridized with cDNAs as described previously (23) and then exposed to x-ray films (Kodak XAR) for 1-3 days with an intensifying screen at Ϫ80°C. 18 S rRNA served as internal loading control. Autoradiographs were analyzed by densitometric scanning using the Quantity One system from Bio-Rad.
Microarray-Gene expression was analyzed using Atlas Rat Toxicology Array II (Clontech). Total RNA was extracted from TNF-␣-transfected and parental H4IIE cells using Trizol (Invitrogen). The sample was treated with RNase-free DNase I for 60 min, extracted using phenol-chloroform, and then precipitated using 2.5 volumes ethanol. Integrity of the RNA was confirmed on an agarose gel. Two g of total RNA were reverse-transcribed with nucleotides containing [␣-32 P]dATP; the labeled cDNAs were purified, denatured, and added to 5 ml of Express-Hyb hybridization solution (Clontech). The filters were prehybridized, hybridized, washed according to the manufacturer's recommendations, and exposed to an x-ray film overnight at Ϫ80°C or to a phosphorimaging screen and scanned (BAS-MS 2025 phosphorimaging; Raytest, Germany). Signal intensities of corresponding spots were compared by AIDA software (Raytest).
Cell Lysis and Western Blot Analysis-Cells were lysed by two rounds of freezing and thawing in lysis buffer (50 mM Tris-HCl, 2.5 mM ethylene diamine tetraacetic acid (EDTA), 150 mM NaCl, 1 mM phenylmethyl sulfonyl fluoride, pH 7.4), and then centrifuged at 10 000 ϫ g for 10 min. Protein concentration in cell lysates was determined using a protein assay kit (Bio-Rad). An equal amount of protein for each sample was separated by 10% SDS-PAGE and electrotransferred onto a polyvinylidene fluoride membrane (Roche Diagnostics). Protein loading was assessed by Ponceau S staining of membranes. The membranes were blocked with 3% nonfat dry milk in a buffer containing 20 mM Tris, 500 mM NaCl, and 0.01% Tween 20 (pH 8.0) and incubated with 1:500 -1:1000 dilution of antibody against MnSOD (Upstate, Milton Keynes, UK) and catalase (Calbiochem). Blots were then probed with the appropriate horseradish peroxidase-conjugated secondary antibody. Bound antibody was visualized using enhanced chemiluminescence reagent (Roche Diagnostics).
Analysis of DNA Fragmentation-After incubation with TNF-␣ or H 2 O 2 for 24 h, 5 ϫ 10 6 cells were harvested. Cell pellets were resuspended in lysis buffer (20 mM Tris, 5 mM EDTA, 0.1 M NaCl, 0.5% SDS, RNase A 10 g/ml, pH 8.0) and incubated for 1 h at 37°C. DNA was extracted with an equal volume of phenol:chloroform:isoamyl alcohol (25:24:1) and precipitated with 0.7 volume isopropanol and 0.1 volume 3 M NaCl. The pellets were resuspended in 50 l of Tris/EDTA buffer, and DNA was electrophoresed using a 2% agarose gel at 70 V for 3-4 h. The DNA bands were visualized by ethidium bromide staining and photographed under UV light using a transilluminator.
Anti-histone ELISA-Quantitative measurements of apoptosis were performed using an ELISA that quantifies DNA fragmentation by measuring cytoplasmic histone-DNA fragments (Cell Death ELISA; Roche Diagnostics). Cells were incubated with TNF-␣ or H 2 O 2 as described above; apoptosis was assessed 24 h later according to the manufacturer's recommendations.
Statistical Analyses-All data were analyzed using one-way analysis of variance, followed by least significant difference post hoc analysis to determine statistical significance. p values Ͻ0.05 were considered statistically significant.

RESULTS
Stable transfection of H4IIE cells with TNF-␣ cDNA resulted in the emergence of two clones (Yv2-12 and Yv13-44) that constitutively expressed TNF-␣ to a different extent. These two clones were used for further experimentation. The concentration of total TNF-␣ protein and biologically active TNF-␣ in the medium was one order of magnitude higher in Yv2-12 cells than in Yv13-44 cells. The wild type cells (wtH4IIE cells) did not produce TNF-␣. No difference in cell proliferation and morphology was observed between wtH4IIE cells and transfectants.
WtH4IIE cells, but not Yv2-12 and Yv13-44 cells, were highly susceptible to the cytotoxic action of exogenous TNF-␣ at a concentration of 25 ng/ml. Yv13-44 showed a slight loss of viability at 50 ng/ml (Fig. 1), and Yv2-12 was completely resistant to TNF-␣ up to 500 ng/ml. Apoptosis as detected by oligonucleosomal DNA fragmentation ( Fig. 2A) and by an antihistone ELISA (Fig. 2B) was elicited by treatment with exogenous TNF-␣ in wtH4IIE cells but not in Yv2-12 and Yv13-44 cells. In wtH4IIE cells, but not in Yv2-12 and Yv13-44 cells, a slight increase in p53 and bax expression was found, although bcl-2 expression was not altered. Activation of caspase 3 (Fig.  3A), caspase 8 (Fig. 3B), and caspase 9 (Fig. 3C) upon treatment with exogenous TNF-␣ was found in wtH4IIE cells but not in Yv2-12 cells, whereas Yv13-44 cells displayed intermediate activation of caspase 3 but not of caspase 8 and 9.
TNF receptor expression was measured in order to exclude the possibility that resistance to TNF-␣ in the transfectants is because of a loss of TNFR1 and/or TNFR2 expression. The wtH4IIE cells as well as the two TNF-␣-resistant clones expressed the TNFR1, whereas none of them expressed the TNFR2. TNFR1 functionality was not lost, as shown by the induction of NF-B activation by exogenous TNF-␣ in the resistant clones (data not shown).
Because formation of ROS is assumed to be involved in TNF-␣ cytotoxicity, TNF-␣ resistance might be caused by the up-regulation of antioxidant enzymes. An increase of MnSOD expression has been shown to occur in other TNF-␣-resistant cell types (10,11). It might therefore be assumed that this adaptation process also occurred in Yv2-12 and Yv13-44 cells. Fig. 4 shows that neither the MnSOD mRNA level nor the MnSOD protein level in Yv2-12 and Yv13-44 cells differed from those in wtH4IIE cells. MnSOD induction could not be obtained in either wtH4IIE cells or Yv2-12 and Yv13-44 cells by exogenous TNF-a.
To provide a possible explanation for the TNF-␣ resistance of the clones, we examined whether catalase and GPx-1 expression is increased in Yv2-12 and Yv13-44 cells. Fig. 5 demonstrates that no difference in GPx-1 mRNA expression was found between wild type cells and TNF-resistant cells and that contrary to expectations catalase mRNA was even lower in the TNF-␣-resistant clones than in wtH4IIE cells (Fig. 5B). Catalase protein expression and catalase enzyme activity were also decreased in the transfectants (Fig. 5, C and D). Treatment with exogenous TNF-␣ down-regulated catalase expression and enzyme activity in wtH4IIE cells but not in the TNF-␣-resistant clones (Fig. 6). A slight induction of GPx-1 by exogenous TNF-␣ could be achieved in wtH4IIE cells but not in Yv2-12 and Yv13-44 cells.
To test the conclusion from Figs. 5 and 6 that low catalase activity is associated with TNF-␣ resistance, cells were treated with the catalase inhibitor ATZ. Fig. 7 shows that in wtH4IIE cells ATZ treatment did, indeed, markedly lower the sensitivity to TNF-␣. No influence of ATZ treatment on TNF-␣ sensitivity was observed in the Yv2-12 and Yv13-44 cells.
To test the initial hypothesis that TNF-␣-resistant cells also display resistance toward H 2 O 2 , cytotoxicity of H 2 O 2 was compared in wtH4IIE cells and in Yv2-12 and Yv13-44 cells. Fig.  8A shows that Yv2-12 and Yv13-44 cells are even more susceptible toward H 2 O 2 than wtH4IIE cells. In Fig. 8, B and C, the induction of apoptosis by H 2 O 2 as detected by oligonucleosomal DNA fragmentation and by an anti-histone ELISA was compared in wtH4IIE cells versus Yv2-12 and Yv13-44 cells. The transfectants were considerably more sensitive to apoptosis induced by H 2 O 2 than the wtH4IIE cells.
MnSOD expression as well as catalase expression can be increased by H 2 O 2 in a number of cells (22, 24 -27). In our experiments, similar to the results with exogenous TNF-␣, both in the wild type cell and the TNF-␣-resistant cells, MnSOD is unresponsive to H 2 O 2 treatment. Catalase mRNA expression and enzyme activity could be increased by H 2 O 2 treatment both in wtH4IIE cells and in Yv2-12 and Yv13-44 cells.
A DNA microarray comparing wtH4IIE cells and Yv2-12 cells was performed to support the hypothesis that TNF-␣ resistance is paralleled by a shift toward a more pro-oxidative state and to identify additional genes differentially regulated in the resistant cells. Of 450 genes, 8 were up-regulated and 31 were down-regulated in the Yv2-12 cells by a factor of at least 2. Preliminary evaluation of the microarray data indeed points to a down-regulation of genes for antioxidant enzymes and of genes involved in glutathione synthesis and in detoxifying phase II metabolism (Table I). At present, the down-regulation of copper and zinc containing superoxide dismutase (CuZn-SOD) indicated by the microarray has been verified by Northern blotting in the Yv2-12 and Yv13-44 cells (data not shown). Moreover, down-regulation of CuZnSOD was obtained in wtH4IIE cells, but not in Yv2-12 and Yv13-44 cells, upon treatment with exogenous TNF-␣. DISCUSSION With the experiments reported here, we have collected evidence that resistance to TNF-␣-induced apoptosis can be associated with a down-regulation of antioxidant proteins. For this, we have constructed two TNF-␣-producing stable transfectants from a wild type rat hepatoma cell that does not express TNF-␣. The wtH4IIE cell is highly sensitive to TNF-␣-induced apoptosis, whereas the transfectants are resistant to TNF-␣induced apoptosis.
TNF-␣ induces apoptosis via the death receptor pathway, and down-regulation of TNF-␣ receptors could theoretically be the mechanism of TNF-␣ resistance. This is, however, not the main mechanism in our transfectants because (a) both the wild type cell and the transfectants express the TNFR1 to a similar extent, (b) both the wild type cell and the transfectants do not express the TNFR2 at all, and (c) NF-B activation by TNF-␣, assumed to require TNF-R1 functionality, can be induced in the transfectants. In vivo, Sass et al. (28) also did not find a down-regulation of TNFR1 in TNF-␣-resistant mice.
A mitochondrial component is involved in TNF-␣-induced cytotoxicity (13,14). This has supported the assumption that fortification of mitochondrial defense against oxidative stress by up-regulating MnSOD could be a mechanism of TNF-␣ resistance. In contrast to our initial expectations, up-regulation of MnSOD expression did not occur in our TNF-␣-producing transfectants and thus could be excluded as a mechanism of TNF-␣ resistance in this cell type. MnSOD is expressed to the same extent in the wtH4IIE cell and in the TNF-␣-resistant clones derived from it and is unresponsive to induction by exogenous TNF-␣. Induction of the enzyme in wtH4IIE cells and in the TNF-␣-resistant clones could not be achieved with H 2 O 2 either. In contrast, MnSOD can be induced by TNF-␣ (29,30) and H 2 O 2 (23) in primary rat hepatocytes. The reason for the intrinsic inability of the H4IIE cell and its descendents to up-regulate MnSOD has not yet been elucidated.
It might have been expected that the expression of other antioxidant enzymes in the TNF-␣-resistant cells was increased as a substitute for the lack of up-regulation of MnSOD. However, gene expression analysis revealed that in the transfectants none of the antioxidant enzymes examined is up-regulated, whereas two antioxidant enzymes, catalase and CuZn-SOD, are even down-regulated. This expression pattern is mimicked by TNF-␣ treatment of the TNF-␣-sensitive wtH4IIE cell: Exogenous TNF-␣ leads to a decrease of catalase and CuZnSOD expression in these cells. As a consequence of low catalase activity, the sensitivity of the transfectants to H 2 O 2 was increased. A decrease in catalase activity (31) and catalase mRNA (32) upon TNF-␣ treatment has also been observed in vivo. The finding that exposure to endogenous TNF-␣ in the TNF-␣-resistant clones and to exogenous TNF-␣ in the TNF-␣sensitive wild type cell leads to a decrease in catalase expression and activity suggests that the down-regulation of this antioxidant enzyme is not merely an epiphenomenon of the transfection process. Additionally, we performed a pharmacological experiment that strongly indicates a causal relationship between low catalase activity and TNF-␣ resistance: Exposure of TNF-␣-sensitive wtH4IIE cells to the catalase inhibitor ATZ renders these cells partially resistant to exogenous TNF-␣.
We conclude from these findings that a pro-oxidant signal linked to down-regulation of antioxidant defense contributes to the TNF-␣-resistant state. This signal may transfer information via mediator concentrations of ROS too low to exert cytotoxic effects. The conclusion that a shift of the pro-oxidantantioxidant balance of the cell has occurred in the TNF-␣resistant clones is further supported by the preliminary results from a cDNA microarray analysis that suggest the down-regulation of a number of proteins with antioxidative properties in the Yv2-12 cell, including thioredoxin-2 and metallothionein-1 as well as a number of enzymes involved in glutathione supply and function.
The conclusion that low activity of antioxidant enzymes may be involved in TNF-␣ resistance is at a first glance not consistent with two interpretations brought forward in the literature. First, it is commonly assumed that the antioxidant enzyme MnSOD, induced via the NF-B survival pathway, is the main protective principle against TNF-␣-induced cytotoxicity (8). In our cells, MnSOD cannot contribute to protection against TNF-␣ because the enzyme is intrinsically unresponsive to induction. Even in cells in which MnSOD is induced, protection may not be linked to the antioxidant character of the enzyme. The reaction product is H 2 O 2 , and the antioxidant potential of the MnSOD reaction can only be realized by subsequent reduction of H 2 O 2 to H 2 O by glutathione peroxidases or catalase. Siemankowski et al. (33) have reported that in MCF-7 cells TNF-␣ induces MnSOD without any change in the expression of other antioxidant enzymes, including GPx-1 and catalase. This may indicate that the mechanism by which MnSOD protects against TNF-␣ could be pro-oxidative in nature because the steady state level of H 2 O 2 increases. It has been concluded from experiments in cRel overexpressing HeLa cells that H 2 O 2 produced via MnSOD induction creates an oxidative injury leading to mitochondrial degeneration, proliferation arrest, and resistance to TNF-␣-induced cytotoxicity (34). On the other hand, the down-regulation of the cytoplasmic CuZnSOD observed in the present work, not only in the TNF-␣-resistant clones but also in the TNF-␣-treated wild type H4IIE cells, is not consistent with this view because the resulting pattern of SOD activities would decrease H 2 O 2 delivery to GPx and catalase. We assume that a coordinated response of H 2 O 2 -delivering and H 2 O 2 -consuming enzymes to TNF-␣ exposure is required for the maintenance of a well dosed pro-oxidant signal to the TNF-␣-dependent signaling cascade. Secondly, in the mouse, transcriptional and/or translational inhibitors sensitize cells and animals to the toxic action of TNF-␣ (35). This suggests that the de novo synthesis of protective proteins is needed for TNF-␣ resistance and down-regulation (e.g. of catalase) as a protective mechanism does not seem to fit in this picture. However, catalase is known to be in part regulated post-transcriptionally via the degradation of the catalase mRNA by a protein (36). If in the presence of a transcriptional or translational inhibitor (a situation linked to high TNF-␣ sensitivity) this protein is no longer formed, this would result in an apparent up-regulation of catalase while uninhibited de novo synthesis of this regulatory protein would be required to achieve a state of TNF-␣ resistance.
In the literature, a number of findings support the view that a pro-oxidant environment confers resistance to death receptordependent apoptosis. Overexpression of catalase increases the sensitivity of HepG2 cells to TNF-␣ (37). From experiments with radiation-exposed leukemia cells, it has been concluded that the primary role of catalase in multicellular organisms could be the regulation of apoptosis rather than the detoxification of H 2 O 2 (38). In lymphocytic cell lines, glutathione depletion prevents TNF-␣-induced apoptosis (39). NF-B activation, a process shown to be a survival pathway in TNF-␣-treated cells (40), is a candidate target for a pro-oxidative anti-apoptotic action because a co-stimulatory role of reactive oxygen species in NF-B activation has been postulated (41).
In the experiments on TNF-␣ resistance reported here, we have focused on the antioxidant enzyme system and on the existence of a pro-oxidant signal maintaining the resistant state. This, of course, does not exclude further mechanisms, either related or unrelated to the redox state of the cell, that may also be involved in the process of TNF-␣ resistance in hepatoma cells as well as in other cell types. In the mouse in vivo, it has been excluded that interleukin-1␤, inducible nitric oxide synthase, and heme oxygenase-1 confer to resistance TNF-␣, although they are induced in the resistant animals (28). By introduction of a randomized hybrid ribozyme library into MCF-7 cells, a number of genes functional in TNF-␣induced apoptosis were recently identified (42). It may be worthwhile testing whether these genes are down-regulated in a TNF-␣-resistant cell.
In conclusion, we have shown that in TNF-␣-resistant transfectants derived from a TNF-␣-sensitive rat hepatoma cell line a down-regulation of certain antioxidant proteins occurs in the absence of any increase in the expression of the other antioxidant proteins examined. We suggest that a pro-oxidant signal linked to the down-regulation of antioxidant defense could be associated with the TNF-␣-resistant state.