2-Acetylaminofluorene Up-regulates Rat mdr1bExpression through Generating Reactive Oxygen Species That Activate NF-κB Pathway*

Overexpression of multidrug resistance genes and their encoded P-glycoproteins is a major mechanism for the development of multidrug resistance in cancer cells. The hepatocarcinogen 2-acetylaminofluorene (2-AAF) efficiently activates rat mdr1b expression. However, the underlying mechanisms are largely unknown. In this study, we demonstrated that a NF-κB site on the mdr1b promoter was required for this induction. Overexpression of antisense p65 and IκBα partially abolished the induction. We then delineated the pathway through which 2-AAF activates NF-κB. 2-AAF treatment led to the increase of intracellular reactive oxygen species (ROS) which causes activation of IKK kinases, degradation of IκBβ (but not IκBα), and increase in NF-κB DNA binding activity. Consistent with the idea that ROS may participate inmdr1b regulation, antioxidant N-acetylcysteine inhibited the induction of mdr1b by 2-AAF. Overproduction of a physiological antioxidant glutathione (GSH) blocked the activation of IKK kinase complex and NF-κB DNA binding. Based on these results, we conclude that 2-AAF up-regulates mdr1b through the generation of ROS, activation of IKK kinase, degradation of IκBβ, and subsequent activation of NF-κB. This is the first report that reveals the specific cis-elements and signaling pathway responsible for the induction of mdr1b by the chemical carcinogen 2-AAF.

A major problem in cancer chemotherapy is the development of resistance to a wide range of structurally and functionally unrelated anti-cancer drugs. One major mechanism in the development of multidrug resistance (MDR) 1 is by overexpressing MDR1 gene and its encoded P-glycoprotein. It is generally believed that the overexpressed P-glycoprotein facilitates the efflux of anticancer drugs from the cytoplasm, thereby reducing the intracellular drug content to sublethal level. In humans, there are two classes of MDR genes: MDR1, which is involved in multidrug resistance, whereas MDR2, in the lipid transport. There are three mdr gene homologues in rodents, but only mdr1a and mdr1b confer multidrug resistance; while mdr2 functions as a lipid transporter.
Levels of MDR1 expression is frequently elevated in human hepatocellular carcinoma (HCC) (1,2). Elevated expression of mdr gene transcripts and their encoded P-glycoprotein is also seen in rodent HCC (3,4). However, mechanisms of the elevation of MDR expression in HCC are largely unknown. In the present study, we investigated the mechanisms of elevated hepatic mdr1b expression in rats induced by hepatocarcinogen 2-acetylaminofluorene (2-AAF). 2-AAF is a hepatocarcinogen that has been frequently used in the development of HCC in experimental animals. AAF is a genotoxic agent. Reaction of electrophilic 2-AAF derivatives with nucleophilic DNA results in the formation of DNA adducts (5). In the treated animals, DNA adducts are proportional to dose in both target tissues, liver and bladder; whereas tumor formation increases linearly with response to dose only in the liver (6). It is believed that 2-AAF also induces liver cancers through non-genotoxic effects, such as the promotion of cell proliferation (7). We (4) and others (8,9) have previously demonstrated that rat HCC induced by 2-AAF exhibited elevated expression of mdr1b. Using a rat hepatoma cell line H4-II-E, we demonstrate here that the induction of mdr1b expression by 2-AAF is mediated by the activated NF-B, which recognizes a cis-acting element located upstream of mdr1b promoter. We also demonstrated that the activation is mediated through oxidative stress induced by 2-AAF, as evidenced by the observations that induction of mdr1b expression can be regulated by the redox modulators.
RNase Protection Assay-The rat mdr1b and 18 S rRNA probes were synthesized by in vitro transcription as described previously (4). Either 20 or 1 g of total RNA was used for mdr1b or 18 S rRNA, respectively. The protected RNA products were analyzed on a 7% denaturing PAGE gel and quantified using a densitometer (Molecular Dynamics, Sunnyvale, CA).
Cell Culture, 2-AAF Treatment, and Transfection-The rat hepatoma cell line H-4-II-E (ATCC 1548) and human embryonic kidney cell line 293 (CRL-1573) were obtained from the American Type Culture Collection. ␥-GCS overexpressing cell line H9 derived from H4-II-E cells has been described elsewhere (15). For 2-AAF (or 2-AAAF) treatment, 2 ϫ 10 6 cells were plated in 10 ml of growth medium on 10-mm Petri dishes 16 h prior to the addition of drugs. Transfection of H4-II-E cells was performed by LipofectAMINE Plus method (Life Technologies Inc., Rockville, MD), following the manufacturer's manual. For stable transfection, 0.5 g of pcDNA3-Neo (Invitrogen, Carlsbad, CA) was mixed with 2.5 g of reporter plasmids. After transfection, cells were selected by G418 resistance according to the procedure described previously (15).
Luciferase Assay and CAT Assay-Cells were lysed by reporter lysis buffer supplied with the luciferase assay kits (Promega Co., Madison, WI). 20 g of protein was used for both reporter assays. For luciferase assays, the intensity of luminescence was measured by a luminometer (Tuner Designs TD-20/20). CAT assay procedure was previously reported (10). In transient transfection assays, pCMV ␤-gal was co-tranfected with reporter constructs and the results were normalized by ␤-galactosidase activities.
Preparation of Cytoplasmic Extracts and Nuclear Extracts and Gel Mobility Shift Assay-1 ϫ 10 7 cell pellets were first washed with PBS and then lysed in 100 l of hypotonic buffer (10 mM HEPES, 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 2 g/ml each of aprotinin and leupeptin, and 0.5 mg/ml benzamidine). The supernatants (cytoplasmic extracts) were saved for Western blots. The pellets were resuspended in 40 l of high salt buffer (20 mM HEPES, 400 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, 2 g/ml aprotinin, 2 g/ml leupeptin, and 0.5 mg/ml benzamidine) for 30 min on ice, with occasional vortex. After centrifugation, the supernatants (nuclear extracts) were saved for gel mobility shift assay.
Gel mobility shift assay was performed with 10 g of nuclear proteins in a total volume of 20 l containing 10 mM Tris-HCl, 50 mM NaCl, 0.5 mM dithiothreitol, 10% glycerol, 0.2% Nonidet P-40, 3 g of poly(dI-dC)⅐poly(dI-dC) and radiolabeled DNA probe (an 88-bp fragment containing 5 ϫ B sites of Ig chain promoter).
Western Blotting Analysis of Endogenous and Transient Transfected IB␣ and IB␤-To analyze endogenous IBs, 50 g of cell extracts from 2-AAF-treated H4-II-E cells were resolved on 10% SDS-PAGE gel and transferred to a nitrocellulose membrane. The membrane was sequentially probed with anti-IB␣, anti-IB␤, and anti-␤-actin antibodies. To analyze the transfected IBs, 6 g of pCMV4-F-IB␣ or pCMV4-F-IB␤ was transfected into 293 cells on a 100-mm dish and 15 h after transfection, cells were split onto 6-well dishes and incubated for another 15 h. Then 100 M 2-AAF was added and incubated for different times as indicated in the figure legend. After washing with cold PBS for three times, whole cell extracts were prepared and 20 g were analyzed by Western blotting, using anti-Flag M2 antibody and anti-␤-actin sequentially.
Flow Cytometry Assay of ROS-2 ϫ 10 5 cells were seeded on 6-well dishes 15 h prior to the experiment. 100 M 2-AAF or 10 M 2-AAAF was added to the cell culture medium at different time points and cells were incubated for different periods of time as indicated in the figure legend. Cells were washed three times with cold PBS and incubated with 8 M dichlorofluorescein diacetate in serum-free phenol red-reduced Dulbecco's modified Eagle's medium at 37°C for 30 min. Cells were then washed with cold PBS for three times and scraped from the dishes in 1 ml of cold PBS. The fluorescence intensity of the dichlorofluorescein diacetate-labeled cells was measured on a FACcan flow cytometer (Becton Dickson, San Jose, CA) and for each sample, 1 ϫ 10 4 cells were analyzed using the green fluorescene emission parameter. The mean fluorescent intensity values were calculated from a fourdecade logarithmic scale by the CellQuest (Becton Dickson) software.
IKK Kinase Activity Assay-Cells were washed twice with ice-cold PBS and then lysed in lysis buffer (20 mM Hepes, 200 mM NaCl, 1% Triton X-100, 10 mM ␤-glycerophosphate, 1 mM NaF, 5 g/ml each of aprotinin and leupeptin, and 1 g/ml benzamidine). 300 g of total cell lysate was incubated with 1 g of IKK␣ antibody and 30 l of protein A-G-conjugated agarose beads (Santa Cruz Biotechology, Inc.) at 4°C for overnight and then washed six times with wash buffer (20 mM Hepes, pH 7.4, 200 mM NaCl, and 0.5% Triton X-100). Beads were resuspended in 20 l of kinase mixture containing 2 g of GST-IB␣, 0.5 Ci of [␥-32 P]ATP, 20 mM Hepes, 10 mg MgCl 2 , 2 mM MnCl 2 , 2 mM dithiothreitol, 50 M Na 3 VO 4 , and 10 mM ␤-glycerophosphate at 30°C for 30 min. 20 l of 2 ϫ SDS loading dye was added to each reaction, boiled for 5 min, and one-third of the total volume was separated on 10% SDS-PAGE gel. Gel was dried and exposed to x-ray film. (4) and other (8,9) have previously demonstrated that expression of mdr1b in hepatoma cells and in liver cancers can be induced by 2-AAF. To confirm that induction of rat mdr1b mRNA could be seen in a cultured cell system, we measured mdr1b mRNA expression by RNase protection assay in rat heptoma cell line H4-II-E, following 2-AAF treatment. Fig. 1 shows a concentration-and time-dependent increase of mdr1b mRNA levels. Densitometric analyses revealed that maximal levels (7.5-fold) of induction were at 100 M (panel A). Induction of mdr1b mRNA was seen 3 h after treatment with 100 M 2-AAF, reached maximum at 5 h, and declined thereafter. Treatment of H4-II-E cells with 10 M 2-AAAF, a major active metabolite of 2-AAF, also induces mdr1b expression to similar levels, indicating that 2-AAAF is a much more potent inducer than 2-AAF (data not shown).

2-AAF and 2-AAAF Up-regulate Rat mdr1b mRNA Expression in H4-II-E Cells-We
The NF-B-binding Site Is Required for 2-AAF Induction of mdr1b-To determine the DNA sequences responsible for the induction, a set of progressive deletion constructs Ϫ243 RMI-CAT, Ϫ214RMICAT, and Ϫ163RMICAT (11, 12) were stably transfected into H4-II-E cells and their responses to 2-AAF treatment were measured by CAT assay. Mass cultures each consisted of more than 20 positive clones were used. The reason for using stably transfected cells rather than the transient transfection approach was because mdr1b gene expression is sensitive to cellular stress and the transfection per se is a stress inducing procedure. Moreover, the transfection efficiency in H4-II-E cells is generally low. As shown in Fig. 2, Ϫ243RMI-CAT and Ϫ214RMICAT constructs were responsive to 2-AAF treatment, but Ϫ163RMICAT completely lost it responsiveness. Thus DNA sequences critical for the 2-AAF induction are located between Ϫ214 and Ϫ163 bp, a region containing previously identified binding sites of p53 and NF-B (11,12). To determine whether the NF-B site in this region is involved in the 2-AAF induction, we made a stably transfected cell line containing a reporter construct with mutation at this site. As shown in Fig. 2, mutation of the NF-B site abolished the induction of reporter expression by 2-AAF. These results thus identified that the NF-B site located at Ϫ167 to Ϫ158 bp of the rat mdr1b promoter are responsible for the induction of mdr1b expression by 2-AAF. Silverman and Hill (9) reported that the DNA sequence between Ϫ214 and Ϫ178 bp was important for the basal and carcinogen inducible promoter activity. However, these investigators failed to dissect the critical sequences that are involved within this region.
Down-regulation of NF-B Abolishes the Induction of mdr1b-Luc by 2-AAF-NF-B usually consists of p50 and p65 subunits. In most unstimulated cells, NF-B is tightly controlled by a class of ankyrin containing inhibitors IBs, which bind to NF-B subunits and sequester them in the cytoplasm. To con- firm the involvement of NF-B in 2-AAF induction of mdr1b, we established a stable cell line from 293 cells by transfecting Ϫ245 mdr1b-Luc. We then transiently transfected recombinant plasmids encoding NF-B inhibitors IB␣ and antisense p65, respectively, into this cells line. As indicated in Fig. 3, transfecting both antisense p65 and IB␣ expressing plasmids abolished the induction by 2-AAF. As a control, no effect on 2-AAF induction was observed when the empty vector was used. These results further demonstrated that NF-B is involved in the 2-AAF induction of mdr1b expression.
2-AAF Activates NF-B DNA Binding Activity and NF-Bmediated Transcription-We then examined whether induction of mdr1b expression by 2-AAF is mediated by an increase of NF-B DNA binding activity. H-4-II-E cells were exposed to 2-AAF, nuclear extracts were prepared. As shown in Fig. 4A, DNA binding activity of NF-B became detectable at 2 h, continued to increase until it reached a maximum at 5 h posttreatment and declined thereafter. The binding was verified by virtue that the complex could be competed by an excess of unlabeled wild type mdr1b-B oligonucleotide fragment, but not by a mutant mdr1b-B oligonucleotide. The supershift by anti-p50 antibody and partial reduction of binding using anti-p65 antibody also support the binding specificity (Fig. 4B). The rise and fall of NF-B binding activity is generally consistent with the levels of mdr1b mRNA as detected by RNase protection assay (Fig. 1B).
We also examined the effect of 2-AAF on the transactivation activity of NF-B using an artificial luciferase reporter pNFB-Luc, which contains five copies of NF-B consensus sequence. pFR-Luc, containing five copies of Gal4 consensus sequence, was used as a control. These two reporters were stably transfected into H4-II-E cells and exposed to 20 -100 M 2-AAF for 24 h. As shown in Fig. 4C, 2-AAF treatment led to a dose-dependent induction of pNFB-Luc, but not pFR-Luc. Taken together, these results suggest that 2-AAF is able to activate both DNA binding and transactivation activities of NF-B in H4-II-E cells.
2-AAF Induces Degradation of IB␤, but Not IB␣-TNF-␣ and interleukin-1␤, the two best characterized NF-B activators, activate NF-B through phosphorylation-dependent degradation of IBs, resulting in the release and subsequent nuclear translocation of NF-B (16,17). To investigate how 2-AAF activates NF-B, we performed Western blotting analysis with IB antibodies to examine whether 2-AAF causes degradation of IB proteins (Fig. 5A). Strikingly, no degradation of IB␣ was observed. In the same experiment, exposure of H4-II-E cells to 1 nM TNF-␣ for 20 min led to nearly complete degradation of IB␣, indicating that the phosphorylation and proteolysis mechanisms were functional in these cells. The same membrane was stripped and re-probed with IB␤ antibody. As shown in Fig. 5A, after exposure to 2-AAF, IB␤ became par-tially degraded at 2 h and still remained so after 5 h, following a time course compatible with that for the activation of NF-B DNA binding activity. In several independent experiments, we were able to obtain reproducible results.
To further strengthen the observation that IB␤ but not IB␣ is degraded in response to 2-AAF treatment, we transfected Flag-tagged IB␣ and IB␤ expressing plasmids into 293 cells and examined the effect of 2-AAF on their degradation, using anti-Flag M2 antibody. The reason for using 293 cells for transient transfection was because the transfection efficiency of H4-II-E cells was very poor (1-2%), whereas 293 generally gave high (70 -80%) transfection efficiency. Moreover, as human cells usually lack certain metabolic enzymes to convert 2-AAF into its active metabolite, we used 2-AAAF instead of 2-AAF in this experiment. As shown in Fig. 5B, 2-AAAF treatment had no effect on the stability of F-IB␣, but caused degradation of F-IB␤. The time course of the induced F-IB degradation was earlier than that found in Fig. 5A. This is probably due to the fact that different cell lines (H4-II-E versus 293), different carcinogens (2-AAF versus 2-AAAF), and different IBs (endogenous versus exogenously transfected) were used. These experiments suggest that the signal initiated by 2-AAF causes preferential degradation of IB␤.
2-AAF Induces mdr1b through Generating Intracellular ROS-NF-B has long been regarded as an important sensor of oxidative stress (18,19). Expression of NF-B can be induced by a variety of excellular influences, including growth factors, UV irradiation, heat shock, and anti-tumor drugs. Many of these inducers are also known for the induction of MDR gene expression (20 -23).
To determine whether oxidative stress plays an important role in up-regulation of mdr1b expression, we first applied a strong oxidant H 2 O 2 to H4-II-E cells stably transfected with mdr1b-Luc, pNFB-Luc, and pFR-Luc recombinants, respectively. Six hours after H 2 O 2 exposure, both mdr1b-Luc and pNFB-Luc were highly activated in a dose-dependent manner, but H 2 O 2 had no effect on pFR-Luc (Fig. 6A). These results suggest that the mdr1b promoter, which contains a NF-B binding site, is sensitive to oxidative stress.
To further demonstrate the role of oxidative stress in the regulation of NF-B-mediated mdr1b expression, we next in-vestigated whether antioxidants would down-regulate AAFinduced mdr1b expression using stably transfected cells containing mdr1b-Luc, pNFB-Luc, and pFR-Luc. N-Acetylcysteine (NAC) is a thiol reducing agent, which is commonly used as an anti-oxidant to block the ROS-induced stress. The transfected cells were treated with 100 M 2-AAF in the absence or presence of increasing concentrations of NAC for 24 h and luciferase activity was determined. As shown in Fig. 6B, NAC at concentrations of 1-15 nM was able to partially abolish the 2-AAF induced expression of mdr1b-Luc and pNFB-Luc, but 2-AAF and NAC had no effect on pFR-Luc. These results support the role of ROS in the regulation of mdr1b expression.
GSH is an important physiological antioxidant in mammalian cells and high level of GSH could be achieved by overexpressing the heavy subunit of ␥-glutamylcysteine synthetase, a rate-limiting enzyme of GSH synthesis. A ␥-glutamylcysteine synthetase overexpressing cell line H9 was established from H4-II-E cells and was previously shown to have elevated levels of intracellular GSH (15,24). We treated H9 and H4-II-E cells in parallel with different doses of 2-AAAF and compared their steady-state mdr1b mRNA levels by RNase protection assay. As indicated in Fig. 6C, the induction levels of mdr1b mRNA by 2-AAAF were significantly lower at each concentration in H9 cells than that in H4-II-E cells. In H4-II-E cells, 2-AAAF treatment induced mdr1b mRNA by up to 7.3-fold, however, in H9 cells the maximal induction was only 3-fold. Taken together, these results indicated that 2-AAF (or 2-AAAF) up-regulates mdr1b expression through generation of ROS and inhibition of intracellular ROS was able to block the induction of mdr1b mRNA by 2-AAF. substantiate that 2-AAF or 2-AAAF generate ROS in H4-II-E cells and to investigate whether overproduction of GSH suppresses the generation of ROS, we performed flow cytometric assay to measure intracellular ROS levels. Both H4-II-E and H9 cells were treated with either 100 M 2-AAF or 10 M 2-AAAF for time intervals ranging from 20 min to 6 h and ROS levels were measured using a non-fluorescent dye dichlorofluorescein diacetate, which becomes highly fluorescent upon oxidation by intracellular ROS. A representative set of histograms from untreated controls and 100 M 2-AAF 1 h treatments was shown in Fig. 7A. The average values of the mean fluorescence intensity from triplicate treatments were plotted in Fig. 7B. In H4-II-E cells, the ROS level was rapidly increased and reached a peak at ϳ1 h after treatment. Levels of ROS were then decreased, but reached a second peak at ϳ4 h. This two-wave profile was reproducible. The underlying mechanism of this ROS wave is unclear at present. In contrast, the increase of ROS in H9 cells was much less dramatic, with a maximum only 50% of that of H4-II-E cells. When 2-AAAF was used, we also obtained a similar result (data not shown), indicating that in terms of both mdr1b induction and ROS generation, 2-AAF and 2-AAAF are nondistinguishable.

2-AAF and 2-AAAF Increase Intracellular ROS and Overexpression of ␥-Glutamylcysteine Synthetase Reduces ROS-To
2-AAF Activates NF-B DNA Binding through Generating ROS-Since mdr1b is a NF-B-regulated gene and NF-B is an important oxidative stress responder, we then compared 2-AAF induced NF-B DNA binding activity in H4-II-E (Fig. 8A) and GSH overproducing H9 cells (Fig. 8B). Nuclear extracts were obtained from both cell lines treated with 100 M 2-AAF for different times and gel mobility shift assay was performed. 2-AAF activated NF-B DNA binding in both cell lines, but the maximal induction level in H9 cells were about 40% lower than that in H4-II-E cells. This result indicates that suppression of ROS was correlated with suppression of the 2-AAF induced activation of NF-B signaling.
2-AAF Activates IKK through Generating ROS-Phosphorylation and ubquitinization dependent degradation of IB proteins is one of the most common mechanisms through which various extracellular stimuli activate NF-B (25)(26)(27). IKK kinase complex has been shown to phosphorylate IBs in cells treated with these inducers. To address the questions whether IKK activation is involved in 2-AAF induction of mdr1b and whether suppression of ROS also down-regulates IKK kinase activity, we performed in vitro kinase assay to measure IKK activity in H4-II-E and H9 cells treated with 2-AAF. As shown in Fig. 9, treatment of H4-II-E cells with 2-AAF increased IKK activity 1.3-1.8-fold with a kinetics and induction level compatible with the NF-B DNA binding activity. In a parallel experiment, IKK activity remained unchanged after 2-AAF treatment in H9 cells. The specificity of the phosphorylation of recombinant GST-I␣ substrate was confirmed by the absence of phosphorylation of GST (panel C in Fig. 9). In addition, the induction of IKK activity in H4-II-E cells was not due to elevation of IKK␣ protein synthesis, since Western blotting analysis with IKK␣ antibody failed to reveal significant increase of IKK␣ protein. Thus, we conclude that induction of mdr1b expression by 2-AAF is mediated through the generation of ROS that activate IKK kinase activity and the downstream NF-B signaling.

Activation of NF-B Signaling and Induced mdr Gene Expression by 2-AAF in Rat
Hepatoma-Previous studies have demonstrated that mdr1b expression is frequently up-regulated in HCC developed by various hepatocarcinogenetic programs. However, the mechanisms of this activation are largely unknown. In this study, we studied mechanisms of mdr1b up-regulation in rat hepatoma cells induced by 2-AAF. We identified that the NF-B-binding site on the mdr1b promoter is necessary for the induction by 2-AAF. We further demonstrated that treatment of rat hepatoma cells with 2-AAF increases DNA binding activity of NF-B through activation of IKK, which degrades IB␤ but not IB␣. Our present study reveals a sequence of events involved in the activation of mdr1b by 2-AAF as shown in Fig. 10. This is the first demonstration that activation of mdr1b expression by a hepatocarcinogen is mediated by the NF-B signaling.
In most cases, NF-B activation involves phosphorylation and degradation of IB␣, which in turn releases NF-B and leads to the nuclear translocation of NF-B. However, we were not able to observe IB␣ degradation after 2-AAF treatment, but instead IB␤ was degraded. These observations were also supported using recombinant DNA containing Flag-tagged IB␣ and IB␤ in transfection assays followed by 2-AAF treatments. To the best of our knowledge, this is the first observation showing that activation of NF-B pathway is mediated by IB␤ but not IB␣ degradation.
Unlike IB␣, IB␤ itself is not regulated by NF-B and thus it is not rapidly resynthesized after degradation. For this reason, IB␤ degradation is thought to associate with prolonged activation of NF-B (14,28). In agreement with this notion, our data showed that 2-AAF caused a relatively slow and prolonged degradation of IB␤ and activation of NF-B DNA binding, in comparison with that caused by cytokines, e.g. TNF-␣. The activation of IKK in the 2-AAF-treated cells could be demonstrated by the in vitro kinase assay experiments, suggesting that the mechanisms of preferential degradation of IB␤ may lie downstream from the IKK activation. At present it is unclear why IB␤ is preferentially degraded, but possibilities can be offered: One possibility is that there may exist a 2-AAFinducible inhibitor which preferentially shields IB␣ from phosphorylation by IKK. Alternatively, an IB␤-specific kinase which preferentially recognizes and phosphorylates IB␤ but not IB␣ may be induced by 2-AAF. Recent studies have identified an IB␤-specific interacting protein (29). Moreover, activation of NF-B by multiple distinct IB kinase complexes has been demonstrated (30). There results suggest that there may be additional IB-interacting proteins and/or kinases involved in the activation of NF-B by 2-AAF. Another possibility is that both IB␣ and IB␤ are phosphorylated, but their subsequent proteolytic degradation is differentially regulated. Further studies on the kinetics and extent of IB␣ and IB␤ phosphorylation as well as their degradation will be helpful to differentiate these possibilities.
Recent studies have demonstrated that multiple pathways are responsible for activation of NF-B. In addition to the mechanism reported here which involves the degradation of IB␤, other alternative mechanisms may also play a role in the NF-B activation in the absence of IB␣ degradation. Beraud et al. (31) found that hypoxia, reoxygeneration, and the pervanadate-induced IB␣ phosphorylation at tyrosine 42 and this led to NF-B activation independent of IB␣ proteolysis. They observed that the regulatory subunit of phosphatidylinositol 3-kinase could recognize tyrosine-phosphorylated IB␣ and sequester it from binding to NF-B. Alternatively, several reports have shown that NF-B subunits p65 and p50 are phosphorylated and phosphorylation increases NF-B DNA binding activity or transactivation potential (32,33). Moreover, the degradation of other NF-B inhibitors including IB␥, IB⑀, and p105 may also contribute to NF-B activation by 2-AAF without degradation of IB␣. To define these possibilities, more detailed studies are warranted.
Redox Regulation of NF-B Signaling and mdr Gene Expression-We also presented evidence showing that ROS levels are elevated in H-4-II-E cells treated with 2-AAF. Induction of ROS is much reduced in the GSH overproducing H9 cells. These results are consistent with our previous report demonstrating that overproduced antioxidant GSH in cultured cells suppressed oxidative stress induced by cytotoxic agents, including TNF-␣, phorbol ester, and okadaic acid (24). Concomitantly, we found that induction of IKK activity, NF-B DNA binding activity, and mdr1b expression is retarded in the H9 cells as compared with the parental H-4-II-E cells. These results strongly suggest that, like MRP1 (15), expression of mdr1b can be regulated by redox conditions. The identification of IKK as an upstream redox sensor of NF-B activating signal by 2-AAF is consistent with those reported by Chen et al. (34) demonstrating that the activity of IB kinase ␤ (IKK␤) was significantly elevated in cells exposed to proxidant vanadate. However, our results may differ from those reported by Li and Karin (35). These investigators reported that induction of IKK activity by TNF-␣ (and therefore IB␣ phosphorylation/degradation) was not affected in cells treated with NAC (35).
Aside from the involvement of IKK as a target of redox regulation, oxidative stress may cause protein conformational changes. Particularly, oxidation of methionyl residues may render proteins susceptible to proteasomal degradation (36). In this regard, our observation that 2-AAF triggering ROS formation raises an additional possibility that ROS might modulate the degradation of IBs by direct oxidation. It is possible that IB␤, because of its minor difference in amino acid composition compared with IB␣, is more sensitive to oxidative alterations and thus prone to proteolysis. Further investigations on the redox control of NF-B signaling in different cell settings with different prooxidants, in combination with domain swapping between IB␣ and IB␤, may allow us to elucidate all these possibilities.
ROS are regarded as having carcinogenic potential and have been associated with tumor promotion (37,38). In addition, ROS are also pivotal factors in the genesis of heart disease (37). There is a fine balance between ROS and endogenous antioxidants, and any disturbance of this balance may cause cancer and heart diseases. For this reason, many natural or synthetic antioxidants have been used to prevent carcinogenesis and cardiovascular problems. Our results that carinogen 2-AAF causes increase of ROS and GSH blocks the 2-AAF-initiated ROS signaling, might have implications in the development of novel cancer therapeutic and preventive interventions. In fact, it has been reported that oral administration of the thiol NAC produced a significant decrease of mitochondrial DNA adduct, in the liver of 2-AAF-treated rats and in the lung and liver of rats exposed to cigarette smoke (39).
Activation of NF-B Signaling and Hepatocarcinogenesis-Liver cancers in experimental animals can be induced by many different protocols including chemical carcinogens, steroid hor-mones, dietary intervention, and viral infections (40). Because these different agents have different cellular targets and models of cytotoxicity, it is likely that multiple pathways are involved in the initial events but then converged during hepatocarcinogenesis. The consistent observations of mdr gene upregulation in the various hepatocarcinogenetic programs strongly suggest an overlapping, if not common, pathway between the induction of mdr gene expression and the development liver cancers. The discovery that NF-B signaling is involved in the activation of mdr1b expression in AAF-treated rat hepatoma cells raises an important scenario suggesting that NF-B may plays a role in liver cancer development as well. Several lines of evidence support the involvement of Rel/NF-B signaling in liver cancer. (i) The retroviral v-rel oncogene acutely induces tumors in birds and mammals (41). (ii) Overexpression or constitutive activation of the Rel/NF-B gene family has been noted in many human hematopoietic and solid tumors (42,43). (iii) NF-B signaling can be activated by a wide variety of stimuli, including genotoxins and nongenotoxins. As liver is the major detoxification reservoir of xenobiotics, many of these excellular stimuli are known to induce liver cancers in experimental animals. (iv) Evidence has accumulated that ROS plays an important role in hepatocarcinogenesis in animal models and suppression of ROS retards liver cancer progression in these models (37,44). Greater than 10-fold increases of mdr1b expression in Fisher rats can be induced by 2-AAF (4). On the other hand, liver neoplastic lesions can be chronically induced in these animals by this hepatocarcinogen. The results described in this article thus provide a molecular basis for further investigation on the roles of NF-B signaling in hepatocarcinogenesis and the induction of mdr1b gene expression in the process. These experiments are currently being investigated in this laboratory. These studies may eventually lead to a better understanding on the mechanisms of liver cancer development and the evolution of drug resistance in this devastating disease.