Tumor Necrosis Factor α Induces Spermidine/Spermine N1-Acetyltransferase through Nuclear Factor κBin Non-small Cell Lung Cancer Cells*

Tumor necrosis factor α (TNFα) is a potent pleiotropic cytokine produced by many cells in response to inflammatory stress. The molecular mechanisms responsible for the multiple biological activities of TNFα are due to its ability to activate multiple signal transduction pathways, including nuclear factor κB (NFκB), which plays critical roles in cell proliferation and survival. TNFα displays both apoptotic and antiapoptotic properties, depending on the nature of the stimulus and the activation status of certain signaling pathways. Here we show that TNFα can lead to the induction of NFκB signaling with a concomitant increase in spermidine/spermine N1-acetyltransferase (SSAT) expression in A549 and H157 non-small cell lung cancer cells. Induction of SSAT, a stress-inducible gene that encodes a rate-limiting polyamine catabolic enzyme, leads to lower intracellular polyamine contents and has been associated with decreased cell growth and increased apoptosis. Stable overexpression of a mutant, dominant negative IκBα protein led to the suppression of SSAT induction by TNFα in these cells, thereby substantiating a role of NFκB in the induction of SSAT by TNFα. SSAT promoter deletion constructs led to the identification of three potential NFκB response elements in the SSAT gene. Electromobility shift assays, chromatin immunoprecipitation experiments and mutational studies confirmed that two of the three NFκB response elements play an important role in the regulation of SSAT in response to TNFα. The results of these studies indicate that a common mediator of inflammation can lead to the induction of SSAT expression by activating the NFκB signaling pathway in non-small cell lung cancer cells.

Polyamines are aliphatic cations present in all cells, whose levels are intricately controlled by their transport and metabolic enzymes. Spermidine/spermine N 1 -acetyltransferase (SSAT) 2 is a rate-limiting step in polyamine catabolism, which catalyzes the transfer of the acetyl group from acetyl-CoA to the N 1 position of spermidine or spermine and has a predominant role in the regulation of intracellular polyamine concentrations in mammalian cells (1,2). Decreases in polyamines have been shown to promote decreased growth or apoptosis (3)(4)(5)(6), depending on the cell type and the particular stimulus, suggesting a complex interaction between polyamines, cell growth, and cell death. Therefore, although polyamines are required for cell growth and differentiation, SSAT is thought to prevent overaccumulation of the higher polyamines from becoming toxic to the cell and may play a role in reducing the growth rate by decreasing intracellular polyamines.
Recently, considerable attention has been paid to SSAT as a target for cancer chemotherapy. SSAT activity is highly regulated and is induced rapidly in response to a number of stimuli, including polyamines, polyamine analogues, hormones, physiological stimuli, drugs, and toxic agents (1). It has been shown that the regulation of SSAT by the natural polyamines and the anti-tumor polyamine analogues is through the polyamine response element (7). Further, the superinduction of SSAT by polyamine analogues has been implicated in the cell type-specific cytotoxic response of several important human tumors (8 -13). Various nonsteroidal anti-inflammatory drugs like aspirin, sulindac, and indomethacin and chemotherapy drugs like 5-flurouracil and oxaliplatin have been shown to induce SSAT expression in various cancer cell types (14 -17). Sulindac induces SSAT by inducing peroxisomal proliferator-activated receptors, which, once activated, can bind to the peroxisomal proliferator-activated receptor response elements in the SSAT 5Ј promoter region (18). Recently, we have shown that aspirin can also induce SSAT expression in colon cancer cells, partly by activating the NFB signaling pathway, which leads to the binding of NFB complexes to the NFB response elements in the SSAT 5Ј promoter region (19).
Tumor necrosis factor ␣ (TNF␣) is a potent pleiotropic cytokine and is a major mediator of inflammation with multiple biological functions (20). One of the molecular mechanisms responsible for the biological activities of TNF␣ is the ability to activate nuclear factor B (NFB), which plays critical roles in cell proliferation and survival (21,22). In mammalian cells, five members of the NFB/Rel family are known: NFB1 (p50 or its precursor p105), NFB2 (p52 or its precursor p100), c-Rel, RelA (p65), and RelB (23,24). Heterodimers composed of p65-p50 are the most abundant active form of NFB in most cell types. However, NFB can consist of other homo-and heterodimers that have different abilities to activate target genes (25,26). In quiescent cells, NFB resides in the cytosol in a latent form bound to IB␣. Stimulation of the cell with TNF␣ triggers a series of signaling events that ultimately leads to the phosphorylation and the proteolytic degradation of IB␣ and activation of NFB. The phosphorylation of IB␣ is elicited by an IB␣ kinase (IKK), which can be activated by mitogen-activated protein kinase (27,28), whereas the proteolysis of IB␣ is mediated by the ubiquitin-proteasome pathway of protein degradation. The degradation of IB␣ triggers the translocation of NFB from the cytoplasm to the nucleus, where it regulates the expression of multiple genes.
In this study, we examined the effects of TNF␣ treatment on SSAT expression in representative human non-small cell lung cancer cells and identified the specific NFB response elements responsible for the regulation of SSAT by TNF␣.

EXPERIMENTAL PROCEDURES
Materials-All cell culture reagents, DNA-modifying enzymes, TRIzol reagent (total RNA isolation reagent), and Lipofectamine reagent were purchased from Invitrogen (Carlsbad, CA). All primers and oligonucleotides were custom made by Invitrogen. Recombinant human TNF␣ was obtained from R&D Systems, Inc. (Minneapolis, MN). IB␣ antibody is from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA), and p65 antibody is from Upstate (Waltham, MA). Paclitaxel (Taxol) was obtained from Sigma-Aldrich.
Cell Culture-The non-small cell cancer lines NCI-A549 (adenocarcinoma) and NCI-H157 (squamous) were purchased from American Type Culture Collection (Manassas, VA) and maintained in RPMI 1640 supplemented with 10% (v/v) bovine calf serum and 100 units/ml each of penicillin and streptomycin. Cultures were maintained at 37°C in humidified atmosphere of 5% CO 2 .
Plasmids-Full-SSAT-luc, having a 3.493-kb-long 5Ј-flanking sequence of the human SSAT gene, was cloned into a promoterless pGL2-basic (Promega, Madison, WI) as previously reported (29). A series of smaller SSAT promoter constructs were made from Full-SSAT-luc, using PCR and subcloned into pGL2-basic vector. 197-SSAT-luc, having 283 nucleotides of the 5Ј-flanking region of the SSAT promoter; 358-SSAT-luc, having 441 nucleotides of the 5Ј-flanking region of the SSAT promoter; and 659-SSAT-luc, having 740 nucleotides of the 5Ј-flanking region of the SSAT promoter, were made from Full-SSAT-luc using PCR and subcloned into pGL2-basic vector. NFB 2 -Luc reporter, having two NFB response elements; dNFB 2 -Luc reporter, in which both the NFB response elements have been deleted; and dominant negative-IB␣ (DN/IB) and its control (Ctrl/IB) plasmids were obtained from Dr. Nancy Davidson (Johns Hopkins University, Baltimore, MD). The DN/IB plasmid has a deletion of 36 NH 2 -terminal amino acids containing Ser 32 and Ser 36 phosphorylation sites, which was cloned into the pcDNA3.1 mammalian expression plasmid vector (Invitrogen). mNFB1-SSAT-luc, mNFB3-SSAT-luc, and dmNFB1/3-SSAT-luc were made by mutating the NFB-1, NFB-3, and both the NFB-1 and -3 sites, respectively, in the Full-SSATluc using the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA). Mutations were done by placing four base substitutions in the respective NFB sites to achieve the same mutated sequences as the probes used for the electromobility shift assays.
Stable Transfection-Exponentially growing cells were plated at 1 ϫ 10 6 cells/100-mm plate and cultured in normal medium for 24 h and then transfected using 50 l of Lipofectamine reagent with 10 g of the DNA plasmid. After 6 h of incubation, with Lipofectamine-DNA complex, cells were supplemented with complete medium having 20% bovine calf serum and 2% penicillin and streptomycin and grown overnight, after which the medium was removed, cells were refed complete medium, and cells were allowed to recover for 24 h prior to selection in 0.4 mg/ml G418. Cells were maintained in selection medium until a stable population was achieved. Individual colonies were then isolated, expanded, and used for screening and experiments.
Transient Transfections-Transient transfections were performed using Lipofectamine reagent according to the supplied protocol. Briefly, 5 ϫ 10 5 cells were seeded in a 6-well plate and cultured in normal medium for 24 h. Cells in each well were transfected with 1 g of firefly luciferase reporter construct along with 0.2 g of pCMV-␤-galalactosidase expression plasmid, used as a control for transfectional efficiency. After 6 h, cells were supplemented with complete medium having 20% bovine calf serum and 2% penicillin and streptomycin and grown overnight. 15 h later, the medium was removed, and cells were refed with the medium containing 10 ng/ml TNF␣ for the indicated amount of time. All transfections were performed in triplicates unless otherwise indicated. Transfected cells were washed once with phosphate-buffered saline and lysed, and luciferase activities were measured using 10 l of cell extract and 50 l of luciferase reagent (Promega). ␤-Galactosidase activity was measured using the ␤-galalactosidase assay kit (Invitrogen) according to the supplied protocol.
Quantitative Real Time PCR-Cells were seeded, grown overnight, and then treated as indicated. Total cellular RNA was extracted using TRIzol reagent and RNA isolation according to the method developed by Chomczynski and Sacchi (30). 3 g of total RNA was treated with RNase-free DNase I (Roche Applied Science) for 30 min at 37°C before reverse transcription with Moloney murine leukemia virus reverse transcriptase (Promega, Madison, WI) as described by the manufacturer. The forward and reverse primers were used at 10 M. The resulting cDNA was subjected to quantitative real time PCR. The DNAintercalating SyBr green reagent (Quantitect; Qiagen) and melting temperature profiles were used for detection of the PCR product. Agarose gel electrophoresis confirmed the presence of a single PCR product. The glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as an internal standard. The experimental and GAPDH PCRs were done in separate tubes in triplicates in the MyiQ single color real-time PCR machine (Bio-Rad), and the average threshold cycle (C T ) for the triplicate was used in subsequent calculations. SSAT cDNA (180 base pairs) was amplified using the following primers (Invitrogen): forward, 5Ј-GGATCAAAATTCTGAAGAAT-3Ј; reverse, 5Ј-ACCCTCTTCACTGGACAGATC-3Ј. As a loading control, GAPDH cDNA (188 base pairs) was also amplified using the following primers: forward, 5Ј-GAAGGTGAAGGTCGGA-GTC-3Ј; reverse, 5Ј-GAAGATGGTGATGGGATTTC-3Ј.
Immunoblotting-Total cell extracts were obtained by lysing cells on ice in radioimmune precipitation buffer (phosphatebuffered saline, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, 30 g/ml aprotinin, 100 M sodium orthovanadate, 10 g/ml phenylmethylsulfonyl fluoride) and centrifuging for 20 min at 4°C. 30 g of total protein was loaded per lane and separated on a 10% SDS-polyacrylamide gels for IB␣ immunoblotting. Nuclear protein was isolated, using the NE-PER nuclear extraction reagents (Pierce). 50 g of nuclear protein was loaded per lane and separated on 12% SDS-polyacrylamide gels for p65 immunoblot. The proteins were transferred electrophoretically to immunoblot polyvinylidene difluoride membrane for 1 h. Blots were blocked in Blotto A (5% w/v nonfat dry milk in 20 mM Tris, pH 7.6, 137mM NaCl, 0.1% Tween 20) and probed for IB␣ and p65 proteins. Blots were stripped and redetected with ␤-actin (Santa Cruz Biotechnology) antibody as a loading control. Western blot analyses were repeated at least three times, and a representative blot was chosen for presentation.
In Taxol experiments, Western blot results were quantified using the LICOR immunofluorescence system (LI-COR Biosciences, Lincoln, NE). Briefly, membranes were blocked for 1 h in Odyssey blocking buffer, per the manufacturer's instructions. Rabbit anti-IB␣ and mouse anti-actin (loading control) (Santa Cruz Biotechnology) primary antibodies were then added together at dilutions of 1:1000 and 1:1500, respectively, with 0.1% Tween 20 in blocking buffer for 1 h at room temperature. Following washes with PBS-Tween, blots were incubated with appropriate fluorescent dye-conjugated secondary antibodies (1:4000 each, 0.1% Tween 20, in blocking buffer, protected from light, for 45 min), which allowed detection and quantification of each protein using an Odyssey infrared detection system and software (LI-COR).
Gel Electromobility Shift Assays-Nuclear extracts were prepared from A549 cells as described previously (31). To study the binding of NFB complexes to the putative NFB sites, doublestranded oligonucleotides for each of the three putative NFB sites in the SSAT 5Ј sequence were 32 P-labeled with polynucleotide kinase (Promega, Madison, WI). The oligonucleotide containing the first putative NFB site spanned from Ϫ304 to Ϫ280 of the SSAT 5Ј sequence and had the sequence NFBwild-1 (w) (5Ј-GCTGCAGAGGGAATTACCTTCTTT-3Ј), whereas the corresponding mutant NFB-mut-1 (m) had the sequence 5Ј-GCTGCAGAtaGAATgcCCTTCTTT-3Ј. The oligonucleotide spanning the third putative NFB site spanned from Ϫ1751 to Ϫ1728 of the SSAT 5Ј sequence and had the sequence NFB-wild-3 (w 3 ) (5Ј-CCTAGGGGGATTCCACG-GATCCT-3Ј, whereas the corresponding mutant NFB-mut-3 (m 3 ) had the sequence 5Ј-CCTAGGtGATTCagCGGATCCT-3Ј. The putative NFB sequence is underlined, and the mutated bases are shown in lowercase letters. A 10-l reaction containing 10 g of nuclear extract was incubated for 15 min at room temperature in a buffer containing 50 mM Tris-HCl (pH 7.5), 5 mM MgCl 2 , 2.5 mM EDTA, 250 mM NaCl, 2.5 mM dithiothreitol, 20% glycerol, and 1 g of poly(dI-dC). Following this, 30,000 cpm of the labeled probe was added, and the reaction mixture was incubated for 30 min at room temperature. The DNAprotein complexes were resolved from the free probe by electrophoresis at 4°C on a 5% polyacrylamide gel in 0.75ϫ TBE buffer, pH 8. Band density was quantified using the Typhoon 8600 PhosphorImager and ImageQuant software (Amersham Biosciences).
Chromatin Immunoprecipitation (ChIP)-Chromatin immunoprecipitation assays were performed using the commercially available ChIP kit (Upstate Cell Signaling) with some modifications. A549 cells were seeded at 2 ϫ 10 6 /culture dish and grown overnight. On the following day, they were treated with 10 ng/ml TNF␣ for increasing times. After treatment, the cells were treated with 1% formaldehyde by adding 0.4 ml of 37% formaldehyde directly to 15 ml of culture medium for 15 min at 37°C. The cells were washed twice with cold phosphate-buffered saline containing protease inhibitors (leupeptin, phenylmethylsulfonyl fluoride, and aprotinin) and then suspended in 0.7 ml of SDS-lysis buffer (50 mM Tris-Cl, pH 8.1, containing 1% SDS and 10 mM EDTA) plus protease inhibitors. Twenty 10-s sonication pulses with 30-s intervals were done using the Branson Sonifier 250 (cycle 40%, output ϭ 1.5/2) to shear chromatin to Ͻ1000-bp fragments. The effectiveness of shearing was confirmed by incubating a 100-l aliquot of the extract with 10 l of 5 M NaCl at 65°C for 4 h (to reverse cross-links) and subsequently subjecting it to electrophoresis on a 1% agarose gel. The sonicated sample was centrifuged at 13,000 rpm for 10 min. 0.2 ml of the cell supernatant was diluted by adding 1.8 ml of ChIP dilution buffer (16.7 mM Tris-Cl, pH 8.1, containing 0.01% SDS, 1.1% Triton X-100, 167 mM NaCl, and 1.2 mM EDTA) plus protease inhibitors. 50 l of this sample was stored for later PCR analysis as the input sample control. To reduce nonspecific background, the remaining amount (1950 l) was precleared with 75 l of salmon sperm DNA/proteinagarose 50% slurry (Upstate Cell Signaling, NY) for 1 h at 4°C with agitation. The cleared supernatant was divided into 0.4-ml aliquots and incubated with either no antibody, a nonspecific antibody (rabbit IgG), 4 l of anti-p50, or 4 l of anti-p65 and incubated overnight at 4°C with rotation, followed by the addition of 30 l of salmon sperm DNA/protein-agarose 50% slurry and incubated for an additional 3 h.
SSAT Enzyme Activity Determination-For enzyme activities, cells were grown overnight and then treated with TNF␣ for different times. Cells were harvested after treatment and washed in cold phosphate-buffered saline. The radiochemical assay of the N 1 -SSAT activity was performed by estimation of labeled N 1 -acetylspermidine synthesized from [ 14 C]acetyl-coenzyme A and unlabeled spermidine, as previously described (8). Fold change was calculated by dividing the enzyme activity for the sample by the vehicle. The enzyme assays were performed in triplicate.
Polyamine Analysis-Cell extracts were prepared in 0.1 N HCl (4 ϫ 10 7 cells/900 l). After sonication, the preparation was adjusted to 0.2 N HClO 4 , and the supernatant was analyzed by reverse-phase high performance liquid chromatography with 1,7-diaminoheptane as an internal standard (32). Protein was determined by the BCA assay (33).
Statistical Analysis-All transient transfection experiments were performed in triplicates and were repeated at least three times. Quantitative real time PCR and ChIP assays were done at least three times. Representative experiments or mean values Ϯ S.D. are shown. Statistical differences were determined by Student's t test. A p value of Ͻ0.05 was considered significant.

TNF␣ Induces SSAT mRNA and SSAT Enzyme Activity and Decreases Intracellular Polyamine Contents in Non-small Cell
Lung Cancer (NSCLC) Cells-We have previously demonstrated the presence of functional NFB response elements in the SSAT gene (34). Since SSAT has been implicated as a stress response gene (1), we sought to determine whether TNF␣, acting through NFB, altered the expression of SSAT in the NSCLC cells. TNF␣, at 10 ng/ml, induces SSAT mRNA expression in both A549 and H157 NSCLC cells within 30 min of treatment (Fig. 1A). SSAT is an enzyme that acetylates both spermidine and spermine. The acetylated products can then be either exported or degraded by the action of polyamine oxidase, thereby leading to a decrease in intracellular polyamine content. Therefore, we next determined whether the increase in SSAT mRNA by TNF␣ treatment was accompanied by an increase in SSAT enzyme activity and a decrease in the polyamine content. TNF␣ treatment led to an increase in SSAT enzyme activity in these NSCLC cells (Fig. 1B) and a modest reduction in the intracellular polyamine content (Fig. 2). The decrease in polyamine content is consistent with the -fold induction in the SSAT enzyme activity by TNF␣ in the NSCLC cells.
TNF␣ Activates NFB Signaling in NSCLC Cells-One of the major signaling pathways by which TNF␣ is known to work is by activating NFB signaling in the cell. In unstimulated cells, NFB resides in the cytosol in a latent form bound to IB␣. Stimulation of the cell with TNF␣ triggers a series of signaling events that ultimately leads to the phosphorylation of IB␣ by IKK, which then becomes a target of ubiquitin-proteosomal degradation. The phosphorylation and subsequent degradation of IB␣ triggers the translocation of NFB complex (p65-p50) protein from the cytoplasm to the nucleus, where it regulates the expression of multiple genes. The finding that TNF␣ induces SSAT expression and the fact that TNF␣ functions predominantly via NFB signaling led us to determine the functionality of the NFB signaling in these NSCLC cells. Treatment of A549 cells with TNF␣ produced a rapid reduction in the IB␣ protein (Fig. 3A), with a concomitant rapid increase in nuclear translocation of the p65 protein in the A549 cells (Fig.  3B). Similar effects on IB␣ and p65 protein were observed in the H157 cells (data not shown). Transient transfection experiments were performed with an NFB reporter construct to determine whether translocation of p65 into the cell nucleus was accompanied by a functional increase in the NFB-responsive genes. As shown in Fig. 3C, TNF␣ treatment led to an increase in the luciferase activity from the NFB 2 -luc plasmid in both A549 and H157 cells but had no effect on the control plasmid lacking NFB response elements, suggesting a functional NFB signaling mechanism in these NSCLC cells.
Attenuation of TNF␣-induced SSAT mRNA by Dominant Negative IB␣-To verify that TNF␣ induction of SSAT expression was indeed mediated by the nuclear translocation of the NFB complex, both A549 and H157 cells were stably transfected with the dominant negative IB␣ (DN/IB-A549 and DN/IB-H157)-overexpressing plasmid. DN/IB plasmid lacks the NH 2 -terminal 36 amino acids, thereby removing the Ser 32 and Ser 36 residues that are required for its phosphorylation by IKK and subsequent proteosomal degradation, but maintains the ability to bind NFB complex in the cytosol. Two A549 clones with the highest expression of DN/IB␣ protein (Fig.  4A) were used to study the effects of TNF␣ treatment on NFB and SSAT activation. Two clones from the vector-transfected A549 cells (Ctrl/IB-A549) were used as the control. Treatment with TNF␣ resulted in translocation of p65 into the nucleus in the Ctrl/IB clones, but the dominant negative IB␣ overexpression completely blocked the nuclear translocation of the p65 protein in either vehicle-or TNF␣-treated cells, suggesting the inability of TNF␣ to activate the NFB signaling in the DN/IB-A549 cells (Fig. 4B). Further, DN/IB overexpression either abolished (as in clone 1) or attenuated (as in clone 2) SSAT mRNA induction by TNF␣ in the A549 cells (Fig. 4C). Treatment with TNF␣ also resulted in p65 translocation and induced SSAT mRNA in the Ctrl/IB-H157 clones, but the dominant negative IB␣ blocked both the nuclear translocation of p65 and the induction of SSAT in the DN/IB-overexpressing H157 clones (Fig. 4D).
Taxol Induces SSAT by an NFB-dependent Pathway in A549 Cells-To determine the effects of activating NFB by an agent other than TNF␣ on the expression of SSAT in the A549 cells, the effects of Taxol exposure were examined. Taxol has been previously demonstrated to induce NFB activation through  phosphorylation of IB␣, thereby leading to its degradation, in various cancer cells (35)(36)(37)(38)(39). Further, Taxol has been shown to activate NFB signaling in the A549 cells (40). At both 10 and 50 nM, Taxol reduced IB␣ protein to 0.52 Ϯ 0.11 and 0.54 Ϯ 0.08, respectively, of control. Further, these concentrations of Taxol led to the induction of SSAT mRNA in both A549 parental and Ctrl/IB-A549 cells, but not in the DN/IB-A549 cells (Fig. 5). These results are consistent with the hypothesis that NFB plays a role in the regulation of SSAT, whether the source of signaling is from TNF␣ or the cytotoxic agent, Taxol.
Mapping of TNF␣-responsive Elements in the SSAT Promoter-The inability of TNF␣ to induce SSAT expression in DN/IB-overexpressing cells is consistent with the hypothesis that NFB plays an important role in the regulation of SSAT by TNF␣ in the NSCLC cells. Three putative NFB response elements, NFB-1 (w) at Ϫ286, NFB-2 (w 2 ) at Ϫ594, and NFB-3 (w 3 ) at Ϫ1735 respective to the transcription start site, have been identified in the SSAT 5Ј promoter (19). In order to determine whether these NFB response elements play a role in the induction of SSAT by TNF␣, the 5Ј-flanking sequences of SSAT were tested for their ability to mediate TNF␣-induced transcription of a reporter gene.   fold but did not induce the 197-SSAT-luc and the control pGL2-basic vector (Fig. 6B). TNF␣ did not induce the 659-SSAT-luc reporter construct, which could be due to the presence of inhibitory cis-acting sequences in the stretch of the SSAT promoter, which are not present in the 358-SSATluc promoter construct. These data suggest a role for NFB-1 and NFB-3 in the induction of SSAT by TNF␣ in the NSCLC cells.
TNF␣ Induces NFB Binding to Specific NFB Response Elements in the SSAT Promoter-Electrophoretic mobility shift assays were performed to determine whether TNF␣ exposure induces the binding of NFB complexes to the NFB response elements in the SSAT promoter and produce the observed increase in SSAT expression in the NSCLC cells. Electrophoretic mobility shift assays were performed using the nuclear extract from the A549 cells, which had been treated with TNF␣ for increasing times, and double-stranded oligonucleotides containing the putative wild-type NFB sequences. As shown in Fig. 7A, we observed binding of nuclear proteins to the NFB-1 (w) probe. The presence of two bands suggested two different protein complexes that are bound to the (w) probe. TNF␣ treatment led to a slight increase in the binding of the complexes to the (w) probe as compared with the vehicletreated nuclear extract, suggested by an increase in the upper band intensity. This binding was specific, since binding to the NFB-1 mutant (m) sequence was not observed, and the binding to the labeled wild type probe could be competed by cold wild probe (w). Incubation of nuclear extracts with the labeled probe and an antibody to the p50 subunit of the NFB complex led to the generation of a supershift band, suggesting the presence   30,000 cpm of the 32 P-labeled oligonucleotides containing the NFB-1 wild-type (w) or mutant (m) sequence were incubated with 10 g of A549 nuclear extract from cells treated with either water or TNF␣ for 15 min, 30 min, and 1 h. The cold wildtype probes were used at a 100-fold molar excess for the competition reaction. The supershift control reaction was done with a p50 antibody and labeled probe. The shifted and the supershifted (SS) bands are labeled. B, 30,000 cpm of the 32 P-labeled oligonucleotides containing the NFB wildtype (w 3 ) or mutant (m 3 ) sequence were incubated with 10 g of A549 nuclear extract from cells treated with either water or TNF␣ for 15 min, 30 min, and 1 h. The cold wild-type probes were used at a 100-fold molar excess for the competition reaction. The supershift control reaction was done with a p50 antibody and labeled probe. The shifted and the supershifted bands are labeled.
of p50 protein in the complexes bound to the NFB-1 sequences. This supershift was observed in the absence of the TNF␣ treatment, indicating the presence of a p50-containing complex at the NFB-1 sequence. Using other putative NFB sequences, we found no significant binding of the TNF␣-treated A549 nuclear extract with the NFB-2 (w 2 ) sequence (data not shown) but found that there was binding of NFB complexes to the NFB-3 (w 3 ) sequence (Fig. 7B). This binding was increased by TNF␣ treatment and was specific, since it could be competed by cold (w 3 ) probe and supershifted in the presence of a p50 antibody. No shift was observed when the NFB-3 mutant (m 3 ) sequence was used as a labeled probe.
To confirm the binding of NFB protein complexes on the SSAT promoter in response to TNF␣ in situ, ChIP assays were performed for all three NFB response elements in the SSAT promoter. A549 cells were treated with TNF␣ for increasing times, and then the nuclear protein-DNA complexes were cross-linked and then precipitated with either p50 or p65 antibodies. PCR amplification of the DNA present in the anti-p50 and anti-p65 chromatin immunoprecipitate demonstrated that p50 protein is bound to the NFB-1 sequence in the unstimulated cells, and TNF␣ stimulation leads to the binding of the p65 protein to the NFB-1 sequence (Fig. 8A). ChIP assays done with the NFB-2 sequence in the SSAT promoter showed that p50 protein, but not p65, is bound to this site with or without TNF␣ treatment (Fig. 8B). ChIP assays done with the NFB-3 sequence in the SSAT promoter showed that neither p50 nor p65 proteins are bound to this site in the unstimulated state, and TNF␣ stimulation leads to the binding of both p50 and p65 proteins to the NFB-3 sequence (Fig. 8C).
To further evaluate the roles of NFB-1 and NFB-3 in the regulation of SSAT by TNF␣, we introduced mutations in either or both of the NFB-1 and NFB-3 sites in the Full-SSAT-luc promoter constructs. The mutated plasmids were then transfected into the A549 cells and then treated with TNF␣ for 1 h. Expression of constructs containing the wild type NFB-3 but mutant NFB-1 sequences (mNFB1-SSAT-luc) were induced to a greater level than the constructs containing a mutant NFB-3 and wild type NFB-1 sequences (mNFB3-SSAT-luc) (Fig. 9). This suggests a greater role for NFB-3 in the TNF␣-induced SSAT expression. Expression of constructs containing both mutant NFB-1 and NFB-3 sequences but wild type NFB-2 (dmNFB1/3-SSAT-luc) was not induced by TNF␣ exposure, indicating that both NFB-1 and NFB-3 play a role in the maximal response to TNF␣ exposure in the NSCLC cells. These results also confirm that NFB-2 is not capable of mediating the TNF␣ response, consistent with ChIP results indicating that p65 is not bound to NFB-2 in TNF␣-stimulated cells (Fig. 8C).

DISCUSSION
TNF␣ is a potent pleiotropic proinflammatory cytokine produced by many different cells in response to infection and inflammatory stress. TNF␣ has been shown to have paradoxical roles in the evolution and treatment of malignant disease (41,42). Depending on the source, cell type, and level of TNF␣ produced, it can lead to either cell death or cell survival. Activation of NFB has been associated with inflammation, increased cel-  lular proliferation, and decreased programmed cell death, but there is also evidence for the role of NFB activation in leading to cell death and apoptosis (43,44). The exact action of the NFB activation in a cell depends on the cell type, the type of activating stimuli, and the ratio of different genes induced. Aspirin, a nonsteroidal anti-inflammatory drug, can lead to the activation of NFB in Caco-2 cells, which leads to SSAT induction in these cells (34).
It is demonstrated here that SSAT is one of the genes regulated by TNF␣ in non-small cell lung cancer cells. Treatment of two non-small cell lung cancer cell lines, A549 and H157, with TNF␣ leads to a rapid increase in the SSAT mRNA expression and enzyme activity and to a decrease in the polyamine content. SSAT activity leads to increased acetylation of spermidine and spermine, which can be then be back-converted via the polyamine oxidase pathway or exported from the cells. SSAT has been implicated as a stress response gene whose expression can be induced to a substantial level by endogenous and environmental factors, including a variety of hormones and growth factors, toxic agents, and the polyamines themselves (1). SSAT overexpression has also been implicated in adaptive responses to environmental stress (9), hypoxia and nutrient depletion (13), polyamine analogues (8, 10 -12, 46), and exposure to nonsteroidal anti-inflammatory drugs (16,48). Our data suggest that induction of SSAT by TNF␣ may provide a mechanism by which cells respond to inflammatory stress by reducing intracellular polyamines and slowing growth.
Superinduction of SSAT by specific agents has been shown to be associated with anti-tumor activity in several important human solid tumor models (12,49), with the NSCLCs typically responding to the polyamine analogue treatment with a significant induction of SSAT and subsequent cell death (8,11,50,51). In specific instances, the superinduction of SSAT results in production H 2 O 2 from the constitutive activity of the acetylpolyamine oxidase, polyamine oxidase. The H 2 O 2 thus produced can lead to DNA damage and ultimately cell death (52), suggesting another mechanism by which TNF␣ can lead to apoptosis in lung cancer cells.
Interestingly, it has been proposed that polyamines can directly modulate NFB binding in various cell types. Currently, the data are conflicting as to the precise role of polyamines in regulating NFB activation and DNA binding (53)(54)(55)(56)(57). Our results suggest that a stress stimulus, like TNF␣, leads to a rapid induction in SSAT expression via NFB activation, leading to a reduction in polyamine content. It is possible that this rapid polyamine depletion inhibits NFB DNA binding activity, thereby directly providing feedback regulation of SSAT expression. More experiments will be necessary to determine whether such a feedback mechanism exists in NSCLC cells.
The molecular mechanisms responsible for the multiple biological activities of TNF␣ are due to their ability to activate multiple signal transduction pathways, including those involving extracellular signal-regulated kinase, other mitogen-activated protein kinases, and NFB (21,22). In the systems reported here, the data are most consistent with TNF␣ acting through NFB. Exposure of the NSCLC lines to TNF␣ reduces IB␣ protein, increases nuclear p65 protein, and induces the expression of reporter constructs containing specific NFB response elements. Additional evidence that TNF␣ is acting through NFB is provided by the observed significant reduction in the ability of TNF␣ to induce SSAT in the DN/IB-A549 and DN/IB-H157 cells. It should be noted, however, that the expression of the dominant negative IB does not completely block SSAT induction by TNF␣ in DN/IB-A549 cells, suggesting that TNF␣ may also function through NFB-independent signaling pathways (58 -62).
Taxol has been shown to activate NFB signaling by increasing the phosphorylation of IB␣ in various cell types. This activation of NFB has also been associated with resistance to Taxol in these cells (35,38). Taxol was used here as another NFB activator to study the role of NFB activation in the induction of SSAT in these cells, since Taxol has been specifically demonstrated to activate NFB in the A549 cells (40). We found that Taxol treatment produced an approximately 50% decrease in the IB␣ protein, thereby leading to the activation of NFB and induced SSAT expression in A549 cells. Overexpression of DN/IB significantly attenuated the induction of SSAT observed in the Ctrl/IB-A549 cells. Although Taxol treatment does not result in complete degradation of IB␣, the suppression of SSAT induction in the presence of DN/IB indicates a central role for NFB activation in the induction of SSAT in response to Taxol.
NFB complexes are present as homo-or heterodimers of p50 and p65 proteins, which can then bind to specific NFB response elements in NFB-regulated genes. Three NFB response elements in the SSAT promoter have previously been identified (34). The results presented here from electrophoretic mobility shift assays, reporter constructs, and ChIP experiments indicate that two sites, NFB-1 at Ϫ286 and NFB-3 at Ϫ1735, but not NFB-2 at Ϫ594, are responsible for the TNF␣ response. Mutating either the NFB-1 or NFB-3 response elements decreased the induction of SSAT reporter constructs by TNF␣, and mutations in both of the NFB response elements were required to completely abolish the induction of SSAT reporter constructs by TNF␣ in A549 cells. Further, constructs having mutant NFB-1 but wild type NFB-3 showed higher SSAT promoter activity after TNF␣ treatment than the constructs having mutant NFB-3 but wild type NFB-1 sequences. These data suggest that NFB-3 may play a greater role than NFB-1 in regulation of SSAT by TNF␣ in A549 cells. This preference for NFB-3 over the NFB-1 could be due to the 5-base difference between the two response elements or due to the different contacts made by the transcriptional activation complex because of the ϳ1.5-kb distance between them. Further experiments will be required to characterize the exact role of NFB-3 and NFB-1 in the actions of TNF␣ on the SSAT promoter.
It is important to note that the chromatin immunoprecipitation experiments were particularly informative, demonstrating that p50 was bound to the NFB-1 and NFB-2 sequences with or without TNF␣ treatment. Of the NFB/Rel proteins, only p65 and c-Rel have potent transcriptional activation domains, whereas p50 protein lacks these domains. Hence it is thought that the binding of p50 homodimers functions as a transcriptional repressor (47,63,64), as compared with the p50-p65 heterodimers, which can function as transcriptional activators. The data presented here suggest that in the basal uninduced state, NFB-1 and NFB-2 are bound by p50 homodimers, thereby keeping the SSAT expression level low, but after TNF␣ treatment (or stress stimuli), there is a change in the promoterbound NFB complexes such that the p50 homodimers are replaced by p50-p65 heterodimers at the NFB-1 site and a recruitment of p50-p65 heterodimers at the NFB-3 site, which together lead to the induction of SSAT expression.
The discovery that the expression of SSAT is regulated by a common mediator of inflammation, TNF␣, is highly significant both in the context of understanding the regulation of polyamine homeostasis and in understanding tumor cell responsiveness to various anti-tumor strategies. Depending on the extent, the induction of SSAT has been linked to both growth inhibition and cell death (10,11,46). Thus, TNF␣ stimulation of SSAT may, in some cases, allow cells to reduce their growth rate when exposed to a stressful environment, and the observed cytotoxic drug responses may be an exaggerated extension of this normal response. Although more studies will be necessary to test this possibility, it is intriguing to consider the potential cross-talk that may occur between the polyamine response element (7,45), which is thought to mediate SSAT induction produced by cytotoxic polyamine analogues, and NFB and how such cross-talk might be exploited for therapeutic advantage.