Acrolein Causes Inhibitor κB-independent Decreases in Nuclear Factor κB Activation in Human Lung Adenocarcinoma (A549) Cells*

Acrolein is a highly electrophilic α,β-unsaturated aldehyde to which humans are exposed in various situations. In the present study, the effects of sublethal doses of acrolein on nuclear factor κB (NF-κB) activation in A549 human lung adenocarcinoma cells were investigated. Immediately following a 30-min exposure to 45 fmol of acrolein/cell, glutathione (GSH) and DNA synthesis and NF-κB binding were reduced by more than 80%. All parameters returned to normal or supranormal levels by 8 h post-treatment. Pretreatment with acrolein completely blocked 12-O-tetradecanoylphorbol-13-acetate (TPA)-induced activation of NF-κB. Cells treated for 1 h with 1 mmdiethyl maleate (DEM) showed a 34 and 53% decrease in GSH and DNA synthesis, respectively. DEM also reduced NF-κB activation by 64% at 2 h post-treatment, with recovery to within 22% of control at 8 h. Both acrolein and DEM decreased NF-κB function ∼50% at 2 h after treatment with TPA, as shown by a secreted alkaline phosphatase reporter assay. GSH returned to control levels by 8 h after DEM treatment, but proliferation remained significantly depressed for 24 h. Interestingly, DEM caused a profound decrease in NF-κB binding, even at doses as low as 0.125 mm that had little effect on GSH. Neither acrolein nor DEM had any effect on the levels of phosphorylated or nonphosphorylated inhibitor κB-α (IκB-α). Furthermore, acrolein decreased NF-κB activation in cells depleted of IκB-α by TPA stimulation in the presence of cycloheximide, demonstrating that the decrease in NF-κB activation was not the result of increased binding by the inhibitory protein. This conclusion was further supported by the finding that acrolein modified NF-κB in the cytosol prior to chemical dissociation from IκB with detergent. Together, these data support the conclusion that the inhibition of NF-κB activation by acrolein and DEM is IκB-independent. The mechanism appears to be related to direct modification of thiol groups in the NF-κB subunits.

is produced by a wide variety of both natural and synthetic processes including incomplete combustion or pyrolysis of organic materials such as fuels, wood, synthetic polymers, food, and tobacco. In addition, patients treated with the cytostatic agent cyclophosphamide are exposed to acrolein as a metabolite of the parent drug (2).
Extensive research has been done on the acute biochemical effects of acrolein. However, the effects of subacute exposures have been little studied, particularly at the molecular level. Recent work has shown that acrolein can inhibit cell proliferation at doses that do not cause lethality (3), and such information may have major significance in terms of signal transduction pathways as well as, perhaps, in the control of cell division and apoptosis. As a metabolite of cyclophosphamide, acrolein may also play a role in the unique antineoplastic efficacy of this drug through molecular effects associated with low acrolein doses.
Myriad adverse cellular effects are seen following exposure to acrolein, including growth inhibition, alterations in the levels of glutathione (GSH), protein sulfhydryls, and thiol-containing enzymes, and increased cell membrane permeability (4 -8). The primary source of acrolein's reactivity is its ␣,␤-unsaturated carbon-carbon bond. This molecule will react via a Michael addition in the presence of a nucleophile to form an alkylated adduct. Acrolein's potential role as a carcinogen is based on the observation that it binds GSH (9) and nuclear chromatin (10) and can form a number of adducts with DNA (11)(12)(13).
Some researchers have suggested that acrolein's antiproliferative effects may be the result of its binding to RNA polymerase, thereby serving as a transcriptional restraint (14). However, the fact that GSH appears to play some role in cell division (15,16) raises the possibility that acrolein-mediated alterations in this tripeptide may also be an important factor. Our previous data (3,17) demonstrated that inhibiting the proliferation of human lung adenocarcinoma A549 cells with acrolein correlated with acrolein-induced changes in GSH.
Although a cause-and-effect relationship between acroleininduced changes in GSH and proliferation has not been shown, it is apparent that acrolein can alter redox-regulated cellular pathways. Nuclear factor-B (NF-B) 1 is one of the most widely studied molecules affected by cellular redox status. It was first identified as a factor that activated the Ig -light chain intron enhancer during B-lymphocyte development (18). High levels of interest in this transcription factor are based on its broad role in coordinately controlling a number of genes including those encoding inflammatory cytokines, chemokines, interfer-ons, proteins of the major histocompatibility complex, growth factors, cell adhesion molecules, and viruses (19).
NF-B, which comprises a 50-and 65-kDa heterodimer complex, is the prototype of a family of dimeric transcription factors consisting of monomers that have approximately 300-amino acid Rel regions that bind to DNA and interact with each other (20). These factors are normally bound to a member of a family of inhibitory proteins known as inhibitor B (IB). The inhibitors all have 5-7 ankyrin repeat domains, each with approximately 30 amino acids, that form a unit able to interact with Rel regions. IB-␣, the best characterized member of this family, binds the p50/p65 heterodimer of NF-B and retains it in the cytoplasm. The exposure of cells to NF-B activators, including oxidants, cytokines (such as tumor necrosis factor-␣ or interleukin-1), or the phorbol ester 12-O-tetradecanoylphorbol-13-acetate (TPA), causes phosphorylation of two serine residues of IB-␣ (Ser 32 and Ser 36 ). This phosphorylation is the signal for ubiquitination and degradation of IB-␣ by the 26 S proteasome. NF-B is then released and translocated to the nucleus where it can bind to B sites on the DNA, thereby activating transcription of target genes (21,22).
NF-B is thought to be under redox control at two distinct levels. The activation and nuclear translocation of NF-B involve reactive oxygen intermediates and can be blocked by reducing agents such as N-acetylcysteine and GSH (23,24). In contrast, the DNA-binding activity of NF-B is inhibited by oxidative agents and potentiated by reducing thiols (25,26). These are likely the results of the requirement that cysteine residues present in the DNA-binding domain of all members of the Rel protein family be reduced to bind DNA (26). Acrolein's reactivity with nucleophiles suggests that it may interfere with NF-B binding either by altering the redox balance of the nucleus or by forming adducts with NF-B. In this study, we describe acrolein's attenuation of NF-B activation in A549 human lung adenocarcinoma cells in a manner independent of IB-␣ and consistent with the formation of acrolein-NF-B conjugates.

EXPERIMENTAL PROCEDURES
Materials-Dulbecco's modified Eagle's medium (DMEM), acrolein (90%; water and dimers make up the other 10%), diethyl maleate (DEM), and o-phthalaldehyde were obtained from Sigma. Fetal bovine serum was purchased from Summit Biotechnology (Fort Collins, CO). [ 3 H]Thymidine (55 Ci/mmol) was obtained from ICN (Costa Mesa, CA). All antibodies were secured from Santa Cruz Biotechnology (Santa Cruz, CA) or New England Biolabs (Beverly, MA). The double-stranded NF-B consensus oligonucleotide was purchased from Promega Corp. (Madison, WI). Phosphorylated and nonphosphorylated IB control cell extracts were obtained from New England Biolabs.
Cell Culture-A549 human lung adenocarcinoma cells, obtained originally from the American Type Culture Collection (Manassas, VA), were cultured in DMEM (pH 7.4) supplemented with 10% (v/v) fetal bovine serum, 3.7 g/liter sodium bicarbonate, and 100 mg/liter gentamicin. Cells were maintained at 37°C with 5% CO 2 . Cultures were passaged at confluency (approximately every 3 days) and were removed from monolayer stock cultures with trypsin-EDTA. Cells were counted with a T-890 Coulter counter (Miami, FL) and plated in either Falcon 6-well dishes (9.6 cm 2 /well) or Corning 10-cm tissue culture plates (55 cm 2 ) with a medium volume of 2 ml/well and 10 ml/plate, respectively.
Cell Proliferation-Changes in cell growth were monitored by the uptake of [ 3 H]thymidine. Cells were seeded 48 h before treatment. DEM was dissolved in 100% ethanol and added to culture dishes at an amount equivalent to 0.1% (v/v) of the medium. For treatment with acrolein, cells were washed twice in one volume/wash of Earl's balanced salt solution (EBSS). Cells were then incubated for 30 min at 37°C with 5% CO 2 in sterile EBSS containing the desired dose of acrolein. Incubation in EBSS was essential because of the reactivity of acrolein with components of DMEM (27). Following treatment, the cells were replenished with fresh DMEM ϩ 10% fetal bovine serum. Additional washes to remove acrolein were not incorporated because any residual acrolein would rapidly react with nucleophiles present in the complete medium.
The uptake of exogenous 3 H-labeled thymidine was measured in cells treated with acrolein, DEM, or vehicle. Cells were pulsed for 2 h with 2.5 Ci/ml [ 3 H]thymidine before isolating the DNA (28). DNA was quantitated by fluorescence after treatment with ethidium bromide (29).
Cell Counts-Cell counts at the time of treatment (48 h post-seeding) were obtained using the CyQuant cell proliferation assay. This assay has a linear detection range of 50 -50,000 cells/200 l and is dependent on a green dye (CyQuant-GR) that fluoresces when bound to cellular nucleic acids. Cell monolayers were washed twice with phosphatebuffered saline, trypsinized, suspended in phosphate-buffered saline, and pelleted at 200 ϫ g. The supernatant was carefully removed and the cells frozen at Ϫ80°C. At the time of the assay, cells were thawed at room temperature and lysed in buffer containing the CyQuant-GR dye prepared according to manufacturer's instructions. Fluorescence was measured (excitation, 480 nm; emission, 520 nm) and compared with a standard curve for cell number determination.
Total Glutathione Measurement-Previous work in our laboratory showed that acrolein treatment does not significantly alter the level of glutathione disulfide (17). Therefore, only total glutathione (GSH ϩ glutathione disulfide) was measured by HPLC (30) or enzymatically (31). Briefly, cells were seeded at 5000 cells/cm 2 in six-well plates and treated with acrolein or DEM. Cell monolayers were washed twice with PBS and lysed with 1 ml of 20 mM EDTA followed by sonication for 1 min. For HPLC analyses, 250 l of the lysate were combined with 83 l of 25 mM NaH 2 PO 4 , pH 7.0. Samples were then processed, derivatized with o-phthalaldehyde, and analyzed as described previously (17). For the enzymatic assay, 100 l of cell lysate were combined with 600 l of 0.2 M KH 2 PO 4 and 5 mM EDTA (pH 7.4) and analyzed. Total protein in the lysates was determined (32) and compared with a bovine serum albumin standard curve.
Electrophoretic Mobility Shift Assays-Electrophoretic mobility shift assays were carried out after the method of Denison et al. (33) as modified by Bowes et al. (34). Briefly, cells were rinsed twice and lysed in ice-cold HEGD (25 mM HEPES, pH 7.6, 1.5 mM EDTA, 10% glycerol, 1 mM dithiothreitol, 0.1 mg/ml phenylmethylsulfonyl fluoride, 0.75 mM spermidine, 0.15 mM spermine) by homogenization. The homogenate was centrifuged at 12,000 ϫ g for 10 min at 4°C. Experiments examining the activation of cytosolic latent NF-B by detergents used the 12,000 ϫ g supernatant fraction. Cytosol (4 g of protein) was pretreated with 0.8% (w/v) sodium deoxycholate and 1.1% Nonidet P-40 for 10 min on ice before incubation with the labeled oligonucleotide probe (35). NF-B was assessed using the 12,000 ϫ g pellet extracted with 20 l of HEGDK (HEGD ϩ 0.5 M KCl) for 1 h on ice. Extracted pellets were centrifuged at 16,000 ϫ g for 10 min at 4°C, and the supernatant containing the nuclear extracts was collected and assayed for protein content (32). Extracts were frozen using liquid nitrogen and stored at Ϫ80°C until analyzed. 5-20 g of extracted protein were incubated in a reaction mixture consisting of 18.8 mM HEPES, 1.1 mM EDTA, 7.5% glycerol, 0.75 mM dithiothreitol, and 62.5 ng/ml poly(dI-dC) for 15 min at room temperature to reduce interference by nonspecific DNA-binding proteins.
To determine NF-B binding activity, 0.1 ng of NF-B labeled with [␥-32 P]ATP (3000 -5000 Ci/mmol; NEN Life Science Products, Boston, MA) was added to the nuclear or cytosolic extracts for 15 min. The specificity of the binding reaction was assessed using unlabeled NF-B, which competitively eliminated the induced band, or with an excess of a non-NF-B competitor oligonucleotide, which was without effect. Bound NF-B was separated from the free probe on a 4% polyacrylamide nondenaturing gel for 2 h at 120 V. Gels were dried under vacuum and exposed to Kodak XAR-5 film (Sigma) for 1-4 h at Ϫ80°C with intensifying screens. Gels were also evaluated with a Packard Instant Imager and Packard imaging software (version 2.02, Packard Instrument Co.).
NF-B Reporter Assay-A549 cells were transfected with the pNF-B secreted alkaline phosphatase (SEAP) vector (CLONTECH Laboratories, Palo Alto, CA). Induction of the NF-B pathway enables it to bind to the enhancer element located in the promoter region of the vector, thus activating transcription of the reporter gene and leading to increases in alkaline phosphatase activity in the culture medium. The alkaline phosphatase assay was done using the Great EscAPe SEAP fluorescence detection kit (CLONTECH) per the manufacturer's instructions.
Conditions for transfection of A549 cells were optimized using Fu-GENE transfection reagent (Roche Molecular Biochemicals, Indianapolis, IN). 2 g of DNA were used for each transfection. 24 h posttransfection, the cells were washed with EBSS and treated with acrolein or DEM for 30 min as described previously. After treatment, cells were washed with EBSS, and fresh DMEM medium with fetal bovine serum containing 100 ng/ml TPA was added. 100 l of media were collected after 2 h for the alkaline phosphatase assay and stored at Ϫ20°C.
Western Analyses-Monolayer cells (10 6 ) were lysed in 300 l of lysis buffer (10 mM Tris-HCl (pH 7.4), 10 mM NaCl, 3 mM MgCl 2 , 1 mM EDTA, 0.1% (v/v) Nonidet P-40, 100 g/ml phenylmethylsulfonyl fluoride, 30 l/ml aprotinin, and 1 mM sodium orthovanadate). The lysate was collected and incubated on ice for 15 min. Samples were centrifuged at 16,000 ϫ g for 10 min, and the supernatant was collected, assayed for protein (36), and stored at Ϫ20°C. Thawed supernatants were mixed 1:3 with loading dye (4% (w/v) SDS, 20% (w/v) glycerol, 4% (w/v) ␤-mercaptoethanol, 0.2 M Tris-HCl (pH 6.8), and 0.02% (w/v) bromphenol blue) and separated on SDS-polyacrylamide gels (8 -15%). Protein was transferred to polyvinylidene difluoride membranes (Millipore, Bedford, MA) and blocked from 1 h to overnight in 5% (w/v) nonfat dry milk (Bio-Rad) in TBS-T (25 mM Tris-HCl (pH 7.6), 0.2 M NaCl, and 0.15% Tween 20 (v/v)). Membranes were incubated with a polyclonal antibody specific for the protein of interest (1:1500 dilution in TBS-T) for 1 h. After washing in TBS-T, the membranes were rinsed and incubated with a horseradish peroxidase-conjugated secondary antibody (1:3000 dilution in TBS-T; Amersham Pharmacia Biotech) for 1 h. After the secondary antibody incubation, the membranes were rinsed with TBS-T, and bound antibodies were detected using enhanced chemiluminescence (ECL) with a kit from Amersham Pharmacia Biotech. Developed film was scanned, and individual band densities were integrated using NIH Image public domain software. Immunoblots following the various treatments were run a minimum of two times. Representative blots are shown in the figures.
Statistics-Data are expressed as means Ϯ S.E. Comparisons between groups were done with a one-way analysis of variance followed by the Student-Newman-Keul's test. A p value of less than 0.05 was considered significant.

Acrolein and DEM Reduce Proliferation in A549
Cells-Sixwell plates were seeded at 5000 cells/cm 2 (48,000 cells/well) and incubated 48 h before treatment for 30 min with 45 fmol of acrolein/cell (6.7 M) in EBSS or for 1 h with 6.7 pmol of DEM/cell (1 mM). DNA synthesis was reduced to 30 and 63% of vehicle-treated cells 2 h after acrolein or DEM exposure, respectively (Table I). DNA synthesis in acrolein-treated cells recovered to supranormal levels by 8 h post-treatment, whereas growth in DEM-treated cells remained significantly suppressed, reaching only 54% of the level of growth in vehicletreated cells (64% of control cells) at 24 h.
Acrolein and DEM Decrease Cellular Glutathione-Under slightly different conditions, we have shown previously that acrolein rapidly decreases total cellular GSH (3). In the current study, cells were treated with 45 fmol of acrolein/cell or 6.7 pmol of DEM/cell. The level of GSH in acrolein-treated cells declined to 13% of that in vehicle-treated cells immediately following treatment and recovered to normal or supranormal levels by 8 h post-treatment (Fig. 1). DEM-treated cells showed a smaller decline in GSH to 63% of the level in vehicle-treated cells with recovery again occurring by 8 h post-treatment (Fig. 1).

Acrolein and DEM Attenuate NF-B Activation-Treating
A549 cells with 35 fmol of acrolein/cell caused a significant decrease in NF-B activation relative to TPA-treated or serumdeprived controls after as little as a 5-min exposure. This binding inhibition increased with the time of exposure. However, NF-B activation in serum-deprived, vehicle-treated cells also began to decline at 2 h post-treatment (Fig. 2). To minimize the effects of serum deprivation, 30-min acrolein treatments were selected for further studies. With this length of treatment, both constitutive and TPA-stimulated NF-B activation were inhibited by acrolein (Fig. 2). Nonspecific binding was evident in this gel but did not correlate with either time or acrolein treatment.
A time-response study of NF-B activation in which cells were treated with 45 fmol of acrolein/cell for 30 min (Fig. 3) showed inhibition and recovery patterns very much like those seen when examining changes in DNA synthesis (Table I) and total GSH (Fig. 1). Acrolein caused a dramatic decline in NF-B binding at 30 min and at 2 h post-treatment. Some recovery of NF-B activation was evident at 4 h, and the inhibitory effect of 30 min of acrolein treatment was fully reversed by 8 h post-treatment (Fig. 3).
DEM also caused a decrease in NF-B activation. There was a clear dose-response relationship with NF-B binding increasingly reduced after 1-h exposures to DEM doses from 0.125 to 2 mM (Fig. 4). Interestingly, the inhibition seen in NF-B activation with 0.125 mM DEM was profound, yet little or no change in total GSH was evident at this dose (data not shown). Treating cells with 3.33 pmol of DEM/cell (1 mM) for 1 h resulted in a NF-B activation time response (Fig. 5) that was almost identical to that obtained following acrolein exposure (Fig. 3), and again paralleled changes in GSH (Fig. 1). NF-B binding decreased dramatically at 1-2 h post-treatment and showed a recovery to near normal binding by 8 -12 h. Nonspecific binding was more intense in this gel, but again did not correlate with either time or DEM treatment.
The SEAP reporter assay confirmed that both acrolein and DEM diminished the transcriptional activity of NF-B. Two h after adding TPA to cells pretreated for 30 min with either 45 fmol/cell acrolein or 1 mM DEM, SEAP activity was decreased by 51 and 45%, respectively.   1. Glutathione levels after treatment with acrolein or DEM. Cells were seeded at 5000 cells/cm 2 in six-well plates and incubated for 48 h at 37°C in 5% CO 2 before treatment with 45 fmol of acrolein/cell (30 min) or 1 mM DEM (1 h). Total glutathione was determined at the indicated times as measured from the start of treatment. Data are expressed as the mean percentage of total glutathione relative to vehicle-treated cells Ϯ S.E. (n ϭ 5 for acrolein, n ϭ 2 for DEM). 45 fmol of acrolein/cell or for 1 h with 3.33 pmol of DEM/cell (Fig. 6, A and B). To further examine this phenomenon, changes in the level of phosphorylated IB-␣ were examined following treatment with acrolein. Once again, no changes in the levels of this protein were observed (Fig. 6C), suggesting that acrolein blocks NF-B activation by an IB-independent mechanism.

Role of IB in Reduced NF-B Binding-Changes
A more thorough analysis of the possibility of IB-independent changes in NF-B activation involved examining the degradation of IB following stimulation with TPA, a tumor promoter that up-regulates protein kinase C. Treatment with 100 ng/ml TPA caused a temporary decrease in the levels of IB (Fig. 7A). By stimulating cells with TPA in the presence of the protein synthesis inhibitor cycloheximide (CHX), IB levels were almost completely abrogated at 2 h post-treatment (Fig.  7B). NF-B activation was also checked at this time to ensure that the treatment with CHX had not affected NF-B binding. Stimulating cells with TPA in the presence of CHX resulted in maximum NF-B activation at 2 h post-treatment (Fig. 8), the same time that IB levels were at their nadir. Finally, cells that had been stimulated with TPA for 1.5 h in the presence of CHX were treated with 45 fmol of acrolein/cell for 30 min. Under these conditions, acrolein still caused a 63% decline in NF-B activation (Fig. 9), indicating that the effect of acrolein was independent of IB.
Cytosolic versus Nuclear Effects-To determine whether acrolein blocked NF-B activation by acting in the cytoplasm or in the nucleus, cytosolic extracts were obtained from cells pretreated for 30 min with 45 fmol/cell acrolein or vehicle. Treatment of vehicle extracts with deoxycholate dissociated NF-B from IB, yielding binding to the consensus sequence (Fig. 10,  lane 4). In contrast, extracts from acrolein-treated cells exhibited greatly diminished binding (Fig. 10, lane 5), suggesting that cytosolic binding occurs. Acrolein may also bind to NF-B in the nucleus as shown by an experiment in which cells were first treated with TPA to stimulate the nuclear translocation of NF-B followed by treatment with either acrolein or DEM. This experiment revealed that both acrolein and DEM decreased NF-B activation under these conditions (data not shown), suggesting that both can act within the nucleus.
Analysis of the inhibition of activation of NF-B was studied further by treating the nuclear extracts of acrolein-and DEMtreated cells with either 100 M GSH or 1% ␤-mercaptoethanol (␤-ME) in vitro to determine whether the inhibitory effects were reversible. These data show that neither GSH nor ␤-ME restores NF-B binding in cells treated with acrolein or DEM (Fig. 11). Pretreatment with TPA before exposing the cells to acrolein or DEM had no effect on the ability of GSH or ␤-ME to restore binding. Interestingly, the nuclear extracts from untreated control cells showed a 24% increase in NF-B binding following the addition of ␤-ME but not GSH. DISCUSSION The data presented here provide several avenues of support for the hypothesis that both acrolein and DEM inhibit NF-B FIG. 8. NF-B binding in the presence of TPA and CHX. Cells were seeded at 10,000 cells/cm 2 in 10-cm plates and incubated for 48 h at 37°C in 5% CO 2 before treatment. Cells were treated with 100 ng/ml TPA in the presence of 100 g/ml CHX (2 h). Nuclear fractions were harvested at the indicated times as measured from the beginning of treatment. 10 g of total protein were loaded per lane. C, control. activation by an IB-independent mechanism. Most importantly, acrolein and DEM block NF-B binding without altering the cellular levels of either the phosphorylated or nonphosphorylated forms of IB-␣. Acrolein and DEM also block binding in the absence of IB making it clear that they must be interfering with the activation of NF-B by some mechanism other than modulation of IB. Two possible mechanisms exist: the direct inactivation of NF-B; or the scavenging of nuclear reducing equivalents required for NF-B binding, such as GSH, thereby lowering its binding affinity for DNA. This latter mechanism is unlikely, however, because the effect of DEM occurred at doses that did not alter GSH. Further, in cells treated with acrolein or DEM alone, or when treatment followed TPA stimulation, NF-B activation could not be restored by the addition of either GSH or ␤-ME. An interesting observation was that ␤-ME increased NF-B binding in untreated control cells. The nuclear redox status could be such that some molecules of NF-B are not in the reduced form necessary to facilitate DNA binding. The addition of a strong reducing agent in vitro may have forced the reduction of this pool of NF-B, resulting in the observed increase in binding.
Acrolein-mediated modifications of NF-B can occur in the cytosol of intact cells as shown by the inability of detergents to release an active form following acrolein treatment. This finding also indicates that the site of modification is not blocked by IB, matching previous data from studies with other thiol reactive agents (37,38). NF-B-like transcription factors belong to the Rel protein family and have a highly conserved N-terminal region that is critical for DNA binding. Within this region, the Cys 62 residue on the p50 subunit and the Cys 38 residue on the p65 subunit are essential for DNA binding (26,39,40). Thiols present in these regions may form conjugates with acrolein via a Michael addition. A mechanism of this type has been described previously when various compounds directly reacted with the p50 subunit to inhibit NF-B binding (38). In a similar way, acrolein inhibits the DNA repair enzyme O 6 -alkylguaninine-DNA alkyltransferase by acting at an acrolein-sensitive thiol residue required for the catalytic activity of the enzyme (41).
The translocated NF-B must have a reduced environment if it is to bind DNA (26,37,42). Therefore, in addition to forming conjugates with NF-B, acrolein may inhibit binding by depleting nuclear GSH. In support of this mechanism, we observed a strong correlation between the time necessary to recover intracellular GSH and the time it takes for NF-B binding to return to normal following an acrolein insult. However, this seems unlikely to be the sole mechanism based on the inability of deoxycholate to activate NF-B after acrolein.
Most if not all agents activating NF-B tend to trigger the formation of reactive oxygen species or are oxidants by themselves (43). Despite the increased NF-B activation seen in the presence of oxidants, studies in two human T-cell lines have shown that a reduction in GSH by buthionine sulfoximine inhibits the activation and translocation of NF-B (25,44). It was also found that the addition of cysteine caused a reversible oxidation of NF-B (25), presumably by raising intracellular glutathione disulfide levels. These data may point to the existence of a critical intracellular redox balance. Increasing this reactive oxygen species pool in the cytosol appears to activate NF-B, whereas increasing it in the nucleus blocks NF-B binding, which may enhance apoptosis (45,46). In general, acrolein appears to affect NF-B binding directly by conjugating its protein subunits and indirectly by alkylating GSH, thereby altering the nuclear redox balance.
In the same way that acrolein affects NF-B, our data suggest that DEM decreases activation both directly and indirectly. The fact that DEM blocks NF-B binding in a manner not reversible with either GSH or ␤-ME indicates that it has direct interactions with the NF-B subunits. Furthermore, because DEM is known to conjugate GSH through the actions of glutathione S-transferase, it may act on NF-B indirectly by lowering the nuclear levels of GSH.
The key features of NF-B transcriptional control are that it is fast, versatile, and involved in many different gene systems (47). Of particular interest to the current study is the involvement of NF-B with genes that regulate cell proliferation. The importance of NF-B subunits in cell proliferation is suggested by several studies. Snapper et al. (48) showed that B cells from p50 Ϫ/Ϫ knockout mice proliferated normally in response to some stimuli but showed no response to other stimuli that were mitogenic in control cells. Mice lacking the p50/p105 subunits developed normally but exhibited defects in immune responses involving B cells (49). A more critical role seems to belong to the p65 subunit. Vascular smooth muscle cells treated with p65 antisense oligonucleotides showed a concentration-dependent inhibition of both adherence and proliferation (50). Even more dramatic is the report that p65 (Rel A) null mice exhibited a dramatic phenotype-embryonic lethality, apparently because of widespread apoptosis in the liver (51). This function in the regulation of cell growth may be mediated through c-myc, which is known to have two NF-B sites in its promoter/enhancer region (52). Furthermore, NF-B is implicated in the transcriptional regulation of the p53 gene (53).
In the current study, treating cells with 45 fmol of acrolein/ cell for 30 min caused an 85% decline in DNA synthesis, which corresponded to a 55% decrease in NF-B binding and an 87% loss of GSH. In all instances, the measured parameters returned to normal levels or higher by 8 h post-treatment. Although electrophoretic mobility shift assay data are only semiquantitative, the correlation between acrolein-mediated reductions in NF-B activation and DNA synthesis support a functional link. It may also be that acrolein inhibits growth protein(s) in a manner similar to the proposed inhibition of NF-B activation, i.e. by conjugation with thiol residue(s) critical to their function.
Results from DEM studies support the role of NF-B in mediating changes in DNA synthesis. Although DEM-treated cells posted a full recovery of GSH by 8 h post-treatment, the 8-h levels of DNA synthesis and NF-B activation reached only FIG. 11. Cells were seeded at 10,000 cells/cm 2 in 10-cm plates and incubated for 48 h at 37°C in 5% CO 2 before treatment. Cells were treated with 45 fmol of acrolein/cell (30 min) or 1 mM DEM (1 h). Some cells were pretreated with 100 ng/ml TPA before incubation with acrolein or DEM. Nuclear fractions were harvested 2 h after the beginning of treatment. GSH or ␤-mercaptoethanol was added to the nuclear extracts in vitro to obtain 100 M or 1% final concentration, respectively. 10 g of total protein were loaded per lane. 77 and 74% of the level in vehicle-treated cells, respectively. However, NF-B activation in DEM-treated cells did return to normal by 12 h post-treatment, whereas DNA synthesis remained suppressed to 24 h.
In conclusion, acrolein causes a dramatic decline in NF-B binding by an IB-independent mechanism, which most likely involves alkylation of critical thiol sites within the DNA binding domain of the NF-B subunits. Identification of acrolein/ DEM-p50 and/or acrolein/DEM-p65 conjugates will provide conclusive evidence of the mechanism by which these chemicals alter NF-B binding. It is apparent that GSH is intimately involved with NF-B activation and that both GSH and NF-B play a role in cell growth and apoptosis. The observed attenuation of DNA synthesis following acrolein insult is probably the downstream result of the effects of acrolein on GSH and NF-B. In addition, NF-B is involved in regulating the expression of several growth-related genes. As a result, the effects of acrolein on these factors could have unrecognized toxic consequences, including a role in the effectiveness of the anticancer drug cyclophosphamide.