3-Deazaadenosine, a S-Adenosylhomocysteine Hydrolase Inhibitor, Has Dual Effects on NF-κB Regulation

Previously we reported that 3-deazaadenosine (DZA), a potent inhibitor and substrate forS-adenosylhomocysteine hydrolase inhibits bacterial lipopolysaccharide-induced transcription of tumor necrosis factor-α and interleukin-1β in mouse macrophage RAW 264.7 cells. In this study, we demonstrate the effects of DZA on nuclear factor-κB (NF-κB) regulation. DZA inhibits the transcriptional activity of NF-κB through the hindrance of p65 (Rel-A) phosphorylation without reduction of its nuclear translocation and DNA binding activity. The inhibitory effect of DZA on NF-κB transcriptional activity is potentiated by the addition of homocysteine. Taken together, DZA promotes the proteolytic degradation of IκBα, but not IκBβ, resulting in an increase of DNA binding activity of NF-κB in the nucleus in the absence of its transcriptional activity in RAW 264.7 cells. The reduction of IκBα by DZA is neither involved in IκB kinase complex activation nor modulated by the addition of homocysteine. This study strongly suggests that DZA may be a potent drug for the treatment of diseases in which NF-κB plays a central pathogenic role, as well as a useful tool for studying the regulation and physiological functions of NF-κB.

drug for treatment of many human diseases, the action mechanism of DZA is not yet fully understood.
Previously, we reported that DZA inhibits LPS-induced TNF-␣ and IL-1␤ transcription in mouse macrophage RAW264.7 cells (26). In this study, elementary experiments confirming DZA inhibition of TNF-␣ transcription in human monocytic macrophage THP-1 cells and human peripheral blood monocytes (PBMC) encouraged us to investigate the effect of DZA on NF-B activation with the intention of understanding the cellular mechanism of DZA. DZA potently inhibited the transcriptional activity of NF-B through the hindrance of p65 phosphorylation without reduction of nuclear translocation and DNA binding activity of NF-B. In RAW 264.7 cells, DZA promoted the proteolytic degradation of IB␣ resulting in an increase of DNA binding activity of NF-B in the nucleus. These results strongly suggest that DZA may serve as a potent drug for the treatment of diseases in which NF-B plays an important pathogenic role, as well as a useful tool for studying the regulation and physiological functions of NF-B. Cell Culture-Mouse macrophage RAW 264.7, human monocytic THP-1, and SV40-transformed African green monkey kidney COS-7 cells were obtained from the American Type Culture Collection (ATCC, Manassas, VA). PBMC were isolated from defibrinated blood using Ficoll-Hypaque (Amersham Pharmacia Biotech, Uppsala, Sweden) by density-gradient separation, followed by adherence to tissue culture dish for 2 h at 37°C. Nonadherent cells were removed by washing the monolayer four times with phosphate-buffered saline. All cells were cultured in RPMI 1640 medium supplemented with 20 mM HEPES, 25 mM sodium bicarbonate, 50 g/ml gentamicin, and 10% heat-inactivated (56°C for 30 min) fetal bovine serum (Hyclone Laboratories Inc., Logan, UT) at 37°C in an atmosphere of 5% CO 2 .

Chemicals
Northern Blot Analysis-Total cellular RNA was isolated from cells at 2 h after LPS (1 g/ml) stimulation according to a previously described method (31). Five g of total RNA was separated on a 1% agarose gel containing 2.2 M formaldehyde. Northern blot analysis was performed as described (26). All of the digoxigenin (Roche Molecular Biochemicals, Mannheim, Germany)-labeled probes were prepared from each specific primer and cDNA fragments of amplimer sets (CLONTECH, Palo Alto, CA) by polymerase chain reaction.
Preparation of Cytoplasmic Fraction and Nuclear Extract-After RAW 264.7 cells were stimulated with LPS (1 g/ml) for 1 h and COS-7 cells were stimulated with recombinant GST-human TNF-␣ protein (50 ng/ml) for 1 h, cells were washed twice with phosphate-buffered saline. Cytosolic fraction and nuclear extract for Western blot analysis and electrophoretic mobility shift assay (EMSA) were prepared as described by Dignam et al. (32). Concentration of protein was determined using Coomassie Plus Protein Assay Reagent (Pierce).
EMSA-Five g of nuclear protein and 1 g of poly(dI-dC) (Roche Molecular Biochemicals) per reaction were incubated for 15 min at room temperature with NF-B consensus sequence (Santa Cruz Biotechnology, Santa Cruz, CA), which was 3Ј-end labeled with 32 P, in binding buffer (20 mM HEPES, pH 7.6, 1 mM EDTA, 10 mM (NH 4 ) 2 SO 4 , 1 mM DTT, 0.2% (w/v) Tween 20, 30 mM KCl). After the binding reaction, samples were analyzed by electrophoresis on a 6% native polyacrylamide gel that was run in 0.5ϫ Tris borate-EDTA (TBE) buffer, pH 8.0, and then the dried gel was subjected to autoradiography. For competition, 50-fold unlabeled NF-B or AP-1 consensus sequence (Santa Cruz Biotechnology) were used.
Plasmids and Transient Transfection Analysis-Plasmid J16 containing two copies of the wild-type NF-B-binding site, and J32 containing two copies of the mutant NF-B-binding site, upstream of a truncated c-fos promoter which was linked to the chloramphenicol acetyltransferase (CAT) gene (33), were kindly provided by Dr. David Baltimore of California Institute of Technology, Pasadena, CA. pSVL65 expressing p65 of NF-B from a SV40 promoter was kindly donated by Dr. Mahnhoon Park of Mogam Biotechnology Research Institute, Yongin, Kyunggi-do, Korea. pSVL (Amersham Pharmacia Biotech) was used as a control taking the place of pSVL65. To assess variations in transfection efficiencies, a control plasmid pCMV␤ (CLONTECH) which expresses the LacZ gene was used. Transfection of cells was performed using FuGENE 6 Transfection Reagent (Roche Molecular Biochemicals) according to a procedure recommended by the manufacturer. Cells transfected with either J16 or J32 were cultured for 18 h, and then exposed to LPS or TNF-␣ in the presence or absence of DZA alone or DZA with Hcy together. For the pSVL65 transfection, cells were exposed to 100 M DZA at 6 h after transfection, and further incubated for 18 h. Induction of CAT expression was determined using CAT enzyme-linked immunosorbent assay (Roche Molecular Biochemicals) according to the manufacturer's instructions, and standardized to constitutive levels of ␤-galactosidase activity.
In Vivo Labeling and Immunoprecipitation-Confluent monolayers of RAW 264.7 cells in 10-cm tissue culture dishes were washed twice with phosphate-free Dulbecco's minimum essential medium and incubated in phosphate-free Dulbecco's minimal essential medium for 1 h. Media was replaced by fresh phosphate-free Dulbecco's minimal essential medium containing 100 Ci of 32 P i per ml (ICN, Costa Mesa, CA), and the cells were incubated for 2 h. Cells were pretreated with or without DZA (100 M) for 1 h, and stimulated by the addition of LPS (1 g/ml). After incubation for 1 h, the cells were washed, and p65 was recovered by immunoprecipitation with anti-p65 (C-20, Santa Cruz Biotechnology) and protein A-Sepharose (Amersham Pharmacia Biotech). The immunopellets were fractionated by SDS-PAGE, and the gel was stained with Coomassie Brilliant Blue R-250 to confirm that the equal amount of protein was loaded in each wall. Dried gel was subjected to autoradiography.
IKK Assay-RAW 264.7 cells were treated with 100 M DZA for various time intervals, and subjected to IKK assay. IKK activity was measured as described (34), using recombinant protein of GST-human IB␣ as a substrate. A polyclonal antibody to IKK␣ (M-280) was obtained from Santa Cruz Biotechnology.

LPS-induced Expression of TNF-␣ mRNA Is Inhibited by
DZA-Transcription of TNF-␣ and IL-1␤ was inhibited by DZA in a dose-dependent manner in RAW 264.7 cells stimulated by LPS (26). To investigate the effects of DZA on the expression of TNF-␣ in other types of cells, THP-1 cells and PBMC were pretreated with increasing concentrations of DZA for 1 h, and stimulated by the addition of LPS at a final concentration of 1 g/ml. As shown in Fig. 1, LPS stimulation increased the steady-state levels of TNF-␣ mRNA as early as 2 h. Treatment of DZA inhibited the LPS-induced expression of TNF-␣ mRNA dose dependently in both THP-1 cells and PBMC in the same manner as in RAW 264.7 cells. Expression of ␤-actin mRNA as a control was not affected by DZA in any of the cells. These results show that DZA efficiently inhibits the LPS-induced expression of TNF-␣ mRNA and this potency is not only restricted to murine macrophage cells.
DNA Binding Activity of NF-B Is Increased by DZA-Since the activation of transcription factor NF-B has been shown to be indispensable for TNF-␣ expression induced by LPS (35), we examined the effect of DZA on DNA binding activity of NF-B in the nucleus of RAW 264.7 cells by EMSA, using 32 P-labeled NF-B specific oligonucleotides. An inducible protein-DNA complex was observed in the nuclear extracts from LPS-stimulated RAW 264.7 ( Fig. 2A). Unexpectedly, cells pretreated with DZA revealed no significant decrease of the LPS-induced DNA binding activity of NF-B, even though expression of the TNF-␣ gene was potently inhibited by exposure to DZA. Moreover, DNA binding activity was potentiated in nuclear extracts dose dependently by DZA, irrespective of LPS stimulation. In competition experiments, a 50-fold amount of unlabeled NF-B-specific oligonucleotide absolutely inhibited typical binding activities, but the same amount of unlabeled AP-1 oligonucleotide failed to inhibit binding activities, confirming their specificities (Fig. 2B). These results suggest that the inhibition of TNF-␣ gene expression by DZA occurred without down-regulation of NF-B DNA binding activity, and that DZA increases DNA binding activity of NF-B in RAW 264.7 cells. Since the presence of a reducing agent such as DTT in preparation of the nuclear extract and the execution of EMSA could mask lost DNA binding activity of NF-B, we repeated the EMSA assay without DTT (Fig. 2C). Removal of DTT from the assay could not modulate each of the DNA binding activities which were presented by DTT in an assay with DTT, indicating that the DZA inhibition of TNF-␣ gene expression is not caused by a loss of NF-B DNA binding activity.
Nuclear Translocation of NF-B Is Increased by DZA-We examined the effects of DZA on the translocation of p65 (Rel-A), p50 (NF-B1), and c-Rel into the nucleus of RAW 264.7 cells stimulated with or without LPS using Western blot analysis. LPS stimulation increased protein levels of each NF-B subunit in the nucleus. Treatment of DZA at a concentration of 100 M induced an increase of Rel family proteins in the nucleus of RAW 264.7 cells regardless of LPS stimulation (Fig. 3A). These results are in exact agreement with the results in Fig. 2A, which shows that DZA induces nuclear translocation of NF-B and potentiates its LPS-induced translocation in RAW 264.7 cells. We next established the modulation of p65 in the nucleus of RAW 264.7 cells by DZA at either various concentrations or various time intervals. Similar to the results in Fig. 2A, DZA at 100 M concentration increased the levels of p65 in the nucleus regardless of LPS stimulation (Fig. 3B). Fig. 3C shows the modulation of p65 level in the nucleus by treatment of DZA at various time intervals. p65 in the nucleus increased remarkably at 60 and 120 min after treatment with DZA, even though there was a slight decrease at 15 min. The total amount of cellular p65 was not modulated by DZA, whereas the amount of cytoplasmic p65 decreased proportionally to the increase of nuclear p65 by DZA as measured in a Western blot analysis of total cellular extracts and cytoplasmic fractions (data not shown). These results demonstrate that DZA increases the amount of p65 nuclear translocation in RAW 264.7 cells, and that the increase of p65 in the nucleus by DZA is caused only by (1 g/ml). After 1 h, nuclear extracts were prepared. 10 g of each nuclear protein was fractionated by SDS-PAGE, and then subjected to Western blot analysis with specific antibodies against p65, p50, c-Rel, respectively. B, dose-response effect of DZA. RAW 264.7 cells were pretreated with DZA at the indicated concentrations (M) for 1 h, and then stimulated by the addition of LPS (1 g/ml) or not stimulated. After 1 h, nuclear extracts were prepared. 10 g of each nuclear protein was separated by SDS-PAGE, and then Western blot analysis with specific antibody against p65 was performed. C, time-response effect of DZA. RAW 264.7 cells were treated with 100 M DZA for the indicated times. At the end of the times, nuclear extracts were prepared. 10 g of each nuclear protein was separated by SDS-PAGE, and then Western blot analysis with specific antibody against p65 was performed. Data illustrated are from a single experiment and are representative of a total of three separate experiments. nuclear translocation, and not by the enhancement of p65 expression.

LPS-induced NF-B Transcriptional Activity Is Inhibited by DZA, and This Inhibitory Effect Is Augmented by the Addition of Hcy-Since
DZA is known to inhibit the expression of TNF-␣ mRNA without a diminution of NF-B nuclear translocation and DNA binding activity, we proved the effects of DZA on the NF-B transcriptional activity by transient transfection experiments of RAW 264.7 cells using reporter gene constructs carrying two copies of the wild-type (J16) NF-B binding sequence or mutant (J32) NF-B binding sequence in front of the CAT gene, which have been shown to specifically respond to NF-B activation (33). LPS stimulation of cells transfected with J16 resulted in a 4.7-fold induction of CAT expression (Fig. 4A), whereas no induction of CAT expression by LPS was observed in cells transfected with J32. Treatment of the cells with 100 M DZA for 1 h before LPS stimulation resulted in a drastic inhibition of LPS-induced CAT expression. LPS-induced transcriptional activity of NF-B was dose dependently inhibited by DZA (Fig. 4B). These results strongly suggest that DZA potently inhibits NF-B transcriptional activity in RAW 264.7 cells stimulated by LPS, even though it more intensely provokes the nuclear translocation and DNA binding activity of NF-B. The effect of Hcy on DZA inhibition of NF-B transcriptional activity was examined. There was a tendency to inhibit NF-B transcriptional activity more than the inhibition observed by DZA alone (Fig. 4B). These results imply that DZA inhibition of NF-B transcriptional activity is involved in the accumulation of DZA-Hcy in RAW 264.7 cells. To determine the effect of DZA on CAT expression induced by expression of p65, RAW 264.7 cells were co-transfected with CAT-reporter plasmids and pSVL65 that express large amounts of p65. As a control, cells were transfected with pSVL substituting for pSVL65. After 6 h of transfection, one group of cells was exposed for 18 h to 100 M DZA and the other was not (Fig. 4C). Expression of p65 induced a 5.9-fold increase of CAT protein, driven from J16 under the absence of DZA treatment, but did not induce CAT protein expression from J32. The expression of CAT protein was completely inhibited from the cells transfected with pSVL65 when DZA was treated into cells at the concentration of 100 M, although the induction of CAT expression in the cells transfected with pSVL 65 was higher than that in the cells only stimulated with LPS. The production of p65 was not affected by DZA as monitored by Western blot analysis using anti-p65 (data not shown). These results, together with the results in Fig. 4, A and B, showing DZA inhibition of LPS-induced NF-B transcriptional activity, suggest that DZA directly inhibits the transcriptional activity of p65 but not at the signal cascade of LPS-induced NF-B activation. To elucidate whether any of the observed effects could be due to DZA-Hcy, cells to be exposed to 100 M DZA for 1 h before LPS stimulation were pretreated with a more specific and potent inhibitor of S-adenosylhomocysteine hydrolase, DZAri, to block the intracellular accumulation of DZA-Hcy (Fig. 4D). Treatment of DZAri almost completely abrogated the inhibition of NF-B transcriptional activity in cells treated with 100 M DZA, indicating a central role for DZA-Hcy in the DZA inhibition of NF-B transcriptional activity. No inhibition of NF-B transcriptional activity was observed in cells treated with DZAri alone. To investigate whether exogenous DZA-Hcy is capable of inhibition, cells were exposed to 100 M DZA-Hcy for 1 h before LPS stimulation. As shown in Fig. 4D, exogenous DZA-Hcy failed to inhibit NF-B transcriptional activity in the cells stimulated with LPS, demonstrating that DZA-Hcy is incapable of inhibition when given exogenously.
LPS-induced Phosphorylation of p65 Is Inhibited by DZA-To determine if DZA might inhibit the functional activation of NF-B by modifying phosphorylation of p65, we examined the phosphorylation of p65 in LPS-stimulated RAW 264.7 cells in the presence or absence of DZA. As shown in Fig.  5, LPS induced a strong phosphorylation of p65. In cells pretreated with 100 M DZA, however, LPS-induced phosphorylation of p65 was markedly reduced. A low constitutive phosphorylation of p65 was observed in unstimulated cells, which was reduced by treatment of the cells with DZA alone. These results strongly suggest that DZA inhibits the transcriptional activity of NF-B by the hindrance of p65 phosphorylation required for the functional activation of NF-B. TNF-induced NF-B Transcriptional Activity Is Inhibited by DZA-We tried to determine if DZA could inhibit NF-B transcriptional activity without affecting nuclear translocation in non-hematopoietic cells stimulated by other inducers of NF-B. TNF-␣ stimulation of COS-7 cells prominently induced the nuclear translocation of NF-B and the DNA binding activity of NF-B, which was not observed in unstimulated cells (Fig. 6A). The addition of 100 M DZA to COS-7 cells before TNF-␣ stimulation had no effect on nuclear translocation or DNA binding activity of NF-B, suggesting that DZA did not affect TNF-␣-induced nuclear translocation of NF-B in COS-7 cells. DZA alone did not induce the nuclear translocation of NF-B in the cells. In transient transfection experiments of COS-7 cells using reporter gene constructs (J16 or J32), TNF-␣ stimulation of cells transfected with J16 resulted in CAT expression that was 2.6-fold greater (Fig. 6B). No induction of CAT expression by TNF-␣ was observed in cells transfected with J32. Treatment of these cells with 100 M DZA for 1 h before TNF-␣ stimulation resulted in complete inhibition of CAT protein expression. These results reveal that DZA is a common and potent inhibitor of NF-B transcriptional activity induced by distinct stimuli. The effect of Hcy on DZA inhibition of NF-B transcriptional activity was examined. As shown in Fig. 6C, DZA dose dependently inhibited the transcriptional activity of NF-B in cells stimulated by TNF-␣. Combination of DZA and Hcy more potently inhibited NF-B transcriptional activity than inhibition observed by DZA alone. These results show that DZA inhibition of NF-B transcriptional activity is involved in the accumulation of DZA-Hcy in COS-7 cells stimulated with TNF␣. Cells co-transfected with pSVL65 and CAT-reporter plasmid (J16 or J32) showed a 3.0-fold expression of CAT protein driven from J16, but not from J32 (Fig. 6D). 100 M DZA strongly inhibited the induction of CAT protein in cells expressing p65, suggesting that DZA directly inhibits the transcriptional activity of p65.
DZA Induces Proteolytic Degradation of IB␣, but Not IB␤-Translocation of NF-B into the nucleus is linked to proteolytic degradation of IB proteins (1), and as we have described above, DZA increases NF-B nuclear translocation in RAW 264.7 cells. To test whether DZA promotes proteolytic degradation of IB proteins, RAW 264.7 cells were pretreated with DZA following stimulation with or without LPS. As shown in Fig. 7A, the level of IB␣ protein was notably reduced by treatment of DZA alone for 2 h at a concentration of 100 M. The IB␣ protein was fully recovered at 1 h after LPS stimulation without DZA, since IB␣ is autoregulated through the activation of NF-B (3,36). Notably, DZA pretreatment before LPS prevented synthesis of IB␣ which was recovered when treated with LPS alone, indicating that DZA interfered with the synthesis of IB␣ because the transcriptional activity of NF-B was inhibited by DZA. Resynthesis of IB␣ after stimulation of LPS was more potently prevented by the addition of Hcy plus DZA (data not shown). We also examined the effect of DZA on IB␣ levels in RAW 264.7 cells at various time inter-vals. IB␣ levels were reduced from their level at 60 min to their level at 120 min after treatment with DZA, and this reduction was continued to 240 min (Fig. 7B). Thus, these results demonstrate that DZA promotes the degradation of IB␣ and prevents its resynthesis. Furthermore, these results reasonably support that DZA increases nuclear translocation and DNA binding activity of NF-B in RAW 264.7 cells. A different IB isoform, IB␤ is one of the major regulators of NF-B activity, whose kinetics is slower than that of IB␣ in response to NF-B inducers such as LPS (6). We investigated the effect of DZA on IB␤ in RAW 264.7 cells (Fig. 7, C and D). Stimulation of LPS for 1 h without pretreatment of DZA led to the disappearance of IB␤, which would normally exist in Western blot analysis using unstimulated RAW 264.7 cells (Fig. 7C). The disappearance of IB␤ was also observed in cells pretreated for 1 h with DZA at increasing concentrations following stimulation with LPS. Interestingly, treatment of 100 M DZA alone for 2 h at several concentrations could not reduce the level of IB␤ (Fig. 7C), and IB␤ remained continuously until 4 h after exposure of 100 M DZA (Fig. 7D), implying that DZA cannot lead to IB␤ down-regulation, and the down-regulation of IB␤ is not involved in enhanced nuclear translocation of NF-B, mediated by DZA in RAW 264.7 cells.

DZA-induced Proteolytic Degradation of IB␣ Is Neither Modulated by Hcy nor Involved in IKK Complex Activation-
The addition of 100 M Hcy potentiated the inhibitory effect of DZA on NF-B transcriptional activity. To investigate whether Hcy can potentiate the effect of DZA on proteolytic degradation of IB␣ in RAW 264.7 cells, cells were treated with increasing concentrations of DZA in the presence or absence of 100 M Hcy. As shown in Fig. 8A, additional treatment of Hcy could not modulate the proteolytic effect of DZA on IB␣. The treatment of Hcy alone could not reduce IB␣ at all. In addition, the synergic effect of Hcy was not observed in cells treated with 100 M DZA for increasing times (Fig. 8B). These results indicate that DZA-induced proteolytic degradation of IB␣ in RAW 264.7 cells is dissociated from the accumulation of DZA-Hcy. Recently, it was reported that IKK complexes consisting of IKK␣ and IKK␤ control phosphorylation of IB proteins in cytokine-induced NF-B activation pathway (7)(8)(9). We tried to determine if IKK complexes can participate in the down-regu-

DISCUSSION
In this study, we demonstrated the dual effects of DZA on the regulation of NF-B, inhibition of NF-B transcriptional activity, and promotion of IB␣ degradation. A hypothetical diagram of the effects of DZA is illustrated in Fig. 9. There have been many reports that described some interesting properties of DZA, including the inhibitory effect on the expression of cytokines and cell adhesion molecules, the induction of apoptosis and an anti-HIV effect. The involvement of NF-B in the expression of numerous cytokines and adhesion molecules which promote several diseases is well known (1). Our previous study (26) and this paper shows that DZA potently inhibited the transcription of cytokines such as TNF-␣ and IL-1␤. Additionally, other researchers reported that the adherence of cells was inhibited through the suppression of adhesion molecule expression of CAT. Determination of CAT expression was performed using CAT enzyme-linked immunosorbent assay kit. Values are mean Ϯ S. D. (n ϭ 3).

FIG. 6. TNF-induced NF-B transcriptional activity is inhibited by DZA in COS-7 cells.
A, the effect of DZA on NF-B binding activity. The cells were pretreated with or without 100 M DZA for 1 h before stimulation, and then stimulated by the addition of recombinant GST-human TNF-␣ (50 ng/ml). Nuclear extracts were prepared at 1 h after stimulation. 5 g of each nuclear protein were subjected to a DNA binding reaction with 32 P-end-labeled NF-B consensus sequence, and then DNA-protein complexes were separated by native polyacrylamide gel electrophoresis. B, the effect of DZA on NF-B transcriptional activity. The cells were transiently transfected with reporter gene constructs carrying two copies of the wild type (J16, f) or mutant type (J32, expression by DZA (27,28). Even though they provided for the possibility of DZA as a potent therapeutic agent for diseases in which these cytokines and adhesion molecules play a central pathogenic role, the mechanism has not been elucidated. Since these cytokines and adherence molecules are included in specific target genes of NF-B, this study successfully explains how DZA inhibits the expression of cytokines and adhesion molecules.
Beg et al. (37) reported that a phenotype of the p65 Ϫ/Ϫ mice died during embryonic development through massive apoptosis of hepatocytes, suggesting that NF-B might play an important role in protecting cells from apoptosis, instead of promoting growth of cells. Additionally, recent reports (38 -41) have described a role for NF-B in blocking apoptosis induced by TNF.
Interestingly, this protective function of NF-B was apparent not only against TNF but also against cells treated with ionizing radiation or chemotherapeutic agents (39). On the other hand, the role of the Rel proteins in oncogenic transformation is quite well established, and it is now clear that Rel proteins malignantly transform and immortalize cells by inducing expression of specific genes. These findings suggest the possibility that some of the target genes for Rel-mediated oncogenesis and for NF-B-mediated blockage of apoptosis may be identical. Thus, it is strongly suggested that agents which inhibit NF-B transcriptional activity may have both direct anti-cancer effects and effects that facilitate apoptosis in tumor cells. Recently, we reported that DZA induced apoptosis in lymphocytic mouse leukemia L1210 cells (30). The report showed that even though DZA raised the NF-B binding activity during apoptosis induction, c-myc transcription that is a downstream target gene of NF-B (1,42) was markedly reduced at the early stage of apoptosis induction. Similarly, the reduction of c-myc transcription in DZA-induced apoptosis was also observed in human promyelocytic leukemia HL-60 cells by other researchers (43,44). Collectively, it is reasonable that DZA inhibition of NF-B transcriptional activity may entirely serve as a potent inducing factor or may participate at least in part of the induction of apoptosis by DZA in tumor cells. Finally, these findings point to DZA as a useful chemotherapeutic agent which may prevent malignant transformation of normal cells into tumor cells and induce apoptosis of tumor cells through the inhibition of NF-B transcriptional activity.
NF-B, whose enhancer elements are located in HIV-long terminal repeats, is a key molecule of HIV replication for the pathogenesis of AIDS (45), suggesting that NF-B may be a target for the therapy of AIDS. Previously, other researchers described that DZA may have an anti-HIV effect (21,22), but the mechanism of this effect has not been understood. The finding of DZA inhibition of NF-B transcriptional activity described in this study explains how DZA can exert an anti-HIV effect as described by other researchers. The inhibition of NF-B transcriptional activity by DZA suggests that DZA may inhibit the replication of HIV, implying that DZA may be a useful drug for the therapeutic treatment of AIDS.
Naumann and Scheidereit (46) described that p65 is strongly phosphorylated during the activation of NF-B in vivo. A recent study demonstrated that the transcriptional activity of NF-B is regulated through phosphorylation of p65 by cAMPindependent activation of a catalytic subunit of protein kinase A (PKAc) associated with IB proteins by a mechanism which signals caused degradation of IB proteins results in the activation of PKAc and subsequent phosphorylation of p65 (47). In this study, we showed that DZA disturbed the phosphorylation of p65 induced by LPS. So, we suggest a possibility that accumulated cellular DZA-Hcy, caused by treatment of cells with DZA, might inhibit the activation of PKAc associated with IB proteins, or interrupt the PKAc-mediated phosphorylation of p65 directly or indirectly. SB203580, a specific inhibitor of p38 mitogen-activated protein kinase, and elevated intracellular cAMP has been represented as an inhibitor of NF-B, which inhibit NF-B transcriptional activity already bound to DNA in the nucleus (16,17). Even though, DZA and these inhibitors share a similarity that all of them inhibit NF-B transcriptional activity without affecting its nuclear translocation and DNA binding activity, there is a definite difference in the mechanism of NF-B inhibition. In contrast to SB203580 and elevated cellular cAMP, DZA inhibits the phosphorylation of p65.
This study indicates that DZA inhibits NF-B transcriptional activity through a hindrance of p65 phosphorylation by FIG. 8. DZA-induced proteolytic degradation of IB␣ is neither modulated by Hcy nor involved in IKK complex activation. A, RAW 264.7 cells were treated with increasing concentrations of DZA for 2 h in the presence or absence of 100 M Hcy, and then cytosolic fractions were prepared. 10 g of each protein was subjected to Western blot analysis using specific antibody against IB␣. B, the cells were treated for the indicated times with 100 M DZA, in the absence or presence of Hcy (100 M). At the end of the times, cytosolic fractions were prepared. 10 g of each protein was separated by SDS-PAGE, and then Western blot analysis with specific antibody against IB␣ was performed. C, the cells were treated with 100 M DZA or LPS (1 g/ml) for the indicated times. At the end of the times, total cell lysates were prepared, and IKK complexes were recovered by immunoprecipitation. IKK assay was accomplished using recombinant protein of GST-IB␣ as a substrate. Similar results were obtained in an independent experiment. the accumulation of DZA-Hcy in cells. Nevertheless, we cannot totally rule out that DZA independent of cellular DZA-Hcy accumulation, may have interfered with NF-B transcriptional activity. In the report by Backlund et al. (48), treatment with 100 M DZA resulted in significant accumulation of DZA-Hcy in RAW 264.7 cells. The addition of Hcy to DZA increased the accumulation of DZA-Hcy to about 5 times the level caused by DZA alone. In contrast to a greatly increased accumulation of DZA-Hcy by the addition of Hcy to DZA treatment in the cells, we observed a slightly potentiated inhibition of NF-B transcriptional activity by the addition of Hcy to DZA.
We described that DZA promotes the proteolytic degradation of IB␣, but not IB␤ in RAW 264.7 cells, leading to an increase of nuclear translocation and DNA binding activity. Increased DNA binding activity of NF-B in the nucleus by DZA was also observed in DZA-induced apoptosis in L1210 cells (30), and DZA might reduce IB␣ in this case. However, a DZA-mediated increase of NF-B DNA binding activity in the nucleus was not discernable in COS-7 cells, suggesting that DZA-mediated IB␣ degradation is cell-type specific. A number of kinases have been suggested to phosphorylate IB␣, and recently a high molecular mass, approximately 700 kDa, kinase complex termed IKK (7-9) was identified. However, DZA mediates IB␣ degradation through an IKK-independent pathway and DZA can stimulate other kinase pathways. We conclude this because IKK was not activated at all by DZA while it was intensely activated by LPS in this study. One notable fact is that DZA stimulates proteolytic degradation of IB␣ but not IB␤. Among kinases which have been reported to induce degradation of IB proteins, including IKK, none can selectively differentiate between IB␣ and IB␤. Thus, we expect that DZA might be a valuable probe to identify kinases that function with only IB␣.
In conclusion, we demonstrate that DZA has dual effects on NF-B regulation. One is an inhibitory effect on NF-B transcriptional activity already bound to target DNA, and another is the promotion of proteolytic degradation of IB␣. Although the wide variety of biological properties observed with DZA has emphasized that DZA is to be an effective drug for treatment of human diseases including inflammation, infections, and tumors, the mechanism and target molecules of DZA's action have never been elucidated. This study is the first to demonstrate that NF-B is a specific target molecule of DZA and suggests that DZA may be a potent drug for the treatment of diseases in which NF-B plays an important pathogenic role, as well as a useful tool for studying the regulation and physiological functions of NF-B.