Inhibitors of nuclear factor kappa B cause apoptosis in cultured macrophages

The precise role of the transcription factor nuclear factor kappa B (NF- κB) in the regulation of cell survival and cell death is still unresolved and may depend on cell type and position in the cell cycle. The aim of this study was to determine if three pharmacologic inhibitors of NF-κB, pyrrolidine dithiocarbamate, N-tosyl-L-lysl chloromethyl ketone and calpain I inhibitor, induce apoptosis in a murine macrophage cell line (RAW 264.7) at doses similar to those required for NF-κB inhibition. We found that each of the three inhibitors resulted in a dose- and time-dependent increase in morphologic indices of apoptosis in unstimulated, LPS-stimulated and TNF-stimulated cells. Lethal doses were consistent with those required for NF- κB inhibition. We conclude that nuclear NF-κB activation may represent an important survival mechanism in macrophages.


Introduction
Nuclear factor kappa B (NF-kB) is a complex of dimeric transcription factors (p50, p65, c-rel, Rel B) activated by tumor necrosis factor (TNF), reactive oxygen species, phorbol myristate acetate (PMA), lipopolysaccharide (LPS), as well as serine/threonine and tyrosine kinases. NF-kB plays a signi cant role in immunity and in ammation; genes regulated by NF-kB include interleukins 1, 2, 6 and 8, inducible nitric oxide synthase, the major histocompatibility complex Class I, the adhesion molecules, ICAM, VCAM and ELAM, and the immunoglobulin k light chain (for a review see reference 1). In addition, there is now data to suggest that NF-kB may also play a role in the regulation of cell survival, cell death and oncogenesis: 2 for example, inhibition of p65, a transcriptionally active subunit of NF-kB, with antisense messenger RNA has been shown to inhibit tumorigenicity of transformed cell lines and tumor growth in animals; 3 Bcl-3, a molecule related to I-kB and which is unable to inhibit nuclear binding of the p65 subunit, has been linked to the development of B-cell lymphomas. 4 The precise role of NF-kB in the regulation of cell death is still a matter of controversy. Initially, several lines of evidence appeared to indicate a role for the excessive activation of NF-kB as a cause of apoptosis: nuclear binding of transcriptionally active NF-kB p50/ p65 heterodimers was noted to follow initiation of apoptosis in a variety of cell lines, [5][6][7][8] and three distinct pharmacologic inhibitors of NF-kB, pyrrolidine dithiocarbamate (PDTC), N-tosyl-L-lysyl chloromethylketone (TPCK) and N-acetyl cysteine, were shown to inhibit apoptosis in HL-60 cells, human thymocytes, prostate carcinoma and neuroblastoma cell lines. 9,10 In some of these experiments, however, the doses of these agents required to inhibit apoptosis were lower than those required to inhibit NF-kB, suggesting that their protective effects were not necessarily due to their ability to inhibit NF-kB. 11 The ability of the TNF-receptor both to activate NF-kB and to execute an apoptosis program were initially suspected to be causally related, futher supporting a pro-apoptotic function for NF-kB. Recent studies, however, have demonstrated that the initiation of apoptosis and the activation of NF-kB by TNF are separate, and possibly, mutually exclusive events. 12 -14 In spite of the fact that various compounds which share the ability to inhibit NF-kB including PDTC, N-acetyl cysteine, genistein, dexamethasone and staurosporine can cause apoptosis at doses similar to those required for NF-kB inhibition, 15 -21 inhibition of NF-kB itself had not been proposed as a mechanism of apoptosis until very recently. The rst evidence in favor of a pro-survival role for NF-kB came from a study of mice with a homozygous deletion of the gene for the p65 component of NF-kB who underwent early embryonic death characterized by apoptosis of liver cells. 22 A second piece of evidence came from the nding of an NF-kB binding site on the promoter region of A20, a zinc nger protein which inhibits apoptosis. 23 Finally, three separate reports were published recently which demonstrated prevention of TNF-mediated apoptosis by induction of NF-kB in vitro. 24 -26 In a fourth study, the NF-kB inhibitor, TPCK was found to induce apoptosis in unstimulated WEHI B-cell lymphoma cells. 27 Although this evidence supports a pro-survival role for NF-kB, other recent studies have demonstrated that the precise relationship between NF-kB and cell death is still unresolved. Grilli et al. for example, have reported that aspirin and sodium salicylate protect against glutamate-induced apoptosis by inhibiting NF-kB in preparations of primary neurons. 28 Similarly, Tsai et al. have shown that PDTC protects neurons from cell death at the same doses at which it causes apoptosis in smooth muscle cells. 15 These con icting results point to the need for further studies to de ne the conditions and cell types in which NF-kB inhibitors either cause or inhibit apoptosis. The answers to these questions are of particular importance because NF-kB inhibitors have been proposed as potential therapeutic agents for in ammatory disease 29 and cancer. 25 The aim of the current study was to determine if three pharmacologically distinct inhibitors of NF-kB (PDTC, TPCK, and Calpain I inhibitor) induce apoptosis in a transformed murine macrophage cell line at doses required for the inhibition of NF-kB. We also sought to determine if activation of the cells with either TNF, LPS or PMA could reverse the effects of NF-kB inhibitors.

Materials and Methods
Reagents RAW 264.7 cells, from a transformed murine peritoneal macrophage cell line, were obtained from ATCC (Rockville, MD); recombinant murine TNF-a was purchased from R&D Systems (Minneapolis, MN); [c-32 P]ATP and Hyper lm were obtained from Amersham Life Sciences (Arlington Heights, IL); calpain I inhibitor was purchased from ICN (Costa Mesa, CA); Cellular DNA Flow Cytometric Analysis Kit was obtained from Boehringer Mannheim (Indianapolis, IN); polynucleotide kinase and kinase buffer for electrophoretic mobility shift assays were obtained from Promega (Madison, WI); NF-kB consensus oligonucleotide was obtained from Stratagene (La Jolla, CA); fetal calf serum was purchased from HyClone (Logan, UT). All other reagents were purchased from Sigma (St. Louis, MO).

Electrophoretic mobility shift assay
Nuclear extracts were prepared as previously described. 30 After determining the protein content of nuclear assays using a standard assay (Bio-RAD, Hercules, CA), 2 mg of each sample was incubated for 15 min with 5 ml of binding buffe r (5 3 TBE; 25% glycerol), 1 ml poly dI:dC (0·25 mg/ ml) and 1 ml 50 mM dithiothreitol. Samples were then incubated for 30 min with a 32 P-labeled consensus NF-kB binding oligonucleotide. To ensure speci city of binding, negative control samples were coincubated with unlabeled oligonucleotide ('cold competitor'). One ml of gel 10 3 loading buffer (250 mM Tris-HCl, pH 7.5, 0.2% bromophenol blue, 40% glycerol) was added to each sample and samples were loaded on a 9% polyacrylamide nondenaturing gel and run at 190 V for 3 ±4 h in 0.25 3 TBE. Gels were exposed to X-ray lm at -708 C overnight.

Cell death assays
Cell death by apoptosis was quanti ed by one of two methods. Cells from all treatment groups were resuspended in PBS containing 5 mg/ ml of acridine orange and examined under a uorescent microscope for morphologic changes consistent with apoptosis (nuclear condensation and fragmentation, membrane blebbing). The presence of apoptosis was con rmed by performing the TUNEL end-labeling assay which measures double-stranded DNA breaks by ow cytometry. 31 Brie y, cells were washed twice in cold PBS/1% BSA and then xed with 100 ml of 4%paraformaldehyde sollution for 30 min. Cells were centrifuged at 188 3 g , washed twice with PBS/1% BSA, and resuspended in 100 ml of permeabilization solution (0.1% Triton X-100 in 0.1% sodium citrate) for 2 min on ice. After washing twice, cells were resuspended in 50 ml of TUNEL reaction mixture (terminal deoxynucleotidyl transferase and nucleotide mixture) and incubated for 60 min at 378 C in a dark, humidi ed chamber. After washing samples twice in PBS, cells were resuspended in 250 ± 500 ml of PBS and analyzed by ow cytometry. Cells undergoing apoptosis were expressed as a percentage of total cells.

Statistical analysis
Comparisons among treatment groups were made by ANOVA and the Tukey ±Kramer Multiple Comparisons Test using Instat (GreaftPad Software, San Diego, CA) software.

Induction of apoptosis by NF-kB inhibitors
Apoptosis was assessed at 18 h for each inhibitor. Treatment with each of the three inhibitors resulted in an increase in the percentage of cells undergoing apoptosis as assessed by acridine orange (Fig. 1) and TUNEL staining (Fig. 2). Compared to no treatment, PDTC, TPCK, and calpain I inhibitor each caused signi cant, dosedependent increases in apoptosis as assessed by acridine orange staining (Fig. 1). Calpain I inhibitor was the most effective of the three agents at killing cells with an EC 50 of between 5 and 10 mM. PDTC was the least effective of the three agents with an EC 50 of approximately 100 mM. The EC 50 for TPCK was approximately 40 mM (Fig. 1). Apoptosis induction by each of the inhibitors at 18 h was con rmed using ow cytometry and the TUNEL assay (Fig. 2).

Time course of apoptosis
The mean percentage of cells undergoing apoptosis as assessed by acridine orange staining at 6, 12, 18 and 24 h for each inhibitor was calculated (Fig. 3). Maximal induction of apoptosis by TPCK (50 mM) and PDTC (100 mM) occurred at 18 h. Calpain I inhibitor (10 mM) induced maximal killing at 12 h. Apoptosis became detectable at about 6 h post-treatment and began to decrease by 24 h post-treatment for all inhibitors. There was little morphologic evidence of apoptosis prior to 6 h for any of the inhibitors tested (data not shown).

Effect of PMA
Experiments were conducted to determine if induction of apoptosis by NF-kB inhibitors could be overcome by concurrent signals which normally induce NF-kB. Neither TNF nor LPS afforded signi cant protection against cell death caused by any of the three inhibitors (data not shown). PMA (10 nM) led to a reduction in the amount of apoptosis induced by PDTC   (100 mM), TPCK (50 mM) and calpain I inhibitor (10 mM) (Fig. 4). Only in the case of calpain I inhibitor, however, was this reduction signicant ( P , 0·001).

Effect of cycloheximidine
Cycloheximide has been reported to both prevent and induce apoptosis, presumably depend-ing on whether it inhibits transcription of death genes or of survival genes. 32,33 In our experiments, cycloheximidine (10 mg/ ml) offered no protection against apoptosis caused by PDTC (100 mM), TPCK (50 mM) and calpain I inhibitor (10 mM) at 18 h (Fig. 5). By itself, however, cycloheximide also induced extensive apoptosis at 18 h. These results suggest that cycloheximide may be blocking induction of a prosurvival gene in an additive fashion with NF-kB inhibitors.

Effects of inhibitors on NF-kB binding by EMSA
To determine if apoptosis occurred at the same doses as NF-kB inhibition by PDTC, TPCK and calpain I inhibitor, electrophoretic mobility shift assays were performed using nuclear extracts from cells incubated for 18 h with different doses of each of the three inhibitors (Fig. 6). PDTC, at doses of 50 mM and higher, effectively inhibited the upper band of the two-band NF-kB characteristically seen on EMSA gels. The upper band corresponds to p50 ±p65 heterodimers. The lower band, which was not inhibited by PDTC, corresponds to p50 ±p50 homodimers, considered to be transcriptionally inactive. At doses above 10 mM, TPCK inhibited the bands corresponding to both the p50 ±p65 heterodimer and the p50 ±p50 homodimer. Interestingly, in some experiments, there was formation of a new band of higher molecular weight than the p50 ±p65 heterodimer which appeared at doses  of TPCK of 10 mM and higher. Calpain I inhibitor inhibited NF-kB (both p50 ±p50 and p50 ± p65 dimers) binding at doses of 1 mM and higher. As with TPCK, an unidenti ed higher molecular weight band appeared at doses of 10 mMcalpain I inhibitor and higher.
These results show that the EC 50 for the induction of apoptosis by all three inhibitors was slightly higher than the minimum dose required for NF-kB inhibition, indicating that NF-kB inhibition and apoptosis occur within the same dose range. Calpain I inhibitor was the most effective inhibitor of NF-kB, and PDTC was the least effective, re ecting the order of ef cacy of these compounds in inducing apoptosis. Inhibition of NF-kB (maximal at 2 h) occurred prior to morphologic evidence of apoptosis (maximal at 18 h). This suggests that an event downstream of NF-kB inhibition (such as inhibition of gene transcription) would be required to provide a causal link between the two events.

Discussion
We have shown that three pharmacologically distinct inhibitors of NF-kB ± PDTC, TPCK and calpain I inhibitor ± each causes apoptosis of RAW 264.7 murine macrophages in a dosedependent fashion. Further, doses of these inhibitors which cause apoptosis are similar to those which inhibit NF-kB binding as detected by electrophoretic mobility shift assay. Simultaneous activation of the cells with either LPS or TNF fails to reverse apoptosis. Simultaneous treatment with PMA partially reverses apoptosis but the effect is only signi cant in calpain I inhibitor-treated cells.
The ability of TPCK, calpain I inhibitor and PDTC to cause apotosis has been previously documented in other cells lines, 14, 34 -37 but an association with inhibition of NF-kB had not been made until very recently when four simultaneous reports were published demonstrating a direct association between NF-kB inhibition and apoptosis. 24 -27 Three reports (based on studies conducted in broblasts, macrophages, Jurkat T cells, human bladder cancer cells and human brosarcoma cells) showed that NF-kB activation suppresses TNF-a-induced apoptosis and that NF-kB inhibition (either by transfection with a dominant-negative I-kB alpha or by 'knock-out' of RelA), augments TNF-induced apoptosis. 24 -26 Our study con rms these results in macrophages and extends them to unstimulated, and LPS-stimulated, cells. In a fourth study, inhibition of NF-kB by TPCK induced apoptosis in WEHI 121 B-cell lymphoma cells even in the absence of TNF-stimulation. 27 Interestingly, as in the case of RAW cells, WEHI cells have constitutive NF-kB activation.
TPCK and calpain I inhibitor are both protease inhibitors. TPCK is an inhibitor of serine chymotrypsin-like proteases and has been shown to inhibit both the phosphorylation and proteolytic degradation of I-kB. 38 Calpain I inhibitor is an inhibitor of neutral cysteine proteases and is known to prevent I-kB proteolytic degradation. 39 There is some evidence to suggest that calpain itself may be involved in the regulation of I-kB degradation. 40 However, it is also possible that the apoptosis-inducing effects of these two inhibitors could derive from their ability to inhibit the 26S proteasome, a multifraction protease which degrades and recycles a variety of intracelullar proteins. Indeed, the ability of PMA partially to reverse apoptosis induced by calpain I inhibitor may be related to its ability to activate the proteasome, most likely downstream from the effects of calpain I inhibitor. The proteasome's role in NF-kB activation is to degrade phosphorylated, ubiquitinconjugated, cytoplasmic I-kB, therby releasing NF-kB dimers for nuclear binding. 41 The proteasome is also involved in the degradation of other proteins which are critical for cell cycle regulation including cyclins, p53 and protein kinase C. 42 -44 Thus, it is not possible to determine from our data whether the effects of the two protease inhibitors tested here are due to the inhibition of NF-kB per se, or to inhibition of the degradation of another protein such as p53 36  that the effects of TPCK and calpain I inhibitor on apoptosis may be directly related to their ability to inhibit NF-kB transloction to the nucleus.
In some cell types, both TPCK and calpain I inhibitor have been shown to be effective inhibitors rather than inducers of apoptosis. 11, 45 -49 Their dual effects on apoptosis may derive from their ability to inhibit proteases involved in the characteristic nucleosomal cleavage of apoptosis independently of their ability to inhibit NF-kB. 45,46,49 Alternatively, it is possible that both underactivation and overactivation of NF-kB may each cause apoptosis, giving NF-kB inhibition a dual role in apoptosis. It is well known, for example, that both excessive cellular activation ('activation-induced' cell death) and insuf cient cellular stimulation (growthfactor withdrawal) can each lead to cell death, most likely by different mechanisms. Finally, the divergent effects of these protease inhibitors on apoptosis may be dependent both on cell type and position in the cell cycle. The possibility exists, for example, that cycling cells are susceptible to apoptosis by these agents whereas terminally differentiated cells arrested in G 0 / G 1 are not.
The mechanism by which PDTC inhibits NF-kB binding is still unclear. However, it has been suggested that its inhibitory effects of NF-kB are due to its metal-binding and/or antioxidant properties. 50,51 Oxidants may activate NF-kB through a mechanism independent of the proteasome. An oxidant-sensitive tyrosine kinase had recently been shown to release I-kB from NF-kB without associated I-kB degradation. 52 It is unlikely that PDTC's actions are similar to those of either TPCK or calpain I inhibitor. In unstimulated RAW cells, TPCK and calpain I inhibitor inhibited both bands of the NF-kB complex, presumably corresponding to p50 ± p65 and p50 ±p50 dimers. However, in our hands, PDTC only inhibited the higher molecular weight band corresponding to the p50 ±p65 dimer. In addition, administration of TPCK and calpain I inhibitor results in the appearance of a new unidenti ed band on EMSA of higher molecular weight than the p50 ±p65 heterodimer. The fact that PDTC only inhibits p50 ± p65 dimers but still induces apoptosis, supports the hypothesis, raised by ndings in the p65 kock-out mouse, 21 that p65 but not p50 is critical for cell survival.
It is interesting that PDTC was substantially less effective than TPCK or calpain I inhibitor in inducing apoptosis in our cell line. One reason may be that higher doses of PDTC are required to inhibit NF-kB. In addition, anti-oxidants such as PDTC may be less potent inducers of apoptosis than protease inhibitors by virtue of their ability to scavenge free radicals. Finally, PDTC (but not, to our knowledge, TPCK or calpain I inhibitor) has the ability to induce metallothionein (unpublished work), a metalregulatory protein which itself inhibits apoptosis. 53 Our ndings regarding PDTC's ability to induce apoptosis are consistent with the recently reported results of Tsai et al. who found that PDTC, at doses of 25 ±150 mM, induced apoptosis in rat and human aortic smooth muscle cells. 15 These authors concluded that PDTC's ability to induce apoptosis might be related to its antioxidant properties because similar results were obtained with N-acetyl cysteine. When serum-deprived PC12 cells were used in the same study, however, PDTC was protective at doses of 100 mM. It is clear from these divergent results that the actions of PDTC, like TPCK and calpain I inhibitor, are dependent on the cell type and cell cycle conditions. PDTC's ability to induce death in smooth muscle cells and macrophages, two cell types which proliferate on conditions of chronic in ammation, (while sparing nerve cells and possibly other terminally differentiated cells) suggests that PDTC and other NF-kB inhibitors might be useful anti-in ammatory agents.
Although the three compounds tested in this study inhibited NF-kB within 2 h, there was a time delay of 6 h before the appearance of apoptosis. This suggests that they are regulating gene transcription and protein translation of downstream effectors of cell survival or death. We used cycloheximide to test the hypothesis that the three compounds might be inducing the transcription of a programed cell death gene, since cycloheximide has been reported to block apoptosis caused by other agents. 33 Cycloheximide not only failed to block apoptosis caused by each of the three NF-kB inhibitors but instead exacerbated it. This suggests the alternative, and more likely, hypothesis that PDTC, TPCK and calpain I inhibitor are inhibiting the transcription of a gene required for cell survival. One candidate for the putative survival gene being repressed by the three inhibitors is A20, a zinc nger binding protein with an NF-kB site on the promoter known to protect cells from undergoing apoptosis. 24 Recent data from Beg and Baltimore, however, show that transfection of Rel A-de cient 3T3 cells with A20 is unable to prevent TNF-induced apoptosis. 24 In contrast, transfection of WEHI cells with c-myc and Bcl-x L can attenuate apoptosis induced by TPCK 27 suggesting that Bcl-x L and/or c-myc may be downstream targets of NF-kB which promote cell survival.
In summary, our data from a transformed murine peritoneal macrophage cell line add to a recent and rapidly growing body of evidence linking NF-kB inhibition to induction of apoptosis in in ammatory and immune cells. Induction of apoptosis in immune effector cells may be a therapeutic mechanism of NF-kB inhibitors.