Original Paper | Published:

Nonsteroidal anti-inflammatory agents differ in their ability to suppress NF-κB activation, inhibition of expression of cyclooxygenase-2 and cyclin D1, and abrogation of tumor cell proliferation

Oncogene volume 23, pages 92479258 (09 December 2004) | Download Citation

Subjects

Abstract

Nonsteroidal anti-inflammatory drugs (NSAIDs) such as aspirin have been shown to suppress transcription factor NF-κB, which controls the expression of genes such as cyclooxygenase (COX)-2 and cyclin D1, leading to inhibition of proliferation of tumor cells. There is no systematic study as to how these drugs differ in their ability to suppress NF-κB activation and NF-κB-regulated gene expression or cell proliferation. In the present study, we investigated the effect of almost a dozen different commonly used NSAIDs on tumor necrosis factor (TNF)-induced NF-κB activation and NF-κB-regulated gene products, and on cell proliferation. Dexamethasone, an anti-inflammatory steroid, was included for comparison with NSAIDs. As indicated by DNA binding, none of the drugs alone activated NF-κB. All compounds inhibited TNF-induced NF-κB activation, but with highly variable efficacy. The 50% inhibitory concentration required was 5.67, 3.49, 3.03, 1.25, 0.94, 0.60, 0.38, 0.084, 0.043, 0.027, 0.024, and 0.010 mM for aspirin, ibuprofen, sulindac, phenylbutazone, naproxen, indomethacin, diclofenac, resveratrol, curcumin, dexamethasone, celecoxib, and tamoxifen, respectively. All drugs inhibited IκBα kinase and suppressed IκBα degradation and NF-κB-regulated reporter gene expression. They also suppressed NF-κB-regulated COX-2 and cyclin D1 protein expression in a dose-dependent manner. All compounds inhibited the proliferation of tumor cells, with 50% inhibitory concentrations of 6.09, 1.12, 0.65, 0.49, 1.01, 0.19, 0.36, 0.012, 0.016, 0.047, 0.013, and 0.008 mM for aspirin, ibuprofen, sulindac, phenylbutazone, naproxen, indomethacin, diclofenac, resveratrol, curcumin, dexamethasone, celecoxib, and tamoxifen, respectively. Overall these results indicate that aspirin and ibuprofen are least potent, while resveratrol, curcumin, celecoxib, and tamoxifen are the most potent anti-inflammatory and antiproliferative agents of those we studied.

Introduction

Owing to its analgesic and anti-inflammatory effects, aspirin (acetylsalicylic acid), first synthesized in 1897, was approved for the treatment of rheumatoid arthritis in 1899 (Botting, 1999). Since then several other nonsteroidal anti-inflammatory drugs (NSAIDs) have been synthesized and approved for human use. The anti-inflammatory and analgesic effects of NSAIDs have been shown to be due to their ability to inhibit the enzymatic activity of cyclooxygenases (COXs), which convert arachidonic acid to prostaglandins (PGs) (Vane, 1971). Aspirin is one of the most widely used drugs, with an average yearly consumption of 30 g/person in the industrialized countries (Roth and Calverley, 1994) and a daily consumption in the United States alone of 35 000 kg (Jack, 1997).

In most cases, aspirin or another NSAID either prevents or delays the onset of the disease. Extensive research has shown that aspirin has a potential in the treatment of a wide variety of inflammatory diseases including cancer, arthritis, cardiovascular diseases, and Alzheimer's disease (O'Neill, 1998). For instance, aspirin has been found effective against colon cancer, especially familial adenomatous polyposis coli, a hereditary precancerous disease caused by the loss of the tumor suppressor gene adenomatous polyposis coli (Giercksky, 2001). Both preventive and therapeutic affects of aspirin have been reported against colorectal adenomas (Baron et al., 2003; Sandler et al., 2003). Moysich et al. (2002) found that regular aspirin use may be associated with reduced risk of lung cancer. When the analyses were restricted to former and current smokers, participants with the lowest cigarette exposure tended to benefit most from the potential chemopreventive effect of aspirin.

How aspirin and other NSAID mediate their effects is not fully understood. Whether inhibition of COX-1 (constitutive) or COX-2 (inducible) activity is a major mechanism of its action has been questioned. For instance, Goel et al. (2003) showed that growth inhibition of human colon cancer cells by aspirin is through COX-independent mechanisms and is due to the change in expression of DNA mismatch repair proteins (Goel et al., 2003). Similarly, Williams et al. (2000) showed that the cytotoxic effects of celecoxib, another NSAID, were independent of COX-2 inhibition because similar effects were observed in COX-2 (+/+), COX-2 (+/−) and COX-2 (−/−) fibroblasts.

Generally, a strong correlation has been found between inflammation and cancer (Coussens and Werb, 2002), suggesting that cancer is an inflammatory disease. Celecoxib, however, has been shown to reduce pulmonary inflammation but not lung tumorigenesis in mice (Kisley et al., 2002). Thus, novel targets to explain the effects of NSAID are being sought. Nuclear factor NF-κB is one such target that has been shown to mediate inflammation and suppress apoptosis, and it is commonly overexpressed in cancer cells (Bharti and Aggarwal, 2002). The expression of most genes that are involved in inflammation (e.g. COX-2) or in cellular proliferation (e.g. cyclin D1) are regulated by NF-κB. Thus, both antiproliferative and anti-inflammatory effects of NSAID can be explained through their ability to suppress NF-κB. Although aspirin was the first NSAID shown to suppress NF-κB (Kopp and Ghosh, 1994), now several NSAID have been shown to suppress NF-κB activation (Kazmi et al., 1995; Scheuren et al., 1998; Yamamoto et al., 1999; Chuang et al., 2002; Bryant et al., 2003).

How various NSAID differ in their ability to suppress NF-κB activation, NF-κB-regulated gene expression, and proliferation has not been examined and is the focus of the current study. We examined 11 different NSAID for inhibition of tumor necrosis factor (TNF)-induced NF-κB activation, IκBα degradation, NF-κB reporter gene expression, expression of cyclin D1 and COX-2, and antiproliferative effects.

Results

In the present study, we investigated the relative anti-inflammatory and antiproliferative effects of most commonly used NSAIDs by examining their effects on TNF-induced NF-κB activation, IκBα kinase, IκBα degradation, NF-κB-regulated reporter gene expression, NF-κB regulating proteins, and proliferation of tumor cells. We examined the effect of the oldest (aspirin) and the most recently identified (celecoxib) NSAIDs. In all, 11 different NSAIDs, along with dexamethasone, were investigated. The similarities and dissimilarities in the chemical structure of these NSAIDs are shown in Figure 1.

Figure 1
Figure 1

Chemical structure of NSAIDs. The molecular weight of each NSAID is indicated within parentheses

NSAIDs inhibit TNF-induced NF-κB activation

Activation of NF-κB is one of the earliest events that mediate inflammation induced by various stimuli in most cells (Bharti and Aggarwal, 2002). Whether all NSAIDs inhibit NF-κB activation induced by TNF, a most potent inflammatory stimulus, was investigated. Cells were pretreated with various concentrations of NSAID either for 4 or 8 h (phenylbutazone, resveratrol and diclofenac), and then activated for NF-κB by treatment with 0.1 nM TNF for 30 min. Nuclear extracts were prepared and analysed for NF-κB activation by electrophoretic mobility shift assay (EMSA). As shown in Figure 2, none of the NSAIDs by themselves activated NF-κB. TNF-induced NF-κB activation, however, was inhibited by all NSAID in a dose-dependent manner. The 50% inhibitory concentration required was 5.67, 3.49, 3.03, 1.25, 0.94, 0.60, 0.38, 0.084, 0.043, 0.027, 0.024, and 0.01 mM for aspirin, ibuprofen, sulindac, phenylbutazone, naproxen, indomethacin, diclofenac, resveratrol, curcumin, dexamethasone, celecoxib, and tamoxifen, respectively (Table 1). Based on IC50, aspirin was least potent and tamoxifen was most potent (567-fold). The IC50 of celecoxib was comparable with dexamethasone (236- vs 210-fold).

Figure 2
Figure 2

Effect of NSAIDs on TNF-induced NF-κB activation. (a) KBM-5 cells were pretreated with various concentrations of NSAIDs either for 4 or 8 h (phenylbutazone, resveratrol, and diclofenac), and then treated with 0.1 nM TNF for 30 min. Nuclear extracts were prepared and analysed for NF-κB activation by EMSA as described in Materials and methods. (b) NF-κB induced by TNF is composed of p65 and p50 subunits. Nuclear extracts from untreated or 0.1 nM TNF-treated KBM-5 cells were incubated with different antibodies, unlabeled NF-κB oligo-probe or mutant oligo-probe, and then assayed for NF-κB activation by EMSA. Aspirin (c) and curcumin (d) had no effect on Oct-1 in KBM-5 cells. Cells were preincubated with the indicated concentrations of aspirin or curcumin for 4 h, and then treated with 0.1 nM TNF for 30 min. Nuclear extracts were prepared and analysed for Oct-1 activation by EMSA. Result shown are representative of two or three separate experiments

Table 1: NSAIDs inhibit TNF-induced NF-κB activation and on cell proliferation

Since NF-κB is a family of proteins, various combinations of Rel/NF-κB protein can constitute an active NF-κB heterodimer that binds to a specific sequence in DNA (Ghosh and Karin, 2002). To show that the retarded band visualized by EMSA in TNF-treated cells was indeed NF-κB, we incubated nuclear extracts from TNF-activated cells with antibodies to either the p50 (NF-κB1) or the p65 (RelA) subunit of NF-κB. Both shifted the band to a higher molecular mass (Figure 2b), thus suggesting that the TNF-activated complex consisted of p50 and p65 subunits. Preimmune serum (PIS) had no effect on NF-κB complex. Excess unlabeled NF-κB oligonucleotide (100-fold) caused complete disappearance of the band and no effect by using mutant NF-κB oligonucleotide.

To further determine the specificity of the effect of NSAIDs on NF-κB, we examined its effect on the transcription factor Oct-1. For this, cells were treated with the indicated concentrations of aspirin or curcumin for 4 h, the nuclear extracts were prepared, and the extracts examined for binding to labeled oligonucleotides containing specific binding sites for Oct-1 by EMSA. Oct-1 bound to their specific DNA sequence, and the treatment of cells with neither aspirin (Figure 2c) nor curcumin (Figure 2d) affected the activity of Oct-1. These results indicate that the effects of NSAIDs on NF-κB activation are specific.

NSAIDs inhibit TNF-dependent IκBα degradation and IκB kinase activity

The activation of NF-κB is known to require the phosphorylation, ubquitination, and degradation of IκBα, the natural inhibitor of NF-κB (Ghosh and Karin, 2002). To determine whether inhibition of TNF-induced NF-κB activation was due to the inhibition of IκBα degradation, we pretreated cells with various concentrations of NSAIDs for either 4 or 8 h (phenylbutazone, resveratrol, and diclofenac), and then exposed them to 0.1 nM TNF for 30 min. The cell extracts were then examined for IκBα status in the cytoplasm by Western blot analysis. TNF induced IκBα degradation, but in NSAID-pretreated cells, TNF failed to induce the degradation of IκBα (Figure 3a). The effect was dose-dependent. Based on IC50, like NF-κB activation, aspirin again was least potent and tamoxifen was most potent in suppressing IκBα degradation.

Figure 3
Figure 3

Effect of NSAIDs on TNF-induced IκBα degradation and IKK activation. KBM-5 cells were pretreated with various concentrations of NSAIDs either for 4 or 8 h (phenylbutazone, resveratrol, and diclofenac), and then treated with 0.1 nM TNF for 15 min (a) or 1 nM TNF for 5 min (b). Cytoplasmic extracts were prepared, resolved on 10% SDS–PAGE, and then electrotransferred to a nitrocellulose membrane. Western blot analysis was performed using anti-IκBα and anti-β-actin antibodies as described in Materials and methods (a). Whole-cell extracts were prepared, immunoprecipitated with anti-IKK-α antibody, and then subjected to IKK kinase assay as described in Materials and methods (b). For loading control, whole-cell extracts were also resolved on 7.5% SDS–PAGE, and performed Western blot analysis using anti-IKK-α antibody. Results shown are representative of two or three separate experiments

Since IκBα degradation requires IκBα kinase (IKK) activation, we also examined the effect of various NSAIDs on TNF-induced IKK activity. Cells were treated with NSAIDs as described above and then treated with 1 nM TNF for 5 min. Whole-cell extracts were prepared, immunoprecipitated with anti-IKK-α antibody, and then subjected to IKK kinase assay. The results indicate that TNF activated IKK and all NSAIDs inhibited TNF-induced IKK kinase activity (Figure 3b).

NSAIDs inhibit TNF-induced NF-κB-dependent reporter gene expression

Although we have shown that NSAIDs block the DNA-binding step in NF-κB activation, DNA binding alone does not always correlate with NF-κB-dependent gene transcription, suggesting that there are additional regulatory steps (Nasuhara et al., 1999). To determine the effect of NSAIDs on TNF-induced NF-κB-dependent reporter gene expression, we transiently transfected A293 cells with the NF-κB-regulated secretory alkaline phosphatase (SEAP) reporter construct, incubated them with NSAIDs, and then stimulated the cells with TNF. An almost eight-fold increase in SEAP activity over the vector control was observed upon stimulation with 1 nM TNF (Figure 4). NSAIDs repressed NF-κB-dependent reporter gene expression induced by TNF in a dose-dependent manner. In most cases, the dose required to suppress NF-κB-dependent reporter activity was similar to that needed for inhibition of NF-κB DNA-binding activity. Thus, these results suggest that NSAIDs also suppressed TNF-induced NF-κB-dependent gene expression.

Figure 4
Figure 4

Effect of NSAIDs on TNF-induced gene expression. A293 cells were plated in 12-well plates and transiently transfected with NF-κB-SEAP plasmid along with β-actin-Renilla plasmid (transfection control) by the FuGENE 6 as described in Materials and methods. Then cells were pretreated with various concentrations of NSAIDs either for 4 or 8 h (phenylbutazone, resveratrol, and diclofenac) and exposed to 1 nM TNF. The cell culture medium was harvested after 24 h and analysed for SEAP activity as described in Materials and methods. The cells were also lysed and analysed for luciferase activity as described in Materials and methods. The results are expressed as fold with basal NF-κB activity in the untreated samples as one. The activity was normalized against the β-actin-Renilla for transfection control. The results shown are the average±s.d. of two separate experiments. The significance of results when compared with the TNF-treated cells are *P>0.1, **P<0.1, ***P<0.01, ****P<0.001

NSAIDs inhibit TNF-induced and NF-κB-regulated protein expression

Through activation of NF-κB, TNF induces cyclin D1 and COX-2, both of which have NF-κB-binding sites in their promoters (Yamamoto et al., 1995; Guttridge et al., 1999; Hinz et al., 1999). The suppression of COX-2 is essential for the anti-inflammatory effects of NSAIDs, whereas the suppression of cyclin D1 explains their antiproliferative effects. To determine whether NSAIDs inhibit TNF-induced cyclin D1 and COX-2, we pretreated the cells with NSAIDs, and then exposed them to TNF. TNF induced cyclin D1 expressions, and NSAIDs blocked TNF-induced expression of this gene product in a dose-dependent manner (Figure 5). Similarly, TNF also induced the expression of COX-2 protein, and NSAIDs blocked TNF-induced gene expression in a dose-dependent manner (Figure 6). These results further support the conclusion that NSAIDs block NF-κB activation and NF-κB-dependent inflammatory gene expression.

Figure 5
Figure 5

Effect of NSAIDs on TNF-induced cyclin D1 expression. KBM-5 cells were pretreated with various concentrations of NSAIDs either for 4 or 8 h (phenylbutazone, resveratrol, and diclofenac), and then treated with 1 nM TNF for 24 h. Whole-cell extracts were prepared, resolved on 10% SDS–PAGE, and then electrotransferred to a nitrocellulose membrane. Western blot analysis was performed using anti-cyclin D1 and anti-β-actin antibodies as described in Materials and methods. Results shown are representative of two or three independent experiments

Figure 6
Figure 6

Effect of NSAIDs on TNF-induced COX-2 expression and PGE2 secretion. (a) KBM-5 cells were pretreated with various concentrations of NSAIDs either for 4 or 8 h (phenylbutazone, resveratrol, and diclofenac), and then treated with 1 nM TNF for 24 h. Whole-cell extracts were prepared, resolved on 7.5% SDS–PAGE, and then electrotransferred to a nitrocellulose membrane. Western blot analysis was performed using anti-COX-2 and anti-β-actin antibodies as described in Materials and methods. Results shown are representative of two or three independent experiments. (b) Macrophages were pretreated with various concentrations of NSAID for 4 h, and then treated with 1 nM TNF for 12 h. Cell culture media was collected and analysed for PGE2 levels using with PGE2 ELISA kit. The results expressed are percentage of untreated control (100%). Results shown are representative of two independent experiments. The significance of results when compared with the TNF-treated cells are *P>0.1, **P<0.1, ***P<0.01, ****P<0.001

NSAIDs inhibit TNF-induced prostaglandin E2 (PGE2) synthesis

PGE2 is formed through COX-mediated conversion of arachidonic acid (Lawrence et al., 2002). Whether suppression of COX-2 expression by NSAIDs correlates with suppression of PGE2 synthesis was examined. To determine this, we pretreated mouse macrophage with the indicated concentrations of aspirin or curcumin for 4 h, and then cells were stimulated with 1 nM TNF for 12 h, collected the culture media, and analysed for PGE2 secretion. Figure 6b shows that TNF induced PGE2 secretion and aspirin suppressed it in a dose-dependent manner (left panel). Similar results were observed with curcumin (right panel). These results thus suggest that NSAIDs inhibit both COX-2 protein expression as well as its enzymatic activity.

NSAIDs inhibit the proliferation of tumor cells

As NSAIDs inhibited the expression of cyclin D1, which is a cell cycle regulator, we also examined the effect of NSAIDs on the proliferation of human lung adenocarcinoma H1299 cells and on human myeloid KBM-5cells. KBM-5 cells were incubated for 72 h in the presence of different concentrations of NSAIDs, and then cell viability was assayed. All NSAIDs decreased cell viability in a dose-dependent manner (Figure 7a). The 50% inhibitory concentration required in KBM-5 cells was 6.09, 1.12, 0.65, 0.49, 1.01, 0.19, 0.36, 0.012, 0.016, 0.047, 0.013, and 0.008 mM for aspirin, ibuprofen, sulindac, phenylbutazone, naproxen, indomethacin, diclofenac, resveratrol, curcumin, dexamethasone, celecoxib, and tamoxifen, respectively (Table 1). Based on the IC50, aspirin was least potent and tamoxifen was the most potent (761-fold) antiproliferative agent. H1299 cells were also incubated for the indicated days in the presence of different concentrations of NSAIDs, and then cell viability was assayed by the 3-(4, 5-dihydro-6-(4-(3, 4-dimethoxy benzoyl)-1-piperazinyl)-2(1H)-quinolinone (MTT) uptake method. All NSAIDs also decreased H1299 cell viability in a time- and dose-dependent manner (Figure 7b).

Figure 7
Figure 7

Effect of NSAIDs on cell viability, cell proliferation, and apoptosis. (a) In all, 5000 cells of KBM-5 were treated with various concentrations of NSAIDs for 72 h. Then, cell viability was determined by MTT incorporation as described in Materials and methods. (b) A total of 5000 H1299 cells were treated with various concentrations of NSAIDs for the indicated days. Then, cell viability was determined by MTT incorporation. (c) Effect of NSAIDs on [3H]thymidine incorporation. In total, 10 000 KBM-5 cells were treated with various concentrations of aspirin or curcumin for 18 h, and then labeled them with [3H]thymidine for further 6 h, and analysed [3H]thymidine uptake as described in Materials and methods. The results expressed are percentage of untreated control (100%). Results shown are representative of two independent experiments. The significance of results when compared with the each untreated control are *P>0.1, **P<0.1, ***P<0.01, ****P<0.001. (d) KBM-5 cells were incubated with various concentrations of aspirin or curcumin for 24 h. Whole-cell extracts were prepared, resolved on 7.5% SDS–PAGE, and then electrotransferred to a nitrocellulose membrane. Western blot analysis was performed using anti-PARP and anti-β-actin antibodies as described in Materials and methods. Results shown are representative of two independent experiments

To determine whether the effect of NSAIDs on cell viability is due to a decrease in cell proliferation or to an increase in apoptosis, we performed [3H]thymidine incorporation assay and PARP-cleavage assay as an indication of caspase activity. KBM-5 cells were treated with the indicated concentrations of aspirin or curcumin for 18 h, labeled them with [3H]thymidine for further 6 h, and analysed [3H]thymidine uptake (Figure 7c). We found that both curcumin and aspirin inhibited thymidine incorporation in a dose-dependent manner.

Whether NSAIDs induce apoptosis was also examined. For this, KBM-5 cells were incubated with the indicated concentrations of aspirin or curcumin for 24 h, prepared whole-cell extract, and performed Western blot analysis using anti-PARP antibody (Figure 7d). These results indicated that NSAIDs induced PARP cleavage in a dose-dependent manner, thus suggesting an induction of apoptosis. Thus, NSAIDs can both inhibit cell proliferation and induce apoptosis.

Discussion

The aim of the current study was to investigate the relative efficacy of classical NSAIDs and those which are COX-2 specific (such as celebrex) in suppressing NF-κB activation and NF-κB-regulated protein expression, and cell proliferation. Our results clearly demonstrate that all the commonly used NSAIDs can suppress TNF-induced DNA binding of NF-κB by inhibiting IKK activation and IκBα degradation. They also inhibited NF-κB-dependent reporter gene expression and NF-κB-regulated protein expression. Both COX-2, which regulates inflammation, and cyclin D1, which regulates proliferation, were downregulated by all NSAIDs. All the NSAIDs inhibited cell proliferation in a dose-dependent manner as well.

That both TNF and NF-κB play a major role in inflammation is well established (Bharti and Aggarwal, 2002; Aggarwal, 2003). Our results indicate that all NSAIDs can block TNF-induced NF-κB activation, in agreement with reports that aspirin (Hass et al., 1992; Kopp and Ghosh, 1994; Yamamoto et al., 1999), dexamethasone (Hass et al., 1992; Kopp and Ghosh, 1994; Yamamoto et al., 1999), sulindac (Hass et al., 1992; Kopp and Ghosh, 1994; Yamamoto et al., 1999), curcumin (Singh and Aggarwal, 1995; Manna et al., 2000), and resveratrol (Singh and Aggarwal, 1995; Manna et al., 2000) block NF-κB activation. Very little is known, however, about the effect of ibuprofen, phenylbutazone, naproxen, indomethacin, diclofenac, celecoxib, and tamoxifen on the NF-κB activity (Kazmi et al., 1995; Scheuren et al., 1998; Chuang et al., 2002; Bryant et al., 2003). Scheuren et al. (1998) showed that R(−)-ibuprofen as well as the S(+)-enantiomer inhibited the activation of NF-κB in response to T-cell stimulation. One recent report showed that low doses of celecoxib inhibit IL-1-induced NF-κB activation and high doses stimulate it (Niederberger et al., 2001), whereas in our system, celecoxib alone did not activate NF-κB. Another recent report showed that aspirin at 5–10 mM activated NF-κB in colorectal cancer HRT 18 cells but not in other colorectal cell lines (Stark et al., 2001), suggesting a cell type specificity. Callejas et al. (2002) also showed that salicylate, aspirin, indomethacin, ibuprofen, and 5,5-dimethyl-3(3-fluorophenyl)-4-(4-methylsulfonyl)phenyl-2(5H)-furanone, a fluorinated derivative of rofecoxib, had no effect on IKK activity, the processing of NF-κB, or the expression of NF-κB-dependent genes, such as NOS-2 in hepatocytes. Tegeder et al. (2001) also showed that indomethacin failed to inhibit NF-κB activation. Our results differ from those reported in these studies. These differences could be linked with cell type used, type of NF-κB inducer employed, NF-κB assay used (DNA binding vs reporter) and its dose, or the dose of NSAID examined.

Our results also indicate that various NSAID suppress NF-κB activation through inhibition of IKK activity, leading to suppression of IκBα degradation. These results are in agreement with reports that aspirin, sulindac, and sodium salicylate can inhibit the activity of IKK (Yin et al., 1998; Yamamoto et al., 1999). Curcumin and resveratrol have also been shown to inhibit IKK (Singh and Aggarwal, 1995; Manna et al., 2000). Our finding that indomethacin can suppress TNF-induced IKK activity, however, differ from that of Yamamoto et al. (1999), who reported that indomethacin had no effect on IKK. The difference again may be due to differences in cell type or activator.

Our results also indicate that among all the NSAID used, tamoxifen is the most potent inhibitor of NF-κB activation. Based on IC50, it is 567-fold more potent than aspirin. There is no prior report on the suppression of NF-κB activation by tamoxifen. Tamoxifen is a nonsteroidal antiestrogen that is widely used for chemoprevention. Since NF-κB suppression has been closely linked with chemoprevention (Bharti and Aggarwal, 2002), it is possible that chemopreventive effects of tamoxifen are due to NF-κB inhibition as described here. Interestingly, the effects of tamoxifen do not appear to be mediated through the estrogen receptor (ER), as neither KBM-5 nor A293 cells used here are known to express ER. These results are in agreement with reports that tamoxifen can induce apoptosis in ER-negative human cancer cell lines (namely T-leukemic Jurkat and ovarian A2780 cancer cells) (Ferlini et al., 1999).

We found that celecoxib, a specific COX-2 inhibitor, is 236 times more potent than aspirin in inhibiting NF-κB activation. This inhibition of NF-κB was again mediated through suppression of IKK and IκBα degradation.

We found that all NSAID downregulated NF-κB-mediated cyclin D1 and COX-2 protein expression. Aspirin and curcumin also suppressed TNF-induced PGE2 secretion. However, the dose needed to suppress PGE2 secretion was lower than that for COX-2 protein expression. These results are in agreement with a previous report (Fernandez de Arriba et al., 1999; Callejas et al., 2002). These results also point to a more complex regulation of COX-2 expression besides NF-κB (Inoue et al., 1995; Miller et al., 1998).

We found that all the NSAIDs were quite effective in suppressing the proliferation of tumor cells. Both NF-κB and NF-κB-regulated COX-2 and cyclinD1 have been associated with cell proliferation. Therefore, it is possible that downregulation of NF-κB, COX-2, and cyclin D1 mediate the antiproliferative effects of NSAID. We found that NSAIDs (e.g. aspirin and curcumin) can inhibit both DNA synthesis and induce apoptosis as indicated by caspase-induced PARP cleavage.

There are reports that suggest COX-independent actions of COX inhibitors (Tegeder et al., 2001). NSAIDs such as sodium salicylate, sulindac, and ibuprofen can induce anti-inflammatory and antiproliferative effects independent of COX activity. These effects were, however, found to be mediated through inhibition of NF-κB. Cyclin D1 is required for progression of cells from G1 to S phase of the cell cycle and thus is critical for cell proliferation. The downregulation of cyclin D1 by NSAIDs may explain their antiproliferative effects against various tumor cells. Among all the NSAID, once again tamoxifen was most potent (761-fold) in its antiproliferative effects, and this effect must be ER-independent. Celecoxib was almost 468 times more potent than aspirin in suppressing cell proliferation.

When compared with the IC50 of NSAIDs against COX-1 and COX-2 (Warner et al., 1999), the IC50 of NSAIDs for NF-κB inhibition or for antiproliferative effects suggest that the NSAID effects described here are COX-independent. For instance, tamoxifen is a weak inhibitor of COX-1 and COX-2 (see Table 1), yet it is most potent in suppression of NF-κB and cell proliferation. Similarly, celecoxib, which is 100 times more potent than tamoxifen in inhibiting COX-2 activity (Warner et al., 1999), is equally potent for suppression of NF-κB and cell growth (see Table 1). Overall our results suggest that all NSAID can inhibit NF-κB activation and NF-κB-regulated gene expression, and this may contribute to their anti-inflammatory and antiproliferative effects. Since NF-κB has been implicated in chemoresistance and radioresistance (Jung and Dritschilo, 2001; Bharti and Aggarwal, 2002), NF-κB inhibitory activity of NSAIDs can be exploited to overcome chemoresistance and radioresistance.

Materials and methods

Materials

Sulindac, indomethacin, aspirin, naproxin, ibuprofen, tamoxifen, dexamethasone, phenylbutazone, diclofenac, MTT and anti-β-actin antibody were purchased from Sigma Chemical (St Louis, MO, USA). Curcumin, with purity greater than 98%, and celecoxib were purchased from LKT laboratory (St Paul, MN, USA). Resveratrol, with purity greater than 95%, was purchased from Alexis (San Diego, CA, USA). All NSAIDs were dissolved in 100% DMSO at different concentrations. Penicillin, streptomycin, Iscove's modified Dulbecco's medium, minimum essential medium, RPMI 1640, and FBS were obtained from Invitrogen (Grand Island, NY, USA). Antibodies to IκBα, COX-2, and cyclin D1 were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Antibody to IKK-α was a kind gift from Imgenex (San Diego, CA, USA).

Cell lines and culture

We used the human leukemic cell line KBM-5, which is phenotypically myeloid with monocytic differentiation, A293, a human embryonic kidney cell line, and H1299, a human lung cancer cell line. KBM-5 cells were maintained in Iscove's modified Dulbecco's medium supplemented with 15% FBS, A293 cells were maintained in minimum essential medium supplemented with 10% FBS, and H1299 cells were maintained in RPMI 1640 supplemented with 10% FBS. Culture media contained 100 U/ml penicillin and 100 μg/ml streptomycin.

EMSA

NF-κB activation was analysed by EMSA as described previously (Takada et al., 2003). In brief, 15 μg of nuclear extracts prepared from TNF-treated or -untreated cells was incubated with 32P end-labeled 45-mer double-stranded NF-κB oligonucleotide from the human immunodeficiency virus-1 long terminal repeat (5′-TTGTTACAA GGGACTTTCC GCT GGGGACTTTCC AGGGAGGCGTGG-3′; underlined sequence is binding site) for 30 min at 37°C, and the DNA–protein complex resolved in a 6.6% native polyacrylamide gel. For supershift assays, nuclear extracts prepared from TNF-treated cells were incubated with antibodies against either p50 or p65 of NF-κB for 15 min at 37°C before the complex was analysed by EMSA. The specificity of binding was also examined by competition with the unlabeled and mutant oligonucleotide. PIS was included as negative controls. The effect of aspirin and curcumin on the binding of Oct-1 to DNA was determined by incubating 15 μg of nuclear extracts with 16 fmol of 32P end labeled with the octamer-binding protein (Oct-1) consensus oligonucleotide 5′-TGTCGAATGCAAATCACTAGAA-3′ (underline indicates Oct-1 binding site) for 30 min at 37°C, and then analysed using 5% native polyacrylamide gel. The radioactive bands from the dried gels were visualized and quantitated by PhosphorImager (Molecular Dynamics, Sunnyvale, CA, USA) using ImageQuant software.

Western blot analysis

In all, 30–50 μg of cytoplasmic protein or whole-cell extract was prepared as described (Ashikawa et al., 2002) and resolved by SDS–polyacrylamide gel electrophoresis (PAGE). Then, the proteins were electrotransferred to a nitrocellulose membrane, blocked with 5% nonfat dry milk, and probed with primary antibodies against either IκBα, cyclin D1, COX-2, or PARP for 2 h at 4°C. The blotting membrane was washed, exposed to horse radish peroxidase-conjugated secondary antibodies for 1 h, and the blots finally detected by chemiluminescence (ECL, Amersham Pharmacia Biotech. Arlington Heights, IL, USA).

IKK assay

The IKK assay was performed by a method described previously (Takada et al., 2004). Briefly, IKK complex from whole-cell extracts were precipitated with antibody against IKK-α, followed by treatment with protein A/G–Sepharose beads (Pierce, Rockford, IL, USA). After a 2 h incubation, the beads were washed with lysis buffer and then assayed in kinase assay mixture containing 50 mM HEPES (pH 7.4), 20 mM MgCl2, 2 mM DTT, 20 μCi [γ-32P]ATP, 10 μM unlabeled ATP, and 2 μg of substrate GST-IκBα (aa 1–54). After incubation at 30°C for 30 min, the reaction was terminated by boiling with SDS sample buffer for 5 min. Finally, the protein was resolved on 10% SDS–PAGE, the gel was dried, and the radioactive bands were visualized by PhosphorImager. To determine the total amounts of IKK-α and IKK-β in each sample, 50 μg of the whole-cell protein was resolved on 7.5% SDS–PAGE, electrotransferred to a nitrocellulose membrane, and then blotted with either anti-IKK-α or anti-IKK-β antibodies.

NF-κB-dependent reporter gene transcription assay

The effect of NSAIDs on TNF-induced NF-κB-dependent reporter gene transcription was analysed by SEAP assay as previously described (Takada and Aggarwal, 2003a). To examine TNF-induced reporter gene expression, A293 cells (5 × 105 cells/well) were plated in 12-well plates and transiently transfected by the FuGENE 6 with pNF-κB-SEAP (0.5 μg) and β-actin-Renilla (12 ng). After 24 h, cells were washed and then treated with NSAIDs: aspirin, ibuprofen, sulindac, naproxen, indomethacin, curcumin, dexamethasone, celecoxib, and tamoxifen for 4 h or phenylbutazone, resveratrol, and diclofenac for 8 h. The medium was then changed, and exposed them to 1 nM TNF. The cell culture medium was harvested after 24 h and analysed for SEAP activity according to the protocol essentially as described by the manufacturer (Clontech, Palo Alto, CA, USA) using a 96-well plate reader (Vecter 3, Perkin–Elmer Life & Analytical Sciences, Boston, MA, USA) with excitation set at 360 nm and emission at 460 nm. The cells were also lysed and analysed for Renilla activity according to the protocol (Promega, Madison, WI, USA) using a 96-well plate reader (Vecter 3).

PGE2 secretion

The effect of NSAIDs on the PGE2 secretion was studied. One million mouse macrophage were seeded into six-well plate, pretreated with the indicated concentrations of NSAID for 4 h, then stimulated with 1 nM TNF for 12 h and collected culture media. The concentration of PGE2 was determined using PGE2 ELISA kit purchased from R&D Systems (Minneapolis, MN, USA).

Cytotoxicity assay (MTT assay)

The cytotoxic effects of NSAIDs were determined by the MTT uptake method as described (Takada and Aggarwal, 2003b). Briefly, 5000 cells were incubated in duplicate in 96-well plates in the presence of NSAIDs at 37°C. Thereafter, MTT solution was added to each well. After 2 h incubation at 37°C, extraction buffer (20% SDS, 50% dimethylformamide) was added, the cells were incubated overnight at 37°C, and then the optical density was measured at 570 nm using a 96-well multiscanner (Dynex Tech., MRX Revelation; Chantilly, VA, USA).

[3H]thymidine incorporation assay

The effect of NSAIDs on DNA synthesis was monitored by the [3H]thymidine incorporation method (Bharti et al., 2004). In all, 10 000 KBM-5 cells in 100 μl of medium were cultured in triplicate in 96-well plates in the presence or absence of the indicated concentrations of NSAID for 18 h. Then, cells were pulse treated with 0.5 μCi [3H]thymidine for 6 h, and the uptake of [3H]thymidine was monitored using a Matrix-9600 β-counter (Packard Instruments, Downers Grove, IL, USA).

References

  1. . (2003). Nat. Rev. Immunol., 3, 745–756.

  2. , , , , and . (2002). J. Immunol., 169, 6490–6497.

  3. , , , , , , , , , , , , , , , , , , , , , , and . (2003). N. Engl. J. Med., 348, 891–899.

  4. and . (2002). Biochem. Pharmacol., 64, 883–888.

  5. , , and . (2004). J. Biol. Chem., 279, 6065–6076.

  6. . (1999). Drugs Today (Barc.), 35, 225–235.

  7. , and . (2003). Am. J. Vet. Res., 64, 211–215.

  8. , , and . (2002). Hepatology, 35, 341–348.

  9. , , , , , and . (2002). Biochem. Pharmacol., 63, 1709–1716.

  10. and . (2002). Nature, 420, 860–867.

  11. , , , , , , , , and . (1999). Br. J. Cancer, 79, 257–263.

  12. , , , , , , and . (1999). Mol. Pharmacol., 55, 753–760.

  13. and . (2002). Cell, 109 (Suppl), S81–S96.

  14. . (2001). Best Pract. Res. Clin. Gastroenterol., 15, 821–833.

  15. , , , and . (2003). Clin. Cancer Res., 9, 383–390.

  16. , , , and . (1999). Mol. Cell. Biol., 19, 5785–5799.

  17. , , , and . (1992). Biochem. Pharmacol., 44, 1569–1576.

  18. , , , , and . (1999). Mol. Cell. Biol., 19, 2690–2698.

  19. , , , and . (1995). J. Biol. Chem., 270, 24965–24971.

  20. . (1997). Lancet, 350, 437–439.

  21. and . (2001). Semin. Radiat. Oncol., 11, 346–351.

  22. , , , , and . (1995). J. Cell. Biochem., 57, 299–310.

  23. , , , , and . (2002). Carcinogenesis, 23, 1653–1660.

  24. and . (1994). Science, 265, 956–959.

  25. , and . (2002). Nat. Rev. Immunol., 2, 787–795.

  26. , and . (2000). J. Immunol., 164, 6509–6519.

  27. , , , , , and . (1998). J. Cell Biochem., 69, 392–413.

  28. , , , , , , , and . (2002). BMC Cancer, 2, 31.

  29. , , , and . (1999). J. Biol. Chem., 274, 19965–19972.

  30. , , , , , , , and . (2001). FASEB J., 15, 1622–1624.

  31. . (1998). Nature, 396, 15–17.

  32. and . (1994). Blood, 83, 885–898.

  33. , , , , , , , , , , and . (2003). N. Engl. J. Med., 348, 883–890.

  34. , , , and . (1998). Br. J. Pharmacol., 123, 645–652.

  35. and . (1995). J. Biol. Chem., 270, 24995–25000.

  36. , , and . (2001). FASEB J., 15, 1273–1275.

  37. and . (2003a). J. Immunol., 171, 3278–3286.

  38. and . (2003b). J. Biol. Chem., 278, 23390–23397.

  39. , and . (2004). J. Biol. Chem., 279, 26287–26299.

  40. , , , , and . (2003). J. Biol. Chem., 278, 24233–24241.

  41. , and . (2001). FASEB J., 15, 2057–2072.

  42. . (1971). Nat. N. Biol., 231, 232–235.

  43. , , , , and . (1999). Proc. Natl. Acad. Sci. USA, 96, 7563–7568.

  44. , , , , and . (2000). Cancer Res., 60, 6045–6051.

  45. , , and . (1995). J. Biol. Chem., 270, 31315–31320.

  46. , , and . (1999). J. Biol. Chem., 274, 27307–27314.

  47. , and . (1998). Nature, 396, 77–80.

Download references

Acknowledgements

We thank Mr Walter Pagel for carefully proof-reading the manuscript and providing valuable comments. Dr Aggarwal is a Ransom Horne Jr, Distinguished Professor of Cancer Research. This work was supported partially by the Clayton Foundation for Research (to BBA), Department of Defense US Army Breast Cancer Research Program Grant BC010610 (to BBA), a PO1 Grant (CA91844) from the National Institutes of Health on Lung Cancer Chemoprevention (to BBA), and a P50 Head and Neck SPORE Grant from the National Institutes of Health (to BBA).

Author information

Affiliations

  1. Cytokine Research Laboratory, Department of Experimental Therapeutics, The University of Texas M.D. Anderson Cancer Center, Box 143, 1515 Holcombe Boulevard, Houston, TX 77030, USA

    • Yasunari Takada
    • , Anjana Bhardwaj
    • , Pravin Potdar
    •  & Bharat B Aggarwal

Authors

  1. Search for Yasunari Takada in:

  2. Search for Anjana Bhardwaj in:

  3. Search for Pravin Potdar in:

  4. Search for Bharat B Aggarwal in:

Corresponding author

Correspondence to Bharat B Aggarwal.

Glossary

NSAIDs

nonsteroidal anti-inflammatory drugs

TNF

tumor necrosis factor

IκB

inhibitory subunit of NF-κB

IKK

IκBα kinase

EMSA

electrophoretic mobility shift assays

SEAP

secretory alkaline phosphatase

COX

cyclooxygenase

About this article

Publication history

Received

Revised

Accepted

Published

DOI

https://doi.org/10.1038/sj.onc.1208169