Sulindac Inhibits Activation of the NF-κB Pathway*

Sulindac is a non-steroidal anti-inflammatory agent that is related both structurally and pharmacologically to indomethacin. In addition to its anti-inflammatory properties, sulindac has been demonstrated to have a role in the prevention of colon cancer. Both its growth inhibitory and anti-inflammatory properties are due at least in part to its ability to decrease prostaglandin synthesis by inhibiting the activity of cyclooxygenases. Recently, we demonstrated that both aspirin and sodium salicylate, but not indomethacin, inhibited the activity of an IκB kinase β (IKKβ) that is required to activate the nuclear factor-κB (NF-κB) pathway. In this study, we show that sulindac and its metabolites sulindac sulfide and sulindac sulfone can also inhibit the NF-κB pathway in both colon cancer and other cell lines. Similar to our previous results with aspirin, this inhibition is due to sulindac-mediated decreases in IKKβ kinase activity. Concentrations of sulindac that inhibit IKKβ activity also reduce the proliferation of colon cancer cells. These results suggest that the growth inhibitory and anti-inflammatory properties of sulindac may be regulated in part by inhibition of kinases that regulate the NF-κB pathway.

Sulindac is a non-steroidal anti-inflammatory agent that is related both structurally and pharmacologically to indomethacin. In addition to its anti-inflammatory properties, sulindac has been demonstrated to have a role in the prevention of colon cancer. Both its growth inhibitory and anti-inflammatory properties are due at least in part to its ability to decrease prostaglandin synthesis by inhibiting the activity of cyclooxygenases. Recently, we demonstrated that both aspirin and sodium salicylate, but not indomethacin, inhibited the activity of an IB kinase ␤ (IKK␤) that is required to activate the nuclear factor-B (NF-B) pathway. In this study, we show that sulindac and its metabolites sulindac sulfide and sulindac sulfone can also inhibit the NF-B pathway in both colon cancer and other cell lines. Similar to our previous results with aspirin, this inhibition is due to sulindac-mediated decreases in IKK␤ kinase activity.

Concentrations of sulindac that inhibit IKK␤ activity also reduce the proliferation of colon cancer cells. These results suggest that the growth inhibitory and anti-inflammatory properties of sulindac may be regulated in part by inhibition of kinases that regulate the NF-B pathway.
The NF-B 1 pathway regulates the cellular response to a variety of stimuli including cytokines, bacterial and viral infection, and activation of cellular stress pathways (reviewed in Refs. 1 and 2). The NF-B pathway is also critical for the control of cellular growth properties (3)(4)(5)(6)(7). For example, disruption of the gene encoding the p65 member of the NF-B family leads to severe hepatic apoptosis indicating that at least in some cell types that NF-B is a critical anti-apoptotic factor (5). Inhibition of apoptosis is likely mediated by NF-B induction of cellular genes such as cellular inhibitor of apoptosis-1 and -2 which inhibit the apoptotic process (4,6). The fact that high constitutive levels of NF-B are found in the nucleus of some tumors further implicates activation of this pathway as a potential mechanism to stimulate cellular growth (8,9). NF-B is comprised of a family of related proteins that can either heterodimerize or homodimerize to facilitate their binding to a consensus DNA element to result in activation of gene expression (reviewed in Refs. 1 and 2). The DNA binding and dimerization properties of NF-B are mediated by a conserved domain in these proteins known as the Rel homology domain. NF-B is normally sequestered in the cytoplasm of cells where it is bound by a family of inhibitory proteins known as IB (2,10). These proteins that include IB␣, IB␤, and IB⑀ contain multiple ankyrin repeats that are critical for their inhibitory function. The ability of the IB to mask the nuclear localization signal of NF-B prevents the nuclear translocation of these proteins (11). A variety of stimuli including treatment of cells with TNF␣, interleukin-1, phorbol esters, lipopolysaccharide, and the viral protein Tax results in activation of the NF-B pathway (2). These stimuli modulate signal transduction pathways that lead to the ability of upstream kinases including NIK and MEKK1 to activate the IB kinases IKK␣ and IKK␤ kinases (12)(13)(14)(15)(16). Stimulation of the activity of these kinases results in their ability to phosphorylate two conserved serine residues in the amino terminus of the IB proteins (17)(18)(19)(20)(21)(22)(23). This resultant phosphorylation of IB leads to its ubiquitination and its degradation by the proteasome resulting in the nuclear translocation of NF-B (19,20).
Aspirin and sodium salicylate, but not several other antiinflammatory agents including indomethacin, can inhibit activation of the NF-B pathway (24 -26). Inhibition of the NF-B pathway by aspirin and salicylate is the result of their specific binding to IKK␤ which inhibits its kinase activity (26). The effects of aspirin and salicylate prevent IB degradation and the nuclear translocation of NF-B (26). In addition to its role as an anti-inflammatory agent, aspirin can also help to prevent the development of colon cancer (27)(28)(29)(30)(31).
Sulindac is a non-steroidal anti-inflammatory agent that is structurally related to indomethacin and inhibits cyclooxygenase activity to prevent prostaglandin synthesis (32)(33)(34)(35). In the colon, sulindac is converted by bacteria to the metabolites sulindac sulfide and sulindac sulfone. Sulindac sulfide is the most active metabolite of sulindac and is concentrated in the colonic epithelium at concentrations that are at least 20-fold higher than those seen in the serum which are about 10 -15 M (33). Sulindac sulfide, but not sulindac sulfone, blocks prostaglandin synthesis by non-selective inhibition of cyclooxygenase 1 and cyclooxygenase 2 (33). However, alternative mechanisms of sulindac action other than inhibition of prostaglandin function have been suggested. For example, sulindac sulfone can inhibit mammary carcinogenesis (36), and sulindac sulfide inhibits the proliferation of colon cancer cell lines that do not express cyclooxygenases (37).
Sulindac, like aspirin, has anti-inflammatory properties and has also been demonstrated to induce the regression of adenomatous colon polyps to help prevent the development of colon cancer (27-31, 38 -43). The effects of sulindac on preventing colon cancer are likely mediated by stimulating cellular apo-ptotic pathways (37, 44 -48). Since sulindac and aspirin have similar pharmacologic properties, we investigated whether sulindac-like aspirin could also inhibit kinases that regulate the NF-B pathway to mediate in part its anti-inflammatory and pro-apoptotic properties. In this study, we demonstrate that sulindac and its metabolites, sulindac sulfide and sulindac sulfone, inhibit the activation of the NF-B pathway by inhibiting IKK␤ kinase activity. This result suggests that inhibition of components of the NF-B pathway may at least in part be involved in the anti-inflammatory properties and the growth properties inhibitory of these agents.

MATERIALS AND METHODS
Cell Culture and Treatment-COS, HCT-15, and HT-29 cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum in a humidified 5% CO 2 and 95% air incubator at 37°C. Aspirin, sodium salicylate, and sulindac were obtained from Sigma and dissolved in 1 M Tris-HCl, pH 8.0, to make 1 M stock solution. In transfection assays, aspirin and salicylate were used at a concentration of 5 mM sulindac; sulindac and sulindac sulfone were used at a concentration of 1 mM; sulindac sulfide was used at a concentration of 200 M, and indomethacin and ibuprofen were used at a concentration of 25 M. Indomethacin and ibuprofen were dissolved in ethanol to make a 100 mM stock solution (26). All the anti-inflammatory agents were added into cell culture medium for 16 -18 h for the luciferase reporter gene expression assay or for 2 h prior to TNF␣ treatment and assays of IKK activity (26).
DNA Plasmids and Transient Transfections-The expression plasmids hemagglutinin-tagged IKK␣ and FLAG-tagged IKK␤, and their constitutively active mutants (SS/EE), were described previously (49). The HIV-1 long terminal repeat luciferase reporter contains the human immunodeficiency virus long terminal repeat with two binding sites for NF-B. Approximately 80% confluent COS, cells in 60-mm plates were transfected with the 3 g of DNA using Fugene 6 (Roche Molecular Biochemicals). Cells were harvested 24 -36 h after the transfection.
Assay of IKK Kinase Activity-Histidine and FLAG-tagged baculovirus-produced IKK␣ and IKK␤ proteins were purified by nickel-agarose chromatography and immunoprecipitated with the M2 monoclonal antibody and then assayed in kinase assays as described (26). Transfected COS cells were suspended in lysis buffer containing 40 mM Tris, pH 8.0, 500 mM NaCl, 0.1% Nonidet P-40, 5 mM EDTA, 5 mM EGTA, 10 mM ␤-glycerophosphate, 10 mM NaF, 1 mM sodium vanadate, and protease inhibitor tablet (Roche Molecular Biochemicals). Immunoprecipitation was performed with 300 ng of M2 FLAG antibody or 50 l of 12CA5 supernatant to precipitate the wild-type and mutant hemagglutinin-IKK␣ and wild-type and mutants FLAG-IKK␤ using 200 g of transfected cell lysate. This was followed by the addition of 20 l of protein A-agarose. After washing with the lysis buffer, the immunocomplex was then incubated with a kinase assay buffer containing 50 mM Tris, pH 8.0, 100 mM NaCl, 10 mM MgCl 2 , 1 mM dithiothreitol, 10 M ATP, 10 mM ␤-glycerophosphate, 10 mM NaF, 1 mM Na 3 VO 4 , 5 Ci of [␥-32 P]ATP, and 5 g of glutathione S-transferase-IB␣ (aa 1-54) at 30°C for 30 min. The kinase reaction mixture was resolved on a SDS-polyacrylamide gel electrophoresis and detected by autoradiography (26).
Calculation of the IC 50 Value of Sulindac-COS cells were treated with various concentrations of sulindac in vivo for 2 h before TNF␣treatment and assays of endogenous IKK activity. To assay endogenous IKK activity, cell lysates (200 g of protein) were immunoprecipitated with a rabbit polyclonal antibody directed against IKK␣ (Santa Cruz Biotechnology) that immunoprecipitates the IKK␣/IKK␤ heterodimer followed by assays of kinase activity using the GST-IB substrate (26).
Apoptosis Assays of HCT-15 Cells-Apoptosis assay was measured with Cell Death Detection ELISA PLUS kit (Roche Molecular Biochemicals) using procedures suggested by the manufacturer with modifications. Briefly, HCT-15 cells were cultured in 6-well plates at a density of 2 ϫ 10 6 cells per well overnight. Cells were treated with reagents including aspirin (5 mM), salicylate (5 mM), sulindac (1 mM), ibuprofen (25 M), and indomethacin (25 M) for the indicated periods. After incubation, the cells were washed twice with PBS and harvested from the plates. Cells were lysed 30 min at room temperature, and the lysates were centrifuged at 200 ϫ g for 5 min. Then 20 l of the supernatant was transferred into streptavidin-coated 96-well plates provided by the manufacturer, and 80 l of the immunoreagents containing anti-histone conjugated with biotin and anti-DNA conjugated with peroxidase was added to each well. After incubation for 2 h at room temperature, plates were washed three times with wash buffer. Perox-idase was determined photometrically with 2,2Ј-azinodi[3-ethylbenzthiazoline sulfonate] as substrate using microtiter enzyme-linked immunosorbent assay plate reader at wavelength of 410 nm. The values from duplicate samples were averaged and subtracted from those of background (samples without lysates). The results were expressed as the enrichment factor using the following formula: absorbance of the sample (dying/dead cells)/absorbance of the control cells (without treatment).

Sulindac Inhibits NF-B-directed Gene Expression-Aspirin
and sodium salicylate, but not indomethacin, inhibit NF-Bdirected gene expression (24,25,49). Recently, we extended these studies and demonstrated that the effects of aspirin and sodium salicylate are mediated by the direct binding of these agents to IKK␤ to decrease its kinase activity (26). Since aspirin and sulindac both have anti-inflammatory and anti-proliferative effects, we investigated whether sulindac could also inhibit activation of the NF-B pathway. Such a result could help explain previous data demonstrating that sulindac sulfide can induce apoptosis in the colon cancer cell line HCT-15 which lacks cyclooxygenases (37).
First, we compared the effects of sulindac, aspirin, salicylate, indomethacin, and ibuprofen on the expression of an HIV-1 long terminal repeat reporter construct that contains two NF-B sites. The concentration of sulindac necessary to inhibit the proliferation of colon cancer cell lines ranges from 400 to 1200 M (44,45,47). At these concentrations, sulindac does not cause cell death, and its effects on cell proliferation are reversible after removing the drug (44,45,47,48). Approximately 4 -6-fold lower concentrations of the sulfide metabolite of sulindac (100 -200 M) inhibit cell proliferation (44,45,47). In the analysis described below, we used a concentration of 1,000 M for both sulindac and sulindac sulfone and a concentration of 200 M for sulindac sulfide. These concentrations of sulindac and its metabolites resulted in reversible effects on cellular proliferation (data not shown). The concentrations of aspirin (1-5 mM), indomethacin (25 M), and ibuprofen (25 M) were based on the pharmacologic concentrations of these agents in the serum of patients required for their anti-inflammatory properties (26).
An NF-B reporter was transfected into COS cells that were treated with known activators of the NF-B pathway. These included either TNF␣ (50), the MAP3 kinases NIK (51) or MEKK1 (52,53), or the human T-cell leukemia virus type I transactivator Tax (49). Each of these activators stimulated NF-B-directed gene expression from 10-to 40-fold (Fig. 1, A  and B). These activators did not stimulate the expression of an HIV-1 reporter containing mutated NF-B sites (data not shown). Treatment of the cells with either aspirin or salicylate reduced TNF␣-directed NF-B gene expression approximately 4-fold (Fig. 1A). Treatment of the cells with either sulindac or sulindac sulfone also reduced TNF␣ induction of NF-B gene expression about 4-fold (Fig. 1A). Sulindac sulfide resulted in approximately a 10-fold decrease in NF-B-directed gene expression (Fig. 1A). Similar levels of inhibition were seen with these agents when NF-B gene expression was activated by NIK expression in cells treated with either aspirin, sodium salicylate, sulindac, or the sulfide and sulfone metabolites of sulindac (Fig. 1A). In contrast, two other non-steroidal antiinflammatory agents, indomethacin and ibuprofen, did not reduce TNF␣ stimulation of NF-B-directed gene expression (Fig. 1A).
MEKK1 and Tax activation of the NF-B pathway was also inhibited 4 -6-fold by treatment of cells with either aspirin, sodium salicylate, sulindac, or sulindac sulfone (Fig. 1B). Sulindac sulfide consistently resulted in a 2-3-fold greater level of inhibition of the NF-B pathway than seen with aspirin, salic-ylate, or sulindac sulfone (Fig. 1, A and B). Neither indomethacin nor ibuprofen reduced Tax or MEKK1 induction of NF-Bdirected gene expression (Fig. 1B). These results suggest that sulindac and its sulfone and sulfide metabolites, like aspirin and sodium salicylate, inhibit activation of the NF-B pathway in response to a variety of well characterized inducers of this pathway.
Sulindac Inhibits NF-B Nuclear Translocation-Next it was important to address the mechanism by which sulindac inhibited the NF-B pathway. Nuclear extract was prepared from either untreated COS cells, COS cells treated with TNF␣, or COS cells transfected with NIK. Gel retardation analysis was performed with nuclear extract prepared from these cells using oligonucleotides corresponding to either NF-B-or SP-1binding sites. TNF␣ treatment of cells induced binding to the wild-type NF-B oligonucleotide ( Fig. 2A, lanes 1 and 2) but not to a mutant NF-B oligonucleotide (data not shown). Stimulation of NF-B binding in response to TNF␣ was inhibited by incubation of the cells with either aspirin, salicylate, or sulindac ( Fig. 2A, lanes 3-5) but not ibuprofen or indomethacin ( Fig.  2A, lanes 6 and 7). Similar inhibition of NF-B binding by treatment of cells with either aspirin, salicylate, or sulindac was seen when the NF-B pathway was activated by NIK transfection (Fig. 2A, lanes 9 -14). None of these agents altered SP1 binding in the gel retardation assay ( Fig. 2A, lower panel). These results indicated that sulindac, like aspirin and salicylate, inhibited NF-B nuclear translocation.
Assays were performed to investigate whether sulindac altered the degradation of IB␣ when COS cells were treated with either TNF␣ or following transfection of COS cells with a NIK expression vector. TNF␣ treatment markedly reduced IB␣ levels at 30 min post-treatment when compared with the IB␣ levels seen in untreated COS cells (Fig. 2B, lanes 1 and 2). Treatment of the COS cells with either aspirin, salicylate, or sulindac prevented TNF␣-induced IB␣ degradation (Fig. 2B,  lanes 3-5). In contrast, treatment of cells with either indomethacin or ibuprofen resulted in significant TNF␣-induced IB␣ degradation (Fig. 2B, lanes 6 and 7). Similar inhibition of IB␣ degradation was seen in NIK-transfected cells treated with either aspirin, salicylate, or sulindac (Fig. 2B, lanes 10 -12). These results suggested that decreases in either the kinase activity of IKK or the ubiquitination or proteasome-mediated degradation of IB␣ could be responsible for sulindac-mediated inhibition of the NF-B pathway.
Cytoplasmic extracts prepared from the COS cells used in Fig. 2, A and B, were also assayed for IKK activity. Immunoprecipitation was performed with an IKK␣ antibody that results in the isolation of the IKK␣/IKK␤ heterodimer (49). These immunoprecipitates were then assayed for kinase activity using a GST-IB␣ substrate extending from amino acids 1-54. TNF␣ treatment of cells stimulated IKK activity (Fig. 2C, lanes   1 and 2), and this stimulation was inhibited by treatment of cells with either aspirin, salicylate, or sulindac (Fig. 2C, lanes  3-5). In contrast, indomethacin and ibuprofen did not inhibit IKK activity (Fig. 2C, lanes 6 and 7). Similar results with these agents were seen in extracts prepared from NIK-transfected COS cells (Fig. 2C, lanes 8 -14). A GST-IB␣ mutant at serine residues 32 and 36 was not phosphorylated by IKK (data not shown). Western blot analysis indicated no change in IKK␣ protein levels in these extracts (Fig. 2C, lower panel). These results indicate that sulindac, like aspirin and salicylate, could inhibit IKK activity to prevent IB␣ degradation and subsequent NF-B nuclear translocation.
Sulindac Inhibits IKK Activity in Colon Carcinoma Cell Lines-It was important to analyze the concentrations of sulindac that were necessary to inhibit endogenous IKK activity. Sulindac concentrations ranging from 10 M to 12 mM were incubated with COS cells followed by treatment of these cells with TNF␣ (Fig. 3). Endogenous IKK kinase activity was then analyzed following immunoprecipitation with an IKK␣ antibody that precipitates the IKK␣/IKK␤ heterodimer and a GST-IB␣ substrate. This analysis indicated that the IC 50 of sulin-  1-7) or NIK-transfected (lanes 8 -14) cells from A were collected and used to perform Western blot assays with a polyclonal antibody against IKB␣. C, cytosolic extracts were prepared from TNF␣-treated (lanes 1-7) or NIK-transfected (lanes 8 -15) cells and used to assay for endogenous IKK kinase activity in untreated cells (lanes 1 and 8) or cells treated with the indicated agents (top panel). Cell lysate containing 200 g of protein was immunoprecipitated with a polyclonal antibody directed against IKK␣, and 20 l of protein A-agarose was added to precipitate the endogenous IKK kinase complexes. Kinase assays contained GST-IB␣ (aa 1-54) as a substrate. The endogenous IKK was also analyzed by Western blot analysis using a polyclonal antibody against IKK␣ (lower panel). untr., untreated; PBS, phosphate buffered saline; ASA, aspirin; Sal, sodium salicylate; Sulin, sulindac; Indo, indomethacin; Ibpr, ibuprofen.
dac was approximately 200 M which is in the lower concentration range where sulindac inhibits cellular growth properties (Fig. 3).
Next we assayed whether sulindac was able to prevent TNF␣-mediated increases in IKK activity in two colon cancer lines, HCT-15 and HT-29. The growth of HCT-15 and HT-29 cells is inhibited by treatment with either sulindac or sulindac sulfide (37,44). HCT-15 cells do not produce prostaglandins, and sulindac sulfide inhibition of their proliferation is due to a mechanism independent of inhibiting cyclooxygenases (37). Both sulindac and aspirin, but not indomethacin, markedly reduced TNF␣-mediated increases in IKK activity in HCT-15 (Fig. 4A, lanes 1-5) and HT-29 (Fig. 4A, lanes 6 -10) cells. There was no change in IKK␣ protein levels in these cells when incubated with aspirin or sulindac (Fig. 4A, lower panel). These results indicate that both sulindac and aspirin can alter TNF␣ induction of IKK activity in colon carcinoma cell lines.
The ability of these anti-inflammatory agents to induce apoptosis in HCT-15 cells was next assayed. Since HCT-15 cells do not contain cyclooxygenases, we would expect the mechanism of apoptosis to be prostaglandin-independent and thus not affected by treatment with ibuprofen or indomethacin. In contrast, agents that inhibit the NF-B pathway including sulindac, aspirin, and salicylate might be expected to induce apoptosis in HCT-15 cells. As shown in Fig. 4B, sulindac potently induced apoptosis in approximately 20% of HCT-15 cells after 24 h of incubation. Following 48 h of treatment, aspirin and salicylate also stimulated apoptosis of HCT-15 cells (Fig.  4C). However, neither indomethacin nor ibuprofen induced significant amounts of apoptosis in the HCT-15 cells after either 48 (Fig. 4C) or 72 h (data not shown) of incubation with these agents. These results would be consistent with potential inhibition of the NF-B pathway leading to increased apoptosis in HCT-15 cells that is independent of changes in prostaglandin synthesis.
Sulindac Specifically Inhibits IKK␤ Kinase Activity-To determine whether sulindac inhibited both IKK␣ and IKK␤ kinase activity, we transfected epitope-tagged cDNAs encoding either wild-type or constitutively active forms of these kinases into COS cells. The transfected COS cells were treated with either aspirin, salicylate, sulindac, indomethacin, or ibuprofen (Fig. 5). Following immunoprecipitation of each of these kinases with epitope-specific monoclonal antibodies, they were assayed for their ability to phosphorylate the GST-IB␣ substrate. None of these agents altered the kinase activity of either the wild-type or constitutively active IKK␣ (Fig. 5A, top panel). In contrast, treatment of cells with aspirin, salicylate, or sulindac, but not indomethacin or ibuprofen, resulted in 4 -7-fold inhibition of both the wild-type (Fig. 5B, lanes 1-6) and the constitutively active IKK␤ (Fig. 5B, lanes 7-12). These results indicate that sulindac, like aspirin and salicylate, specifically inhibited IKK␤ kinase activity. There was no effect of these agents on the level of either the transfected IKK␣ or IKK␤ proteins (Fig. 5, A and B, lower panels).
In addition, we assayed the effects of sulindac and its metabolites sulindac sulfide and sulindac sulfone on preventing NIK-mediated increases of both endogenous IKK activity and IKK␤ activity. NIK strongly induced endogenous IKK activity (Fig. 5C, lanes 1 and 2), and this was somewhat inhibited by treatment of cells with aspirin and sulindac (Fig. 5C, lanes 3   FIG. 3. IC 50 for in vivo sulindac treatment. Sulindac concentrations ranging from 10 M to 12 mM were added to COS cells for 2 h prior to treatment of the cells with TNF␣ (20 ng/ml) for 10 min. Immunoprecipitation with IKK␣ antibody was then performed followed by assays of IKK kinase activity using a GST-IB␣ substrate. Kinase activity was quantitated by PhosphorImager scans and plotted against the concentration of sulindac in three separate experiments. and 4) but not indomethacin (Fig. 5C, lane 7). Both sulindac sulfone (Fig. 5C, lane 5) and sulindac sulfide (Fig. 5C, lane 6) potently inhibited IKK kinase activity. The effects of these agents on IKK␤ kinase activity were then assayed following immunoprecipitation of an epitope-tagged IKK␤ protein. NIK strongly induced IKK␤ kinase activity (Fig. 5D, lanes 1 and 2), and its effects were inhibited by 1 mM but not 200 M of sulindac (Fig. 5D, lanes 3 and 4). Sulindac sulfide inhibited IKK␤ activity at 200 but not 40 M (Fig. 5D, lanes 5 and 6). Sulindac sulfone inhibition of IKK␤ activity, like that of sulindac, was seen at 1 mM but not 200 M (Fig. 5D, lanes 7 and 8), whereas indomethacin treatment did not inhibit IKK␤ activity (Fig. 5D, lane 9). There was no effect of sulindac or its metabolites on IKK␤ protein levels (Fig. 5D, lower panel). These results indicate that sulindac and its metabolites inhibit IKK␤ kinase activity.
We next assayed the effects of sulindac, aspirin, salicylate, and indomethacin on the kinase activity of cDNAs encoding the MAP kinases p38, ERK2, or SAPK (54 -56). Thus we could address the ability of sulindac to inhibit a variety of other cellular kinases. Each of these epitope-tagged kinases was transfected into COS cells, and the cells were treated with either anisomycin or 12-O-tetradecanoylphorbol-13-acetate to stimulate the activity of these kinases. Anisomycin treatment increased the kinase activity of both p38 (Fig. 6, lanes 1-6) and SAPK (Fig. 6, lanes 13-18), whereas 12-O-tetradecanoylphorbol-13-acetate treatment increased the activity of ERK2 (Fig. 6,   lanes 7-12). Neither aspirin, salicylate, sulindac, or indomethacin treatment of COS cells altered the activity of these kinases (Fig. 6, top panel) or the levels of these epitope-tagged proteins (Fig. 6, lower panel). These results indicate that the ability of sulindac and aspirin to inhibit IKK␤ was specific in that the activity of a variety of other kinases was not altered by these agents.
Sulindac Inhibits Baculovirus-produced IKK␤-To determine whether sulindac or other anti-inflammatory agents directly inhibited IKK␤ kinase activity, recombinant IKK␣ and IKK␤ proteins produced in baculovirus were assayed in kinase reactions. Neither aspirin, salicylate, sulindac, indomethacin, nor ibuprofen inhibited IKK␣ kinase activity (Fig. 7A lanes  1-6). However, aspirin, salicylate, and sulindac, but not indomethacin or ibuprofen, inhibited IKK␤ kinase activity from 3to 5-fold (Fig. 7B, lanes 1-6). Western blot analysis indicated that equivalent amounts of these kinases were present in each of these reactions (Fig. 7, A and B, lower panels). These results indicate that IKK␤ is a direct target for inhibition of the NF-B pathway by sulindac, aspirin, and salicylate. DISCUSSION Previously, we demonstrated that the anti-inflammatory agents salicylate and aspirin inhibit the NF-B pathway by direct binding of aspirin and salicylate to IKK␤ resulting in their competition for its binding to ATP (26). In contrast, the anti-inflammatory agent indomethacin does not alter NF-B- Western blot analysis is shown in the lower panels using the monoclonal antibody, 12CA5, for IKK␣ and the M2 FLAG monoclonal antibody for IKK␤ to assay the expression of these kinases. directed gene expression. In the current study, we addressed whether the anti-inflammatory agent sulindac, which is structurally related to indomethacin, and its metabolites sulindac sulfide and sulfone can inhibit NF-B-mediated gene expression. Surprisingly, we find that sulindac and particularly its sulfide metabolite are potent inhibitors of the NF-B pathway. The mechanism of action of these agents is due to inhibition of IKK␤ but not IKK␣ kinase activity. These results suggest that several non-steroidal anti-inflammatory agents are potent inhibitors of the NF-B pathway and function by inhibiting IKK␤ kinase activity.
Aspirin and sulindac, in addition to being anti-inflammatory agents, are inhibitors of cellular proliferation (37,44,45,47,48). For example, both sulindac and aspirin have been demonstrated to reduce the incidence of developing colon cancer in patients with adenomatous polyps (27, 29 -31, 39, 40, 42, 43). Although both the anti-inflammatory and growth inhibitory properties of aspirin and sulindac may be due to their inhibition of cyclooxygenases and subsequent reduction in prostaglandin synthesis (45,47,(57)(58)(59), other regulatory pathways may also be targeted by these agents. Since the NF-B pathway is involved in both the pathogenesis of the inflammatory response and in cellular growth control (reviewed in Refs. 1 and 2), this pathway is also a potential target for inhibition by aspirin and sulindac. Our data suggest that these agents inhibit IKK␤ activity to prevent IB␣ degradation and thus prevent NF-B-mediated increases in gene expression. Furthermore, the concentrations of these agents needed to inhibit the NF-B pathway do not result in cellular toxicity. The serum concentration of aspirin in patients with arthritis treated with chronic aspirin therapy is similar to that used in this study (60). Sulindac and its sulfide metabolite have relatively low concentrations in the serum but substantially higher levels in the colonic epithelium where the sulfide metabolite is concentrated at least 20-fold (32,33). The concentrations of sulindac and sulindac sulfide used in this study likely reflect the levels seen in the colonic epithelium.
At the present time, we cannot determine whether the effects of sulindac and aspirin on inducing apoptosis in HCT-15 cells are predominantly due to effects on inhibiting the NF-B pathway or potentially other regulatory pathways. The fact that HCT-15 cells are defective in the generation of prostaglandins (37), yet undergo apoptosis in response to treatment with either sulindac or aspirin, suggests that alternative pathways other than the one mediated by cyclooxygenases likely exist. Both sulindac and sulindac sulfide can reduce the proliferation rate, change the morphology of cells, and cause G 0 /G 1 cell cycle arrest and subsequent apoptosis of colon cancer cell lines (45,47). Sulindac has also recently been demonstrated to cause microsatellite stability in cells isolated with hereditary nonpolyposis colon cancer syndrome (48). The concentrations of sulindac needed to inhibit the growth of colon cancer cell lines are similar to those that inhibit the NF-B pathway. Although sulindac-mediated effects on a variety of other cellular regulatory factors have been demonstrated (45,47), our results are consistent with a role for sulindac inhibition of the NF-B pathway as a potential mechanism that may be involved in inducing apoptosis. The relationships between the different pathways that are inhibited by sulindac treatment will need to be further investigated to understand better the mechanisms of its growth inhibitory properties.