Indomethacin-induced Radiosensitization and Inhibition of Ionizing Radiation-induced NF-κB Activation in HeLa Cells Occur via a Mechanism Involving p38 MAP Kinase¹

Eukaryotic cells invoke adaptive responses to multiple forms of environmental stress, including IR,⁴ by initiating genetically preprogrammed signaling pathways (1, 2). These adaptive responses include the activation of cellular machinery involved in DNA repair, gene induction, cell cycle arrest, apoptosis, and lethality (3, 4). These signaling pathways are involved in the transmission of inter and intracellular information through multiple cellular signal transduction pathways. The physiological role of the transient induction of signaling pathways regulating transcription factors in response to IR remains unclear (5, 6, 7), but it has been suggested that one function is the activation of preprogrammed reparative or protective cellular processes responding to the damaging effects of IR (8, 9, 10). This hypothesis fits well with the growing idea that transcription factors, such as NF-κB, play central roles in the cellular response to stress (11, 12, 13).

NF-κB is a ubiquitous, pleotropic, multisubunit eukaryotic transcription factor activated in response to both inflammatory and noninflammatory exogenous stimuli (14, 15), indicating that NF-κB is involved in multiple cellular processes (16, 17). In most cells, NF-κB exists as an inactive heterodimer, the predominant form of which is composed of p50 and p65 (Rel A) subunits (18, 19) and is sequestered within the cytoplasm by association with an inhibitory protein, I-κB (20, 21). Interaction of I-κB with a NF-κB dimer prevents the nuclear uptake of the NF-κB DNA-binding subunits through the masking of nuclear localization signals (22, 23, 24). Phosphorylation, ubiquitination, and degradation of I-κB by a variety of stimuli result in NF-κB nuclear translocation and the subsequent activation of downstream target genes, including several cell adhesion molecules, inflammatory mediators, and perhaps the cytoprotective response to IR (25, 26, 27). Numerous environmental stresses, including IR, induce NF-κB DNA binding and gene expression by this mechanism (28, 29), a transient response that appears in a wide range of tumor cell lines exposed to IR (11, 28, 29). Furthermore, the induction of NF-κB by IR is often accompanied by an inflammatory response, such as that manifested in the irritation detected on the skin and oral mucosa of head and neck patients receiving therapeutic irradiation (11). Studying the relationship of inflammation and NF-κB activation in response to IR, therefore, seems a logical pursuit in determining the role of transcription factors after exposure to environmental insults (30).

NSAIDs, traditionally, have been used to inhibit cyclooxygenase activity and thereby alleviate clinical cases of pain and inflammation (31). Recently, aspirin, sodium salicylate, and several other NSAIDs have been shown to inhibit the activation of NF-κB, but this effect appears to be cell line and NSAID specific, just as it is dependent on the exogenous agent (e.g., TNF, IL-1, H₂O₂) that is used to induce NF-κB activity (32, 33, 34, 35). In addition, whereas NSAIDs inhibit cyclooxygenase activity at similar relative low doses, the concentrations required to inhibit NF-κB activation are generally of a magnitude greater (32, 33, 34, 35). At these relatively high concentrations, it has been shown that NSAIDs can function as agents that potentiate the cytotoxic effects of IR (36). As such, these results suggest that inhibition of the activation of NF-κB and the coordinating cellular radiosensitization both occur through mechanisms other than the inhibition of cyclooxygenase.

The MAPKs are proline-directed serine/threonine kinases that are important mediators of the reaction of cells to a variety of stimuli by transducing extracellular signals into cellular responses (37). Exposure of mammalian cells to heat shock, strong oxidants, UV irradiation, and other stressful conditions activates a family of these homologous stress-activated protein kinases, including p38 MAPK (38, 39). Recently, it has been shown that sodium salicylate and several other NSAIDs also rapidly activate p38 MAPK (37). Additionally, through the use of the pyridinyl imidazole compound SB 203580, which behaves as a specific p38 inhibitor (40), others have shown that NF-κB inhibition by sodium salicylate or other NSAIDs exposure is abolished through p38 inhibition and the resulting abrogation of NSAID-induced inhibition I-κB phosphorylation and degradation (37, 41). SB 203580 appears to be a specific inhibitor of p38, as it was demonstrated recently that no effect is observed for 12 other common cytoplasmic kinase signaling factors (42). From these studies, it is understood that NSAIDs can alter the activity of upstream cytoplasmic signaling factors regulating the activity of nuclear transcription factors. These experiments also show that p38 MAPK, in addition to its previously established role in the stimulation of signaling factors, can inhibit cytoplasmic signaling factors as well.

With these findings in mind, we investigated the relationship between NSAIDs, p38 MAPK, the NF-κB transcription factor, and cell survival response after IR-based cytotoxic insult to HeLa cells. Specifically, from results published previously and from our preliminary data, we hypothesized that: (a) IR-induced activation of NF-κB can be inhibited by treatment with the NSAID indomethacin; (b) treatment with indomethacin may introduce a radiosensitizing effect in immortalized cell lines at concentrations similar to those required for NF-κB inhibition; (c) pretreatment with SB 203580 will prevent activation of p38 by the NSAID, thereby preventing the inhibition of NF-κB by indomethacin and indomethacin-induced radiosensitization; and (d) permanent cell lines that are genetically unable to activate NF-κB will demonstrate increased sensitivity to IR-induced cytotoxicity. These ideas were experimentally confirmed, and the uses of other NSAIDs, including sodium salicylate and sulindac, offered similar outcomes, as did indomethacin. Although indomethacin-induced radiosensitization and inhibition of IR-induced NF-κB activation are significantly restored by the inhibition of p38 MAPK, the lack of complete restoration presented by the data indicates that indomethacin and other NSAIDs do not use the p38 MAPK pathway exclusively.

HeLa (human cervical carcinoma) cells were grown in MEM (α modification), supplemented with 10% heat-inactivated (56°C, 30 min) calf serum, penicillin (100 units/ml), and streptomycin (100 μg/ml) in a humidified, 5% CO₂ incubator at 37°C. For clonogenic cell survival assays, cells were seeded into 100-mm tissue culture dishes at a density of 2 × 10⁵ and grown to 75% confluence before experimental treatment. To obtain cellular extracts from irradiated samples, cells were seeded at a density of 2.5 × 10⁶ cells/dish, and serum was starved in medium containing 1% calf serum for 48 h before treatment.

The NSAIDs indomethacin [1-(p-chlorobenzoyl)-5-methoxy-2-methylindole-3-acetic acid], sulindac [(z)-5-fluoro-2-methyl-1-(p-[methylsulfinyl] benzylidene) indene-3-acetic acid], and sodium salicylate were obtained from Sigma Chemical Co. (St. Louis, MO), and stock solutions were dissolved in 100% ethanol. Stocks of the specific p38 MAPK inhibitor SB 203580 (Calbiochem, La Jolla, CA) were made in DMSO (Sigma Chemical Co.). Indomethacin (250, 500, 600, or 1000 μm), sulindac (750 or 1000 μm), or sodium salicylate (5, 10, or 20 mm) were added to the growth media and incubated for 1 h at 37°C without or with a 1-h pretreatment of SB 203580 (40 μm). NSAID- and SB 203580-containing media remained in contact with the specified cells throughout the experiment. Corresponding volumes of the ethanol and/or DMSO vehicles were added to designated sham controls. HeLa cells were exposed to IR (2–10 Gy) in a Pantak high frequency 220 kV and 10 mA X-ray generator. The exposure chamber of the X-ray machine supports a 5% CO₂ atmosphere at 37°C, and control cells were placed into a similar environmentally controlled chamber adjacent to the X-ray machine. After irradiation, cells were returned to 37°C for time intervals specified.

The super-repressor form of I-κBα that contains serine-to-alanine substitution mutations at residues 32–36, I-κB SS32/36AA, was a kind gift from Dr. Warner Green, University of California, San Francisco (42). The I-κBα was PCR amplified from the parent CMV-I-κB SS32/36A plasmid and cloned into pTet-On (CLONTECH, Inc.) and sequenced to confirm the integrity of I-κB SS32/36AA gene sequence. pTet-I-κB SS32/36AA was transfected into HeLa Tet-On Cervical epithelioid carcinoma cells expressing the reverse tTA protein (CLONTECH, Inc.), followed by selection with 1 mg/ml G418. The pTet-On system uses the pTet-On regulatory plasmid that expresses the reverse tTA and actives transcription in the presence of tetracycline. pTet-On also expresses a Neomycin-resistance gene contained in the plasmid. All pTet-I-κB SS32/36AA-overexpressing cell lines used in the study are pools of ≥25 independent clones. Control permanently transfected cell lines expressing the parent plasmid pTet-On alone were also isolated. For clonogenic cell survival experiments, cells were treated with tetracycline for 6 h before exposure to IR.

Nuclear and cytoplasmic extracts were obtained from serum-starved cells via a method modified from Dignam (43) and overviewed in Curry et al. (11). Total protein concentrations were determined via a Bradford analysis (Bio-Rad Laboratories, Hercules, CA) on a Beckman (Fullerton, CA) DU-640 spectrophotometer. After preparation and quantification, all samples were stored at −80°C and thawed on ice.

The relative NF-κB DNA-binding activities of treated HeLa cells were determined by the EMSA assay overviewed in Curry et al. (11). Briefly, equal amounts of nuclear protein (15–20 μg) from treated cells were incubated with 100,000 cpm of a double-stranded oligomer containing an NF-κB-specific binding domain (Promega, Madison, WI) that was end-labeled with γ-³²P (NEN Radiochemicals, Boston, MA). After electrophoresis, gels were dried, exposed to a phosphorscreen, and analyzed via a Storm 840 phosphorimager (Molecular Dynamics, Sunnyvale, CA) using ImageQuant 5.1 software. Each experiment was performed in its entirety at least three times, and the results presented are representative.

Cells were assayed for the effect(s) of NSAIDs, SB 203580, and/or IR on cell survival according to established methods of performing the clonogenic assay (44). Briefly, HeLa cells were seeded at densities of 2 × 10⁵ cells/100-mm tissue culture dish and allowed to grow in a 37°C incubator until they reached 75% confluence. After attaining sufficient growth, SB 203580 and/or NSAIDs were added to the growth media as described earlier, after which the plates were exposed to IR at doses of 2–10 Gy. The cell survival curves were normalized to the minor cytotoxicity induced by indomethacin alone. Immediately after IR exposure, cells were trypsinized, diluted, counted, and seeded into 60-mm cloning dishes at densities of 200–20,000 cells/dish. Colonies were allowed to form from surviving cells in a humidified, 5% CO₂, 37°C environment for 7–10 days, after which they were fixed, stained, and counted. Individual assays were performed at multiple dilutions with a total of six plates per data point repeated twice for a total of three identical experiments. The results of these trials are shown in a linear-logarithmic plot relating IR exposure to survival.

The NF-κB reporter plasmid, 4x-κB-tk-Luc, contains 4 NF-κB binding elements upstream of a minimal thymidine kinase promoter placed upstream of the luciferase reporter gene (CLONTECH, Inc.). Full-length constitutively active human MKK6b(E) gene that constitutively activates p38^MAPK has been described elsewhere (37, 44). Transfections were performed in HeLa cells, plated at a density of ∼1 × 10⁶ cells/100-mm plate. HeLa cell were serum starved (1% calf serum) for 8 h and transfected via calcium phosphate precipitation (30). In each transfection, 1 μg of 4x-κB-Luc with either 3 μg of pcDNA3 or 3 μg of pcDNA3-MKK6b(E) was used. As a control, 1 μg of the β-galactosidase expression plasmid (pCMV-β-gal) was used. Transfected cells were exposed to 10 Gy of IR 36 h after transfection without or with pretreatment with indomethacin for 1 h before IR and harvested after 10 h. Luciferase activity was determined using a luminometer (Zylux Corp., Maryville, TN). β-galactosidase activity was determined (Promega), and the relative-fold induction of luciferase activity was calculated by normalizing to the β-galactosidase activity.

The activation of NF-κB after exposure to IR has been established previously in several tumor cell types (11, 28, 29). To additionally investigate this occurrence, we determined the time course of IR-induced NF-κB activation in HeLa cells. HeLa cells were exposed to 10 Gy of IR, nuclear cell extracts were prepared, and NF-κB DNA-binding activities were assessed by EMSA using an oligomer containing a consensus NF-κB site (κB). As demonstrated by others (11, 28, 29) and shown in Fig. 1,A, the DNA-binding activity of NF-κB increased roughly 8-fold at 1–3 h (Fig. 1,A, Lanes 2–4) after IR exposure, and this induction returned to baseline levels at 5 h post-IR (Fig. 1,A, Lane 6), relative to untreated control samples (Fig. 1 A, Lane 1). Super shift assays and EMSA were done to confirm the presence of p50 and p65 binding to the κB DNA-binding element, as well as competition with cold, nonradiolabeled κB probe (data not shown). Sequentially lower doses of IR exposure (2, 5, and 8 Gy) produced less NF-κB-DNA binding in a dose-dependent fashion (data not shown) as shown by others (11, 28).

It has been shown that the activation of NF-κB by several exogenous agents (including TNFα, IL-1, and H₂O₂) can be inhibited by specific NSAIDs (32, 33, 34, 35). As such, the effect of pretreatment with indomethacin on IR-induced activation of NF-κB was determined. As shown in Fig. 1 B, HeLa cells collected at 1 h after treatment with IR only (Lane 2) resulted in an increase in NF-κB activity relative to the sham-treated control cells (Lane 1). When HeLa cells were pretreated with 600 (Lane 4) or 1000 μm (Lane 5) of indomethacin for 1 h before IR exposure, we observed an inhibition of IR-induced NF-κB that was not seen after the 250 μm indomethacin and IR treatment (Lane 3). Treatment of HeLa cells with indomethacin alone at 250, 600, and 1000 μm for this time interval (Lanes 6–8 versus Lane 1) or for exposure times of ≤3 h (data not shown) had no effect of NF-κB DNA binding. The results of these experiments suggest that indomethacin alters a signaling pathway involved in the regulation of IR-induced activation NF-κB DNA binding, and the threshold dose for this process is clearly greater than that required to inhibit cyclooxygenase activity.

Previous results have shown that NSAIDs potentiate the cytotoxic response of in vivo tumors to IR at relatively high concentrations (36). To reaffirm these results in our in vitro model system, we assayed HeLa cells for survival after pretreatment with indomethacin and exposure to IR. Indomethacin was added to HeLa cells and incubated at 37°C for 1 h, after which they were exposed to 2, 4, 6, or 8 Gy of IR (Fig. 2, A and B). Immediately after IR exposure, the cells were trypsinized from the treatment dishes, counted, and plated for clonogenic cell survival as described previously. HeLa cells assayed for clonogenic survival became markedly radiosensitized with 600 and 1000 μm indomethacin pretreatments (Fig. 2,A, ▴ and •, respectively), relative to sham-treated control samples (). Pilot trials using lesser concentrations of 250 (Fig. 2,B, ⋄) and 500 μm (Fig. 2 B, ○) of indomethacin before IR, however, did not result in any significant or repeatable radiosensitization, relative to the control treatment (). The clonogenic cell survival curves shown were normalized to the minor cytotoxicity induced by indomethacin alone (data not shown). The results of these experiments indicate that indomethacin behaves as a radiation sensitizer in HeLa cells at concentrations similar to that required to inhibit IR-induced activation of NF-κB.

Recently, it has been demonstrated that sodium salicylate and NSAIDs can rapidly activate p38 MAPK and inhibit the activation of NF-κB (37), but the mechanism, specific NSAID, and cell line all appear to be somewhat variable. As such, we determined if SB 203580, a chemical inhibitor specific to p38 MAPK (40), would regulate the ability of indomethacin to inhibit IR-induced activation of NF-κB DNA-binding activity. To address this idea, HeLa cells were pretreated with DMSO vehicle or SB 203580 before treatment with indomethacin and/or IR. HeLa cells treated with SB 203580 only (Fig. 3 A, Lane 1), sham-treated (Lane 2), IR only (Lane 3), and SB 203580 before IR (Lane 4) are shown as controls. HeLa cells treated with indomethacin at concentrations of 600 (Lane 5) and 1000 μm (Lane 6) for 1 h before IR demonstrated an inhibition of NF-κB DNA-binding activity similar to that described earlier (Lanes 4 and 5). In contrast, cells treated with 40 μm SB 203580 before and during indomethacin and IR exposures showed a substantial but not complete reversal of indomethacin-induced inhibition of NF-κB DNA-binding activity at 600 μm indomethacin (Lane 7). In addition, the indomethacin-induced inhibitor effect on IR induction of NF-κB is also diminished by SB 203580 at higher NSAID concentrations (1000 μm; Lane 7 versus Lane 8). These results suggest that p38 MAPK is upstream of and involved in at least one signaling cascade involved in indomethacin inhibition of IR-induced activation of NF-κB.

The aforementioned results demonstrate that pretreatment of HeLa cells with indomethacin inhibits IR-induced activation of NF-κB and enhances the cytotoxic effects of IR. These results also show that treatment with SB 203580 reverses the ability of indomethacin to inhibit IR-induced activation of NF-κB. As such, it seems logical to determine whether SB 203580 will also reverse the radiosensitizing effect of indomethacin. To address this idea, clonogenic cell survival experiments were performed. Designated HeLa cells were treated with 40 μm SB 203580 before and during indomethacin exposure and subsequent irradiation at 2–8 Gy. These experiments (Fig. 3,B) demonstrated an increase in IR-induced cytotoxicity with 1000 μm indomethacin (Fig. 3,B, •). In contrast, the addition of SB 203580 before indomethacin treatment and IR exposure significantly, but not completely, reduced the increase in tumor cell killing observed with the NSAID alone (Fig. 3 B, ⋄ versus •). The results presented here, when considered with the previous trials, suggest that p38 MAPK is a target molecule for both indomethacin inhibition of IR-induced activation of NF-κB in response to IR and indomethacin-induced radiosensitization. These results also indicate that the mechanism by which indomethacin behaves as a radiosensitizer is not entirely p38 MAPK dependent but also appears to use other cellular machinery involved in the cytotoxic response to environmental insults.

The induction of NF-κB DNA binding after IR exposure in HeLa cells has been described previously (45, 46, 47) to involve a mechanism resulting in I-κB degradation and the subsequent nuclear localization of NF-κB. Thus, we determined if inhibition of IR-induced activation of NF-κB by indomethacin prevented NF-κB nuclear localization and/or I-κB degradation. To examine this issue in detail, HeLa cells were exposed to 10 Gy of IR without or with indomethacin pretreatment, and subcellular extracts were prepared 1 h after irradiation. Western analysis showed an increase in immunoreactive NF-κB nuclear protein levels in response to IR (Fig. 4 A, Lane 3) when compared with sham-treated (Lane 1) or indomethacin-alone treated controls (Lane 2). Preirradiation treatment with 600 (Lane 5) and 1000 μm (Lane 6) of indomethacin inhibited the IR-induced increase in nuclear NF-κB protein levels (Lane 3). Lesser concentrations of indomethacin (250 μm; Lane 4) had no marked effect on NF-κB expression.

Through a protein-protein interaction, I-κB blocks recognition of the NF-κB nuclear localization sequence, thus preventing transport of NF-κB into the nucleus; it is only when I-κB is degraded that NF-κB is transported into the nucleus (25, 26, 27). With this in mind, Western analyses performed on the cytoplasmic subcellular extracts isolated above were used to determine the immunoreactive I-κB cytoplasmic protein levels after exposure to IR. These results clearly demonstrate that indomethacin inhibits IR-induced degradation of I-κB (Fig. 4,B, Lane 3 versus Lanes 4–6). The results of this experimental series demonstrate a positive correlation of I-κB degradation with the nuclear localization of NF-κB, just as they show that indomethacin inhibits IR-induced degradation of I-κB at similar concentrations observed for radiosensitization (Fig. 2 A).

To additionally investigate the potential role of p38 in the mechanism of the inhibition of NF-κB in response to IR, a transient cotransfection model was established. This system was used to determine: (a) if indomethacin inhibits IR-induced activation of NF-κB-dependent reporter gene expression; and (b) if the genetic activation of p38 will inhibit IR induction of NF-κB in the absence of indomethacin. The constitutive activation of p38 is achieved using a long splice variant of MKK6(E) that has been shown previously to activate endogenous p38 in a dose-dependent manner and decreased TNFα-induced phosphorylation of I-κB (37, 44). As such, HeLa cells were cotransfected with a 4x-κB-Luc reporter plasmid and either pcDNA3 or pcDNA3-MKK6b(E). These results demonstrated that: (a) IR induces NF-κB reporter gene expression, consistent with that observed for NF-κB DNA-binding activity (Fig. 5, Lane 1 versus 2); (b) indomethacin inhibits NF-κB reporter gene expression (Lane 2 versus 3); and (c) cotransfection of pcDNA3-MKK6b(E) significantly reduces IR induction of NF-κB, whereas no effect is seen with cotransfection with pcDNA3 (Lanes 5 versus 4). These results, combined with those above (Fig. 3 A), demonstrate that chemical inhibition of p38 prevents indomethacin inhibition of NF-κB, whereas genetic activation of p38 significantly reduces IR-induced activation of NF-κB in the absence of indomethacin. The results of these experiments suggest that p38 is an intermediate signaling factor in a pathway activated by indomethacin that inhibits IR induction of NF-κB.

The results presented above suggest that NF-κB may have a radioprotective function in how HeLa cells respond to the cytotoxicity of IR. To more rigorously address this question, permanent cell lines were constructed that inducibly express a I-κBα super-repressor gene that contains serine-to-alanine mutations at residues 32 and 36 (I-κB SS32/36AA). This dominant negative gene prevents phosphorylation, ubiquitination, and proteasome-dependent degradation of I-κB after treatment with agents that activate the upstream kinase complex that phosphorylates I-κB (42). In cell lines expressing I-κB SS32/36AA gene, NF-κB remains bound to the I-κB mutant protein preventing NF-κB nuclear localization and activation of gene expression, resulting in a functional NF-κB knockout-like cell line. To address these issues, a permanent HeLa cell line containing an tetracycline-inducible expression plasmid containing I-κB SS32/36AA (pTet-I-κB SS32/36AA) was G418 selected, and roughly 25–50 colonies were pooled.

Initially these cells were examined for their ability to inhibit IR-induced activation of NF-κB DNA-binding activity. When HeLa cells containing pTet-I-κB SS32/36AA were irradiated in the absence of tetracycline, an 8-fold induction of NF-κB DNA binding was observed (Fig. 6,A, Lane 1 versus 2), similar to that seen above (Fig. 1). Permanent cell lines containing the control vector only activated NF-κB DNA binding in both the absence and presence of tetracycline (data not shown). In contrast, HeLa cells treated with tetracycline for 6 h before IR (Lane 3 and 4) failed to activate NF-κB DNA binding (compare Lanes 2 versus 4), confirming that after tetracycline treatment, these permanent cell lines inhibit the activation of NF-κB in response to IR. Finally, these cell lines were used to determine whether NF-κB plays a cytoprotective role in response to IR. When cells expressing pTet-I-κB SS32/36AA were treated with tetracycline before IR, there was increase in IR-induced cytotoxicity (Fig. 5 B) as measured by clonogenic cell survival. There was no change in permanent cell lines containing the control vector pTet-On either in the absence or presence of tetracycline. As such, the results of these experiments appear to identify NF-κB as an IR-inducible factor that protects HeLa Cervical Tumor cells from the cytotoxicity of IR.

Although NSAIDs are groups of pharmacological agents with similar physiological end results, the chemical structures and mechanisms of action of this chemical group can be quite diverse. This observation is illustrated by considerable interpatient variability in the therapeutic benefit of NSAIDs and significant differences in side-effect profiles among similar NSAIDs (36, 40). Among the most commonly reported detrimental side effects resulting from NSAID treatment are adverse reactions in the: (a) gastrointestinal tract; (b) kidney; (c) liver; (d) dermis; (e) central nervous system; and (f) hematological problems (40). The NSAIDs sulindac and sodium salicylate both have wide therapeutic windows and only become toxic at relatively higher concentrations, when compared with indomethacin. As such, these NSAIDs are potentially more useful clinical agents than a drug such as indomethacin, which has a relatively poor therapeutic index, particularly at higher concentrations (36).

From this reasoning, it was determined whether other NSAIDs have similar influences upon NF-κB activity as those effects exhibited by indomethacin. HeLa cells were exposed to 10 Gy of IR, either alone or with preirradiation addition of sulindac, and then assayed for NF-κB DNA binding. Sham-treated cells (Fig. 7,A, Lane 2), cells treated with SB 203580 only (Lane 1), and cells exposed to IR only (Lane 3) are shown as controls. Sulindac alone had no effect on NF-κB DNA binding (data not shown). IR-induced NF-κB DNA-binding activity was inhibited by progressively increasing doses of sulindac (Lanes 3 versus 4–5), results similar to those observed after using indomethacin. The addition of SB 203580 before sulindac and IR treatments produced a reversal of sulindac-induced inhibition of IR-induced activation of NF-κB activity (Lanes 7–8), also results similar to those yielded by the indomethacin trials. Finally, the effect of sodium salicylate on IR-induced activation of NF-κB was determined. These results show that pretreatment with sodium salicylate also inhibits IR induction of NF-κB (Fig. 7 B, Lanes 4–6), and this effect is reversed by pretreatment with SB 203580 (Lanes 7–8). Sodium salicylate treatment alone had no effect on NF-κB DNA binding at the doses used in these experiments (data not shown). When considered cumulatively, this series of results demonstrates that chemical inhibition of p38 MAPK reverses the effects of sulindac- and sodium salicylate-mediated inhibition of IR-induced activation of NF-κB and suggests that these NSAIDs also inhibit the activation of NF-κB by IR via a mechanism involving p38 MAPK.

The NF-κB transcription factor has emerged as a central component of the inducible cellular signaling machinery that is essential for a variety of functions, such as growth, immunity, T-cell activation, and inflammation (15, 16, 48). Another hallmark of NF-κB is its extraordinary capacity to respond to a diverse range of both physiological and pathological forms of environmental stress, including, but not limited to, IR (5, 6, 7, 8). Extensive research has only begun to uncover the mechanism by which IR and similar oxidative stress-inducing stimuli incite the transient activation of NF-κB DNA binding and gene expression in multiple tumor cell types. Specific aspects of the cellular machinery involved, as well as the physiological role for the induction of signaling pathways regulating NF-κB in response to IR exposure, remain unclear.

A possible role lies in the speculation that such early genes as NF-κB which function as transcription factors may play a part in tumor cell survival after IR-induced oxidative stress (11, 49, 50). This hypothesis fits well with the growing idea that other transcription factors, such as c-Fos, c-Jun, Egr-1, and p53, play similar central roles in the cellular response to stress or stress-inducing stimuli (10, 12, 30). Two lines of previous work suggest a role for NF-κB in cell survival and suggest that early gene overexpression predicts for clinical outcome with definitive radiation therapy (11). First, IL-3 and the oncogenic TEL/platelet-derived growth factor receptor fusion protein appear to prevent cell death via activation of NF-κB after cytokine deprivation or exposure to platelet-derived growth factor receptor inhibitors (48). Second, ataxia telangiectasia cells, distinguished by their exquisite sensitivity to IR-induced death, become markedly more resistant to radiation death by insertion of an I-κBα gene that restores IR-induced activation of NF-κB (29). In this regard, these stress-induced gene products may function in coupled short-term changes in the cellular phenotype by modulating the expression of specific target genes involved in cellular defenses to the damaging effects of IR through the activation of preprogrammed reparative or protective cellular processes (1, 3, 4, 11). These findings additionally suggest that the alteration of cytoplasmic signal transducing machinery may represent a mechanism for inhibiting the activation of NF-κB after IR and its subsequent protective effects. Hence, NF-κB and its prospective upstream activators provide ideal paradigms for studying the roles of signaling pathways and the regulation of transcription factors in the cellular transient response to IR-induced stress.

In this study, we have addressed the idea that NF-κB is a downstream target of indomethacin and that NF-κB is involved in the process of radiosensitization and then partially deciphered a mechanism by which these events occur. Using a biochemical cellular fractionation scheme and clonogenic cell survival technique, we have demonstrated that indomethacin inhibits IR-induced activation of NF-κB and sensitizes HeLa cells to IR-induced cytotoxicity at similar concentrations. In addition, pretreatment of HeLa cells with SB 203580, a compound specific for p38 MAPK inhibition, partially abrogated the ability of indomethacin to inhibit IR-induced activation of NF-κB and significantly reversed the indomethacin radiosensitizing effect. Because inhibition of p38 substantially inhibited indomethacin inhibition of IR-induced activation of NF-κB, it seemed logical to determine whether exogenous activation of p38 using MKK6b(E) would also inhibit the activation of NF-κB. Luciferase reporter assays demonstrated that transient cotransfection of MKK6b(E) inhibited IR-induced activation of 4x-κB-tk-Luc, suggesting that p38 is part of a signaling pathway that prevents the induction of NF-κB-dependent gene expression. Taken together, these results suggest that the cytoplasmic signaling protein p38 MAPK is involved in indomethacin-induced radiosensitization and that NF-κB may be one, but perhaps not the only, downstream target in this process.

We have shown previously that indomethacin exposure can also reverse the resistance phenotype of H₂O₂-resistant tumor cells to the cytotoxicity of hyperthermic radiosensitization (51). However, in contrast to HeLa cells, these H₂O₂-resistant tumor cells contain and activate constitutively elevated AP-1 transcriptional complexes that show no resistance to IR-induced cytotoxicity when compared with the parent cells. These results, as well as those presented above, suggest that tumor cell protective responses to hyperthermia and IR-induced cytotoxicity involved distinct and differing intracellular pathways. Because the cellular biological properties of heat shock and IR are distinctly different, these contrasting results seem logical.

Our work implies that indomethacin-induced activation of p38 MAPK resulting in the inhibition of NF-κB in response to the cytotoxicity of IR may provide a unique paradigm to delineate a novel mechanism for possible mechanisms of tumor cell radiosensitization (Fig. 8). In addition, a more substantial understanding of how signaling factors play a cytoprotective and sensitizing effect in response to cytotoxic agents may provide a model system to explore other chemicals or drugs that may have similar effects on early response genes in the process of radiosensitization.