Combinatorial Effect of Non-Steroidal Anti-inflammatory Drugs and NF-κB Inhibitors in Ovarian Cancer Therapy

Several epidemiological studies have correlated the use of non-steroidal anti-inflammatory drugs (NSAID) with reduced risk of ovarian cancer, the most lethal gynecological cancer, diagnosed usually in late stages of the disease. We have previously established that the pro-apoptotic cytokine melanoma differentiation associated gene-7/Interleukin-24 (mda-7/IL-24) is a crucial mediator of NSAID-induced apoptosis in prostate, breast, renal and stomach cancer cells. In this report we evaluated various structurally different NSAIDs for their efficacies to induce apoptosis and mda-7/IL-24 expression in ovarian cancer cells. While several NSAIDs induced apoptosis, Sulindac Sulfide and Diclofenac most potently induced apoptosis and reduced tumor growth. A combination of these agents results in a synergistic effect. Furthermore, mda-7/IL-24 induction by NSAIDs is essential for programmed cell death, since inhibition of mda-7/IL-24 by small interfering RNA abrogates apoptosis. mda-7/IL-24 activation leads to upregulation of growth arrest and DNA damage inducible (GADD) 45 α and γ and JNK activation. The NF-κB family of transcription factors has been implicated in ovarian cancer development. We previously established NF-κB/IκB signaling as an essential step for cell survival in cancer cells and hypothesized that targeting NF-κB could potentiate NSAID-mediated apoptosis induction in ovarian cancer cells. Indeed, combining NSAID treatment with NF-κB inhibitors led to enhanced apoptosis induction. Our results indicate that inhibition of NF-κB in combination with activation of mda-7/IL-24 expression may lead to a new combinatorial therapy for ovarian cancer.


Introduction
Ovarian cancer represents the most lethal gynecological cancer and the 5 th leading cause of women death related to cancer in the United States [1]. Late diagnosis is one of the main hurdles to treat ovarian cancer, as nearly 70% of women present with an advanced stage of the disease at diagnosis [2]. Several epidemiological studies have suggested that the use of NSAIDs at clinically relevant concentrations reduces colorectal [3], breast [4] and ovarian cancer risks [5,6,7], although controversy still remains [8,9]. One major target of NSAID action is inhibition of cyclooxygenase (COX), which are responsible for the conversion of arachidonic acids into prostaglandins and employ a variety of different mechanisms. The two COX genes, COX-1 and COX-2, are almost identical; however, one relevant difference is that COX-1 expression is constitutive, whereas COX-2 expression is induced by growth factors and pro-inflammatory stimuli [10].
NSAIDs are typically classified as specific COX-2 inhibitors or non-specific COX inhibitors. In ovarian cancer COX-1, but not COX-2 has been found to be overexpressed [11,12], nevertheless other studies reported that COX-2 is also upregulated [13,14]. High COX-1 expression in ovarian cancer strongly correlates with high levels of Vascular Endothelial Growth Factor (VEGF) [15,16,17] and NSAIDs inhibit VEGF production in ovarian cancer cell lines [12,18] indicating that COX-1 may regulate VEGF expression. Angiogenesis and VEGF expression are implicated in ascites formation [19] and metastasis of ovarian cancer [20], while its inhibition prevents ascites formation and inhibits disseminated cancer growth [21].
Our previous study demonstrates that NSAIDs also induce apoptosis of cancer cells via induction of mda-7/IL-24 expression [22], leading to enhanced expression of two members of the Growth Arrest and DNA-Damage 45 (GADD45) family [23]. The GADD45 gene family encodes three structurally highly related growth arrest-and DNA damage-inducible proteins, GADD45 a, b and c play a role in the G2/M checkpoint in response to DNA damage [24]. Under normal physiological conditions, mda-7/IL-24 is expressed in cells of the immune system and normal melanocytes [25]. High levels of mda-7/IL-24 have been demonstrated to specifically induce apoptosis of cancer cells and therefore mda-7/IL-24 has been referred as a ''magic bullet'' [26,27]. Moreover, several studies indicated that overexpression of mda-7/IL-24 by a recombinant adenovirus results in cancer cell apoptosis and therapeutic benefits in ovarian cancer [28,29].
NF-kB/IkB signaling is another pathway that has been implicated in drug resistance and survival in ovarian cancer [30,31]. The NF-kB/IkB pathway is emerging as key player in tumorigenesis, invasion and metastasis for various cancers [30,31] and is a critical step for cancer cells to escape programmed cell death [32]. Furthermore, resistance of cancer cells to chemotherapeutic agents has been associated with deregulated NF-kB activation [33]. Moreover, we have shown that inhibition of constitutively active NF-kB by adenoviral expression of IkB induces apoptosis in prostate cancer cells and inhibits tumor formation in SCID mice [34].
In this report, we compared a broad set of NSAIDs for their efficacies to induce apoptosis of ovarian cancer cells. Since our previous study demonstrated a role for mda-7/IL-24 in NSAIDmediated cell death, we also evaluated whether apoptosis induction by NSAIDs is due to induction of mda-7/IL-24 expression. Viral delivery of mda-7/IL-24 is currently used in clinical trials for various cancers [27]. Identification of drugs that are most efficient in mda-7/ IL-24 induction may provide an alternative to viral delivery for exploiting the anti-neoplastic effects of mda-7/IL-24. We also hypothesized that combining NSAIDs with inhibitors of the NF-kB pathway may enhance the effects of NSAIDs against ovarian cancer. We, therefore, tested pharmacological inhibitors of the NF-kB pathway for their abilities to induce apoptosis in ovarian cancer cells.
Here, we demonstrate that Sulindac Sulfide and Diclofenac are the most potent NSAIDs that induce ovarian cancer apoptosis via mda-7/ IL-24 expression and also reduce tumor growth in vivo. mda -7/IL-24 expression leads to GADD45a and c upregulation and JNK kinase activation. Several pharmacological NF-kB inhibitors also induce apoptosis of ovarian cancer cells and in combination with NSAIDs potentiate the apoptotic effect of NSAIDs.

Results
NSAIDs are potent inducers of mda-7/IL-24 and apoptosis in ovarian cancer cells A broad panel of NSAIDs was tested for their abilities to induce apoptosis and mda-7/IL-24 gene expression in four ovarian cancer cell lines, SKOV-3, CAOV-3, SW626 and 36M2. The concentrations for all NSAIDs drugs used in this study were selected to be comparable to achievable physiological plasma concentrations [35][36][37][38][39][40][41][42][43][44][45][46][47][48][49]. Apoptosis was measured 24 hours after treatment of SKOV-3, CAOV-3, SW626 and 36M2 ovarian cancer cells with this set of NSAIDs, revealing that a variety of, but not all NSAIDs induced apoptosis (Figure 1a and S1a). Consistently strong inducers of apoptosis in all four-cell lines included Sulindac Sulfide, Diclofenac, Ebselen and Naproxen when compared to the solvent controls. Some NSAIDs resulted in significant induction of apoptosis in a subset of the ovarian cancer cell lines including Sulindac Sulfone, Acetaminophen, Aspirin and Flurbiprofen whereas treatment with NS-398, Ibuprofen, Finasteride, Flufenamic Acid and Meloxicam resulted only in marginal or no apoptosis induction (Figure 1a and S1a).
For each of the four most consistent inducers of apoptosis we determined the lowest dose that still induces programmed cell death of ovarian cancer cells. The concentrations of the selected NSAIDs were tested at 2, 5 and 10 times lower concentrations than the physiologically achievable doses used in the experiments for Figure 1a and S1a. Apoptosis was measured in ovarian cancer cells 24 hours after treatment with different doses of the four NSAIDs. Our results show that Sulindac Sulfide, Diclofenac and Naproxen concentrations even at 5 times lower dose still effectively induce apoptosis, while Ebselen can be reduced only 2-fold ( Figure 1b).
Based on our previous observations in other types of cancer we determined whether apoptosis induction by NSAIDs correlates with mda-7/IL-24 induction. We measured mRNA expression levels of mda-7/IL-24 mRNA in response to different NSAIDs in SKOV-3 cells by real time PCR analysis demonstrating that mda-7/IL-24 (maximum of 12-fold induction) expression is commonly induced by NSAIDs that promote apoptosis in ovarian cancer cells (Figure 1c). These results were also corroborated in two additional ovarian cancer cell lines. In CAOV-3 and SW626 ovarian cancer cell lines mda-7/IL-24 is also strongly induced (maximum of 77fold induction in CAOV-3 cells) ( Figure S1b). NSAIDs that strongly enhanced apoptosis such as Sulindac Sulfide, naproxen, ebselen, and diclofenac ( Figure 1a) significantly induced mda7/IL-24 expression (Figure 1c), whereas NSAIDs that only marginally induced apoptosis (Figure 1a) did not significantly enhance mda-7/ IL-24 expression except ibuprofen which induced mda-7/IL-24 but did not induce apoptosis (Figure 1c).
We and others have shown that overexpression of mda-7/IL-24 following infection with an adenovirus carrying the mda-7/IL-24 gene induces apoptosis and inhibits cell proliferation in cancer cells [50,51]. In order to evaluate whether induction of apoptosis in cancer cells by NSAIDs is dependent on mda-7/IL-24 upregulation, we conducted experiments with a lentivirus encoding a siRNA against mda-7/IL-24 that was previously generated by our group [23]. Infection of SKOV-3 cells with the mda-7/IL-24 siRNA lentivirus reduced apoptosis induced by NSAIDs by 40-70% relative to the control lentivirus (Figure 1d), further supporting the notion that NSAID-mediated apoptosis is at least partially dependent on mda-7/IL-24 induction. The same results were observed in CAOV-3 cells (data not shown).

Synergistic effects of NSAID combinations
Using the lowest dose of each NSAID that affected apoptosis of ovarian cancer cells (Figure 1b), we systematically analyzed apoptosis induction upon combining low doses of NSAIDs. A panel of NSAIDs including Diclofenac, Sulindac Sulfide, Naproxen and Ebselen were tested for their abilities to induce apoptosis alone and in combination. SKOV-3 ovarian cancer cells were treated with 10 mM Sulindac Sulfide, 40 mM Diclofenac, 25 mM Ebselen or 40 mM Naproxen and combinations thereof. Apoptosis was measured 24 hours after treatment revealing that the majority of the combinations of NSAIDs tested induced apoptosis significantly more than either of the NSAIDs alone in ovarian cancer cells (Figure 2a). Certain combinations such as Sulindac Sulfide and Diclofenac, Sulindac Sulfide and Naproxen and Diclofenac and Naproxen were more effective in apoptosis induction than others ( Figure 2a). We validated these results in CAOV-3 and SW626 cell lines ( Figure S2). Isobologram analysis using combinations of Diclofenac and Sulindac Sulfide indicates that combinations of the drugs results in a synergistic effect (Figure 2b).

NSAID treatment reduces ovarian cancer xenograft growth in SCID mice
To determine whether NSAIDs reduce tumor growth in vivo, SKOV-3 ovarian cancer cells were injected subcutaneously in SCID mice. The mice were randomly divided into 3 groups and fed one of three diets through the entire experiment: AIN-93G as the control and the AIN-93G diet supplemented with either 200ppm Sulindac Sulfide or 100ppm Diclofenac. Two months later the animals were examined for tumor formation and tumor weight. All mice developed tumors indicating that this particular dose of NSAIDs did not prevent tumor formation. However, as seen in Figure 3, Sulindac Sulfide and Diclofenac treatment reduced the average tumor volume by 30% and 20%, respectively, when compared to the control diet with a p-value ,0.05, confirming its anti-tumor efficacy.

Induction of GADD45 a and c gene expression and activation of JNK in ovarian cancer by NSAIDs
We have previously reported that induction of apoptosis by NSAIDs is tightly linked to induction of mda-7/IL-24 expression and consequently to GADD45 a and c upregulation in several cancer cell lines [23]. As NSAIDs induce mda-7/IL-24 expression in ovarian cancer cells (Figure 1c and S1c), we evaluated changes in GADD45 a and c gene expression.
To evaluate whether regulation of GADD45 genes is involved in NSAID-mediated apoptosis, expression of GADD45 a and c mRNAs was measured by real time PCR in SKOV-3, CAOV-3, and SW626 cells treated with NSAIDs. The NSAIDs with the strongest pro-apoptotic activity including Sulindac Sulfide, Diclofenac, Naproxen and Ebselen strongly enhanced GADD45 a and c expression, indicating that increased GADD45 a and c expression closely correlates with pro-apoptotic activity of NSAIDs ( Figure S3a and S3b). Lentiviral expression of siRNA against mda-7/IL-24 in ovarian cancer cells demonstrated that knockdown of mda-7/IL-24 reduces diclofenac-induced GADD45 a and c gene expression indicating that GADD45a and c induction is at least partially dependent on mda-7/IL-24 expression ( Figure S3c).
Since others and we had shown that JNK activation plays a role in apoptosis induction in cancer cells and GADD45 a and c interact with the upstream kinase of JNK, MTK1 [52], we evaluated the activation of JNK during NSAID-mediated apoptosis. JNK kinase activity was tested in protein extracts obtained from CAOV-3 and SKOV-3 cells treated with Sulindac Sulfide (50 mM), Diclofenac (200 mM) or DMSO for 24 hours by an in vitro kinase assay. Western blot analysis revealed very little JNK activity in untreated control cells and a strong increase in JNK activity in both cell lines upon treatment with Sulindac Sulfide and Diclofenac (Figure 4a and 4b). Corroborating this evidence a weak inducer of mda-7/IL-24, GADD45 a and c and apoptosis, Flurbiprofen, demonstrated only a marginal induction of JNK activity (data not shown). To elucidate the functional relevance of GADD45 a and c and mda-7/IL-24 for NSAIDmediated JNK induction and apoptosis in ovarian cancer, JNK kinase activity was tested in protein extracts obtained from CAOV-3 cells treated with Sulindac Sulfide and Diclofenac and infected with lentivirus encoding siRNA against GADD45 a and mda-7/IL-24 genes. Western blot analysis revealed JNK kinase activation by Sulindac Sulfide and Diclofenac was markedly dependent on GADD45 a and mda-7/IL-24 induction, since JNK kinase activity in Sulindac Sulfide and Diclofenac treated mda-7/ IL-24-/-cells was abolished when compared to mda-7/IL-24+/+ cells (Figure 4b). In order to further characterize the effect of NSAIDS in inducing apoptosis, the levels of PARP activation were

Combinatorial treatment of pharmacological inhibitors of the NF-kB pathway with NSAIDs induce apoptosis in ovarian cancer cells
We investigated the biological relevance of the NF-kB pathway in ovarian cancer cells and determined the functional consequences of its inhibition. Instead of using adenoviral delivery of the IkB inhibitor we moved towards a more clinically relevant model and used pharmacological inhibitors of the NF-kB pathway. Inhibitors of the NF-kB pathway were tested for their abilities to induce apoptosis in ovarian cancer cells. Apoptosis was measured 24 hours after treatment of SKOV-3, CAOV-3 and SW626 ovarian cancer cells with four different inhibitors of NF-kB, 5 nM 6-Amino-4-(4-phenoxyphenylethylamino) quinazoline [53], 50 mM Isohelenin [54], 50 mM IKK-2 inhibitor SC-514 [55], and 200 mM IKK inhibitor II Wedelolactone (7-Methoxy-5,11,12trihydroxy-coumestan) [56] or DMSO (control). 6-Amino-4-(4phenoxyphenylethylamino) quinazoline was an efficient inducer of apoptosis in all three cell lines, and the IKK inhibitor II Wedelolactone (7-Methoxy-5,11,12-trihydroxy-coumestan) induced apoptosis in two of the three cell lines. Treatment with Isohelenin or IKK-2 inhibitor SC-514 resulted only in marginal or no apoptosis induction (Figure 5a). Additionally, Real-time PCR analysis indicates that Wedelolactone (7-Methoxy-5,11,12-trihydroxy-coumestan) induces strong activation of the GADD45 a and c gene expression (Figure S4b), and promotes JNK phosphorylation ( Figure S4c) and cleavage of PARP ( Figure S4a).
To determine whether the NF-kB inhibitors 6-Amino-4-(4phenoxyphenylethylamino) quinazoline and Wedelolactone (7-Methoxy-5,11,12-trihydroxy-coumestan) enhance the pro-apoptotic activities of NSAIDs we combined the lowest doses of each of the four NSAIDs, Sulindac Sulfide, Diclofenac, Ebselen, and Naproxen with the lowest doses of the two NF-kB inhibitors that still induce apoptosis (Figure 1b and 5b, respectively). The NSAIDs and NF-kB inhibitors were tested for their abilities to induce apoptosis alone and in combination. SKOV-3, CAOV-3, and SW626 ovarian cancer cells were treated with 10 mM Sulindac Sulfide, 40 mM Diclofenac, 25 mM Ebselen, 40 mM Naproxen, 1 nM 6-Amino-4-(4 phenoxyphenylethylamino) quinazoline and 200 mM Wedelolactone. Apoptosis was measured 24 hours after treatment revealing that combinations of NSAIDs with the NF-kB inhibitor 6-Amino-4-(4 phenoxyphenylethylamino) quinazoline significantly enhanced apoptosis in ovarian cancer cells compared to either of the drugs alone (Figure 5c). In contrast, Wedelolactone enhanced the pro-apoptotic effects only of some of the NSAIDs and less efficiently in SKOV-3 cells (Figure 5c). Isobologram analysis indicates that the combination of 6-Amino-4-(4 phenoxyphenylethylamino) quinazoline with Sulindac Sulfide results in a synergistic effect (Figure 5d).

Discussion
NSAIDs have emerged as potential drugs for chemoprevention in cancer but their benefits are still in question. Some traditional NSAIDS such as Sulindac are currently being tested in clinical trials for various cancers. Indeed, preclinical studies provide consistent evidence that NSAIDs can effectively inhibit tumorigenesis in particular through inhibition of cyclooxygenase-2 (COX-2). Importantly, aspirin use has been associated with a decreased risk of distant breast cancer recurrence and breast cancer death [57].
Epidemiologic studies indicate inverse associations between use of nonsteroidal anti-inflammatory drugs (NSAID) and the incidence of ovarian cancer [5,6,7]. Several reports suggest that certain NSAIDs induce apoptosis and cell cycle arrest in human ovarian cancer cells but the exact molecular mechanism by which NSAIDs induce antitumorigenic activity is not clear. Previously, our groups described a novel pathway by which NSAIDs induce apoptosis and growth arrest in cancer cells. We demonstrated that induction of the pro-apoptotic cytokine MDA-7/IL-24 by NSAIDs is crucial for programmed cell death induced by NSAIDs [23]. However, this study did not include ovarian cancer cells.
Several reports using an adenovirus encoding the mda-7/IL-24 gene (Ad-mda-7) show its profound and selective anticancer activity in animal models [26,27,50,58,59,60] including a report of selectively induced cell death of ovarian cancer cells that results in suppression of tumor growth in vivo [58]. However, transient expression, potentially adverse immune reactions (mediated by adenovirus) and problems with systemic delivery restrict the generalized use of adenoviral delivery of mda-7/IL-24, particularly when administered systemically as a non-replicating adenovirus.
In this context, our findings that NSAIDs with anti-cancer activity induce high levels of mda-7/IL-24 in ovarian cancer cells provide a new therapeutic strategy to enhance mda-7/IL-24 levels on a systemic level. Indeed, we have obtained a comprehensive overview of the consequences of a whole panel of NSAIDs on ovarian cancer cell survival by comparing their efficacies to induce apoptosis and mda-7/IL-24 expression. The most potent inducers of mda-7/IL-24 gene expression include Sulindac Sulfide and Diclofenac. Our finding corresponds with previous reports that demonstrated that treatment of human lung tumor xenografts in nude mice with Ad-mda-7 in addition to Sulindac reduced tumor growth more efficiently than Ad-mda-7 [50]. Moreover, these results corroborate our previous findings that apoptosis induction of the pro-apoptotic cytokine mda-7/IL-24 mediates induction of GADD45 a and c expression and JNK activity in other types of cancer [23]. While Sulindac Sulfide and Diclofenac themselves may not be the ideal drugs to induce mda-7/IL-24 and apoptosis in ovarian cancer cells, and particularly Diclofenac elicits many adverse effects in patients that limit its use in cancer patients, it should be feasible to generate modified versions of these drugs that are more potent in their anti-cancer activities and with reduced adverse and off-target effects. Indeed a modified version of Sulindac has recently been reported to be more active against cancer cells without inhibiting COX 1 and 2 [61].
Diclofenac has previously been shown to induce apoptosis in colon and squamous cell carcinoma and to inhibit pancreatic tumor growth [62,63]. However, there are no reports about its use in ovarian cancer. Here, we demonstrate that Diclofenac as well as Sulindac Sulfide induce apoptosis and inhibit tumor growth of ovarian cancer. These compelling data reinforce the notion of the potential benefits of NSAID treatment for ovarian cancer.
We also identified Naproxen and Ebselen as moderate inducers of apoptosis and mda-7/IL-24 expression in ovarian cancer cells. While Naproxen helps to prevent urinary bladder and colon carcinogenesis [64], Ebselen has been shown to reduce cisplatin treatment toxicity in rat ovarian cancer models, enhancing anti-tumor activity and improving mortality, morbidity and outcome [65]. As mentioned before, we have reported that induction of mda-7/IL-24 by structurally different NSAIDs is crucial for apoptosis induction of breast, prostate, renal and stomach cancer cells [23]. However, in this previous study, Naproxen and Ebselen had only marginal effects on apoptosis induction. In this report, we observed different drug activities for Naproxen and Ebselen. Ebselen and Naproxen induced apoptosis and mda-7/IL-24 expression in ovarian cancer cells and also synergized with the more potent NSAIDs, Diclofenac and Sulindac Sulfide, suggesting potential clinical utility in ovarian cancer therapy.
We have previously shown that inhibition of NF-kB in cancer cells increases apoptosis without promoting mda-7/IL-24 production [23]. One of the major transcriptional circuits implicated in inflammation is the NF-kB/IkB pathway [66]. Furthermore, NF- kB has been implicated in cancer cell survival and escape from programmed cell death and is activated by chemotherapeutic agents in cancer cells [30,31,33]. Mutations in different genes of the NF-kB pathway and constitutively active NF-kB are frequently observed in various types of cancer [33]. Indeed, ovarian cancer cells frequently contain activated NF-kB prior to therapy and are, therefore, expected to be resistant to chemotherapy a priori. On the other hand, we have demonstrated that inhibition of activated NF-kB in cancer cells induces apoptosis without the addition of a chemotherapeutic agent indicating the central role of NF-kB in cell survival of many cancer cells. These findings also suggest the possibility of enhancing therapeutic outcomes by combining NF-kB inhibitors with chemotherapy or other drugs such as NSAIDs. At high concentrations, NSAIDs have been shown to inhibit the TNF-mediated activation of NF-kB [67]. However, we have previously shown that at achievable plasma concentrations, NSAIDs Sulindac Sulfide has no effect on the NF-kB signaling pathway [23]. Indeed here we demonstrate that NF-kB inhibitors strongly induce apoptosis in ovarian cancer cells. Importantly, NF-kB inhibitors markedly enhanced the efficacy of NSAIDs to induce apoptosis, corroborating our hypothesis that these combinatorial regimens can be utilized even though animal model studies are still necessary to prove its efficacy in vivo.
In summary, our results strongly support the hypothesis that drug treatment regimens that lead to enhanced mda-7/IL-24 expression in cancer cells and block NF-kB may have significant efficacy against ovarian cancer. These results also provide a rationale to screen for additional small molecules, natural compounds or even chemically modified NSAIDs, which selectively and efficiently induce mda-7/IL-24 expression in order to obtain more potent anti-cancer drugs.

Real-time PCR
Total RNA was harvested using QIAshredder (Qiagen, Valencia, CA) and RNeasy Mini kit (Qiagen). Real-time PCR was performed as described [23]. The full description of the method is described in Methods S1. The sequences of the primers are as follows: mda-7/IL-

Western-blot
Western-blot were performed as described in supplementary information by using antibodies against total SAPK/JNK (New England Biolabs), phospho JNK (Cell Signaling), mda-7/IL-24 (kindly provided by Paul B. Fisher, Virginia Commonwealth University, School of Medicine, Richmond, VA), PARP and GAPDH (Santa Cruz Biotechnology).

Apoptotic assays
Apoptosis was assayed by using the Apoptotic Cell Death Detection ELISA (Roche) according to the manufacturer' protocol. Significant statistical difference of control samples against samples treated with NSAIDs was determined by Student' t-test.

Isobologram
The isobologram was calculated using Calcusyn software (Biosoft). The drugs were applied across a range of concentrations and cell proliferation was evaluated using the MTT assay at each dosage. Data from cell proliferation was expressed as the fraction of cells inhibited by drug treatments compared with untreated cells. Interaction between pairs of drugs was determined using the Calcusyn software, which generated the isolobogram. The isobologram is a graphical representation of the interaction between two drugs and is formed by plotting the individual drug doses required to achieve a single agent effect on their respective x and y axes, a line connecting the two points is drawn and the concentrations of the two drugs used in combination to achieve the same effect are plotted on the isobologram. Combination data points that fall on the line represent an additive interaction, whereas points above or below represent antagonism or synergy, respectively.

Lentivirus constructs
The lentiviruses encoding siRNA against the three GADD45 family members have been previously described [34]. The LVsiRNA GFP construct (control) was kindly donated by Dr. Oded Singer (Salk Institute for Biological Studies, San Diego, CA). The lentivirus encoding siRNA against mda-7/IL-24 gene was cloned using Advantage 2 PCR kit (Clontech, Mountain View, CA), and the virus was generated by using a previously described methodology [34]. The following siRNA oligonucleotide for mda-7/IL-24 was used: 59CTGTCTAGACAAAAACTTTGTTCT-CATCGTGTCATCTCTTGAATGACACGATGAGAACAAA-GGGGGATCTGTGGTCTCATACA-39.

Animal experiment
Severe combined immunodeficient (SCID)-beige mice were purchased from Taconic (Germantown, NY) and housed in a pathogen-free environment. The animals were randomly divided into 3 groups (n = 7 per group). 8-week-old female SCID-beige mice were fed with one of the experimental diets supplemented with 200 ppm of Sulindac Sulfide, 100 ppm of Diclofenac or control solvents for 2 weeks. Immediately before implantation, SKOV-3 cells were trypsinized and resuspended in MEM with 10% fetal bovine serum. Cell viability was determined by trypan blue exclusion and a single cell suspension with .90% viability was used for implantation. SKOV-3 cells (2610 6 cells in 50 ml) were carefully injected subcutaneously as described previously [23,34] and animals continued on experimental diets. Tumor size (volume) was measured every 5 days, starting at day 35 after implantation. The experiment was finished when the average tumor weight in the control animals reached 2-5% of the body weight. The diets were prepared by Research Diets, Inc. (New Brunswick, NJ). Body weight and food intake were measured weekly. All procedures with animals were reviewed and approved by the Institutional Animal Care and Use Committee at the Beth Israel Deaconess Medical Center according to NIH guidelines.