Ovatodiolide Targets β-Catenin Signaling in Suppressing Tumorigenesis and Overcoming Drug Resistance in Renal Cell Carcinoma

Dysregulated β-catenin signaling is intricately involved in renal cell carcinoma (RCC) carcinogenesis and progression. Determining potential β-catenin signaling inhibitors would be helpful in ameliorating drug resistance in advanced or metastatic RCC. Screening for β-catenin signaling inhibitors involved in silico inquiry of the PubChem Bioactivity database followed by TCF/LEF reporter assay. The biological effects of ovatodiolide were evaluated in 4 RCC cell lines in vitro and 2 RCC cell lines in a mouse xenograft model. The synergistic effects of ovatodiolide and sorafenib or sunitinib were examined in 2 TKI-resistant RCC cell lines. Ovatodiolide, a pure compound of Anisomeles indica, inhibited β-catenin signaling and reduced RCC cell viability, survival, migration/invasion, and in vitro cell or in vivo mouse tumorigenicity. Cytotoxicity was significantly reduced in a normal kidney epithelial cell line with the treatment. Ovatodiolide reduced phosphorylated β-catenin (S552) that inhibited β-catenin nuclear translocation. Moreover, ovatodiolide decreased β-catenin stability and impaired the association of β-catenin and transcription factor 4. Ovatodiolide combined with sorafenib or sunitinib overcame drug resistance in TKI-resistant RCC cells. Ovatodiolide may be a potent β-catenin signaling inhibitor, with synergistic effects with sorafenib or sunitinib, and therefore, a useful candidate for improving RCC therapy.


Introduction
Renal cell carcinoma (RCC) is the most lethal genitourinary cancer, and the worldwide incidence and mortality rates of RCC have increased annually. In 2008, the incidence was 4/100,000 and the mortality 1.6/100,000 people worldwide. The incidence is 3.2/100,000 and mortality 1.7/100,000 people in Taiwan [1,2]. Most advanced RCC is highly refractory to chemotherapy and radiation therapy and has reduced the 5year survival to 0-20% [3,4].
Six targeted agents for treating advanced or metastatic RCC are now approved and in clinical use. Three are tyrosine kinase inhibitors (TKIs), including sunitinib, pazopanib, and sorafenib. TKIs could improve the overall survival of RCC patients [5,6]. Other agents include an antivascular endothelial growth factor, monoclonal antibody bevacizumab, and 2 Evidence-Based Complementary and Alternative Medicine 2 mammalian targets of rapamycin inhibitors, temsirolimus and everolimus [7][8][9][10][11][12]. However, limited efficacy has been reported for these drugs [5], and more potent compounds that target specific signaling pathways of RCC pathogenesis are needed to improve the high rate of refractory disease.
The -catenin signaling pathway is intricately involved in RCC carcinogenesis and progression. Several -catenin signaling components have been examined in RCC recently, and -catenin signaling may be constitutively active in RCC [13,14]. Aberrant activation of -catenin signaling is involved in RCC carcinogenesis and progression and in the overexpression or overactivation of -catenin [13,14] and oncogenic WNT10A ligand [15] as well as genetic or epigenetic dysregulation of WNT antagonists [16,17]. -Catenin overexpression in RCC was associated with increased incidence and poor prognosis [16,[18][19][20]. The investigation of canonical -catenin signaling and RCC has focused on genetic and epigenetic changes of WNT antagonistic genes [14]. For instance, Dickkopf 2 (DKK2) rs17037102 and DKK3 rs1472189 polymorphisms were found associated with RCC prognosis [21]. The epigenetic silencing of WNT antagonistic genes, including secreted Frizzled-related proteins, DKKs, and WNT inhibitory factor 1, was highly correlated with poor RCC prognosis [14]. Some biologic and small-molecule inhibitors of -catenin signaling have been used to develop novel cancer therapeutic agents [22,23] but scantily for RCC treatment and chemoresistance [24]. To our knowledge, only two pharmaceutical -catenin inhibitors, RX-8243 [25] and BC2059 [26], had been reported to reduce cell proliferation in RCC cell lines.
High-throughput screening to identify modulators of molecular targets has been used with crude extracts or pure compounds isolated from Chinese herbs. The technique can comprehensively delineate relationships among compound structures and biological activities for biological and clinical relevance in specific diseases. Here, we used an easily accessed database for direct cytotoxic screening of compounds that might be effective in RCC. There are ∼25% metastatic RCC patients appeared no clinical benefit of TKIs or responded to TKIs initially but go onto disease progression after a median of 5-11 months [27][28][29][30][31][32]. Discovery of specific inhibitors of other critical signaling pathways would help improve RCC therapy. In this study, we screened for inhibitors of -catenin signaling and evaluated the biological effects of the revealed ovatodiolide in 4 RCC cell lines in vitro and 2 RCC cell lines in a mouse xenograft model. We also examined the synergistic effects of ovatodiolide and sorafenib or sunitinib in 2 TKIresistant RCC cell lines.

Cell Lines.
The RCC lines 786-O, Caki-1, ACHN, and A498, the nontumorigenic human kidney epithelial cell line HK-2, and the human embryonic kidney cell line HEK293T were obtained from the Bioresource Collection and Research Center (Taiwan). 786-O, Caki-1, and ACHN cell lines were maintained in RPMI-1640 and A498, and HEK293T cells were maintained in Dulbecco's Modified Eagle medium (DMEM), all with 10% fetal bovine serum, 1 g/mL penicillin, and 1 g/mL streptomycin (Invitrogen). HK293T cells were grown in keratinocyte serum-free medium supplemented with 50 ng/mL bovine pituitary extract and 5 ng/mL epidermal growth factor. All cells were maintained at 37 ∘ C in a 5% CO 2 atmosphere. To obtain drug-resistant RCC cell lines, 786-O and ACHN were cultured with 5 M sorafenib or 5 M sunitinib malate (Santa Cruz Biotechnology) and fresh complete RPMI-1640 was replaced every 3 days. Cells were cultured for 6 months. The parental controls were performed on 786-O and ACHN cells with similar passage numbers with the only difference being the presence of sorafenib or sunitinib malate in the media.

Compound
Screening. All compounds for screening were offered by Professor Wen-Liang Chang. After disregarding of the redundant compounds, a total number of 21 pure compounds from Citrus reticulata Blanco, 16 from Hibiscus syriacus L., and 23 from Anisomeles indica L. were selected. In brief, the dried and powdered fruit skins of C. reticulata, root barks of H. syriacus, or stems of A. indica (1.0 kg/each) were extracted sequentially with acetone (5 L, 3 times), methanol (5 L, 3 times), 5 L of ethanol (95%, 60%, and 20%), and water (2 L) under reflux for 2 h. The crude extracts were then defatted with n-hexane, partitioned with chloroform and nbutanol, and chromatographed on a silica gel column by eluting with n-hexane/ethyl acetate gradient, with increasing polarity. Ovatodiolide was prepared as described previously and confirmed by high-performance liquid chromatography (HPLC) (column: RP C18e4.6× 250 mm, 5 mm (Merck)). The mobile phase consisted of acetonitrile and 0.1% trifluoroacetic acid (TFA) in water, 64:36 (UV detection at 265 nm) [33]. HPLC revealed the purity of compounds to be ∼95% pure. Representative HPLC chromatogram of ovatodiolide was shown in Figure S1E in supplementary Matrial available online at http://dx.doi.org/10.1155/2013/161628. We evaluated the cytotoxic potential of each compound. First, in silico drug screening involved the use of the PubChem BioActivity database to select each Active outcome in any BioAssay for human tumor cell growth inhibition or antiproliferative activity, in vivo antitumor or anticancer activity, induction of apoptosis, or cytotoxicity (summarized in Table S1). We selected 5 pure compounds for C. reticulata Blanco, 4 for H. syriacus L., and 2 for A. indica L. Second, we used transcription factor/lymphoid enhancer factor (TCF/LEF) reporter assay with these 11 compounds to compare repression of -catenin signal transduction. Psoralen, an abundant pure compound of Psoralea corylifolia L., was used as acatenin signaling control [34]. After 24 hr of transfection with TOPFlash or FOPFlash plasmids, cells were treated with each compound (20, 40 M) for an additional 24 hr and luciferase activities were measured to evaluate the inhibitory effects of compounds on endogenous -catenin signaling (Table  S1). Dimethyl sulfoxide (DMSO) stock solution was kept at −20 ∘ C and freshly diluted to the desired concentrations with cell culture medium immediately before use. The final concentration of DMSO in culture medium was 0.1%. . To detect the activity ofcatenin signal transduction, we used the TCF/LEF reporter  assay with luciferase reporter plasmids (Super 8x TOPFlash  with the wild-type TCF binding sites; Super 8X FOPflash  with the mutant-type TCF binding sites (Addgene clone  M50, M51)). In addition, pGL3-NFAT luciferase (Addgene 17870), CRE-Luc, and NF B reporter plasmids (Qiagen) were used to compare the regulatory effects of ovatodiolide in NF-AT-or cAMP-response-element-(CRE-) regulated promoters. The pGL4.71 renilla luciferase vector (Promega) was cotransfected in a 1/40 molar ratio to normalize transfection efficiency with Lipofectamine 2000 (Invitrogen). After 24 hr of transfection, cells were exposed for 24 hr to DMSO or 20 M ovatodiolide with recombinant human WNT3a (rhWNT3a, 25 ng/mL) or LiCl (20 mM) for TOPFlash, ionomycin (1 g/mL) for NF-AT luc, forskolin (10 M) for CRE luc, and tumor necrosis factor (TNF ; 10 ng/mL) for NF B luc activity controls. Assay of luciferase activity at 48 hr involved use of a Dual-Luciferase reporter assay system (Promega). All experiments were performed in triplicate.

RNA Preparation and Quantitative
Real-Time PCR. RNA was isolated from treated cells by the use of TRIzol (Invitrogen). RNA samples were treated with RQ1 RNase-free DNase (Promega) to remove any genomic contamination. In all, 5 g treated RNA sample was used for reverse transcription with SuperScript III (Invitrogen). Quantitative real-time PCR involved the StepOne Real-Time PCR System (Applied Biosystems) with the GM SYBR qPCR Mix Kit (GeneMark) and GAPDH was used as an internal control. Besides analysis of the melting curve, real-time PCR products were analyzed by gel electrophoresis to confirm single PCR products. Primer sets are as in Table S2 [35].

Flow
Cytometry. Subconfluent cells were trypsinized, washed with 1x PBS, and adjusted to 2 × 10 6 cells/mL. A total of 1 × 10 6 cells were fixed with 100% EtOH for 10 min then incubated with 1 mg/mL propidium iodide for 10 min at room temperature. Cells were analyzed within 20 min by the use of BD FACSCalibur (BD Biosciences).

Cell Migration and Invasion Assay.
For wound healing assay, cells were plated in 6-well plates and cultured to 90% confluence. Each RCC cell was treated with 10, 20, or 40 M ovatodiolide and scraped with a p200 tip (time 0). Before imaging, suspended cells were washed off. The distance of migrating cells was measured from images (5 fields) at 24 and 48 hr after ovatodiolide treatment. The results of wound healing assay were normalized as ratio of wound repaired area to the nontreated control set to 100%. Transwell assay of each RCC cell was assessed by use of 8 m inserts (BD Biosciences). In all, 1 × 10 4 cells were loaded into upper wells, and both upper (200 L) and lower (1 mL) chambers were filled with complete medium (RPMI-1640 or DMEM supplemented with 10% fetal bovine serum (FBS)) containing 20 M ovatodiolide or 0.1% DMSO. For invasion assay, each insert was coated with 1 mg/mL Matrigel at 37 ∘ C for 5 hr. An amount of 1 × 10 4 cells was loaded into a coated insert, and both upper (200 L) and lower (1 mL) chambers were filled with complete medium containing 20 M ovatodiolide or 0.1% DMSO. The migration and invasion chambers were incubated in a humidified 5% CO 2 incubator at 37 ∘ C for 24 hr. Cells were fixed with 500 L methanol/acetone (1 : 1) for 15 min; the inner surface of the upper chambers was wiped with a cotton swab to remove the unmigrated cells for migration assay or Matrigel was scraped off for invasion assay. The chambers were washed with 500 L 1x PBS and stained with 500 L hematoxylin for 1 min at room temperature. After washing with 1x PBS, the transwell membranes were torn off and placed on slides. The stained cells were analyzed by the use of ImageJ software and 5 random fields were counted at 100x magnification. All data represent the mean of triple independent transwell assays.
2.10. Zymography. The enzymatic activities of MMP-2 and MMP-9 were determined by gelatin zymography. In brief, conditioned media were prepared with standard SDS gel-loading buffer containing 0.01% SDS withoutmercaptoethanol or DTT and not boiled before loading. An amount of 50 g conditioned media underwent SDS-PAGE with 0.1% gel. After electrophoresis, gels were washed twice with 1x TBST then incubated with 2% Triton X-100 for 30 min at room temperature to remove SDS, then in 30 mL reaction buffer (40 mM Tris-HCl, pH 8.0, 10 mM CaCl 2 , and 0.02% NaN 3 ) for 24 hr at 37 ∘ C. Before scanning, gels were stained with Coomassie Brilliant Blue R-250 for 30 min and destained with destaining solution (30% methanol, 10% acetic acid, and 60% water).

Detection of -Catenin Nuclear Translocation and Stabil-
ity. An amount of 2 × 10 6 of each RCC cell line was treated with DMSO or ovatodiolide (10, 20, and 40 M) for 24 hr, and cytoplasmic and nuclear extracts were separated with the use of NE-PER Nuclear and Cytoplasmic Extraction Reagents (Thermo Scientific). In all, 20 g nuclear and cytoplasmic extracts were used for SDS-PAGE and Western blot analysis to compare -catenin nuclear and cytoplasmic distribution.
-Catenin degradation was well documented with the 26S proteasome pathway. We compared the ovatodiolide effects on -catenin stability in RCC cells by treating cells with 40 M ovatodiolide and translation inhibitor cyclohexamine (CHX, 100 g/mL) or 26S proteasome inhibitor MG-132 (10 or 20 M) for 48 hr, and 20 g total cell lysates were assessed for SDS-PAGE and western blot analysis.
2.13. Colony-Forming Assay. In vitro tumorigenicity was evaluated by colony-forming assay. In brief, 2 mL of 0.5% agarose in complete RPMI-1640 was used as bottom agar in a 6cm dish, and 2 × 10 4 cells were mixed with 0.3% agarose in complete RPMI-1640 containing 20 M ovatodiolide or 0.1% DMSO. Cells were maintained in a humidified 5% CO 2 incubator at 37 ∘ C for 15 days with fresh medium replaced every 3 days. At the 15th day, cells were stained with crystal violet for 1 min and destained with tap water for 15 min. Colonies in each dish were counted by the use of ImageJ with triplicated repeats for each condition.
2.14. Xenografting. Female Balb/c nude mice (6 weeks old) were purchased from the National Laboratory Animal Center (Taiwan) and acclimated for 1 week. In brief, 1 × 10 7 786-O or ACHN cells in 100 L of 1x PBS with ∼10 mg/mL Matrigel (BD Biosciences) were implanted into the right flank of each mouse. RCC cells at passage 8 were used for xenografting. Before xenografting, cells were tested for mycoplasma by the use of the e-Myco Mycoplasma PCR Detection Kit (Intron). 786-O cells were xenografted into 25 nude mice, and 18 mice showed ∼50 mm 3 tumors after 7 days. ACHN cells were xenografted into 25 nude mice, and 20 mice revealed ∼50 mm 3 tumors after 7 days. We chose 18 mice from each group and randomly divided them into 3 groups (6 mice/group) for systematic treatment: DMSO, 50 mg/kg ovatodiolide, or 100 mg/kg ovatodiolide in 60 L PBS with 0.5% DMSO by intraperitoneal injection daily. Control mice were intraperitoneally injected with 60 L PBS and 0.5% DMSO daily. Tumor size was measured every 2 days with the use of calipers and calculated by (length × width 2 )/2. Tumors were removed at 22 days for 786-O cells and 30 days for ACHN cells, because the body weight of some 786-O xenografted mice was lower than the regulation of The Laboratory Animal Center of National Defense Medical Center and mice should be sacrificed. After measuring tumor weight, a small part of each tumor was flash-frozen in liquid nitrogen for western blot analysis and other parts were fixed with formalin for immunohistochemistry.

Screening for -Catenin Signaling Inhibitory Compounds in RCC Cell Lines.
To identify potential compounds suppressing -catenin signaling activity, we performed a twostep screening of 21 pure compounds of C. reticulata Blanco, 16 compounds of H. syriacus L., and 23 compounds of A. indica L. The first step consists of in silico drug screening involving the PubChem Compound database to search for human tumor cell line growth inhibition/antiproliferative activity, in vivo antitumor/anticancer activity, induction of apoptosis, or cytotoxicity (summarized in Table S1). In all, 11 compounds were selected, including 5 pure compounds of C. reticulata Blanco, 4 compounds of H. syriacus L., and 2 compounds of A. indica L. In the second step, these 11 compounds were used to examine -catenin signaling activity by the use of the TCF/LEF reporter assay with 786-O and Caki-1 RCC cell lines, which feature high endogenous -catenin signaling. Psoralen, a pure compound of Psoralea corylifolia L., was used as a -catenin signaling inductive control [34]. Ovatodiolide ( Figure 1, CAS no.: 3484-37-5, structure obtained from ChemSpider) had the highest inhibitory efficiency for luciferase activity, and the TOP/FOP luciferase ratios were reduced 89% and 92% after 24 hr treatment of 786-O and Caki-1 cells, respectively, with 40 M ovatodiolide ( Figure  S1A). Ovatodiolide, 40 M, cotreated with rhWNT3a or LiCl significantly decreased rhWNT3a or LiCl-induced -catenin signaling, respectively, with TOP/FOP ratios reduced ∼80% in both cell lines, respectively ( Figure S1B).
To confirm the specificity of ovatodiolide in suppressing -catenin signaling, we compared the ovatodiolide effects with NF-AT, CRE, and NF B luciferase reporter assays, with their agonistic compounds used as inductive controls. Ovatodiolide specifically inhibited the luciferase activity of TOP-flash but had no effect in NF-AT, CRE, and NF B reporters (Figure 1(b)). The suppressive effects of ovatodiolide were further evaluated with -catenin/TCF/LEF downstream genes by immunocytochemistry. The staining for nuclearcatenin and its downstream genes cyclin D1 and survivin was less in ovatodiolide-treated RCC cells than in DMSO vehicle controls ( Figure S1C). In 4 RCC cell lines (786-O, Caki-1, A498, and ACHN), ovatodiolide reduced levels of active -catenin (Ser552 phosphorylation on -catenin) and its downstream genes (c-myc, cyclin D1, and survivin) but not other WNT molecules (TCF4, LRP5/6 and its active phosphorylated form, Axin1, and disheveled) (Figures 1(c) and S1D). However, ovatodiolide had no inhibitory effects in HEK293T, a low constitutive WNT signaling cell, or in normal kidney epithelial HK-2 cells ( Figure S1D). Ovatodiolide treatment at 10, 20, and 40 M reduced mRNA levels ofcatenin-signaling target genes Axin2, Sp5, and Nkd1 [35] by 60% to 80% in both RCC cells (Figure 1(d)).

Ovatodiolide Reduces Cell Viability and Induces Apoptosis in RCC Cells.
To evaluate the cytotoxicity of ovatodiolide in RCC and normal kidney cell lines, we analyzed cell viability. Ovatodiolide had a significantly higher cytotoxic effect in 4 RCC cell lines (786-O, Caki-1, ACHN, and A498) but less effect in HK-2 cells (Figures 2(a), S2A, and S2B). The IC 50 with 48 hr treatment for HK-2 cells was 88.20 M, which is much higher than that for RCC cells (16.09, 38.6, 28.40, and 25.81 M for Caki-1, 786-O, ACHN, and A498 cells). With 48 hr treatment, ovatodiolide significantly increased the sub-G1 cell population by ∼5to 6-fold in RCC cells than in controls (Figures 2(b) and S2C). G2/M arrest was increased ∼ 1.5-fold in ovatodiolide-treated cells, perhaps associated with survivin downregulation [36,37]. The apoptosis-inductive effects were also confirmed; cleaved caspase 3 and cleaved PARP level were markedly increased in ovatodiolide-treated cells because of the induction of both intrinsic and extrinsic apoptotic pathways (Figures 2(c) and S2D); cleaved caspase 9 and 8 levels were increased and therefore upregulated apoptotic proteins (Bax, Bid, and PUMA) and downregulated antiapoptotic proteins (Bcl-2, Bcl-xL, and survivin) (Figures  2(c) and S2D). To avoid the ovatodiolide inhibitory effect on -catenin signaling (Figure 1(c)) was a result of highdose (exceeded the IC 50 ) induced cell apoptosis, a sub-IC 50 concentration (15 M) was also examined in Caki-1 and 786-O for 24 h and 48 h. As in Figure S2E, 15 M ovatodiolide also reduced levels of active -catenin (p-Ser552--catenin) and its downstream genes (c-myc, cyclin D1, and survivin) but not other WNT molecules (TCF4), LRP5/6 and its active phosphorylated form, Axin1, and dishevelled.

Ovatodiolide Reduced RCC Aggressiveness by Suppressing
-Catenin Signaling. To examine the inhibitory effects of ovatodiolide on RCC aggressiveness, we evaluated its effects on cell migration, invasion, and tumorigenicity. After 48 hr of 40 M ovatodiolide treatment, migratory ability was reduced >50% in each RCC cell line as compared with controls ( Figure  S3A). Similar inhibitory effects were observed by transwell assay (Figure 3(a) upper). For cell invasion, 24 hr ovatodiolide treatment significantly reduced >65% of invasive cell numbers as compared with controls ( Figure 3(a) lower). Ovatodiolide treatment reduced the protein expression of invasion factors MMP-2 and MMP-9 and therefore reduced their digestive activities ( Figure S3B).
Tumorigenicity of ovatodiolide was evaluated with in vitro colony-formation assay and in vivo xenografting. Treatment with 20 M ovatodiolide for 20 days significantly reduced colony forming ability ∼60 to 80% in cell lines (Figure 3(b)). Balb/c nude mice were subcutaneously injected with 1 × 10 7 786-O or ACHN cells, two higher tumorigenic RCC cell lines. Tumor size reached ∼50 mm 3 after 7 days. Intraperitoneal injection of 50 or 100 g/kg for 22 days in mice with 786-O xenografts and 30 days in mice with ACHN xenografts, hence, systematic treatment, was a prior way for the smallest molecule drug delivery. Ovatodiolide significantly reduced in vivo tumorigenicity of 786-O or ACHN cells, especially with 100 g/kg ovatodiolide (Figures  4(a) and S4A). Treatment with 100 g/kg ovatodiolide significantly reduced both tumor volume and tumor weight compared to controls (Figures 4(b) and S4B). Ovatodiolidetreated mice showed no distinguishable body weight loss or systemic toxicity (Figures S4B and S4D). However, in 786-Oxenografted mice, DMSO significantly reduced body weight after 17 days of 786-O cell injection ( Figure S4C).

Ovatodiolide Reduced -Catenin Stability by Inhibiting AKT Activation and Reducing GSK3 Phosphorylation.
To explore the ovatodiolide inhibition of -catenin signaling, we further investigated its effects on -catenin stability and related regulatory molecules. Ovatodiolide treatment did not modify the mRNA level -catenin in each RCC cell ( Figure S3C). However, -catenin nuclear translocation was dose-dependently decreased after 24 hr ovatodiolide treatment ( Figure S3D). Consequently, RCC cells were cotreated with ovatodiolide, the translation inhibitor CHX, and 26S Evidence-Based Complementary and Alternative Medicine proteosome inhibitor MG-132 to confirm the suppression of -catenin stability. Cotreatment with CHX decreased most of the -catenin protein level, and MG-132 treatment abrogated this inhibitory effect of ovatodiolide (Figure 3(c)).
Ovatodiolide promoted -catenin degradation through the 26S proteosome pathway but not lysosome-associated protein degradation pathway ( Figure S3E). The interaction between E-cadherin, -catenin, TCF4, and -catenin was further compared by coimmunoprecipitation. TCF4--catenin interaction but not E-cadherin-catenin interaction was remarkably reduced in each cell ( Figure S3F). Despite the Caki-1, 786-O, and A498 being socalled E-cadherin-negative cells, the faintly immunoprecipitated E-cadherins were unchanged after ovatodiolide treatment. -Catenins phosphorylated by GSK3 at residues T41, S37, and S33 are recognized by the -TrCP E3 ubiquitin-ligase complex, ubiquitinylated, and ultimately degraded by the 26S proteosome [38]. GSK3 (S9) phosphorylated by active AKT (i.e., AKT S473 phosphorylated form) inhibits GSK3 kinase activity [39]. Otherwise, -catenin phosphorylated at S552 by active AKT enhances -catenin protein levels and nuclear signaling [40,41]. We addressed these possible regulators and -catenin phosphorylation status with ovatodiolide treatment. Ovatodiolide dose-and time-dependently reduced both phosphorylated AKT (S473) and GSK3 (S9) levels ( Figure 3(d)). Therefore, phosphorylated -catenin S552 forms were decreased but phosphorylated S33/37/T41 forms were increased. Treatment with the AKT inhibitor VIII (5 M) induced similar effects, and constitutively active AKT abrogated the ovatodiolide-induced inhibition ofcatenin signaling (Figures S5A and S5B). The effect of constitutively active Akt also partially rescued the OVA-induced cell death ( Figure S6A). Besides, ovatodiolide treatment did not modify other downstream molecules of AKT, including p-Foxo3a (T32), p-mTOR (S2448), and p-p70S6K (T389) levels ( Figure S5A). Therefore, ovatodiolide inhibitedcatenin signaling by reducing -catenin activity and stability. With ovatodiolide treatment of xenografted mice, levels of phosphorylated -catenin, cell cycle markers Ki-67 and cyclin D1, and survival marker survivin were decreased as compared with controls (Figures 4(c) and 4(d)) and levels of phosphorylated AKT (S473) and GSK3 (S9) were decreased (Figure 4(c)). Thus, ovatodiolide reduced -catenin signaling in vivo and reduced RCC cell tumorigenicity. The physical binding between ovatodiolide and -catenin was simulated on the molecular docking website PATCHDOCK with the 3D structure files for ovatodiolide (PubChem: CID 6451060) and -catenin (PDB: 1QZ7). As in Figure S6B, the ovatodiolide inserted into the -catenin molecule enclosing by the AKT phosphorylation site, Ser-552 residue, and may result in a stereochemical change to reduce its activation. However, there is no proper 3D structure including N-terminus ofcatenin and it is uneasy to evaluate whether ovatodiolide also bound to the GSK3 targeting Ser33, Ser37, or Thr41 residues.  Figure 5(a)). Drug-resistant cells showed greater viability ( Figure 5(b)) and increased levels of cyclin D1 and antiapoptotic Bcl-2 but also lower levels of apoptotic proteins (cleaved caspase 3, cleaved PARP, and PUMA) as compared to their parental cells. Combined ovatodiolide and sorafenib or sunitinib treatment significantly increased the cytotoxic effect in both drug-resistant 786-O and ACHN cells as compared with their treatment alone ( Figure 5(b)).

Ovatodiolide Synergistically Increased Sensitivity of RCC
Assessment of the synergistic activity of 20 M ovatodiolide with 5 M sorafenib or sunitinib involved the isobolographic method [42] for drug-resistant 786-O and ACHN cells (Figure 5(c)). To confirm that the synergistic cytotoxicity was caused by specific inhibition of receptor tyrosine kinase and -catenin signaling, we compared the inhibitory effects of these drugs combined on the RAS-RAF-MEK1-ERK1 signaling pathway, a typical target of sorafenib or sunitinib, and the AKT-GSK3 --catenin axis. Ovatodiolide with sorafenib or sunitinib synergistically reduced levels of phosphorylated RAF1, MEK1, and ERK1 in drug-resistant 786-O and ACHN cells ( Figure 5(d)). Combined treatment synergistically reduced phosphorylation of -catenin (S552). STAT3 (Y775) is another target of sorafenib or sunitinib [43]. Phosphorylation of STAT3 (Y775) was also reduced with ovatodiolide combined with sorafenib or sunitinib. In addition, ovatodiolide alone inhibited -catenin signaling without affecting RAS-RAF-MEK-ERK signaling or STAT3 status. Ovatodiolide conferred a synergistic effect that resensitized sorafenib-or sunitinib-resistant cells towards these chemotherapeutic agents.

Discussion
This study demonstrated that ovatodiolide is an anti-catenin signaling compound, at least in RCC, as evidenced by its ability to reduce -catenin stability and suppresscatenin activation in vitro and in vivo. More importantly, when combined with sorafenib or sunitinib, ovatodiolide could enhance the treatment response and overcome drug resistance. We successfully used compound screening with the PubChem BioActivity database combined with TOP/FOP reporter assays to target -catenin signaling in RCC with ovatodiolide. And this procedure should be easily and quickly possessed for screening of specific signaling inhibitors among most purified compounds or synthetic chemicals.
Ovatodiolide is a cemsbrane-type diterpenoid and one of the major components of A. indica L. It can reduce lipopolysaccharide-induced nitric oxide and cytokine levels in macrophages [44] and blood pressure in anaesthetized dogs [45] and is responsible for the anti-inflammatory and antihypotensive effects of A. indica. Ovatodiolide also has cytotoxic effects in some human cancer cell lines by inducing apoptotic pathways [46,47] and has antimetastatic effects by downregulating c-Jun N-terminal kinase, p38 mitogenactivated protein kinase, and PI3K/AKT signaling pathways, therefore inhibiting NF B-MMP-9 axis activation [48].
Combined treatment with ovatodiolide and sorafenib or sunitinib overcame drug resistance in TKI-resistant RCC cells. Sorafenib and sunitinib are approved for treatment of advanced RCC by the US Food and Drug Administration. Both have been reported as multikinase inhibitors of vascular endothelial growth factor receptor, platelet-derived growth factor receptor, RAF, and several different tyrosine kinases and therefore are used in treating several cancers [59]. RCC patients treated with these TKIs show prolonged progressionfree survival and/or overall survival, but resistance to therapy is inevitable [60]. Despite permanent genetic or epigenetic changes in the tumor or selection of drug-resistant clones, resistance of these TKIs is attributed to resumption of angiogenesis, or "angiogenic escape, " according to reversible gene expression in the tumor and/or its microenvironment [61] or accompanied by alternative signaling pathways such as reduction in level of interferon -related angiostatic chemokines [62]. Current strategies to maximize the effectiveness of treatment have mostly focused on sequential, combination, adjuvant, or novel targeted therapy [63].
We also found that RAS-RAF-MEK1-ERK1 signaling is reversibly expressed in sorafenib-or sunitinib-resistant 786-O and ACHN cells. In addition, we found reversed phosphorylated STAT3 status, another target of these TKIs in different cancer types [43]. After combined treatment of ovatodiolide with sorafenib or sunitinib, the reversible gene expression was abrogated and the cytotoxic response was greater than in controls. The -catenin signaling dysregulation involved in the drug resistance of RCC cells [64] and the synergistic effects of -catenin signaling inhibitors and TKIs may be useful to enhance the treatment response of highly chemorefractory, advanced RCC. For evaluation of bioavailability in human, the human equivalent dose (HED) [65] was calculated as follows: HED (mg/kg) = 0.1 mg/kg × 3 (mouse Km)/37 (adult human Km) = 0.008 mg/kg = 0.486 mg/60 kg adult human. And it may be achievable for ovatodiolide treatment in an adult human with a quaque die administration.
In conclusion, ovatodiolide is a potent inhibitor ofcatenin signaling and therefore inhibits cell viability, migration, invasion, and both in vitro and in vivo tumorigenicity