Discovery of a Novel Orally Active Small-Molecule gp130 Inhibitor for the Treatment of Ovarian Cancer Free

This article is featured in Highlights of This Issue, p. 829

Ovarian cancer is the fourth-leading cause of death in women with gynecologic diseases in the United States. About 70% of the ovarian cancer cases are diagnosed at the late and distant stage and therefore are poorly treatable (1). Prolonged use of paclitaxel and carboplatin in the standard treatment often induces drug resistance, causes ovarian cancer relapse, and eventually the death of patients (2). In this context, there is an urgent need for breakthrough drugs with effective therapeutic impact on ovarian cancer.

Gp130 is part of the receptor signaling complexes for at least 8 cytokines [interleukin (IL)-6, IL-11, IL-27, LIF, CNTF, OSM, CT-1, and CLC; ref. (3)]. Ligand binding induces the association of gp130 with a cytokine-specific receptor-α chain, followed by the activation of downstream signaling cascades including JAK/STAT, RAS/RAF/MAPK, and PI3K/AKT pathways (4). It has been shown that phosphorylation of gp130 at Ser782 downregulates cell surface expression of gp130 (5, 6). As a ubiquitously expressed receptor, gp130 is involved in a wide range of important biologic processes including inflammation, autoimmunity, cancer (7), stem cell maintenance, and embryonic development (8). In recent years, increasing attention has been paid to the relationship between gp130 and cancer progression. However, the role of gp130 in ovarian cancer progression remains elusive, and there is a dearth of small-molecule inhibitors selectively targeting gp130.

Previously, we discovered novel quinoxalinhydrazides to have desirable physicochemical and drug-like properties and a broad-spectrum anticancer activity (9–11). Here, we show that the lead compound, SC144, is a first-in-class small-molecule gp130 inhibitor with oral activity in ovarian cancer.

SC144, a quinoxalinhydrazide derivative, was synthesized as described (9). Stattic, a Stat3 inhibitor (12), was purchased from Sigma-Aldrich. Stock solutions of 10 mmol/L SC144 and Stattic were prepared in dimethyl sulfoxide (DMSO) and stored at −20°C. Further dilutions were made fresh in cell culture medium. IL-6, IFN-γ, and platelet-derived growth factor (PDGF) were purchased from PeproTech. LIF was purchased from Santa Cruz Biotechnology. SDF-1α plasmid was a generous gift from Dr. Ghalib A. Alkhatib (Indiana University, Indianapolis, IN) and the protein was expressed and purified as described (13). Anti-phospho-Stat3 (Y705, 9145), anti-Stat3 (9132), anti-gp130 (3732), anti-phospho-Stat1 (Y701, 9171), anti-Stat1 (9172), anti-Cyclin D1 (2978), anti-Bcl-X_L (2764), anti-Bcl2 (2870), anti-phospho-Akt (T308, 2965), anti-phospho-Akt (S473, 4060), anti-Akt (9272), and anti-Ape1 (4128) were purchased from Cell Signaling Technology. Anti-phospho-gp130 (Ser782, 12978-R), anti-actin (1616-R), normal rabbit IgG (3888), and goat anti-rabbit IgG-HRP (2030) were purchased from Santa Cruz Biotechnology. Anti-gp130 (B-R3, 852-060-000) was purchased from Cell Sciences. Mouse IgG2a (010-0141) and goat anti-mouse IgG-HRP (610-1302) were purchased from Rockland. Goat anti-mouse Cy5-linked IgG and goat anti-rabbit Cy5-linked were purchased from GE Life Sciences. gp130 siRNA, Stat3 siRNA, control siRNA, siRNA transfection reagent, and siRNA transfection medium were purchased from Santa Cruz Biotechnology. Trypan Blue 0.4% solution was purchased from Lonza Inc.

OVCAR-8, OVCAR-5, and OVCAR-3 cells (National Cancer Institute, Developmental Therapeutics Program, Bethesda, MD) were maintained in RPMI-1640 supplemented with 10% heat-inactivated FBS (Gemini-Bioproducts). NCI/ADR-RES cells (National Cancer Institute) were maintained in RPMI-1640 supplemented with 10% heat-inactivated FBS and 5 mmol/L l-glutamine. HEY and Caov-3 cell lines were kindly provided as gifts by Dr. Louis Dubeau (University of Southern California, Keck School of Medicine, Los Angeles, CA). HEY cells were maintained in Dulbecco's Modified Eagle Media (DMEM) supplemented with 10% heat-inactivated FBS and 5 mmol/L l-glutamine. Caov-3 were maintained in MEM supplemented with 10% heat-inactivated FBS. Human normal kidney epithelial and human normal endometrial epithelial cells were kindly provided by Dr. Alan Epstein (University of Southern California, Keck School of Medicine) and maintained in RPMI-1640 supplemented with 10% heat-inactivated FBS. Cells were grown as monolayers at 37°C in a humidified atmosphere of 5% CO₂. To remove adherent cells from the flask for subculture and counting, cells were washed with PBS without calcium or magnesium, incubated with a small volume of 0.25% trypsin–EDTA solution (Mediatech, Inc.) for 5 to 10 minutes, resuspended with culture medium, and centrifuged. All experiments were carried out using cells in the exponential growth phase. Cells were routinely checked for mycoplasma contamination by using PlasmoTest (InvivoGen).

Colony formation assays were conducted as described (14). Briefly, OVCAR-8 cells (600 cells/well) and Caov-3 cells (800 cells/well) were seeded in 24-well plates and allowed to attach. After 48 hours, serial dilutions of the corresponding compounds or antibodies were added to the culture medium and incubated for 48 hours. Cells were cultured until colonies were formed (10 days for OVCAR-8 and 15 days for Caov-3), then subsequently washed, stained with crystal violet solution (2%) for 1 hour, and thoroughly washed with water.

Cells (only attached cells were collected) or xenograft tumor tissues were washed with ice-cold PBS and lysed in cold lysis buffer (20 mmol/L Tris-HCl, 150 mmol/L NaCl, 1 mmol/L EDTA, 1% Triton X-100, pH 7.5) with 1× protease and phosphatase inhibitors. Protein concentration was determined by BCA protein assay (Thermo Scientific). Proteins were resolved on 8% or 10% SDS-PAGE and electrotransferred to Immun-Blot PVDF membrane (Bio-Rad). After blocking with 5% milk in Tris-buffered Saline with Tween-80 (TBST), membranes were probed with the indicated primary antibodies, subsequently with horseradish peroxidase–conjugated secondary antibody, and developed using Dura Extended Duration Substrate (Thermo Scientific). Immunoreactive proteins were visualized with the Chemi-Doc System (Bio-Rad).

Growth inhibition was assessed using an MTT assay as described earlier (15). Cells were seeded in 96-well microtiter plates, allowed to attach 24 hours before siRNA transfections or the addition of corresponding compounds to the culture medium. After 72 hours, cells were incubated with 0.3 mg/mL MTT (Amresco) for an additional 3 hours at 37°C. After removal of the supernatant, DMSO was added to the wells and the absorbance was read at 570 nm. All assays were conducted in triplicate. Percentage of cell growth inhibition was expressed as: (1 − A/C) × 100% (A and C were the absorbance values from experimental and control cells, respectively). Inhibitory concentration 50% (IC₅₀) values were determined for each drug from a plot of log (drug concentration) versus percentage of cell growth inhibition. SDs were calculated on the basis of the IC₅₀ values obtained from at least 3 independent experiments.

The Drug Affinity Responsive Target Stability (DARTS) assay was optimized and used to assess the binding of SC144 to gp130 via the protease protection from pronase (Roche Applied Science, Inc.) as previously described (16, 17). OVCAR-8 cells were lysed using M-PER (Thermo Scientific, Inc.) supplemented with protease and phosphatase inhibitors. The supernatant of cell lysate containing 4 to 6 μg/μL total proteins was incubated with SC144 at indicated concentrations at room temperature for 1 hour, followed by proteolysis with 1 μg pronase to every 9,600 μg of lysate for 30 minutes at room temperature. Final concentration of DMSO was 1% in all samples. To stop proteolysis, 5× SDS sample loading buffer [Tris-HCl 0.25 mol/L, pH 6.8, SDS 10%, glycerol 50%, bromophenol blue 0.5%, dithiothreitol (DTT) 100 mmol/L] was added to each sample at a 1:4 ratio, mixed well, and boiled at 100°C for 5 minutes. Samples were analyzed by Western blotting.

OVCAR-8 and Caov-3 cells (2 × 10⁵) were seeded in 35-mm dishes, allowed to attach overnight, and then received indicated treatments. Cells were washed with cold PBS. Both floating and attached cells were collected, stained with the Annexin V–FITC Apoptosis Detection Kit (BioVision) according to the manufacturer's protocol. The resulting fluorescence was measured by an LSR II flow cytometer (Becton Dickinson).

OVCAR-8 cells were grown on poly-l-lysine (Sigma-Aldrich)–coated glass coverslips and subjected to serum-free RPM1-1640 overnight. With or without compound pretreatments, cells were stimulated with 50 ng/mL IL-6 for 20 minutes (for Stat1) or 10 minutes (for Stat3), 50 ng/mL LIF for 15 minutes, or 50 ng/mL IFN-γ for 20 minutes. Cells were then fixed with 3.7% formaldehyde in PBS for 15 minutes at room temperature and permeabilized for 10 minutes with ice-cold 100% methanol at −20°C. After blocking with 5% goat serum, slides were incubated with anti-Stat1 (1:200) or anti-Stat3 (1:1,000) antibodies overnight at 4°C, followed by simultaneous incubation with Cy5-linked goat anti-mouse or goat anti-rabbit secondary antibody (1:1,000) and 100 nmol/L SYTOX Green nucleic acid stain (Invitrogen) at room temperature for 1 hour. Images were collected on a Leica TCS SP confocal system with ×40 or ×100 magnification.

siRNAs for gp130 (siGP130.1: Santa Cruz, #sc-29333; siGP130.2, siGP130.3, siGP130.4: OriGene, #SR302381) and Stat3 (siSTAT3.1: Santa Cruz, #sc-29493; siSTAT3.2: Santa Cruz, #sc44275) were used in this study. Subconfluent OVCAR-8 cells were transfected in 6- or 96-well plates with 80 nmol/L gp130, Stat3, or control siRNA according to the manufacturer's protocol. Protein expression was tested by immunoblotting after 72 hours of siRNA exposure. Cell viability was evaluated by MTT staining after 48, 72, and 96 hours of siRNA exposure.

Athymic mice (Charles River Laboratories) were used for in vivo efficacy studies. Mice were fed ad libitum and kept in air-conditioned rooms at 20°C ± 2°C with a 12-hour light–dark period. Animal care and manipulation were in agreement with the USC institutional guidelines, which were in accordance with the Guidelines for the Care and Use of Laboratory Animals.

OVCAR-8 cells in the logarithmic growth phase from in vitro cultures were implanted in athymic mice (5 × 10⁵ cells in 100 μL of PBS/mouse) under aseptic conditions as previously described (18). Tumor growth was assessed by biweekly measurement of tumor diameters with a Vernier caliper. Tumor volume was calculated according to the formula: Tumor volume (mm³) = D × d²/2, where D and d are the longest and shortest diameters, respectively. For intraperitoneal (i.p.) administration, tumors were allowed to grow to an average volume of 70 mm³. Mice were then randomly assigned into 2 groups, 4 mice per group, for daily vehicle control or SC144 (10 mg/kg) treatment. To prepare SC144 solution for i.p. administration, 200 mg/mL DMSO stock solution of SC144 was diluted to 20 mg/mL in propylene glycol and further added to 0.9% NaCl with 40% propylene glycol. For oral administration, tumors were allowed to grow to an average volume of 40 mm³. Mice were then randomly assigned into control (5 mice) and treatment (4 mice) groups. SC144 (100 mg/kg) was orally delivered every day to the treatment group. To prepare SC144 solution for oral administration, 300 mg/mL DMSO stock solution of SC144 was mixed with sesame oil. Mouse body weight and tumor volume were measured twice a week.

Upon autopsy, tumors, kidneys, livers, pancreas, spleens, hearts, lungs, and brains from mice were collected, fixed in formalin, embedded in paraffin, and sectioned. Sections (4 μm) were stained with hematoxylin and eosin to facilitate histologic examination.

Immunohistochemistry was conducted on formalin-fixed, paraffin-embedded sections, following antigen retrieval. Internal hydrogen peroxidase activity was blocked with 3% hydrogen peroxide for 15 minutes. Tissue sections were subsequently blocked by Power block for 10 minutes and 10% goat serum for 30 minutes, followed by incubation with indicated antibodies at 4°C overnight. Anti-von Willebrand factor antibody (Dako, A0082) was kindly provided by Dr. Florence M Hofman (University of Southern California, Keck School of Medicine). Anti-Ki67 mAb (Thermo Scientific, RM-9106-S) was kindly provided by Dr. Bangyan Stiles (University of Southern California, School of Pharmacy). Anti-CD31 antibody (Santa Cruz Biotechnology, Inc., sc-8306) was kindly provided by Dr. Jason Wu. Anti-cleaved caspase-3 antibody was purchased from Cell Signaling, Inc. (9661). Tissue sections were then incubated with biotin-conjugated secondary antibody (BioGenex, HK337-5G) for 20 minutes and peroxidase-conjugated streptavidin (BioGenex, HK330-5K) for 20 minutes. TSA Plus Cyanine 3 Evaluation Kit (PerkinElmer, Inc., NEL744E001KT) was used to generate fluorescent signals, whereas nucleus was stained by 4′,6-diamidino-2-phenylindole (DAPI). Terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) staining was conducted according to manufacturer's instruction (Roche, 11 684 817 910).

The Student t test was used for statistical analysis and P value determination via SPSS 16.0 (SPSS Inc.) and Prism 5 (GraphPad Software, Inc.). Differences were considered statistically significant at P < 0.05.

SC144 inhibits cell growth in a panel of human ovarian cancer cell lines with IC₅₀ values in a submicromolar range (Table 1, Fig. 1A). The potency of SC144 toward NCI/ADR-RES (paclitaxel- and doxorubicin-resistant, IC₅₀ = 0.43 μmol/L; refs. 19–21) and HEY (cisplatin-resistant, IC₅₀ = 0.88 μmol/L) suggests an ability to overcome drug resistance in ovarian cancer. SC144 also inhibited colony formation in Caov-3 and OVCAR-8 cells (Supplementary Fig. S1A). We compared the abilities of SC144 to induce apoptosis and cell death in human ovarian cancer cells (OVCAR-8 and Caov-3) with human normal epithelial cells (kidney epithelial cells and endometrial epithelial cells). The kidney epithelial cells and the endometrial epithelial cells do not express cancer specific marker survivin (Supplementary Fig. S1B; ref. 22). SC144 induced substantial apoptosis in OVCAR-8 and Caov-3 cells, but no considerable apoptotic effect was induced in the kidney epithelial cells or the endometrial epithelial cells (Fig. 1B). Consistently, trypan blue staining showed that SC144 significantly caused cell death in OVCAR-8 and Caov-3 cells, whereas no significant cell death was observed with SC144 treatment in the kidney epithelial cells or the endometrial epithelial cells (Fig. 1C). These results suggest that SC144 is preferentially cytotoxic to human ovarian cancer cells.

SC144 treatment substantially increased the phosphorylation of gp130 (S782) in both OVCAR-8 and Caov-3 cells in a time- (Fig. 2A) and dose-dependent manner (Fig. 2B). In both cell lines, total gp130 protein was detected as 2 individual bands in immunoblotting (Fig. 2A). Silencing of gp130 using siRNAs decreased the abundance of both bands (Supplementary Fig. S2A), confirming that both bands are gp130. Treatment of OVCAR-8 and Caov-3 cells with brefeldin A (BFA), an inhibitor of protein glycosylation produced a time-dependent shift of gp130 into the lower band (Supplementary Fig. S2B), indicating that the upper band represents glycosylated gp130 and the lower band represents unglycosylated or minimal core glycosylated gp130. After 6-hour continuous SC144 treatment in OVCAR-8 and Caov-3 cells, the upper bands of total gp130 shifted to the lower band (Fig. 2A), indicating that SC144 treatment induced gp130 deglycosylation following the stimulation of gp130 phosphorylation. Also, SC144 is mechanistically different from BFA, which did not induce gp130 phosphorylation (Supplementary Fig. S2C). In addition, we found that gp130 ligands IL-6 and LIF substantially enhanced Stat3 (Y705) phosphorylation, whereas the phosphorylation of gp130 (Ser782) was only increased by LIF treatment (Supplementary Fig. S2D). These results suggest that phosphorylation at Ser782 is cytokine-specific and is not required for gp130 activation. Serine phosphorylation of surface receptors is often associated with receptor internalization. Therefore, we further examined the effect of SC144 treatment on cell surface gp130 abundance. SC144 significantly induced gp130 internalization at 10 μmol/L with 6- and 24-hour treatment, whereas at 0.5 and 2 μmol/L, no significant changes were detected at cell surface gp130 (Supplementary Fig. S2E).

Induction of gp130 phosphorylation and deglycosylation suggest that SC144 targets gp130. Binding of a small-molecule compound to its target protein can result in conformational changes and proteolytic stabilization of the protein by decreasing its sensitivity to proteases (17, 23). Similar in concept to DNA footprinting assay, DARTS assay (16, 17) was used to test the binding of SC144 to gp130. Incubation of SC144 with OVCAR-8 cell lysate protected gp130 from digestion in a dose-dependent manner as shown by the increased abundance of the gp130 band (Fig. 2C). Especially, 100 and 1,000 μmol/L of SC144 significantly increased gp130 abundance by 63% (P = 0.035) and 102% (P = 0.020) compared with vehicle control. These results show the direct binding of SC144 to gp130 and suggest that SC144 may induce conformational changes in gp130 and affect its activity.

Treatment of OVCAR-8 and Caov-3 cells with an anti-gp130 (B-R3) antibody (Fig. 3A) or silencing of gp130 using siRNAs in OVCAR-8 cells (Supplementary Fig. S3A) substantially decreased constitutive Stat3 (Y705) phosphorylation. This suggests that the constitutive Stat3 activation is mainly maintained by extracellular gp130 ligands constitutively produced by ovarian cancer cells via an autocrine mechanism (24). Therefore, Stat3 activation could be used as a predictive marker for gp130 activity. SC144 treatment suppressed the constitutive phosphorylation of Stat3 (Y705) in a time- (Fig. 3B and C) and dose-dependent manner (Fig. 3D) in OVCAR-8 and Caov-3 cells. The total Stat3 level was decreased after 24 hours of SC144 treatment (Fig. 3C), perhaps due to an autoregulatory role of Stat3 (25). In comparison, Stattic, a previously reported Stat3 inhibitor (12), inhibited Stat3 (Y705) phosphorylation in OVCAR-8 cells in a time-dependent manner (Supplementary Fig. S3B). Moreover, Stattic exerted no substantial effect on phosphorylated and total gp130, indicating that SC144-induced gp130 phosphorylation and deglycosylation do not result from the direct inhibition of Stat3 (Y705) phosphorylation.

The activation of gp130/Stat3 promotes the expression of multiple genes (26), including Bcl-2 (27), Bcl-X_L (28), Cyclin D1 (29), survivin (29), Ape1/Ref-1 (30), and MMP-7 (31), which are involved in cell growth, survival, cell-cycle progression, proliferation, and metastasis. The role of gp130/Stat3 in promoting the expression of these genes in OVCAR-8 cells was confirmed by silencing of gp130 (Supplementary Fig. S4A) and Stat3 (Supplementary Fig. S4B). Importantly, silencing of Stat3 also reduced gp130 expression, which is consistent with a previous study unveiling a STAT-binding element located in the gp130 promoter region (29). To assess the effects of SC144 on the expression of gp130/Stat3-regulated genes, we treated OVCAR-8 cells with SC144 at various times. Survivin and MMP-7 were the most sensitive to SC144 treatment; their abundance was significantly decreased during the first 24 hours (Fig. 3E), whereas cyclin D1, Bcl-X_L, Bcl-2, and Ape1/Ref-1 expression started to decrease after 24 hours. Consistent with the Stat3 knockdown data, gp130 level was also significantly reduced within the first 24 hours of SC144 treatment. It is also important to point out that SC144 is mechanistically different from BFA due to the fact that total gp130 level was decreased after SC144 treatment but was increased after BFA treatment (Supplementary Fig. S2B). In addition, Stat1 was not affected by SC144 treatment. The inhibition of the gp130/Stat3 downstream effectors (Bcl-2, BclX_L, survivin, cyclin D1, Ape1/Rel-1, and MMP-7) further validates the mechanism of SC144 cytotoxicity in ovarian cancer cells. In addition, the cytotoxicity of SC144 (Supplementary Fig. S5A) as well as its inhibitory effects on gp130/Stat3 pathway (Supplementary Fig. S5B) were also observed in 2 pancreatic cancer cell lines, Panc-1 and Mia-Paca-2, suggesting its potential use for other cancers.

Taken together, these results confirm that SC144 inhibits the activation of the Stat3 signaling pathway downstream of gp130.

We further assessed whether the inhibition of gp130/Stat3 signaling contributes to the cytotoxicity of SC144 in ovarian cancer cells. By using gp130 siRNAs or Stat3 siRNAs, we found that silencing of gp130 or Stat3 significantly inhibited OVCAR-8 cell growth (Fig. 3F). While anti-gp130 (B-R3) antibody substantially reduced the ability of OVCAR-8 cells to form colonies in a dose-dependent manner (Fig. 3G), it only induced limited apoptosis (Supplementary Fig. S6A) compared with that triggered by SC144 (Fig. 1B), suggesting there may be additional mechanisms contributing to SC144-induced apoptosis. Furthermore, pharmacologic suppression of Stat3 activity by Stattic inhibited cell growth in OVCAR-8, Caov-3, and NCI/ADR-RES cells (Supplementary Fig. S6B). Taken together, these results show that the gp130/Stat3 inhibition results in cytotoxicity in human ovarian cancer cells.

We compared the effects of SC144 on the activation of downstream signaling stimulated by gp130 cytokines, including LIF and IL-6, and non-gp130 cytokines, including an IFN family member IFN-γ, a GPCR (CXCR4) ligand SDF-1α, and a growth factor PDGF. SC144 inhibited LIF-induced phosphorylation of Stat3 (Y705) and Stat1 (Y701) in a dose- and time-dependent manner (Fig. 4A and Supplementary Fig. S7A). SC144 also inhibited LIF-induced Akt activation. Similar inhibitory effects of SC144 were observed on IL-6–stimulated phosphorylation of Stat3, Stat1 and Akt (Fig. 4B). We further examined the inhibitory effect of SC144 against different concentrations of IL-6. IL-6 stimulates Stat3 phosphorylation in a dose-dependent manner (Supplementary Fig. S7B). When 50 ng/mL IL-6 was used to induce Stat3 phosphorylation, SC144 exhibited the inhibitory effect at over 10 μmol/L (Supplementary Fig. S7C). When 3 ng/mL IL-6, which is close to the IL-6 concentration in the human body (24), was used to induce Stat3 phosphorylation, SC144 showed inhibitory effect at 2 μmol/L and a complete inhibition at 5 μmol/L. These results suggest that SC144 inhibits the downstream signaling activation elicited by gp130 ligands. In contrast, SC144 did not affect the activation of phosphorylation of Stat3, Stat1, and Akt stimulated by IFN-γ (Fig. 4C). Akt phosphorylation induced by SDF-1α (Fig. 4D) or PDGF (Fig. 4E) was not affected by SC144. Neither SDF-1α nor PDGF showed significant effects on Stat3 phosphorylation in OVCAR-8 (Fig. 4D and E). Our results show that the inhibitory effect of SC144 on the downstream signaling is gp130-specific and gp130-dependent.

The selectivity of SC144 in inhibiting the activation of downstream signaling stimulated by gp130 cytokines was also confirmed by examining the nuclear translocation of Stat3 and Stat1 induced by LIF, IL-6, or IFN-γ, which occurs following tyrosine phosphorylation. Immunofluorescence microscopy showed that SC144 treatment effectively inhibited LIF- and IL-6-induced nuclear translocation of Stat3 and Stat1 but was not able to affect STATs nuclear translocation induced by IFN-γ in OVCAR-8 cells (Fig. 5 and Supplementary Fig. S7D).

To determine in vivo efficacy of SC144, we tested its effect on established tumors after subcutaneous inoculation of human ovarian cancer OVCAR-8 cells in the flank of nude mice. Compared with vehicle control treatment, daily i.p. administration of SC144 (10 mg/kg) for 58 days significantly inhibited tumor growth by about 73% (from 835.2 ± 228.1 mm³ to 228.1 ± 92.5 mm³; P = 0.032; Fig. 6A). To assess the oral bioavailability of SC144, we tested SC144 through daily oral administration. After we sacrificed the mice in the control group on day 69 due to their large tumors, we continued SC144 treatment for another 35 days. The average tumor volume in mice receiving daily oral administration of SC144 (100 mg/kg) was about 82% smaller than that in the control group (on day 69, 975.8 ± 247.9 mm³ vs. 175.4 ± 80.8 mm³, P = 0.014; Fig. 6B). No significant body weight loss was detected during both treatment periods (Supplementary Fig. S8A and S8B). In addition, no substantial toxicity was observed in liver, kidney, spleen, lung, heart, pancreas, and brain (Supplementary Fig. S8C and S8D), further showing that SC144 did not produce significant adverse effects in mice at its effective anticancer dosage.

While the tumors excised from control mice appeared red, the ones from SC144-treated mice were pale (Fig. 7A), suggesting that SC144 treatment effectively inhibited tumor angiogenesis. SC144-induced reduction in tumor microvessel density was further confirmed by the significant decrease in both CD31 and von Willebrand factor (vWF) staining (Fig. 7A and Supplementary Fig. S8E). SC144 treatment also resulted in extensive areas of necrosis in the tumor (Fig. 7B and Supplementary Fig. S8F). Consistent with our in vitro data, SC144 treatment triggered apoptosis and suppressed cell proliferation in tumor, as is confirmed by the significant increase in TUNEL and cleaved caspase-3 staining and significant decrease in Ki67 staining, respectively (Fig. 7C and Supplementary Fig. S8G). The levels of gp130, MMP-7, Bcl-X_L, Bcl-2, and Ape1/Ref-1 in the tumor were substantially decreased in the SC144 treatment group (Fig. 7D), which is consistent with our in vitro data (Fig. 3E). Phospho-gp130 (S782) was decreased after SC144 treatment, which may result from the downregulation of total gp130 protein. The decrease in Bcl-X_L, Bcl-2, and Ape1/Ref-1 explains the extensive necrosis in the tumor tissue from the SC144 treatment mice due to the roles of these gp130 effectors in protecting cancer cells from apoptosis (32, 33), whereas the reduced level of MMP-7 contributes to both apoptosis and anti-angiogenesis (34, 35).

IL-6 and Stat3 are involved in cancer progression and drug resistance in a variety of human cancers including ovarian cancer. Inhibition of IL-6 secretion by CDDO-Me induced apoptosis and sensitized drug-resistant ovarian cancer cells to paclitaxel and cisplatin (36). CNTO 328, a monoclonal antibody to IL-6, is currently in a phase I clinical trial for ovarian cancer. On the other hand, Stat3 is persistently active in ovarian cancer cell lines but not in normal ovarian epithelial cells (37). In vitro and in vivo studies confirmed that inhibition of constitutive Stat3 activation by using a Jak-selective inhibitor, a dominant-negative Stat3, or short hairpin RNA (shRNA) suppressed the growth and development of ovarian cancer (38–41). Considering that gp130 is positioned at the junction of this oncogenic signaling network and is decisive for the network activation, blocking gp130 activity using nonpeptidic small molecules is an attractive therapeutic approach to gp130-dependent cancers. However, so far no selective small-molecule gp130 inhibitors have been developed. In our study, we developed a quinoxalinhydrazide compound SC144 that inhibits gp130 cytokine signaling, as well as the downstream gene expression, and shows in vitro and in vivo potency in ovarian cancer. Because of the fact that the gp130 cytokine, IL-6, is constitutively produced and the key effectors (Stat3, Akt) in the gp130 downstream signaling pathways are persistently active in ovarian cancer (24, 42–44), ovarian cancer cells might be more sensitive to gp130 inhibitors than normal cells. This concept is in accord with our results that SC144 is more cytotoxic in human ovarian cancer cells than human normal epithelial cells and inhibits human ovarian tumor growth in a mouse xenograft model without obvious toxicity to normal tissues. To our knowledge, SC144 is the first-in-class small-molecule gp130 inhibitor with oral activity in ovarian cancer. In addition, our results showing that gp130 persistently activates Stat3 in ovarian cancer cells, promoting cell growth and proliferation, are consistent with a similar discovery in colitis-associated tumorigenesis (45). Although the identification of amino residues on gp130 that SC144 interacts and the elucidation of off-targets are beyond the scope of this present study, the following evidence cumulatively show that gp130 is directly inhibited by SC144: (i) SC144 binds to gp130 (Fig. 2C) and induces gp130 (S782) phosphorylation and deglycosylation (Fig. 2A and B); (ii) anti-gp130 antibody is able to completely suppress constitutive Stat3 activation (Fig. 3A), indicating that Stat3 activation as an indicator for the effect of SC144 on gp130; (iii) SC144 inhibits Stat3 activation and the expression of downstream genes (Fig. 3B–E); and (iv) SC144 potently inhibits the activation of the downstream signaling pathways stimulated by gp130 ligands but shows no effect on the activation of the downstream signaling pathways stimulated by non-gp130 ligands (Figs. 4 and Fig. 5 and Supplementary Fig. S7).

The role of phospho-S782 in gp130 was first documented in 2000 to downregulate gp130 cell surface expression (5). Recently, it has been reported that phospho-S782-associated downregulation of gp130 cell surface expression is mediated by immediate internalization (46). Indeed, phosphorylation of many transmembrane receptors at Ser/Thr residues is commonly associated with the negative modulation of receptor activity (47). The induction of gp130 deglycosylation and the decrease in total gp130 protein level may also contribute to Stat3 inhibition at a later time. However, it required 10 μM of SC144 to induce gp130 internalization (Supplementary Fig. S2E), a concentration higher than for triggering gp130 deglycosylation. This may suggest that the amount of deglycosylation induced by SC144 at 2 μmol/L may not be enough to immediately initiate receptor internalization. However, the in vivo concentration of SC144 exceeds 10 μmol/L under our experimental conditions. Moreover, inhibition of gp130/Stat3 amplifies the cytotoxicity of SC144 by downregulating the expression of several downstream genes, such as Bcl-2, Bcl-X_L, cyclin D1, MMP-7, Ape1/Ref-1, and survivin, in ovarian cancer cells that results in the submicromolar IC₅₀ values after 72-hour SC144 treatment (Table 1).

On the basis of results presented in this study, we propose a plausible model for the anticancer mechanism of SC144 in ovarian cancer (Fig. 8). SC144 binding induces conformational changes and activity loss in the cell surface–bound gp130 receptor, exposing the Ser782 residue to protein kinases such as CaMKII, CaMKIV, and MK2 (6, 46). The decrease in gp130 activity leads to the inactivation of downstream signaling pathways involving Stat3 and Akt. The suppression of gp130/Stat3-mediated gene expression amplifies the potency of SC144 due to the inhibition of cell-cycle progression and angiogenesis and the promotion of apoptosis and cell death. In addition, the downregulation of the PI3K/Akt pathway also contributes to apoptosis and cell death. Our working model may serve to identify a unifying anticancer mechanism likely to be shared by future gp130 inhibitors.

Clinically, in patients with advanced ovarian cancer, high plasma levels of IL-6 correlate with poor prognosis (48, 49), and increased levels of IL-6 are also present in malignant ascites (50). Therefore, biopsy detection of IL-6 levels appears to be one strategy to identify subjects/patients who may benefit most from IL-6/gp130-blocking therapies. In addition, other components of the gp130 receptor complexes (e.g., OSMR-β, LIFR-β) as well as other gp130 cytokines (e.g., OSM) have been reported to be highly expressed in human ovarian carcinoma tissues and have the ability to activate downstream signaling pathways (e.g., Stat3 pathway; ref. 51). On the basis of such observations, these molecules can also be used as markers for selecting target subjects/patients. In fact, this evidence also supports that gp130 antagonists may be more effective for ovarian cancer treatment compared to IL-6 inhibitors (e.g., anti-IL-6 mAb), as other gp130 cytokines may compensate the effects of IL-6.

No potential conflicts of interest were disclosed.

Conception and design: S. Xu, N. Neamati

Development of methodology: S. Xu, F. Grande, A. Garofalo, N. Neamati

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): S. Xu, F. Grande, A. Garofalo, N. Neamati

Writing, review, and/or revision of the manuscript: S. Xu, N. Neamati

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): N. Neamati

Study supervision: N. Neamati

The authors thank Drs. Roger Duncan, Ebrahim Zandi, and Joe Miller for their comments and suggestions and also thank Dr. Bangyan Stiles for her kind assistance for immunohistochemical staining.

This study was supported by funds through the Department of Defense Ovarian Cancer Program (OCRP) Idea Award (W81XWH-07-1-0414) to N. Neamati.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.