Purified PSII, isolated from Rhizoma paridis (25), was provided by the Department of Pharmacology at Sichuan University (Chengdu, Sichuan, China). VEGF was obtained from R&D Systems (Minneapolis, MN, USA). VEGF ELISA kit was purchased from R&D Systems. Growth factor-reduced Matrigel was from BD Biosciences (San Jose, CA, USA). Antibodies against VEGFR2, VEGF, Bcl-2 and Bcl-xL were obtained from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA, USA). β-actin antibody was purchased from Sigma Chemical Co. (St. Louis, MO, USA). CD31 antibodies were from Epitomics, Inc. (Burlingame, CA, USA).

Primary human umbilical vascular endothelial cells (HUVECs) were from Sichuan university, China. The human high-grade serous ovarian cancer SKOV3 cell line was purchased from the American Type Culture Collection (ATCC; Rockville, MD, USA). SKOV3/vector and SKOV3/IκBαM cells were obtained from the University of Texas M.D. Anderson Cancer Center. HUVECs were maintained in endothelial cell culture medium (ECM) (Lonza, Walkersville, MD, USA) as per the manufacturer’s protocol. The SKOV3 cell line was cultured in RPMI-1640 medium (Gibco-BRL, Life Technologies Gaithersburg, MD, USA) supplemented with 10% fetal bovine serum (FBS) (HyClone, Logan, UT, USA) at 37°C under a humidified 95:5% (v/v) mixture of air and CO₂.

Cells (5×10³ cells/well) were directly incubated with indicated concentrations of PSII for indicated durations. Carrier dimethyl sulfoxide (DMSO) (<0.1%) was used as a negative control. The cell viability was examined by the 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide (MTT; Sigma) assay (26). The cytotoxic effects of PSII were expressed as the 50% inhibitory concentration (IC₅₀).

The terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling assay was performed as previously described (25). Briefly, cells were treated with PSII after seeding on coverslides for 24 h. All cell layers were fixed by a 4% paraformaldehyde solution, washed in phosphate-buffered saline (PBS), and then permeated by a permeabilization solution (0.1% Triton X-100 solution) for 2 min at 4°C. The apoptosis index was determined as the percentage of TUNEL-positive cells/1,000 4,6-diamidino-2-phenylindole (DAPI)-stained nuclei. Random fields were recorded using fluorescence microscopy (confocal scanning laser microscopy, Leica TCS4D; Germany).

The chemotactic motility was carried out using a Transwell migration assay with 8.0-μm pore polycarbonate filter inserts as previously described (27). Briefly, the well was coated with 0.1% gelatin. HUVECs (2×10⁴ cells/well) in ECM medium supplemented with 0.5% FBS were placed in the top chamber. In the bottom chamber, VEGF (50 ng/ml) in ECM medium was used as a positive control. Conditioned media from SKOV3/vector or SKOV3/IκBαM cells cultured in the presence or absence of PSII were used as a chemoattractant. After 18 h of incubation, the migrated cells were fixed with 4% paraformaldehyde and stained with 1% crystal violet. Images were captured using an Olympus inverted microscope (Olympus; magnification, ×200). The migrated cells were counted. Three independent experiments were carried out.

A tube formation assay was carried out as previously described (28). Briefly, each well of pre-chilled 24-well plates was coated with 100 μl reduced growth factor Matrigel (2.8 mg/ml; BD Biosciences). Matrigel was polymerized at 37°C for 45 min. SKOV3 cells were treated with or without PSII and then incubated with fresh media without PSII for 24 h. The conditioned media were collected and used to study the in vitro tube formation assay. Endothelial cells (1.5×10⁴ cells/ml) in ECM medium supplemented with 0.5% FBS were placed onto the layer of Matrigel. VEGF (50 ng/ml) was used as a positive control for tube formation. After 20 h of incubation at 37°C, tubulogenesis was fixed and photographed using a Zeiss microscope equipped with a digital camera and Matrox Intellicam imaging software. Tubular structures were quantified using AngioSys software (TCS Cellworks) as per the manufacturer’s instructions.

The whole-cell lysates were obtained using radioimmunoprecipitation assay (RIPA; Sigma) buffer (20 mmol/l Tris, 2.5 mmol/l EDTA, 1% Triton X-100, 1% deoxycholate, 0.1% SDS, 40 mmol/l NaF, 10 mmol/l Na₄P₂O₇ and 1 mmol/l phenylmethylsulfonyl fluoride). Protein concentrations were determined using the Bradford assay (Bio-Rad, Hercules, CA, USA) and equalized before loading. Twenty five micrograms of total proteins were resolved using 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). Primary antibodies against Bcl-2 (1:1,000), Bcl-xL (1:1,000), VEGF (1:2,000), IκBα (1:1,000) p-IκBα (1:1,000, Ser32), phosphorylated P65 (1:2,000, Ser536), P65 (1:2,000), IκκB (1:1,000) (all from Cell Signaling, Danvers, MA, USA), and HRP-conjugated secondary antibodies were used for the studies. Relative optical density of the protein signals of interest were qualified by ImageJ software (NIH).

EMSA study was carried out as previously described (13). Briefly, SKOV3 cells were pretreated with various concentrations of PSII for 24 h. Nuclear extracts were prepared. EMSA was carried out using Ds-Cold-NF-κB probes: 5′-AGT TGA GGG GAC TT CCC AGG C-3′ and 5-TGG GGA ACC TGT GCT GAG TCA CTG GAG-3′. The positive control was the nuclear extracts from the SKOV3 cells treated with 50 ng/ml TNF for 30 min. The negative control was the extracts from the SKOV3 cells without any treatment.

The IP and in vitro kinase assays were carried out as previously described (29). Briefly, reaction mixtures (25 μl) contained 40 mM β-glycerophosphate, 7.5 mM MgCl₂, 5% glycerol, 7.5 mM EGTA, [γ-³²P] ATP (0.2 mM, 1 μCi), 1 mM orthovanadate, 50 mM NaF and 0.1% (v/v) β-mercaptoethanol. The cytoplasmic extract (2 mg) immunoprecipitated with the appropriate antibody was used for the phosphorylation reaction and was washed with lysis buffer containing 50 mM Tris-HCl (pH 7.5), 120 mM NaCl, 5 mM EDTA, 50 mM NaF, 0.2 mM Na₃VO₄, 1 mM DTT, 0.5% NP-40 and protease inhibitors (Protease Inhibitor Cocktail Tablets; Boehringer, Mannheim; 1 tablet/50 ml) or with 1 μg of purified recombinant GST-IκBα (Cell Signaling) at 37°C for 1 h. Reactions were stopped by adding 1 volume of Laemmli sample buffer containing 5% β-mercaptoethanol and resolved on a 4–20% SDS/PAGE gel. Gels were autoradiographed and bands were counted using a Molecular Dynamics PhosphorImager software.

Paraffin tumor sections (5 μm) were derived from solid tumor sections that were resected, fixed with 10% formaldehyde and embedded in paraffin. Sections were treated with 0.3% hydrogen peroxide at room temperature to block endogenous peroxidase activities ensued by a 5% bovine serum albumin incubation. Tumor sections were exposed to antibodies against CD31 (1:100) and VEGFR2 (1:100). Images were captured by a Leica DM 4000B photomicroscope (Solms, Germany; magnification, ×200 and ×400). The microvessel density was based on CD31 immunohistochemical signals calculated by Image-Pro Plus 6.0 program (Media Cybernetics) (n=5). Hematoxylin and eosin (H&E) staining was carried out using standard techniques.

The human ovarian mouse tumor model has been previously described (26). Briefly, 4- to 6-week-old female Balb/c nude mice (Chengdu Experimental Animal Center, Chengdu, China) were randomly divided into 6 groups (n=5). SKOV3/vector or SKOV3/IκBαM cells (5×10⁶ cells/100 μl) were injected subcutaneously into the mice. One week after the implantation, mice were treated with PSII (15 and 25 mg/kg) by daily intraperitoneal injections. The administrations were carried out on 4 consecutive days/week for 4 weeks (between day 8 and 35). Control groups received the control solution containing the same amount of DMSO (v <0.1%) without PSII. The body weights and tumor volumes were recorded twice weekly. The volume (V) of the solid tumors was measured by a caliber and calculated according to the formula: V = length × width² × 0.52. The implanted tumors and vessels were monitored by a Philips HD11 ultrasound scanner (Philips Medical Systems, Best, The Netherlands) equipped with a 11 MHz linear array transducer. The volume of solid tumors (expressed in millimeters) was documented in three dimensions (length, width and height). The minimum diameter of the lesion that can be detected by ultrasound is 0.01 cm. At the termination of the present study, all mice were euthanized using carbon dioxide asphyxiation. All experiments were conducted based on the National Institutes of Health Guidelines for the Care and use of Experimental Animals. All protocols were approved by the Animal Investigation Committee of the Institute for Nutritional Sciences (Shanghai, China).

Statistical significance of differences between groups was performed using one-way ANOVA followed by the Student’s t-test. Data are presented as mean ± standard error (SE). IC₅₀ values were calculated by SPSS software version 13.0 (SPSS, Inc., China). A value of p≤0. 05 was considered to indicate a statistically significant result.

In our previous study, we demonstrated that PSII had no effects on the survival of non-tumorigenic human vascular smooth muscle, human bronchial or ovarian surface epithelial (OSE) cells. However, PSII decreased the cell viability and inhibited the growth of human tumor cell lines including a high-grade serous ovarian cancer cell line SKOV3 in a concentration-dependent manner (28). Here, we further demonstrated that PSII exhibited more potent antitumorigenic activity than VP16-etoposide. Kinetic studies showed that PSII treatment inhibited SKOV3 cell growth in a concentration-dependent manner. Most important, at the same concentration and at the same time point, PSII killed more cells than VP16, a positive control (Fig. 1). PSII treatment resulted in lower IC₅₀ values (20.99, 10.44, 8.83 and 6.98 μM, day 1–4, respectively) than those of VP16 (82.04, 17.18, 11.80 and 8.01 μM, day 1–4, respectively).

A previous study showed that NF-κB plays a crucial role in SKOV3 tumor cell growth (13). The transfection of SKOV3 cells with IκBαM reduced the DNA binding and gene transcription activity of NF-κB. using this model system, we wanted to further characterize the antitumorigenic property of PSII under a condition wherein the NF-κB signaling pathway is compromised. Kinetic studies (Fig. 1) demonstrated that the combination treatment (PSII and the transfection of IκBαM into SKOV3 cells) rendered marked inhibitory effects on tumor cell growth. The antitumorigenic effect was more effective compared to the PSII only treatment. Notably, this combination (PSII and IκBαM transfection) was more potent than a combination treatment of IκBαM transfection and VP16, an antimicrotubule agent (our positive control). For example, the maximum inhibitory ratio achieved with 10 μM PSII treatment and IκBαM transfection following a 4-day treatment was 82% compared with the 65% inhibition produced by 10 μM VP16 treatment and IκBαM transfection. In this model, PSII treatments also yielded lower IC₅₀ values compared to those of the VP16 treatments. Since our interest focused on the effects of PSII on NF-κB activity, VEGF-mediated angiogenesis, and the molecular mechanisms underlying such effects but not cell death, the above studies allowed us to identify the sub-cytotoxic concentrations of PSII to be used for most of the subsequent studies (2.5 and 5 μM).

Cell motility and tubulogenesis are crucial steps in angiogenesis (30). Therefore, we wanted to determine whether PSII treatment affects the ability of tumor cells to induce endothelial cell migration and tube formation. First, using the Boyden chamber assay, we showed that conditioned media from the PSII-treated SKOV3 cells failed to induce HUVEC migration as compared to the control-treated group (carrier DMSO, <0.1%) and the VEGF-induced migration group (Fig. 2A and B). We further demonstrated that PSII treatment also modulated SKOV3 cell-induced tubulogenesis. As shown in Fig. 2D, conditioned media from the PSII-treated SKOV3 cells also failed to induce tube formation in a concentration-dependent manner compared to the control treatment (VEGF only, 50 ng/ml; and DMSO carrier, <0.1%). The disruption of tube formation was observed in HUVECs following treatment with cytotoxic levels of PSII (2.5 μM) (28). We also measured the VEGF level in the conditioned medium obtained from the PSII-treated SKOV3 cells. A marked reduction in VEGF levels was observed in the conditioned media of the PSII-treated SKOV3 cells compared to the control (Fig. 2C). We further examined the possibility that the reduced VEGF level may be attributed to cell death. As shown in Fig. 3C, TUNEL staining studies showed a minimal increase in the percentage of TUNEL-positive SKOV3 cells following treatment with 2.5 μM PSII compared to the control. Together, the results suggest that PSII treatment modulates VEGF levels in the tumor microenvironment leading to reduced capillary formation.

Since a previous study (13) demonstrated that NF-κB plays an important role in SKOV3 tumor cell growth and PSII treatment affects the levels of SKOV3-secreted VEGF, a target of the NF-κB signaling pathway (31), we wanted to examine whether PSII modulates NF-κB activity in SKOV3 cells. In Fig. 3A, EMSA data indicated that the binding of NF-κB to its promoter DNA consensus sequence was reduced in the PSII- treated cells at a concentration as low as 2.5 μM. Notably, PSII treatment (5 μM) and the transfection of IκBαM into SKOV3 cells yielded a comparable reduction in DNA binding. Such observable effects of PSII on NF-κB transcriptional activity led us to examine potential changes in the expression of downstream NF-κB targets. In Fig. 3B, we showed that PSII treatment modulated the expression of well-known NF-κB targets, e.g. Bcl-2, Bcl-xL and VEGF (31). PSII treatment led to the downregulation of these targets in a concentration-dependent manner. This finding supports the previous experiment showing the reduced levels of secreted VEGF in the conditioned media of PSII-treated cells (Fig. 2C). Together, the results indicated that PSII treatment compromises NF-κB activation and reduces the expression of several NF-κB-downstream targets known to play pivotal roles in tumor growth biology.

Since the binding of NF-κB to its promoter DNA consensus sequence and the expression of several NF-κB-downstream targets were altered by the treatment of PSII, we wanted to determine whether IκBα or P65 levels were altered in PSII-treated cells. Protein signals of IκBα, phosphorylated-IκBα, P65, phosphorylated-P65 and IKKβ were assessed. Protein signals from the extracts of SKOV3, SKOV3/vector and SKOV3/IκBαM cells were used as controls. The results from Fig. 4A showed that while there was no change in the total IκBα and P65 levels in all conditions, unexpectedly, PSII treatment (2.5 μM) appeared to reduce IKKβ expression. Consistent with this observation, PSII treatment also reduced phosphorylation of IκBα and P65 on Ser32 and Ser536, respectively.

Next, we examined whether PSII treatment also affects IKKβ kinase activity on its substrate IκBα. Immunoprecipitated IKKβ from the extracts of the PSII-SKOV3 treated cells and carrier DMSO-treated cells were used in an in vitro kinase assays. Fig. 4B showed that control cells had stronger IKKβ kinase activity as compared to that of the PSII-treated SKOV3 cells. Indeed, PSII treatment suppressed IKKβ activity in a concentration-dependent manner. Together, these results indicate that PSII targets IKKβ leading to a reduction in NF-κB signaling and the expression of NF-κB-downstream targets.

Our previous study showed that PSII suppresses tumor growth in a xenograft mouse model of ovarian cancer (25). To characterize the anti-angiogenic property of PSII in vivo and the relationship with NF-κB signaling pathway, we continued to use this xenograft model employing SKOV3/vector and SKOV3/IκBαM cells. Consistent with our previous finding (28), PSII treatments suppressed tumor growth rates and reduced the tumor weights and tumor sizes as compared to the control treatment (Fig. 5A and C). In our previous study, we demonstrated that color Doppler ultrasound can be used as a non-invasive method to assess angiogenesis (28). using this approach (Fig. 5B), here, we confirmed that, on day 35, before the termination of the in vivo tumor growth study, PSII rendered profound inhibitory effects on neovascularity refected by the stark reduction in spectral and color Doppler signals in the PSII-treated groups compared to that of the control mice. The maximum diameter of blood vessels and microvessel density were significantly reduced (p<0.05) (Fig. 5C). The aforementioned reductions were also correlated with the suppression of tumor growth evident by a >50% reduction in tumor wet weights as compared to that of the controls (p<0.05) (Fig. 5C).

Notably, in combination with the transfection of the super-engineered repressor of IκBα, i.e. IκBαM (S32A and S36A) known to inhibit NF-κB activity, the antitumorigenic effects of PSII were enhanced significantly. While the effects of PSII treatment alone or in combination with IκBαM rendered similar inhibitory effects on tumor wet weights and tumor volumes, the combination treatment markedly suppressed the growth rate and reduced the number of vessels and the maximum diameter of blood vessels (by >50%) compared to that of the PSII only treatment (Fig. 5C). Clearly, the combination treatment exerts marked and extremely effective anti-angiogenic effects. To confirm this observation, we used immunohistochemistry to assess signals of angiogenesis markers, i.e. VEGFR2 and CD31, on tumor sections of the grafts from the treated or control mice. Our results confirmed that PSII treatment clearly reduced the expression of VEGFR2, the key player of VEGF signaling and microvessel density refecting by CD31 signals (Fig. 6A). These data also suggest the possibility of using PSII in combination with a drug that inhibits NF-κB signaling. Together, the data in the present study allowed us to elucidate a novel antitumorigenic and anti-angiogenic feature of PSII (Fig. 6B). For the first time, we provide evidence showing that PSII potently inhibits angiogenesis and the growth of human ovarian cancer by suppressing NF-κB signaling in cancer cells leading to the suppression of pro-angiogenic factors of the tumor microenvironment, e.g. VEGF.