Oridonin Inhibits Tumor Growth and Metastasis through Anti-Angiogenesis by Blocking the Notch Signaling

While significant progress has been made in understanding the anti-inflammatory and anti-proliferative effects of the natural diterpenoid component Oridonin on tumor cells, little is known about its effect on tumor angiogenesis or metastasis and on the underlying molecular mechanisms. In this study, Oridonin significantly suppressed human umbilical vascular endothelial cells (HUVECs) proliferation, migration, and apillary-like structure formation in vitro. Using aortic ring assay and mouse corneal angiogenesis model, we found that Oridonin inhibited angiogenesis ex vivo and in vivo. In our animal experiments, Oridonin impeded tumor growth and metastasis. Immunohistochemistry analysis further revealed that the expression of CD31 and vWF protein in xenografts was remarkably decreased by the Oridonin. Furthermore, Oridonin reinforced endothelial cell-cell junction and impaired breast cancer cell transendothelial migration. Mechanistically, Oridonin not only down-regulated Jagged2 expression and Notch1 activity but also decreased the expression of their target genes. In conclusion, our results demonstrated an original role of Oridonin in inhibiting tumor angiogenesis and propose a mechanism. This study also provides new evidence supporting the central role of Notch in tumor angiogenesis and suggests that Oridonin could be a potential drug candidate for angiogenesis related diseases.


Introduction
Tumor neoangiogenesis not only supplies nutrients and oxygen to enhance tumor growth but also, especially, provides the principal route of tumor metastasis which is the main cause of morbidity and mortality in most cancers [1]. During tumor metastasis, cancer cells escape from the primary tumor, enter into lymphatic or blood circulation (intravasation), and then cross the vessel endothelial cell layer to enter the parenchyma of the target organ. The vasculature endothelial cell layer is a natural barrier for tumor cell trans-endothelial migration and invasion [2]. Therefore, transendothelial migration (TEM) is a critical step in the metastatic dissemination of malignant cells from a primary tumor to distant vital organs [3]. Inhibition of TEM may be an effective strategy of suppressing tumor growth and metastasis.
Among the complex signaling pathways regulating endothelial cell-cell contacts, which determine the permissibility of TEM, the Notch signaling pathway is critical [4]. Notch signaling is an evolutionarily conserved pathway that regulates cell fate decisions during various developmental processes [5]. In mammals, there are five ligands (Jagged 1, 2, Delta-like 1, 3, 4) and four Notch receptors (Notch 1-4), which are expressed on the cell surface [5]. Upon Notch ligand binding to a receptor on an adjacent cell, the intracellular portion of the receptor is cleaved and translocates into the nucleus, leading to the expression of downstream genes such as Hes-1 and HESR1.
A growing body of evidence suggests that Notch is an attractive target to block tumor metastatic progression [6]. Notch mediates communication and interactions between endothelial cells and tumor cells, and promotes tumor angiogenesis [7,8,9]. In the tumor microenvironment, Jagged ligands can be induced by tumor-associated growth factors such as VEGF [10], followed by activating Notch expressed in tumor endothelial cells [11]. Curiously, Jagged 2 is the Notch ligand most significantly correlated with overall and metastasis-free survival of breast cancer patients [12]. Vascular endothelial cells express the Notch receptors 1, 2 and 3, and Notch signaling is critical to the proper formation of a functional vasculature [13]. Also, Notch activity was specifically upregulated in the tumor endothelium, suggesting that interfering with Notch activity may negatively affect tumor neoangiogenesis. Several Notch inhibitors such as RO4929097 [14] and MK-0752 [15] have already been used in clinical trials. Therefore, targeting the Notch pathway in endothelial cells might provide a valid strategy for antiangiogenic therapies [16].
Oridonin, an effective diterpenoid component isolated from the medicinal herb Rabdosia nervosa (Hemsl) [17], has various antibacterial, anti-inflammatory, proapoptotic and anti-tumor activities as well as other pharmacological properties [18,19,20,21,22]. Linda C. Meade-Tollin et al. reported Oridonin inhibits formation of capillary-like networks, which implied Oridonin exhibites antiangiogenesis activity [23]. However, the mechanism of Oridonin action on tumor anigiogenesis remains unknown. In this study, we investigated the mechanism of Oridonin in suppressing tumor growth and metastasis through inhibiting tumor angiogenesis by blocking the Jagged-Notch signaling pathway.

Cell Culture and Proliferation Assay
HUVECs (ScienCell Research Laboratories, San Diego, CA, USA) were purchased from Science Research Laboratories and cultured in complete ECM (Sciencell) supplemented with 5% FBS. HCT116 cells were obtained from the American Type Tissue Collection (ATCC, Manassas, VA, USA) and maintained in DMEM supplemented with 10% FBS (Gibco BRL Life Technologies, Eggenstein, Germany). 4T1 mammary carcinoma cell was purchased from ATCC and maintained in RPMI-1640 medium supplemented with 10% FBS, 1% Glutamax-1 and 1% penicillin-streptomycin. All cells were maintained at log phase at 37˚C with 5% carbon dioxide. Cell proliferation was determined by the Promega CellTiter 96 (Promega, Madison, WI, USA) nonradioactive cell proliferation assay according to manufacturer's instruction [24]. All experiments were performed in triplicate and repeated at least three times.

Tube Formation and Migration Assay
In vitro angiogenesis was assessed with tube formation and migration assays. Briefly, 1610 4 HUVECs were seeded on Matrigel with or without different concentrations of Oridonin followed by the addition of VEGF (20 ng/ml). After about 8 hours, photomicrographs were taken with an OLYMPUS inverted microscope. Tubular structures were quantified by Image-Pro Plus 6.0 software, and the inhibition percentage was expressed using untreated wells as 100%. HUVEC migration was determined with a wound healing migration assay and a modified Boyden chamber assay. Confluent HUVECs were pretreated with Mitomycin C for 2 hours before incubating with VEGF (20 ng/ml) and Oridonin for about 8-12 hours. Migrated cells were photomicrographed and counted manually. The modified Boyden chamber model (Transwell, 8.0 mm pore size; Costar) was used as previously described [25]. HUVECs at 80% confluence were serum-starved for 24 hours. Next, HUVECs (4610 4 ) in 100 ml of serum-free medium were plated in the transwell insert. Then 600 ml of fresh basic ECM medium containing 10 ng/ml vascular endothelial growth factor (VEGF) and different concentrations of Oridonin were added into the bottom well. After 4 hours of incubation, cells were washed with PBS to remove the un-invaded cells, fixed with 4% paraformaldehyde and stained with crystal violet. Migrated cells were imaged using Olympus IX70 inverted microscope connected to a DXM1200 digital camera and counted manually.

Aortic Ring Spreading Assay
Aortic ring spreading assay was performed as previously described with modification [26]. 48-well plates were coated with 100 ml of Matrigel per well and polymerized at 37˚C for 30 minutes. Aortic rings harvested from 6 to 8-week-old Sprague-Dawley rats were plated in the wells and overlaid with 100 ml of Matrigel for sealing. VEGF (100 ng/ml) in serum-free culture medium, with or without 5 mM Oridonin was added. The medium was changed every two days. Microvessel-like properties of sprouting structures were observed and photographed on the 7 th day and counted by Image-Pro Plus 6.0 software.

Mouse Corneal Micropocket Assay
Corneal angiogenesis was assessed as described previously [27]. Briefly, in one eye of 6 to 8-week-old male C57BL/6 mice, corneal micropockets were created using a modified needle. Micropellet containing 250 ng VEGF was implanted into each corneal pocket. After injecting 5 mg/kg Oridonin intraperitoneally for seven days, eyes were imaged under the microscope. The length and clock number of new blood vessels were surveyed, and the area of neovascularization was calculated using the formula Area (mm 2 ) 50.263.146VL6CN, where VL is the maximal vessel length extending from the limbal vasculature toward the pellet, CN is the clock hours of neovascularization and 1 clock hour equals 30 degrees of arc.

Xenograft Mouse Tumor Model
Male nude mice, 5-6 weeks old, were inoculated subcutaneously in the right flank with 5610 6 HCT116 cells suspended in 50 ml PBS. When tumor volume reached ,150 mm 3 , mice were randomly assigned to the Oridonin treatment group (n58) or the control (DMSO) group (n58). Oridonin (7.5 mg/kg) was administered daily by intraperitoneal injection. Tumor volume was determined using digital vernier caliper measurements and the formula: A6B 2 60.52, where A is the longest diameter of the tumor and B is the shortest diameter of the tumor.

Spontaneous Metastasis Model and Hematoxylin & Eosin Staining
The metastasis model was as previously reported [28], orthotopic injections were performed under anesthesia by using sodium pentobarbital. An incision was made through the abdominal muscles between 4 th and 5 th papilla to expose the mammary fat pat (MFP). 4T1 cells (1610 5 ) in 50 ml PBS were injected into the MFP. Based on the primary tumor size, mice were segregated into groups for the appropriate treatments. On the day 7, Oridonin was injected intraperitoneally every two days. Lungs were dissected and fixed in 10% formaldehyde when the mice were agonal. Paraffin-embedded lungs were cut for H & E staining. Slides were examined under a light microscopy and pictures were taken at 2006 magnification.

Immunohistochemistry
Paraffin-embedded sections of tumor tissues and lungs were used for immunohistochemistry staining performed as previously described [29,30]. Immunohistochemistry analysis was performed on 5-mm sections of formalinfixed paraffin-embedded colon tumors derived from mice injected with HCT116 cancer cells. Then 50 serial sections were cut from each tissue, and every 10th tumor section was stained for expression of the endothelial cell-specific marker CD31 and Von Willebrand Factor (vWF) by using an anti-CD31 (Lifespan Biosciences, Inc., Seattle, WA, USA) and anti-vWF polyclonal antibody (Millipore Corporation, Temecula, California, USA). Lung sections were stained with anti-CD31 and anti-Pericentrin (Abcam plc. Temecula, California, USA) polyclonal antibodies. Images were recorded under the Leica DM 4000B photomicroscope and integrated optical density (mean IOD) of blood vessels in tumor sections was analyzed using Image-Pro Plus 6.0 software. Angiogenetic vessels consisting of at least two visible cells in a crosssectional view were scored as positive.

Immunofluorescence
After treating with Oridonin, HUVECs grown on gelatin-coated glass coverslips were fixed in 4% (w/v) paraformaldehyde and permeabilized with 0.1% Triton-X 100 in PBS. F-actin was visualized by sequential incubation with indicated antibody overnight at 4˚C (Jackson Immunoresearch Laboratories, West Grove, PA, U.S.A.). Pictures were acquired using a confocal microscope (Leica Instruments, IN, U.S.A.) controlled by the LAS AF advanced 1.9.0 software (Diagnostic Instruments).

Transendothelial Migration of Tumor Cells
Tumor cell transendothelial migration was performed as previously described [2]. Approximately 2610 5 HUVECs were added to matrigel-coated 24-well Transwell inserts and grown to confluence for 48 hours, with daily replacement of fresh culture medium. Then, 1610 5 4T1-GFP cells with different concentration of Oridonin were added onto the HUVEC monolayer and incubated for 18-20 hours. Then the non-migrated cells were removed using a cotton swab. To assay migration, the migrated cells were photographed and counted manually.

Real-time Quantitative Polymerase Chain Reaction
HUVECs were starved for 12 hours and exposed to various concentrations of Oridonin for 12 hours and then induced with VEGF (100 ng/ml) for 10 minutes. Total RNA was isolated from treated HUVECs with Trizol (Invintrogen). 1 mg of total RNA was high performance reverse transcribed into cDNA in a volume of 20 ml with reverse transcriptase and oligo dT primers (Promega, Madison, USA) according to the manufacturer's manual. The cDNA of b-actin served as the endogenous control. A quantitative fluorogenic SYBR Green (Takara, Otsu, Shiga, Japan) used for real-time quantification in an ABI prism 7000 sequence detection system. Primer sequences for Jagged 1, Jagged 2, Dll-1, Notch1, Hes-1 and HESR1 are summarized in the Table 1 [31]. PCR cycles were 5 minutes at 95˚C, followed by 40 cycles with an annealing temperature of 56˚C.

Statistical Analysis
Statistical analysis was performed with Microsoft Excel software. The values in vitro experiments were expressed as means ¡SE. The Student t-test for independent analysis was applied to evaluate the difference between treatment and control groups. All experiments were repeated at least three times except animal experiment. A value of p,0.05 was considered statistically significant.

Oridonin Inhibited Endothelia Cell Proliferation
To detect the effect of Oridonin on angiogenesis in vitro, we first examined the potential anti-proliferation activity of Oridonin on HUVECs with MTS assay. Result showed that Oridonin significantly inhibits VEGF induced HUVEC proliferation in a dose-dependent manner, with the IC50 at about 2.5 mM (Fig. 1A right panel).

Oridonin Inhibited Tube Formation and Migration of HUVECs in vitro
As tube formation is a critical step in angiogenesis, we investigated whether Oridonin affected HUVEC angiogenic activity in vitro with tube formation assay. We found that tubular structures on Matrigel decreased over 90% after treating with Oridonin (Fig. 1B). Then we examined the effects of Oridonin on migration of HUVECs in vitro. When incubated with VEGF and Oridonin, HUVEC wound healing and transwell migration were suppressed in a dose-dependent manner with the IC50 at 2.5 mM (Fig. 1C and D). Therefore, Oridonin inhibits endothelial cell proliferation, migration, and tube formation, three activities essential for angiogenesis.

Oridonin Suppressed Microvessel Formation in the Rat Aortic Ring ex vivo
In the rat artic ring assay, the effect of Oridonin on the sprouting of vessels from aortic rings upon VEGF (20 ng/mL) stimulation was examined. After incubation with 0.1% DMSO and Oridonin for seven days, there were significantly fewer neomicrovessels from aortic rings in the Oridonin-treated group than those in control group. 5 mM Oridonin effectively block VEGF-induced microvessel sprouting, with a reduction of more than 90% (Fig. 2A).

Oridonin Inhibited VEGF-induced Angiogenesis in vivo in the Mouse Corneal Neovascularization Model
To further investigate whether Oridonin inhibited VEGF-induced angiogenesis in vivo, Oridonin was tested in the mouse cornea model. We found that Oridonin markedly inhibited VEGF-induced neovascularization (Fig. 2B). Quantification analysis revealed the significant anti-neovascularization effect of Oridonin, with an 88% reduction of vascularization area (N58, P,0.01) (Fig. 2C right panel). At the same time, clock number decreased by 66.7% and vessel length by 60% upon treatment with Oridonin ( Fig. 2C left panel and middle panel). These findings demonstrated that Oridonin was a potent inhibitor of angiogenesis in vivo.

Oridonin Suppressed Tumor Growth and Angiogenesis in Solid Tumor
Our results above show significant effects of Oridonin on in vitro endothelial cell parameters relevant to angiogenesis (Figs. 1 and 2) as well as in vivo inhibition of corneal vascularization ( Fig. 2B and C). However, the tumor environment is complex and characterized by down-regulation of multiple signal transduction pathways that may impede the efficacy of compounds that function in more normal settings. Therefore, to investigate the value of Oridonin as an antiangiogenic therapeutic compound in tumors, we evaluated the ability of Oridonin to block tumor growth and angiogenesis in the subcutaneously implanted HCT116 colorectal carcinoma tumor model [32]. When tumors became palpable, mice were treated with either Oridonin (7.5 mg/kg per day) or DMSO for 15 days. We found that tumor size (Fig. 3A, right panel and Fig. 3B-3C, left panels) was significantly reduced in the Oridonin treated group. The mean tumor weight of the Oridonin -treated group was also much smaller that of the control group ( Fig. 3B right panel). At the same time, there was no significant difference between control and Oridonin-treated mouse body weight (Fig. 3C, right panel), implying the potential for few side-effects of Oridonin at the therapeutic dosage. Tumor sections stained with anti-CD31 and anti-vWF antibodies revealed that Oridonin inhibited new blood vessels as well as pruned preexisting tumor vessels (Fig. 3D). To test the effect of Oridonin on HCT116 tumor cells, additional analysis with MTS proliferation assay was performed, and results showed that Oridonin inhibited HCT116 tumor cells proliferayion. However the IC50 of Oridonin on HCT116 was 10-20 mM (Fig. 3A left panel), which was more than that on HUVECs (IC50 ,5 mM) (Fig. 1A right panel), suggesting that Oridonin was more effective in inhibiting proliferation in HUVECs than in HCT116 cells at the same concentration. These results suggested that Oridonin reduced tumor growth mainly via inhibiting tumor angiogenesis.

Oridonin Block Tumor Cell Extravasation across the Endothelium
It is well known that malignant tumors depend on neovascularization for their metastasis [33]. To test whether Oridonin affects tumor metastasis, we performed tumor metastasis assays, injecting 4610 5 4T1 breast cancer cells into the mammary fat pad of BALB/c mice. In all injected BALB/c mice, the primary tumor burden was high, with no significant difference between treated and control mouse tumor size, 754 mm 3 and 793 mm 3 , respectively at the 23 rd day (data not shown). To our surprise, metastatic node formation was inhibited in the Oridonin group (51 nodes in average) compared to the control group (72 nodes in average); a representative lung from each experimental group was shown in Fig. 4A. In additional analysis, normal tissue and metastatic nodules in the lung were obvious after H &E staining, as being sparse or compact, respectively (Fig. 4B). Metastases comprised 26% of lung area in control mice, but was significantly reduced in the Oridonin group (12%, n56, p,0.01). Typical areas of H&E stained lungs in control and Oridonin-treated groups are shown in Fig. 4C. Individual tumor cells in the Oridonin-treated mice were found inside the blood vessels (data not shown). To confirm that these cells were tumor cells, we performed immunofluorescence staining for pericentrin, a centrosomal protein used to detect neoplastic cells [34]. Breast tumor cells in control lungs with invasive metastases show a high level of centrosomal abnormalities by immunofluorescence staining for pericentrin (S1 Figure and Fig. 4D). In Oridonin-treated mice, tumor cells (green) with centrosomal abnormalities were only located inside CD31-stained blood vessels (red) (Fig. 4D). No tumor cells were found outside the vessels. These results may imply that Oridonin inhibits metastasis formation in vivo by inhibiting the migration/invasion of tumor cells through the endothelium.

Oridonin Regulated HUVEC Cell-cell contacts and Inhibited Tumor Cell Transendothelial Migration In Vitro
Cell-cell contact is indispensable during cell motility [35] and actin fibers underlie early cell-cell contacts [36]. To investigate the mechanism of Oridonin inhibition of EC migration and tumor cell TEM, we first performed F-Actin and detected the expression of VE-cadherin. VE-cadherin is crucial for the maintenance and control of endothelial cell-cell contacts [37]. Results showed that the morphology and cell junction of HUVECs were changed after treating with Oridonin. There are more cell-cell connections with Oridonin ( Fig. 5A and B). Previous studies showed that Notch signal pathway contributes to cell-cell contact [38,39]. Therefore, we treated HUVECs with the Notch inhibitor DAPT and found that the effects of HUVEC treatment with 10 mM DAPT was similar to that of Oridonin ( Fig. 5A and B). These results suggested that the molecular mechanism of Oridonin biological effects in HUVECs may involve Notch pathway signaling. Tumor cell transendothelial migration is a crucial step in both the intravasation of cells from primary tumors into the vasculature and in extravasation of circulating tumor cells into suitable secondary sites leading to tumor metastatic colonization [40]. To further study whether Oridonin affects TEM, we measured the transendothelial migration of 4T1 breast cancer cells. In order to exclude the confounding effects of differing HUVEC densities, the monolayer in each  chamber was confirmed to be confluent by visual inspection for each experiment. As showed in Fig. 5C and D, 4T1 tumor cell transendothelial migration could be enhanced by FBS or VEGF and Oridonin decreased VEGF-induced migration more effectively than FBS-induced trans-HUVEC migration, with an IC50 of less than 1 mM. These data showed that Oridonin inhibits an important step in breast cancer cell transendothelial migration in vivo.

Oridonin Blocked the Jagged-Notch Signaling Pathway in Endothelial Cells
One published signaling cascade mediating VEGF signaling is the Notch pathway. We were interested to determine whether Oridonin could affect Notch signaling in endothelial cells. First, we verified inhibition of Notch signaling in HUVECs using a known Notch inhibitor, DAPT. To examine the effects and concentrationdependence of DAPT on Notch signaling in HUVECs, we performed Q-PCR and western blot assays. 20 mM DAPT suppressed the mRNA expression of Dll-1, Jagged2, Notch1 and Hes-1, which are involved in the Notch pathway ( Fig. 6A left panel). At the same time, DAPT inhibited VEGF-induced Notch activity (Fig. 6A right panel). Thereafter we used 20 mM DAPT as the positive control in the gsubsequent assays to measure Oridonin effects on Notch signaling. As expected, Oridonin remarkably inhibited key ligands, receptor and downstream genes and protein expression involved in Jagged-Notch Signaling Pathway ( Fig. 6B and C). In particular, note the inhibition of both Jagged2 ligand and VEGF-stimulated Notch activation at 10 mM Oridonin, suggesting potency of VEGF/Notch inhibition by Oridonin at concentrations relevant to effects on endothelial cell functions in vitro. Therefore, Oridonin inhibition of angiogenesis may occur through blockage of VEGF stimulation of Jagged2 and Notch activation.

Discussion
Angiogenesis is required for tumor growth and metastasis [25]. Identification of novel angiogenesis inhibitor will benefit drug discovery for treating angiogenic diseases, such as tumor. Here, we found that Oridonin not only inhibited angiogenesis in the mouse cornea assay, the golden standard for angiogenesis, but also suppressed rat artery ring sprouting. In addition, Oridonin inhibited tube formation which was consistent with a previous result that Oridonin could inhibit capillary-like networks formation of human dermal microvascular endothelial cells [23].
Over 90% of deaths from solid tumors are attributable to tumor metastasis [41]. Cancer metastasis is a multistep process involving many types of cell-cell interactions. In the tumor metastatic cascade, tumor cells initially attach to vascular endothelium and intravasate, followed by circulation and extravasation from the blood stream. To form metastases, migrating tumor cells must overcome the endothelial barrier [42] which is accompanied by drastic changes in cytoskeletal organization in neighboring endothelial cells [43]. Therefore, transendothelial migration of tumor cells is a crucial step of metastasis and inhibitors of transendothelial migration may be an attractive way of blocking metastasis. Here, we found that Oridonin promoted HUVECs cell-cell connections (Fig. 5A), which may have prevented tumor cells from crossing the endothelial cell layer. At the same time, data presented here provided evidence that Oridonin more significantly inhibited VEGF-induced transendothelial migration than that of FBS-induced in breast cancer cell (Fig. 5B), implying that Oridonin has potential specificity for a downstream target of VEGF signaling. These results, together with the effects of tumor metastasis and growth inhibition, suggested that inhibition of transendothelial migration of tumor cells is an important aspect of blocking breast tumor metastasis by Oridonin.
Among the molecular mechanisms involved in tumor angiogenesis and metastasis, accumulating evidence suggests that Notch is a key regulator of tumor angiogenesis and metastasis [6,44,45]. Disruption of Notch has been implicated in multiple tumor types [46]. In breast cancer, evidence from in vitro experiments, mouse models and human tumor samples indicated that Notch plays a predominantly oncogenic role [46], suggesting the importance of Notch signaling pathways during tumor development. VEGF and Notch are interacting signaling pathways in tumor angiogenesis [47]. It is well established that the VEGF acts as a potent activating stimulus for angiogenesis, whereas Notch appears to help guide cell fate decisions that appropriately shape this activation. These reports suggest that VEGF and Notch may be valid therapeutic targets in cancer [48]. In the present study, we demonstrated that Oridonin inhibited the expression of Jagged2, Notch, and their downstream genes (Fig. 6B), similar to the effect of the Notch inhibitor DAPT (Fig. 6A). Furthermore, we found that Oridonin downregulated the activity of Notch (Fig. 6C). These results suggested that the above biological functions of Oridonin in inhibiting tumor angiogenesis and metastasis was accomplished through blocking the Jagged-Notch pathway.
In conclusion, our results indicate that Oridonin inhibited tumor growth and metastasis via suppressing tumor angiogenesis by blocking the VEGF induced Jagged-Notch signal pathway. Oridonin may be an effective anti-tumor candidate therapeutic by regulating Jagged-Notch activity in cancer therapy.