Bmi-1 Promotes Glioma Angiogenesis by Activating NF-κB Signaling

Angiogenesis in glioma is associated with the poor prognosis of the disease and closely correlates with the highly invasive phenotype of glioma cells, which represents the most challenging impediment against the currently glioma treatments. Bmi-1, an onco-protein, has been implicated in the progression of various human cancers, including gliomas, whereas its role in glioma angiogenesis remains unclear. Our current study examined the effects of Bmi-1 on glioma angiogenesis in vitro as well as in vivo. We found that overexpression of Bmi-1 enhanced, whereas knockdown of Bmi-1 diminished, the capability of glioma cells to induce tubule formation and migration of endothelial cells and neovascularization in chicken chorioallantoic membrane. In vivo, Bmi-1 overexpression and knockdown, respectively, promoted and inhibited angiogenesis in orthotopically transplanted human gliomas. Furthermore, NF-κB activity and VEGF-C expression was induced by Bmi-1 overexpression, whereas Bmi-1 knockdown attenuated NF-κB signaling and decreased VEGF-C expression. Additionally suppression of NF-κB activity using a specific chemical inhibitor abrogated the NF-κB activation and the pro-angiogenic activities of glioma cells. Together, our data suggest that Bmi-1 plays an important role in glioma angiogenesis and therefore could represent a potential target for anti-angiogenic therapy against the disease.


Introduction
Angiogenesis, the formation of new blood vessels from preexisting ones, occurs in many physiological and pathological conditions such as development, wound healing and cancer [1]. In the past decades, tumor angiogenesis has become an important field of research in cancer biology and clinical oncology, as neovascularization is required for the growth and metastasis of various types of human cancers including gliomas [2]. Tumor cells produce angiogenesis inducers, represented by vascular epithelial growth factors (VEGF), which play a crucial role in endothelial survival, proliferation and differentiation, as well as new vessel sprouting [3]. Endothelial cells of environmental capillaries activated by tumor cells-released proangiogenic factors are commonly subjected to migration and mitotic division, consequently forming a network of new vessels connected with the existing vasculature [4]. Its essential role in cancer development and progression makes angiogenesis, as well as molecular promoters underlying this important process, a promising target for cancer therapy. Glioma, the most common type of primary brain tumors, represents one of the most aggressive and lethal human cancer types [5]. The prognosis of gliomas has remained poor, and the cumulative 1-year survival rate is below 30% [6,7]. Of note, malignant gliomas, such as glioblastoma multiforme (GBM), are characterized by highly aggressive phenotypes, including prominent vascularization underlied by upregulation of pro-angiogenic factors and endothelial proliferation [8]. In this context, the poor patient survival is largely attributable to the high invasiveness and accelerated growth of the tumor [9]. It has been reported that invasion and proliferation of glioma cells are angiogenesisdependent, and that gliomas are among the cancer types with the highest degree of vascularization, which also represents an independent prognostic factor for the disease [9][10][11]. Increased tumor microvascular density (MVD) has been identified as a hallmark in gliomas with a shorter patient survival [12,13]. Understanding the molecular mechanisms mediating the development and progression of glioma angiogenesis will help identify diagnostic/prognostic biomarkers as well as therapeutic targets against the deadly condition.
B-cell-specific Moloney murine leukemia virus integration site 1 (Bmi-1) is a member of the polycomb gene family and has been found upregulated in a variety of human cancer types including acute myeloid leukemia and breast, colon, lung, ovarian and nasopharyngeal cancers, suggesting a potential role of Bmi-1 as an oncogene [14][15][16][17][18][19]. Dysregulated expression of Bmi-1 has been found in various biological processes associated with cancer development and progression, including cell proliferation, invasion and repression of apoptosis or senescence [18,[20][21][22][23]. Previous studies have revealed that Bmi-1 interacts with a number of cancer-related signaling pathways, such as activating NF-kB signaling, suppressing the p16/Rb and/or p19 ARF /MDM2/p53 tumor suppressive pathways, and inducing the activation of Hedgehog signaling [24][25][26][27]. It has been reported that Bmi-1 induces epithelial-mesenchymal transition (EMT) in nasopharyngeal carcinoma cells through stabilizing Snail, a transcriptional repressor associated with EMT, via modulation of the PI-3K/Akt/ GSK-3b pathway [14]. Furthermore, Bmi-1 suppresses transcription of the tumor suppressor phosphatase and tensin homolog deleted on chromosome ten (PTEN) via directly associating with the PTEN gene locus [14]. Moreover, Bmi-1 suppresses p16 INK4a and p14 ARF , and promotes cell proliferation by inhibiting the p16/ Rb and/or p14 ARF /MDM2/p53 signaling pathways [27]. Our previous study has also demonstrated that Bmi-1 is overexpressed, correlates with poor overall survival of patients with gliomas and plays an important role in growth and survival of glioma tumor cells [22]. Whether Bmi-1 has a biological function in glioma angiogenesis however, remains unclear.
In the present study, we report that Bmi-1 promotes glioma angiogenesis in vitro as well as in vivo. We find that Bmi-1 activates the NF-kB signaling pathway and induces expression of NF-kB target genes, including VEGF-C that plays a pivotal role in tumor vascularization. These results reveal a mechanism by which Bmi-1 enables glioma neovascularization via activation of NF-kB signaling.

Ethics statement
This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. All animal studies were conducted with the approval of the Sun Yatsen University Institutional Animal Care and Use Committee. All surgery was performed under sodium pentobarbital anesthesia, and all efforts were made to minimize suffering.

Real-time RT-PCR and data analysis
Real-time RT-PCR and data analysis were performed as previously described [28]. Briefly, total RNA was extracted from cells by using the Trizol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's instruction. Two micrograms of RNA from each sample was used for cDNA synthesis primed with random hexamers. Real-time RT-PCR was performed using the Applied Biosystems 7500 Sequence Detection system. PCR amplification of cDNA was performed with a initial denaturation step at 95uC for 10 minutes, followed by 28 cycles of denaturation at 95uC for 60 seconds, primer annealing at 58uC for 30 seconds, and primer extension at 72uC for 30 seconds, and completed with a final extension at 72uC for 5 minutes. The expression of various genes were defined based on the threshold cycle (Ct), and relative expression levels were calculated as 2 2[(C Western Blotting (WB) analysis WB analysis was performed according to a standard method as previously described [22]. The blot membrane was probed with a 1:500-diluted rabbit anti-Bmi-1 antibody (Abcam, Cambrige, MA, USA). The membranes were stripped and re-probed with mouse monoclonal anti-b-actin antibody (1:1,000; Sigma, Saint Louis, MO, USA) as a loading control.

HUVEC tubule formation assay
HUVEC tubule formation and migration assay was performed as previously described [29]. Briefly, 200 ml growth-factor-reduced Matrigel (Collaborative Biomedical Products) was added into wells of a 24-well plate and polymerized for 30 min at 37uC. HUVEC (2610 4 ) in 200 ml of conditioned medium were added to each well and incubated at 37uC in 5% CO 2 for 20 h. Images were captured under bright-fields using a Zeiss Axio Imager A1 microscope with 1006 final magnification. The induced tube-like structures under each condition in triplicate were quantified by determining their length.

Chicken chorioallantoic membrane (CAM) assay
To evaluate the direct effect on angiogenesis, CAM assay was performed at the 8th day of development of fertilized chicken eggs as previously described [30]. A 1-cm diameter window was opened in the shell of each egg with 8-day-old chicken embryo (Yueqin Breeding Co. Ltd, Guangdong, China). The surface of the dermic sheet on the floor of the air sac was removed to expose the CAM. A 0.5-cm diameter filter paper was first placed on top the CAM, and 100 ml conditioned medium harvested from transduced or treated glioma cells was added on the center of the paper. After the window was closed with sterile adhesive tape, the eggs were incubated at 37uC under 80-90% relative humidity for 48 hours. Following fixation with stationary solution (methanol: acetone = 1:1) for 15 min, the CAM was cut and harvested, and gross photos of each CAM were taken under a digital camera (Panasonic, Osaka, Japan). The effect of conditioned media was evaluated by the number of second-and third-order vessels in comparison with that treated with the medium harvested from the vector-control group. Statistical analysis of second-and third-order vessels was carried out using 2-tailed student's t-test.

Transwell migration assay
Various cells (1610 4 in 50 ml) to be tested were plated on the top side of polycarbonate Transwell filter and incubated at 37uC for 22 hrs, followed by removal of cells inside the upper chamber with cotton swabs. Migrated cells on the lower membrane surface were fixed in 1% paraformaldehyde, stained with hematoxylin, and counted (10 random 2006 fields per well). Cell counts were expressed as the mean number of cells per field of view. Three independent experiments were performed and the data are presented as mean 6 standard deviation (SD).

Intracranial brain tumor xenografts and IHC staining
Various cells (5610 5 in 5 ml) were stereotactically implanted into individual nude mouse brains. Brain glioma-bearing mice were euthanized 2 weeks post-implantation as previously described [29]. Brains of various mice were removed, cut into 6 mm thickness sections and stained immunohistochemically (IHC) staining using anti-CD31 (Invitrogen, Carlsbad, CA, USA) and an anti-VEGF-C antibodies (Cell signaling, Danvers, MA, USA). The images of IHC stained brain sections were captured using the AxioVision Rel.4.6 computerized image analysis system (Carl Zeiss Co. Ltd., Jena, Germany).

Luciferase assay
Fifteen thousand cells were seeded in triplicates in 24-well plates and allowed to settle for 24 h. 100 ng of luciferase reporter plasmids or the control-luciferase plasmid, plus 10 ng of pRL-TK renilla plasmid (Promega, Madison, WI, USA), were transfected into glioma cells using the Lipofectamine 2000 reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's recommendation. Luciferase and renilla signals were determined 48 h after transfection using the Dual Luciferase Reporter Assay Kit (Promega, Madison, WI, USA) according to a protocol provided by the manufacturer. Three independent experiments were performed and the data are presented as mean 6 SD.

Enzyme-linked immunosorbent assay (ELISA)
ELISA was performed according to the manufacturer's manual (Keygentec Co., Shanghai, China). Briefly, standard (diluted) and tested samples (Bmi-1 overexpressing and silencing cells, and the vector control cells), including a negative control, are added to appropriate wells of the plate. Incubated at 36uC for 90 min, and wash the unbound samples off the plate with DI water and PBS-Triton. Incubation with specific antibody (anti-VEGFC) performed at 36uC for 60 min. After washing step, add secondary antibody and incubate for 60 min. And then add the substrate and incubate at room temperature for 60 min. Stop the reaction and read plates on an ELISA plate reader. Colorimetric measurement was recorded as OD 450 readings.

Statistical analysis
Statistical analyses were performed using the SPSS 11.0 software package. Data are presented as mean 6 SEM. P values of 0.05 or less were considered statistically significant.

Ectopic expression of Bmi-1 induced angiogenesis in glioma cells in vitro
To investigate the effect of Bmi-1 on glioma angiogenesis, previously constructed pMSCV/Bmi-1-derived retrovirus [22] was used to transduce LN382T and T98G glioma cells to establish Bmi-1-overexpressing stable cell lines. Expression of Bmi-1 was verified by WB ( Figure 1A). Subsequently, the effect of Bmi-1 on enhanced ability of glioma cells in inducing in vitro angiogenesis was examined using an in vitro tube-like structure formation assay. Conditioned medium (CM) collectd from Bmi-1-overexpressing glioma cells was added to Matrigels embedded with HUVEC cells. As shown in Figure 1B, CM derived from Bmi-1-expressing glioma cells induced more tube-like structure formed by HUVEC as well as in shorter time than that by control CM. Additionally, ectopic expression of Bmi-1 by glioma cells also potently induced the formation of second-and third-order vessels in chicken chorioallantoic membranes (CAM assay) ( Figure 1C). Additionally, HUVEC migration was markedly induced by CM of Bmi-1overexpressing glioma cells compared with control CM ( Figure 1D). Taken together, the data suggested that Bmi-1 enhanced the capability of glioma cells to stimulate neovascularization in vitro.

Bmi-1 promoted the expression of VEGF-C in glioma cells in vitro
It has been reported that new vessel formation driven by glioma cells is mechanically associated with increase expression of VEGF-C [29]. Toward an understanding of the mechanism by which Bmi-1 promotes the angiogenesis of glioma cells, we investigated whether Bmi-1 induces the expression VEGF-C in glioma cells. As shown in Figure 2A, VEGF-C mRNA was upregulated in Bmi-1overexpressing glioma cells when compared that in control cells. In addition, levels of secreted VEGF-C protein in the supernatants of Bmi-1-overexpressing glioma cells was also elevated ( Figure 2B). These results indicated that Bmi-1 increased expression and secretion of VEGF-C in glioma cells.

Silencing Bmi-1 reduced angiogenesis and VEGF-C expression in glioma cells
To establish an experimental model in which expression of endogenous Bmi-1 is knocked down, Bmi-1-specific RNA interference (RNAi) oligonucleotides were cloned into a retroviral transfer vector pSuper-retro-puro as described previously [22]. In this model, attenuated expression of Bmi-1 was verified by WB ( Figure 3A). Consequently, as shown in Figure 3B, CM of Bmi-1-shRNA-transfected cells showed decreased abilities to induce formation of tube-like structure by HUVEC cells in vitro when compared with the control CM, suggesting that inhibition of endogenous Bmi-1 in glioma cells markedly reduced glioma cellinduced in vivo angiogenesis. The stimulated HUVEC migration was also reduced by CM derived from Bmi-1-silencing glioma cells when compared with control CM ( Figure 3C). Consistent with these observations, shRNA depletion of Bmi-1 in glioma cells also significantly reduced VEGF-C mRNA and protein expression ( Figure 3D and E).

Bmi-1 induced angiogenesis and VEGF-C expression in vivo
To assess whether Bmi-1-induced glioma angiogenesis in vivo, we utilized an orthotopic brain glioma tumor model and determined whether Bmi-1 modulates glioma angiogenesis in the brain. As shown in Figure 4A, when compared with the controls, gliomas that stably overexpressed Bmi-1 markedly induced vascellum on the surface in the tumor-bearing brain whereas surface vascellum was significantly decreased in brains bearing Bmi-1-knockdown gliomas. Furthermore, IHC analyses demonstrated markedly elevated expression of CD31, enhanced microvascular outgrowth and increased MVD in Bmi-1-overexpressing gliomas, as opposed to those Bmi-1-knockdown gliomas (P,0.05, Figure 4A). Additionally, when compared with control gliomas, Bmi-1 overexpression up-regulated VEGF-C levels whereas knockdown of Bmi-1 resulted in decreased VEGF-C expression in brain gliomas ( Figure 4B). Taken together, our data demonstrated that Bmi-1 induced tumor angiogenesis and VEGF-C expression in brain gliomas.

Bmi-1 promoted the pro-angiogenic potency of glioma cells via the NF-kB pathway
We have previously shown that Bmi-1 induces NF-kB activation in glioma cells [22]. To determine whether NF-kB signaling mediates Bmi-1-promoted glioma angiogenesis, we first examined whether Bmi-1 expression enhances the transcriptional activity of NF-kB in LN382T and T98G glioma cells. As shown in Figure 5A, Bmi-1 overexpression induced transcription activity of NF-kB whereas siRNA knockdown of Bmi-1 reduced NF-kB activities. Furthermoe, the levels of several NF-kB target genes, TNF-a, IL-6, CCND1, MYC, BcL-XL and MMP-9, were upregulated in Bmi-1overexpressing, but downregulated in Bmi-1-silenced glioma cells ( Figure 5B). Additionally, compared to that in control cells, the activity of a VEGF-C promoter that contains a NF-kB binding site was significantly increased in Bmi-1-overexpressing but markedly decreased in Bmi-1-silenced glioma cells. Either a VEGF-C promoter fragment lacking the NF-kB binding site or mutation of the NF-kB binding site in the VEGF-C promoter region showed no NF-kB activity in either Bmi-1-overexpressing and -silenced  Figure 5C). Taken together, these results suggest that Bmi-1 expression modulates the transactivation activity of the VEGF-C promoter harboring the NF-kB binding site in glioma cells.
Lastly, we examined whether Bmi-1 induced angiogenesis of glioma cells via activating NF-kB signaling. As shown in Figure 6 A and B, CM derived from Bmi-1 overexpressing glioma cells stimulated tube-like structures formed by HUVEC cells and HUVEC cell migration. Conversely, Bmi-1 stimulation of in vitro HUVEC angiogenic activities was markedly impeded by a specific NF-kB inhibitor JSH-23 that attenuates the transcriptional activity of NF-kB [31]. Furthermore, JSH-23 treatment also attenuated Bmi-1 induction of VEGF-C expression in glioma cells ( Figure 6 C and D). Additionally, the HUVEC tube formation and migratory ability induced by Bmi-1 could be significantly reduced by treatment with a VEGF-C inhibitor (VEGFR3-Fc) ( Figure S1 A  and B). Collectively, the data suggest that functional NF-kB activation and the consequent VEGF-C expression are essential for the development of the pro-angiogenic phenotype of glioma cells induced by Bmi-1.

Discussion
The major finding of this study is that upregulated Bmi-1 promotes angiogenesis in human gliomas in vitro and in vivo. Our results also suggest that the pro-angiogenic activity of Bmi-1 overexpression is associated with an activation of NF-kB signaling and the consequent upregulation of VEGF-C.
Previous studies have reported that knocking down of endogenous Bmi-1 diminishes secretion of VEGF from lung adenocarcinoma cells in vitro, leading to promoted angiogenesis in lung adenocarcinoma [32]. Moreover, co-expression of Bmi-1 and RasV12 promotes hepatocellular carcinoma (HCC) formation in mice, and the induced muring tumors resemble human HCC. Additionally, Bmi-1 promotes hepatic angiogenesis in vivo through cooperating with other oncogenic genes [33]. In this study, we demonstrate for the first time that Bmi-1 also promotes tumor angiogenesis in malignant gliomas. Our findings that Bmi-1 modulates the NF-kB/VEGF-C signaling pathway to mediate an angiogenesis phenotype in human gliomas further suggests that Bmi-1 may represent a prospective therapeutic target for antiangiogenic strategies against the disease.  It is established that poor survival of patients with glioma can be largely attributed to the highly invasive nature of tumor, and the invasive progression is closely attributed to tumor angiogenesis [2]. Moreover, both invasion and angiogenesis result in destruction of surrounding brain tissue, and these characteristics of glioma cells correlates with patients prognosis [34][35][36][37]. While many studies have been conducted to investigate the oncogenesis and progression of glioma, the molecular mechanisms involved in regulating glioma angiogenesis remain incompletely understood. VEGF-C, a member of the VEGF family, plays an important role in angiogenesis, lymphangiogenesis and endothelial cell growth and survival. VEGF-C has been found to be upregulated in gliomas and associated with the progression and prognosis of the disease [38]. Of note, VEGF-C is also a direct target regulation of NF-kB signaling, and VEGF-C stimulates neovascularization through binding and activating VEGFR-2-and VEGFR-3-mediated cellular signaling [39]. Thus, our finding that Bmi-1 upregulates VEGF-C expression in glioma cells through NF-kB activation further underscores the importance of VEGF-C in glioma angiogenesis and provides a basis for using VEGF-C as a surveillance biomarker during anti-Bmi-1 therapy.
On the other hand, MMP-9 and IL-8, both of which are downstream target genes of NF-kB, have also been found to be upregulated in glioma cells [40,41]. Furthermore, MMP-9 and IL-8 expression are further increased in Bmi-1-overexpressing glioma cells ( Figure 5, and Figure S2). It has been reported that expression of MMP-9 closely correlates with tumor angiogenesis [42]. MMP-9 has also been shown to trigger an angiogenic switch during tumor progression by releasing matrix-bound VEGF, making the latter available for interaction with VEGF receptors [43]. In addition, IL-8 is expressed and secreted at high levels in human gliomas and is critical to gliomas neovascularity and progression Bmi-1 Promotes Glioma Angiogenesis PLOS ONE | www.plosone.org [44]. Inducing IL-8 expression in glioma cells could promote glioma growth and angiogenesis through NF-kB signaling or other inflammatory stimuli [41]. Although the mechanism via which MMP-9 and IL-8 induce inflammatory stimuli glioma angiogenesis remains to be better understood, it would be of great interest to further evaluate whether and how in gliomas they cooperates with VEGF-C to promote neovascularization, due to the fact that these key angiogenic factors are all subject to the modulation of one upstream regulator, Bmi-1.
In conclusion, we demonstrate in this study that Bmi-1 induces glioma angiogenesis both in vitro and in vivo, and upregulation of Bmi-1 leads to increased VEGF-C expression through activating NF-kB signaling. Conversely, inhibition of NF-kB activity markedly reduces Bmi-1 stimulated angiogenic activities and attenuates Bmi-1 upregulation of VEGF-C. Our data provides new insights to the development of novel strategies to inhibit tumor angiogenesis in glioma by targeting Bmi-1, suggesting a necessity of further investigations of the anti-glioma therapeutic value of Bmi-1.