Application of luteolin nanomicelles anti-glioma effect with improvement in vitro and in vivo

Glioblastoma multiforme (GBM) is one of the most common and malignant tumor. Luteolin, a polyphenolic compound, has been proposed to have anti-tumor activity against various cancers. However, the greatest obstacle in the administration of luteolin is its hydrophobicity as well as the low oral bioavailability. In this study, we formulated luteolin-loaded MPEG-PCL (Luteolin/MPEG-PCL) micelles aiming to improve its solubility in aqueous solution and investigate the anti-tumor effect on glioma in vitro and in vivo. The spherical Luteolin/MPEG-PCL micelles were completely dispersible in normal saline and could release luteolin in a sustained manner in vitro. We demonstrated that Luteolin/MPEG-PCL micelles had stronger cytotoxicity and induced a higher percentage of apoptosis in C6 and U87 cells than free luteolin in vitro. The immunohistochemical study revealed that Luteolin/MPEG-PCL micelles induced more glioma cell apoptosis than free luteolin and inhibited neovascularization in tumor tissues. The Pro-caspase9 and Bcl-2 down-regulation and cleaved-caspase9 and Bax up-regulation suggested that luteolin induced apoptosis via the mitochondrial pathway in vitro. What is more, we found the drug could cumulated much more in the nano-drug group than free drug group through imaging in vivo. In conclusion, the Luteolin/MPEG-PCL micelles have the potential clinical application in glioma chemotherapy.


INTRODUCTION
Glioma is the most common primary malignant brain tumor in children and adults, and has been notoriously known to be a great challenge for clinical treatment. Glioblastoma (GBM), which is classified as Grade IV by the World Health Organization (WHO), is the most aggressive and deadly subtype and accounts for more than 50 percent of glioma [1]. With the development of neuroimaging, GMB can be early diagnosed and more details could be obtained from MRI. However, treated with comprehensive therapy which consists of surgical resection followed by radiotherapy and chemotherapy, the median survival time of patients with GBM is only 15-17 months [2][3][4].
GMB features a high proliferation activity and a diffuse nature having a tendency to invade surrounding brain tissues. Due to several factors, the effect of conventional chemotherapy is unsatisfactory. First, the effective accumulation of chemotherapy agents in the central nervous system is limited because of the blockage of blood-brain barrier (BBB). Secondly, recent research showed that glioma cells over-express the multi-drug resistance (MDR) transporters which lead to the extruding of agents resulting in the phenomenon of chemoresistance [5]. Moreover, glioma cells are able to survive in extreme conditions such as glucose deprivation or stress by www.impactjournals.com/oncotarget/ Oncotarget, 2017, Vol. 8, (No. 37), pp: 61146-61162 Research Paper www.impactjournals.com/oncotarget autophagy or modulating multiple signaling pathways [6][7][8][9]. Hence, more effective chemotherapy drugs are sorely in need.
Luteolin is an important natural polyphenolic compound, which mainly derived from diverse fruits and vegetables, including pomegranate, cabbage, celery, spinach and so forth. It has been widely concerned and researched extensively due to its diverse pharmacological effects on anti-virus, anti-inflammatory and anti-tumor [10][11][12][13][14]. Additionally, with the dietary supplement of luteolin, the diet-induced insulin resistance in mice was obviously ameliorated [15,16]. Studies have confirmed its effectiveness in suppressing a wide variety of human tumor, including colorectal cancer, cervical cancer, lung cancer, skin cancer, etc [12,14,[17][18][19]. Compared with conventional chemotherapy agents, luteolin possesses a prominent advantage of low toxicity. It has been reported that luteolin selectively acts on tumor cells and normal cells are not affected [20]. Another strong evidence demonstrating the safety of the agent is the widespread use of luteolin in diet by millions of people in many countries. However, the clinical application of luteolin is hindered by its intrinsic poor water-solubility as well as the low oral bioavailability.

Figure 1: Preparation and identification of Luteolin/MPEG-PCL micelles. (A) Preparation of Luteolin/MPEG-PCL micelles.
Luteolin/MPEG-PCL micelles were prepared by self-assembly methods. First, luteolin and MPEG-PCL were co-dissolved in acetone. Then the mixture was added into water, and in this process luteolin and MPEG-PCL assembled into micelles with luteolin and PCL inside and PEG outside. The utilization of nanomaterials in cancer treatment is of intriguing interest. Different nanomaterials such as carbon, metal, or lipid-based nanoparticles are of diverse sizes, structures, physicochemical properties, surface properties and biocompatibilities, so as to meet specific requirements.
Various nanomaterials have been formulated to deliver chemotherapy agents to tumor sites due to highly hydrophobic nature of these agents. We have previously reported the efficiently delivery of polyphenolic compound by the biodegradable monomethyl poly(ethylene glycol)poly(ε-caprolactone) (MPEG-PCL) micelles in cancer therapy [21]. After encapsulation, the hydrophobic core part of MPEG-PCL micelles functions as a container of agents while the hydrophilic shell part greatly increases its solubility in aqueous media. In this research, luteolin loaded MPEG-PCL (Luteolin/MPEG-PCL) micelles were prepared by a self-assembly method. The morphology, zeta potential, encapsulation efficiency and pharmacokinetics of Luteolin/MPEG-PCL micelles were investigated. In the next step, the anti-tumor activity of Luteolin/MPEG-PCL micelles was evaluated in vitro and vivo (nude mice and zebrafish).

Preparation and characterization of Luteolin/ MPEG-PCL micelles
Luteolin/MPEG-PCL micelles were synthesized by a single-step precipitation method, as previously reported [21]. Luteolin/MPEG-PCL micelles contains a ball-shaped hydrophilic PEG shell and a hydrophobic PCL core where luteolin was encapsulated into the core part of micelles and surrounded with PEG to improve its water solubility ( Figure 1A).
The DL and EE of Luteolin/MPEG-PCL micelles were 6% and 98.5%, respectively. As shown in Figure  1B, the particle size of Luteolin/MPEG-PCL micelles measured by DLS was 36.5 nm (PDI=0.12). Moreover, the size was confirmed by TEM ( Figure 1C), which was consistent with the DLS results. Since Luteolin/MPEG-PCL micelles measured by DLS was in aqueous phase and structure of these amphiphilic particles was usually loose in solution, the DLS size was always larger compared to TEM size. The surface of Luteolin/MPEG-PCL micelles was negatively charged, with the zeta potential of -3.2 mv.
The in vitro release profile of free luteolin and luteolin from Luteolin/MPEG-PCL micelles was examined in PBS (PH7.4) at 37 °C. As shown in Figure  1D, free luteolin exhibited a rapid release and the accumulated release reached a peak of 79.2%±1.9% in 12h. In comparison, about 46% luteolin was released from nano-micelles in the first 10h and another 34% of drug was released sustainedly over 70h. The cumulative curve indicated that the drug-release process of Luteolin/MPEG-PCL micelles was stable and sustained in vitro (p < 0.01, Luteolin/MPEG-PCL micelles versus free luteolin).
Concentration-time profiles for free luteolin and Luteolin/MPEG-PCL micelles after intravenous injection are shown in Figure 1E. The T max , T 1/2 , C max and AUC of free luteolin were 15 min, 0.7 h, 27.32 μg/ml and 14.76 mg L −1 h −1 respectively, while the T max , T 1/2 , C max and AUC of Luteolin/MPEG-PCL micelles were 15 min, 0.9 h, 63.27 μg/ml and 106.7 mg L −1 h −1 (p < 0.01, Luteolin/MPEG-PCL micelles versus free luteolin) respectively suggesting that MPEG-PCL micelles could improve the bioavailability and pharmacokinetics of luteolin in vivo.

Apoptosis induced by luteolin
As apoptosis is an important mechanism of anticancer chemotherapy, we detected cell apoptosis of U87 cells induced by free luteolin and Luteolin/MPEG-PCL micelles with annexin V-FITC and PI apoptosis assay. The ability of free luteolin and Luteolin/MPEG-PCL micelles to induce apoptosis of U87 cells is shown in Figure 4. In the NS and Blank micelle group, the apoptosis rate was 5.02% and 4.83%, respectively. In the free luteolin and Luteolin/MPEG-PCL micelle group, the apoptotic cells increased in a concentration-dependent manner. The apoptosis rates were 11.2% and 14.02%, 14.3% and 17.4%, 19.8 and 32.3% (p < 0.01, Luteolin/MPEG-PCL micelles versus free luteolin) when the U87 cells were treated with free luteolin and Luteolin/MPEG-PCL micelles at a concentration of 6.25 μg/ml, 12.5 μg/ml and 50 μg/ml, respectively. It is evident that Luteolin/MPEG-PCL micelle was capable of inducing more U87 cell apoptosis than other groups.

Western blotting
To further investigate the underline mechanism of the anti-cancer effect of Luteolin, gene expressions related to cell apoptosis and proliferation following treatment of free luteolin and Luteolin/MPEG-PCL micelles were examined. As demonstrated in Figure 5A, the Bcl-2 protein expression was decreased while the protein expression of Bax increased after treatment of free luteolin and Luteolin/MPEG-PCL micelles, indicating that luteolin induced apoptosis through the mitochondrial pathway. As demonstrated in Figure 5B, the levels of MAPK and p-MAPK after luteolin treatment were also examined showing obviously decreased p-MAPK after 48h treatment of Luteolin/MPEG-PCL micelles, whereas the total MAPK were unchanged.

Luteolin inhibited tumor growth in animal models
The transgenic zebrafish model and nude mice xenograft model were employed to evaluate the antiglioma effect of free luteolin and Luteolin/MPEG-PCL micelles. Three days after injection of perivitelline into the zebrafish, U87 cells formed a solid tumor (green fluorescence). Shown in Figure 6, antitumor effects were observed in FL, and LM nano-micelles groups when compared with the control and blank micelles group. As detected by confocal microscopy, the tumors in the Luteolin/MPEG-PCL micelles group were smaller than FL group. No antitumor effects was observed in blank micelles group. The average tumor volume in the LM nano-micelles group was significantly smaller than that of NS, EM and FL groups. These data indicated that MPEG-PCL could enhance the anti-glioma cancer activity of luteolin in vivo.
In mice bearing C6 subcutaneous glioma model, the blank micelles had no anti-glioma effects. The tumor volumes decreased by 40% and 78% respectively under exposure of free luteolin and Luteolin/MPEG-PCL micelles in C6 glioma model with the same interventions in the Figure 7A   of subcutaneous gliomas were shown in Figure 7C. The tumor weights in the Luteolin/MPEG-PCL micelle group were lower than that of other groups in both animal models as well in the Figure 7D. The tumor weights in the Luteolin/MPEG-PCL micelle group were lower than that of other groups in the Figure 7D (p < 0.01, Luteolin/ MPEG-PCL micelles versus free luteolin).
In mice bearing U87 subcutaneous glioma model, the tumor volumes decreased by 48.7% and 79.5% respectively under exposure of free luteolin and Luteolin/ MPEG-PCL micelles in U87 glioma model with the same interventions in the Figure 8A (p < 0.01, Luteolin/ MPEG-PCL micelles versus free luteolin). Free luteolin and Luteolin/MPEG-PCL micelles exerted significant anti-giloma effects without obviously impact of mice body weight in the Figure 8B. And the representative photographs of subcutaneous gliomas were shown in Figure 8C. The tumor weights in the Luteolin/MPEG-PCL micelle group were lower than that of other groups in both animal models as well in the Figure 8D. The tumor weights in the Luteolin/MPEG-PCL micelle group were lower than that of other groups in the Figure 8D (p < 0.01, Luteolin/MPEG-PCL micelles versus free luteolin).
The results of both animal models were consistent and the representative photographs of subcutaneous gliomas were shown in Figure 7 and Figure 8. The tumor weights in the Luteolin/MPEG-PCL micelle group were lower than that of other groups in both animal models as well. It suggested that the anti-glioma activity of luteolin was enhanced with the help of MPEG-PCL micelles delivery system.

Histological analysis by TUNEL assay, Ki67 and CD31
Tumor sections were stained with TUNEL for the investigation of cell apoptosis in four groups, which were treated with NS, blank micelles, free luteolin and Luteolin/ MPEG-PCL micelles respectively. As shown in Figure 9, no considerable positve nuclei was found in the NS and blank micelles group while substantial amount of TUNEL stains were observed in the Luteolin/MPEG-PCL micelle group. The apoptotic index was 2.3% ±1.2% in the NS group, 3% ±1.73% in the blank micelles group, 39% ±4.4% in the free luteolin group (p < 0.01, free luteolin versus NS and blank micelles) and 71% ±10.5% in the Luteolin/MPEG-PCL micelle group (p < 0.01, Luteolin/ MPEG-PCL micelles versus free luteolin, NS and blank micelle). The results indicated that luteolin encapsulated in the MPEG-PCL micelles was more capable of inducing apoptosis of U87 cells in vivo.
The U87 cell proliferation was evaluated by immunohistochemical staining. Tumor sections from four groups were stained with Ki-67. As shown in Figure 10, the positive cells which emitted red fluorescence were abundant in the NS group, blank micelle group and free luteolin group. In comparison, few positive cells were observed in the Luteolin/MPEG-PCL micelle group. The Ki-67 labelling index, which is usually related to the clinical course of tumor, is significantly lower in Luteolin/ MPEG-PCL micelle group (22% ±6.5%) than other groups (88% ±7.2% in NS group, 89.3% ±3.2% in blank micelle group, 53% ±6% in free luteolin group) (p < 0.01, Luteolin/MPEG-PCL micelles versus free luteolin, NS and blank micelle). It is evident that Luteolin/MPEG-PCL micelles effectively suppressed tumor proliferation.
Furthermore, the angiogenesis of U87 glioma was explored as well. According to Figure 11, few microvessels with red fluorescence were observed group treated with Luteolin/MPEG-PCL micelles. The number of microvessels was significantly lower in the Luteolin/ MPEG-PCL micelle group (19±2.2) compared with NS group (67.8±7.8), blank micelle group (66.4±12.2) and free luteolin group (37.2±5.16) (p < 0.01, Luteolin/MPEG-PCL micelles versus free luteolin, NS and blank micelle). We propose that the mechanism of anti-angiogenesis may play an important role in inhibiting tumor growth by Luteolin/MPEG-PCL micelles.

Imaging in vivo
Mice bearing gliomas were injected with free drug or Drug/MPEG-PCL nano-micelles through the tail vein and imaged from the dorsal side at pre-determined time points using a bio-imaging system ( Figure 12). Nontreated mice were used as the control group. The average fluorescent signal in the tumors of tumor-bearing mice injected with the drug/MPEG-PCL nano-micelles was higher than that of the tumor-bearing mice injected with free drug at the same time and same dose, which indicated that nano-micelles could enhanced more accumulation of drug in tumor.

DISCUSSION
The anti-tumor effects of luteolin in various human cancers have been documented in previous published studies. However, few researches involve in checking the effectiveness of luteolin to suppress glioblastoma, which is charaterized by high proliferation, invasion and resistant to conventional chemotherapy. Recent study showed that luteolin effectively sensitized the cancer cells to several chemotherapeutic drugs by inhibiting Nrf2 activity [12]. In addition, luteolin has been reported to attenuate oxidative stress, brain edema and neurological deficits, which were common comorbidities in glioblastoma [22,23]. Therefore, luteolin would become an ideal anti-glioma agent if it efficaciously block tumor growth or directly kill glioma cells.
Oral administration of luteolin is not reasonable due to its low bioavailability caused by the poor absorption and first pass effect. Recent research showed that the oral absorption rate of flavonoids is less than 10% and the maximum plasma concentration of luteolin was merely 3.79 μg/ml in rats which received a dose of 100 mg/kg orally [24,25]. Moreover, luteolin is prone to be eliminated in circulation. Therefore, effective plasma concentration of luteolin applied either orally or by injection is hard to achieve.
Effective chemotherapy requires a continuously high local concentration of drugs in the tumor site. The diameter of Luteolin/MPEG-PCL micelles measured by DLS in our research was 36 nm. Since the interendothelial junctions of tumor mainly vary from 40 nm to 80 nm, luteolin/MPEG-PCL micelles were able to enter the tumor region through passive diffusion with little hindrance. Together with the impaired lymph drainage of nanocarriers in tumor, the enhanced penetration and retention (EPR) effect was greatly enchanced.
Another important aspect is the high stability of drug-loaded nanomaterials. Previous researches have demonstrated that PEGylated nanomaterials were able to avoid protein absorption in circulation, which is a key feature for drug delivery [26,27]. Additionally, as the luteolin/MPEG-PCL micelles gradually degrades, encapsulated luteolin will be released slowly, which is determined by the core-shell structure of MPEG-PCL micelles. In this study, the result of drug release in vitro and pharmacokinetics in the xenograft nude mouse model suggested a more sustained release and prolonged circulation of Luteolin/MPEG-PCL compared with free luteolin. All these characteristics significantly contribute to the tumor accumulation of Luteolin/MPEG-PCL. With the help of MPEG-PCL micelles, phytochemical luteolin obtained higher bioavailability, improved tumor retention and lower off-target toxicity.
To evaluate the therapeutic efficacy of luteolin against GBM in vivo, U87 GBM transgenic zebra fish model, U87 GBM nude mice model and C6 nude mice model were employed. As shown in Figure 7 and Figure  8, Luteolin/MPEG-PCL micelles dramatically inhibited glioma growth in all animal models. Meanwhile, the Ki-67 expressions of glioma cells were detected by immunohistochemical staining. The treatment of Luteolin/ MPEG-PCL micelles led to a markedly inhibition of cell proliferation. In order to explore the mechanism, tumor sections were stained with CD31 to examine the neovascularization of glioma. The microvessel density in the Luteolin/MPEG-PCL micelles group was much lower than other control groups, indicating that antiangiogenesis  , n=5) (B). The LM treatment caused inhibition of angiogenesis in the tumors. This implies that anti-angiogenic effect may be a mechanism of inhibiting U87 cells by FL and LM in vivo, and that LM inhibited more angiogenesis than other groups. may play an important role in suppressing glioma by luteolin in vivo. Moreover, TUNEL assay was applied to evaluate cell apoptosis. In comparison with other groups, more apoptosis cells were significantly observed through electron microscope in group treated with Luteolin/ MPEG-PCL micelles. The molecular mechanism of glioma cell apoptosis induced by Luteolin/MPEG-PCL micelles was further investigated in this research.
The effect of luteolin on the proliferation of glioma cells was observed in vitro with MTT and CFSE assay. It is evident that Luteolin/MPEG-PCL micelle suppresses glioma growth well in vitro. Also, a markedly decrease of phosphorylated p44/42 MAP kinase was found in the Luteolin/MPEG-PCL micelle group, which implies the p44/42 MAP kinase pathway may be involved in the tumor growth inhibition by luteolin. The phosphorylated p44/42 MAP kinase has been proved to promote proliferation of tumor cells [28].
Since some studies showed the anti-tumor effect of luteolin may be related to apoptosis, we employed FCM to determine the level of apoptosis and confirm the expression of proteins related to apoptosis, such as bcl-2, bax, caspase-9 and so forth [29][30][31]. The apoptosis induced by Luteolin/MPEG-PCL micelles was detected by FCM in our research. Meanwhile, as shown in Figure  5. , the expression of bcl-2 decreased significantly whereas markedly increases of bax, cleaved caspase-9 after Luteolin/MPEG-PCL micelles treatment for 48 hours. It suggests that Luteolin/MPEG-PCL micelle induces glioma apoptosis by the mitochondrial pathway.
In this research, we formulated Luteolin/MPEG-PCL micelles by a one step self-assembly method. The encapsulation of luteolin significantly improved its solubility and made it dispersible in water. Moreover, this drug delivery system ameliorated the pharmacokinetics. Together with the enhanced EPR effect, Luteolin/MPEG-PCL micelles prominently inhibited glioma growth in vitro and vivo. Luteolin suppresses tumor by inducing apoptosis via the mitochondrial pathway in vitro. RAS-RAF-MEK-MAPK pathway was involved in the cell growth inhibition by luteolin. The Luteolin/MPEG-PCL micelle has advantages of high anti-glioma activity, convenient manufacture and low toxicity. It has the potential of further clinical application in glioma chemotherapy. Nano-drug accumulation inside of the tumor was significantly more than free drug. And free drug in the body was removed very soon. www.impactjournals.com/oncotarget
The MPEG (molecular weight = 2000) (Sigma-Aldrich, St. Louis, MO, USA) was dried in a one-necked flask under vacuum and magnetically stirred at 105 °C for 90 min before use.
MPEG (2000)-PCL(2000) diblock copolymer with a molecular weight of 4000 was synthesized by ring opening of ε-caprolactone, which was initiated by MPEG as descripting previously [21]. Briefly, MPEG and ε-CL were introduced into a dry glass ampoule under a nitrogen atmosphere. Sn(Oct)2 was then added into the reaction vessel under mild agitation, and the reaction system was kept at 130 °C for 6 h. The purified MPEG-PCL copolymer was kept in a desiccator before further use.

Preparation of Luteolin/MPEG-PCL micelles
Luteolin/MPEG-PCL micelles were synthesized by a single-step self-assembly method. In brief, 6 mg luteolin and 94 mg MPEG-PCL diblock copolymer were dissolved in 2 ml acetone and added into 4 ml distilled water under constant stirring for 5 minutes. The above mixture was then evaporated under reduced pressure to remove acetone. The resultant formulation was called Luteolin/MPEG-PCL micelles solution. Finally, the prepared Luteolin/MPEG-PCL micelles were lyophilized and stored at 4°C.

Characterization of Luteolin/MPEG-PCL micelles
We dissolved lyophilized Luteolin/MPEG-PCL micelles (10 mg) in methanol (0.1 ml) to determine the concentration of luteolin. The content of luteolin in the formulation was detected by high performance liquid chromatography (HPLC, Shimadzu LC-20AD, Japan). Drug loading (DL) and encapsulation efficiency (EE) of Luteolin/MPEG-PCL micelles were calculated by the following formulas (1) and (2): (1) EE = Experimental Luteolin loading Theoretical Luteolin loadin ng ×100% Dynamic light scattering (DLS, Malvern Nano-ZS 90) was used to measure the particle size and zeta potential of Luteolin/MPEG-PCL micelles at 25 °C after previous dilution of samples with distilled water. This process was performed in triplicate. Transmission electron microscopy (TEM, H-6009IV, Hitachi, Japan) was applied to determine the morphological characteristics of Luteolin/MPEG-PCL micelles. In brief, diluted suspension of Luteolin/ MPEG-PCL micelles was placed drop-wise on a copper TEM grid. Then the grid was negatively stained with phosphotungsten acid (2%, w/v) for 20 minutes and allowed to dry. Thirty particles were observed to calculate the mean particle diameter.
The Then the luteolin content in samples were quantified by HPLC system. This process was repeated in triplicate.

Pharmacokinetics
Pharmacokinetics investigation was conducted in Male Sprague-Dawley rats. The rats were randomly divided into two groups of five animals each. The two groups were injected with free luteolin (50 mg/kg) or Luteolin/MPEG-PCL micelles (50 mg/kg) respectively through the tail vein injection. At predetermined time points (predose, 15min, 30min, 1h, 2h, 4h, 8h, 12h, 24h), the blood was collected from carotid artery and serum concentration of luteolin was analyzed by HPLC. This process was repeated in triplicate.

MTT test
The effects of free luteolin and Luteolin/MPEG-PCL micelles on C6 and U87 cell viability were evaluated by MTT method. In brief, C6 and U87 cells were seeded at a density of 5×10 3 cells per well in 96-well plates and allowed to grow in DMEM added with 10% FBS for 24h. Cells were then treated with free luteolin or Luteolin/MPEG-PCL micelles at different concentrations respectively for 24h or 48h. At the end of incubation, fresh medium containing 500 μg/ml MTT was added. Thereafter cells were incubated for 3h at 37 °C. The formazan was then dissolved in DMSO during the process and the absorbance was measured at 570 nm using a plate www.impactjournals.com/oncotarget reader (OPTImax, Molecular Dynamics). This process was repeated in triplicate.

Cell proliferation assay
Cell proliferation was tested by CFSE staining method. U87 cell was stained with CellTrace™ CFSE staining solution and incubated for 20 minutes in a 37°C water bath. And U87 cells were seeded at density of 5×10 3 cells per well in 6-well plates. The next day, cells were treated with free luteolin or Luteolin/MPEG-PCL micelles at different concentrations for 48h. After incubation cells were harvested and analyzed in a flow cytometry. This process was repeated in triplicate.

Apoptosis study
Induction of apoptosis of U87 cells by free luteolin and Luteolin/MPEG-PCL micelles was investigated by flow cytometry (FCM, BD FACSCalibur). In brief, U87 cells were seeded at at density of 5×10 5 cells per well in 6-well plates. The next day, cells were treated with free luteolin or Luteolin/MPEG-PCL micelles at different concentrations for 48h. After incubation cells were harvested and stained with 5 μl Annexin-V-FITC and 5 μl PI (Annexin V-FITC/PI Apoptosis Detection Kit). Then the apoptosis of stained cells were analyzed in a flow cytometry. This process was repeated in triplicate.

Western blotting
Western blotting was performed to analyse the expression level of proteins involved in cell apoptosis and proliferation. U87 cells were seeded at a density of 5×10 4 cells per plate and allowed to attach overnight. Cells were then treated with free luteolin, Luteolin/ MPEG-PCL micelles at a concentration of 40 μg/ml, 20 μg/ml, 10 μg/ml, 5 μg/ml and EM for 48h. Cells treated with normal saline or blank micelles served as control groups. In the next step cells were lysed in RIPA (radioimmunoprecipitation assay) buffer. The protein extracts were separated by SDS-PAGE (sodium dodecyl sulfate-polyacrylamide) and transferred to PVDF (polyvinylidene fluoride) (Bio-Rad, Hercules, CA) membranes. The membranes were then probed with primary antibody against pro-caspase-9, cleavedcaspase-9, Bax, Bcl-2, MAPK, p-MAPK and β-actin followed by incubation with secondary antibodies. The blots were visualized by an enhanced chemiluminescence detection system (Amersham Biosciences Corp., Piscataway, NJ). This process was repeated in triplicate.

Anti-glioma effect in vivo
The anti-glioma effects of Luteolin/MPEG-PCL micelles were studied in transgenic zebrafish model and nude mice xenograft model.
In the transgenic zebrafish model, U87 tumor cells were transfected with EGFP by lentiviral. Zebrafish embryos were stripped off egg sheath and anesthetized with 0.01% tricaine at 48 hpf. Under a Zeiss Stemi 2000-C dissecting microscope (Carl Zeiss Microimaging Inc., Thornwood, NY), zebrafish was then injected with 300 U87 tumor cells in perivitelline space using a Cell Tram Vario injector (Eppendorf, USA) with a glass micropipette (L = 50 mm, diameter of the needle opening about 25 μg/ ml). At 24h after the implantation of tumor cells into zebrafish, Holtfreter's solution (control), blank micelles, free luteolin and Luteolin/MPEG-PCL micelles were added into the incubating Holtfreter's solution at a final luteolin concentration of 5 μg/ml. Finally, images of tumor were taken using a confocal microscope (DM6000 CS, Leica, Germany) at 5 days post-inoculation. This process was repeated in triplicate.
In the nude mice xenograft model, nude mice were injected subcutaneously with 0.1 mL of C6 or U87 cells containing 1×10 7 cells in the right flank. As the tumor grows to 0.1 cm 3 , the mice were randomly assigned to 4 groups and treated with normal saline (NS), empty micelles (EM), free luteolin (FL) (50 mg/kg body weight) or Luteolin/MPEG-PCL (LM) (50 mg/kg body weight) micelles everyday respectively. The diameter of tumor was assessed by caliper in two dimensions every other day and the body-weight of mice was measured in the meanwhile. Tumor volume was calculated according to the following formula: V=0.52*(ab 2 ), where a represented the length while b represented the width. Mice were sacrificed at predetermined time and tumors were immediately weighed. This process was repeated in triplicate.

TUNEL assay
Apoptosis of U87 tumors induced by Luteolin/ MPEG-PCL micelles was detected by TUNEL staining. When tumor tissues were sectioned, a TUNEL kit (Promega, Madison, WI, USA) was used to analyze apoptotic cells within tumors following the manufacturer's protocol. Five equal-sized tumor sections were detected. TUNEL positive cells were observed under a fluorescent microscope (×400).

Immunohistochemical examination of Ki67 and CD31
Tumor proliferation activity and neovascularization in tumor tissues were investigated by Immunohistochemical analysis with antibody Ki67 and CD31. Tumor sections were stained with rabbit anti-mouse ki67 polyclonal antibody (1: 50; BD PharmingenTM, USA) or rabbit anti-mouse CD31. Then they were washed twice with PBS, and incubated with FITC-or Rhodamine-conjugated secondary antibody respectively for 1h (Abcam, USA). Ki-67 positive and total cells were counted in each tumor section under microscope and Ki-67 labeling index was calculated. The CD31 positive microvessels were counted under high power field (×400).

Imaging in vivo
For imaging in vivo animal experiments, 6-weekold female nude BALB/c mice were used. Glioma tumors were induced in mice by inoculating them in the back with U87 cells (1.0 ×10 7 ). When the tumor volume reached approximately 150-200 mm 3 , Coumarin/MPEG-PCL micelles or free coumarin (1 mg/kg) were intravenously injected through the tail vein. After i.v. injection of the drug, optical fluorescence imaging was performed by positioning each mouse on an animal plate in the vivo imaging system that was heated to 37 °C (Quick View 3000). The images were obtained 0.5, 1, 2, 4, 6, 8 and 24h after injection of the test drugs.

Statistical analysis
Results were expressed as mean ± SD. Statistical analysis was performed with one-way analysis of variance (ANOVA) using SPSS 15.0 software (Chicago, IL, USA). Values of P < 0.05 are indicative of statistically significant.