Synergistic Effect of Autophagy Inhibitor Si-beclin1 Combined With the Doxorubicin Nano-delivery System Against Advanced Prostate Cancer


 Apoptosis tolerance is an important mechanism of tumor resistance in tumor therapy. Autophagy can prevent apoptosis induced by antitumor drugs and promote tumor resistance. The purpose of the presentn study was to improve the sensitivity of chemotherapeutic drugs and enhance the efficacy by inhibiting autophagy. In the present study, hydrophobic doxorubicin-hydrazone-caproyl-maleimide (DOX-EMCH) and autophagy-inhibitory si-Beclin1 were simultaneously delivered via the amphiphilic peptide micelle system(Co-MPs) using ploy(L-arginine)-poly(L-histidine)-DOX-EMCH as the copolymer building unit. It was found that the constructed micelle system promoted the escape of si-Beclin1 from endosomes and the release of DOX into the nucleus and the Co-MPs exhibited 2.7–fold higher cytotoxicity and apoptosis, in PC3 cells than DOX treatment alone did, which demonstrates the si-Beclin1 inhibited the autophagy activity of prostate cancer (PCa) cells by targeting the type III PI3K pathway and improve the sensitivity to the chemotherapy drug DOX. In addition, the peptide micelles successfully targeted DOX and si-Beclin1 passively to the tumor tissue. Compared with DOX or si-Beclin1 treatment alone, the Co-MPs showed a 3.4-fold greater tumor inhibition in vivo, which demonstrated a synergistic anti-proliferative effect in vivo. Our results suggest that the Co-MPs developed in this study may prove to be a promising combination method to provide autophagy inhibition and chemotherapy in cancer treatment, especially for PCa.


Background
The rising incidence of prostate cancer (PCa) has seriously endangered human health in men. According to the 2019 data from the communications authority (CA), PCa is the leading cause of cancer-related death in men in the United States [1]. Endocrine therapy is the mainstay of treatment for advanced PCa.
Autophagy is a process of cell self-degradation, which plays an important role in the regulation of metabolic stress and maintenance of genome integrity and stability of the internal environment.
Autophagy is closely associated with cancer and is bidirectional to tumor cells. Autophagy can prevent apoptosis induced by antitumor drugs, and promote tumor resistance [15][16][17].Some studies [18][19][20] reported that treatment with Docetaxel signi cantly up-regulated autophagy in lung cancer, and that autophagy inhibitors could regulate the autophagic activity, thus improving the e cacy of Docetaxel, suggesting that inhibition of autophagy to promote the sensitivity of chemotherapeutic drugs may also be an alternative treatment for PCa.

Page 4/17
The condensation ability of the complexes was determined by agarose gel electrophoresis. As shown in Fig. 2D, with the N/P ratio increasing from 0.5 to 20, the condensing ability of DOX-conjugated micelles was enhanced gradually. When the N/P ratio was greater than 10, the si-Beclin1 was completely condensed, indicating that the compression ability of DOX-conjugated micelles was improved by electrostatic interaction [38]. To con rm this nding, DTT, a reducing agent, was used to break this disul de bond. As shown in Fig. 2D, the Co-MPs showed a weaker si-Beclin1 binding a nity in the presence of DTT because of the depolymerization of Co-Mps. Based on the above results, we hypothesized that the broken of inter-molecular disul de bond can release DOX and si-Beclin1 by reducing the condition in the cytoplasm, thus reducing the a nity of CO-MPs. This reductive sensitive drug delivery system remain stable extracellularly but not stable in the cytoplasm, which can maintain effective release.

Release at different pHs of DOX
The DOX release pro les of Co-MPs were investigated at pH5.5 and pH 7.4 at 37℃, respectively. As shownin Fig. 2E, the release of DOX from the micelles was pH-sensitive. For example, DOX release from Co-MPs reached 78.9% at pH 5.5 but only 54.8% at pH 7.4 (P < 0.01) when observed for 48 h, which is most likely due to the release of the DOX from the disul de bond under acidic conditions and protonation of the histidine within the structure of the micelles at endolysosomal pH (about 5.0) [39,40]. The proton sponge effect of histidine can destroy the internal structure of micelles, thereby promoting the release of DOX [41].

Cellular uptake
Knowing that effective cellular uptake is essential for drug delivery [42], we labeled si-Beclin1with FAM probe, and incubated DOX, DOX-MPs and Co-MPs with PC3 cells for 4 h. The positive cells were quantitatively assessed by ow cytometry after 4-h incubation. As shown in Fig. 3A, B, the uorescent signal of FAM increased with the increase of the N/P ratio, and when the N/P ratio was 80, the number of positive FAM cells was 1.79-fold and 1.32-fold as high as that at an N/P ratio of 20 and 40(p < 0.05). The uptake of DOX in PC3 cells showed a dose-dependent manner, and the positive cells at 0.5 µg/ml, 1 µg/ml and 2 µg/ml accounted for 45.57%, 54.11% and 88.66% respectively. But the DOX-MPs had higher cellular uptake at a low concentration. When the concentration was 0.5 µg/ml, the percentage of positive cells was 93.01% (Fig. 3C, D). It was shown that doxorubicin modi ed by the membrane penetrating peptide was easier to enter the cell after being made into micelles. It is noteworthy that Co-MPs showed the optimal gene uptake e ciency at an N/P ratio of 40.
To observe the intercellular location of DOX, DOX-MPs and Co-MPs, PC3 cells were observed under a confocal laser scanning microscope (CLSM). DAPI (4′,6-diamidino-2-phenylindole) was used to stain the nucleus. Figure 3E shows the intracellular distribution of the micelle in PC3 cells for 4 h after transfection.
In the free DOX, red uorescence was distributed in the nucleus and cytoplasm in coincidence with blue uorescence. In the DOX-MPs group, red uorescence was also distributed in the nucleus and cytoplasm, and the uorescence intensity was brighter than that of the simple group, indicating that the DOX could enter the nucleus to play the cytotoxic effect. In Co-MPs, the red uorescence and green uorescence were partly distributed in the nucleus and cytoplasm, indicating that DOX and si-Beclin1 were successfully transported into cells by nanomicelles, and FAM-si-Beclin1 successfully escaped from the endosome and entered the nucleus [43,44]. These results suggest that the co-delivery system could promote endosomal escape, and carry the genetic and chemotherapeutic drug into the nucleus. This nding was in good agreement with the ow cytometry data.

In vitro evaluation of autophagy inhibition
To evaluate the autophagy inhibition capability of Co-MPs in PC3 cells, ow cytometry and confocal microscopy were used to monitor autophagy. As shown in Fig. 4A, the autophagy level of free DOX was signi cantly higher than that in DOX-MPs and Co-MPs groups when the concentration of DOX was 0.5 µg/mL and si-Beclin1 was 100 nM. The percentage of autophagy-postive cells in Co-MPs group was lower than that in DOX group (6.77% vs18.00, P < 0.05), indicating that si-Beclin1 could be effectively codelivered by Co-MPs and inhibited autophagy caused by nanomaterials.
In addition to observing autophagosomes, the aggregation of uorescent spots in autophagic ux observed by confocal microscopy was also used. Tandem uorescent-tagged LC3(mRFP-EGFP-LC3) was used to monitor autophagic ux based on the different sensitivities of EGFP and mRFP uorescent proteins to acidic pH [45]. The weakening of GFP could indicate the fusion of lysosomes and autophagosomes to form autophagolysosomes. At this time, GFP uorescence was quenched and only red uorescence was displayed. As shown in Fig. 4B, the uorescence intensity of the autophagy marker protein LC3II increasedgradually in GFP-labeled DOX group as compared with the control group, and punctate red uorescence distribution was also increased, suggesting that autophagy PC3 cells were activated. The red uorescence intensity in DOX-MPs group was decreased as compared with that in DOX group, and the uorescence intensity in Co-MPs group was the weakest, indicating that autophagy in the total load group with the expression of si-Beclin1 was inhibited, thus reducing the DOX-enhanced autophagy.
In addition, the dynamic process of autophagy was observed by TEM. The presence of autophagic vacuoles in the cytoplasm indicates the emergence of autophagy [46]. Figure 4C shows the treatment of the cells with si-Beclin1, DOX, DOX-MPs and Co-MPs(DOX:0.5 µg/mL and si-Beclin1: 100 nM). In the control group, the cell structure was clear, the nuclear membrane was smooth, and the cytoplasmic organelles were rich. But the crista of mitochondria in DOX-MPs and DOX groups was ruptured, swollen and denatured, and the number of autophagosomes was increased signi cantly as compared with si-Beclin1 and Co-MPs groups, indicating that DOX was able to induce more autophagicvacuoles, and less autophagy was produced in nanoparticles of DOX.
The result of LC3II/I qRT-PCR analysis showed that LC3II/I proteins was downregulated by 41.3% and 10.3% in the Co-MPs group compared with that in the DOX and DOX-MPs (Fig. 4D). Furthermore, to explore the role of beclin1 in regulating doxorubicin-induced autophagy, the expression levels of beclin1, LC3 and P62 was detected by Western blot. Compared with the control group, DOX treatment induced a signi cant increase in the expression of the autophagy marker protein LC3II and a signi cant decrease in the expression of p62 protein (Fig. 4D). In addition, the expression of LC3II was weak in Co-MPs group, and p62 accumulated in the Co-MPs, indicating that DOX could induce PC3 cell autophagy, and wrapping of si-Beclin1 nanoparticles could reduce the incidence of autophagy.

Cell viability assay
The cytotoxicity in PC3 cells of different groups was investigated. Figure 5A, B show the cell viability after treatment with free DOX, DOX-MPs and Co-MPs at the concentration of DOX of 0 µg/mL to 2 µg/mL and si-Beclin1 concentration of 100 nM for 24 and 48 h. The proliferation of PC3 cells was inhibited in the three groups in a concentration-dependent manner. CCK-8 assay demonstrated that the IC50 value in PC3 cells at 48 h was 0.95 µg/mL,0.70 µg/mL and 0.45 µg/mL in free DOX, DOX-MPs, and Co-MPs group respectively. The IC50 value in Co-MPs group was about 2.1-fold lower than that in free DOX group (P < 0.05), indicating that co-delivery of DOX and si-Beclin1 could accelerated the intracellular DOX release and improved the endo/lysosomal escape and the release of si-Beclin1, making the Co-MPs an e cient co-delivery system with good synergy between DOX and si-beclin1, and enhanced the cytotoxic of DOX. Figure.5C shows PC3 cell apoptosis in each group. No signi cant apoptosis was observed in PC3 cells exposed to si-Beclin1 after 48-h treatment. Compared with the control group, cell apoptosis was reduced by 19.87% in DOX-MPs group, which was approximately 1.35-fold higher than that in free DOX group(14.67%).In Co-MPs group, cell apoptosis was reduced by 39.77%, which was approximately 2.0and 2.7-fold higher than that in DOX-MPs group and free DOX group. These results suggest that Co-MPs could silence the si-Beclin1 gene and inhibit the autophagy caused by the chemotherapy drug doxorubicin.

In vivo distribution
The investigation of in vivo distribution is essential for the evaluation of safety and effectiveness of nano-micellar delivery [47]. Doxorubicin has its own uorescence, which can be used as an indicator to reduce the interference of animal auto uorescence [48,49]. The tumors and major organs were excised to investigate the biodistribution of DOX and DOX-PMs. As shown in Fig. 6A, the tumors and the anatomical organs were scanned at 24 h post-administration, and images were obtained. A uorescence signal was observed in the liver, lung, kidney, and tumor in the naked DOX. In DOX-MPs group, signi cant uorescence accumulation was observed in the tumor site as a result of the EPR effect by MPs due to their nanosize. The quantitative analysis indicated that the signal intensity of DOX in the tumor tissue of DOX-MPs group was 5-fold higher than that in free DOX group (P < 0.01), and 3.2-fold lower than that in the liver of free DOX group (P < 0.01). This was due to the EPR effect and the avoidance of the reticuloendothelial system (RES) by the nanosize of the micelles. Above all, these results indicate that DHRss micells have the ability of high e cient drug delivery into tumor tissues via the EPR effect [50,51].

In vivo antitumor effect
The in vivo antitumor effect of Co-MPs and the synergistic effects of DOX and si-Beclin1 were investigated in PC3 human androgen-independent prostate tumor-bearing nude mice. The tumor volume and body weight of the nude mice were monitored at regular intervals. As shown in Fig. 6B, no inhibitory effect on tumor growth was observed in the normal saline group andsi-Beclin1 group. The tumor volume was the smallest and the inhibitory effect was the most obvious in Co-MPs group as compared with the other groups (P < 0.01). The tumors were 4.8-,3.4-and 2.1-od smaller in Co-MPs group than those in the groups treated with si-beclin1, DOX, and DOX-MPs, respectively at day 21, which con rms the synergistic effect of DOX and si-Beclin1 against PC3 solid tumors in vivo. There was also con rmed by the difference in the average weight of isolated tumors at day 21( Fig. 6C)(p < 0.05). This is consistent with the antitumor effect in vitro.
The safety of Co-MPs was evaluated by body weight changes before and after administration of DOX in nude mice [52]. As shown in Fig. 6D, the body weight of the nude mice in DOX group increased slowly and then decreased in the whole process, indicating that DOX-MPs, si-Beclin1 and Co-MPs did not induce signi cant systemic toxicity during the experimental period. However, signi cant weight loss was observed in free DOX group due to the systemic toxicity of DOX (p < 0.01) [53].

TUNEL and immunohistochemistry analysis
TUNEL staining assay was used to determine whether the Co-MPs could induce the cancer cell apoptosis which was a key factor in the inhibition of cell proliferation. As shown in Fig. 7A, an increased brown staining could be obtained in the group treated with Co-MPs. Thus, the anti-proliferative mechanism of Co-MPs was identi ed to associate with the induction of cell apoptosis. The immunohistochemical result showed that Co-MPs signi cantly inhibited the expression of LC3 protein, which was represented by the reduction of Ki67 (Fig. 7B), a key points in cell proliferation. The above results indicate that co-loaded micelles enhanced the anti-tumor activity of chemotherapeutics and were related to the inhibition of autophagy of tumor cells.

Histological analysis
Histological study was performed to further evaluate the in vivo antitumor safety. As shown in Fig. 7C, the HE staining results of tissues and organs showed that in DOX group, the myocardial cells were damaged, the myocardial bers were ruptured, and the intermuscular space was widened. But no signi cant organ toxicity was observed in si-Beclin1, DOX-MPs and Co-MPs groups. The results showed that DOX delivered by Co-MPs has remarkably reduced cardiotoxicity compared with free DOX.

Evaluation of in vivo autophagy inhibition
The process of autophagy was observed in vivo by TEM. The presence of autophagic vacuoles in the cytoplasm indicates the emergence of autophagy. As shown in Fig. 7D, the number of autophagosomes in DOX and DOX-MPs groups increased signi cantly as compared with si-Beclin1 and Co-MPs groups. This is consistent with the previous in vitro cytology experiments, indicating that inhibiting autophagy gene beclin1 could down-regulate the autophagy level in tumor tissues.

Discussion
Autophagy is an important mechanism for cells to adapt to environmental changes, prevent the invasion of pathogenic microorganisms and maintain the stability of internal environment. Autophagy activity changes in a variety of human tumors. Autophagy plays a dual role in promoting and inhibiting tumor development. In case of tumorigenesis, cancer cells confer stress tolerance such as acid environment, chemotherapy, hypoxia and de ciency of nutrition by utilizing autophagy. In some cases, the cancer cells also activate autophagy in response to various chemotherapeutic drugs, which resist cell death and decrease the curative effect [54,55].
In this study, PC-3 cell line was used as a model to investigate the synergistic e cacy of the combination of chemotherapy and autophagy inhibition by co-delivery nanomicelle. We have successfully synthesized a amphiphilic peptide micelle system for co-delivery of si-beclin1 and DOX, which can effectively silence beclin1 gene suppress DOX-induced autophagy, and consequently showed obvious tumor killing effect. The results of transmission electron microscopy showed that the number of autophagosomes was signi cantly reduced by co-delivery system. Laser confocal results show that aggregation of mRFP-LC3 spots means the aggregation of autophagic bubbles. Western blot also con rm this conclusion. In addition, the combination of si-Beclin1 and DOX signi cantly induced apoptosis and necrosis of tumor cells. The antitumor effect of Si Beclin1 and DOX in vivo showed a signi cant synergistic effect.
Berlin 1, which is homologous with yeast autophagy gene AT96, is a key factor that mediates the localization of other autophagy proteins in preautophagosomes, and participates in the regulation of mammalian autophagy formation. Beclin 1 combined with type III P13K and vps34 to induce autophagy [56]. Therefore, Beclin 1 is an important gene regulating autophagy. In our study, we successfully developed a co-delivery system with good target ability. The results showed this co-delivery system can effectively silence Beclin1gene expression and suppress DOX-induced autophagy, which could underlie its therapeutic antitumor effect and reduce systemic toxicity.

Conclusions
In the present study, we demonstrated successful application of self-assembling polypeptide-based cationic micells for co-delivery of si-Beclin1 and the chemotherapeutic drug DOX into androgenindependent prostate cancer cells both in vitro and vivo. Our results show that the synthesized peptide micelles could effectively encapsulate gene drug si-Beclin1 and chemotherapy drug doxorubicin. In vitro experiments showed that peptide micelles could load DOX and si-Beclin1 into PC3 cells, and enhance cell anti-proliferation and apoptosis by inhibiting autophagy caused by DOX. In addition, in vivo experiments showed that peptide micelles could be targeted to the tumor site through the EPR effect, and effectively inhibit the growth of PC3 xenograft tumors in nude mice. It is expected to become an effective means of combined tumor therapy.

Cells and cell culture
Human PC3 cells (Institute of Biochemistry and Cell Biology, Shanghai, China) were cultured in RPMI 1640 with 10% FBS and 1% penicillin-streptomycin, and incubated under 5% CO 2 atmosphere at 37℃.

Plasmid and transfection
mRFP-EGFP-LC3 plasmid was from Addgene. Cells were transiently transfected with the plasmid using lipofectamine 2000 (Invitrogen) according to the manufacture's instructions.

Synthesis of DHRss
First, a histidine-arginine peptide (CH3CR6, HR) was synthesized using the method of F-moc-solid phase peptide synthesis (SPPS). Second, DOX was coupled to Terminal carboxyl group of HR peptide to obtain DOX-HR (DHR) through condensation reaction. The products were puri ed by reverse HPLC. Then, the DHR(50 mg) and L-cysteine hydrochloride (0.58 mg) were dissolved in 1.8 mL distilled water (pH 7.0). Then, 0.2 mL 5% hydrogen peroxide was added dropwise to the mixed solutions with stirring and incubated for 12 h. Then, the mixture was dialyzed in water for 12 h and freeze-dried for another 24 h to produce DHRss.

Preparation of DHRss polymer micelle (DOX-PMs), si-Beclin1 loaded HRss polymer micelle (Co-MPs)
The DOX Conjugated DHRss polymer micelle (DOX-PMs) was prepared using the probe-based ultrasonication technique. DHRss (5 mg) was dissolved in 8 mL distilled water. Then, 2 mL dichloromethane was injected into the DHRss solution dropwise, followed by ultrasonication at 200W for 1 min in an ice bath using a probe-based sonicator (JY92-IIN, Xinzhi Scienti c Co., Ltd., Ningbo, China). Then, the mixed solution was immediately stirred overnight at room temperature to eliminate dichloromethane. Finally, a Millipore (MW = 3000) was used to remove the monomer in the micellar dispersion (Fig. 1). DOX conjugated and si-Beclin1-loaded DHR polymer micelles (Co-PMs) were prepared by adding an appropriate amount of si-Beclin1 to DOX-PMs at N/P = 40, followed by vortexing for 30 s and incubated for 30 min at room temperature before use. 5.6 Complex characterization 5.6.1 Particle size, zeta potential and transmission microscopy (TEM) The particle size and zeta potential of the DOX-PMs (1 mg/mL) and Co-PMs (N/P = 40, 2 µg/mL siBeclin1) were measured by dynamic light scattering (Zetasizer Nano ZS90, Malvern) at 25 ℃.Co-PMs morphology was examined using a transmission electron microscopy (TEM, Hitachi, Japan) using an acceleration voltage of 75 kV.

Agarose gel electrophoresis
The condensation ability of the complexes was determined by agarose gel electrophoresis. The complexes were prepared at different N/P ratios (0. . After 30-min incubation, the complex (1 µg siBeclin1) was added to the pores of an acetic acid-EDTA buffer and TAE-containing 1% agarose gel. The gel was run at 100V for 30 min. The nucleic acid framework was irradiated under UV. The DNA release ability was evaluated by salt separation. Complexes with an N/P ratio of 20 were prepared and incubated in 25 nM DTT at 37 °C for 2 h. Samples were analyzed with agarose gel electrophoresis under the same conditions.

In vitro drug release study
The pH-dependence of the drug release behavior of the Co-PMs was shown using a GloMax-Multi Jr Single Tube Mutiode Reader. To examine the pH-dependent dye release, 10% Co-PMs solution was treated with solutions of disodium hydrogen phosphate citrate buffer at different pHs (pH 5.5, and pH 7.4). The amount of DOX-EMCH released at each time point was determined by uorescence detector analysis. 5.7 Cellular uptake assay si-Beclin1 and DOX uptake by PC-3 cells was analyzed using ow cytometry. PC3 cells were seeded into 12-well plates at 3 × 10 5 cells per well and incubated for 24 h at 37℃ in 5% CO2. To determine the cellular uptake of si-Beclin1 by Co-MPs, FAM-labeled si-Beclin1(FAM-si-Beclin1) was complexed with DOX-MPs at N/P of 10, 20, 40 and 80 to obtain Co-MPs and incubated for 30 min. For DOX uptake, DOX solution and DOX-MPs were added to PC3 cells, with a nal DOX concentration of 0.5 µg/mL, 1 µg/mL and 2 µg/mL. After 4-h incubation, cells were washed, trypsinized, centrifuged, re-suspended in 300 µl PBS, and nally analyzed on a FACScan ow cytometer (Becton Dickinson, SanJose, CA, USA). The experiment was repeated 3 times.
For confocal laser scanning microscopy (CLSM), PC3 cells were seeded into glass-bottom 24-well plates at a density of 1 × 10 5 cells per well and incubated for 24 h. After replacing the culture medium, the free DOX solution, DOX-MPs, Co-MPs were added to PC3 cells with a nal FAM-1-si-Beclin1 concentration of 100 nM and a nal DOX concentration of 0.5 µg/mL. After 4-h incubation, the medium was discarded, the cells were xed using 4% paraformaldehyde and treated with 4,6-diamidino-2-pheylinole dihydrochoride to stain the nucleus. Then, the cells were washed, sealed with mounting medium, and imaged using a CLSM.

Western Blot assay
PC3 cells were seeded in 6-well plates at a density of 5 × 10 5 cells per well, and incubated for 24 h. The cells were treated with different groups for 24 h, and subsequently harvested and resuspended with RIPA lysis buffer, followed by 30-min incubation on ice. Then, the protein sample was collected. An equal amount of protein was denatured by boiling for 5 min, separated by SDS-PAGE, transferred onto PVDF membrane (Millipore, USA), and blocked using 5% nonfat dried milk at RT for 1 h, then probed with antibodies P62, Beclin1, LC3. Bands were quanti ed by Image J software.

Cytotoxicity assay
To evaluate the cytotoxicity of DOX and si-Beclin1, a CCK-8 assay was performed. Brie y, PC3 cells were seeded into 96-well plates at a density of 1 × 10 4 cells per well, and incubated for 24 h. The medium was then replaced with fresh culture medium containing various concentrations of the polymer. Cells without treatment were used as a control. After 24-h and 48-h incubation, fresh medium containing a 10% CCK-8 solution was added. The absorbance of each well was measured at 450 nm using a microplate reader (Thermo Fisher Scienti c, Waltham, MA, USA). The absorbance of the untreated cells was set at 100%, and cell viability was expressed as the percentage relative to the absorbance of the untreated cells. The experiment was repeated three times.

Cell apoptosis
To determine the effect of Co-MPs on cell apoptosis, PC3 cells were seeded into 12-well plates ( The uorescence of DOX was used to investigate the distribution of micelles in vivo. The absorbance of DOX was measured at 488 nm. A subcutaneous tumor model was generated by injection of 0.1 mL of PC3 cells suspension(1 × 10 6 ) into the right axilla of nude mice. The tumors were allowed to grow to approximately 100 m3 before the experiment. To determine the tissue distribution of DOX, 18 female nude mice bearing PC3 prostatic cancer were equally randomized to three groups and injected with DOX and DOX-MPs (5 mg/kg). The mice were sacri ced 24 h later to excise the heart, liver, spleen, lung, kidney and the tumor. The excised organs and tumors were washed with cold saline and imaged using the FX Pro in vivo imaging system (Carestream Health, USA).
An in vivo anti-tumor effect assay was carried out as follows: 30 mice bearing visible PC3 tumors were equally randomized into saline, si-Beclin1, DOX, DOX-MPs and Co-MPs groups. The mice were intravenously administered with the respective formulation daily for three days at a dose of 5 mg/kg DOX and 2 mg/kg si-Beclin1. The body weight and tumor volumes ([major axis]×[minor axis]2/2, measured by calipers) were monitored and recorded twice per week for 21 days. Then, the mice were sacri ced, and their tumors were excised, weighed and photographed. Tumor volume(V) was calculated as: V = A × B 2 /2.

12 Tunel And Immunohistochemical Analysis
Para n-embedded tumor tissue sections (5 µm) were subjected to TUNEL analysis, and immunohistochemistry according to standard protocols provided by the manufacturers. Apoptotic signals in tissue sections were visualized by microscopy. Immunohistochemical analyses of LC3II/I, Ki67 and Paxillin were performed. Brie y, sections were permeabilized with incubated with the LC3, Ki67 or Paxillin antibody(Cell Signaling Technology, Danvers, MA) overnight at 4℃. After washing with PBS, samples were incubated with HRP-conjugated secondary antibody (Cell Signaling Technology, Danvers, MA). LC3, Ki67 and TUNEL-postive cells were captured with a Nikon E-800 M microscope (Tokyo, Japan).

13 Histologic Analysis
After the nude mice were sacri ced, hearts, livers, spleen, lung and kidney were collected and xed in 4% paraformaldehyde for 24 h and subsequently embedded in para n. Tissue sections (5μm) were subjected to H&E staining.
5.14 in vivo TEM: the tumor tissues sections (2-3mm) were subjected to observe autophagy by transmission electron microscopy. Brie y, tissues were collected and xed with 2% glutaraldenhyde solution immediately, and xed overnight at 4℃, dehydrated in increasing concentrations of ethanol and acetone, embedded in Araldite, sliced into (5-7 nanometers) sections, post-stained with uranyl acetate and lead citrate, and nally examined under a Hitachi H7650 transmission electron microscope.

Statistical analysis
All values are presented as the mean±SD. Each value is the mean of at least three repetitive experiments in each group. The statistical signi cance was determined using Student's t-test. The differences were considered signi cant for *p<0.05.