Aptamer-Functionalized Dendrimer Delivery lncRNA MEG3 Enhances Castration-Resistant Prostate Cancer Gene Therapy

Androgen castration therapy is an effective treatment method for prostate cancer patients who cannot be completely cured by surgery. However, drug resistance often leads to treatment failure and poor prognosis. For castration-resistant prostate cancer (CRPC), some new treatment methods have been gradually put into clinical validation or use. Considering the side effects of traditional treatment, gene therapy has been applied in recent studies to combat prostate cancer (PCa). A 19-nt RNA aptamer, termed as EpDT3, is endocytosed when bound to epithelial cell adhesion molecule EpCAM-overexpressing cells, including various types of prostate cancer cells. Therefore, poly (amidoamine) dendrimer (PAMAM) conjugated with EpDT3 on the surface was developed for gene delivery targeting the CRPC. Moreover, the PAMAM-PEG-EpDT3 vehicles were accumulated in CRPC cells. The CRPC-inhibitory effect of PAMAM modied with EpDT3 was signicantly higher than that of the unmodied construct when loading the plasmid that encodes long noncoding (lnc) RNA MEG3 (pMEG3). In summary, the above results suggested that PAMAM-PEG-EpDT3/pMEG3 nanoparticles (NPs) have great potential for enhancing CRPC gene therapy. on the ecacy of PAMAM-PEG-EpDT3/pMEG3 NPs against CRPC, we further investigated the mechanism of PAMAM-PEG-EpDT3/pMEG3 NPs against CRPC by immunohistochemistry. The expressions of Ki67, Bcl-2, Cyclin D1, and p53 proteins in tumor tissues of different drug delivery groups were analyzed. highest level of p53, followed by the PAMAM-PEG-EpDT3/pDNA, PAMAM-PEG/pMEG3, and PAMAM-PEG-EpDT3/pMEG3 groups. The level of p53 was reduced, indicating that lncRNA MEG3 signicantly inhibits the proliferation of CRPC cells. Compared to the PAMAM-PEG/pMEG3 group, PAMAM-PEG-EpDT3/pMEG3 group showed relatively shallow staining and low H-Score. The low positive expression of mutated p53 indicated that lnc MEG3 was transferred to CRPC cells through the targeting role of EpDT3 aptamer, thereby inhibiting cell proliferation.

host of aptamers selected and generated from large synthetic libraries through the systemic evolution of ligands by exponential enrichment (SELEX) are optimal candidates for drug delivery by binding to speci c targets [15]. Shigdar et al. developed a 19-nt RNA aptamer that interacted speci cally with cancer cells that overexpress EpCAM on the cell surface and could be endocytosed after binding to the molecule [16,17]. The aptamer, named EpDT3, can further target EpCAM-overexpressing PCa cells by modifying speci c drug or gene carriers, thereby effectively and accurately ghting against PCa.
Noncoding RNA (ncRNA) is a general term for the functional RNAs that cannot encode and translate proteins, including long noncoding RNAs (lncRNA) and short ncRNAs (such as siRNA and miRNA).
Noncoding RNA plays a major role in chromosome transcription and inactivation, gene expression and shutdown, cell cycle, and apoptosis [18]. Currently, in gene therapy studies, the use of short noncoding RNAs, such as siRNA and miRNA, are preferred for the treatment of various cancers [19]. In recent years, lncRNA has been found to be closely related to the occurrence and development of tumors [20]. Among these, lncRNA MEG3 is expressed in most normal tissues, but can hardly be detected in human tumor cells. It is speculated that MEG3 inhibits tumor growth through ectopic expression [21]. Further studies on MEG3 showed that it inhibited the synthesis of DNA in meningioma cells by stimulating p53-mediated transcription, thereby inhibiting the proliferation of cancer cells [22,23]. Luo et al. also con rmed that lncRNA MEG3 inhibited the proliferation and induced apoptosis of of CRPC cells by affecting the expression of p53 [24]. In conclusion, lncRNA MEG3 is a promising tumor suppressor gene for clinical application of CRPC.
However, several studies have reported the mechanism of lncRNA with few application in the treatment of cancer [25]. Unfortunately, only a few lncRNAs exert tumor suppressor function, and due to the excessive length, they are di cult to synthesis and presented poor stability [20,26,27]. Therefore, we constructed plasmid-encoding lncRNA MEG3 (pMEG3) to overcome the stability defect of lncRNA. In addition, there is no suitable and effective vector to deliver these lncRNAs into target cells. How to deliver lncRNA into tumor cells, such as PCa, in vivo, through targeted the nano-drug delivery system has become the focus of translational medicine research.
Polyamine (PAMAM) dendrimers are the earliest and most widely distributed dendrimers with layered 3D structures that are extensively used in medical applications. Similarly, the surface groups on PAMAM can be used as anchor points which can bind or absorb different types of reagents to achieve a variety of functions, such as improving target capability, modifying solution behavior, and reducing toxicity [28,29].
Herein, EpDT3 and polyethylene glycol (PEG) were attached on the surface of PAMAM to develop a new vector PAMAM-PEG-EpDT3 for pMEG3 delivery in CRPC cells. CRPC cell lines and its xenograft on nude mice were selected as CRPC model to evaluate the targeting ability and cellular uptake mechanism of PAMAM-PEG-EpDT3, as well as the therapeutic effect of PAMAM-PEG-EpDT3/pMEG3 nanoparticles (NPs) loaded with pMEG3 ( Fig. 1).

Results And Discussion
Preparation and characterization of PAMAM-PEG-EpDT3/pMEG3 NPs The purity of commercially synthesized EpDT3 was determined by high-pressure liquid chromatography (HPLC), and the data showed >90% purity. Mass spectrometry (MS) determined the molecular weight of EpDT3 as 6285.6 and 5'Cy3-EpDT3 as 6853.8 (Figure 2A and 2B).
The synthesis of PAMAM-PEG and PAMAM-PEG-EpDT3 is illustrated in Figure 2C. The characteristic group of PAMAM-PEG-EpDT3 was veri ed by 1 H-NMR spectra. As shown in Figure 2D, PAMAM skeleton peak appeared at 2.2-3.4 ppm. In addition, the methylene characteristic absorption peak (δ 3.6) was veri ed in the 1 H-NMR spectra of PAMAM-PEG and PAMAM-PEG-EpDT3, indicating that PEG was conjugated with PAMAM. The amino group on the surface of PAMAM reacted with the succinimide of MAL-PEG-NHS, while the disappearance of the characteristic peak of MAL in the 1 H NMR spectrum of PAMAM-PEG-EpDT3 further suggested that EpDT3 was linked to PEG. The introduction of PEG not only partially blocks the positive electricity of PAMAM and reduces toxicity but also makes PAMAM-PEG as the gene carrier with prolonged circulatory effect [30]. Consequently, PAMAM-PEG-EpDT3 was formed due to the reaction between the sulfhydryl of EpDT3 and MAL on one end of the PEG chain ( Figure 2C). Since MAL and SH react under mild conditions, PAMAM-PEG-EpDT3 was synthesized while EpDT3 activity was maintained.
Newly prepared PAMAM, PAMAM-PEG, and PAMAM-PEG-EpDT3 were incubated with varying amounts of pMEG3 to achieve the pMEG3-loaded NPs. Agarose gel electrophoresis was used to evaluate the binding ability of pDNA. As shown in Figure 3A, the vehicles encapsulate pMEG3 without any leakage when the N/P ratio is >10. Next, we selected PAMAM-PEG-EpDT3/pMEG3 NPs prepared at N/P ratio 15 for further evaluation. Figure 3B shows that the mean particle size of PAMAM-PEG-EpDT3/pMEG3 NPs was 180±0.15 nm with a polydispersity of 0.236 and zeta potential 19.7±0.23 mV ( Figure 3C). The appropriate particle size and zeta potential ensure that the NPs are enriched in tumor tissues through the enhanced permeability and retention effect (EPR) effect and taken up by tumor cells [31,32].
Evaluation of the targeting ability of EpDT3 and PAMAM-PEG-EpDT3 PC-3 and DU-145 cells expressing EpCAM were analyzed as CRPC cell models in vitro. Shigdar et al. suggested that the aptamer EpDT3 binds speci cally to EpCAM expressed on the surface of CRPC cells [11,12] and mediates endocytosis. The localization of EpDT3 was investigated by CLSM. As shown in Figure 4A, the signals of Cy3-EpDT3 in PC-3 and DU-145 cells were distributed in both cytoplasm and cell membranes, suggesting that EpDT3 is successfully initialized by CRPC cells.
The selective and effective uptake of nanomaterials by target cells is crucial to the ultimate therapeutic effect [33,34]. Therefore, EpDT3 increases the uptake of nanomaterials in CRPC cells. To further clarify the uptake mechanism, we rst loaded the commercially available uorescent probe BODIPY onto PAMAM-PEG-EpDT3, and then incubated PC-3 and DU-145 cells with different concentrations of BODIPY-PAMAM-PEG-EpDT3. Subsequently, the concentration-dependent cellular uptake of PAMAM-PEG-EpDT3 was evaluated in PC-3 and DU-145 cells qualitatively ( Figure 4B). It was observed that the BODIPY signal increases in a concentration-dependent manner, indicating that EpDT3 increases the cellular uptake of PAMAM. When the concentration of BODIPY-PAMAM-PEG-EpDT3 was 0.05-1 µM, the uptake in PC-3 and DU-145 cells was positively correlated with the concentration of the vehicle.
Quantitative analysis by ow cytometry also revealed that compared to BODIPY-PAMAM, the cellular uptake of BODIPY-PAMAM-PEG-EpDT3 is increased. The increase in BODIPY-PAMAM-PEG-EpDT3 concentration led to an increase in the positive rate of BODIPY uorescence in PC-3 cells from 77.97% to 92.14% while in DU-145 cells from 64.68% to 97.22%. The cellular uptake of PAMAM-PEG-EpDT3 was higher than that of PAMAM-PEG, which might be attributed to the ability of EpDT3 to combined with CRPC cells ( Figure 4C). This phenomenon was consistent with the results observed by CLSM.

Endocytosis mechanism study
The mechanism of the endocytosis of BODIPY-PAMAM-PEG-EpDT3 in PC-3 cells was examined. As shown in Figure 5A, we observed that the green uorescence intensity in each group declined after the treatment of endocytosis inhibitors. Compared to the control group, the uorescence intensity was signi cantly decreased in lipin and phenylarsine oxide-treated groups, while colchicine had a lower in uence on cellular uptake. When BODIPY-PAMAM-PEG-EpDT3 was incubated with cells at 4 ℃, the cellular uptake of the vehicles was reduced signi cantly, indicating that the endocytosis of BODIPY-PAMAM-PEG-EpDT3 was an energy-dependent pathway.
The mechanism of the cellular uptake of PAMAM-PEG-EpDT3/pMEG3 by PC-3 cells is shown in Figure  5B. Interestingly, the decrease in the red uorescence intensity of EMA was observed in all the endocytosis inhibitor-treated groups. Similarly, the conditions presented in the study of endocytosis mechanism of BODIPY-PAMAM-PEG-EpDT3-, lipin-, and phenylarsine oxide-treated groups showed lower red uorescence as compared to the control group, while the decrease in the colchicine group was not signi cant. Thus, it can be deduced that this process was energy-dependent. In addition, the endocytosis mediated by EpCAM and EpDT3 and the electrostatic interaction between the cationic surface of PAMAM-PEG-EpDT3 and cell membrane might also mediate endocytosis. After the treatment with excessive EpDT3, the cellular uptake of PAMAM-PEG-EpDT3/EMA-DNA declined markedly, suggesting that EpDT3 speci cally binds to the cell membrane and enhances the cellular uptake of PAMAM-PEG-EpDT3/pMEG3.
In vitro anticancer e cacy PAMAM and PAMAM-PEG-EpDT3 empty vectors were incubated with PC-3 and DU-145 cells, and their effects and potential toxicity were studied by CCK-8 analysis. Notably, at the same concentration, the cell viability of the PAMAM group was signi cantly lower than that of the PAMAM-PEG-EpDT3 group. This indicated that the modi cation of PEG and EpDT3 markedly reduces the toxicity of PAMAM, thereby improving the biocompatibility of the NPs ( Figure 6A and 6B).
According to the previous reports [24], the expression of lncRNA MEG3 in PCa tissues was signi cantly downregulated as compared to that of the adjacent normal prostate tissues. LncRNA MEG3 inhibits tumor cell proliferation or induces tumor cell apoptosis by stimulating p53-dependent transcription. Therefore, we constructed pMEG3-expressing lncRNA MEG3 and constructed PAMAM-PEG-EpDT3/pMEG3 NPs to study its gene therapy effect on CRPC cells.
In vivo anticancer e cacy and safety evaluation Nude mice were injected with PC-3 cells subcutaneously to establish a CRPC model. Then, the anti-CRPC e cacy of PAMAM-PEG-EpDT3/pMEG3 NPs was tested in tumor-bearing nude mice. The tumor images of CRPC tumor-bearing nude mice are shown in Figure 7A. The Saline group and the PAMAM-PEG-EpDT3/pDNA group had relatively large tumor volumes, followed by the PAMAM-PEG/pMEG3 group.
Among the four groups, PAMAM-PEG-EpDT3/pMEG3 group exhibited the smallest tumor volume. Intriguingly, the tumor volume data showed that PAMAM-PEG/pMEG3 and PAMAM-PEG-EpDT3/pMEG3 effectively inhibits the growth of PC-3 transplanted tumors. Moreover, PAMAM-PEG-EpDT3/pMEG3 exerted a better anticancer effect than PAMAM-PEG/pMEG3, which could be attributed to the strong tumor targeting of PAMAM-PEG/pMEG3, facilitating an easy entry of the NPs into tumor cells, which is bene cial in gene therapy. Figure 7B shows the weight of the tumor isolated after the experiment. Compared with the saline group, the tumor inhibition rate of the PAMAM-PEG-EpDT3/pDNA group was only 5.35%, and no obvious antitumor effect was shown. The tumor inhibition rates were 40.95% and 63.34%, respectively. The latter had a signi cantly smaller tumor weight and a higher tumor suppression rate, which had a stronger effect on inhibiting the growth of CRPC cell transplanted tumors. The above results indicate that the main reason why PAMAM-PEG/pMEG3 and PAMAM-PEG-EpDT3/pMEG3 inhibit tumor growth is the expression of LncRNA MEG3, not the role of the nanocarrier itself. In addition, the modi cation of EpDT3 aptamer can make more NPs accumulate in the CRPC site, improve the e ciency of pMEG3 expressing LncRNA MEG3 in tumors, and thus play a better antitumor effect. Figure 7C shows the average weight change of nude mice with transplanted tumor in each group. The weight of the nude mice in each group was relatively stable during the rst 13 days, and the weight increased slightly, with an average weight distribution of 22-25g. In the later period of the experiment, the weight of the nude mice in each group began to decline. Among them, the body weight of nude mice in the saline group and PAMAM-PEG-EpDT3/pCDNA group decreased signi cantly, down to about 19 g. The weight of nude mice in the PAMAM-PEG/pCDNA MEG3 group and PAMAM-PEG-EpDT3/pCDNA MEG3 group decreased slowly and remained above 20g. This is because with the proliferation of tumor blood vessels, the tumor tissue began to absorb nutrients from the nude mouse body, so the weight of nude mice showed a downward trend in the later period. However, in the PAMAM-PEG/pMEG3 group and PAMAM-PEG-EpDT3/pMEG3 group, the tumor growth rate was slower and the tumor volume was relatively small due to the drug treatment, so the weight loss of nude mice was not obvious. At the same time, the results of changes in body weight also indicate that the nanocarrier itself has little side effects.
To assess the safety of the treatment system, we injected the NPs intravenously in healthy mice and collected blood samples 24 hours later. Aspartate transferase (AST) and glutamic acid alanine transferase (ALT) are two important indicators of liver function, and creatinine (CRE) is an important indicator of kidney function. Figure 7D shows the ALT, AST and CRE test results of different mouse groups. There is no signi cant difference between the data of each group, indicating that neither nanocarriers nor gene-carrying NPs will damage the liver and kidney functions of nude mice and are safe gene delivery systems.
Histological analysis Figure 8 shows the hematoxylin and eosin (H&E) staining results of tumor tissues of different drug delivery groups. The saline and PAMAM-PEG-EpDT3/pDNA groups showed good tumor growth, no areas of necrosis, large tumor cell nuclei, and relatively clear edges, indicating that the nanocarriers had no obvious toxicity and side effects. The PAMAM-PEG/pMEG3 group and the PAMAM-PEG-EpDT3/pMEG3 group had loose or vacuolar tumor tissue, and deep nuclear staining (as shown by the black arrow) was also observed in PAMAM-PEG-EpDT3/pMEG3 or Fragmentation and dissolution (as indicated by the red arrow), accompanied by in ltration of many neutrophils. These results indicate that LncRNA MEG3 has obvious CRPC inhibition, while EpDT3 aptamer-modi ed NPs have a stronger anti-CRPC effect.
After a series of studies on the e cacy of PAMAM-PEG-EpDT3/pMEG3 NPs against CRPC, we further investigated the mechanism of PAMAM-PEG-EpDT3/pMEG3 NPs against CRPC by immunohistochemistry. The expressions of Ki67, Bcl-2, Cyclin D1, and p53 proteins in tumor tissues of different drug delivery groups were analyzed.
Ki67 is a nuclear antigen associated with proliferating cells. Its function is closely related to mitosis and is indispensable in cell proliferation. Ki67 can be used to label cells outside the G0 phase of the cell cycle, which is termed as the proliferation index of cells [35][36][37]. The higher the positive expression of Ki67, the higher the proportion of cells in the proliferation cycle and the faster the rate of tumor growth. The Hscore of Ki67 protein in the PAMAM-PEG-EpDT3/pDNA, PAMAM-PEG/pMEG3, and PAMAM-PEG-EpDT3/pMEG3 groups was 54.77, 34.71, and 12.64, respectively. Furthermore, PAMAM-PEG-EpDT3/pDNA group showed an intense staining, the highest H-score, and the maximal expression of Ki67 ( Figure 8). Compared to the other groups, PAMAM-PEG-EpDT3/pDNA group showed super cial staining, low H-score, and low expression of Ki67, indicating that lncRNA MEG3 signi cantly inhibits the proliferation of CRPC cells. Additionally, EpDT3 aptamer enabled the drugs to accumulate in tumor sites, which improved the transfection e ciency of pMEG3 and inhibited the proliferation of cancer cells.
The expression of Bcl-2 protein is an indicator of apoptosis [38][39][40]. As shown in Figure 8, both the Saline and the PAMAM-PEG-EpDT3/pDNA groups showed an intense staining and high positive expression of Bcl-2, while the PAMAM-PEG/pMEG3 and PAMAM-PEG-EpDT3/pMEG3 groups both showed a reduced staining intensity and low positive expression of Bcl-2, indicating that LncRNA MEG3 signi cantly promotes the apoptosis of CRPC cells.
Immunohistochemical analysis of Cyclin D1 protein in tumor tissues of different groups are shown in Figure 8. Cyclin D1 protein is expressed in all the drug delivery groups. The Saline group and PAMAM-PEG-EpDT3/pDNA group exhibited a high Cyclin D1 expression, while the PAMAM-PEG-/pMEG3 and PAMAM-PEG-EpDT3/pMEG3 groups showed reduced staining intensity and H-score, indicating a downregulated expression. Cyclin D1 is a key protein regulating the G1 phase of the cell cycle, and its main function is to promote cell proliferation. Also, it is a proto-oncogene, and its overexpression leads to uncontrolled cell proliferation and malignancy of tumors [41][42][43]. These results indicated that lncRNA MEG3blocks cell cycle and inhibits the proliferation of CRPC cells.
Finally, we detected the expression of p53 protein in all groups. The wild-type P53 gene is a tumor suppressor that inhibits the proliferation and differentiation of cancer cells and regulates the G1 phase of the cell cycle [44]. When mutated or inactivated, p53 becomes an oncogene, causing malignant proliferation of tumor cells, which is closely related to the occurrence and evolution of tumors [45]. The wild-type p53 protein is unstable and has a short half-life, rendering it di cult to be detected by immunohistochemistry, and mutations in the TP53 have previously been con rmed a strong correlation with upregulated p53 expression as measured by IHC [46][47][48].
As shown in Figure 8, the Saline group had the highest level of p53, followed by the PAMAM-PEG-EpDT3/pDNA, PAMAM-PEG/pMEG3, and PAMAM-PEG-EpDT3/pMEG3 groups. The level of p53 was reduced, indicating that lncRNA MEG3 signi cantly inhibits the proliferation of CRPC cells. Compared to the PAMAM-PEG/pMEG3 group, PAMAM-PEG-EpDT3/pMEG3 group showed relatively shallow staining and low H-Score. The low positive expression of mutated p53 indicated that lnc MEG3 was transferred to CRPC cells through the targeting role of EpDT3 aptamer, thereby inhibiting cell proliferation.

Conclusion
In this study, PAMAM-PEG-EpDT3 was developed as a novel carrier for targeted delivery of lncRNA MEG3 to CRPC cells. PAMAM-PEG-EpDT3 presented more excellent targeting ability to CRPC than that without the modi cation of EpDT3. In addition, we elucidated the potential mechanism underlying the cellular uptake of PAMAM-PEG-EpDT3/pMEG3 NPs. In vitro and in vivo models con rmed the anti-CRPC effect of PAMAM-PEG-EpDT3/pMEG3 NPs, which indicate its great potential in gene therapy for CRPC.
Cell culture PC-3 and DU-145 cells considered to be CRPC cell lines were provided by the Chinese Academy of Sciences (Shanghai, China). The cells were cultured in RPMI 1640 medium containing 10% fetal bovine serum (FBS), 100 µg/mL streptomycin, and 100 U/mL penicillin under 5% CO 2 at 37 °C humidi ed atmosphere.

Synthesis of PAMAM-PEG-EpDT3
PAMAM G5 was solubilized in methyl alcohol, followed by nitrogen-drying. The PEG with two functional groups (NHS-PEG3500-MAL) was diluted in PBS (pH 8.0) to 10 mg/mL. The mixture of PAMAM and NHS-PEG-MAL at a molar ratio of 1:2 was stirred at room temperature for 2 h in the dark. The resulting mixture was concentrated two times by ultra ltration device (MWCO 10 kDa) at 12000 rpm for 20 min at 4 °C and resolubilized in PBS (pH 7.0). The nal product was lyophilized to obtain PAMAM-PEG-MAL.
The thiolated EpDT3 was coupled to MAL on the above synthesized PAMAM-PEG-MAL. Brie y, the mixture of PAMAM-PEG-MAL and thiolated EpDT3 at a molar ratio of 10:1 was stirred in a nitrogen atmosphere at room temperature for 24 h in the dark. Further puri cation was carried out by ultra ltration to obtain PAMAM-PEG-EpDT3. The resulting synthesized product was identi ed by 1 H-NMR analysis on a Varian Mercury Plus NMR 400 MHz spectrometer (Varian, Salt Lake City, UT, USA).

Evaluation of cellular uptake of PAMAM-PEG-EpDT3
The binding capacity of EpDT3 to CRPC cells was evaluated. PC-3 and DU-145 cells were seeded in the chambered coverslips in 24-well plates at a density of 7 × 10 4 /mL and cultured to 50% con uency before the culture medium was removed. Cy3-EpDT3 was added to evaluate the localization of EpDT3 in cells cultured for 30 min, while those without treatment were adopted as the blank control. Subsequently, the culture medium was removed, and the cells were xed with paraformaldehyde for 20 min. Subsequently, the nuclei were stained with Bisbenzimide Hoechst 33342 (3 µg/mL) for 15 min. Finally, the cells were washed, and the intracellular location of EpDT3 was captured via Leica SP8 Inverted Confocal Microscope (Leica, Heidelberg, Germany) under × 100 magni cation.
Fluorescence microscopy was used to evaluate the cellular uptake of PAMAM-PEG-EpDT3 by PC-3 and DU-145 cells in a concentration-dependent manner. The cells were seeded in 24-well plates at a density of 7 × 10 4 /mL and cultured to 90% con uency. Then, BODIPY-PAMAM-PEG-EpDT3 at the PAMAM concentrations of 0.05, 0.1, 0.2, 0.5, and 1 µM was added to the culture medium. The cells were incubated for 30 min, washed three times with PBS, and observed with IX53 Inverted Fluorescence Microscope (Olympus, Tokyo, Japan) to evaluate the cellular uptake of PAMAM-PEG-EpDT3.
Flow cytometry was also used to study the cellular uptake of PAMAM-PEG-EpDT3 in PC-3 and DU-145 cells. The cells were seeded in 6-well plates at a density of 2 × 10 5 /mL and cultured to 90% con uency. Then BODIPY-PAMAM-PEG-EpDT3 or BODIPY-PAMAM was added at the concentration of 0.5 µM. After incubation for 30 min, the cells were washed, trypsinized, and resuspended in 200 µL PBS (pH 7.0). The uorescence intensity of BODIPY was measured using a ow cytometer (NovoCyte, ACEA Biosciences Inc., Hangzhou, China) to determine the cellular uptake. The cells without treatment were used as negative control.
Preparation of PAMAM-PEG-EpDT3/pMEG3 NPs The plasmid encoding lncRNA MEG3 (pMEG3) was chosen to construct PAMAM-PEG-EpDT3/pMEG3 NPs to investigate the inhibitory effect of CRPC cells. pMEG3 was diluted to 100 µg/mL with 50 mM Na 2 SO 4 , which was mixed with newly synthesized PAMAM, PAMAM-PEG, and PAMAM-PEG-EpDT3 at different phosphorus (N/P) ratio for 30 s and incubated at room temperature for 30 min.

Endocytosis inhibition study
Ethidium monoazide bromide (EMA) was used to label the plasmid vector pMEG3, according to the manufacturer's protocol. In order to elucidate the mechanism underlying the cellular uptake, PC-3 cells were seeded on the coverslips in 24-well plates at a density of 7 × 10 4 /mL and cultured to 90% con uency. Endocytosis inhibitors, such as lipin (inhibitor of caveolin-mediated endocytosis), phenylarsine oxide (inhibitor of clathrin-mediated endocytosis), and colchicine (inhibitor of macropinocytosis), were added to the culture medium, respectively. After 30 min pre-treatment, the cells were incubated with a new culture medium containing BODIPY-PAMAM-PEG-EpDT3 (equivalent to 1 µM PAMAM) for an additional 30 min. In order to assess if the uptake was energy-dependent, the cells were incubated with BODIPY-PAMAM-PEG-EpDT3 at 4 °C for 30 min. To investigate the speci city of the cellular uptake of PAMAM-PEG-EpDT3, PC-3 cells were pretreated with 10 µM EpDT3 for 30 min to saturate the EpCAM receptor, followed by washing with PBS and examination by uorescence microscopy.
Moreover, the subsequent steps were carried out as described above to investigate the cellular uptake of PAMAM-PEG-EpDT3/EMA-pMEG3.

Cytotoxicity assays in vitro
The cytotoxicity of PAMAM and PAMAM-PEG-EpDT3 vehicles was studied to determine its safety by CCK- Twenty nude mice with PC-3 cell xenograft tumors of uniform shape and size were randomly divided into four groups (5 mice per group) and treated with A) PBS, B) PAMAM-PEG-EpDT3/pDNA, C) PAMAM-PEG/pMEG3, and D) PAMAM-PEG-EpDT3/pMEG3, respectively. The injection was administered on days 1, 3, 5, 7, and 9 at a dose of 2.5 mg/kg pDNA or pMEG3, followed by observation for 10 days. The bodyweight of the mice were monitored every 2 days. Subsequently, the mice were sacri ced, the tumors and vital organs were excised for analysis. In addition, 1.5 ml whole blood was taken to measure ALT, AST and CRE.

Histological analysis
Tumor tissues were xed with 4% paraformaldehyde for 24 h. After dehydration with alcohol gradient, the tumor tissues were embedded in para n, sliced into 4-µm-thick sections, dewaxed, and stained with H&E. Images were acquired using a Nikon (Tokyo, Japan) Eclipse CI microscope equipped with DS-U3 imaging system for histopathological evaluation.
The expression of proteins Ki67, Bcl-2, Cyclin D1, and P53 proteins in tumor tissues was detected by immunohistochemistry by an experienced pathologist.
Statistical analysis SPSS 18.0 software was adopted for statistical analysis, and data are presented as mean ± SD. One-way analysis of variance (ANOVA) was used to analyze the signi cance between groups. P < 0.05 indicated a signi cant difference.