Chemotherapy priming of the Pancreatic Tumor Microenvironment Promotes Delivery and Anti-Metastasis Efficacy of Intravenous Low-Molecular-Weight Heparin-Coated Lipid-siRNA Complex.

Pancreatic ductal adenocarcinoma (PDAC) is a type of malignant tumor with high lethality. Its high tumor cell-density and large variety of extracellular matrix (ECM) components present major barriers for drug delivery.
Methods: Paclitaxel-loaded PEGylated liposomes (PTX-Lip) were used as a tumor-priming agent to induce tumor cell apoptosis and decrease the abundance of ECM to promote cellular uptake and tumor delivery of nanodrugs. Paclitaxel exerts anti-cancer effects but, paradoxically, exacerbates cancer metastasis and drug resistance by increasing the expression of apoptotic B-cell lymphoma-2 protein (BCL-2). Thus, low-molecular-weight heparin-coated lipid-siRNA complex (LH-Lip/siBCL-2) was constructed to inhibit cancer metastasis and silence BCL-2 by BCL-2 siRNA (siBCL-2).
Results: Significant tumor growth inhibition efficacy was observed, accompanied by obvious inhibition of cancer metastasis in vivo.
Conclusion: These results suggested our sequential delivery of PTX-Lip and LH-Lip/siBCL-2 might provide a practical approach for PDAC or other ECM-rich tumors.

Introduction

Pancreatic ductal adenocarcinoma (PDAC), also known as pancreatic cancer, is a type of malignant tumor with high lethality and a low survival rate 1. Compared to other solid tumors, pancreatic tumors always contain a large variety of extracellular matrix (ECM) components such as collagen I, hyaluronic acids and cancer-associated fibroblasts, which are associated with the development and metastasis of cancer. Moreover, ECM increases tumor stroma pressure. The dense stroma forms a physical barrier that makes nanocarriers more difficult to deliver 2.

Previous studies have shown that the rate and extent of penetration of small-molecule drugs or nanocarriers into solid tumors is related to the composition and structure of the tumor tissue and the density of the tumor cells 3. Tumor priming is an emerging strategy that reduces the density of tumor cells using chemotherapeutics, and promotes the delivery and efficacy of nanodrugs in solid tumors 4. However, chemotherapy is a double-edged sword: paradoxically, paclitaxel (PTX) kills cancer cells to exert anti-cancer effects but exacerbates cancer metastasis. Yi et al. demonstrated that PTX improves the tumor microenvironment for cancer cells to metastasize via Atf3 gene, a stress-inducible gene 5. Furthermore, chemotherapy can cause serious side effects because of its cytotoxicity to normal tissues following systemic administration.

In our study, paclitaxel-loaded PEGylated liposomes (PTX-Lip) at a low dosage were used as a tumor-priming agent to enhance the delivery of nanocarriers into pancreatic tumors. PTX-Lip was used to instead of free PTX, which was used as the tumor-priming agent in a previous study 6. In this study, a low dose of PTX-Lip (50 mg/m2) significantly reduced myelosuppression and hepatotoxicity compared to free PTX (100 mg/m2) but exerted a similar effect for tumor priming. Furthermore, PTX-Lip regulated the tumor microenvironment by inducing tumor cell apoptosis and decreasing interstitial fibrosis of the tumor.

In order to further minimize the negative factors of drug resistance and cancer metastasis caused by chemotherapeutics, we further constructed low- molecular-weight heparin-coated cationic liposomes (LH-Lip) carrying B-cell lymphoma-2 protein siRNA (LH-Lip/siBCL-2), which were administered following low dose pretreatment with PTX-Lip. B-cell lymphoma-2 protein (BCL-2) is an important regulatory protein in the apoptosis pathway that promotes cell survival mainly by inhibiting cell apoptosis 7. Most chemotherapeutics including PTX and doxorubicin can trigger the apoptosis pathway and activate a cellular defense mechanism of antiapoptosis by increasing the expression of BCL-2, which prevents cell death and induces drug resistance 8. Moreover, emerging evidence has revealed that BCL-2 family proteins are regulators associated with cancer cell invasion and metastasis, and BCL-2 expression levels are positively correlated with cancer metastasis 9. Therefore, BCL-2 siRNA (siBCL-2) was chosen to silence the BCL-2 expression of pancreatic cancer cells for cancer therapy.

In designing LH-Lip/siBCL-2, cationic liposomes were coated with low-molecular-weight heparin to reduce their toxicity and enhance anti-metastasis efficacy. Cationic liposomes are widely used in gene delivery as they are biodegradable due to their biological membranous structure and exhibit lower immunogenicity and toxicity compared to viral carriers 10. However, they may cause potential toxicity both in vitro and in vivo. Wei et al. demonstrated that cationic liposomes induced cell necrosis and caused a serious inflammatory response 11. Introduction of excessively positively charged nanoparticles into systemic circulation causes hemolysis and embolism, and the nanoparticles adsorb various plasma proteins, resulting in nanoparticle aggregation 12. Mucopolysaccharides, including hyaluronic acid, chondroitin sulfate, and heparin, constitute the main components of connective tissue. Researchers have found that encapsulating chondroitin sulfate on the surface of cationic carriers greatly reduces their toxicity and increases their stability 13. Except chondroitin sulfate, low- molecular-weight heparin (LH) has been applied as a good alternative to unfractionated heparin with fewer side effects and exhibits potent anti-metastatic efficacy in nanoformulations 14. The results of our study showed that LH-Lip played a major role in inhibiting cancer metastasis and prolonged the survival of tumor-bearing mice.

Here, we designed a novel dosing regimen that consists of a low dose pretreatment with PTX-Lip accompanied by sequential administration of LH-Lip/siBCL-2 (Scheme ). In step 1, a low dose of PTX-Lip is used as a tumor-priming agent to regulate the tumor microenvironment and promote the delivery of nanodrugs. In step 2, LH-Lip/siBCL-2 is sequentially administrated to inhibit cancer metastasis and downregulate the expression of BCL-2. Antitumor and antimetastasis efficacies were evaluated both in vitro and in vivo.

Methods

Chemicals and reagents

Cholesterol was purchased from Chengdu Kelong Chemical Company (Chengdu, China). Soybean phospholipids (SPC) were purchased from Shanghai Taiwei Chemical Company (Shanghai, China). 1, 2-dioleoyl-3-trimethylammonium-propane (DOTAP) was obtained from Shanghai Advanced Vehicle Technology (AVT) Ltd. (Shanghai, China). DSPE-PEG2000-OMe and 1, 2-dioleoyl-sn-glycero-3- phosp-hoethanolamine-N-(carboxyfluorescein) (CFPE ) were purchased from Avanti Polar Lipids (Alabaster, AL, USA). Protamine sulfate (fraction X from salmon) and calf thymus DNA were purchased from Sigma-Aldrich (St. Louis, MO, USA). Paclitaxel (PTX) and low-molecular-weight heparin (Enoxaparin sodium, LH) were purchased from Melonepharma (Dalian, China). BCL-2 siRNA, scrambled siRNA and Cy5-labeled siRNA were purchased from Gene Pharma Ltd. (Shanghai, China). The BCL-2 siRNA (siBCL-2) sense sequence is 5'-GUG AUG AAG UAC AUC CAU UdTdT-3', and the anti-sense sequence is 5'-AAU GGA UGU ACU UCA UCA CdTdT-3'. The scramble siRNA (siCTL) sense sequence is 5'-UUC UCC GAA CGU GUC ACG UTT-3', and the anti-sense sequence is 5'-ACG UGA CAC GUU CGG AGA ATT-3'.

Preparation and characterization of lipid-siRNA complex (lipoplex)

Lipoplex was prepared using the thin film hydration method. 2.0 mg of cholesterol and 3.6 mg of DOTAP (molar ratio = 1:1) were dissolved in 1 mL of chloroform. Then, the organic solvent was evaporated by rotary evaporation under a vacuum, and the film was formed and hydrated in 1 mL DEPC (diethyl pyrocarbonate)-treated deionized water (ddH2O) at 37 °C for 20 min and sonicated at 80 W for 100 s. siRNA-loaded naked cationic liposome complex (N-Lip/siRNA) was prepared by mixing equal volumes of suspension A (8.3 mM liposome and 200 mg/ mL protamine) and suspension B (160 mg/mL siRNA and 160 mg/mL calf thymus DNA) and incubating for 10 min at room temperature. Then, N-Lip/siRNA was filtered using ultra-filtration tubes (30 kDa, Pall Corporation, Ann Arbor, Michigan, USA), and the free siRNA was removed by centrifugation at 9300 rcf for 30 min. Likewise, CFPE-labeled lipoplex was prepared by adding 400 μL of CFPE (100 μg/mL) to the organic solution before the lipid film was formed.

LH-Lip/siRNA was prepared by mixing N-Lip/ siRNA with an equal volume of LH solution (0, 1, 2, 3, 4, 5, 6, 7, 8, 9, and 10 mg/mL, dissolved in ddH2O). LH solution was added dropwise into N-Lip/siRNA (4 mM lipid) and then magnetically stirred for 10 min at room temperature. The hydrodynamic diameters and zeta potentials of lipoplexes were detected using a Malvern Zetasizer Nano ZS90 (Malvern Instruments Ltd., Malvern, UK). The morphology of LH-Lip/ siBCL-2 was observed under a transmission electronic microscope (TEM; JEM 100CX, JEOL Ltd., Tokyo, Japan). The entrapment efficiency (EE) of siRNA was determined using Cy5-labeled siRNA as previously described 15. The ability of N-Lip to form self- assembled complexes with siRNA was measured by agarose gel electrophoresis. N-Lip/siRNA at charge ratios N/P varying from 1/1 to 10/1 were prepared.

Cellular uptake

BXPC-3 cells were seeded on 6-well plates at a density of ~5×105 cells/well and incubated overnight. PTX-Lip was added at a final PTX concentration of 0.3 μg/mL. After 24 h, CFPE-labeled N-Lip/siBCL-2 and LH-Lip/siBCL-2 were added at a final CFPE concentration of 2 μg/mL. After 2 h, the cells were washed twice with phosphate buffered saline (PBS), and were trypsinized and resuspended in 0.3 mL PBS. The fluorescence intensity of cells was measured by flow cytometry (Cytomics™ FC 500, Beckman Coulter, Miami, FL, USA).

To evaluate the cellular uptake after priming with different concentrations of PTX, BXPC-3 cells were seeded and pretreated with PTX-Lip (0, 0.1, 0.3, 0.6, and 1 μg/mL PTX). After 24 h, CFPE-labeled LH-Lip/siBCL-2 was added as described above. The fluorescence intensity of cells was measured by flow cytometry (Cytomics™ FC 500, Beckman Coulter, Miami, FL, USA).

Silencing effect of siBCL-2-loaded lipoplex

BXPC-3 cells were seeded and pretreated with PTX-Lip (0, 0.1, 0.3, 0.6, and 1 μg/mL PTX) for 24 h as described above. Then, the cells were collected and washed twice using staining buffer (PBS containing 2% bovine serum albumin and 0.1% NaN3). Next, the cells were incubated with FITC-labeled goat anti-rabbit BCL-2 antibody. After 1 h, the cells were washed and resuspended in 0.3 mL staining buffer for flow cytometry analysis (Cytomics™ FC 500, Beckman Coulter, Miami, FL, USA).

For the silencing study, formulations with siRNA (50 nM) were added after the incubation with PTX-Lip (0.3 μg/mL). After 24 h, the cells were processed as described above.

Tumor penetration

Tumor-bearing mice were randomized into 4 groups (6 mice/group): PBS, PTX-Lip (50 mg/m2), PTX-Lip (100 mg/m2), and free PTX (dissolved in ethanol-Cremophor ELP 35 mixture, v/v = 1:1,100 mg/m2). 24 h and 48 h after injection of the above formulations, mice were injected with DID-labeled LH-Lip. After 24 h, mice were euthanized via cardiac perfusion with PBS and 4% paraformaldehyde. The tumors were collected and imaged using the IVIS spectrum system (Caliper Life Sciences, Hopkinton, MA, USA) and cryosectioned at a thickness of 10 μm. The sections were firstly stained with primary anti-CD34 antibody and then with FITC-labeled secondary antibody. All the sections were stained with DAPI and imaged by CLSM (LSM800, Carl Zeiss, Germany).

Preliminary toxicity of PTX with different dosages and formulations in vivo

To evaluate the preliminary toxicity of PTX, blood cell levels and serum biomarkers in mice were measured. Mice were randomized into 4 groups (3 mice/group): PBS, PTX-Lip (50 mg/m2), PTX-Lip (100 mg/m2), and free PTX (dissolved in ethanol-Cremophor ELP 35 mixture, v/v = 1:1,100 mg/m2). After injection of the different formulations three times, mice were euthanized and blood samples were obtained for a blood cell assay and a serum chemistry assay.

Evaluation of anti-tumor activity in vivo

BXPC-3 subcutaneous tumor-bearing mice were randomized into 6 groups: PBS, single agents (50 mg/m2 PTX-Lip per dose, 0.35 mg/kg LH-Lip/siBCL- 2 per dose), or their combinations (PTX-Lip plus N-Lip/siBCL-2, PTX-Lip plus LH-Lip/siCTL, and PTX-Lip plus LH-Lip/siBCL-2). PTX-Lip was given on day 7, day 9, and day 11, and other formulations were given on day 8, day 10 and day 12. Animal body weight and tumor volumes were monitored every 3 days. On the 22nd day, half of the mice were euthanized and their tumors were collected. The survival time of each group was continuously recorded. The expressions of BCL-2 and Caspase-3 in tumors were measured by western blotting. Tumor histology was performed after hematoxylin and eosin (H&E), BCL-2 and Caspase-3 staining. To quantify the relative BCL-2 mRNA expression in the tumor, mice were euthanized at 48 h after treatment with the different formulations. Total RNA was then harvested from tumors and qRT-PCR was performed.

A BXPC-3 metastatic tumor mouse model was established by intravenous injection of a suspension of 1×106 BXPC-3 cells (100 μL) via the tail vein into 4-week-old female nude mice. Before cancer cell injection, mice were injected with PTX-Lip (50 mg/m2 per dose) three times. N-Lip/siBCL-2, LH-Lip/siCTL and LH-Lip/siBCL-2 (0.35mg/kg siRNA per dose) were given on day 10 three times after cancer cell injection. Animal body weight was monitored every 3 days. On the 30th day, half of the mice were euthanized and their lungs and livers were collected and histology was performed after H&E staining. Lung histology was performed after Ki67 staining. The survival time of each group was continuously recorded.

Statistical analysis

All data are expressed as mean ± SD. Statistical comparisons were performed using one-way ANOVA for multiple groups (GraphPad Prism, V5.01, GraphPad, La Jolla, CA, USA). Two-tailed Student's t-test was used to compare two groups. Significant differences between or among groups are indicated by *P < 0.05, **P < 0.01, and ***P < 0.001, respectively.

Results

Characterization of lipoplex

To reduce the toxicity of cationic liposomes and improve their stability, LH-coated lipid-siRNA complex was prepared. After coating with LH (1 mg/mL), LH-Lip exhibited a negative charge and the size increased from 100 nm to 200 nm (Figure ). When the concentration of LH increased (from 1 to 4 mg/mL), the particle diameter decreased and zeta potential increased. Negatively charged LH was coated on the surface of the cationic liposomes by electrostatic attraction. More LH wrapped tightly around the cationic liposomes with increased concentration of LH, so the size of the lipoplex gradually decreased until it was stable at around 110 nm. At the same time, the zeta potential of the lipoplex increased, which might be due to the addition of LH changing the electrostatic repulsion between particles. To take advantage of the enhanced permeability and retention (EPR) effect, the diameters of nanoparticles should be smaller than about 150 nm 16. In addition, too much negative charge from LH could block delivery of siRNA 17. Therefore, we chose a relatively small diameter of LH-Lip/siBCL-2 as the optimal formulation. 8 mg/mL was chosen as the appropriate concentration of LH for preparing LH-Lip with a zeta potential of -31.20 ± 0.85 mV. The size of LH-Lip was 111.7 ± 0.99 nm, slightly larger than that of the uncoated liposomes (N-Lip) (Figure ). LH-Lip displayed an approximately spherical shape (Figure ). The binding efficiency of LH was around 50% with no significant change within 24 h (Figure ). The results of agarose gel electrophoresis showed that siRNA could be absolutely loaded onto N-Lip at charge ratios (N/P) greater than 1/1 (Figure ), and the siRNA encapsulation efficiencies of both lipoplexes were greater than 95% (N-Lip/Cy5-siRNA, 97.04% ± 0.00%; LH-Lip/Cy5-siRNA, 98.02% ± 0.00%).

The size of PTX-Lip was 98.23 ± 2.3 nm, and the zeta potential of PTX-Lip was -5.08 ± 0.23 mV. The drug-loading efficiency of PTX was 77.67% ± 0.02%, and the drug-loading content of PTX was 2.49 ± 0.00%. Figure shows that PTX-Lip exhibits sustained release of PTX compared to free PTX from dialysis tubes to the release media, accompanied with long blood circulation and more accumulation in the tumor (Figure ).

Stability of lipoplex in vitro

Figure shows that LH-Lip/siCTL retards siRNA degradation in serum condition, but free siCTL or free siCTL plus free LH degrade after 60 min, suggesting that LH-Lip/siCTL complex protects siRNA from degradation by serum nucleases. Although free siRNA released siRNA faster from dialysis tubes to the release media (Figure ), LH-Lip/siRNA protected siRNA against degradation.

Cationic liposomes can adsorb a variety of plasma proteins in the blood leading to decreased stability 18, so the lipoplex was incubated with fetal bovine serum (FBS) and mouse serum. No obvious changes to N-Lip/siBCL-2 and LH-Lip/siBCL-2 were observed 30 min after incubation with FBS (Figure ). After incubation with mouse serum, aggregation of N-Lip/siBCL-2 was observed accompanied by an increase in size (Figure ), but no obvious change was observed for LH-Lip/siBCL-2. The size of N-Lip/siBCL-2 was greater than that of LH-Lip/ siBCL-2, indicating that coating with LH prevented adsorption of serum proteins and liposome aggregation.

Hemolysis assay is often used to evaluate the toxicity and safety of materials 19. The final lipid concentration of lipoplex was 300 μΜ. The hemolysis rate of LH-Lip/siBCL-2 at this concentration was lower than 5% and the structure of erythrocytes was little changed. However, the hemolysis rate of N-Lip/ siBCL-2 was higher than 60% at this concentration (Figure ).

Flow cytometry of pancreatic cancer BXPC-3 cells showed that pretreatment with PTX-Lip increased cellular uptake of LH-Lip/siBCL-2, but did not significantly enhance N-Lip/siBCL-2 uptake (Figure ). This might be because cationic liposomes are easily internalized and pretreatment with PTX-Lip had little effect on their uptake. The cellular uptake of LH-Lip/siBCL-2 increased with increasing PTX concentration (0-1 μg/mL) and reached a maximum at a concentration of 0.3 μg/mL (Figure ). Our previous study demonstrated that pretreatment with docetaxel promoted the uptake of nanoparticles, and it was mainly caused by the G2/M retention effect 20. The G2/M retention effect describes the fact that the uptake ability of cells is strongest in the G2/M phase rather than other phases 21. The percentage of BXPC-3 cells in G2/M phase increased with increasing PTX concentration (Figure ).

Subcellular localization microscopy showed that free siRNA was minimally internalized by the cells (Figure ). However, fluorescence signal was obviously observed in the LH-Lip/siRNA group, indicating that LH-Lip/siRNA promoted uptake of siRNA into BXPC-3 cells. This enhanced uptake might result from the LH coating. Fibroblast growth factors (FGFs) have a high affinity for glycosaminoglycan heparin and promote the combination of heparin and fibroblast growth factor receptor (FGFR) 22. The BXPC-3 cells exhibited high expression levels of FGF2 (a classical isomer of the FGF family) (Figure ). In addition, the pretreatment with PTX-Lip further promoted cellular internalization of LH-Lip/siRNA (Figure ).

Silencing effect of siBCL-2-loaded liposomes on BCL-2 and cytochrome C expression

FITC-labeled rabbit anti-human BCL-2 was used to detect BCL-2 expression. The pretreatment with PTX-Lip significantly increased the expression of BCL-2 (Figure ). LH-Lip/siBCL-2 significantly inhibited the expression of BCL-2 induced by chemotherapeutics, while N-Lip/siBCL-2 slightly inhibited BCL-2 levels (Figure ). This might be because N-Lip induced high levels of apoptosis (Figure ), and expression of BCL-2 is closely related to apoptosis 7. To investigate the relationship between concentration of siBCL-2 and silencing efficiency, BXPC-3 cells were treated with different concentrations of siBCL-2-loaded LH-Lip, and BCL-2 mRNA was measured using qRT-PCR. The result showed that the pretreatment with PTX-Lip significantly increased the BCL-2 mRNA level, and LH-Lip/siBCL-2 reduced the BCL-2 mRNA level in a dose-dependent manner (Figure ). The siBCL-2 concentration that produced 50% inhibition was about 50 nM, so we chose this concentration for the further studies.

BCL-2 exerts anti-apoptotic effects by preventing release of mitochondrial cytochrome C 23. Therefore, the expression levels of cytochrome C in mitochondria and cytoplasm could further demonstrate the silencing effect of siBCL-2. Figure shows that PTX-Lip induces higher cytochrome C levels in mitochondria. However, LH-Lip/siBCL-2 inhibited the cytochrome C levels in mitochondria and showed approximately the same levels in the cytoplasm as control.

Cellular apoptosis assay and cytotoxicity study

Apoptosis induced by the different lipoplexes was measured using an Annexin V-FITC/PI Apoptosis Detection Kit. LH-Lip/siBCL-2 induced significantly less apoptosis and necrosis in LO2 cells (Figure and Figure ) but significantly more BXPC-3 cell necrosis (Figure and Figure ) than did N-Lip/siBCL-2. This result might be due to the fact that the toxicity of free LH to BXPC-3 cells was significantly greater than that to LO2 cells (Figure ). On the other hand, the anti-proliferation effect was enhanced with the increasing PTX concentration, and the LH-Lip/siBCL-2 group showed much lower cell viability compared to the other groups (Figure ).

Effects of tumor priming

Three-dimensional tumor spheroids were prepared to evaluate the solid tumor penetration ability of the lipoplex after tumor priming. The results showed that tumor spheroid uptake increased at different depths with increasing PTX concentration and reached a maximum at a concentration of 0.3 μg/mL (Figure ). Additionally, the penetration ability of Cy5-labeled siRNA was also enhanced after priming with PTX-Lip (Figure ). Compared to other tumors, PDAC is rich in ECM components, such as collagen I, hyaluronic acids and cancer-associated fibroblasts, which hinder the delivery of nanoparticles 24-26. Transwell chambers were used to evaluate penetration of a tumor stroma barrier composed of fibroblasts. In the upper chamber, cellular uptake by NIH-3T3 cells (Figure ) was the same as that shown in Figure . In the lower chamber, cellular uptake of the lipoplex by BXPC-3 cells was significantly increased with pretreatment compared to the non-pretreated group (Figure ), indicating that tumor priming could increase the penetration ability of lipoplex through fibroblasts.

BXPC-3 subcutaneous tumor-bearing nude mice were used to estimate tumor penetration with tumor priming in vivo. The pretreatment with PTX increased the penetration of DID-labeled LH-Lip compared to the PBS group (Figure and Figure ), and stronger fluorescence was observed with pretreatment for 24 h vs. 48 h (Figure ). After injection of PTX-Lip at 50 mg/m2 for 24 h, the penetration ability of Evans Blue in tumor tissues was enhanced by 63% compared to the free PBS group, and was stronger than that of the free PTX group (Figure ). Additionally, there were no significant differences in penetration ability between groups pretreated with PTX-Lip at the doses 50 mg/m2 and 100 mg/m2. Therefore, pre-injection of PTX-Lip at 50 mg/m2 for 24 h was selected for further study. Immuno-histochemical images showed that the pretreatment with PTX-Lip increased cell apoptosis in the tumor, but decreased the expression of α-SMA and collagenous fibers (Figure ), which are components of tumor stroma 27.

A previous study demonstrated that pretreatment with free PTX could promote the interstitial transport, penetration and dispersion of nanoparticles 6. However, free PTX causes severe myelosuppression, and Cremophor induces anaphylaxis 28. PTX liposomes can avoid Cremophor as a solvent and can reduce the therapeutic dose and toxicity of PTX. Table shows that mice treated with free PTX (100 mg/m2) displayed a significant decrease in white blood cells (WBC), platelets (PLT) and neutrophile granulocyte (Gran) compared to the PBS group, indicating severe myelosuppression of the blood system. No significant differences in WBC, PLT and Gran were observed between mice treated with PTX-Lip (50 mg/m2) and PBS.

Glutamic-pyruvic transaminase (ALT) and alkaline phosphatase (ALP) are measured to monitor hepatic function 29, and urea nitrogen (BUN) and creatinine (CREA) are measured to assess renal function 30. Table indicates that the level of ALT and ALP significantly increased in the free PTX (100 mg/m2) group and the PTX-Lip (100 mg/m2) group, but changes in the PTX-Lip (50 mg/m2) group were negligible. Additionally, no obvious influence on BUN and CREA was observed among all groups. Therefore, a low dose (50 mg/m2) of PTX-Lip showed lower toxicity than a higher dose (100 mg/m2) of PTX-Lip and free PTX.

Inhibitory effect on cell migration and invasion

A previous study demonstrated the inhibitory effect of LH on melanoma cell migration and invasion 14. Here, wound healing assay was used to evaluate the migration of BXPC-3 cells. Both LH-Lip/siBCL-2 and pretreated groups hindered the migration of BXPC-3 cells, and the combination of pretreatment and LH-Lip/siBCL-2 exhibited the strongest inhibitory effect (Figure and Figure ). N-Lip/ siBCL-2 also hindered cell migration, attributed to the anti-metastatic effect of siBCL-2 9. Matrigel-coated transwell chambers were used to evaluate cell invasion. As shown in Figure and Figure , BXPC-3 cells were aggressive and could easily penetrate the matrigel layer. The combination of pretreatment and LH-Lip/siBCL-2 exhibited the strongest inhibitory effect.

Interactions of tumor cells with platelets and actin cytoskeleton

Emerging evidence indicates that platelets promote cancer metastasis via platelet-tumor cell interactions, and induce an epithelial-mesenchymal- like transition (EMT) 31. A previous study demonstrated that LH could inhibit the adhesion of platelets and melanoma cells 14, but its inhibitory effect on pancreatic cancer cells has not been evaluated. The results of this study suggested that formulations with LH could significantly inhibit the adhesion of platelets and BXPC-3 cells (Figure and Figure ). EMT is pivotal for metastasis of PDAC 32, and the switch between E- and N-cadherin is a classical feature of EMT 33. After incubation with platelets, the expression of E-cadherin decreased but N-cadherin levels increased, indicating that platelets could induce EMT of BXPC-3 cells. However, BXPC-3 cells treated with the formulation with LH expressed more E-cadherin and less N-cadherin (Figure ). The tensile forces of actin stress fibers encourage tumor cells to invade and metastasize 34. CLSM imaging indicated that formulations with LH attenuated the formation of actin filaments and lamellipodia (Figure ).

Anti-tumor and anti-metastasis efficacy

BXPC-3 subcutaneous tumor-bearing mice were injected with free Cy5-siRNA and Cy5-siRNA-loaded liposomes. In the tumor, a higher fluorescence signal was observed in the LH-Lip/Cy5-siRNA group than in the free Cy5-siRNA group (Figure ). This is logical because the LH-coated lipoplex should have prolonged circulation compared to cationic liposomes 35. Additionally, the pretreatment with PTX-Lip increased the tumor penetration of Cy5-siRNA compared to the un-pretreated group. The tumor-to- liver ratio of LH-Lip/Cy5-siRNA was highest among all groups (Figure ). This suggested that LH-Lip/Cy5-siRNA exhibited better behavior in vivo, with more tumor accumulation and relatively less distribution to the liver than the other formulations.

For evaluating the therapeutic effect, mice were treated as shown in Figure . Malignant tumor always increases body energy consumption, resulting in body weight loss 36. However, body weight increased gradually during the treatment (Figure ), which might be caused by rapid increases in tumor weight during the early stage of tumorigenesis. Treatment with LH-Lip/siBCL-2 alone did not significantly prolong the survival time (27 days), but pretreatment with a low dose of PTX-Lip before administration of LH-Lip/siBCL-2 significantly prolonged the survival time to 39 days (Figure ). The pretreatment with PTX-Lip before administration of nanodrugs (including LH-Lip/siBCL-2, LH-Lip/ siCTL, and N-Lip/siBCL-2) also showed significant reduction in tumor growth (Figure ). The expressions of BCL-2 and Caspase-3 were detected using western blotting and immuno-histochemical staining. Figure shows that treating with PTX-Lip induced high BCL-2 levels in tumor tissues, but the combination with siBCL-2 reduced the expression of BCL-2, indicating that the dramatic silencing effect of siBCL-2-loaded liposomes. The relative expression levels of BCL-2 mRNA (Figure ) showed the same result as western blotting. Additionally, PTX-Lip in combination with LH-Lip/siBCL-2 induced significant apoptosis according to the apoptosis marker Caspase-3 (Figure ).

A previous study demonstrated that free PTX exacerbates cancer metastasis 5, which is a risk factor for chemotherapy. To test whether PTX-Lip affects cancer metastasis, mice were treated following the scheme presented in Figure . Mice showed continuous weight loss from day 18, which was obvious in the PBS and PTX-Lip treatment groups (Figure ). No change in survival time was observed in groups treated with PBS or PTX-Lip alone (Figure ). Metastatic nodules in the lungs and liver significantly increased in the PTX-Lip-treated group (Figure ), indicating that pretreatment with PTX-Lip exacerbated pancreatic cancer metastasis. However, our ideal formulation could reduce the risk of cancer metastasis and prolong the survival time. H&E-stained images of lungs show that this combination therapeutic strategy inhibited edema in the parenchyma (Figure ), resulting in inhibition of the inflammatory response. Additionally, Ki67-stainied images of lungs show active cell proliferation activity in metastatic nodules (Figure ).

Discussion

Cationic liposomes are widely used in gene delivery, but they change the cell membrane potential and induce cytotoxicity by rupturing the membrane 37. When cationic liposomes enter the blood circulation, they bond with negatively charged plasma protein in the blood, resulting in aggregation and sedimentation 18. All these adverse factors impede the clinical application of cationic liposomes. In our study, the combination of LH and cationic liposomes made the surface potential of the carrier switch from positive to negative (Figure ), thus significantly reducing aggregation and precipitation (Figure ). Cationic liposomes, N-Lip, obliviously caused rapid red blood cell rupture. However, LH-coated lipoplex exhibited little hemolytic toxicity at the same concentration range (Figure ). Additionally, LH-Lip significantly reduced the cell apoptosis and necrosis of normal LO2 cells caused by N-Lip (Figure ). The combination of LH and cationic liposomes to form a lipoplex was through electrostatic interactions, which are unstable compared to chemical bonds but require a simple preparation process. The outer LH is easily removed or degraded by heparanase 38, minimizing obstruction of siRNA transfection. BCL-2 silencing results showed that LH-Lip/siBCL-2 effectively transferred siRNA to BXPC-3 cells (Figure ) and tumor tissues (Figure ).

Pancreatic tumor always contains multiple ECM components, which impede the penetration of nanodrugs 39. In our study, tumor priming was used to improve the tumor microenvironment, by inducing tumor cell apoptosis and decreasing the abundance of ECM (Figure ). Tumor priming through PTX-Lip promoted the cellular uptake (Figure ) and delivery (Figure and Figure ) of nanoparticles. Compared to free PTX, PTX-Lip exhibited lower toxicity (Table and Table ) and the same penetration effect on solid tumor (Figure ) at a lower dosage (50 mg/m2).

However, pretreatment with PTX may increase drug resistance or exacerbate cancer metastasis. Drug resistance is regulated via various signaling pathways. The anti-apoptotic protein BCL-2 is overexpressed in PDAC and is related to acquired chemotherapy resistance 40. After pretreatment with PTX-Lip, the expression of BCL-2 increased (Figure and Figure ). However, blockade of BCL-2 activity using siBCL-2 decreased this overexpression and promoted the therapeutic effect of PTX (Figure and Figure ). On the other hand, silencing BCL-2 also played a role in inhibiting cancer cell invasion and metastasis, but this anti-metastasis effect was weak using N-Lip/siBCL-2.

A previous study demonstrated that LH exhibits anti-metastasis activity to melanoma 14. Similarly, LH-coated lipoplex also exhibited obvious anti-metastasis ability to pancreatic cancer in our study. Interestingly, PTX-Lip alone also inhibited cell migration and invasion in vitro because PTX-induced cytotoxicity reduced cell viability. The combination of PTX-Lip and LH-Lip/siBCL-2 obviously inhibited BXPC-3 cell migration and invasion (Figure ). Furthermore, the adhesion between BXPC-3 cells and platelets was inhibited using LH-Lip/siBCL-2, resulting in inhibition of the EMT process (Figure ). In addition, LH-Lip/siBCL-2 effectively reduced the formation of cancer metastases in vivo (Figure ).

Conclusion

In summary, considering the specific tumor microenvironment of PDAC, we developed a novel therapeutic strategy to enhance anti-cancer and anti-metastasis efficacy. First, PTX-Lip was used as a tumor-priming agent to induce tumor cell apoptosis and decrease the abundance of ECM, resulting in enhanced the delivery of lipid-siRNA complex. Additionally, PTX-Lip showed lower toxicity and the same penetration effect on solid tumor at a lower dosage (50 mg/m2). Then, we further designed a LH-coated lipid-siRNA complex to load BCL-2 siRNA (LH-Lip/siBCL-2). This LH-coated lipoplex could not only reduce the cytotoxicity and poor stability of the cationic carriers, but also inhibit cancer metastasis in vivo. Sequential delivery of PTX-Lip and LH-Lip/ siBCL-2 is a promising therapeutic approach for PDAC or other tumors with rich ECM and high metastatic ability.

Supplementary figures.

Click here for additional data file.

Table 1

Blood cell levels in mice after treatment with different PTX formulations (n = 3).

Groups	WBC (109/L)	PLT (109/L)	Gran (109/L)
PBS	3.63 ± 0.25	398.3 ± 15.0	1.57 ± 0.15
PTX-Lip (50 mg/m2)	3.67 ± 0.15	366.3 ± 15.0	1.37 ± 0.32
PTX-Lip (100 mg/m2)	3.50 ± 0.10	282.7 ± 11.8*	1.20 ± 0.20
Free PTX (100 mg/m2)	0.97 ± 0.35*	250.0 ± 9.54*	0.43 ± 0.15*

WBC: white blood cells; PLT: platelets; Gran: neutrophile granulocyte.

*P < 0.001 vs. PBS group.

Table 2

Serum biomarkers in mice after treatment with different PTX formulations (n = 3).

Groups	ALT (U/L)	ALP (U/L)	BUN (mM)	CRE (μM)
PBS	42.25 ± 4.03	114.1 ± 12.0	6.78 ± 0.35	34.5 ± 2.12
PTX-Lip (50 mg/m2)	40.05 ± 4.17	106.3 ± 16.3	6.80 ± 0.67	37.0 ± 2.83
PTX-Lip (100 mg/m2)	114.2 ± 4.88**	140.8 ± 9.69*	6.46 ± 0.71	32.3 ± 3.18
Free PTX (100 mg/m2)	369.8 ± 3.11***	191.5 ± 16.3*	6.66 ± 0.96	34.5 ± 2.12

ALT: glutamic-pyruvic transaminase; ALP: alkaline phosphatase; BUN: urea nitrogen; CRE: creatinine.

*P < 0.05, **P < 0.01, and ***P < 0.001 vs. PBS group.