Intraperitoneal siRNA nanoparticles for augmentation of gemcitabine efficacy in the treatment of pancreatic cancer

Abstract

Pancreatic ductal adenocarcinoma is a deadly disease with limited treatment options due to late diagnosis and resistance to conventional chemotherapy. Among emerging therapeutic targets, the CXCR4 chemokine receptor and polo-like kinase 1 (PLK1) play critical roles in the progression, metastasis and chemoresistance of pancreatic cancer. Here, we tested the hypothesis that combining CXCR4 inhibition by a polymeric CXCR4 antagonist PAMD-CHOL with PLK1 knockdown by siRNA, will enhance the therapeutic effect of gemcitabine in orthotopic model of metastatic pancreatic cancer. We formulated nanoparticles with cholesterol-modified PAMD and siPLK1 and found strong synergism when combined with gemcitabine treatment in vitro in both murine and human pancreatic cancer cell lines. Biodistribution of the nanoparticles in orthotopic pancreatic cancer models revealed strong accumulation in primary and metastatic tumors, with limited hepatic disposition. The cholesterol-containing nanoparticles showed not only increased tumor accumulation than the cholesterol-lacking control but also deeper penetration to the tumors. In a therapeutic study in vivo, the triple combination of PAMD-CHOL/siPLK1 and gemcitabine showed superior anticancer activity when compared with single and dual combination controls. In conclusion, PAMD-CHOL/siPLK1 nanoparticles synergistically enhance anticancer activity of gemcitabine in pancreatic cancer and represent a promising addition to the treatment arsenal.

Graphical Abstract

INTRODUCTION

Pancreatic ductal adenocarcinoma (PDAC) is the fourth leading cause of cancer-related death with current survival rates about 8%.¹ PDAC may become the second leading cause of cancer deaths by 2030.² The early stage of PDAC usually goes undiagnosed because of the lack of symptoms and thus in most cases patients present with advanced stages at the time of diagnosis, making them ineligible for curative surgical resection.^{3, 4} Only 20% of the PDAC patients are eligible for surgery and even in those cases, 80% develop metastases.^{5, 6} Chemotherapy is thus the main treatment method, including the drug cocktail FOLFIRINOX, gemcitabine (GEM), and albumin-bound paclitaxel (Abraxane).^{7, 8} GEM is a nucleoside analog of deoxycytidine ⁹ that is phosphorylated in cells, incorporates into new DNA strands, and causes apoptosis.^{10, 11} However, GEM failed to increase the survival rate in PDAC because most patients develop resistance.^{12, 13} Therefore, it is urgent to find new strategies to reverse the GEM resistance and improve the overall therapeutic effect.

Multiple targets have been explored to overcome GEM resistance in PDAC. CXCR4 is a chemokine receptor that is overexpressed in many cancers, including PDAC.^{14, 15} The tumor cells can directionally migrate to different organs based on the chemotactic gradient of the CXCR4 ligand CXCL12.¹⁶ The CXCR4 pathway plays an important role in the migration of pancreatic cancer cells^{17, 18} and development of immunosuppressive tumor microenvironment.¹⁹ Evidence also suggests CXCR4 involvement in PDAC GEM resistance by activating FAK, ERK, and Akt signaling pathways, increasing the expression of survival proteins, and enhancing transcription of β-catenin and NF-κB.^{20, 21} Polo-like kinase 1 (PLK1) is a family member of serine and threonine protein kinases that is highly expressed in PDAC^{22, 23} and high PLK1 expression is correlated with GEM resistance.^{24, 25} The resistance is related to the phosphorylation of origin recognition complex 2 (Orc2) by PLK1 to facilitate DNA replication, phosphorylation of Hbo1 (Orc1) that increases the expression of cFos, and the increased level of cFos affecting the target multidrug resistance 1 (MDR1).²⁶ In the absence of effective small molecule inhibitors, the use of siRNA against PLK1 is a viable option as long as the challenges of delivery to desmoplastic PDAC tumors can be overcome.^27-30 Overall, CXCR4 and PLK1 represent suitable targets to overcome GEM resistance in PDAC that function by affecting complementary signaling pathways.

The goal of this study was to evaluate if simultaneous inhibition of CXCR4 and knockdown of PLK1 sensitizes PDAC to the treatment with GEM. To test the hypothesis, we have used combination polyplexes consisting of CXCR4-inhibiting polymer (PAMD-CHOL) and siRNA against PLK1.^31-33 We evaluated the gene silencing ability, anti-migration and anti-proliferation effect of the polyplexes with and without GEM in vitro in mouse KPC8060 and human S2-013 PDAC cell lines. The biodistribution of the polyplexes after intraperitoneal administration was tested in both mouse and human orthotopic PDAC models before proceeding to the evaluation of the anticancer efficacy of the combination therapy.

MATERIALS AND METHODS

Materials.

N,N’-Hexamethylenebisacrylamide (HMBA) was purchased from Pfaltz & Bauer (Waterbury, CT). AMD3100 was purchased from BioChemPartner (Shanghai, China). N,N-diisopropylethylamine (DIPEA) was from Acros Organics (New Jersey, US). Cholesteryl chloroformate were from Sigma-Aldrich (St. Louis, MO). Dulbecco’s Modified Eagle Medium (DMEM), Fetal Bovine Serum (FBS), penicillin/streptomycin (Pen-Strep), Dulbecco’s Phosphate Buffered Saline (PBS) were from Thermo Scientific (Waltham, MA). Carboxyfluorescein (FAM) labeled siRNA (FAM-siRNA, sense sequence 5’-6FAM AUGAACGUGAAUUGCUCAAUU-3’), Cy5.5 labeled siRNA (Cy5.5 siRNA, sense sequence 5’-Cy5.5-AUGAACGUGAAUUGCUCAAUU-3’), negative control siRNA that nonspecific to any gene (siNC, sense sequence 5’-AUGAACGUGAAUUGCUCAA-3’), siRNA targeting mouse PLK1 (siPLK1, sense sequence 5’-CAACACGCCUGAUUCUCUAdTdT-3’), siRNA targeting human PLK1 (siPLK1, sense sequence 5’-UAAGGAGGGUGAUCUUCUUCAdTdT-3’) were from Dharmacon (Lafayette, CO). GEM was purchased from the hospital compounding pharmacy at the University of Nebraska Medical Center.

Polymer synthesis and characterization.

PAMD and PAMD-CHOL were synthesized as previously described^{32, 34} by Michael-type polyaddition. Equal molar ratio of AMD3100 and HMBA were dissolved in 70% MeOH (v/v) and reacted for 3 days at 37°C under nitrogen. After dialyzing the reaction mixture against methanol (3500 MWCO) for 3 days, the solvent was evaporated. Then, PAMD (200 mg) and DIPEA (62.2 mg) were dissolved in anhydrous methylene chloride and cholesteryl chloroformate (43.2 mg) solution in methylene chloride was added drop wise to the ice-cold solution over 1 h. After 24 h, the mixture was dialyzed against ethanol/water (v/v 1/1) by adjusting pH to 4.0 followed by dialyzing under water for 2 days. The polymer structure and the content of cholesterol were analyzed by ¹H-NMR using Varian INOVA (400 MHz). Fluorescently labeled polymers were obtained by conjugating Cy3 NHS ester following the manufacturer’s protocol for protein labeling.

Preparation and characterization of polyplexes.

Polyplexes (polymer/siRNA) were prepared by mixing 1:1 volume ratio of polymer and siRNA in 20 mM HEPES (pH 7.4) and incubated at room temperature for 30 min. The siRNA condensing ability of polymers was evaluated by agarose gel electrophoresis assay. Different w/w ratio polymer/siRNA were loaded to 2% agarose gel (0.5 μg/mL ethidium bromide) and ran at 110 V in 0.5 × Tris/Borate/EDTA buffer for 15 min. The gel was imaged with E-Gel Imager (Life Technologies, CA). Hydrodynamic size and zeta-potential of the polyplexes were measured by Dynamic light scattering (DLS) using a NanoBrook Omni (Brookhaven Instruments, NY). The morphology of polyplexes was tested by TEM (Tecnai G2 Spirit, FEI Company, USA). For stability study, polyplexes were incubated with freshly isolated mouse ascites at 37 °C. The hydrodynamic diameter of polyplexes was measured at different time point.

Cell culture.

KPC8060 pancreatic cancer cell line derived from the genetically engineered PDAC mouse model (Kras^LSL−G12D/+; Trp53^{LSL−R172H/+}; Pdx-1-Cre) was a kind gift from Dr. Hollingsworth, University of Nebraska Medical Center. S2-013 human pancreatic tumor cell line derived from a liver metastasis was kindly provided by Dr. Singh, University of Nebraska Medical Center. Both cell lines were cultured in high-glucose DMEM with 10% FBS and Pen-Strep (100 U/mL, 100 μg/mL) at 37 °C with 5% CO₂ in a humidified incubator.

Cytotoxicity.

Cytotoxicity of the tested agents was determined in KPC8060 and S2-013 cell lines by CellTiter-Blue Cell Viability Assay (Promega, Madison, WI). The cells (8000 cells/well) were seeded in 96-well plate and incubated overnight before incubation with different concentrations of the test articles. The medium was removed after 24 h and the CellTiter-Blue reagent added for 2 h incubation. The fluorescence intensity [FI] was measured at λ_ex/λ_em= 560/590 nm by a microplate reader (Molecular Devices, CA). The relative cell viability (%) was [FI]_treated/[FI]_untreated × 100. IC₅₀ was calculated by GraphPad Prism. Combination effect of the GEM and polyplex treatment was analyzed using CompuSyn software. The calculated combination index (CI) indicated synergism (CI < 1), additive effect (CI = 1), or antagonism (CI > 1).

Transwell migration.

KPC8060 and S2-013 cells were trypsinized, washed, and resuspended in serum free medium. KPC8060 cells (80,000) or S2-013 cells (100,000) suspended in 300 μL serum-free media were added to transwell inserts (8.0 μm pores, Fisher Scientific, Pittston, PA) after pretreatment with different polymer or AMD3100 concentrations for 30 min. The lower chamber was immersed in 600 μL of serum-free media with or without 20 nM CXCL12. After incubating for 16 h, the cells on the top of the inserts were removed using cotton swabs. The cells at the bottom of the inserts were fixed in methanol, stained with 0.2% Crystal Violet, and counted using EVOS xl microscope under 20 × magnification.

Cell uptake.

KPC8060 or S2-013 cells were seeded in 8-well glass chamber (Catalog #: 155409, Fisher Scientific) and incubated overnight. Polyplexes prepared with fluorescently labeled siRNA (FAM-siRNA, 100 nM, w/w = 2.5) were added to each well for 4 h, washed with PBS, and fixed in 4% paraformaldehyde. The cells were stained with Hoechst 33342 and imaged using confocal microscopy (LSM800 Laser Scanning Microscope, Zeiss, Jena, Germany). For cell uptake by flow cytometry, the cells were seeded in 12-well plate, treated for 4 h, washed with PBS, trypsinized, resuspended in PBS, and analyzed by FACS Calibur (BD Bioscience, Bedford, MA).

Spheroid penetration.

KPC8060 cells (10,000) were seeded in ultralow attachment 24-well plates (Corning, NY) and cultured for 7 days. When the diameter of the formed spheroids reached ~400 μm, they were incubated with polyplexes containing FAM-siRNA (200 nM, w/w = 2.5) for 12 h. The spheroids were washed twice using PBS and imaged using confocal microscope with the Z-stack 10 μm. Transformation of the florescence images by ImageJ was used to create surface plots.

Real-time polymerase chain reaction (RT-PCR).

5 × 10⁴ KPC8060 cells or 1 × 10⁵ S2-013 cells were seeded in 6-well plate. Cells were incubated with polymer/siNC or siPLK1 polyplexes (100 nM, w/w = 2.5) in serum-free medium for 4 h, followed by GEM (25 nM) for another 24 h. After removing serum-free medium, cells were incubated with culture medium for 24 h. Total RNA were extracted following protocol. The mRNA expression level of PLK1 was quantified by SYBR Green RT-PCR. 0.5 μg siRNA was transcribed to cDNA using High Capacity cDNA transcription kit (Applied Biosystems). PCR reaction was conducted on Rotor-Gene Q (QIAGEN) equipment with iTaq Universal SYBR Green Supermix (Bio-Rad) and following primers. mGAPDH primers (forward: 5’-CAATGACCCCTTCATTGACC; reverse: 5’-GATCTCGCCCTGGAAGATG), hGAPDH primers (forward: 5’-ACAGTTGCCATGTAGACC; reverse: 5’-TTGAGCACAGGGTACTTTA), mPLK1 primers (forward: 5’-CTCAATAAAGGTGTGGAGAAC; reverse: 5’-TGTAGCAAGTCACTAAGGTG) and hPLK1 primers (forward: 5’-ATTTCCGCAATTACATGAGC; reverse: 5’-TCCTGGAAGAAGTTGATCTG). mRNA levels were calculated relative to internal control based on the comparative threshold value (Ct) method.

Colony formation.

KPC8060 cells (200/well) or S2-013 cells (600/well) were seeded in 12-well plates and incubated overnight. The cells were treated with PAMD-CHOL/siPLK1 or PAMD-CHOL/siNC polyplexes (100 nM, w/w=2.5) for 24 h, followed by GEM (25 nM) for another 24 h. The medium was replaced with fresh one and the cells cultured for another 7 days, when they were fixed in 4% paraformaldehyde, stained with 0.2% Crystal Violet, and washed with PBS. The areas of the formed colonies were determined using ImageJ software.

Cell cycle.

KPC8060 cells (50,000/well) or S2-013 cells (100,000/well) were seeded in 6-well plate and incubated overnight. The cells were treated with the polyplexes (100 nM siRNA, w/w=2.5) for 24 h, followed by GEM (25 nM) for another 24 h. The cells were trypsinized, washed with PBS, and fixed in 70% ethanol for 1 h at 4 °C. The fixed cells were stained with FxCycle™ PI/RNase Staining Solution (Catalog # F10797, Thermo Fisher Scientific) for 30 min at room temperature. Cells in different phase of cell cycle were analyzed by FACS Calibur (BD Bioscience, Bedford, MA).

Orthotopic PDAC model.

Animal experiments followed the protocol approved by the University of Nebraska Medical Center Institutional Animal Care and Use Committee. Male C57BL/6 mice (7 weeks old, Charles River Laboratories) or male athymic nude mice (NCr-nu/nu, 7 weeks old, Charles River Laboratories) were anesthetized by IP injection of ketamine/xylazine. The incision was made in the peritoneum at the left-center abdomen region and 2.5 × 10⁴ of KPC8060 cells or 5 × 10⁵ of S2-013 cells in 40 μL Matrigel/PBS (1:1) solution were injected into the pancreas. 5-0 chromic catgut and soft staples were used to close the abdomen. The staples were removed 10 days after surgery.

Biodistribution.

After implantation of tumor cells, ultrasound was used to monitor tumor growth and the tumor volume was quantified by Vevo LAB software. When the tumor volume reached ~200 mm³, the mice were injected intraperitoneally with fluorescently labeled PAMD/siRNA and PAMD-CHOL/siRNA (n=3, 20 μg siRNA/mouse, w/w=2.5). The polymers were labeled with Cy3 and siRNA was labeled with Cy5.5. Mice were sacrificed 24 h after injection and tumors and major organs were harvested for ex vivo imaging using Xenogen IVIS 200 (Ex=535 nm, Em=580 nm for Cy3 and Ex=675, Em=720 for Cy5.5). The fluorescence intensity from tumors and each organ was analyzed using the provided IVIS instrument software. Primary tumors were embedded in O.C.T compound, cut into frozen sections (10 μm), and stained with DAPI before imaging the intratumoral distribution by confocal microscopy.

Anticancer activity.

Two weeks after the KPC8060 cell pancreas implantation, the mice were randomly assigned into six groups (n=5) and given intraperitoneal injection of PBS, PAMD-CHOL/siNC, PAMD-CHOL/siPLK1, GEM, GEM+PAMD-CHOL/siNC, and GEM+PAMD-CHOL/siPLK1. The mice were injected with polyplexes containing 2.5 mg/kg siRNA at day 10, 12, 15, 17, 20, 22, and 25. GEM (25 mg/kg) was injected at day 14, 19, 24 by intraperitoneal injection. The body weights were recorded every other day and the mice were sacrificed on day 28. Primary tumors were weighed and macroscopic metastases in isolated organs observed by dissecting microscope. Major organs were fixed with 4% paraformaldehyde, embedded in paraffin, sectioned and stained with H&E. Tumors were also collected for RT-PCR assay. The apoptosis of tumor cells was studied by Caspase 3 immunohistochemical staining.

Statistical analysis.

Data are presented as mean ± SD. One-way ANOVA was used to analyze differences among three or more groups. Student’s t-test is used to analyze the significance between two groups. P < 0.05 was considered a minimal level of statistical significance. All statistical analysis was performed with GraphPad Prism 8.

RESULTS AND DISCUSSION

In the last decade, we have been developing a class of polycationic CXCR4 antagonists (PAMD) as vectors for delivery of nucleic acids.^{33, 35, 36} Here, we have used cholesterol-modified PAMD (PAMD-CHOL) that was recently reported as a highly effective vector for intraperitoneal siRNA/miRNA delivery in PDAC.³¹ The PAMD-CHOL and control PAMD were synthesized as described previously.³¹ The structure of PAMD-CHOL was confirmed by ¹H NMR (Figure S1) and the content of cholesterol (17 wt%) calculated based on the ¹H NMR integral intensity of methyl group from cholesterol and aromatic phenylene protons of PAMD. PAMD-CHOL could bind with siRNA to form nanoparticles (polyplexes) at polymer/siRNA w/w ratio ≥ 2 as shown by the agarose gel assay in Figure S3A. The hydrodynamic size of PAMD/siRNA and PAMD-CHOL/siRNA particles was 100 nm and 130 nm, respectively. The zeta potential was 18.8 mV and 10 mV, respectively (Figure S3C). According to TEM of PAMD-CHOL/siRNA nanoparticles shown in Figure S3B, the polyplexes had spherical shape. Moreover, we tested the stability of the two polyplexes in ascites fluid. PAMD-CHOL/siRNA polyplexes were much more stable with size around 300 nm after incubating with ascites for 8 h while the size of PAMD/siRNA polyplexes control increased to 700 nm (Figure S3D).

Inhibition of pancreatic cancer cell migration by PAMD-CHOL

First, cytotoxicity of PAMD-CHOL and PAMD was evaluated in KPC8060 and S2-013 PDAC cell lines (Figure S2A). The IC₅₀ of PAMD-CHOL was 85.6 μg/ml in KPC8060 and 33.2 μg/ml in S2-013 cells. In comparison, PAMD showed IC₅₀ = 29.1 μg/ml in KPC8060 cells and IC₅₀ = 18.5 μg/ml in S2-013 cells, confirming previous observation that the conjugation of cholesterol decreases cytotoxicity of polycations (Figure S2B). This is likely because the conjugation of cholesterol decreases the exposed positive charges that could interact with cell membranes and cause damage.³⁷ PAMD-CHOL showed CXCR4 antagonism in dose-dependent manner with EC₅₀ of 0.12 μg/mL, well below its toxic concentration.³⁴

The effect of CXCR4 inhibition with PAMD-CHOL on the migration of the two PDAC cell lines was tested using CXCL12 as the chemoattractant (Figure 1). PAMD-CHOL effectively inhibited migration of both the mouse KPC8060 and human S2-013 cells. When compared with a small molecule CXCR4 antagonist AMD3100 at optimal and safe concentrations, PAMD-CHOL showed either better (in KPC8060 cells) or the same (in S2-013) inhibitory activity as this positive control.

Figure 1. Inhibition of pancreatic cancer cell migration in mouse KPC8060 (A) and human S2-013 (B) cells.

Cell migration induced by CXCL12 chemokine gradient. Inhibition with 300 nM AMD3100 and 2 μg/ml PAMD-CHOL. Mean # migrated cells ± SD (n=3). ****P < 0.0001.

Nanoparticle uptake in cell monolayers and penetration in multicellular spheroids

The cellular uptake of the nanoparticles in cell monolayers was observed by confocal microscopy and quantified using flow cytometry in both KPC8060 and S2-013 cell lines. After incubation with nanoparticles prepared with fluorescently labeled siRNA (PAMD-CHOL/FAM-siRNA, 100 nM, w/w= 2.5) for 4 h, the cells showed strong intracellular fluorescence, which was significantly higher than in cells treated with the control PAMD/FAM-siRNA nanoparticles (Figure 2A, B). Flow cytometry was then used to quantify the uptake. In both KPC8060 and S2-013 cell lines, PAMD-CHOL/FAM-siRNA showed the greatest percentage of cells that internalized the particles as well as the largest amount of particles per cell as judged from the mean fluorescence intensity (MFI). The MFI in the PAMD-CHOL group was 2 times higher than in the PAMD group in KPC8060 cells (Figure 2C) and 1.4 times higher in S2-013 cells (Figure 2D).

Figure 2. Nanoparticle uptake in cell monolayers and multicellular spheroids.

(A) Confocal microscopy images of (A) KPC8060 and (B) S2-013 cells treated with nanoparticles containing FAM-siRNA (100 nM, w/w=2.5) for 4 h. Scale bar = 20 μm. Flow cytometry quantification of the 4 h cell uptake in the KPC8060 (C) and S2-013 (D) cells. (E) Surface plots of KPC8060 tumor spheroids treated with nanoparticles containing FAM-siRNA (200 nM, w/w=2.5) for 12 h.

To better model nanoparticle uptake in tumors, we have evaluated cell uptake and penetration of the PAMD-CHOL and PAMD particles in 3D tumor spheroids (Figure 2E). The KPC8060 spheroids were treated with PAMD/FAM-siRNA and PAMD-CHOL/FAM-siRNA (200 nM, w/w= 2.5) for 12 h and imaged by confocal microscopy. As shown in Figure 2E, the PAMD-CHOL/FAM-siRNA nanoparticles exhibited higher uptake and more uniform distribution through the spheroid than the cholesterol-free PAMD particles, which were localized mostly at the outer layers. As in the above 2D cultures, the presence of cholesterol improves interactions of the particles with cell membranes, enhancing the overall uptake.^{38, 39} The mechanism of the deep penetration of PAMD-CHOL particles in the spheroids however remains unclear. Diffusion is unlikely to be responsible for the penetration given the size and positive surface charge of the particles, which make the transport through the matrix between the cancer cells extremely challenging. Our working hypothesis relies on transcellular mechanism of transport in which cells in the spheroid periphery take up the nanoparticles, which are then exocytosed and passed on to cells in the center of the spheroids.⁴⁰

Potentiating effect of PAMD-CHOL/siPLK1 nanoparticles on GEM activity

We examined the transfection efficacy of the PAMD-CHOL nanoparticles (w/w=2.5) prepared with siPLK1 and negative control siRNA (siNC). The relative mRNA expression of PLK1 in both cell lines was tested by RT-PCR. The PAMD-CHOL/siPLK1 particles achieved nearly 48% target knockdown in the KPC8060 cells (Figure 3A) and 68% target knockdown in the S2-013 cells (Figure 3A). To evaluate how the combined CXCR4 inhibition and PLK1 knockdown affects sensitivity of the cells to GEM, we have used CellTiterBlue assay to measure the effect of the PAMD-CHOL/siPLK1 and GEM combination on viability of the KPC8060 and S2-013 cells. Both KPC8060 and S2-013 cell lines were relatively resistant to GEM action with IC₅₀ of 4.9 μM and 58 μM, respectively (Figure S2B). In the absence of GEM, the PAMD-CHOL/siPLK1 particles caused about 20% cell death in KPC8060 and about 40% cell death in S2-013. We have used suboptimal GEM concentrations up to 100 nM to evaluate the combination effect. As shown in Figure 3B, the combination of GEM and PAMD-CHOL/siPLK1 showed significantly improved cell killing effect than either treatment alone or a combination of GEM with PAMD-CHOL/siNC control. At 25 nM GEM, ~20% cell killing was observed, which was increased to 40% when PAMD-CHOL/siNC treatment was added to the GEM regimen, validating the beneficial effect of CXCR4 inhibition on GEM activity. As expected, the triple combination of GEM with PAMD-CHOL/siPLK1 showed the highest cell killing with ~60% cell death in KPC8060 cells. As suggested by better PLK1 knockdown, the combination performed even better in the S2-013 cells where it showed nearly 80% cell death. In contrast, the GEM and PAMD-CHOL/siNC combination showed ~40% cell death and GEM alone only achieved ~ 20% cell killing. These data confirm the potentiating effect of the combined CXCR4 inhibition and PLK1 knockdown on anticancer activity of GEM. To determine whether the combination effect was synergistic, we calculated combination index (CI). The GEM and PAMD-CHOL/siPLK1 combination had CI = 0.17 in the KPC8060 cells and CI = 0.09 in the S2-013 cells (Figure S2C). The CI values show strong synergistic effect in both PDAC cell lines. The combination effect was further validated by apoptosis analysis using DAPI staining and evaluation of the nuclear morphology (Figure S5). Cells treated with the GEM and PAMD-CHOL/siPLK1 combination had more apoptotic nuclei than the control treatments. In conclusion, the PAMD-CHOL/siPLK1 nanoparticles potentiated PDAC cells to the anticancer activity of GEM in a strongly synergistic manner.

Figure 3. Combination effect of PAMD-CHOL/siPLK1 and GEM in KPC8060 and S2-013 cell lines.

(A) PLK1 knockdown by PAMD-CHOL/siPLK1 nanoparticles in KPC8060 and S2-013 cells (100 nM siRNA, w/w=2.5, 48 h). (B) Cell viability after treatment with PAMD-CHOL/siPLK1 (100 nM, w/w=2.5) for 48 h and different concentrations of GEM for 24 h. (C) Colony formation assay in cells treated with PAMD-CHOL/siPLK1 (100 nM, w/w=2.5) for 48 h and GEM (25 nM) for 24 h. Colony areas were quantified by ImageJ software. All data are shown as mean ± SD (n=3). * P < 0.05, ** P < 0.01, *** P < 0.001, **** P < 0.0001.

The effect of the combination treatment on tumorigenic potential of the cancer cells was investigated by the colony formation assay (Figure 3C). CXCR4 inhibition alone (G2 - PAMD-CHOL/siNC) had negligible effect on the colony formation but PLK1 knockdown with PAMD-CHOL/siPLK1 particles (G3) showed strong effect on the colony formation in both KPC8060 and S2-013 cells. As expected, GEM (G4) alone showed about 40% and 60% decrease in the colony formation as evaluated by the area of colonies in KPC8060 and S2-013 cells, respectively. Interestingly, the combination treatment with GEM and PAMD-CHOL/siPLK1 (G6) showed nearly complete suppression of colony formation in both cell lines. As above, combining GEM with CXCR4 inhibition alone (G5) demonstrated intermediate activity.

GEM treatment and PLK1 knockdown cause apoptosis due to the cell cycle arrest.^{41, 42} To better understand the mechanism of the anticancer activity of our combination, we analyzed cell cycle distribution after the treatment (Figure S4A, B). Compared with PAMD-CHOL/siNC group, PAMD-CHOL/ siPLK1 significantly increased the cell population in G2/M phase in both KPC8060 and S2-013 cell lines. Cells treated with GEM had higher S phase population. Therefore, the cell apoptosis caused by GEM because of S phase cell cycle arrest and the cell death induced by the downregulation of PLK1 expression was due to G2/M phase cell cycle arrest.

Biodistribution of nanoparticles in orthotopic PDAC models

Because of the promising in vitro results with the combination of GEM and PAMD-CHOL/siPLK1, we proceeded to in vivo testing of this treatment modality. First, biodistribution study was conducted to determine the ability of the nanoparticles to accumulate in orthotopic KPC8060 and S2-013 pancreatic tumors. When the primary tumors reached ~200 mm³, dually labeled nanoparticles containing Cy5.5-labeled siRNA and Cy3-labeled PAMD-CHOL were given by intraperitoneal injection and tumor and organ distribution analyzed after 24 h. PAMD/siRNA particles dually labeled as the PAMD-CHOL ones were used as a control to validate the importance of cholesterol for the ability of the PAMD-CHOL polyplexes to penetrate tumors in vivo similarly to the above spheroid penetration experiment. Mice injected with PBS were used as negative control (Figure S6).

As shown in the ex vivo fluorescence images of the siRNA signal in the nanoparticles (Figure 4A), PAMD-CHOL particles achieved greatly enhanced accumulation in the orthotopic pancreatic tumors when compared with the PAMD particles. Quantification of the fluorescence in the tumor and major organs revealed that the PAMD-CHOL particles achieved 3.3-fold enhancement of the siRNA accumulation in the primary tumors when compared with PAMD (Figure 4B). Increased siRNA accumulation was also found in the spleen (2.6-fold) and the liver (1.2-fold). Somewhat unexpectedly, lung also showed significant siRNA accumulation. This is likely the result of the partial clearance of the polyplexes from the peritoneal cavity by the lymphatic system through the thoracic duct before entering blood circulation. At the used tumor stage, metastases to peritoneal organs including liver, spleen, and kidney were already present. Thus, at least part of the increased splenic and hepatic accumulation is explained by the ability of the PAMD-CHOL particles to accumulate in the metastatic lesions found in these organs (Figure 4E), as first shown in our recent study.³¹ The fact that the fluorescence signal at the metastatic lesions was significantly higher than in the surrounding healthy tissues indicated that the PAMD-CHOL/siRNA could limit off-target toxicity of the treatment. The pronounced tumor accumulation is because normal organs are covered with a layer of compact mesothelial cells (mesothelium) which greatly limits the accumulation of the nanoparticle. In contrast, peritoneal tumors and metastases are covered with broken mesothelium, which allows the particle distribution into tumors.³¹

Figure 4. Biodistribution of nanoparticles in orthotopic KPC8060 PDAC model in immunocompetent C57BL/6 mice.

Nanoparticles were prepared with Cy5.5-labeled siRNA and Cy3-labeled polymer (PAMD or PAMD-CHOL) and given by intraperitoneal injection (siRNA 2.5 mg/kg, polymer 6.25 mg/kg) when tumors reached 200 mm³. (A) Ex vivo tumor and organ images of Cy5.5-siRNA fluorescence after 24 h. (B) Semiquantitative analysis of Cy5.5-siRNA fluorescence intensity in the tumors and major organs at 24 h. (C) Tumor/kidney ratio of Cy5.5-siRNA mean fluorescence intensity after 24 h. (D) Tumor/liver ratio of Cy3-polymer mean fluorescent intensity after 24 h. (E) Accumulation of Cy5.5-siRNA in the metastatic tumors in the liver, spleen and kidney. (F) Confocal images of frozen tumor sections (DAPI stained nuclei showed in blue). Scale bar = 20 μm. Data are shown as mean ± SD (n = 3). * P < 0.05, ** P < 0.01.

Non-covalently assembled particles like polycation/siRNA polyplexes are susceptible to disassembly and release of siRNA, which is then rapidly excreted by renal filtration. To estimate the relative stability of the two nanoparticles and how it may affect their ability to accumulate in the pancreatic tumors, we have calculated tumor/kidney (T/K) ratio for the siRNA in the particles (Figure 4C). The T/K ratio was 3.8 for the PAMD-CHOL/siRNA particles but only 1.1 for the PAMD/siRNA particles, clearly suggesting the important role of cholesterol modification on stabilizing the particles in the peritoneum. The retention time of polyplexes in peritoneal cavity is very important for the intraperitoneal treatment as polyplexes are susceptible to clearance from peritoneal cavity by lymphatic drainage into thoracic duct.^{43, 44} Because of the strong tendency of polycations to accumulate in the liver, we have also calculated the tumor/liver (T/L) ratio for the labeled polymer in the nanoparticles (Figure 4D). The T/L ratio was 1.6 for the PAMD-CHOL nanoparticles indicating enhanced tumor selectivity relative to the PAMD particles, which showed preferential liver uptake indicated by the T/L ratio of 0.6.

To evaluate the distribution of the nanoparticles within the tumor, the frozen tumor tissues slides were imaged using confocal microscopy (Figure 4F). In the PAMD/siRNA group, both siRNA and the polymer accumulated only at the tumor periphery. In contrast, both the siRNA and the polymer were found both at the periphery and in the central regions of the tumors in case of PAMD-CHOL/siRNA. Importantly, significant colocalization of the siRNA and PAMD-CHOL fluorescence was observed, suggesting that largely intact nanoparticles were able to penetrate into the tumors.

In addition to the KPC8060 syngeneic mouse model of PDAC, we validated the biodistribution data also in the orthotopic S2-013 human xenograft model. We first established the growth rate of the S2-013 tumors using ultrasound imaging to monitor the tumor growth (Figure S7A). The tumor volume was quantified using the Vevo LAB software (Figure S7B).

When the tumors reached ~200 mm³, biodistribution of dually labeled PAMD-CHOL/siRNA particles was assessed as above. We have observed preferential accumulation of the particles in the primary tumors (Figure 5A) as determined from the presence of the fluorescence of both the siRNA and PAMD-CHOL. As above, accumulation was also observed in metastatic lesions in the spleen and the liver (Figure S7C). Similar T/K ratio for the siRNA (T/K = 3.7) as in the KPC8060 model was observed also in the human S2-013 model. Improved T/L ratio was found for both the siRNA and polymer distribution, with T/L = 2.9 for PAMD-CHOL and T/L = 2.6 for siRNA. Deep tumor penetration of the PAMD-CHOL/siRNA nanoparticles was also confirmed in this tumor model (Figure 5D).

Figure 5. Biodistribution of nanoparticles in orthotopic S2-013 human xenograft PDAC model in athymic nude mice.

(A) Ex vivo tumor and organ images of Cy5.5-siRNA fluorescence after 24 h (nanoparticles prepared with Cy5.5-labeled siRNA and Cy3-labeled polymer (PAMD or PAMD-CHOL) and given by intraperitoneal injection (siRNA 2.5 mg/kg, polymer 6.25 mg/kg) when tumors reached 200 mm³). Semiquantitative analysis of Cy5.5-siRNA (B) and Cy3-polymer (C) fluorescence intensity in the tumors and major organs at 24 h. (D) Confocal images of frozen tumor sections (DAPI stained nuclei showed in blue). Scale bar = 20 μm. Data are shown as mean ± SD (n = 3). * P < 0.05, ** P < 0.01.

Anticancer activity in vivo.

Considering the role of CXCR4 in the PDAC tumor immune microenvironment, we have selected the syngeneic KPC8060 model to conduct the therapeutic efficacy study due to the presence of intact immune system in the test animals. After establishing the orthotopic tumors, mice were treated with seven intraperitoneal doses of PAMD-CHOL/siPLK1 or PAMD-CHOL/siNC (siRNA 2.5 mg/kg, PAMD-CHOL 6.25 mg/kg). Three doses of GEM (25 mg/kg) were given also by intraperitoneal injection, starting after the second nanoparticle dose (Figure 6A). Monitoring of the animal body weight revealed no significant change due to the treatment (Figure S8A). The lack of apparent treatment toxicity was also later confirmed after autopsy from H&E staining of major organs (Figure S9). At day 28 post-tumor implantation, the mice in all the treatment groups were sacrificed and the tumors weighted (Figure 6B and C). Combining GEM with the PAMD-CHOL/siPLK1 treatment (G6) showed the best anticancer activity confirming the benefits of this triple combination. Strong activity was also observed when GEM treatment was combined with PAMD-CHOL/siNC nanoparticles (G5), which only inhibit CXCR4 – confirming that CXCR4 inhibition can potentiate GEM anticancer activity even in vivo. Treatment with the nanoparticles alone (G3) provided intermediate anticancer activity. Moreover, the combination treatment also decreased the metastasis frequency. There was almost no metastasis in the combination treatment group (Figure S8B). The expression of PLK1 mRNA was evaluated using RT-PCR. The PAMD-CHOL/siPLK1 nanoparticles downregulated the expression of PLK1 mRNA nearly 40% (Figure 6D). The apoptosis of tumor cells in different treatment groups was evaluated using caspase-3 immunohistochemical staining (Figure 6E). As expected, the combination of GEM and PAMD-CHOL/siPLK1 showed the greatest extent of apoptosis in the tumors among the tested treatments.

Figure 6. Anticancer activity of the combination treatment with PAMD-CHOL/siPLK1 nanoparticles and GEM in orthotropic KPC8060 PDAC model in immunocompetent C57BL/6 mice.

(A) Treatment scheme of the therapeutic study (2.5 mg/kg siRNA, 6.25 mg/kg PAMD-CHOL, 25 mg/kg GEM). (B) Primary tumor weight on day 28 (* vs. PBS group). Data shown as mean ± SD (n = 5). (C) Primary tumor images on day 28. (D) PLK1 mRNA expression in primary tumors on day 28. Data shown as mean ± SD (n=5) *P < 0.05, ** P < 0.01. (E) Caspase- 3 immunohistochemistry analysis. Scale bar = 100 μm.

CONCLUSION

In this study, we demonstrated that intraperitoneal nanoparticles capable of combined CXCR4 inhibition and PLK1 siRNA knockdown potentiates anticancer activity of GEM in metastatic PDAC. The results show that the overall improvement in GEM activity relies on unique biopharmaceutical and pharmacological properties of the nanoparticles. First, intraperitoneal administration of the particles leads to strong accumulation in primary and metastatic pancreatic tumors, while simultaneously minimizing the typical hepatic distribution observed with similar cationic nanoparticles. Modification with cholesterol has proved to be vital for the observed biodistribution behavior as nanoparticles lacking such modification showed less selective tumor delivery and more pronounced disassembly and premature release of the delivered siRNA. The in vivo biodistribution findings were strong validation of the predictive power of the used 3D spheroid model of PDAC, which showed similar deep penetration for the cholesterol modified nanoparticles. In vitro studies showed that both CXCR4 inhibition and PLK1 knockdown alone can sensitize PDAC cells to GEM, however combining both CXCR4 and PLK1 inhibition resulted in strong synergism. Overall, this study points to the potential of the combination nanoparticles as part of the combination treatment of PDAC. Our future studies will focus on dissecting the mechanism of the sensitizing effect and on evaluating the treatment in additional PDAC and other peritoneal cancer models.