Performance of Pyridylthiourea‐Polyethylenimine Polyplex for siRNA‐Mediated Liver Cancer Therapy in Cell Monolayer, Spheroid, and Tumor Xenograft Models

Medical application of siRNAs relies on methods for delivering nucleic acids into the cytosol. Synthetic carriers, which assemble with nucleic acids into delivery systems, show promises for cancer therapy but efficiency remains to be improved. In here, the effectiveness of pyridylthiourea‐polyethylenimine (πPEI), a siRNA carrier that favors both polyplex disassembly and endosome rupture upon sensing the acidic endosomal environment, in 3 experimental models of hepatocellular cancer is tested. The πPEI‐assisted delivery of a siRNA targeting the polo‐like kinase 1 into Huh‐7 monolayer produces a 90% cell death via a demonstrated RNA interference mechanism. Incubation of polyplex with Huh‐7 spheroids leads to siRNA delivery into the superficial first cell layer and a 60% reduction in spheroid growth compared to untreated controls. Administration of polyplexes into mice bearing subcutaneous implanted Huh‐7Luc tumors results in a reduced tumor progression, similar to the one observed in the spheroid model. Altogether, these results support the in vivo use of synthetic and dedicated polymers for increasing siRNA‐mediated gene knockdown, and their clinical promise in cancer therapeutics.

The cell line was grown in an adherent state onto a plastic substrate (175 cm 2 Falcon tissue culture flask) in high-glucose Dubelcco's modified eagle medium (DMEM) supplemented with 10 % fetal bovine serum (FBS) (Perbio, Brebières, France), 1 % Non-essential amino acids, 50 U/mL penicillin G and 50 μg/mL streptomycin. When needed, cells were suspended in solution by a trypsin treatment. The siRNA were purchased from Eurogentec (Seraing, Belgique), annealed at 90 µM (or 1.23 µg.µL -1 ) and stored in aliquots at -80°C. The sense (S) and antisense sequences (AS) of the siRNA duplexes were as followed. Untargeted siRNA (siC): 5'-GAUU AUGU CCGG UUAU GUAU U (S) and 5'-UACA UAAC CGGA CAUA AUCU U (AS). The siC sequences are the ones of Luc-U that was described by Judge et al. [2] . Important to mention, this 2'OMe modified siRNA did not target and silence the luciferase gene of Huh-7luc cells. It did not trigger as well an immune response [2] .

In vitro siRNA delivery into 2D culture
The Huh-7 cells were seeded the day before the experiments. Transfection assays were performed either in 6-well plate (Costar, 100000 cells/well, 2 mL medium) or in 12-well plate (Costar, 50000 cells/well, 1 mL medium). This procedure is described for assays at final concentrations of 100 µM PEI and 20 nM siRNA. These concentrations correspond to an ethylenimine Nitrogen (N) to siRNA Phosphate (P) N/P ratio of 125. In experiments using other PEI concentrations, the PEI input was scaled accordingly. For polyplex formation, the 100 µM siRNA stock solution was freshly diluted in 4.5% glucose solution to 2.2 µM. This 2.2 µM siRNA solution (20 µL) was then rapidly added to 200 mM PEI (1.1 µL). After a gentle vortex, the mixture was diluted with DMEM (180 µL) and immediately added into wells by dilution into the complete cell culture medium. Added volumes were 200 µL for the 6-well plate, 100 µL for the 12-well plate. The living cells were stained with Hoechst 33342 for nuclei imaging by incubation into cell culture medium containing serum for 30 minutes (1 µg/mL final concentration). The cell culture was carefully removed and replaced with PBS for immediate imaging.  The experiments were carried out in quintuplicate in 96-well plate using the spheroids as previously described. The siRNA polyplexes as previously prepared in 4.5% glucose were diluted 20 times in complete cell culture medium. The polyplexes N/P 14 (11 µL) were then immediately added to each spheroid by further dilution into the cell culture medium (100 µL).
Final concentrations were 180 nM siRNA and 100 µM PEI. The cell culture medium was replaced at days 2, 5 and 7 with complete cell culture medium containing polyplexes.
The spheroid growth was examined using a Nikon Eclipse TS100 microscope and a 10 X objective. Images were recorded at days 0, 2, 5, 7 and 9 with a Nikon Color CCD DS-Fi2 Camera and a DS-L3 camera controller. The area (A) of each spheroid was measured using the NIH Image J software and reported in µm 2 . The volume (V) was then calculated using the formula: V = 4/3(A) 3/2 The MCTS growth (G) was calculated relative to the initial volume (V0; day 0). An ANOVA statistical analysis followed by a Bonferroni's test was performed between groups at the days 7 and 9. Comparison showed statistic differences (p < 0.005) between the siPLK/PEI group and the controls (siC/PEI or untreated groups). No statistical variation was seen between the two controls.

Subcutaneous hepatocellular carcinoma model
The subcutaneous and orthotopic cell-Derived Xenograft HCC mouse models have been previously characterized and reported. [3] The Huh-7 cell line (Japanese Collection of Research Bioresources Cell Bank, Osaka, Japan) was stably transformed to express a firefly luciferase by transduction of firefly luciferase-encoding pCLNCX vector (Dr. Lorang, NIH, Bethesda, MD, USA) with a retrovirus. Important to mention, this firefly luciferase was not silenced by SiC. These Huh-7luc cells (10 6 in 50 µL PBS) were subcutaneously injected on the back of NMRI-nu mice. Three weeks later, when the subcutaneous tumors reached at least 5 mm in one dimension, mice were anesthetized by isoflurane gaz inhalation and they were intraperitoneally injected with a solution of D-luciferin (Caliper Lifesciences) (2 mg in 100 µL PBS). The bioluminescence was monitored using an IVIS 50 in vivo imaging system (Caliper Lifesciences, Roissy, France) and, after bioluminescence acquisition, mice were allocated to the different experimental or control groups as described [4] . The polyplexes N/P 14 (20 µg siRNA in 50 µL 4.5% glucose solution) were injected into the tumors of still anesthetized mice. The mice were then let to awake. Longitudinal luminescence monitoring and intratumoral injections were repeated thrice a week for two weeks. Luminescence was calculated using the Living Image 3.1 software (Caliper Lifesciences) and was expressed as photons.second -1 .cm -2 .steradian -1 (p.s -1 cm -2 .sr -1 ). The relative tumor growth (RTG) was calculated by dividing the luminescence at each time point to the initial value. Results were expressed as mean  standard error (SE). One-way analysis of variance (ANOVA) was completed with a post hoc multiple comparison analysis between groups using the Mann-Whitney or Fisher tests. Differences were considered significant at p <0.05.
Alternatively, three weeks after intra-hepatic Huh-7-Luc cell implantation (10 6 in 50 µL PBS), PBS or 20 or 40 µg fluorescent Cy5-siRNA/PEI were intravenously injected. 24h later, mice were injected with luciferin, livers were harvested and separated from the hepatic tumors. The tumor bioluminescence and the fluorescence of the siRNA were then recorded.

Histology
Tumors were harvested from euthanized mice and were fixed in 10% buffered formalin.
Tissues were then embedded in paraffin, sectioned in 7 µm slices and stained with Hematoxylin and Eosin (H&E). The necrosis level was blindly scored from 0 (no necrosis) to 5 (complete necrosis). Figure S1. Cellular response to siRNA-mediated PLK1 mRNA degradation 3 days after addition of siRNA/πPEI onto Huh-7 2D culture. The morphology of the cell nuclei was observed after staining with Hoechst 33342 and fixation with 2.5% glutaraldehyde. Final concentrations were 100 µM PEI and 20 nM siRNA and the polyplexes were added by simple dilution into the complete 10% FBS cell culture medium. Images showed nucleus fragmentation signaling apoptosis. NMRI-Nude mice with orthotopic Huh-7-Luc tumor were intravenously injected with PBS or the Cy5-siRNA/PEI (N/P 14) at doses of 20 µg and 40 µg siRNA/mouse. 24h later, mice were injected with luciferin, livers were harvested and separated from the hepatic tumors. The tumor bioluminescence (upper panels) and then the fluorescence of the siRNA (lower panels) were recorded. Fluorescence is seen in the liver and in the tumors with clear colocalization with bioluminescence (arrows) in part of tumor. Figure S4. Quantification of relative tumor growth following administration of polyplexes in a mice model of hepatocellular cancer. The indicated polyplexes containing 20 µg nucleic acids were repeatedly injected into tumors over two weeks at days 0, 2, 4, 7, 9 and 11. A. Histograms A and B report the relative tumor growth following administration of siRNA/PEI polyplexes (N/P10) or sticky siRNA/PEI polyplexes, respectively. The sticky siRNA is supposed to facilitate siRNA delivery [5]