The pH-Triggered Triblock Nanocarrier Enabled Highly Efficient siRNA Delivery for Cancer Therapy.

Small interfering RNA (siRNA) therapies have been hampered by lack of delivery systems in the past decades. Nowadays, a few promising vehicles for siRNA delivery have been developed and it is gradually revealed that enhancing siRNA release from endosomes into cytosol is a very important factor for successful delivery. Here, we designed a novel pH-sensitive nanomicelle, PEG-PTTMA-P(GMA-S-DMA) (PTMS), for siRNA delivery. Owing to rapid hydrolysis in acidic environment, PTMS NPs underwent hydrophobic-to-hydrophilic transition in endosomes that enabled combination of proton sponge effect and raised osmotic pressure in endosomes, resulting in vigorous release of siRNAs from endosomes into cytosol. In vitro results demonstrated that PTMS/siRNA complexes exhibited excellent gene silencing effects in several cell lines. Their gene silencing efficiency could reach ~91%, ~87% and ~90% at the N/P ratio of 50/1 in MDA-MB-231, A549 and Hela cells respectively, which were better than that obtained with Lipofectamine 2000. The highly efficient gene silencing was then proven from enhanced siRNA endosomal release, which is mainly attributed to pH-triggered degradation of polymer and acid-accelerated siRNA release. In vivo experiments indicated that NPs/siRNA formulation rapidly accumulated in tumor sites after i.v. injection. Tumor growth was effectively inhibited and ~45% gene knockdown efficacy was determined at the siRRM2 dose of 1mg/kg. Meanwhile, no significant toxicity was observed during the whole treatment. We also found that PTMS/siRNA formulations could lead to significant gene silencing effects in liver (~63%) and skin (~80%) when injected by i.v. and s.c., respectively. This research work gives a rational strategy to optimize siRNA delivery systems for tumor treatments.

Introduction

Since its discovery in 1998 1, RNA interfering (RNAi) provides huge potentials in treatment of many kinds of diseases, such as pan class="Disease">cancers 2, n>an class="Disease">glaucoma 3, hyperlipidemia 4 and skin diseases 5. As the most important effector in RNAi world, small interfering RNA (siRNA) however faces great challenge in realizing such potentials, and the biggest barrier is the lack of effective delivery carriers for siRNA delivery 6, 7. Recent years, many researchers have attempted to solve the problem and a few delivery nanoparticles (NPs) such as polymer-based 8, 9, lipid-based 10, 11 and conjugation-based vehicles 12 were fabricated and exploited. About twenty nano-based siRNA drugs are under the clinical trials at present 6, 13 (and summarized from https://clinicaltrials.gov), which makes siRNA therapies hopeful in treating human diseases.

Cationic pan class="Chemical">polymers dispn>lay advantages for siRNA delivery, such as low cost, simpn>le to prepn>are and high delivery potency 14, 15. In 2008, the first cationic n>an class="Chemical">polymer based siRNA drug (CALAA-01), targeted to rrm2 gene, was used for cancer therapy in clinical trial, and the phase Ia/Ib data was published few years later 16. Another type of polymer, Dynamic PolyConjugate (DPC), had been deployed for delivery of siRNA tackling HBV infection since 2013 17 and approved in clinical trial in 2014 (https://clinicaltrials.gov). These developmental milestones provided great affirmation for polymers in siRNA therapies and deeply encouraged their further exploration.

In past several years, a few works revealed that siRNA endosomal escape seemed to be more important than endocytosis for successful delivery. Sahay et al. 18 found that ~70% internalized siRNAs, delivered by LNP (pan class="Chemical">lipid nanopn>articles), were exocytosed to extracellular milieu from endosomes/lysosomes, resulting in poor gene silencing effects. However, LNP significantly enhanced silencing efficiency in n>an class="Disease">NPC1-deficient cells, which was believed to increase siRNA retention time in endosomes/lysosomes from knockout of NPC1 gene. Almost the same time, Gilleron et al. 19 demonstrated that only few internalized siRNAs (1~2%) could be released from endosomes into cytosol, and that was very close to ~3.5% revealed by Wittrup et al. in 2015 20. Also, several pH-responsive nanoparticles have been successfully developed to enhance cytosolic siRNA delivery for tumor therapy in recent years 21-23. These works suggested that enhancing siRNA endosomal release would be one of the feasible ways for improving gene silencing potency.

For this reason, we prepared a novel pH-sensitive nanopan class="Species">micelle using poly(ethylene glycol)-co-poly[(2, 4, 6-trimethoxybenzylidene-1, 1, 1-tris(hydroxymethyl)] ethane methacrylate-co-poly(n>an class="Chemical">dimethylamino glycidyl methacrylate), PEG-PTTMA-P(GMA-S-DMA), termed PTMS (Fig. ). Herein, PTTMA was the hydrophobic core of this amphiphilic nanomicelle. Due to hydrolysis reaction of acetal group, PTTMA polymers underwent hydrophobic-to-hydrophilic transition in acidic buffer, such as pH 5.0 and pH 4.0 24. Considering the acidic microenvironment in tumors, few TTMA based delivery systems of chemotherapy drugs had been reported for effective cancer treatment 24, 25. Therefore, we believed that PTMS NPs would exhibit pH-sensitive property in endosomes/lysosomes (pH ~5.5 26), leading to disassembly of NPs there. Combining the collapse of NPs and proton sponge effect triggered by dimethylamino (DMA) groups 15, osmotic pressure of endosomes was dramatically increased. Meanwhile, a large number of hydrolysis products further raised up the pressure of endosomes 27. Finally, endosomes were swollen, ruptured and siRNAs was heavily released into cytosol, inducing to highly efficient gene knockdown efficiency (Scheme ). To our knowledge, this is the first research work using TTMA-based polymer for siRNA delivery. In this work, we systematically studied gene knockdown efficiency in vitro and in vivo mediated by PTMS NPs, and explained the delivery mechanism. We hope this work will give a clue for rational design of siRNA delivery systems and make contributions in siRNA therapies.

Materials and Methods

Materials

Methoxy poly (ethylene glycol) (mPEG, Mn=2kDa), 1,1,1-tris(hydroxymethyl) ethane, 2,4,6-trimethoxybenzaldehyde, p-toluenesulfonic acid monohydrate, methacryloyl chloride, N,N'-Dicyclohexylcarbodiimide (DCC), 4-dimethylamiopryidine (DMAP), azobisisobutyronitrile (AIBN), glycidyl methacrylate (GMA) and 2-(Dimethylamino) ethanethiol (DMA-SH) were purchased from Sigma-Aldrich. Diethyl ether, tetrahydrofuran (THF), dichloromethane (DCM), triethylamine (TEA) and N,N-Dimethylformamide (DMF) were purchased from Jiangtian company (Tianjin, China). S-1-dodecyl-S′-(α,α′-dimethyl-α″-acetic acid) trithiocarbonate (CTAm) was synthesized according to previous reported 28.

Dulbecco's modified Eagle's medium (DMEM), fetal bovine serum (FBS), lipofectamine 2000 (lipo 2000), trypsin, Opti-MEM, and penicillin-streptomycin were purchased from Invitrogen Corporation (Carlsbad, CA). Agarose was purchased from GEN TECH (Hong Kong, China). Ethidium bromide (EB), 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide (MTT), and dimethyl sulfoxide (DMSO) were also purchased from Sigma-Aldrich. Plk1 antibody was obtained from cell Signaling Technology, Inc. (Danvers, MA) and GAPDH antibody was obtained from Zhongshan Goldenbridge Biotechnology Co., Ltd. (Beijing, China). Cy5-labeled siRNA (Cy5-siRNA), Cy3-labeled siRNA (Cy3-siRNA), negative controlled siRNA (siNC), anti-firefly siRNA (siFL), anti-PLK1 siRNA (siPLK1), anti-ApoB siRNA (siApoB), anti-SCD1 siRNA (siSCD1) and anti-RRM2 siRNA (siRRM2) were supplied by Suzhou Ribo Life Science Co., Ltd. (Suzhou, Jiangsu Province, China), Table to check the detailed sequences. All other reagents were purchased from Sigma-Aldrich unless specially marked.

Synthesis of TTMA monomer

2,4,6-Trimethoxybenzylidene-1,1,1-tris(hydroxymethyl) ethane methacrylate (TTMA) was synthesized according to previous works (Fig. ) 24, 29. Briefly, 1,1,1-tris(hydroxymethyl) ethane (12g) and 2,4,6-trimethoxybenzaldehyde (20g) were dissolved in THF (150ml), then p-hydroxybenzenesulfonic acid (1g) as the catalytic reagent was added into the mixture and the reaction mixture was shaken at room temperature for 24h. Hereafter, the dried product (14.9g) and TEA (7.8g) were dissolved in DCM (50ml), methacryloyl chloride (6.84g) was added drop wise to the solution, following to react at room temperature for 24h. The final product TTMA (13.5g) was purified by column chromatography.

Synthesis of PEG-CTAm

Macroinitiator PEG-CTAm was obtained by following steps (Fig. ). At the presence of DCC (2.06g) and DMAP (1.22g), dried PEG2000 (10g) and CTAm (3.66g) were dissolved in DCM (50ml) for 48h to completely react. The residue was rinsed three times with diethyl ether and PEG-CTAm was obtained.

Synthesis of PEG-PTTMA (PT)

PEG-PTTMA (PT) was synthesized by Reversible Addition-Fragmentation Chain Transfer (RAFT) polymerization according to published paper (Fig. ) 24. RAFT agent PEG-CTAm (236.5mg), TTMA (1281mg), AIBN (3.28g) and DMF (3ml) were added into a Schlink tube, the solution was degassed three cycles and filled with nitrogen. Then mixtures were placed in a thermostatic water bath at 70ºC for 24h under the protection of nitrogen. At last, product was dialyzed against distilled water and lyophilized.

Synthesis of PEG-PTTMA-PGMA (PTM) and PEG-PTTMA-P(GMA-S-DMA) (PTMS)

PEG-PTTMA-PGMA (PTM) was synthesized as the same as PEG-PTTMA polymer but with PEG-PTTMA (758.8mg) and glycidyl methacrylate (GMA) (427mg) as the reactants. To obtain PEG-PTTMA-P(GMA-S-DMA) (PTMS), PTM (1g) and 2-(dimethylamino) ethanethiol (DMA-SH) (0.5g) were dissolved in DMF (10ml) to react for 24h at 50ºC under the protection of nitrogen (Fig. ). PTMS polymer was lyophilized after dialysis. To verify the structures, PT, PTM and PTMS polymers were analyzed by 1H-NMR spectra using Varian Unity spectrometer (300 MHz) (Fig. ).

Agarose gel retardation assay

Gel retardation was tested to evaluate siRNA binding ability of PTMS. As we did in previous work 30, siNC (0.3ug) was incubated with PTMS solutions at room temperature for 20min to achieve various N/P ratios from 1:1 to 8:1. N/P ratio means molar ratio of amino groups (N) of polymer to phosphate groups (N) of siRNA. The NPs/siRNA complexes were adjusted to final volume (16μl) and mixed with 6×loading buffer (4μl, Takara Biotechnology, Dalian, Liaoning Province, China). The mixture was loaded onto 2% agarose gel containing 5μg/ml ethidium bromide (EB). Electrophoresis was carried out at a voltage of 120mV for 20min in 1×TAE running buffer. Finally, the results were recorded at UV light wavelength of 254nm by image master VDS thermal imaging system (Bio-Rad, Hercules, CA).

Nile red release profile in different pH buffer

PTMS nanomicelles were prepared via self-assembly processes. Briefly, dried PTMS polymer (20mg) was dissolved in DMSO (1ml) and dialyzed against distilled water (10ml) for 24h. During this time, dialysate was frequently replaced with fresh water every three hours. After that, PTMS nanomicelles were formed and stored in 4℃. To evaluate pH-sensitive property, PTMS (20mg) and Nile Red (50μg) were dissolved in DMSO (1ml) and formed Nile Red loaded micelles after dialyzed against distilled water. Fluorescence intensities of the PTMS NPs were measured at given time-points by Varian fluorescence spectrophotometer in various buffer, pH from 7.4 to 4.0. The hydrolysis behavior of PTMS NPs was estimated by fluorescent changes.

Size and Zeta-potential of PTMS NPs

The sizes and zeta potential of empty NPs and PTMS/siNC complexes were measured by dynamic light scattering (DLS) (Zetasizer Nano ZS, Malvern, U.K.) at a wavelength of 633nm with a constant angle of 173º. PTMS/siNC complexes (contained 2μg siNC) were prepared as reported in our previous works 31. Various PTMS/siNC samples were formed with N/P ratios of 10:1 to 50:1 by electrostatic interaction. Before detecting, the sample was adjusted to the final volume (1ml) using distilled water.

Luciferase assay

For all involved in vitro transfection assays in this work, Lipo 2000/siRNA formulation was used as the positive control. Lipo 2000/siRNA formulation was prepared based on standard manufacturer's instructions but fixed weight/volume ratio of siRNA to Lipo 2000 reagent with 1µg/3µl.

MDA-MB-231-Luc (stably expressing luciferase gene) cells were seeded in 24-well plates with 5×104 cells per well. After 24h, the medium was replaced with Opti-MEM (0.5ml) and added PTMS/siFL complexes with different N/P ratios for another 4h, handling the final siRNA transfection concentration of 50nM. Opti-MEM medium was replaced by fresh DMEM (1ml) and further incubated for 20h. Cells were washed three times with 1×PBS (0.5ml) and lyzed with 1×passive lysis buffer (100μl) per well. After violently shaking for 30 min, cell supernatant (10μl) was collected for luminescence measurements with adding substrates (50μl) by Multi-Mode Microplate Reader (Synergy HT, BioTek, USA).

Cell viability

MDA-MB-231-Luc cells were seeded in 96-well plates at 1×104 cells per well the day before transfection. Cells were treated with various complexes for another 24h at the same siRNA concentration as luciferase assay. Ten microliters MTT solution (5mg/ml in 1×PBS) was added into each well to incubate for 4h at 37℃. The solution was removed gently and replaced with DMSO (50μl). After the formazan crystals was dissolved completely, the absorbance of 540nm (OD540) was measured by Multi-Mode Microplate Reader with a reference wavelength of 650nm (OD650). Untreated cells were marked as control sample (mock). Cell viability was calculated according to following equation:

In vitro gene knockdown measurement

Real time PCR (RT-PCR) and western blot were performed to assess gene silencing effect. For RT-PCR test, MDA-MB-231-Luc, A549 and Hela cells were used. Cells were seeded in 6-well plates at 2×105 cells per well. Twenty-four hours later transfected with PTMS/siPLK1, total RNA was extracted using TRIzol Reagent (Invitrogen) according to standard manufacturer's instructions. After finishing reverse transcription with total RNA (1μg), cDNA (50ng) was quantified by RT-PCR system using SYBR Green PCR Master Mix. Herein, we used GAPDH (or β-Actin) gene as the internal control. All primers used in this work were listed in Table .

For western blot, MDA-MB-231-Luc cells were seeded in 6-well plates at 2×105 cells per well. After 48h treated with PTMS/siPlk1, total protein was collected and the concentration was measured by BCA protein assay kit. Fifty microgram protein was loaded and separated by SDS-PAGE, following transferred to PVDF membranes. Protein coated membranes were blocked with 5% BSA buffer for 1h then incubated with primary antibody (1:1000) overnight at 4℃. PVDF membrane was probed with horseradish peroxidase (HRP)-labeled secondary antibody (1:5000) for another 1h. At last, the membranes were treated with ECL kit (Pierce, Rockford, IL) and imaged by Bio-Rad Universal Hood Ⅱ (Bio-Rad Laboratory, Bossier City, LA).

Flow cytometry detection

MDA-MB-231-Luc cells were seeded in 6 well-plates at 2×105 cells per well. After 24h, the medium was replaced with Opti-MEM and cells were treated with PTMS/Cy5-siRNA complexes for 4h at the final siRNA concentration of 50nM. Cells were digested by 0.25% trypsin, washed three times by cold 1×PBS (1ml), suspended in 1×PBS (400μl), and detected by flow cytometry (Becton Dickinson, San Jose, CA).

To test endocytic pathway of NPs, pharmacological inhibitors were used as did in our previous work 15. Briefly, amiloride (100µM), chlorpromazine (30µM) and genistein (1mM) were added into cells for 0.5h before treatment with several formulations. Four hours later, cells were detected by flow cytometry as described above. Here, 4℃ treated sample was the positive control, and normalized results were presented based on non-pretreated cells.

Confocal observation

For studies of the subcellular localization of NPs/siRNA complexes, MDA-MB-231-Luc cells were seeded into 35mm dishes at 2×105 cells each well. Twenty-four hours later, cells were transfected with PTMS/Cy5-siRNA at the final siRNA concentration of 50nM. After 4h and 10h, cells were washed three times with 1×PBS and stained by Lysotracker Green (indicated endosome/lysosome). Then cells were fixed by paraformaldehyde (4%) for 40min and stained by DAPI (indicated nucleus). At last, cells were imaged by Zeiss confocal microscope (LSM700, Carl Zeiss, Germany).

pH-triggered degradation of PTMS

For pH-sensitive measurement of pan class="Chemical">PTMS NPs, n>an class="Chemical">1H NMR was tested first. Briefly, PTMS polymer (20mg) was dissolved in DMSO-d6 (0.6ml) with 5% trifluoroacetic acid (TFA, 30µl) for 2h, following 1H NMR spectra of sample was recorded and analyzed. To determine degradable behavior directly, UV-spectroscopy was measured at the wavelength of 295nm, indicating the relative amount of 2,4,6-trimethoxybenzaldehyde monomer in given solution. PTMS NPs solution (10µg/ml) was prepared as mentioned above and pH was adjusted to 6.0, 5.0 and 4.0, respectively. The solutions were shaken at 37℃, and wavelength of 290-280nm was monitored at each designed time-point within 36h.

Size changes and in vitro siRNA release

The pH of PTMS NPs solution (10µg/ml) was adjusted to 7.4, 6.0, 5.0 and 4.0, respectively, shaking at 37℃. Size changes were recorded and analyzed by DLS in following 32h. TEM images were also observed to verify the size changes treated in pH 5.0 buffer for 24h.

The pH-triggered siRNA release was performed as did as before 31 with slight modification. pan class="Chemical">PTMS/n>an class="Chemical">Cy5-siRNA (N/P=15) formulation was prepared as mentioned above, with 40µg of the final amount of Cy5-siRNA per sample. The Cy5-siRNA solution (110µl) was added into PB buffer (1.2ml) with different pH values (n=3), and shaken at the speed of 100rpm at 37℃. At designed time-points, sample was centrifuged for 10min at the speed of 15000g, 200µl supernatant was collected for testing and equal volume of fresh buffer was added. Cy5-siRNA fluorescence was detected with 630nm excitation and 680nm emission. Accumulated release of Cy5-siRNA was finally calculated based on calibration curve.

In vivo tumor accumulation and tumor inhibition activity

All involved animals in this work were maintained in Peking University Laboratory Animal Center (an AAALAC accredited experimental animal facility). All procedures were performed in accordance with protocols approved by the Institutional Animal Care and Use Committee (IACUC) of Peking University.

For tumor targeting evaluation, MDA-MB-231 cells (5×106 cells) were suspended in 1×PBS (100μl) and subcutaneously injected in right axillary fossa of female BALB/c nude mice (16-18g). When the tumor grew to about 400 mm3, the mice were randomly divided into three groups. Mice were i.v. injected with 1×PBS, Free Cy5-siRNA, and PTMS/Cy5-siRNA (N/P=15) at the final siRNA dose of 1mg/kg, respectively. Cy5 signals of mice whole body was captured using animal imaging system among 48h. At the end of experiment, mice were sacrificed by cervical dislocation, main organs were harvested and further imaged.

To evaluate in vivo tumor growth inhibition of PTMS formulation, tumor model of female BALB/c nude mice were prepared as described above. When the tumor grew to about 40mm3, mice were randomly divided into four groups and treated with PBS, naked siRRM2, PTMS/siNC (N/P=15) and PTMS/siRRM2 (N/P=15) by i.v. injection, respectively. Mice were injected every three days for 4 times at the dose of 1mg/kg. Body weight was monitored during entire experiment. Tumor volume was recorded every day until day 12 using caliper and calculated with the following formula: tumor volume (mm3) = 0.5 × length × width2. At the end of experiment, rrm2 knockdown efficiency were tested and four kinds of serum cytokines (IL-2, IL-6, IFN-γ and TNF-α) were measured using Luminex Technology (4A Biotech Co., Ltd., Beijing, China).

In vivo biodistribution and gene knockdown measurement

For biodistribution, male C57BL/6 mice (18-20g) were purchased from the Academy of Military Medical Science of China and divided into three groups. Mice were intravenously (i.v.) injected with PBS, free Cy5-siRNA and PTMS/Cy5-siRNA (N/P=15) at the final siRNA dose of 1mg/kg, respectively. At 6h and 24h after treatment, main organs were harvested and imaged using Kodak in vivo imaging system (Kodak In-vivo Imaging System FX Pro, Carestream Health, USA).

Pharmacokinetics study was performed as did in previous work 31. Briefly, male C57BL/6 mice were i.v. injected with PBS, naked Cy3-siRNA and PTMS/Cy3-siRNA (N/P=15) at the dose of 1mg/kg. Blood samples were continuously collected at various time points (1, 5, 15, 30, 60 and 120min) post injection. Cy3 fluorescence signal was detected by Multi-Mode Microplate Reader and normalized results were presented based on the starting point (1min).

To assess in vivo gene silencing efficiency, we selected two kinds of injection ways. One was intravenously (i.v.) administration. Male C57BL/6 pan class="Species">mice were treated with n>an class="Chemical">PTMS/siApoB for 48h. Liver were collected and ApoB mRNA level was detected by RT-PCR. The other one was subcutaneous (s.c.) injection. Mice were administrated with PTMS/siSCD1 for 48h. Skin, at the injected site, was harvested and expression of scd1 gene was tested by RT-PCR. For both experiments, the final injected dose of siRNA was 1mg/kg and the N/P ratio was 15/1.

Statistical analysis

Statistical analysis was taken using GraphPad Prism 5 software. Results were presented as mean ± S.E.M. or mean ± S.D. For statistical comparison, One-Way ANOVA followed by Tukey's test was used and P < 0.05 was considered statistically significant.

Results and Discussions

Formation and characterization of NPs

pan class="Chemical">PTMS n>an class="Chemical">co-polymer was synthesized by multistep reactions (Fig. ). The final structural formula was given out using H1-NMR spectra, PEG45-PTTMA30-P(GMA-S-DMA)60 (Mw=27770), where the number represented degree of polymerization (DP) (Fig. ) shown by image of agarose gel (Fig. ), siRNA was loaded easily with increasing N/P ratios and completely captured at the N/P ratio of 5. Therefore, PTMS/siNC formulations with N/P ratios above 5/1 were formed and prepared to use. Size and zeta-potential of PTMS/siNC formulations were analyzed by DLS (Fig. ). Interestingly, compared with empty NPs (PTMS) (43.89 ± 0.22nm), the size of PTMS/siNC (N/P=5) was much larger (1052.5 ± 75.66nm). But with N/P ratios further increased, the size gradually decreased, showing about 70-120nm diameters at the higher N/P ratios of 10/1 to 50/1. It was believed that certain surface charge of NPs was helpful to protect NPs from aggregation because of electrostatic repulsion among NPs 31. From empty NPs to PTMS/siNC (N/P=5), zeta-potential declined from +5.35 ± 0.63mV to -0.14 ± 0.35mV, causing disappearance of electrostatic repulsion and heavy aggregation. When the charge increased to around +10 mV at N/P ratios of 10/1 to 50/1, electrostatic repulsion recovered and made NPs/siRNA stable in solutions.

pH-triggered disassembly of PTMS NPs

As reported, pan class="Chemical">acetal groupn> was sufficiently stable in pH 7.4 but prone to be rapn>idly hydrolyzed in acidic buffer 24, 25. We supn>posed that n>an class="Chemical">PTMS NPs were completely destroyed in acidic environment because of hydrophobic-to-hydrophilic transition of hydrophobic core (TTMA) (Fig. , B). To verify this, Tyndall effect was observed by shining a laser beam (Fig. ). As expected, Tyndall effect was significantly weakened when treated with HCl, demonstrating PTMS NPs disassembled in the solution. Nile Red release profile was a popular method to track disassemble behavior of NPs 32, the fluorescence signal of Nile Red was strongly influenced by surrounding environment. In hydrophobic environment, fluorescence signal of Nile Red was very strong, but in hydrophilic environment, fluorescence intensity was too weak to be detected. Here, we prepared Nile Red loaded PTMS to test pH-sensitive feature (Fig. ). Fluorescence signal was not weakened obviously even after 24h of incubation in pH 7.4 buffer, whereas it gradually decreased in acidic solutions, with pH deceased from 6.0 to 4.0. What's more, almost no signal of Nile Red was detected in pH 4 solution after 20h. These results suggested that PTMS NPs were endowed pH-triggered disassembly behavior in solutions.

In vitro gene knockdown effect

To assess in vitro gene pan class="Disease">knockdown efficiency, luciferase assay was performed first. n>an class="CellLine">MDA-MB-231-Luc (stably expressing luciferase gene) cells were transfected with various PTMS/siFL formulations at the final concentration of 50nM (Fig. ). At high N/P ratios (above 20/1), PTMS/siFL complexes exhibited better gene silencing efficiency than Lipo 2000 (30.76 ± 1.74%), even close to 95% at the N/P ratio of 50. We also found that PTMS/siFL (N/P=5) didn't show significant gene silencing activity, that was probably resulted from its heavy aggregation. Importantly, no obvious cytotoxicity was observed for all formulations (Fig. ). To further evaluate delivery potency in vitro, Polo-like Kinase 1 (plk1) gene was selected for RT-PCR measurement. As displayed in Fig. , plk1 gene was dramatically inhibited by PTMS/siPLK1 complexes in MDA-MB-231 cells, consistent with luciferase assay. Compared with Lipo 2000 (41.24 ± 18.33%), the remaining activities of PTMS/siPLK1 polyplexes were much lower, that were 25.75 ± 14.53% (N/P=20), 18.64 ± 7.32% (N/P=30) and 9.01 ± 3.32% (N/P=50). Western blot assay also supported results in MDA-MB-231 cells (Fig. ). Besides, we expanded plk1 gene knockdown tests in other two cell lines. In both A549 and Hela cells, both RT-PCR (Fig. ) and western blot (Fig. ) demonstrated that gene knockdown efficiency mediated by PTMS was much better than Lipo 2000 at high N/P ratios. These results indicated that PTMS nanomicelle was a promising delivery vehicle for siRNA due to its safety and effective gene silencing efficiency in several cell lines.

In vitro cellular uptake

To explore the mechanism why PTMS showed excellent siRNA delivery in vitro, internalization of PTMS/Cy5-siRNA was detected by flow cytometry. Compared with control groups (Lipo 2000 and naked siRNA), the fluorescence peaks of PTMS/Cy5-siRNA complexes were obviously right shifted (Fig. ), reflecting PTMS NPs delivered much more siRNAs into cells. It was showed more clearly in quantified data (Fig. ), all formulations showed about 100% cellular uptake (blue dots), but mean fluorescence intensity (MFI) of PTMS/Cy5-siRNA were about 4.1 times (N/P=50) to 7.5 times (N/P=10) than Lipo 2000 sample. Interestingly, MFIs of PTMS formulations tended to decline with increasing N/P ratio, which seems to be contrary to gene silencing efficiency results showed in Fig. .

To answer this question, endocytic pathway of PTMS/Cy5-siRNA formulation was then evaluated (Fig. ). Endocytic inhibitors, including amiloride, chlorpromazine and genistein, were selected to block macropinocytosis, clathrin-mediated endocytosis and caveolae-mediated endocytosis, respectively. Definitely, PTMS/Cy5-siRNA complexes, with various N/P ratios, showed energy-dependent endocytosis as their cellular uptakes were dramatically hindered under cold shock at 4℃. Interestingly, NPs with different N/P ratios exhibited energy dependence: the higher N/P ratio, the less energy-dependency. Also, these polyplexes showed different inhibitors-dependent endocytosis. Showed in Fig. , both chlorpromazine and genistein significantly impeded cellular uptake of PTMS/Cy5-siRNA polyplexes, indicating clathrin and caveolae-mediated endocytosis were two major endocytic pathways. But NPs with different N/P ratios showed different dependency on these two endocytosis pathways. After treatment with chlorpromazine, the inhibition rates were about 89%, 70% and 38% for N/P=10, N/P=20 and N/P=30, respectively, and the inhibition rates were around 18%, 25% and 23% for genistein treated cells. This demonstrated that dependency on different endocytic pathways may contribute partly to the difference of cellular uptake for NPs of various N/P ratios.

Subcellular localization analysis

To answer the question why less internalization showed better gene silencing efficacy, subcellular localization of pan class="Chemical">PTMS/n>an class="Chemical">Cy5-siRNA was then studied using confocal microscopy (Fig. ). Cy5 signals (red color) of PTMS/Cy5-siRNA complexes were observed clearly in cytosol at 4h and 10h after treatment. It was observed that signal of lipo 2000 was much weaker than PTMS formulations, consistent with flow cytometry data (Fig. ). More importantly, we found that Cy5 signals were diffused punctate at 10h rather than aggregated punctate at 4h (indicated by white arrows in Fig. ), implying NPs promoted siRNA release from endosomes into cytosol 15, 33. MFI and co-localization ratios were quantified from randomly selected four images (Fig. ). At the same treatment times, PTMS/Cy5-siRNA with higher N/P ratio exhibited lower co-localization ratio. The value of co-localization ratio was significantly declined from 4h to 10h. They were 88.15 ± 0.53% (N/P=10), 76.40 ± 3.85% (N/P=20), 69.02 ± 1.90% (N/P=30) in 4h, and 78.70 ± 3.81% (N/P=10), 64.78 ± 4.44% (N/P=20), 55.13 ± 3.93% (N/P=30) in 10h. As we know, the lower co-localization ratio, the more siRNA endosomal escape, the better gene silencing effect. This might be the reason why PTMS showed better gene knockdown efficiency at higher N/P ratios. Due to rapid hydrolysis of TTMA in acidic environment, PTMS/siRNA formulation disintegrated quickly in endosome, where the pH was about 5.5 26 and proton sponge effect was induced. At the same time, colloidal osmotic pressure in endosomes was further raised up because of a large number of hydrolyzed products 27. Finally, endosomes were swollen, ruptured and released siRNA into cytosol. In addition, MFIs of samples were calculated (Fig. ). MFI decreased with increasing N/P ratios, exhibiting the similar tendency as observed in flow cytometry. Compared with 4h, the slightly reduction of MFI was detected for each sample in 10h, which may be contributed by exocytosis 18. Combining consideration of flow cytometry test and confocal images together with gene silencing efficiency results, we supposed that enhancing siRNA endosomal escape was much more important than improving endocytosis for successful delivery.

pH-triggered degradation of NPs and siRNA release

To further understand high-efficiency siRNA delivery mediated by pan class="Chemical">PTMS, we here analyzed in vitro n>an class="Chemical">polymeric degradation and siRNA release in acidic buffer, mimicking endosomal microenvironment. Firstly, hydrolytic study of acetal was performed using 1H NMR spectra according to reported method 34. As showed in Fig. , the sharp peak at 6.25-6.50 ppm (b') was observed clearly for TFA incubated sample, which was contributed by 2,4,6-trimethoxybenzaldehyde monomer. And the strong signal around at 3.75-4.0 ppm (c') was presented correspondingly, indicating rapid hydrolytic reaction of TTMA under acidic buffer. Similarly, acetal hydrolysis in acidic solutions was determined by UV-spectroscopy with monitoring the wavelength of 295nm that was characteristic absorbance of 2,4,6-trimethoxybenzaldehyde (Fig. ). Typically, pH-dependent hydrolytic rate was exhibited, which was agreement with previous reports 24. The hydrolytic rate of acetal was much faster in pH 4.0 than pH 6.0 and pH 5.0 buffer, further confirming the pH-triggered degradation of PTMS NPs.

In addition, the behavior of PTMS NPs was tracked in pH 5.0 buffer, mimicking endosomal microenvironment. Size changes were tested within 32h (Fig. ). Because of hydrophobic-to-hydrophilic transition of PTMS NPs under acidic conditions, significant changes of sizes were observed. Parts of NPs expanded to ~5µm at 8h, and the percentage of larger size increased in 12h and 18h. TEM images also showed dramatically aggregation (Fig. ), size changed from ~40nm to ~1µm under pH 5.0 buffer for 24h. Small sizes (~3nm) of NPs were detected after 32h in pH 5.0 buffer showed in Fig. indicating water-soluble unimers resulted from completely acetal hydrolysis of TTMA 24. Moreover, size changes exhibited pH-dependency, larger NPs were observed earlier in pH 4.0 buffer than pH 5.0, 4h and 8h, respectively. Few aggregations were detected in pH 6.0 condition and no aggregation was showed in pH 7.4 buffer (Fig. ), suggesting PTMS NPs was relative stable without acetal hydrolysis in physiological environment. To simulate endosomal escape of siRNA, in vitro siRNA release profile was performed (Fig. ). As expected, Cy5-siRNA release was dramatically promoted in acidic conditions compared with neutral buffer, while pH 4.0 showed the fastest. In pH 5.0 condition, about 70% of accumulated release was measured within 10h, which ensured quickly endosomal escape of siRNA after endocytosis. Take these results into consideration, it was believed that high-efficiency silencing effects of PTMS/siRNA formulation were resulted from effective endosomal escape of siRNA driven by fast hydrolysis of acetal.

In vivo tumor distribution and inhibition

Encouraged by potent gene pan class="Disease">knockdown efficiency in vitro, In vivo expn>eriments were performed. With the sizes of 70-120nm, n>an class="Chemical">PTMS/Cy5-siRNA formulations were endowed with tumor enhanced permeability and retention (EPR) effect 35. To explore the possibility of cancer therapy in mice level, in vivo tumor accumulation was detected. Cy5 signals of whole body were captured at different time-points during 48h (Fig. ). Both naked siRNA and PTMS/Cy5-siRNA complex quickly reached tumor sites (indicated by black circles) within 1h. However, fluorescence intensity of tumor sites fell down with time flies, and it dropped much faster in naked siRNA group. Cy5 signal was too weak to be detected in naked siRNA group after 48h. At the end of experiment, tumors were separated, imaged and analyzed quantification (Fig. ). Both fluorescent image and quantified data indicated that PTMS NPs delivered much more siRNAs into tumor compared with naked siRNA group. Further in vivo antitumor activity was tested in mice model (Fig. ). Significant tumor growth inhibition was observed in PTMS/siRRM2 group, but not shown in other groups. The rrm2 gene expression in tumors was tested later by both mRNA (Fig. ) and protein levels (Fig. ). Knockdown efficiency was only obtained in PTMS/siRRM2 group (~45%), demonstrating that tumor growth inhibition was resulted from silence of rrm2 gene. Moreover, no significant changes of body weight were observed during entire experiment (Fig. ) and serum cytokines (Fig. ) were not induced significantly, suggesting PTMS/siRRM2 formulation was well-tolerated in mice.

In vivo organ distribution and gene knockdown

Besides, we tested the biodistribution of pan class="Chemical">PTMS compn>lexes in healthy C57BL/6 n>an class="Species">mice (Fig. ). Cy5 signals of PTMS/Cy5-siRNA were much stronger than naked siRNA group in both monitoring points, suggesting PTMS NPs prolonged siRNA half-life in body. There were at least two factors for this result. One was the PEG motif of PTMS NPs, which was believed to enable NPs to avoid nonspecific absorption of serum proteins and increase blood circulation time 36, 37. That was indeed observed in pharmacokinetics study (Fig. ). The half time (T1/2) of naked siRNA was only several minutes (< 5min) but it was around 30min for PTMS formulation. Area under the curve (AUC) of PTMS formulation was also much larger than naked siRNA. The other factor was size distribution of PTMS/Cy5-siRNA. Some works revealed that polyplexes with >10nm size effectively reduced siRNA excretion from kidney 38. Liver, lung and spleen were observed to show extensive accumulation of polyplexes, indicating PTMS NPs had delivery potential in these organs. We then selected liver for in vivo functional study and apoB (Apolipoprotein B) as the target gene. The remaining activity of apoB gene was 37.1 ± 4.6% for PTMS/siApoB group, much better than naked group (90.98 ± 16.86%) (Fig. ). We also tested delivery potency of PTMS formulation by local administration. Recently, siRNA therapies in skin diseases showed huge application prospects and commercial value. It was worth mentioning that therapeutic siRNA, named TD101, has completed phase Ib clinical trial for pachyonychia congenita (PC) treatment 39. To explore delivery potency of PTMS/siRNA formulation in skin, mice were s.c. injected with PTMS/siSCD1 formulation for 48h. The skin of injected site was then collected and gene silencing efficiency was detected by RT-PCR (Fig. ). Stearoyl-CoA desaturase-1 (scd-1) gene expression was dramatically inhibited by PTMS/siSCD1 complexes (~80%). But no significant gene knockdown efficiency was observed in both naked siSCD1 group and PTMS/siNC group. These results were mostly related with high endocytosis and effective siRNA endosomal escape of the formulation.

Conclusion

In summary, we here reported a novel pH-sensitive nanopan class="Species">micelle, named n>an class="Chemical">PTMS. Because of pH-triggered feature of TTMA group, PTMS NPs effectively enhanced siRNA endosomal escape and mediated excellent gene silencing effects in vitro, better than positive control (Lipo 2000). In vivo experiments indicated that PTMS delivered much more siRNA into tumor sites and mediated effective tumor growth inhibition without significant toxicity. Besides, obvious gene inhibitions were detected in liver and skin by i.v. and s.c. injection, respectively. Hence, PTMS/siRNA nanodrugs showed therapeutic prospects in tumors, liver-related diseases and skin-related disorders. In this work, we tried to improve siRNA delivery potency by rational design of carriers, our data revealed that siRNA endosomal release was more important than endocytosis for successful delivery. The pH-sensitive feature is perfectly matched with enhancing siRNA endosomal release, which could be an important principle for optimizing siRNA delivery systems.

Supplementary figures and tables.

Click here for additional data file.