Delivery of ENaC siRNA to epithelial cells mediated by a targeted nanocomplex: a therapeutic strategy for cystic fibrosis

The inhibition of ENaC may have therapeutic potential in CF airways by reducing sodium hyperabsorption, restoring lung epithelial surface fluid levels, airway hydration and mucociliary function. The challenge has been to deliver siRNA to the lung with sufficient efficacy for a sustained therapeutic effect. We have developed a self-assembling nanocomplex formulation for siRNA delivery to the airways that consists of a liposome (DOTMA/DOPE; L), an epithelial targeting peptide (P) and siRNA (R). LPR formulations were assessed for their ability to silence expression of the transcript of the gene encoding the α-subunit of the sodium channel ENaC in cell lines and primary epithelial cells, in submerged cultures or grown in air-liquid interface conditions. LPRs, containing 50 nM or 100 nM siRNA, showed high levels of silencing, particularly in primary airway epithelial cells. When nebulised these nanocomplexes still retained their biophysical properties and transfection efficiencies. The silencing ability was determined at protein level by confocal microscopy and western blotting. In vivo data demonstrated that these nanoparticles had the ability to silence expression of the α-ENaC subunit gene. In conclusion, these findings show that LPRs can modulate the activity of ENaC and this approach might be promising as co-adjuvant therapy for cystic fibrosis.

the cell membrane 6,7 . ENaC biodistribution is quite variable among different cells and tissues 8 . A fourth subunit, δ-ENaC, has also been found in several human tissues, which, together with the βand γ-subunits, can form a functional ion channel 9,10 . However, the expression of the α-ENaC subunit appears to be essential for full channel function, while residual channel activity can be measured in the absence of the two others (β and γ) 11 . ENaC dysregulation due to a genetic mutation of the subunits has been linked with the pathogenesis of Liddle's syndrome and salt-sensitive hypertension (pseudoaldosteronism) 12 . There is increasing evidence implicating ENaC dysregulation and hyperactivity in CF disease 3 .
The hypothesis of utilizing an ENaC-blocking molecule to facilitate restoration of the airway surface liquid (ASL) volume sufficiently to allow normal mucociliary clearance is of interest in the management of lung disease in CF patients 13 . The present study aimed to develop small interfering RNA (siRNA) nanoparticle formulations to be delivered to the airways. SiRNA is a feasible silencing strategy and we have shown that is possible to silence ENaC. Short interfering RNAs (siRNAs) are double-stranded RNAs 20-25 nucleotides long, which naturally regulate gene expression by degrading the messenger RNA 14 . Since the introduction of siRNA knockdown expression of specific genes in mammalian cells in 2001 14,15 , there is a growing number of therapeutic applications to treat a wide range of diseases by gene silencing.
To deliver the siRNA against the α-subunit of ENaC, we have developed nanoparticle formulations (LPRs) consisting of a cationic liposome (DOTMA/DOPE), a peptide with a targeting motif and siRNA. The targeting moieties SERSMNF in peptide E and YGLPHKF in peptide Y have been identified by biopanning a phage peptide library 16 . The peptide motif Y displays close similarity to a targeting protein expressed by the intracellular pathogen Legionella pneumophila while peptide E closely resembles the receptor binding proteins of two intracellular pathogens, rhinovirus and Listeria monocytogenes [16][17][18] and both have been shown to provide better targeting specificity and efficiently mediate delivery to airway cells [19][20][21] . LPRs were evaluated for their transfection efficiency in silencing the α-ENaC gene in vitro and in vivo. Biophysical characteristics including size, charge and shape were determined before and after nebulisation to assess their stability to shear forces. The current study allowed us to understand in more detail the potential of LPRs as tools for siRNA delivery to the airways and to develop them further as promising therapeutic applications for CF.

Targeted silencing of ENaC in cell lines transfected with LPR detected by qRT-PCR. LPR
nanocomplexes, formulated with peptide E (K 16 GACSERSMNFCG) and containing either targeting ENaC or non-targeting/irrelevant siRNA at 50 nM, were used to transfect 16HBE14o-epithelial cells. Following siRNA treatment, the level of silencing of the α-ENaC gene, determined by qRT-PCR was 42.8% (Fig. 1A), normalised to the GAPDH housekeeping gene.
The efficiency of gene silencing of LPR nanoparticles was further examined in another cell line (A549). It showed an average silencing of 46.6% at 50 nM and of 32.5% at 100 nM (Fig. 1B). The 100 nM concentration was shown to be more toxic (see below), which may explain the lower level of silencing in these cells compared to 50 nM.

Transfection of primary normal and CF airway epithelial cells and air-liquid interface (ALI) cultures.
We next compared the silencing efficiency of LPR nanoparticles formulated with peptide E at 50 nM siRNA in primary normal human bronchial epithelial cells (NHBE). The expression of α-ENaC in submerged NHBE showed a silencing efficiency of 78.5% ( Fig. 2A). As expected, silencing was less efficient (66.2%) when the same cells were transfected in air-liquid interface (ALI) cultures (Fig. 2B). In primary CFBE cells transfected in submerged culture using LPRs formulated with the targeting peptide Y, the efficiency of silencing was 59.5% at a siRNA concentration of 50 nM and 72.9% at 100 nM ( Fig. 3A) which was slightly lower than the silencing seen with LPRs formulated with peptide E ( Fig. 2A). A decrease in knockdown efficiency mediated by peptide Y-LPRs was observed in CF cells grown at ALI (Fig. 3C), as shown for NHBE cells, with expression of the α-ENaC subunit transcripts reduced by 28.2% at 30 nM and 43.3% at 50 nM.
The effect of these formulations on CFBE cell viability in comparison to cells that were not transfected, ranged between 72% at 50 nM and 57% at 100 nM for α-ENaC-targeted siRNA, whereas the cell survival was higher at the same concentrations of non-targeted/irrelevant siRNA (Fig. 3B), although there was no statistical difference among the different groups.

Biophysical properties and transfection efficiencies of nebulised LPRs. Aerosol delivery is an
attractive approach for therapeutic interventions in the affected epithelial cells in the lungs of CF patients and we have previously shown that LPRs can deliver genes to the airway epithelium by nebulisation 19,22 . LPRs were formulated in water, nebulised through an AeroEclipse II BAN jet nebuliser and the aerosol was collected. When loading the nebuliser with a small volume only a maximum of 50% of the suspension can be released as aerosol 23 , however, the amount of LPR suspensions nebulised here was 68.8 ± 4.7% (n = 3, Supplementary Table S1). In the same experiments the yield of siRNA in the collected nebulised LPRs, expressed as percentage of the siRNA in the pre-nebulisation suspension, was 17.5 ± 4.6%. Samples, prior to and after nebulisation, were stored refrigerated then investigated for biophysical properties and efficiency of silencing (Fig. 4A). The LPR sizes showed no significant variation after nebulisation at 200 ± 21 nm pre-nebulisation and 192 ± 4 nm after nebulisation. The ζ potential also displayed a similar colloidal stability, showing values of 25.6 ± 2.6 mV for the pre-nebulised LPRs versus 23.6 ± 0.4 mV for the aerosolised ones. Also TEM images showed that the LPR suspension pre-and post-nebulisation were homogeneous and quite similar (Fig. 4C).
In order to determine the ability of post-nebulised LPRs, formulated with either peptides (Y and E), to knockdown the transcript for the α-ENaC gene in vitro, equal amounts of freshly made LPRs, pre-nebulised nanoparticles suspension and nebulised material were added to cultures of 1HAEo− cells (Fig. 4B). The silencing efficiency Scientific RepoRts | 7: 700 | DOI:10.1038/s41598-017-00662-2 of Y-LPRs was 77.7% for the pre-nebulisation suspension and 78.5% for the post-nebulisation formulation, compared to 85.4% for their freshly made counterparts. Similar α-ENaC mRNA knockdown efficiencies were shown by the E-LPRs; both the freshly made preparation and the pre-nebulisation formulation displayed a similar silencing ability of 79.8%, whereas the expression of α-ENaC showed a transcript knockdown of gene equal to 74.8% by the post-nebulisation suspension.
Detection of α-ENaC protein by confocal microscopy and western blotting. Confocal microscopy experiments were performed on NHBE cells grown in ALI culture ( Fig. 5A-C), transfected with LPRs containing peptide E and siRNA at 50 nM concentration then stained for α-ENaC subunit (green), actin (red) and DAPI (blue). Although quantification was not possible due to the non-homogeneous growth of the cells in ALI cultures and the worsening of the spherical aberration when focusing in the depth of the transwell, a decrease in the fluorescent staining of the α-subunit was detected.
The silencing of the protein was then investigated by western blotting assessing the amount of α-ENaC at the plasmalemma level and in the intracellular pool. Despite the observation that mRNA was efficiently silenced at a concentration of 50 nM siRNA, a decrease in the protein level of the 90 kDa α-subunit was seen only with 100 nM siRNA ( Fig. 6A and B). A decrement was also seen in the 65 kDa band at least in the cytosolic fraction of ENaC siRNA treated cells compared to the irrelevant siRNA (Fig. 6A). Quantification of the 90 kDa bands that were detected on the western blots ( Fig. 6B) showed that the percentage of knockdown normalised to the non-targeting siRNA was 50% and 25% in plasma membrane and cytosolic fractions, respectively. SiRNA delivery in vivo. We then evaluated the in vivo knockdown of ENaC α-subunit transcripts. C57BL6 mice were treated with LPRs containing peptide Y and 27 μg of either ENaC siRNA or irrelevant/non-targeted control siRNA by oropharyngeal administration and the expression of α-ENaC was examined 48 h later. A statistically significant reduction (*P < 0.05) using Student's t-test in the expression of α-ENaC (Fig. 7A) was observed between the ENaC-treated group and the control groups (either treated with non-targeting/irrelevant siRNA or untreated). H&E staining of the lungs showed that LPR administration to the lungs of mice, containing either ENaC-targeting or non-targeting siRNAs, induced a mild peribronchial proteinaceous and predominantly mononuclear cell infiltrate compared with untreated controls (Fig. 7B-D).

Discussion
Despite the remarkable personalised approaches brought to CF therapy by the CFTR modulators, gene therapy still offers a valid option for patients with mutation classes where these new therapeutics are lacking. SiRNA approaches to silence ENaC may also be a co-adjuvant therapy not only for CF, but to all those treatments aiming to restore appropriate airway surface liquid volume and mucus hydration by regulating sodium hyperabsorption. Gene therapy for CF, targeting fully differentiated airway epithelial cells, has to take into account the dynamic airway plasticity that requires repeated applications. We have previously shown that LPR nanoparticles can be used to target different cell-types 18,21,[24][25][26] . The aim of this study was to establish whether we could reduce the Na + reabsorption in the airway epithelial cells and in vivo in the mouse lungs by silencing the sodium channel ENaC. The α-ENaC expression in NHBE submerged cultures (n = 5) treated with 50 nM siRNA was 21.5% relative to GAPDH housekeeping gene. The difference was statistically significant (*** P ≤ 0.001). (B) Air-liquid interface cultures of NHBE (n = 3) treated with LPRs containing 50 nM siRNA showed a remaining expression of 33.8% giving therefore a silencing of 66.2% relative to GAPDH housekeeping gene. The remaining expression of the α-ENaC gene is expressed as percentage relative to non-targeted/irrelevant siRNA-treated cells.
ENaC is a complex target because the channel is formed of different subunits and because the α-subunit can be alternatively spliced, can undergo post-translational modifications and can be highly glycosylated and ubiquitinylated [27][28][29] . It is important also to consider that ENaC is differentially expressed in cells and tissues 8 and that the amount of intracellular protein seen by confocal microscopy in some cells could be due to the constant replacement of the subunit due to the proteolytic cleavage as part of one of ENaC regulatory mechanisms [29][30][31] .
We showed here that the LPR formulation has the ability to knockdown the α-ENaC subunit at various degrees in different airway epithelial cells. Although primary cells are usually more difficult to transfect, in general we found that better silencing was achieved in these cells rather than in cell lines, where a knockdown level of 50% or better was achieved. Although silencing efficiencies were slightly lower when the same cells were maintained in ALI culture, a more physiologically-representative model of the airway respiratory epithelium, here the knockdown levels in CF primary epithelial cells were extremely encouraging, considering that ALI cultures are notoriously difficult to transfect. Interestingly, the amount of α-ENaC mRNA did not correspond to protein levels, which may be due to translational regulation of the mRNA, protein stability or the α-subunit forming a conducting channel with different regulatory characteristics. Our data also demonstrated that either targeting peptides were equally able to mediate the delivery of liposomal nanoparticles to the airway epithelial cells. Consistent with our previous results 18 , the LPRs had the desired biophysical characteristics for a delivery vector in terms of size and stability. The biophysical analysis revealed that all the LPR formulations displayed sizes suitable for cellular uptake 24 and showed enhanced colloidal stability, which render them less prone to aggregation or flocculation. LPR formulations made in water and nebulised by AeroEclipse II BAN jet nebuliser retained their biophysical properties and their ability to silence the α-ENaC gene in airway epithelial cells suggesting that LPRs presented in the current work may also be useful for clinical translation.
We further demonstrated that LPRs have the ability to knockdown the gene encoding the α-ENaC subunit in C57BL6 mouse lungs. The experiments showed that there was a certain degree of silencing of the gene in the ENaC siRNA-treated mice. A complete inactivation of the α-ENaC gene in mouse has been reported to abolish the Na + transport in airway epithelia causing the development of respiratory distress due to the inability to clear Overall, our study demonstrates that LPRs can be used to effectively deliver siRNA to the airway epithelium and that the expression of α-ENaC can be modulated in vitro and in vivo. However, an optimal dose is required to achieve the therapeutic benefits without potential side effects. Receptor-targeted liposome-peptide-siRNA nanoparticles represent an efficient and safe non-viral siRNA delivery system that could be used to modulate sodium hyperabsorption, thus helping to restore the correct volume of surface liquid, mucus hydration and mucociliary clearance in the airways, in cystic fibrosis.   for a further 48 h prior to RNA extraction to measure the silencing efficiency by qRT-PCR, or lysed to detect the knockdown of the protein by western blotting or rinsed with PBS and fixed for confocal microscopy.

Cell lines, primary cells and air-liquid interface (ALI) cultures. The 16HBE14o-and 1HAEo− cells
Aliquots of the LPR nanocomplexes were prepared and diluted in H 2 O (Life Technology) to a final volume of 500 µl. Complex sizes and ζ potentials were measured by dynamic light scattering and by laser Doppler anemometry using a Nano ZS Zetasizer (Malvern Instruments, Malvern, UK) with the settings described previously 19 . A total of three measurements per sample were taken automatically. For TEM analysis, the LPR nanocomplexes were prepared in water and applied onto a glow-discharged 300-mesh copper grid coated with a Formvar/carbon support film (Agar Scientific Ltd, Stansted, UK). After a few seconds, the grid was dried by blotting with filter paper. The sample was then negatively stained with 1% (w/v) uranyl acetate for a few seconds, before blotting with filter paper and air-dried. Imaging was carried out under a Philips CM120 BioTwin Transmission Electron Microscope and operated at an accelerating voltage of 120 KV.
Cell Viability. Cell viability was assessed by measuring the release of lactate dehydrogenase (LDH) from damaged cells with the CytoTox-ONE Homogeneous Membrane Integrity Assay kit (Promega, Southampton, UK). Seeded cells were transfected with LPRs as above, then the culture supernatant was collected and mixed with an equal volume of the CytoTox-ONE reagent, according to the manufacturer's instructions. The fluorescence (excitation λ 560 nm -emission λ 610 nm) was measured on a FLUOstar Optima spectrophotometer (BMG Labtech, Aylesbury, UK). Cell viability for each complex was expressed as a percentage of the viability of untreated cells.

LPR nebulisation.
For the nebulisation studies, 2 ml of LPR nanoparticles containing 100 μg of siRNA, prepared in H 2 O (Life Technologies), were added to the sample chamber of the AeroEclipse II BAN nebuliser, a kind gift from Trudell Medical International Europe Ltd (Manchester, UK). The nebulizer was then connected to the compressor Hikoneb ® AeroCare II (Piston Type Nebuliser, Antek Electronic Ltd, Ankara, Turkey). The mouthpiece of the nebuliser was modified and connected by an adaptor to a large semi-rigid plastic tube (about 15 cm long) inserted two-thirds of its length inside a 50 ml tube through its cap, so that the aerosolised material would remain in the bottom. Also in the cap a barrel of a luer lock 1 ml syringe was inserted upside-down and attached to a tubing that helped to direct the flow out of the barrel into the water (placed in a Erlenmeyer flask), but could also allow some of the condensed material to fall back into the tube. The tube was pre-chilled and kept on ice to minimise evaporation. The nebulisation was performed for 15 min with the AeroEclipse set on continuous operational mode. Pre-and post-nebulisation samples were collected with a precision pipette and stored at 4 °C until required for further analysis. The concentrations of siRNA in LPRs pre-nebulisation, post-nebulisation and also from the left-over material in the sample chamber of the nebuliser were measured using Nanodrop 1000 (Thermo Fisher Scientific Inc., Wilmington, USA). The biophysical properties were then measured and their silencing activities were analysed in cellular transfections added at a concentration of 50 nM per well.

RNA extraction from cells and mouse tissues. Lysates from cells washed in PBS, scraped and frozen at
−80 °C until use, were first homogenized with Precellys (2 cycles × 5600 rpm, 30 sec per cycle) and then mixed with 1 vol of TRIreagent (Sigma). Mice lungs in 1-1.4 mL of TRIreagent (Sigma) were directly homogenized with the Precellys (4 cycles × 5600 rpm, 30 sec per cycle). The homogenates were centrifuged at 1000 g for 10 min at 4 °C and the supernatants were collected. After 5 min 1/5 vol of chloroform was added, following 3 min incubation the samples were centrifuged at 12000 g for 5 min at 4 °C. The supernatants were collected, added to an equal volume of 70% ethanol and further purified using the ISOLATE II RNA Micro Kit (Bioline, London, UK) according the manufacturer's instructions. Samples were frozen at −80 °C until required for analysis.

Reverse transcription and Real-time PCR.
Total RNA was defrosted, treated with DNase (1 μg) and either converted into first-strand cDNA using random hexamers and MuLV reverse transcriptase (Life Technologies) for 1 h at 42 °C and cDNA was subsequently heated at 80 °C for 5 min prior to qRT-PCR assay or used in a one-step qRT-PCR (SensiFAST ™ Probe Hi-ROX Kit, Bioline) combining the reverse transcription reaction with the qPCR reaction. Human or mouse ENaC, β-actin or GAPDH were quantified by Taqman primers and probes (Hs01013028_m1, Hs99999903_m1 and Hs03929097_g1, or Mm01182998_g1, Mm00607939_s1 and Mm99999915_g1, all from Life Technologies). The qRT-PCR assays were performed in an ABI PRISM ® 7000 with the following default PCR parameters set at 50 °C for 2 min, 95 °C for 10 min and then 40 cycles at: 95 °C for 15 s and 60 °C for 1 min. The default parameters for one-step qRT-PCR were: 45 °C for 20 min, 95 °C for 2 min and then 40 cycles at: 95 °C for 15 s and 60 °C for 1 min. Only for A549 cells PrimePCR ™ Probe Assays (Bio-Rad Laboratories Ltd, Hemel Hempstead, UK) were used: Human SCNN1A (qHsaCEP0024081 FAM) and Human GAPDH (glyceraldehyde-3-phosphate dehydrogenase, qHsaCEP0041396 Hex). GAPDH mRNA, in comparison with β actin, seemed to be a more suitable reference gene from in vitro transfections because showed minimal changes in expression levels between the individual samples and experimental conditions. Western blotting. Proteins from cells and ALI cultures were isolated and separated into a cytosolic and a membrane fraction using the ProteoExtract Transmembrane Protein Extraction Kit (Millipore Ltd, Watford, UK) according to the manufacturer's instruction. The proteins were then concentrated by centrifugation at 12000 g for 30 min using Amicon Ultra-0.5 mL Centrifugal Filters (Millipore UK Ltd). For western blotting, the concentration of protein in the lysates was first measured by Pierce ™ BCA Protein Assay (Pierce Biotechnology, Rockford, lL, USA) reading the absorbance at 595 nm in a FLUOstar Optima luminometer (BMG Labtech, Aylesbury, UK).
Equal amounts of protein were then added into 4-12% NuPAGE ® bis-tris SDS-PAGE gels (Life Technologies, Paisley, UK) and the gels were run for 35 min with MOPS buffer. Proteins were then wet-transferred onto nitrocellulose (Ø45 µm, Schleicher & Schuell Protran ™ BA85 Nitrocellulose Membranes, Sigma). Membranes were left for 1 h in TBS-1% semi-skimmed milk then washed twice with Tris-buffered saline/Tween (TBST). The membrane was then probed overnight with a custom anti-α-ENaC antibody diluted 1:500 in 1% semi-skimmed milk/TBST. The custom anti-α-ENaC antibody, generated by Cambridge Research Biochemicals, was a kind gift of Ken Clark, Therapeutic Oligo DPU, Glaxo SmithKline, Stevenage, UK. After washing in TBST, the membrane was incubated for 1 h at room temperature with the goat anti-rabbit IgG conjugated with horseradish peroxidase (Life Technologies, Paisley, UK) diluted 1:10,000 in TBST. Following further washing the blot was developed with SuperSignal West Pico Chemiluminescent Substrate (Thermo Fisher Scientific Ltd, Cramlington, UK). The western blot densities, following the rolling ball background subtraction set at 50, were analysed using ImageJ programme (NIH, Bethesda, MD, USA, NIH.gov).
After stripping with Restore ™ western blot stripping buffer (Thermo Fisher Scientific -Pierce), GAPDH bands were detected by probing the membranes overnight with a mouse anti-GAPDH antibody (Santa Cruz Biotechnology, Inc., Heidelberg, Germany) 1:500 in TBST, followed by a horseradish peroxidase-labelled rabbit polyclonal anti-mouse immunoglobulins (1:1000 in TBST; Dako UK Ltd, Ely, UK).
Confocal Microscopy. Primary cells were grown in ALI cultures and transfected as above. Transwell inserts were then rinsed with PBS and fixed with 4% paraformaldehyde, washed with PBS, permeabilised with PBS containing 0.1% Triton for 10 min and blocked for 1 h with 3% BSA in PBS containing 0.1% Triton. The transwell membranes were stained with anti-α-ENaC rabbit polyclonal IgG antibody (Thermo Scientific) diluted 1:100 in 2% BSA and 0.03% Triton in PBS for 2 h. Transwell membranes were then washed thoroughly and incubated with Alexa Fluor 468 Phalloidin (Life Technologies) to stain F-actin diluted 1:100 and with goat anti-rabbit IgG conjugated with Alexa Fluor 488 (Life Technologies) diluted 1:1000 in 1% BSA and 0.03% Triton in PBS for 1 h. The inserts were finally washed with PBS and with distilled water prior adding the mounting medium (VectaShield, Vector Laboratories Ltd, Peterborough, UK) containing DAPI to stain the nucleus. The inserts were then mounted onto slides and kept in the dark at 4 °C. Images were acquired using a Zeiss LSM 710 confocal microscope (equipped with 7 laser lines: 405, 458, 488, 514, 561, 594 and 633 nm, and with 63x Plan-Apochromat NA 1.4 Oil objective) and processed using ImageJ software (NIH.org).
Oropharyngeal delivery and histological staining of the lungs. For in vivo studies, 55 μl of LPR suspensions containing 27 μg ENaC targeting or irrelevant siRNA were administered through the oropharynx of 6-8 weeks old C57BL6 female mice (Charles River, Margate, UK) as described [35]. The third group of untreated mice was used as a control. Animals were killed 48 hours later and the lungs either excised and quickly frozen in liquid nitrogen for mRNA expression analysis as described above or alternatively the lungs were inflation fixed in situ via cannulation of the trachea with 4% paraformaldehyde at a pressure of 20 cm for histological analysis. Following inflation fixing, the lungs were excised, dehydrated and processed to paraffin wax as described previously 36 . Histological sections were dewaxed, rehydrated, and stained with haematoxylin and eosin as described previously 22 . All animals were handled in strict accordance with UCL animal care policies and all the procedures were carried out in accordance with the United Kingdom Animals (Scientific Procedures) Act 1986 (UK) and E.U. regulations. The work was approved and carried out under Home Office Project Licenses 70/6149 and 70/7073. Statistical Analysis. All graphs display mean and standard error of the mean (SEM). Statistical analysis was performed using the Student's t-test to calculate statistically significant differences and individual P values using