1. Introduction
Small interfering RNA (siRNA) shows immense potential in treating previously un-druggable diseases via gene-specific knockdown as demonstrated by the FDA approval of ONPATTRO
® (partisiran) [
1] and GIVLAARI
® (givosiran) [
2]. While the chemical modification of siRNA increased the nuclease resistance, serum stability, target affinity and half-life of the molecule, encapsulation into lipid nanoparticles (LNPs) or conjugation with targeting ligands (e.g., N-acetylgalactosamine) improved the therapeutic window and altered the distribution and pharmacokinetics profiles [
3]. LNPs and GalNac siRNA conjugates are mostly restricted to the liver following intravenous (I.V.) administration and require steroidal anti-inflammatory treatment preceding LNP dosing to limit cytokine production. Delivery systems that meet criteria such as colloidal stability, high encapsulation efficiency, low toxicity/immune stimulation, and siRNA delivery efficiency to extrahepatic organs are critically needed.
Chitosan (CS), a family of cationic bio-copolymers composed of β (1-4) linked N-acetyl glucosamine (GlcNAc) and D-glucosamine (Glc), gained attention for nucleic acid (NA) delivery due to its low toxicity, simple production, and ease of chemical modification [
3,
4]. It can be tweaked for specific fractions of protonable Glc vs. GlcNAc, average molecular weights (Mw and Mn), and assembly into polyelectrolyte complexes (nanoparticles) via spontaneous electrostatic interactions. Several in vitro studies showing siRNA delivery with chitosan have been published previously [
4,
5,
6,
7,
8,
9,
10,
11,
12,
13,
14,
15,
16,
17,
18,
19,
20,
21,
22,
23,
24,
25]. In vivo, the nasal administration and intratracheal catheter administration of chitosan/siRNA formulations led to effective RNA interference in the lungs of transgenic EGFP mice [
15,
21]. A sustained and effective siRNA accumulation of chitosan/siRNA formulation was shown within the kidneys of I.V. injected mice [
26,
27]. Folic acid/chitosan conjugates were used to deliver siRNA to activated macrophages [
28]. Chitosan/siRNA nanoparticles were shown to knock down COX-2 specifically in macrophages, which might prevent kidney injury induced by unilateral ureteral obstruction [
29]. Since the in vivo induction of cytokines was never extensively characterized in previous work, a systemic study with accurately characterized chitosans that investigates hemocompatibility, in vivo acute toxicity and demonstrates knockdown following the I.V. administration of nanoparticles (NPs) is needed.
Here, we investigated the effect of chitosan polymer length, dose, and surface modification with hyaluronic acid (HA) on the hemolytic potential, acute and organ toxicity, cytokine induction, in vivo biodistribution and target knockdown efficacy, compared chitosan NPs with commercially available cationic LNPs (Invivofectamine®) for siRNA delivery, and extensively assessed acute toxicity, biodistribution in live animal imaging, and target knockdown efficacy in mice. We hypothesized that the administration of sub-hemolytic doses of chitosan is non-toxic compared to LNPs and induces potent gene knockdown at the site of accumulation. We tested three specific hypotheses in this study: (1) chitosan nanoparticles accumulate extra hepatically, (2) HA-coated nanoparticles have a different biodistribution pattern vs. uncoated formulations, and (3) HA nanoparticles demonstrate higher knockdown efficiency in accumulated sites due to improved hemocompatibility and increased doses.
2. Materials and Methods
2.1. Materials
Medical grade hyaluronic acid (HA, 866 kDa, HA1M-1) was purchased from Life Core Biomedical (Life Core Biomedical LLC, Chaska, MN, USA). This particular HA was chosen based on our previous work on nanoparticle stability [
30]. A lipopolysaccharide (LPS) serotype O55:B5 (TLRgrade™) from Enzo Life Sciences (Enzo life sciences, Farmingdale, NY, USA), isoflurane (Forane™) from Baxter (Baxter Canada, Mississauga, ON, Canada), BD vacutainer SST Gold from VWR international (VWR International, Mont-Royal, QC, Canada), IDEXX green top Lithium–Heparin and yellow top serum microtainers from IDEXX Laboratories (IDEXX Laboratories, Markham, ON, Canada), Altogen in vivo transfection kit from Altogen Biosystems (Altogen Biosystems, Las Vegas, NV, USA), Invivofectamine
® 2.0 and 3.0, phosphate-buffered saline (PBS), UltraPure™ DNase⁄RNase-Free water, 10% Neutral Buffer Formalin, AlexaFluor 546 phalloidin with ProLong
® Diamond antifade containing DAPI and nuclease free water from Life technologies (Burlington, ON, Canada) were all used. D-trehalose, L-histidine, diethyl pyrocarbonate (DEPC), 1N HCl, and RNaseZAP™ were purchased from Sigma-Aldrich (Sigma-Aldrich, Oakville, ON, Canada). Rabbit monoclonal anti-GAPDH (Ab181602) and biotinylated goat anti-Rabbit IgG (Ab97049) were purchased from Abcam (Abcam, Cambridge, UK). Serum vials (223685, 223686 and 223687) were purchased from Wheaton (Wheaton, Millville, NJ, USA), and butyl stoppers (73828A-21) were purchased from Kimble Chase (Kimble Chase, Rockwood, TN, USA). PVDF filters (0.22 μm) and Amicon Ultra-15 centrifugal filter units were purchased from EDM Millipore (EDM Millipore Ltd., Etobicoke, ON, Canada). The native and 2′O methyl (2′OMe) modified anti-ApoB siRNA sequences were custom synthesized by Dharmacon Inc (GE Dharmacon, Lafayette, CO, USA). The anti-GAPDH siRNA was purchased from Life technologies as a predesigned Ambion
® In Vivo GAPDH Positive Control siRNA (Life technologies, Burlington, ON, Canada). Chitosans were obtained from Marinard, (Laval, QC, Canada).
2.2. siRNA Sequences and Chitosan Characterization
All siRNA sequences came in a lyophilized format following HPLC purification and subjected to quality control (QC) (e.g., endotoxin content, LC-MS, PAGE and UV/Vis spectrophotometric analysis). The sequences of siRNAs are summarized in
Table S3.
Chitosans were depolymerized with nitrous acid with the aim of obtaining chitosans of number-average Mn of 10 and 120 kDa. Those target Mn were chosen based on our previous work using chitosan to deliver siRNA [
31]. The actual chitosan number and weight-average molecular weights (Mn and Mw) (
Table 1) were then determined by gel permeation chromatography (GPC) using a Shimadzu LC-20AD isocratic pump coupled with a Dawn HELEOS II multi angle laser light scattering detector (Wyatt Technology Co., Santa Barbara, CA, USA), an Optilab rEX interferometric refractometer (Wyatt Technology Co.), and two Tosoh TSKgel (G6000PWxl-CP and G5000PWxl-CP; Tosoh Bioscience LLC, King of Prussia, PA, USA) columns. Chitosans were eluted at pH 4.5 using an acetic acid (0.15 M)/sodium acetate (0.1 M)/sodium azide (4 mM) buffer. The injection volume was 100 μL at an 0.8 mL/min flow rate at 25 °C. The dn/dc value was determined as 0.208 at 658 nm. The degree of deacetylation (DDA) was determined by 1H NMR.
2.3. Preparation of Chitosan-Based Nanoparticles
Low (10 kDa) and high (120 kDa) molecular weight chitosans were dissolved overnight in nuclease-free water (NFW) and 1N HCl, using a glucosamine to HCl ratio of 1:1, to a final concentration of 5 mg/mL. HA was prepared by dissolving sodium hyaluronate in NFW at a concentration of 1 mg/mL. The stock solutions were sterile filtered using a 0.22 μm PVDF filter and used to prepare solutions containing 0.83% w/v trehalose and 5.83 mM histidine (toxicity) or 1% trehalose and 3.8 mM histidine (efficacy) at a specific amine: phosphate: HA carboxyl molar ratio (N:P:C = 2:1:1.5) by dilution in nuclease-free water, 4% w/v trehalose and 28 mM histidine (pH 6.5). Before complexation, anti-ApoB (native and 2′Ome modified) and anti-GAPDH siRNA stock solutions were diluted to 0.2 mg/mL in the same buffer as chitosan and/or HA (0.83% trehalose and 5.83 mM histidine or 1% trehalose and 3.8 mM histidine).
2.4. Preparation, Lyophilization, and Reconstitution of Uncoated and HA-Coated Anti-ApoB Nanoparticles for the Assessment of In Vivo Toxicity
Uncoated and HA-coated anti-ApoB nanoparticles were prepared at a final N:P:C ratio of 5:1:0 and 2:1:1.5, respectively, using the advanced Automated In-line Mixing System (AIMS) as described before [
7]. Chitosan at a specific N:P ratio (5:1 or 2:1) was mixed using a closed and sterile system comprising an LS14 Pharmapure tubing (1/16”) and two Masterflex L/S digital peristaltic pumps (Cole-Parmer, Montreal, QC, Canada), with siRNA (0.2 mg/mL) using a Y-connector and a mixing flow rate of 150 mL/min (Re = 4000). Anti-ApoB nanoparticles prepared at N:P = 2 were HA coated to a final N:P:C ratio of 2:1:1.5. Chitosan–siRNA nanoparticles (N:P:C ratio of 2:1:0) were mixed with HA at a 1:2 vol:vol ratio and a mixing flow rate of 150 mL/min (nanoparticles) and 75 mL/min for HA. Nanoparticles were incubated for 30 min at room temperature (RT) before analyses or freeze-drying. To inactivate possible nucleases, the whole closed system was treated with diethylpyrocarbonate (DEPC), autoclaved, and flushed with nuclease-free water.
Anti-ApoB nanoparticles were lyophilized under sterile conditions, using a 3-day cycle. Nanoparticle volumes of 2 and 5 mL were freeze-dried (FD) using a Laboratory Series Freeze-Dryer PC/PLC (Millrock Technology, Kingston, NY, USA). Samples were backfilled with Argon, stoppered, crimped, and stored at 4 °C until reconstitution. All freeze-dried samples were reconstituted to 12× initial concentration (208/417 μL to 5/10 mL serum vials respectively) and then incubated at RT for 5–10 min, and the concentration was adjusted by a nearly isotonic aqueous solution of 10% w/v trehalose and 70 mM histidine (pH 6.5) so that the desired dosage (mg siRNA/kg animal body weight) would be reached upon the injection of 10 μL of nanoparticle suspension per gram of body weight (BW).
2.5. Preparation of Uncoated and HA-Coated Anti-GAPDH Nanoparticles for Assessment of In Vivo Target Knockdown
Anti-GAPDH siRNA (0.2 mg/mL), low Mn chitosan (10 kDa), high Mn chitosan (120 kDa) and HA working solutions were prepared in the same way as described in the previous section. Uncoated chitosan–siGAPDH NPs were prepared at an N:P ratio of 5 by electrostatic mixing at a 1:1 vol:vol. HA-coated NPs were prepared at an N:P ratio of 2.5:1 by manual mixing (1:1 vol:vol), incubated at RT for 15 min, and coated with HA by mixing 1 part of HA working solution (0.4 mg/mL) to 2 parts of chitosan–siGAPDH NPs for a final N:P:C ratio of 2.5:1:2. The final volume never exceeded 1 mL, and chitosan was pipetted into siRNA. NPs were kept at RT for 20–30 min before administration to animals.
2.6. Preparation of Invivofectamine®-siRNA LNPs
Invivofectamine® 2.0 and 3.0 were prepared as per the manufacturer’s recommendation. First, 250 μL of anti-ApoB siRNA (3 mg/mL) was diluted 1:2 in complexation buffer, mixed with 500 μL of Invivofectamine®2.0, vortexed for 30 s, incubated at 50 °C for 30 min, diluted with 14 mL of phosphate-buffered saline and concentrated at 4000 g using an Amicon Ultra-15 centrifugal filter unit (EDM Millipore Ltd., Etobicoke, ON, Canada) to a final volume of 872 μL (0.8 mg/mL siRNA).
For Invivofectamine®3.0, an anti-GAPDH siRNA (2.4 mg/mL) was mixed with a complexation buffer at 1:1 ratio and immediately added to Invivofectamine®3.0 at a 1:1 vol:vol ratio, vortexed for 30 s, incubated at 50 °C for 30 min and diluted to 0.25 mg/mL siRNA. All LNPs were subjected to QC (Dynamic Light Scattering (DLS), Doppler velocimetry, UV measurements and sterility assessment) and stored at 4 °C for 10–16 h before administration into mice. Invivofectamine®3.0 was used as replacement for Invivofectamine®2.0, which was discontinued at the time of the efficacy study.
2.7. Determination of Size and Surface Charge
The size and surface charge (ζ-potential) of NPs were determined by DLS and Laser Doppler velocimetry using a ZetaSizer Nano ZS device (Malvern Instruments Ltd., Malvern, UK). Measurements (N = 2–3, n = 6–9) were performed at a detector’s scattering angle of 173 at 25 °C using the viscosity of water as the sample diluent. NPs were diluted to 1× their initial concentration using NFW, which was followed by a dilution 1:4 and 1:8 using sterile 1% trehalose solution before determination of size and ζ-potential, respectively. A Smoluchowski equation was used to calculate the ζ-potential from the measured electrophoretic mobility.
2.8. Hemocompatibility
The hemolytic and hemagglutination properties of uncoated and HA-coated NPs were tested according to ASTM E2524-08 [
32] and Evani et al. [
31], respectively. The influence of dose, Mn, N:P ratio and HA coating on erythrocyte aggregation (hemagglutination) was investigated to better understand chitosan–blood interaction and limit potential in vivo toxicity. Blood was collected from healthy human donors following protocol approval by the Polytechnique Montreal Ethics Committee. Anti-ApoB NPs were prepared as described above, FD in the presence of 0.83%
w/
v trehalose, and 5.8 mM histidine (pH 6.5), and rehydrated to 12× the pre-FD concentration for the highest tested concentration (or dose) at iso-osmolality and then serially diluted to final siRNA concentrations of 0.1, 0.25, 0.5, and 0.8 mg/mL. Plasma-free hemoglobin (PFH) in the blood was measured at 0.49 mg/mL prior to assay. Total blood hemoglobin (TBH) was adjusted to a concentration of 10 ± 1 mg/mL (dTBH). NPs were diluted in dTBH at a 1:7:1 volumetric ratio with 100 μL of NPs at the target concentration pipetted into 700 μL PBS and 100 μL of blood (dTBH 10 ± 1 mg/mL). For colorimetric determination of hemolysis, 700 μL of samples was incubated at 37 °C for 3 h and visually inspected every 30 min for nanoparticle flocculation, dispersion, sinking or floating. Supernatant was collected following centrifugation at 800×
g for 15 min, and absorbance was measured at 540 nm (Tecan Systems, Mannedorf, Switzerland). A four-parameter regression algorithm was used to obtain the calibration curve to calculate the hemoglobin concentration in the supernatant of each PFH sample. The percentage of hemolysis was computed as:
Hemolysis (%) = 100 × (
PFHsample/
dTBH). For hemagglutination, the remaining 200 μL of each sample prepared above was pipetted in 96-well assay plates, incubated for 3 h, and visualized using an Axiovert light microscope, and the area covered by red blood cells was estimated and scored.
2.9. In Vivo Studies
In vivo experiments were randomized double blinded and approved by the University of Montreal Ethics Committee (CDEA) and the Montreal Heart Institute Research Center Ethics Committee. Mice (Charles River, Quebec, QC, Canada) were acclimatized in a pathogen-free facility with unrestricted access to water and food. Mice had body condition scores (BCSs) of 3 [
31] with BW in the 20–25 g range at the time of injection. Injection volumes were calculated as 10 μL/g of BW and injections were performed within 10–15 s. Mice were euthanized by cardiac puncture followed by cervical dislocation.
2.10. Determination of Chitosan–siRNA Biodistribution Using Ex-Vivo Organ Imaging
Balb/c nude female (♀) mice aged 6 weeks weighing 20–22 g were used for biodistribution experiments. Test articles (naked siRNA, Invivofectamine®2.0 and chitosan-based NPs) formulated at an N:P:C ratio of 5:1:0 or 2:1:1.5 (Mn 10 and 120 kDa) were injected at a dose of 0.25 mg/kg DY647-labeled siRNA, except for the HA-coated NPs, which were administered at 0.165 mg/kg. DY647 fluorophore was administered at a dose of 0.5 mg/kg. Mice were euthanized 4 h post-administration and immediately perfused using PBS (1 × 20 mL) and 10% Neutral Buffer Formalin (NBF, 1 × 40 mL). Ex vivo imaging on collected organs was performed using a whole animal imaging system mounted with an EMCCD EM N2 camera (NUVU Cameras, Montreal, QC, Canada). Controls included PBS, naked DY647-labeled siRNA, DY647 alone, and commercially available lipid control Invivofectamine®2.0.
2.11. Determination of Chitosan–siRNA Nanoparticle In Vivo Toxicity
Unlike LNPs or liposomes, information on liver (or systemic) toxicity following the administration of uncoated (positively charged) and HA-coated (negatively charged) chitosan NPs is lacking. CD-1® (ICR) female (♀) and male (♂) mice aged 4–5 weeks and weighing 22–24 g were administered test and control articles for toxicity study (7/group; 4 ♀ and 3 ♂). Mandibular blood was collected prior to and 4 h post-administration to prepare serum. Two out of seven mice from each group were euthanized at 4 h (1 ♀ and 1 ♂), and the remaining five (4 ♀ and 1 ♂) were euthanized 24 h post-administration. At each time point (4 versus 24 h), total circulating blood volume (tCBV) was collected by intra-cardiac puncture, and organs were harvested and washed in PBS. One half was immediately stored in liquid nitrogen (LiqN), and the second half was fixed in 10% NBF.
2.12. Hematological and Serological Parameters
The total circulating blood volume was split into lithium heparin and serum separation tubes, serum separated, for the comprehensive complete blood count and the “CC4” clinical chemistry panels using a Sysmex XTV 2000 (Sysmex, Mississauga, ON, Canada) and Beckman AU680 analyzers (Beckman Coulter Ltd., Mississauga, ON, Canada).
2.13. Determination of Cytokine Levels
Serum samples collected at 0 (baseline) and 4 h post-administration of test articles were assayed for pro-inflammatory cytokines (TNF-α, IL-1β, IL-6, KC and IFN-γ) using the Luminex® technology. Plates were designed using the Bio-Plex® assay builder (Bio-Rad Laboratories, Mississauga, ON, Canada), which was followed by the manufacturer’s QC. For each plate, a standard curve was prepared by diluting the Bio-Plex® Pro Mouse Cytokine Standard 23-Plex in the Bio-Plex® in standard diluent followed by 4-fold serial dilutions from 1:4 to 1:65536 in the same diluent. Samples were thawed on ice, cleared by centrifugation (10,000× g, 10 min, 4 °C), and diluted 1:4 using the Bio-Plex® Sample diluent (Bio-Rad Laboratories, Mississauga, ON, Canada), and a volume of 20 μL was transferred to assay plates prefilled with pooled capture antibodies. The plates were incubated for 30 min under orbital shaking (800 rpm, RT), washed as per the manufacturer’s recommendation using a Bio-Plex® Pro II Wash Station (Bio-Rad Laboratories, Mississauga, ON, Canada), incubated with biotinylated detection antibodies (30 min, 800 rpm, RT), washed and revealed post-incubation for 10 min with streptavidin–phycoerythrin (800 rpm, RT). Data were acquired on a Bio-Plex® 200 system using the RP1 PMT setting (Bio-Rad Laboratories, Mississauga, ON, Canada) with a minimum of 50 beads per region analyzed. A standard curve was prepared using serial dilutions, and a 5-parameter regression algorithm was used to fit the data and interpolate each cytokine value in serum samples. To account for inter-plate variability, two samples (e.g., one LPS and one Invivofectamine®2.0 (8 mg/kg) sample) were used as inter-plate calibrators.
2.14. Determination of Chitosan–siRNA Nanoparticle In Vivo Efficacy
Balb/c male (♂) mice aged 6–7 weeks and weighing 22–25 g were used for an efficacy study. Uncoated anti-GAPDH NPs (92-10-5 and 92-120-5) and HA-coated NPs (HA92-10) were administered at 1 mg/kg (uncoated) and 8 mg/kg siRNA (HA-coated) every other day for a total of three injections. Naked anti-GAPDH siRNA (siGAPDH) and Altogen lipid NPs (Altogen LNP) were I.V. administered at 2.5 mg/kg every other day for a total of three injections. The liver-targeting Invivofectamine®3.0 lipid NPs (InvLNP) were I.V. injected at 2.5 mg/kg as a single injection. All mice were euthanized 72 h following the last administration to collect tCBV and organs. tCBV was serum separated, and organs were split into halves and stored in LiqN and fixed in 10% NBF before protein extraction, determination of GAPDH enzymatic activity, Western blotting, histology and immunohistochemistry.
2.15. Assessment of GAPDH Enzymatic Activity Using the KDalert® Assay
Frozen tissues were cut on dry ice, weighed (~20 mg), and then disrupted using the 5 mm steel beads and TissueLyzer® II system (Qiagen Inc, Toronto, ON, Canada) at 2 × 30 Hz, 20 s/cycle. Homogenized tissues were re-suspended in 750 μL of KDalert™ lysis buffer (Life Technologies, Burlington, ON, Canada) and incubated on ice for 30 min with inversions every 10 min. Lysates were clarified by centrifugation (2270× g, 30 min, 4 °C), transferred to new tubes, and diluted (1:20) in KDalert™ lysis buffer. A standard curve was prepared by diluting GAPDH stock solution (26 U/mL) with lysis buffer at a 1:100 ratio (GAPDH:Lysis), which was followed by 2-fold serial dilutions from 1:5 to 1:320. Twenty microliters of diluted samples, and standards, were transferred into 96-well plates and 180 μL of the KDalert™ Master Mix (Life technologies, Burlington, ON, Canada) was pipetted into each well. Plates were incubated for 15 min at RT, and absorbance was measured at 610 ± 10 nm using a TECAN Infinite® F-500 microplate system (Tecan Systems, Mannedorf, Switzerland). The GAPDH activity was computed from the standard curve and normalized to the total protein content of the lysate sample as determined using the BioRad DC Protein assay kit (Bio-Rad Laboratories, Mississauga, ON, Canada).
2.16. Western Blotting
The affinity-purified monoclonal antibodies used were against GAPDH and vinculin. Kidney cortices were excised, homogenized using the TissueLyzer® II system (Qiagen Inc., Toronto, ON, Canada) as described above, suspended in KD Alert lysis solution (Life technologies, Mississauga, ON, Canada), and centrifuged at 2270 g for 30 min at 4 °C. The supernatant was quantified using a BioRad DC Protein assay kit (Bio-Rad Laboratories, Mississauga, ON, Canada) and diluted in SDS buffer containing a final concentration of 62 mM Tris (hydroxymethyl)-aminomethane, 0.1 M SDS, 8.7% glycerol, 0.09 mM bromophenol blue, and 0.04 M dithiothreitol (DTT). The samples were heated for 5 min at 90 °C, loaded into Protean mini TGX SDS-PAGE (4–12%) gradient polyacrylamide gels (Bio-Rad Laboratories, Mississauga, ON, Canada), and overnight wet transferred to Amersham™ HyBond® P PVDF membranes (GE Lifesciences, Mississauga, ON, Canada). Membranes were dried and blocked for 1 h at room temperature in 5% non-fat milk, probed overnight at 4 °C with anti-GAPDH primary antibody (1:1000), washed (3X, 15 min, 1% Triton in the presence of blocking buffer), and incubated with HRP-conjugated anti-rabbit IgG1 secondary antibody (1:500) for 1 h, washed, revealed using the Clarity Max™ ECL substrate (Bio-Rad Laboratories, Mississauga, ON, Canada) and visualized using the ChemiDoc MP™ system (Bio-Rad Laboratories, Mississauga, ON, Canada). Protein band quantification was performed using ChemiDoc MP software.
2.17. Clinical Signs and Body Weight
Mice clinical signs were determined for 4 h post-administration of test articles and at euthanasia. Scores for clinical signs—body condition, general aspect, natural behavior, and provoked behavior—were recorded by trained personnel and qualified animal care technicians. The Mouse Grimace Scale (MGS) was also used for the scoring of clinical signs in case of distress. Bodyweight was recorded prior to each injection and at euthanasia and expressed as a percent change relative to the previous injection.
2.18. Histology and Immunohistochemistry
Samples were fixed in 10% NBF, embedded in paraffin to collect 5 μm sections followed by hematoxylin and eosin staining and immunohistochemical analysis of GAPDH (Ab181602, 1:250 dilution). Prior to immunohistochemistry, antigen retrieval was performed with 10 mM Tris/1 mM EDTA pH 9 at 60 °C. Sections were blocked with 20% (v/v) goat serum/0.1% (v/v) Triton X-100/PBS for 1 h at room temperature and then incubated for 16 h at 4 °C with Rabbit monoclonal anti-GAPDH diluted 1:250 in 10% (v/v) goat serum/0.1% (v/v) Triton X-100/PBS. Sections were then incubated for 1 h at room temperature with biotinylated goat anti-Rabbit IgG (Ab97049) diluted 1:500 in 10% (v/v) goat serum/0.1% (v/v) Triton X-100/PBS. Revelation was performed with the Vectastain Avidin Biotin Complex (ABC)–alkaline phosphatase (ALP) and AP Red substrate kits (Vector Laboratories, Burlingame, CA, USA). Sections were counterstained with a Weigert Iron Hematoxylin prior to dehydration, clearing and mounting. Slides were scanned using a NanoZoomer digital slide scanner (Hamamatsu, Boston, MA, USA) and visualized using the NDP® view 2.0 software (Hamamatsu, Boston, MA, USA).
2.19. Confocal Laser Scanning Microscopy
For the in vivo biodistribution and subcellular localization of DY647-labeled siRNA, organs were cryosectioned (5 μm), actin stained using AlexaFluor 546 phalloidin and mounted with ProLong® Diamond antifade containing DAPI. Sections were imaged in multitrack mode using a Zeiss LSM 510 META confocal Axioplan 200 microscope (Carl Zeiss AG, Feldbach, Switzerland).
2.20. Statistical Analysis
Data were collected and expressed as average ± standard deviation. Statistical analysis was conducted using a GraphPad Prism® 7.0 (GraphPad Software Inc., La Jolla, CA, USA) software package. Unless otherwise stated, one-factor ANOVA followed by Dunnet’s test for multiple comparisons was performed on collected data.
4. Discussion
Here, we investigated the effect of chitosan polymer length, dose, and surface modification with HA on the hemolytic potential, acute and organ toxicity, cytokine induction, in vivo biodistribution and target knockdown efficacy in addition to comparing chitosan NPs with commercially available cationic LNPs (Invivofectamine®). Taken together, our data showed that uncoated chitosans (high and low Mn, 92% DDA) are safe, well tolerated, non-immune stimulating delivery systems that target kidney PTECs to achieve significant functional knockdown in kidney cortices.
Hemolytic and hemagglutination properties have been well characterized for cationic polymers such as PEI [
33,
34] and chitosan [
10] with chitooligosaccharide (Mn < 5 kDa) found to be non-hemolytic but causing dose-dependent erythrocyte aggregation [
35]. Additionally, our previous in vitro study demonstrated the non-genotoxic effect of HA-coated chitosan NPs [
36]. Here, we show that uncoated chitosan NPs display dose- and molecular weight-dependent hemolytic and hemagglutination properties that could be abrogated with the use of NPs prepared at a low N:P ratio or HA coating (
Figure 2), highlighting careful dosing to avoid hemotoxicity and/or embolism. The maximum siRNA dose that could potentially be intravenously administered with chitosan was found to depend on the Mn, N:P ratio and HA coating. According to PEGylated LNPs standard, a hemolytic index below 5% is regarded safe [
34]. Consequently, doses of 5 and 1 mg/kg siRNA could be administered with low and high Mn chitosan, respectively, when formulated at N:P 5, while doses of at least 8 mg/kg siRNA could be used for N:P 2 and HA-coated NPs (
Figure 2). The hemolytic/hemagglutination potential of chitosan could occur through the interaction with negatively charged erythrocyte (RBC) membranes via a pore-forming mechanism, followed by an osmotic shock, and/or through the regulation of the surface protein and increase in surface roughness, as demonstrated before [
35]. Moreover, the interaction between chitosan amino and acidic groups on erythrocytes could promote polyelectrolyte complex formation causing RBC aggregation as seen for other biomaterials [
37]. NP coating with HA, a biocompatible and negatively charged molecule, eliminated both hemolysis and RBC aggregation possibly due to limited interaction with erythrocyte membranes through electrostatic repulsion and reduced interaction with serum components. Unlike uncoated chitosan, LNPs did not show dose dependent hemolysis, which was probably due to surface PEGylation implied by the quasi-neutral ζ-potential ~ 8–10 mV. Shielding with PEG has been the method of choice to limit LNP hemolysis with high PEG density required for improved biocompatibility and reduced cytokine induction [
38] and is incorporated in most LNPs available commercially or in clinical development. Although the exact composition of InvLNPs is not disclosed by the manufacturer, a quasi-neutral surface charge is probably associated with PEGylation or an increased molar ratio of neutral to cationic lipids in the formulation. Unlike PEGylation, electrostatic coating with HA demonstrated a similar protective effect, permitting a dose increase to at least 8 mg/kg.
The immune-stimulating properties of NPs, or their payloads, monitored through the expression of cytokines in plasma, serum or target tissues [
39,
40,
41,
42] represent a major hurdle for clinical translation. Our uncoated and HA-coated chitosan NPs did not induce type-I pro-inflammatory cytokines (IL-1β, TNF-α, INFγ and IL-6) except for a small, statistically insignificant increase in KC, which is a human IL-8 homologue indicating a non-immunogenic effect 4 h post-administration. Since KC has distinct target specificity for neutrophils [
43,
44], the absence of neutrophil invasion, 24 h post-administration, in organs where chitosan had accumulated suggests an epithelial cell-independent mechanism of KC expression.
The adjuvant and immune stimulating effect of CS involves the activation of DCs and the secretion of type-I cytokines through NLRP3 inflammasome activation and the recently discovered cGAS/STING pathway for lower DDA (80%) chitosans [
11,
12,
45,
46,
47]. An apparent contradiction between the lack of cytokine activation here and the literature could be explained by differences in routes of administration, dose, DDAs and priming of immune cells. For instance, most studies demonstrating the anti-allergic properties of chitin and chitosan (Th2 inhibition) via the expression of type-I cytokines have been tested in vitro and/or using the intranasal, intraperitoneal, intraocular and intravaginal routes of administration [
12,
45]. However, in all these studies, priming strategies were used and could explain cytokine induction consistent with the finding that chitosan stimulated significant cytokine release only from primed BMMΦ [
46]. Here, we did not measure cytokine levels at subsequent time points, which could also explain the absence of cytokine induction, which only appeared around 9 h and peaked 24 h post-stimulation [
12]. Other considerations such as Mn, contaminants, particle size may also contribute to the observed difference.
LNPs and liposomes possess potent immune stimulation governed by the lipid and cationic head groups and/or the combination with the nucleic acid payload [
38,
39,
40,
41,
48]. In this study, Invivofectamine
® LNPs demonstrated a dose-dependent induction of INFγ, IL-6 and KC and a minor TNF-α increase in serum. Immune stimulation was abrogated by 2′Ome-modified siRNA, confirming previous results with LNPs [
39,
40] highlighting major differences with our chitosan system where cytokine induction was not observed with any payload. Since TNF-α—a potent cytokine—is activated by the activation of Toll-Like receptors (TLRs) [
49,
50], the TNF-α stimulation observed with LNP used in this study, while not with chitosan, suggests a TLR-based mechanism of immune induction reminiscent of Chol:DSPC:DOTAP (3:1:1) cationic liposomes [
41].
We then examined the acute toxic effects of chitosan, dose, siRNA sequence, and HA coating on hematological and serological parameters. Hematocrit (HCT) and total hemoglobin (Hb) levels were unchanged versus PBS and within the normal reference ranges of CD-1
® (ICR) mice, indicating a relatively safe and non-hemolytic profile for all formulations tested. Lower Hb but not HCT levels compared with the reference range observed intragroup might be due to differences in gender, age and quantification techniques [
51,
52]. However, Hb levels were normal and comparable to the PBS group. Unlike chitosan NP and their HA-coated form, LNPs used in this study sharply decreased platelet counts, which was consistent with previous observations [
40,
48]. Thrombocytopenia was also observed for anti-sense oligonucleotide (ASO) administered at doses above 200 mg/kg, which resulted in a halt in both the IONIS CARDIO-TTR and the NEURO-TTR phase III trials, and could be traced to the phosphorothioate (PS) backbone modification [
53]. Interestingly, lymphocyte counts decreased with both lipid and chitosan-based formulations when formulated with the native immune stimulatory [
39] anti-ApoB sequence.
Chitosan accumulation in the kidney did not impair kidney function, since levels of BUN and creatine remained normal. However, a drawback of our study is the lack of BUN and creatinine measurements in urine, which are more predictive than their serum counterparts, as they permit the computation of the glomerular filtration rate (GFR), which is a clinical indicator of renal function. BUN and creatinine, indirect indicators of liver health, support the absence of liver toxicity indicated by normal ALT, AST and ALP levels. Unlike uncoated and HA-coated NPs, LNPs showed a typical dose-dependent increase in transaminases, indicating transient liver toxicity [
41,
42,
48] further accompanied by a reduction in body weight highlighting systemic (liver) toxicity. The decrease in body weight observed with LNPs [
40,
41,
42,
48] could be attributed to either the lipids [
41,
42,
48] or the properties of the encapsulated nucleic acid payload [
40]. In the present study, the decrease in body weight could be due to the general toxicity induced by the lipid system, since injections were performed with a LNA-modified sequence containing 2′Ome and phosphorotioates (PS). LPS treatment increased BUN and decreased Cr levels in serum typical of catabolic processes following the induction of cytokines in fever like symptoms or infections [
54]. The decrease in BW with LPS treatment could possibly be linked with elevated cytokine levels compared with other groups that had lower (i.e., InvLNP) or no cytokine release (i.e., uncoated and HA-coated NPs). The decrease in alkaline phosphatase (with LPS detoxifying properties) following the I.V. injection of LPS could be due to malnutrition and weight loss and correlates with overt clinical signs and changes in the general appearance of mice (
Table S2).
Organ and tissue toxicity is generally recognized by morphological changes, immune infiltration, apoptosis and/or necrosis. In the current study, no morphological changes, including an absence of infiltrating neutrophils, apoptotic and/or necrotic cells were observed in main organs upon single (
Supplemental Figures S4–S6) and multiple injections, further confirming the safety of uncoated and HA-coated NPs. However, immune infiltration in liver was observed with high doses (8 mg/kg) of Invivofectamine
® 2.0 (
Supplemental Figure S2) supporting immune stimulation data.
Intravenous administration caused chitosan siRNA NPs accumulation in the kidneys and promoted siRNA translocation through the glomerular basement membrane (GBM) evidenced by the intracytoplasmic localization and punctuate pattern of siRNA. The PTEC internalization of chitosan has been previously demonstrated to be dependent on the glucosamine (Glc)–megalin interaction and subsequent endocytosis [
27]. HA coating modified the physicochemical properties of NPs, with a shift in size and ζ-potential indicating effective electrostatic coating, without modifying the kidney-targeted biodistribution pattern possibly via CD44 internalization. Indeed, PTECs express at least five CD44 splice variants playing an important role in HA internalization [
55]. In addition, the HA-dependent colloidal stability of NPs in serum [
23] could decrease in circulation due to shedding, exposing the chitosan–siRNA core (N:P 2) to accumulate in PTEC via megalin-mediated endocytosis. Independent of the observed PTEC accumulation, the mechanism of NP translocation through GBM still remains unclear, since fenestration and ECM restrict the translocation and diffusion of NPs. Alternative delivery through the fenestrated peritubular capillaries could occur but faces diffusion challenges through the negatively charged interstitium.
We next examined the efficacy of our NPs to induce target-specific knockdown. The glyceraldehyde 3-phosphate (GAPDH) gene was selected as a target due to its ubiquitous expression in tissues and the availability of in vivo validated, and chemically modified, siRNA sequences. In this study, functional GAPDH knockdown in the kidney cortex was achieved upon three injections of uncoated NPs. GAPDH enzymatic activity was reduced in the kidney lysate by around 55% and 45% using low (10 kDa) and high (120 kDa) Mn chitosan, respectively, which was confirmed by the Western blot analysis and qualitative immunohistochemistry. Unlike uncoated chitosan, HA-coated NPs accumulated in the kidney but did not induce target knockdown. This result could be explained by the need of excess chitosan (N:P 5 in uncoated vs. 2.5 in HA-coated) to promote endosomal release [
56], which is possibly through the proton sponge effect. Therefore, it is likely that HA-coated NPs formulated at an N:P:C ratio of 2.5:1:2 can translocate to the cytoplasm of PTEC but remain sequestrated in endolysosomal compartments due to the poor endosomal buffering capacity and reduced proton sponge effect. In addition, the negatively charged HA molecule, if co-localizing with chitosan, could contribute to lower endosomal release by masking positive charge in the endosome, therefore reducing the capacity of endocytosed chitosan to mediate endosomal rupture. In contrast to HA-coated NPs (N:P:C ratio of 2.5:1:2), uncoated chitosan formulations prepared at an N:P ratio of 5 contain around 70% free chitosan [
57] that could co-localize in PTEC endosomes and promote endosomal rupture, explaining the observed efficacy. In contrast to chitosan, Invivofectamine
® LNPs accumulated in liver (
Supplemental Figure S2) and induced target knockdown (
Supplemental Figure S7), as seen before [
58]. Lower knockdown levels with LNPs in this study than seen before [
58,
59] could be explained by differences in target gene half-lives (t1/2).
Compared to the potency of LNPs in biopharmaceutical pipelines (~70–90%) [
60,
61], the lower functional target knockdown obtained with our system (~50–60%) could be explained by the half-life of the target gene, potency of the siRNA, and tissue-dependent technical challenges. Given that chitosan accumulates in PTECs (minor cell subtype of the kidney) versus LNPs in hepatocytes (predominant cell-type in the liver), an assessment of target knockdown using conventional techniques (e.g., qPCR, enzymatic activity, Western blotting) that average expression levels across all cell types in the tissue sample is inevitably underestimated. Therefore, the functional knockdown obtained in this report underestimates the true efficiency of our system to silence a target gene in PTECs, highlighting that the precise evaluation of knockdown requires the development of methods capable of estimating knockdown in a specific subset of cells composing an organ.
Taken together, our findings are critically important in revealing that uncoated and HA-coated NPs display no toxicity along with the extrahepatic delivery of siRNA leading to functional knockdown in kidney cortices. The efficacy of our uncoated system in inducing functional target knockdown in PTECs specifically differentiates it from cyclodextrin-based NPs accumulating in the glomerulus and podocytes [
62]. This study also highlights the potential of the HA-coated chitosan hybrid system as a potential system that accumulates in the kidney and could be delivered at high doses without hemolytic and/or adverse events. Further investigation is needed to elucidate the mechanism of PTEC accumulation and lack of knockdown efficacy observed with the HA-coated system in this report despite similar distribution properties.