2023 Volume 46 Issue 10 Pages 1468-1478
Most retinal diseases involve the degeneration of choroidal retinal pigment epithelial (RPE) cells. Because of a blood–retina barrier (tight junction formation), RPE cells restrict the entry of hydrophilic macromolecules (e.g., small interfering RNA (siRNA)) through blood stream and eye drops. A cytoplasm-responsive stearylated (STR) peptide, STR-CH2R4H2C (CH2R4) enables stable siRNA complexation, cell permeation, and intracellular dynamics control. We previously demonstrated how CH2R4-modified liposomes promoted siRNA efficacy. We investigated the influence of amino acid sequences of functional peptides on cellular uptake pathways, siRNA transfection efficacy, and the permeation of peptide-modified liposomes in rat RPE-J cells. Four STR-peptides, consisting of arginine (R), cysteine (C), histidine (H), lysine (K) or serine (S), were designed based on CH2R4. We prepared siRNA-loaded, peptide-modified cationic liposomes (CH2R4-, CH2K4-, CH2S4-, SH2R4-, and SH2S4-lipoplexes). CH2R4-, CH2K4-, and SH2R4-lipoplexes induced cellular uptake by macropinocytosis by activating cytoskeletal F-actin, possibly due to cationic amino acids (arginine, lysine). SH2R4-lipoplexes were trapped in endosomes, whereas CH2R4- and CH2K4-lipoplexes enhanced endosomal siRNA release suggesting cysteine contributes to endosomal escape. Although cationic liposome-based, CH2S4- and SH2S4-lipoplexes (not including arginine and lysine) showed lower siRNA transfection efficiency. This difference may be because siRNAs were retained on both peptide moieties and cationic liposomes in CH2R4-, CH2K4- and SH2R4-lipoplexes, whereas in CH2S4- and SH2S4-lipoplexes, siRNAs were loaded to the cationic liposomes, but not on peptides. In three-dimensional spheroids, CH2R4- and CH2K4-modified liposomes promoted permeation through tight junctions. Thus, cationic amino acids and cysteine within peptide sequences of CH2R4 could be effective for siRNA delivery to the retina using functional peptide-modified liposomes.
Retinal diseases (e.g., age-related macular degeneration (AMD), diabetic retinopathy) are mostly induced by the degeneration of choroidal retinal pigment epithelial (RPE) cells located in the outermost layers of the retina.1–3) What is known as wet AMD is caused by a vascular disorder due to various factors such as genetics, aging, smoking, glycation, and drusen.4–6) Such sustained stresses induce the formation of immature new blood vessels through the secretion of vascular endothelial growth factor (VEGF) from the RPE. This neovascularization promotes hemorrhage from the retina, resulting in visual field loss and blindness.1) The RPE forms a rigid tight junction called the blood–retina barrier (BRB), which strictly prevents the entry of substances (e.g., hydrophilic high-molecular-weight drugs) via the bloodstream and eye drops.7–10) Therefore, the intravitreal administration of anti-VEGF drugs (e.g., anti-VEGF antibody) is the current first choice of drug therapy and the only treatment to prevent the progression of AMD. However, intravitreal administration has risks, such as elevation of the intraocular pressure and blood pressure, retinal detachment, infection, and physical and mental burdens.11–13)
A recent focus on nucleic acid medicine (e.g., small interfering (si)RNA, antisense RNA) as next-generation candidates is to develop the more effective delivery systems. Small interfering RNA can inhibit target protein expression by complementarily binding to mRNA and cleaving it, but has to overcome disadvantages of high clearance, low stability, and low bioavailability in tissues.14,15)
Cationic materials (lipids, polymers, and peptides) improve the stability (nuclease resistance) via complexation, delivery efficiency (cellular uptake, cell penetration), safety, and transfection efficacy of nucleic acid medicines (e.g., antisense RNA, plasmid DNA, siRNA).15–22) The cationic functional peptide, stearyl-CH2R4H2C (CH2R4), is an applicable cytoplasm-responsive, stearylated peptide for effective siRNA delivery via intravenous or transdermal administration.23–26) The CH2R4 consists of arginine (R), histidine (H), cysteine (C), and stearic acid (STR). Arginine forms stable complexes with siRNA due to its positive surface charge and promotes cellular uptake.27–31) Histidine induces the endosomal escape of siRNA through the protonation of an imidazole group via a change in pH from a neutral to acidic environment (pH 7.4 to 5.5). An influx of protons into the endosome leads to explode the endosomal membrane due to an increase in osmotic pressure and releases siRNA (proton sponge effect).16,32–34) Cysteine is bound to both ends of the CH2R4 sequence and enhances the stability of siRNA complexes with CH2R4 by forming an intermolecular disulfide bond. Cysteine dissociates siRNA in the acidic pH of endosomes.31,35–38) We previously indicated that CH2R4-modified liposomes were a promising nanocarrier for siRNA delivery to the posterior segment of the eye by non-invasive instillation.39) Additionally, CH2R4-modified liposomes showed potential for migration to the posterior segment of the eye by rat eye instillation. This finding holds promise for improving drug delivery to the retina, which is generally difficult with eye drops. However, the contributions of CH2R4 in improving cellular uptake, increasing siRNA transfection efficacy, and suppressing VEGF expression in rat RPE-J cells are unclear.
This study aimed to investigate the effect of the amino acid sequences of functional peptides on the cellular uptake pathway, siRNA transfection efficacy, and permeation enhancement of peptide-modified liposomes in rat RPE-J cells. To this end, CH2R4 and four types of STR-peptides based on CH2R4 were used. The cellular uptake pathway and siRNA transfection efficiency of siRNA-loaded, peptide-modified liposomes (peptide-lipoplexes) were evaluated in rat RPE-J cells. Further, permeation by the peptide-lipoplexes was observed using three-dimensional (3D) RPE spheroids that mimic the BRB.
Five types of STR-peptides, consisting of (R), (H), (C), (K), or serine (S), were used: stearic acid (STR)-CHHRRRRHHC (CH2R4), STR-CHHKKKKHHC (CH2K4), STR-CHHSSSSHHC (CH2S4), STR-SHHRRRRHHS (SH2R4), and STR-SHHSSSSHHS (SH2S4) were specially ordered to BEX Co., Ltd. (Tokyo, Japan).31,32) 1,2-Dioleoyl-sn-glycero-3-phosphocholine (DOPC; #850375P), 1,2-Dioleoyl-3-trimethylammonium-propane (DOTAP; #890890C), and cholesterol (CHO; #700100P) were purchased from Avanti Polar Lipids (Alabaster, AL, U.S.A.). ATTO647N DOPE (ATTO-DOPE; #AD647N-16), used for the fluorescent labeling of liposomes, was obtained from ATTO-TEC GmbH (Siegen, Germany). LipoTrust™ EX Oligo (LipoTrust; #LEO-01) was purchased from Hokkaido System Science Co., Ltd. (Hokkaido, Japan). Special-grade chloroform, isopropanol, and ethanol were from Nacalai Tesque Inc. (Kyoto, Japan). Cyanine 5-labeled siRNA (Cy5-siRNA; sense, 5′-Cyanin5 AUC CGC GCG AUA GUA CGU AdTdT-3′, antisense, 5′-UAC GUA CUA UCG CGC GGA UdTdT-3′) and rat VEGF siRNA (siVEGF; sense, 5′-CUU CCA GAA ACA CGA CAA AdTdT-3′, antisense, 5′-UUU GUC GUG UUU CUG GAA GdTdT-3′) were purchased from BIONEER Co. (Daejeon, Korea). Negative control siRNA (siControl; #BIN SN-1003 AccuTarget™ Negative Control siRNA) was purchased from Thermo Fisher Scientific Inc. (Waltham, MA, U.S.A.). SYBR®Green I Nucleic Acid Stain was purchased from Lonza (# 50513, Switzerland). Dextran sulfate sodium salt from Leuconostoc spp. was purchased from Sigma-Aldrich Japan K.K. (Tokyo, Japan). The rat retinal pigment epithelial cells (rat RPE-J cells, CRL-2240™) were obtained from the American Type Culture Collection (Manassas, VA, U.S.A.). Dulbecco’s modified Eagle medium (DMEM; #08458-16, high glucose, Nacalai Tesque Inc.), fetal bovine serum (FBS; #SH30910.03, Cytiva, Tokyo, Japan), penicillin (100 U/mL)–streptomycin (100 µg/mL) solution (#2625384, Nacalai Tesque Inc.), 1% non-essential amino acid solution (NEAA; #0634456, Nacalai Tesque Inc.), heparin (#081-00136, FUJIFILM Wako Pure Chemical Corp., Osaka, Japan), LysoTracker™ Red DND-99 (#L7528, Thermo Fisher Scientific Inc.), SlowFade® antifade reagent (#S36937, Thermo Fisher Scientific Inc.), Hoechst® 33342 (#H342, DOJINDO Laboratories, Kumamoto, Japan), Opti-MEM® (#31985070, Thermo Fisher Scientific Inc.), and Acti-stain™488 Phalloidin (#PHDG1-A, Cytoskeleton Inc., Denver, CO, U.S.A.) were used in cytological assays. Chlorpromazine hydrochloride (CPZ; #033-10581, FUJIFILM Wako Pure Chemical Corp.), 5-(N-ethyl-N-isopropyl) amiloride (EIPA; #14406, Cayman Chemical), and chloroquine diphosphate (CQ; #038-17971, FUJIFILM Wako Pure Chemical Corp.) were used in intracellular dynamics study. The DC™ Protein Assay Kit was purchased from Bio-Rad Laboratories Inc. (#5000112JA, Hercules, CA, U.S.A.). The rat VEGF enzyme-linked immunosorbent assay (ELISA) was R&D Systems Inc. (#RRV00, Minneapolis, MN, U.S.A.). Rabbit anti–zonula occludens-1 (ZO-1) polyclonal antibody was purchased from Proteintech Group Inc. (#1773-1-AP, Rosemont, IL, U.S.A.). Goat anti-rabbit immunoglobulin G (IgG) H&L (Alexa Fluor® 555; #A21428) was purchased from Thermo Fisher Scientific Inc. Blocking one (#03953-95) was obtained from Nacalai Tesque Inc.
Preparation of Functional Peptide-LP–siRNA LipoplexesEach lipid was independently dissolved in chloroform (5 mg/mL) and stored at 4 °C until use. DOTAP, DOPC, and CHO solutions were mixed at a molar ratio of 0.25/1.55/1.2 (µmol). ATTO647N-DOPE (0.1 mol% to total lipid) was added only when evaluating liposome behavior. Each mixture was dried with a rotary evaporator in vacuo.40) The thin lipid films obtained were hydrated with 10 mM N-(2-hydroxyethyl)piperazine-N′-2-ethanesulfonic acid (HEPES) buffer (pH 8.0) and sonicated (BRANSON Digital Sonifier®; Central Scientific Commerce Inc., Tokyo, Japan) for 3 min at 30 W, yielding cationic liposomes (cationic-LP). The STR functional peptides, consisting of different amino acid sequences, were dissolved in nuclease-free water and gently dropped into liposomes while vortexing (2 mol% peptide to total lipid). Free peptides were removed by ultrafiltration (13000 × g for 3 min; Amicon® ultra 100 K; Sigma-Aldrich Japan K.K.), yielding five types of functional peptide-modified liposomes (CH2R4-LP, CH2K4-LP, CH2S4-LP, SH2R4-LP, SH2S4-LP).41) Subsequently, siRNA was added to each liposome mixture and the mixture was then purified by ultrafiltration (13000 × g for 1 min; Amicon® ultra 100 K). Functional peptide-LP–siRNA lipoplexes (CH2R4-lipoplex, CH2K4-lipoplex, CH2S4-lipoplex, SH2R4-lipoplex, SH2S4-lipoplex) were obtained. Each lipoplex was composed of a molar ratio of nitrogen (N) on lipid and peptide to phosphate (P) on siRNA (N/P ratio) of 5 : 1. This N/P ratio was appropriate to produce reproducible lipoplexes without aggregation (data not shown) and enable to load large amount of siRNA for single eye drop in comparison with other N/P ratio.
Physical Properties of LipoplexesPeptide-lipoplexes were diluted to an adequate concentration. The mean particle size, polydispersity index and zeta-potential were measured by a Zetasizer Nano-ZSP (Malvern Panalytical Ltd., Malvern, U.K.). The siRNA loading efficiency was determined by SYBR Green Assay.42) Briefly, each lipoplex sample (50 µL) and 500-fold diluted SYBR® Green (50 µL) were mixed in a well of a 96-well plate for 15 min at room temperature in the dark. The fluorescence intensity (F1) of each sample was measured by a microplate reader (Ex:494 nm, Em:521 nm; Varioskan Flash 2.4; Thermo Fisher Scientific). Subsequently, 5 mM dextran sulfate (25 µL) in 100 mM MES–HEPES buffer (50 mM MES, 50 mM HEPES, 75 mM NaCl, pH 7.2) was added to each sample and the fluorescence intensity (F2) was measured. The F1 reflects the amount of siRNA interacting with cationic liposomes because the SYBR Green is not allowed to intercalate when rigid complexes are formed between siRNA and cationic peptide due to electrostatic interaction.42) The F2 reflects the total amount of siRNA included in the lipoplexes because the polyanion (e.g., dextran sulfate) allows release of siRNA from complexes due to interaction with cationic peptides. Using an absolute calibration curve method (siRNA standard solution conc., 0.125–4.0 µg/100 µL), the amount of siRNA (W1, W2, W3) in the samples were determined. W1, W2 and W3 were calculated as the amounts of siRNA loaded in cationic liposomes (W1), lipoplexes (W2; total siRNA amount in lipoplex), and peptide moieties on the surface of liposomes in the lipoplex (W3 = W2–W1). The siRNA loading efficiency (%) was defined as the percentage of measured W2/or W3 to the formulated amount of siRNA.
Cell Culture and TreatmentsRat retinal pigment epithelial cells were grown in DMEM containing 10% FBS and 1% NEAA for 24 h at 33 °C in a humidified 5% CO2 atmosphere before treatment with samples. When washing cells after treatment with samples, phosphate buffered saline (PBS) containing 20 U/mL heparin was used as a washing buffer to remove lipoplexes attached to the surface of cells after treatment with samples.41,43) In cellular uptake experiments, cells (2.0 × 105 cells/well) in 24-well plates were treated with 100 µL of lipoplex sample (Cy5-siRNA, 0.158 µM or ATTO647N-labeled liposome, 0.2 µM) in 900 µL serum-free DMEM for 1 or 4 h at 33 °C and 5% CO2. After washing with PBS, cells were collected by treatment with 0.25% trypsin-ethylenediaminetetraacetic acid (EDTA). Mean fluorescence intensities (MFI) were measured using a flow cytometry (FACS Celesta; Becton Dickinson & Co., NJ, U.S.A.). The MFIs in 10000 events per sample were used for the evaluation of cellular uptake of lipoplexes and siRNA. In an analysis of cellular uptake pathways (clathrin-mediated endocytosis and macropinocytosis routes), cells (2.0 × 105 cells/well) in 24-well plates were treated with 900 µL serum-free DMEM containing endocytosis inhibitors (CPZ, 14 µM; EIPA, 100 µM) for 30 min at 33 °C and 5% CO2.35,43) and then incubated with 100 µL of lipoplex sample (Cy5-siRNA, 0.158 µM) for 1 h at 33 °C and 5% CO2. After washing with PBS, cells were collected by treatment with 0.25% trypsin-EDTA. The MFI of Cy5-siRNA was measured using flow cytometry.
In observations of endosomal escape, rat RPE-J cells (3.0 × 105 cells/well) grown on glass coverslips in 6-well plates were used. Cells were treated in the presence or absence of inhibitors (CPZ, 14 µM; EIPA, 100 µM; for 30 min) and then incubated with 100 µL of lipoplex sample (Cy5-siRNA, 0.079 µM) in 1900 µL serum-free DMEM for 1 h at 33 °C and 5% CO2.44) Additionally, cells (2.0 × 105 cells/well) in 24-well plates were treated with 1900 µL serum-free DMEM containing CQ (200 µM) and 100 µL of lipoplex sample (Cy5-siRNA, 0.079 µM) for 4 h at 33 °C and 5% CO2. After washing with PBS, treated cells were incubated with serum-free DMEM containing 75 nM LysoTracker™ Red DND-99 for 1 h to stain endosomes (lysosomes).44,45) After washing with PBS, cells were fixed with 4% paraformaldehyde and sealed on glass slides with SlowFade® containing 1% Hoechst® 33342. The intracellular localization of Cy5-siRNA was observed using a confocal laser scanning microscope (CLSM; FV1000D IX81; OLYMPUS, Tokyo, Japan). In observations of F-actin activation (macropinocytosis), cells (3.0 × 105 cells/well) were treated with 100 µL of lipoplex sample (Cy5-siRNA, 0.079 µM) in 1900 µL Opti-MEM for 4 h at 33 °C and 5% CO2. After washing with PBS, the cells were fixed with 4% paraformaldehyde, washed again and incubated with 0.5% Triton X-100 (2 mL) for 15 min at room temperature. After washing with PBS, cells were treated with Acti-stain™488 phalloidin (100 nM) for 1 h at room temperature to stain cytoskeletal F-actin.46–48) The cells were sealed on glass slides with SlowFade® containing 1% Hoechst® 33342. The fluorescence of Cy5-siRNA and F-actin was observed using a CLSM.
Inhibition of VEGF Expression in Rat RPE-J Cells by siVEGF-LipoplexRat retinal pigment epithelial cells (3.0 × 105 cells/well) in 6-well plates were treated with 100 µL of CH2R4-, CH2K4- or SH2R4-lipoplex sample (siVEGF or siControl, 0.079 µM) in 1900 µL Opti-MEM for 4 h at 33 °C and 5% CO2. After washing with PBS, the cells were incubated in 2000 µL Opti-MEM for 72 h at 33 °C and 5% CO2. These conditions enable the induction of VEGF expression in normal rat RPE-J cells under low nutrient and oxidative stress.49,50) The amounts of secreted VEGF in the medium and total protein of the cells were quantified using a rat VEGF ELISA Kit and DC Protein Assay Kit, respectively, according to the manufacturer’s instructions. The amount of VEGF was normalized to the total protein content. Subsequently, relative VEGF protein expression levels were calculated using the following equation:
VEGF protein expression level (%) = VEGF in treated cells ×100/VEGF in untreated cells (control).
Evaluation of Permeability of Lipoplexes in 3D RPE-Cell SpheroidsLipoplexes were developed to deliver siRNA to the retina, which shows strong tight junction barriers (i.e., BRB) formed by RPE cells. To evaluate permeation enhancement by the modification of functional peptide, multicellular 3D spheroids of rat RPE-J cells were used. Retinal pigment epithelial cells (2.0 × 105 cells/100 µL/well) were grown in low-absorption 96-well round bottom plates (Sumitomo Bakelite Co., Ltd., Tokyo, Japan) for 7 d at 33 °C and 5% CO2 to form rat RPE spheroids.51) After washing twice in PBS, RPE spheroids were treated with 20 µL of a CH2R4-, CH2K4- or SH2R4-lipoplex sample (siControl, 0.317 µM; ATTO647 N-labeled liposome, 0.4 µM) in 80 µL of Opti-MEM for 1 or 4 h at 33 °C and 5% CO2. The spheroids were washed twice in PBS containing heparin and fixed with 4% paraformaldehyde. After washing with PBS, spheroids were incubated with 100 µL Blocking One for 1 h at room temperature and washed twice with 0.5% Triton X-100 (100 µL). To immunostain the tight junction formation factor, ZO-1, the spheroids was treated with primary antibody (unlabeled rabbit anti–ZO-1 antibody, dilution ratio 1 : 100), incubated for 2 h at room temperature and then washed twice with 0.5% Triton X-100 (100 µL). The spheroids were incubated with secondary antibody (AlexaFluor555-labeled goat anti-rabbit IgG, dilution ratio 1 : 1000) and 1% Hoechst® 33342 for 2 h at room temperature. After washing twice with 0.5% Triton X-100, ATTO647N-lipoplexes and AlexaFluor555-stained ZO-1 (tight junction formation) were observed using CLSM.
Statistical AnalysisData are presented as the mean ± standard deviation (n = 3). Statistical analysis was performed using Student's t-test or one-way ANOVA. Statistical significance was defined as p < 0.05.
The five types of STR-peptides used with the following sequences, consisting of (R), (C), (H), (K) or (S), were designed based on CH2R4: STR-CHHRRRRHHC (CH2R4), STR-CHHKKKKHHC (CH2K4), STR-CHHSSSSHHC (CH2S4), STR-SHHRRRRHHS (SH2R4), and STR-SHHSSSSHHS (SH2S4). By adding each peptide (2 mol% to total lipid) to cationic DOTAP/DOPC/CHO liposomes (cationic-LP; size 68.4 ± 31.0 nm, zeta-potential 27.0 ± 9.5 mV), functional peptide-modified liposomes (CH2R4-LP, size 62.6 ± 27.2 nm, zeta-potential 46.1 ± 9.8 mV; CH2K4-LP, 64.2 ± 29.4 nm, 38.8 ± 9.8 mV; CH2S4-LP, 76.6 ± 42.2 nm, 34.7 ± 7.0 mV; SH2R4-LP, 66.1 ± 29.8 nm, 44.2 ± 8.6 mV; SH2S4-LP, 64.9 ± 30.2 nm, 33.9 ± 9.1 mV) were prepared. Then, by incubating with siRNA, CH2R4-lipoplex (size 74.2 ± 33.3 nm, zeta-potential +36.9 ± 11.8 mV), CH2K4-lipoplex (81.0 ± 38.9 nm, +29.6 ± 5.2 mV), CH2S4-lipoplex (115.7 ± 52.8 nm, +19.9 ± 4.3 mV), SH2R4-lipoplex (76.8 ± 37.0 nm, +33.6 ± 8.0 mV), and SH2S4-lipoplex (116.6 ± 51.0 nm, +13.0 ± 4.3 mV) were obtained. An siRNA-loaded cationic-LP was used as a control-lipoplex (118.3 ± 53.8 nm, +9.7 ± 3.6 mV). Cationic peptide (CH2R4, CH2K4, SH2R4) -modified lipoplexes had higher surface charges (approx. +30 mV to +40 mV) than, non-cationic peptide (CH2S4, SH2S4) -modified lipoplexes (approx. +10 mV to +20 mV). However, the cell viability of each lipoplex showed almost 100% by ATP assay (data not shown), indicating that no effect of surface charge of peptide-modified lipoplexes on cytotoxicity in this study. Figure 1A shows the total siRNA loading efficiency of the lipoplexes. Figure 1B shows the calculated siRNA loading efficiency that was retained by interacting with peptide moieties on the surface of liposomes in the lipoplex. In this study, we used the N/P ratio of 5 which enable to load larger amount of siRNA for single eye drop with no effect on physical properties of lipoplexes. All of the lipoplexes showed high total siRNA loading efficiency. In particular, CH2R4-, CH2K4- and SH2R4-lipoplexes showed higher total loading efficiencies of more than 90% compared to CH2S4- and SH2S4-lipoplexes (Fig. 1A). Interestingly, the siRNA loading efficiencies to the peptide moieties changed with differences in the amino acid sequence (Fig. 1B). That is, the calculated loading efficiencies of the CH2R4-, CH2K4- and SH2R4-lipoplexes were approximately 50%, indicating that the siRNA was loaded to both the liposomes and peptides at a ratio of about 1 to 1. In contrast, SH2S4- and CH2S4-lipoplexes showed high total siRNA loading efficiencies of around 70–80%, whereas calculated siRNA loading efficiencies were few. These results suggested that the siRNA was retained on cationic peptides consisting of arginine or lysine in the lipoplexes due to electrostatic interactions despite cationic-LP–based lipoplexes. These differences may cause differences in particle sizes among the lipoplexes. That is, CH2R4-, CH2K4- and SH2R4-lipoplexes had smaller uniform particle sizes of around 70–80 nm (zeta-potential: about +30 mV) than SH2S4- and CH2S4-lipoplexes (100–120 nm, about +10 mV). Comparing CH2R4 and SH2R4 using 4% agarose electrophoresis, CH2R4 formed stable complexes with siRNA and released siRNA after the addition of a reducing agent (to mimic an intracellular reducing environment), while SH2R4 complexes showed a weak interaction in that the siRNA easily dissociated in the presence of a polyanion (data not shown). This may be because the siRNA was retained in the complexes with CH2R4 due to disulfide bond formation between thiol groups in a cysteine. A cysteine generally causes dimerization between molecules or in molecules, while the disulfide bond dissociates in a reducing environment. These results suggest that the cysteine in peptide sequences of CH2R4- and CH2K4-lipoplexes played a role in improving complex stability. These properties may affect the siRNA transfection efficiency.
Small interfering (si)RNA loading efficiency was measured by SYBR Green Assay. Each lipoplex (50 µL) and 500-fold diluted SYBR Green (50 µL) sample was mixed in a 96-well plate and incubated for 15 min at room temperature in the dark. The fluorescence intensity (F1) of each sample was measured by a microplate reader (Ex:494 nm, Em:521 nm). Then, after adding 5 mM dextran sulfate (25 µL), the fluorescence intensity (F2) was measured. Using an absolute calibration curve method, the amount of siRNA in the samples was calculated from F1 and F2. The F1 reflects the amount of siRNA interacting with cationic liposomes. The F2 reflects the total amount of siRNA included in each lipoplex sample (A). F2 minus F1 reflects the amount of siRNA predicted to be loaded on the peptide moiety of lipoplexes (B). Data represent the mean ± standard deviation (S.D.). (n = 3).
Arginine, lysine, and cysteine are known to enhance endocytosis. Thus, the influence of peptide sequences on the cellular uptake of lipoplexes and siRNA transfection efficiency in rat RPE-J cells was evaluated using flow cytometry. We confirmed that each lipoplex at the concentration used in this study was not cytotoxic using an adenosine triphosphate assay (data not shown). CH2R4-, CH2K4- and SH2R4-lipoplexes showed effective cellular uptake by about 10-fold and siRNA transfection efficiency by about 2-fold compared to control-lipoplexes (Figs. 2A, B). In particular, CH2R4- and CH2K4-lipoplexes showed an increase in uptake rate with exposure times longer than 3 h (Fig. 2C). This result may be because siRNAs were retained on the peptide moieties of CH2R4-, CH2K4- and SH2R4-lipoplexes (calculated siRNA-loading efficiencies were about 50%, Fig. 1B).
Retinal pigment epithelial (RPE) cells were treated with 100 µL lipoplex (siRNA, 0.158 µM) in serum-free DMEM for 1 h (A, B) or 1–4 h (C) at 33 °C and 5% CO2. The fluorescence intensity of ATTO647N-labeled lipoplexes or Cy5-siRNA was measured using flow cytometry. Data represent the mean ± S.D. (n = 3). ** p < 0.01 by Student’s t-test. ** p < 0.01. MFI, mean fluorescence intensity.
To evaluate cellular uptake pathways, lipoplexes were transfected into rat RPE-J cells treated with endocytosis inhibitors (CPZ inhibits clathrin-mediated endocytosis, EIPA inhibits macropinocytosis). In cells treated with CPZ, each type of lipoplex tested showed no significant difference in the uptake of siRNA compared to untreated cells, showing a non-clathrin–mediated endocytosis pathway (Figs. 3A, C). In contrast, in cells treated with EIPA, CH2R4-, CH2K4- and SH2R4-lipoplexes showed a significant decrease in uptake of siRNA, indicating lipoplexes consisting of peptides with cationic amino acids used the macropinocytosis pathway (Figs. 3B, C). In CH2R4- and CH2K4-lipoplexes, siRNA uptake markedly decreased compared to that of SH2R4-lipoplexes. This decrease suggests that siRNA transfections through cationic amino acids are enhanced by cysteine because cysteine may induce macropinocytosis due to its intermolecular dimerization.35,36) In untreated cells exposed to CH2R4-, CH2K4- and SH2R4-lipoplexes, independent siRNAs were strongly observed, showing that the peptides in such lipoplexes promoted the release of siRNA from endosomes (Fig. 3C). These improvements may be because the combination of histidine and arginine/or lysine is more effective for endosomal escape.31) In cells treated with EIPA (Fig. 3C), siRNA fluorescence was weakly observed, signifying that uptake of CH2R4-, CH2K4- and SH2R4-lipoplexes was through the macropinocytosis pathway, similar to the results in Fig. 3B.
Cy5-small interfering (si)RNA (0.158 µM) was transfected into retinal pigment epithelial (RPE) cells treated/untreated with 14 µM chlorpromazine (CPZ) that inhibits clathrin-mediated endocytosis (A) and 100 µM 5-N-ethyl-N-isopropyl-amiloride (EIPA) that inhibits macropinocytosis (B). The fluorescence intensity of Cy5-siRNA was measured using flow cytometry. Data represent the mean ± S.D. (n = 3). * p < 0.05, *** p < 0.001 by Student’s t-test. (C) Rat RPE-J cells were pretreated with CPZ or EIPA in 100 µL and incubated with lipoplex (Cy5-siRNA, 0.079 µM) for 1 h. After washing, endosomes and nuclei were fluorescently stained with 75 nM LysoTracker™ Red DND-99 and 1% Hoechst® 33342, respectively. Magnification: 2000×; scale bars: 20 µm; blue: nuclear; red: endosomes; yellow: siRNA. (D) Rat RPE-J cells were incubated with 100 µL lipoplex (Cy5-siRNA, 0.079 µM) in Opti-MEM for 4 h at 33 °C and 5% CO2 and then treated with 100 nM Acti-stain™488 phalloidin for 1 h at room temperature to immunostain the cytoskeletal factor, F-actin. The fluorescence of siRNA and F-actin was observed by a confocal laser scanning microscope (CLSM). Magnification: 2000 ×; scale bar: 20 µm; blue: nuclear; green: F-actin; red: siRNA; white arrows: F-actin lamellipodia. (E) Lipoplexes (Cy5-siRNA, 0.079 µM) were incubated with rat RPE-J cells in the presence/absence of 100 µL chloroquine (CQ) for 4 h at 33 °C and 5% CO2. Magnification: 2000×; scale bars: 20 µm; blue: nuclear; red: endosomes; yellow: siRNA. MFI, Mean fluorescence intensity.
Cell-penetrating peptides (CPP) containing arginine and lysine are thought to induce cellular uptake by macropinocytosis through the activation of cytoskeletal F-actin by such amino acids.41,45–48) Thus, morphological changes in F-actin and the localization of siRNA were observed. F-actin in RPE-J cell was fluorescently stained by Acti-stain™488 phalloidin, a heptapeptide derived from natural products that has a specific binding affinity for F-actin.46,48) For unmodified-, CH2S4- and SH2S4-lipoplexes, F-actin was not activated and was arranged in a regular pattern along the cytoskeleton (filamentous structure) as in untreated cells; siRNA localization was not observed (Fig. 3D). That is, such lipoplexes consisting of non-cationic amino acids may not affect the induction of macropinocytosis. In contrast, for CH2R4-, CH2K4- and SH2R4-lipoplexes, lamellipodia-like structures of F-actin polymers that appeared when molecules are internalized by macropinocytosis were observed, and siRNA was markedly distributed.47) Furthermore, for SH2R4-lipoplexes, strong siRNA fluorescence was observed in cells treated with CQ, which promotes endosome destruction. This indicates entrapment of SH2R4-lipoplexes in endosomes and decrease release from endosomes after uptake by cells compared to CH2R4-lipoplexes30) (Fig. 3E). That is presence of cysteine relates to the promotion of endosomal escape. The results indicate that peptides containing cationic amino acids (e.g., arginine, lysine) are effective for the activation of the cellular uptake of liposomes and siRNA, and that this is not only due to a surface potential-dependent interaction with the cell membrane but combination with cysteine.
We previously showed that VEGF mRNA expression was significantly suppressed in rat RPE-J cells through siVEGF-loaded CH2R4-lipoplex by modification of 2.0 mol% CH2R4 to cationic-LP under same experimental conditions this study.39) As described above, CH2R4-, CH2K4- and SH2R4-lipoplexes were superior in improving the internalization of lipoplexes and siRNA. However, SH2R4-lipoplex without cysteine was fewer release of siRNA from endosomes than CH2R4-, and CH2K4-lipoplexes. This result may affect the siRNA activity and suppressing effect of protein expression. In fact, the CH2R4-, and CH2K4-lipoplexes loading siVEGF showed tendencies to decrease in VEGF expression compared to respective siControl-loaded lipoplexes, particularly CH2R4-lipoplex was significant comparable to siVEGF-loaded LipoTrust as a positive control carrier, whereas VEGF expression for SH2R4-lipoplexes did not change with siVEGF and siControl treatments (Fig. 4). These results suggest that amino acid sequence, which induce both cellular uptake and siRNA release from endosome, are essential in composition of our peptide-modified nanocarrier for effective knockdown of target protein.
Retinal pigment epithelial (RPE) cells were treated with lipoplexes (siControl or siVEGF, 0.079 µM) in Opti-MEM for 4 h at 33 °C and 5% CO2. After 72 h culture in Opti-MEM at 33 °C, protein and total protein, and secreted VEGF protein in culture media were measured by DC Protein assay and VEGF enzyme-linked immunosorbent (ELISA) assay, respectively. VEGF protein expression levels (%) are defined as VEGF protein (pg)/Total protein (mg) × 100 (%). Data represent the mean ± S.D. (n = 3). p-Values are indicated: n.s. p > 0.05, * p < 0.05 by Student´s t-test.
Lipoplexes were developed with the aim of delivering siRNA to the retina through the instillation. However, the retina has strong tight junction barriers (i.e., BRB) formed by RPE that restrict the transportation of macromolecules and hydrophilic molecules via blood circulation and the instillation. To evaluate any improvement in the permeation of CH2R4-, CH2K4- and SH2R4-lipoplexes, 3D multicellular spheroids of rat RPE-J cells were used. A spheroid culture system is widely used as a biological evaluation model to replace animal experiments due to its mimicry of 3D structures having barrier functions because of the formation of tight junctions.51,52) We prepared RPE spheroids and confirmed the tight junction formation by observation of the presence of fluorescently immunostained ZO-1 (a tight junction formation factor). Figures 5A and B show the CLSM images and relative mean fluorescence intensity (MFI; MFI ratio of spheroid inner layer/surface) of lipoplexes that penetrated tight junctions in RPE spheroids after treatment with CH2R4-, CH2K4-, and SH2R4-lipoplexes for 1 or 4 h. The fluorescence of ATTO647-labeled lipoplexes were not observed at 1 h, whereas CH2R4-, CH2K4-, and SH2R4-lipoplexes showed significant permeation, which reached the insides of spheroids through tight junction layers at 4 h relative to control-lipoplexes (Fig. 5A). The CH2R4-lipoplexes showed excellent permeation (Fig. 5B), indicating that combinations of arginine, cysteine, and histidine were optimal components of functional peptide-modified liposomes for a retina target.
(A) Retinal pigment epithelial (RPE) spheroids were treated with lipoplexes (siRNA, 0.317 µM) in Opti-MEM for 1 or 4 h at 33 °C and 5% CO2. The fluorescence of ATTO-labeled lipoplexes and AlexaFluor555-stained zonula occludens-1 (ZO-1) in the centers of spheroids was observed using a confocal laser scanning microscope (CLSM). (B) The relative mean fluorescence intensity (MFI; MFI ratio of spheroid inner layer/surface) of lipoplexes that penetrated tight junctions (ZO-1). Data represent the mean ± S.D. (n = 3). * p < 0.05 by Student´s t-test. yellow: ZO-1; red: lipoplex. 3D: three-dimensional; DIC: differential interference contrast.
CH2R4, composed of arginine, histidine, cysteine and stearic acid, is a functional peptide that promotes cellular uptake, escapes from the endosome, and forms stable complexes with siRNA that dissociates in the cytoplasm.24) We previously indicated that the modification of CH2R4 by neutral and cationic liposomes can improve cellular uptake and intracellular dynamics.39) In this study, we investigated the effect of amino acid sequences of functional peptide-modified liposomes on cellular uptake pathways and siRNA transfection efficacy. Four types of peptides were designed based on CH2R4. The arginine and cysteine in the composition of CH2R4 was replaced with lysine or serine (neutral amino acid). CH2R4, CH2K4, CH2S4, SH2R4, and SH2S4 were used for modification with cationic liposomes. The lipoplexes consisting of siRNA and unmodified-cationic liposomes (about +30 mV, around 65 nm) showed a decrease in surface charge (about +10 mV) and an increase in particle size (around 120 nm). This may be because the positive charge of cationic liposomes was offset by the negative charge of the siRNA and the agglomeration thus formed. CH2R4, CH2K4, and SH2R4 were positively charged by basic amino acids (arginine, lysine). Lipoplexes consisting of such peptides had uniform particle sizes of 70–80 nm, high zeta-potentials of around +30 mV, and total siRNA loading efficiencies of 90% or higher. Additionally, siRNAs were retained on both peptide moieties and the surface of cationic liposomes in ratios of about 1 to 1 (Fig. 1). Such lipoplexes significantly enhanced the cellular uptake of siRNA, leading to the speculation that siRNA complexed with a peptide moiety functioned effectively (Fig. 2). In contrast, SH2S4 and CH2S4 had a neutral charge and the surface potentials of their lipoplexes depended on the zeta-potentials of cationic liposomes. Although total siRNA loading efficiencies were around 80%, the cellular uptake of siRNAs was almost the same as those of control-lipoplexes and lower than those of lipoplexes consisting of CH2R4, CH2K4, and SH2R4. This was the reason why SH2S4- and CH2S4-lipoplexes were thought to show a decrease in zeta-potentials and an increase in particle sizes, similar to control-lipoplexes. Additionally, siRNAs were retained by only cationic liposomes, and not by peptide moieties (Fig. 1). Arginine- and lysine-rich peptides are known as CPPs and enhance cellular uptake through endocytosis by interacting with the cell surface or substances expressed on cells.31,35,36) Thus, the cellular uptake pathways of the lipoplexes were evaluated.
Lipoplexes that included cationic peptides (CH2R4, CH2K4, SH2R4) induced macropinocytosis and promoted siRNA release from endosomes, suggesting that CH2R4, CH2K4, and SH2R4 on liposomes functioned as CPPs (Fig. 3). Basic amino acids promote endosomal escape through a proton sponge effect and enhanced interaction with endosomal membranes.31) Macropinocytosis is generally induced by the activation of F-actin through arginine and lysine to induce cytoskeletal deformation.34,37,47)
CH2R4, CH2K4, and SH2R4 may activate F-actin on the cell surface and induce intracellular uptake by macropinocytosis. Interestingly, regardless of if both CH2R4- and SH2R4-lipoplexes contained arginine, CH2R4-lipoplexes demonstrated superior functions compared to SH2R4-lipoplexes, highlighting the contribution of cysteine. A cysteine contained in CPPs was reported to promote macropinocytosis via intramolecular or intermolecular dimerization.35,36) CH2R4-lipoplexes can form intermolecular disulfide bonds. In fact, the cellular uptake of CH2R4-lipoplexes was more inhibited with EIPA treatment than that of SH2R4-lipoplexes (without cysteine), meaning that the cysteine involves macropinocytosis pathway (Fig. 3B). The lamellipodia-structures of F-actin are morphological changes that are observed during cell differentiation and uptake through macropinocytosis.18,29,53–55) Such morphological changes of F-actin were not observed in treatments with control-, SH2S4- and CH2S4-lipoplexes (without basic amino acids), while treatment with CH2R4-, CH2K4-, or SH2R4-lipoplexes led to the activation of F-actin (Fig. 3D). The substances taken up from outside the cell through clathrin-mediated endocytosis or macropinocytosis are generally trapped in endosomal vesicles. The siRNA is cleaved and metabolized in an acidic environment rich in the lysosomal enzymes of endosomes. Therefore, the endosomal escape of siRNA after intracellular uptake is important for the expression of siRNA activity. CH2R4-, CH2K4-, and SH2R4-lipoplexes promoted endosomal escape, which may be because basic amino acid-containing peptides induce the release of siRNA into the cytoplasm by a proton sponge effect. The electrostatic interactions between positively-charged basic amino acids and the negatively-charged endosomal membrane also may promote siRNA release from the endosome. In SH2R4-lipoplexes (without cysteine), a larger number of siRNAs independently localized in the cytoplasm in the presence of CQ than those in untreated cells with CQ, which enhances endosome disruption. This increase suggests that SH2R4-lipoplexes are normally trapped in endosomes and are not easier to release compared to CH2R4-lipoplexes even though siRNA uptake efficiency was high (Fig. 3E). Control-, SH2S4- and CH2S4-lipoplexes (without basic amino acids) did not show endosomal escape. This lack of release of siRNA may be because the lipoplexes were positively charged but surface charges were decreased by complexation with siRNA. Additionally, the siRNA did not interact with the peptide but only with cationic liposomes. CH2R4-, CH2K4- and SH2R4-lipoplexes improved endosomal escape. However, the suppression of VEGF expression was only significant for CH2R4-lipoplexes. Arginine has a guanidino group and can form two hydrogen bonds with cellular and endosomal membranes; it can exist stably even under protonation in acidic conditions.30,34,56) Through these functions, because CH2R4 can strongly interact with the cell surface and endosome membrane, it is thought that CH2R4-lipoplexes were superior in their enhancement of the intracellular uptake of siRNA and endosomal escape, which caused significant VEGF knockdown by siVEGF.57) In contrast, in CH2K4 (that the arginine of CH2R4 replaced with lysine), the lysine also functions as a CPP to promote siRNA transfection efficiency; however, it has less of an effect on cells than arginine because it forms only one hydrogen bond per molecule on the cell surface and endosomes.30) Therefore, although CH2K4-lipoplexes were capable of high siRNA transfection, it is thought that siRNA did not have a significant suppressing effect on VEGF expression. In CH2R4-lipoplexes, CH2R4 forms stable complexes through electrostatic interactions with siRNA and disulfide bonds between cysteines.31) Further, intermolecular dimerization may enhance the interaction of peptides with cellular and endosomal membranes, and enhance endosomal escape (Fig. 3E). These combined functions enable effective siRNA delivery into the cytoplasm by CH2R4-lipoplexes. SH2R4-lipoplexes showed high siRNA transfection, but had no suppressing effect on VEGF production due to the SH2R4 lacking cysteine and thus not being affected by disulfide bonds.
The RPE is located in the outermost layer of the retina and forms a BRB, which restricts drug delivery to the inside of the eyes through the blood circulation and by topical administration. The lipoplexes were designed to be used a non-invasive siRNA delivery system such as in eye drops.39) In general, for in vivo application, there are concern the effect of interactions between cationic lipoplexes and biological components (e.g., tear, cornea, and other ocular tissues). we have confirmed that the composition of 2 mol% peptide-modified lipoplexes in this study showed no cytotoxicity and histological changes (data not shown). We evaluated the permeation of lipoplexes in a 3D RPE spheroid model that mimics the BRB and shows the formation of tight junctions (Fig. 5). CH2R4-, CH2K4-, and SH2R4-lipoplexes showed superior permeation with a longer time and localized inside of ZO-1 layers in RPE spheroids, while SH2S4- and CH2S4-lipoplexes were not observed.58,59) This indicates that a basic amino acid–containing peptide has the potential to promote the permeation of lipoplexes through tight junctions.
In conclusion, this study demonstrated the effective roles of amino acid sequences of promising peptide-modified liposomes for the improvement of siRNA efficacy in rat RPE-J cells. The uptake pathway, siRNA-loading efficiency, the suppression of VEGF protein production, and permeability differed depending on the amino acid sequence of peptides. We elucidated that peptide with sequences containing basic amino acids that enhances cell permeation and that cysteine that enhances siRNA retention and endosomal escape functioned effectively even when such peptides were modified in liposomes in rat RPE-J cells. Peptides with such functions are useful for the construction of siRNA nanocarriers that target the retina where BRB generally restricts drug delivery.
This work was supported by JSPS KAKENHI Grant Number: 20K12652.
S.N. and Y.T. contributed equally as first authors. The authors contributed to this work as follows: S.N.: methodology, validation, formal analysis, investigation, data curation, writing the original draft, and visualization; Y.T.: conceptualization, methodology, formal analysis, investigation, data curation, manuscript review and editing, visualization, supervision, funding acquisition, and project administration; K.E.: data curation, validation, and investigation; H.I.: supervision. All authors contributed to the writing of the manuscript. All authors approved the final version of the manuscript.
The authors declare no conflict of interest.
