Plasmids, antibodies, and reagents

The JFH1 plasmid [25] was used to produce infectious HCV particles. The pRluc-JFH1 plasmid [26] harboring Renilla luciferase (Rluc)-coding gene in the context of genotype 2a full-length HCV JFH1 cDNA clone was received from Dr. Xulin Chen (Institute for Virus Research, Chinese Academic of Sciences, Wuhan, China). The Rluc-J6/JFH1 (FL-J6/JFH-5′C19Rluc2AUbi) plasmid [27], which was derived from the previously described infectious genotype 2a chimeric HCV genome J6/JFH1 [28], was used for the synthesis of a full-length HCV genome that expresses Rluc. The bicistronic vector pDual-IRES, which expresses a cap-dependent Rluc reporter and an HCV IRES-dependent firefly luciferase (Fluc) reporter was described previously [29]. The pcDNA3.1-Ago2 vector expressing the human Ago2 was constructed by RT-PCR-mediated amplification of Ago2 cDNA using the forward primer Ago2-EcoRI-F (5′-gactGAATTCgATGTACTCGGGAGCCGGCCCCGCACT-3′) and the reverse primer Ago2-NotI-R (5′-gactGCGGCCGCTCAAGCAAAGTACATGGTG-3′), followed by cloning of the PCR-amplified product into the EcoRI and NotI-cleaved pcDNA3.1 vector (Invitrogen, Carlsbad, CA, USA). Recombinant human Ago2 was obtained from Sino Biological Inc. (Beijing, China). The antibody against HCV NS5A was obtained from Biodesign International (clone 381; Kennebunk, ME, USA). The mouse anti-α-tubulin antibody was from Calbiochem (La Jolla, CA, USA). The rabbit polyclonal anti-NS5B antibody was generated previously in our laboratory [30]. The goat anti-HCV E2 polyclonal antibody was obtained from Acris Antibodies (GmbH, Hideenhausen, Germany). Polyinosinic:polycytidylic acid [poly(I:C)], a double-stranded RNA (dsRNA) analog, was obtained from Sigma-Aldrich (St. Louis, MO, USA).

Cell culture

The Huh7 human hepatocellular carcinoma and human embryonic kidney 293 (HEK293) cell lines were grown in the Dulbecco’s modified Eagle’s medium (DMEM) as reported previously [31]. The R-1 cell line, in which a selectable HCV subgenomic replicon replicates [32], was also maintained in DMEM supplemented with 1 mg/ml of G418. Human peripheral mononuclear cells (hPBMCs) were isolated by centrifugation over Ficoll-Hypaque (GE Healthcare, Uppsala, Sweden) and grown as previously described [31].

siRNA and transfection

The siRNAs targeting HCV IRES were obtained from ST Pharm (Seoul, Korea). The sequences of siRNAs used in this study are shown in Table 1. Several chemically modified siIE22 derivatives were synthesized and their sequences are shown in the corresponding figure in the results section. In one set of siIE22 derivatives, the 2′-OH group in the ribose sugar backbone was substituted with the 2′-O-methyl (2′O-Me) or 2′-fluoro (2′F) groups, while another derivative carried a replacement of one of the non-bridging oxygen atom from the backbone phosphate between two ribonucleotides with a sulfur atom to create a phosphorothioate (PS) linkage. siRNAs were transfected into cells using Lipofectamine RNAiMAX (Invitrogen) or the lipidoid ND98 as previously reported [33].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. siRNAs used in this study.

https://doi.org/10.1371/journal.pone.0146710.t001

LNP formulation and nanoparticle size analysis

For systemic siRNA delivery, siRNA LNPs were prepared using ND98, cholesterol (Sigma-Aldrich), and polyethylene glycol (PEG) 2000-ceramide C16 (Avanti Polar Lipids, Alabaster, AL, USA), as previously described [33]. The concentration of siRNA encapsulated within LNPs was determined using the Quant-iT RiboGreen RNA assay [33]. LNP particle size was analyzed using a NanoSight LM10 instrument (NanoSight, Amesbury, UK) according to the manufacturer’s instructions. Briefly, a monochromatic laser beam at 405 nm was applied to PBS-diluted (10,000-fold dilution) siIE22 LNPs. Brownian motion of LNPs was tracked and measured from frame to frame using nanoparticle tracking analysis software. The velocity of moving particles was used to calculate particle size from the 2D Stokes–Einstein equation.

Target RNA cleavage assays

An in vitro RNA cleavage assay was performed using a human recombinant Ago2 as described previously [34]. Alternatively, the human Ago2 immunoprecipitated from pcDNA3.1-Ago2-transfected HeLa cells was also used as previously reported [35]. Briefly, Ago2, which forms an active RNA-induced silencing complex (RISC) complex in cells, was immunoprecipitated using a rabbit anti-Ago2 antibody (clone C34C6; Cell Signaling Technology, Danvers, MA, USA) and protein G Dynabeads (Dynal, Oslo, Norway). Target RNAs used for the Ago2 cleavage assays included HCV genome 5′-end region (488-nts including an extra G residue added to the 5′-end) and an HCV IRES region spanning nts 313–343 that were 5′-end ³²P-labeled using [γ-³²P] ATP and T4 polynucleotide kinase. RNA samples were resolved on a denaturing polyacrylamide gel for autoradiography.

HCV infection

Full-length HCV (genotype 2a, JFH1 clone) RNA was prepared by in vitro transcription and electroporated into Huh7 cells according to the protocol reported previously [36]. Infectious HCV particles collected from the culture medium were used for infection of Huh7 cells as previously described [31].

Transient HCV replication assay

Huh7 cells were co-transfected by electroporation with in vitro transcripts of pRluc-JFH1 and pGL3 plasmids (Promega, Madison, WI, USA). After 24 h, siRNAs were transfected using Lipofectamine RNAiMAX (Invitrogen). Rluc and Fluc activities in cell lysates were monitored 48 h after transfection using the Dual-Glo luciferase assay system (Promega). Rluc activity was normalized against Fluc activity to calculate relative luciferase activity.

TaqMan real-time quantitative RT-PCR

The HCV genome copy number was estimated by real-time quantitative RT-PCR (qRT-PCR) using a TaqMan probe specific for the HCV 5′-UTR as previously described [29].

Western blot and northern blot analyses

Western blot analysis of HCV viral proteins was performed as described previously [37]. For HCV genome detection by northern blotting, total RNA (5 μg) run on a denaturing formaldehyde agarose gel was blotted and hybridized with a ³²P-labeled DNA probe [HCV NS3 NS5B (nts 3446–9265)] prepared by using the Ready-To-Go DNA labeling kit (GE Healthcare Life Sciences, Piscataway, NJ, USA).

Interferon β promoter activity assay

HEK293 cells grown in 24-well plates were transfected with 400 ng of IFNβ-pGL3, which expresses Fluc under the control of the IFN-β promoter [38] and 100 ng of the pRL-TK reporter (Promega), which expresses Rluc as an internal control. Six hours after transfection, cells were washed, incubated with fresh medium, and transfected with siRNA (100 nM) or poly(I:C) (1 μg/ml) using Lipofectamine RNAiMAX (Invitrogen). Stimulated cells were harvested 8 h later, and luciferase assays were performed using the Dual-Glo luciferase assay system (Promega) as described above.

Cell viability

The cytotoxicity of siIE22 was measured using MTS [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide] reagent as previously described [31].

siRNA stability test

Serum stability of siRNAs was assessed by northern blot analysis. siRNA (2 μM) was incubated in 45% human plasma at 37°C for indicated times prior to northern blot analysis by a conventional method. Briefly, total RNA extracted with Trizol LS reagent (Invitrogen) was separated by 15% polyacrylamide/7 M urea gel electrophoresis and transferred to a membrane by electroblotting. As a probe, 5′-end ³²P-labeled, synthetic ribo-oligonucleotide complementary to siRNA guide-strand was used. Densitometric quantification of the radioactive hybridization signal was performed using a Fuji BAS-2500 Phosphorimager.

Enzyme-linked immunosorbent assay

The experiment with hPBMCs was approved by the Institutional Review Board of the Yonsei University (Approval No: 1040917-201312-BR-133-01E). hPBMCs were isolated from peripheral blood, which was obtained from healthy volunteers with their written informed consent and provided by the Korean Red Cross, Western Blood Center Blood Services (Approval No: 2823–201312). hPBMCs (1 × 10⁶ cells/well) grown in a 96-well round-bottom plate were transfected with siRNA (10 nM) or poly(I:C) (1 μg/well) using ND98 or LNP. As a positive control for these assays, 50 μg/ml of poly(I:C) was added directly to the culture medium. After stimulation for 16 h, stimulated cultured cells were centrifuged, supernatants were collected by centrifugation, and the amount of IFN-α was determined using a human IFN-α⋅enzyme-linked immunosorbent assay (ELISA) kit (eBioscience, San Diego, CA, USA). The detection limit of the human IFN-α ELISA was 15 pg/ml. Serum IFN-α⋅titer in BALB/c mice (n = 4 per group) that received siRNA or poly(I:C) (1 mg RNA/kg body weight) either complexed with ND98 or formulated in LNPs was determined by ELISA (detection limit: 12.5 pg/ml) using a kit supplied by PBL Interferon Source (Piscataway, NJ, USA).

Immunohistochemistry

Xenograft tissues fixed in 4% paraformaldehyde in PBS were embedded in a frozen section compound (Surgipath FSC22, Leica Microsystems, Wetzlar, Germany). Tissue sections were incubated for 2 h at room temperature with an affinity-purified polyclonal anti-NS5B antibody or a goat anti-HCV E2 polyclonal antibody with gentle rocking. After washing 3 times with PBS, appropriate fluorescent-conjugated secondary antibodies were used for antigen visualization. Nuclei were visualized by staining with 1 μM 4′,6′-diaamidino-2-phenylindole (DAPI) in PBS for 10 min. Confocal images were collected on an LSM 510 META confocal laser scanning microscope (Carl Zeiss, Oberkochen, Germany).

Animal experiments

All animal experiments were performed in accordance with the Korean Food and Drug Administration guidelines. Protocols were reviewed and approved by the Institutional Animal Care and Use Committee of the Yonsei Laboratory Animal Research Center (Permit No: 2012–0076). At the termination of experiments, all mice were euthanized by CO₂ inhalation.

In vivo RNAi assay

The dual luciferase expression vector pDual-IRES (20 μg in 1.6 ml PBS), which was purified with the Endotoxin Free Maxi Kit (Qiagen, Hilden, Germany), was administered to 6–8 weeks old female BALB/c mice by hydrodynamic injection through the tail vein within 5 to 8 s. After 1 h, mice were intravenously (iv) injected with siRNA LNPs at a dose of 1 mg siRNA per kg body weight. After 16 h, livers were homogenized and mixed with an equal volume of the Glo lysis 1× buffer (Promega). Then, the mixture was centrifuged for 5 min at 13,000 g at 4°C to obtain cleared lysate that was analyzed for reporter activity using the Dual-Glo luciferase assay system (Promega). Luciferase expression was reported as relative light units (RLUs) per mg protein.

Evaluation of in vivo siRNA anti-HCV activity

To evaluate in vivo anti-HCV efficacy of siRNA, NOD-SCID mice bearing HCV-replicating Huh7 xenograft were used. HCV RNA-transfected cells mixed with Matrigel (BD Biosciences Discovery Labware, Bedford, MA, USA) were injected subcutaneously (sc) into both flanks of anesthetized immunodeficient NOD-SCID male mice (5 weeks of age, 20–23 g body weight). siIE22 LNPs (1 mg siRNA per kg body weight) were administered by tail vein injection at 3 to 4 weeks after implantation.

Statistical analysis

Statistical analysis was performed using GraphPad Prism (Graphpad Prism Software Inc., La Jolla, CA, USA). Results are presented as the mean ±standard deviation (SD) from at least three independent experiments, unless otherwise stated. Differences between groups were considered statistically significant if P < 0.05.

Identification of potent HCV IRES-targeting siRNAs by siRNA-tiling experiments

To identify siRNA-targetable regions within highly structured HCV IRES, which is also associated with multiple cellular proteins, we utilized siRNA tiling approach as follows. A total of 25 siRNAs were initially tiled with overlapping sequences that were shifted by two nucleotides over the IRES SL region spanning nts 277–343 (Fig 1A, siRNA tiled region is shown in thick gray line; see also siRNA sequences in Table 1 and Fig 1B). We assessed anti-HCV activity of synthetic siRNAs in Huh7 cells transiently transfected with in vitro transcripts of the full-length genotype 2a HCV replicon Rluc-JFH1 encoding an Rluc reporter and identified six highly potent anti-HCV siRNAs (siIE19, siIE21, siIE22, siIE23, siIE24, and siIE25) targeting the SL IIIf and IV regions spanning nts 313–343 (Fig 1C). Other siRNAs targeting sites upstream of these IRES regions showed less potent antiviral activity than the selected siRNAs. RNAi effect of non-selected siRNAs was marginal and sometimes application of these siRNAs had an unexpected boosting effect on virus replication. Among the selected antiviral siRNAs, siIE22 targeting nts 319–337 was identified as the most effective siRNA, reducing the reporter activity by up to 98% at a 50 nM concentration.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Screening for potent HCV IRES-targeting siRNA by siRNA tiling experiments.

(A) The proposed secondary structure of HCV IRES. The IRES region spanning nts 277–343 (shown in gray) was targeted by siRNAs. The target site of the selected potent anti-HCV siRNA siIE22 is shown in blue. The base-pairings in the proposed pseudoknot (PK) structures (PKs 1 and 2) are shown in green. (B) siRNA sequences tiled over HCV IRES. The underlined sequence represents a mapped druggable region (nts 313–343) where the targets of selected potent siRNAs were enriched. (C) Anti-HCV activity of HCV IRES-targeting siRNAs in Huh7 cells transfected with Rluc-JFH1 (top panel), an HCV replicon encoding the Rluc reporter. The Rluc gene was fused in frame to the DNA sequence encoding 17 N-terminal amino acid residues of the HCV core protein. Huh7 cells were electroporated with the Rluc-JFH1 in vitro RNA transcript and pGL3 plasmid used for normalization of transfection efficiency. After 24 h, the cells were transfected with each of the IRES-specific siRNAs or a scrambled (Sc) siRNA (50 nM). At 48 h post-transfection, luciferase activity was measured. (D) Huh7 cells harboring an HCV subgenomic replicon RNA (R-1) were transfected with each of the IRES-specific siRNAs or Sc siRNA (10 nM each). At 48 h post-transfection, HCV RNA levels were quantified by real-time qRT-PCR. *, P < 0.01. (E and F) Dose-dependent inhibition of HCV-replication by siIE22 was assessed in R-1 cells, as described in (D). HCV genome copy number and HCV proteins (NS5A and NS5B) levels were analyzed by qRT-PCR (E) and western blotting (F), respectively.

https://doi.org/10.1371/journal.pone.0146710.g001

Our results described above underscore the importance of siRNA targeting site selection within the highly structured IRES in executing RNAi. To further support this conclusion, we designed two additional siRNAs—siIE21.5 and siIE22.5—targeting the regions around the siIE22 binding site (1-nt shifting toward the 5′ and 3′ direction, respectively; see Fig 1B). For this purpose, we also included two other previously reported potent anti-HCV siRNAs—si313 [39] and siE [40]—for antiviral efficacy assays because the si313 target (nts 315–333) partially overlaps with binding sites of siIE19 (nts 313–331) and siIE21 (nts 317–335). siE, which is 1 nt longer than siIE24 (targeting nts 323–341), targets nts 323–342. We found that siIE22 displayed greater antiviral potency in R-1 cells harboring a genotype 1b HCV subgenomic replicon (Fig 1D), although the siIE22 target sequence overlaps with the target sites of other tested siRNAs described above. Importantly, we observed that siIE21.5 and siIE21 showed lower antiviral activity than siIE22, as their targets were shifted by 1 nt and 2 nts toward the 5′-direction, respectively. Similarly, sequential nucleotide shifting of the siIE22 target site toward the 3′-direction also attenuated the antiviral potency of siIE22 (compare siIE22 with siIE22.5, siIE23, siIE24, siE, and siIE25 in Fig 1C and 1D). These results clearly map a short 31-nt stretch ranging from nt 313 to nt 343 as the druggable target site for potent anti-HCV siRNAs. siIE22 targeted a previously unidentified 19-nts sequence array from nt 319 to nt 337 within the IRES subdomain IIIf (Fig 1A) involved in the formation of the pseudoknot 1 [24]. It showed the greatest anti-HCV potency with a half-maximum effective concentration (EC₅₀) of 0.4 nM in R-1 cells (Fig 1E). Expression of HCV NS5A and NS5B proteins, as assessed by western blot analysis, was decreased dose-dependently by the treatment with siIE22 (Fig 1F). Results of siRNA tiling experiments showed that even a single nucleotide target shift toward either 5′ or 3′ direction could lower the antiviral potency of siIE22 suggesting a specific siRNA positioning effect.

We reasoned that this superior antiviral activity of siIE22 could be mediated by a greater target cleavage activity. To test this possibility, we performed IRES-target cleavage assays using immoprecipitated Ago2 (Fig 2A), which forms an active RISC in cells treated with a duplex form of siRNA [35], and using a recombinant Ago2 protein (Fig 2B), which uses single-stranded antisense siRNA guide-strand for RNAi [34]. We observed that two siRNAs siIE21.5 and siIE22.5, whose targets were shifted by a single nucleotide toward the 5′ and 3′ directions, respectively, relative to the siIE22 target site, as well as siRNAs siIE21 and siIE23, whose targets were similarly shifted by two nucleotides, were capable of digesting the probe equally well in vitro. Subsequent analysis of the stability of siIE21.5 and siIE22.5 in human plasma showed no noticeable differences in siRNA stability between these siRNAs and siIE22 over the period of 60 min incubation at 37°C (Fig 2C). Considering the fact that HCV IRES is bound by various ITAFs as well as by ribosomal complexes [8, 21], these results collectively suggest that the strong antiviral activity of siIE22 might be due to its optimal target accessibility in cellular context. Further supporting this conclusion, although the two siRNAs targeting the pseudoknot 2 [24] (Fig 1A), siIE3 and siIE6, showed only negligible antiviral activity (Fig 1C), both of them were capable of cleaving the target probe as effectively as siIE22 (Fig 2D).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Target RNA cleavage activity and serum stability of the selected IRES-targeting siRNAs.

(A) In vitro cleavage assays were performed with 5′ ³²P-radiolabeled 31-nts long HCV IRES (spanning nts 313–343) using Ago2 complexes immunoprecipitated from HeLa cells transfected with each indicated siRNA (20 nM). Sc, scrambled siRNA. (B) RNA target cleavage assay was performed with 5′ ³²P-radiolabeled HCV IRES substrate (spanning nts 1–487) using a recombinant human Ago2 protein and guide strand of indicated siRNAs. As a negative control, siIE22 passenger-strand was used. RNA samples were analyzed by electrophoresis on a denaturing polyacrylamide denaturing gel followed by autoradiography. (C) Serum stability of siRNAs was assessed by incubating siRNAs in 45% plasma for the indicated time periods and subsequent quantification of remained intact siRNA guide-strand by Phosphorimager analysis of northern blots. (D) The cleavage assay was performed as described in (B). Radioactivity of cleaved products was quantified by Phosphorimager analysis and shown below the autoradiogram (mean ± SD of three independent experiments).

https://doi.org/10.1371/journal.pone.0146710.g002

Both the sequence and the length of siRNA are important in designing potent siRNAs targeting the IRES subdomain IIIf within the pseudoknot 1

We designed three additional synthetic siRNAs—siIE318_27 (27-mer; targeting nts 318–342), siIE320_25 (25-mer; targeting nts 320–342), and siIE322_23 (23-mer; targeting nts 322–342) (Fig 3A; also see their sequences in Table 1)—to test if these siRNAs, which contained the seed sequence of siIE22 (5′ 2–8 nts on the guide strand) but differed in the total length, display antiviral potency similar to that of siIE22. Of these three siRNAs, siIE318_27 showed the greatest antiviral activity with 85% inhibition at 100 pM in a cell line bearing a full-length HCV replicon encoding Rluc reporter (Fig 3B). However, siIE318_27 siRNA showed a lesser effect against the HCV replicon in R-1 cells when compared with gs_PS1 siIE22 (Fig 3C). Moreover, 27-mer siIE318_27 could not digest the IRES target as effectively as 21-mer siIE22 and its stabilized derivative gs_PS1 siIE22 (Fig 3D; compare lanes 3 and 4 with lane 5). The target cleavage assay also revealed that siIE320_25 and siIE322_23 displayed higher cleavage effects than siIE318_27 (Fig 3D, lanes 5–7). These cleavage assay results are consistent with an earlier report showing that siRNAs composed of more than 23 nts are less efficient in RNAi than 21-nt long siRNAs [41]. In this regard, we did not expect to observe that 27-mer siIE318_27 siRNA had a greater antiviral potency than the two smaller size siRNAs, siIE320_25 and siIE322_23. Considering that these siRNAs, as well as siE, all have the same 3′ -end boundary (ending at the nt 342) in their targets with variable 5′ -end boundary sequences (starting at the nt 318, 320, 322, and 323 for siIE318_27, siIE320_25 and siIE322_23, and siE, respectively), it appears that defining a specific binding site within the 31-nt druggable region (nts 313–343) mapped in the present study is critical for the identification of potent anti-HCV siRNAs.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Analysis of anti-HCV potency and RNAi activity of various siRNAs targeting the IRES subdomain IIIf.

(A) Passenger strand sequences of siIE22 and six siRNAs sharing their targets with siIE22 in the IRES subdomain IIIf. (B) Evaluation of anti-HCV efficacy of a set of selected siRNAs (100 pM each) using an HCV replicon expressing an Rluc reporter. Luciferase activity at 48 h post-siRNA treatment is shown. Sc, scrambled siRNA. (C) Antiviral efficacy of gs_PS1 siIE22 and siIE318_27 in R-1 cells harboring an HCV subgenomic replicon. (D) In vitro target cleavage assays were performed as described in Fig 2A. P, 5′ ³²P-radiolabeled 31-nt long HCV IRES probe; CP, cleaved probe. (E) Comparison of antiviral activity of indicated siRNAs targeting the region shared with siIE22 in R-1 cells. In (B), (C), and (E), *, P < 0.05; **, P < 0.01.

https://doi.org/10.1371/journal.pone.0146710.g003

Further, we tested another three 21-mer siRNAs (Table 1; they are all 21-mer siRNAs with a 3′-end two nucleotide overhang, including siIE316, siIE322, and siIE319) that share parts of their target site with siIE22 in antiviral activity comparison experiments. Of note, siIE319 shares 19 nts in its target sequence with siIE22; however, it differs from siIE22 in that the overhang sequences on both guide and passenger strands of siIE319 are derived from the HCV sequence and thus do not have TT or UU residues, the two conventional siRNA overhang sequences. Surprisingly, this difference resulted in a significant 1.78-fold reduction (P < 0.01) in the anti-HCV potency of siIE319 (Fig 3E), demonstrating the important role of overhang sequences in the formation of an active RISC complex with the guide strand chosen based on thermodynamic stability of siRNA termini. During the RNAi catalytic cycle, there is consecutive guide-strand 3′-end binding to and release from the PAZ domain of Ago2 [42]. It is thus likely that unfavorable interactions between 3′-end of the guide strand (GC in siIE319 vs. UU in siIE22) with the PAZ domain of Ago2 might interfere with active RISC formation and result in weaker RNAi. This hypothesis is supported by recent findings of Kandeel and Kitade [43], who demonstrated that weaker binding of siRNAs by the PAZ domain correlates with higher RNAi.

It should be noted that the siIE316 target is shifted by 1 nt toward the 5′-direction relative to siIE21 and that the siIE322 target is shifted by 1 nt toward the 3′-direction relative to siIE23. None of these siRNAs showed a stronger inhibitory effect than siIE22. A shift by three nts of the siIE22 target toward the 5′-direction to reach the 5′ boundary of the siIE316 target reduced antiviral potency. Together, these results further highlight the importance of defining 5′- and 3′-ends target boundaries within the mapped siRNA accessible region in the HCV IRES subdomain IIIf for the identification of potent therapeutic siRNAs.

In vivo inhibition of HCV IRES-mediated translation by siIE22 LNP

Inefficient siRNA delivery to target tissues and cells has been a major obstacle to the clinical development of siRNA therapeutics. We encapsulated siIE22 using a lipidoid nanoparticle formulation composed of the lipidoid ND98, cholesterol, and a PEG-lipid to facilitate the delivery of siRNA to the liver (Fig 4A). The resulting LNPs containing siIE22 (siIE22 LNPs) had a mean diameter of ~100 nm (Fig 4B), as assessed by the nanoparticle tracking analysis. In mice that were hydrodynamically injected with the pDual-IRES plasmid, the siIE22 LNP intravenously injected at a dose of 1 mg siRNA/kg body weight resulted in a significant reduction in the normalized luciferase activity by up to ~2 log₁₀ (P < 0.01), confirming the capacity of LNPs to provide systemic delivery of siRNA (Fig 4C).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Inhibition of HCV IRES-mediated translation by systemically delivered LNP-formulated siIE22.

(A and B) Schematic diagram of siIE22 LNP (A) and LNP particle size analysis (B). (C) Experimental schedule and schematic representation of the pDual-IRES plasmid. The pDual-IRES plasmid was hydrodynamically injected through the tail vein of BALB/c mice (n = 4 per group). After 1 h, mice were iv injected with siIE22 LNP at a dose of 1 mg/kg body weight. The Fluc expression level in the liver was determined 16 h after the injection. Luciferase activity is reported as RLU per mg protein. *, P < 0.01. (D) BALB/c mice (n = 4 per group) were iv injected with indicated siRNAs (1 mg/kg body weight) complexed with ND98 or formulated with LNP. Poly(I:C) (1 mg/kg) complexed with ND98, formulated with LNP, or free form (each in 170 μl) was administered. PBS or LNP vesicles alone were used as control treatments. Two hours later, serum IFN-α levels were quantified by ELISA. The dotted line indicates the detection limit of the assay (15 pg/ml). (E) hPMBCs grown in 96-well plates were transfected with indicated siRNA at 10 nM concentration or with 1 μg/well poly(I:C) using the lipidoid ND98 or stimulated by a direct addition of 50 μg/ml poly(I:C) to the medium. After 16 h, cell culture supernatants from stimulated cells were analyzed for IFN-α by ELISA. Data shown are from one of the two independent experiments with similar results. ND, non-detectable. (F) HEK293 cells were transfected with the luciferase expressing plasmids (IFNβ-pGL3 and pRL-TK) for the IFN-β promoter activity assay. After 6 h, cells were transfected with 100 nM siIE22 or scrambled (Sc) siRNA, or 1 μg/ml poly(I:C). After 8 h, cells were harvested for dual luciferase assays. Fluc activity was normalized to Rluc activity from the pRL-TK plasmid. Normalized luciferase activity (Fluc/Rluc) of mock-treated cells was defined as 100. Data are presented as the mean ± SD of six measurements from two independent experiments.

https://doi.org/10.1371/journal.pone.0146710.g004

The cytosol pattern recognition receptor (PRR), retinoic acid-inducible gene 1 (RIG-I), recognizes cytoplasmic foreign RNAs including a short duplex siRNA with certain types of sequences, and induces IFN-β production [44]. We observed that siIE22 LNPs, when injected through the tail vein, did not increase IFN-α production in mice, whereas synthetic dsRNA poly(I:C), a potent activator of type-I IFN responses [44], increased serum IFN-α levels when it was delivered using ND98 or LNPs (Fig 4D). We found that poly(I:C) induced weaker IFN- α production when it was LNP-formulated compared to the IFN-α induction levels observed when poly(I:C) was administered with ND98. This finding suggests that LNP formulation may reduce the innate immune response-activating activity of siRNAs by influencing their interaction with PRRs. Furthermore, we observed little induction of IFN-α production by siIE22 LNPs in hPBMCs, while poly(I:C), known to be recognized by TLR3, induced IFN-α production when it was added directly to the culture medium, as expected (Fig 4E). Consistent with the results obtained in mice, poly(I:C), which is also recognized by the cytosolic PRRs RIG-I and MDA5 [44–46], however, did not activate IFN-α production in hPBMCs when it was delivered within LNPs. Furthermore, no intrinsic induction of the type I IFN response by siIE22 per se was observed in HEK293 cells by using a luciferase reporter plasmid carrying the binding site of IRF3, a transcription factor activated by the RIG-I or TLR3 pathway [38]. As shown in Fig 4F, siIE22 did not trigger promoter activation, whereas poly(I:C) that was delivered intracellularly by RNAiMax-mediated transfection increased the promoter activity by about 3-fold, as expected. In summary, these results clearly demonstrate that the RNAi potency of siIE22 is not mediated by IFN responses.

Anti-HCV efficacy of chemically modified siIE22 derivatives

Stabilization of synthetic siRNAs is often regarded as a prerequisite for successful in vivo applications. Northern blot analysis showed that the half-life of unmodified siIE22 in human plasma was ~30 min (Fig 5A). We thus chemically modified siIE22 to improve its stability against nucleases. Of several siIE22 derivatives formed with a set of passenger and guide strands of siIE22 with 2′O-Me or 2′F nucleotide modifications, or PS linkages (Fig 5B), we could select four siRNAs that showed improved stability following minimal structural modifications. Those siRNA include gs_All-F and gs_Py-F, as well as gs_PS1 and gs_PS2 siRNAs carrying one and two PS linkages on the guide strand, respectively (Fig 5C). Modified siRNAs with a PS linkage on the passenger strand were excluded because their influence on the stability was significantly lower than the effect of PS modification(s) on the guide strand. When PS-modified passenger strands were duplexed with the similarly modified corresponding guide strands, the stability of the resulting duplexes was not improved (see ps_PS1/gs_PS1 and ps_PS2/gs_PS2 in Fig 5C).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Anti-HCV efficacy of chemically modified siIE22 derivatives.

(A) Non-modified siIE22 (2 μM) was incubated in 45% human plasma for the indicated time periods. RNA extracted from each sample was resolved by electrophoresis on a denaturing 15% polyacrylamide gel and subjected to northern blotting (NB) analysis for detection of siIE22 guide-strand. The Phosphorimager image shown is from one representative experiment of three independent experiments with similar results. Densitometric analysis of siIE22 guide strand signal was done using a Phosphorimager. Relative intensity of signals was plotted using SigmaPlot to estimate siIE22 guide strand half-life. Relative signal (% of signal at time 0) is shown below a representative blot. (B) Sequences of modified guide and passenger strands of siIE22 derivatives used in this study. Modified residues are shown in green or blue. “s”, phosphorothioate linkage. (C) Plasma stability of a set of the selected siIE22 derivatives was evaluated as in (A). (D) Anti-HCV activity of the selected modified siIE22 (1 nM) was evaluated in R-1 cells as in Fig 1D. *, P < 0.01. (E) Analysis of half-life of the gs_PS1 siIE22 as in (A).

https://doi.org/10.1371/journal.pone.0146710.g005

The results of antiviral efficacy test for the four selected siRNAs revealed that although gs_All-F and gs_Py-F siRNAs were more stable, they displayed reduced anti-HCV activity (Fig 5D). Based on the antiviral activity and siRNA stability, we selected gs_PS1 siIE22, which had an increased half-life of 205 min in plasma (Fig 5E) and retained anti-HCV activity of the original siRNA upon a chemical modification, for subsequent in vivo experiments.

Anti-HCV efficacy of gs_PS1 siIE22 LNPs in a mouse model of HCV replication

In vivo anti-HCV activity of siIE22 LNPs was assessed in a mouse model of HCV replication in which Huh7 cells transfected with HCV (JFH1) RNA by electroporation were implanted subcutaneously into flanks of NOD-SCID mice. In xenograft tissues retrieved from the mice at 4 weeks post-xenografting, HCV genome (Fig 6A) and viral proteins (NS5B and E2) (Fig 6B) were detected by northern blotting and immunohistochemical analysis, respectively, indicating a productive replication of HCV in the xenografts. The mice were given LNP-formulated gs_PS1 siIE22 by a single iv injection of 1 mg siRNA/kg body weight. Systemic delivery of siRNA LNPs led to more than 80% reduction in the serum HCV titer on days 3 and 7 demonstrating that sustained antiviral activity lasted for at least 7 days (Fig 6C). Administration of gs_PS1 siIE22 LNPs at a dose of 1 mg/kg did not increase alanine aminotransferase activity (data not shown), which indicated an absence of acute liver toxicity. Furthermore, when siRNA LNPs were administered once every 3 days at a dose of 1 mg/kg for a total of 4 injections, serum HCV RNA titer was reduced by more than 4 log₁₀ (Fig 6D), while LNP-formulated Sc siRNA had little effect. Sequence analysis of 16 independent clones for the region targeted by siIE22 revealed that no mutations that might confer resistance to the siRNA occurred during the course of siRNA treatment (S1 Fig). This observation suggest that the target sequence has a high genetic barrier to the development of resistance.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Anti-HCV efficacy of gs_PS1 siIE22 LNP in a mouse model for HCV replication.

(A) HCV genome in the HCV-replicating Huh7 (+ HCV) or Huh7 (- HCV) sc xenograft was detected by northern blotting. An in vitro transcribed HCV RNA genome was used as a size marker. The 28S rRNA detected by ethidium bromide staining is shown as a loading control. (B) Immunostaining for HCV viral proteins (E2 and NS5B) in the xenograft at 4 weeks post-xenografting. DAPI, nuclear staining. Scale bar, 10 μm. (C) The NOD-SCID mice (n = 3) carrying HCV-replicating Huh7 xenograft were treated with gs_PS1 siIE22 LNP at a dose of 1 mg/kg body weight via tail vein injection. Shown are relative serum HCV RNA titers at the indicated time points. *, P < 0.01. (D) Relative serum HCV RNA titers in the mice treated with LNP-formulated Sc siRNA (Sc LNP) or gs_PS1 siIE22 LNP (1 mg/kg) once in every three days for 4 times. The data were generated from two independent experiments with a total of 6 mice per group. Each differently colored diamond represents an individual mouse. *, P < 0.01.

https://doi.org/10.1371/journal.pone.0146710.g006