Potential conflict of interest: Nothing to report.
RNA interference is highly effective at inhibiting HBV gene expression and replication. However, before small interfering RNA (siRNA) can be used in the clinic, it is essential to develop a system to target their delivery. Antibody-mediated delivery is a novel approach for targeting siRNA to appropriate cells. In this report, we asked whether this siRNA delivery strategy would be effective against HBV. Of 5 candidates, a specific siRNA that effectively inhibited HBV gene expression and replication was determined. Two fusion proteins, s-tP and sCκ-tP, were constructed to contain a single chain of the human variable fragment, scFv, against hepatitis B surface antigen (HBsAg), a truncated protamine (tP), and in the case of sCκ-tP, a constant region of the κ chain (Cκ). S-tP and sCκ-tP were developed to provide targeted delivery of the siRNA, siRNA expressing cassettes (SEC), and siRNA-producing plasmids. Fluorescein isothiocyanate-siRNA, fluorescein isothiocyanate-SEC, and plasmid DNA were specifically delivered into HBsAg-positive cells using the sCκ-tP fusion protein, and effectively inhibited HBV gene expression and replication. HBV gene expression was also inhibited by siRNA or siRNA-producing plasmids in HBV transgenic mice. Conclusion: Our results describe a potential method for the targeted delivery of siRNA or siRNA-producing plasmids against HBV, using anti-HBsAg fusion proteins. (HEPATOLOGY 2007;46:84–94.)
Ribonucleic acid interference is a naturally occurring mechanism for silencing homologous genes.1, 2 Synthetic and plasmid-based siRNA expression systems are both effective at suppressing HBV infection, in vitro and in vivo,3–9 providing the basis for a new therapeutic strategy against HBV. However, the absence of a suitable delivery system presents a major obstacle against their clinical use.10–12 Targeted delivery of small interfering RNA (siRNA) to relevant cell populations should increase the therapeutic index and minimize potential side effects and cellular toxicity.
Antibody-mediated delivery is an effective method of targeting siRNA to particular cells. This involves fusing the nucleic acid–binding domain to a Fab fragment or scFv that recognizes a membrane receptor, resulting in a fusion protein that possesses cell recognition and nucleic acid–binding abilities. The fusion protein can then bind nucleic acids and deliver them into target cells through receptor-mediated endocytosis. This antibody-mediated uptake has been successfully used to deliver plasmid DNA and antisense oligonucleotides.13–17 Recently, Song et al.18 demonstrated that siRNA targeted delivered to human immunodeficiency virus–infected cells and HER2 positive cells by this strategy executed gene silencing. However, whether this promising strategy for siRNA delivery is applicable to other diseases remains unknown.19 In this report, we provide evidence that antibody-mediated siRNA delivery is effectively targeted to HBV-infected cells.
In a previous study, we have obtained 5 human anti-hepatitis B surface antigen (HBsAg) Fab antibodies from a phage display library and have reconstructed them as single-chain antibodies.20, 21 One of the single-chain antibodies, HBs 15, was effectively internalized into HBsAg-positive cells and was reconstructed into s-tP and sCκ-tP fusion proteins to evaluate the applicability of antibody-mediated siRNA delivery strategy in HBV-infected cells.
After evaluating the 5 siRNA-producing plasmids, a single plasmid and its corresponding siRNA and siRNA expression cassettes (SEC) were used to evaluate the inhibition on HBV gene expression and replication. The ability of the two s-tP fusion proteins to effectively deliver siRNA, SEC, or plasmid DNA to HBsAg-positive cells, and inhibit HBV gene expression were assessed both in vitro and in vivo.
FITC, fluorescein isothiocyanate; HBV, hepatitis B virus; HbeAg, hepatitis B e antigen; HbsAg, HBV surface antigen; SEC, siRNA expression cassettes; SDS-PAGE, sodium dodecyl sulfate polyacrylamide gel electrophoresis; siRNA, small interfering RNA.
Materials and Methods
HBsAg-positive HepG2.2.15 cells were grown in Dulbecco's minimum essential medium supplemented with 10% fetal bovine serum and 200 μg/ml G418 (Gibco, Karlsruhe, Germany). HBsAg-negative HepG2 cells were cultured in RPMI1640 supplemented with 10% fetal bovine serum. Cells were maintained at 37°C in a humidified incubator with 5% CO2.
Six- to 8-week-old, pathogen-free, male HBV transgenic mice weighing between 20 and 25 g were purchased from the medical department of Peking University, introduced from Jackson Laborotory,22 and maintained on a 12-hour dark/12-hour light schedule in the animal center.
Construction of siRNA-Producing Plasmids and Design of siRNA.
Five different HBV-specific RNA interference target regions were selected according to the HBV genome sequence (GenBank No. U95551): H1 (X) 5′-tgtcaacgaccgaccttga-3′, H2 (P+X) 5′-tgtgcacttcgcttcac ct-3′, H3 (P+S) 5′-ggctcagtttactagtgcc-3′, H4 (C) 5′-tagagtctccggaacattg-3′, and H5 (P+S) 5′-tcctgctgctatgcctcat-3′. The recombinant plasmid, pSUPER-H1∼H5, encoding the corresponding siRNA, was constructed as described previously.23, 24 Corresponding siRNA-H2 and FITC-siRNA were synthesized and obtained as annealed duplexes (Genepharma, Shanghai, China).
HBsAg and Hepatitis B e Antigen Assay.
HBsAg and hepatitis B e antigen (HBeAg) levels in the culture medium of the cell colonies and in the serum of treated mice were determined using the AXSYM systems kit (Abbott Diagnostics, North Chicago, IL),25 based on Microparticle Enzyme Immunoassay technology.
Real-time PCR was performed using the HBV fluorescence quantitative PCR Diagnostic Kit (DaAn, Guangzhou, China) described previously.26 The detailed procedure was performed according to the manufacturer's protocol.
Construction, Expression, and Purification of scFv and s-tP Fusion Proteins.
The Genbank numbers for the corresponding Fd and light chain used were AF455924 and U91942, respectively. The s-tP fusion sequence was amplified from 2 PCR reactions, with the truncated human protamine 1 encoding sequence designed in the 3′ scFv primers. A human Cκ sequence was amplified and inserted into the s-tP fusion sequence to construct the sCκ-tP fusion gene. Sequences encoding scFv, s-tP, and sCκ-tP were cloned into the pET28a and all three fusion proteins were expressed in Escherichia coli BL21 and purified by Ni2+-NTA agarose (Qiagen, Valencia, CA) according to the manufacturer's protocol.
Immunofluoresence and Immunohistochemistry.
To assess internalization, HepG2.2.15 and HepG2 cells were incubated on coverslips with purified s-tP or sCκ-tP fusion proteins for 30 minutes at 37°C, fixed, and immunofluorescence stained using penta-his monoclonal antibody (Qiagen).
For immunohistochemistry, formalin-fixed mouse livers were embedded in paraffin. Serial 5-μm sections were obtained using a Leica microtome. Immunostaining of HBsAg was performed using an immunohistochemical staining kit (Maixin, Fuzhou, China) with anti-human HBsAg mouse monoclonal antibody (Chemicon, Temecula, CA).
Gel-Shift Assay and EMSA.
HBsAg DNA fragments were radiolabeled by PCR with 32P-dCTP and purified using a PCR product purification kit (Promega, Madison, WI). Ten nanogram labeled DNA fragments or 200 ng unlabeled whole plasmid DNA were incubated with increasing amounts of the s-tP or sCκ-tP fusion proteins in 0.2 M NaCl solution at room temperature for 30 minutes before being run on an 0.8% agarose gel or 12% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). For auto-radiography, the polyacrylamide gel was dried, exposed on radiographic film at −70°C for 24 hours, and developed.
After 24 hours' incubation with a mixture of sCκ-tP or s-tP and FITC-labeled siRNA, SEC, or pEGFP-N3, which was pre-incubated in 0.2 M NaCl for 30 minutes, cells were collected and fixed with ice-cold 70% ethanol in phosphate-buffered saline before flow cytometry.
In Vivo Treatment with siRNA–Antibody or Plasmid–Antibody Complexes.
To detect FITC-siRNA localization, synthesized FITC-siRNA, alone or with sCκ-tP, were intravenously injected into HBV transgenic or C57 mice. The sCκ-tP and FITC-siRNA mixture was incubated for 30 minutes in 0.2 M NaCl before injection. Fifty micrograms FITC-siRNA was administered to each mouse at a molar sCκ-tP:siRNA ratio of 1:6. Mice were killed and tissues snap-frozen for cryosectioning 24 hours after injection. For treatment analysis, 50 μg siRNA-H2 or 40 μg pSUPER-H2 was incubated with sCκ-tP before administration.
Twenty micrograms RNA extracted from the livers of mice was separated by electrophoresis on a 12% agarose-formaldehyde gel and transferred to a Hybond-N nylon membrane (Roche Biochemica, Mannheim, Germany). Transcripts were detected by hybridization with 32P-labeled DNA probes prepared from the S gene region and labeled by PCR. For autoradiography, the gel was dried, exposed on radiographic film at −70°C for 24 hours, and developed.
Analysis of Hepatotoxicity and IFN-α Production.
ALT activity was assayed using the ThermoTrace Infinity ALT Reagent (ThermoDMA, Guangzhou, China) with methods described in the manufacturer's manual. Values are given as units/titer. IFN-α serum levels were measured using a quantitative IFN-α enzyme-linked immunosorbent assay kit (SenXiong, Shanghai, China). Absorbance was read on a Sunrise microplate reader at 492 nm.
All in vitro experiments were performed in two independent experiments performed in triplicate. All in vivo data were calculated from experiments performed in triplicate. The results represent means ± SEM. Statistical analysis was performed using the SPSS 13 software package for Windows (SPSS, Chicago, IL). Statistical significance was based on a P value of 0.05.
Potent and Persistent Inhibition of HBV Gene Expression and Replication by siRNA-Producing Plasmids.
Five target sequences were selected from the HBV genome, and corresponding siRNA-producing plasmids (pSUPER-H1 through 5) driven from the RNA polymerase III-promoter H1 were constructed (Fig. 1A) and transfected into HepG2.2.15 cells. Decreased HBsAg levels in the medium were observed 48 and 72 hours after transient transfection (Fig. 1B). Stable transfectants were established to evaluate the degree of inhibition. All of the 5 plasmids showed significant inhibition effect (Fig. 1C, D), with the pSUPER-H2 plasmid being most effective (Fig. 1E). The specificity of siRNA-H2 was confirmed in stable transfected HBx-EGFP or mutant HBx (mHBx)-EGFP HepG2 cells before being used for further experiments. siRNA-H2 inhibited HBx-EGFP expression in HBx-EGFP stable transfectants, but not mHBx-EGFP transfectants (Fig. 1F, G). In addition, when the stable HepG2.2.15 transfectants were continuously cultured for 10 months, the inhibitory effect was still maintained, as shown by the lower levels of HBsAg in the medium (Fig. 1H).
The s-tP Fusion Proteins Bind Both DNA and HBsAg and Are Internalized into HBsAg-Positive Cells.
The scFv, s-tP, and sCκ-tP encoding sequences were constructed, expressed and purified from E. coli (Fig. 2A, 2B). The s-tP and sCκ-tP fusion proteins retained HBsAg-binding ability as measured by indirect enzyme-linked immunosorbent assay (Fig. 2C). Both fusion proteins were still internalized into HBsAg-positive HepG2.2.15 cells (Fig. 2D). DNA binding activities of s-tP and sCκ-tP were examined by gel-shift assay and EMSA. When plasmid DNA or radiolabeled DNA fragments were mixed with increasing amounts of the s-tP or sCκ-tP fusion proteins, there was a reduction in the migration of whole plasmid DNA (Fig. 2E) and a slower migration of the DNA fragments (Fig. 2F). In contrast, DNA incubated with the scFv showed no significant change in gel mobility.
s-tP Fusion Proteins Specifically Delivered siRNA, SEC, and Plasmids Into HBsAg-Positive Cells.
To determine whether the s-tP fusion proteins could specifically deliver siRNA into HBsAg-positive cells, FITC-siRNA was added to HepG2.2.15 cells or HepG2 cells, alone or in combination with the scFv or s-tP fusion proteins before flow cytometric analysis. HepG2.2.15 cells were unable to take up FITC-siRNA alone, or a mixture of FITC-siRNA and scFv lacking protamine. However, while the HepG2.2.15 cells effectively internalized FITC-siRNA mixed with s-tP fusion protein, HepG2 cells could not. Both the HepG2.2.15 and HepG2 cells were unable to internalize FITC-siRNA that was mixed with the control csCκ-tP fusion protein, which contains a scFv against HER2 (Fig. 3A). Lipofectamine 2000 transfection was used as a positive control for delivery. The delivery efficiency analysis showed that both fusion proteins were most efficient when the siRNA: protein molar ratio was 6:1 (Fig. 3B), with the sCκ-tP fusion protein showing better delivery efficiency (P < 0.05). The sCκ-tP fusion protein was therefore chosen for subsequent experiments.
The ability of sCκ-tP to deliver SEC or plasmid DNA into cells was also evaluated. Fluorescence was detected in HepG2.2.15 cells after incubation with a mixture of sCκ-tP and FITC-SEC or pEGFP-N3 for 24 hours, but not in HepG2 cells (Fig. 3C). FITC-SEC delivery was comparable to FITC-siRNA delivery, but the number of EGFP-expressing cells was much lower (19%) after the sCκ-tP–mediated delivery of pEGFP-N3.
sCκ-tP Delivered siRNA, SEC, or siRNA-Producing Plasmids Inhibited HBV Gene Expression and Replication.
HepG2.2.15 cells were treated with a mixture of sCκ-tP and siRNA-H2, SEC-H2, or pSUPER-H2 that targeted the HBx gene. The levels of HBsAg, HBeAg, and HBV-DNA in the medium were all significantly lower in the 3 treated groups than the untreated control group at days 1, 2, 3, and 4 post incubation (Fig. 4A–C). The suppressed expression of HBsAg protein level was confirmed on day 2 by SDS-PAGE and Western blot analysis (Fig. 4D). Compared with SEC-H2 or pSUPER-H2, sCκ-tP delivered siRNA-H2 showed the strongest inhibitory effect. The mixture of siRNA-H2 and csCκ-tP targeting HER2 did not have an inhibitory effect.
sCκ-tP-Delivered siRNA and siRNA-Producing Plasmids Inhibited HBV Gene Expression in HBV Transgenic Mice.
To evaluate the sCκ-tP–mediated targeted delivery of siRNA in vivo, FITC-siRNA was injected into HBV transgenic mice alone or in combination with sCκ-tP fusion protein. FITC-siRNA was primarily detected where HBsAg-positive cells are located in the liver and kidneys, providing evidence that FITC-siRNA was efficiently targeted (Fig. 5A). Confocal fluorescence microscopy of siRNA and HBsAg showed that FITC-siRNA colocalized with HBsAg-positive hepatocytes, confirming the specificity of delivery (Fig. 5B). In contrast, FITC-siRNA levels were much lower in the livers and kidney of mice injected with FITC-siRNA alone. Faint fluorescence staining was observed in other tissues. The distribution of FITC-siRNA was similar regardless of the presence of sCκ-tP in normal C57 mice (data not shown).
To evaluate the inhibitory effect of the targeted delivered siRNA or siRNA-producing plasmids in vivo, siRNA-H2 with sCκ-tP or csCκ-tP, or pSUPER-H2 with sCκ-tP, or siRNA-H2, pSUPER-H2, or sCκ-tP alone were injected into HBV transgenic mice. Hydrodynamic injections were used in the siRNA or siRNA-producing plasmid only groups. HBV mRNA levels was detected by Northern blot analysis at day 4 post injection. As shown in Fig. 5C, mice treated with sCκ-tP and siRNA-H2 or pSUPER-H2, or siRNA-H2 or pSUPER-H2 alone, had lower HBV RNA levels than untreated controls. Mice that received siRNA-H2 with csCκ-tP or sCκ-tP alone had similar HBV RNA level as untreated mice. Serum HBsAg levels decreased over 11 days in mice treated with sCκ-tP and siRNA-H2 or pSUPER-H2. In contrast, serum HBsAg levels decreased until day 7 and began to increase in mice treated with siRNA-H2 or pSUPER-H2 alone. Mice injected with sCκ-tP alone also exhibited decreased serum HBsAg levels, which may have been a neutralization effect (Fig. 5D). The inhibition of HBsAg expression in the liver was further confirmed at day 7 using immunohistochemistry. In all treatment groups except those receiving sCκ-tP alone, HBsAg expression was significantly lower than the untreated and csCκ-tP control groups (Fig. 5E).
sCκ-tP–Mediated Delivery of siRNA or siRNA Producing Plasmids Does Not Induce an IFN-α Response and Has No Hepatotoxicity in Transgenic Mice.
To assess whether our treatment induced an IFN response, serum IFN-α levels were measured. No obvious increase in IFN-α was detected after siRNA or siRNA-producing plasmid administration, either alone or in the presence of sCκ-tP (Fig. 5F). Hepatotoxicity was also assessed by evaluating the serum ALT levels. ALT levels rose slightly on days 1 and 4 in sCκ-tP and csCκ-tP–treated groups, but decreased to normal levels by day 7. The relatively stable ALT levels indicated that the specific delivery of siRNA or siRNA-producing plasmids mediated by sCκ-tP induced no obvious hepatotoxicity in transgenic mice (Fig. 5G).
Several strategies have been designed to target siRNA delivery,16, 19, 27–30 among which antibody-mediated delivery has the greatest potential. This method of siRNA delivery is particularly attractive for the following reasons: (1) chemical coupling of siRNA with the antibody fusion protein is not required, (2) stability of the siRNA complex in the blood is increased, (3) the complex binds specifically to the cellular plasma membrane, (4) the complex is actively internalized into the target cells, and (5) siRNA induces the desired biological effect on its target cells.
To successfully apply antibody-mediated targeted siRNA delivery to hepatitis B therapy, it is vitally important to identify specific targets on HBV-infected cells. During HBV replication and production, HBsAg is overproduced, such that large amounts are expressed on the surface of infected hepatocytes. Whereas the presence of HBsAg on the plasma membrane provides a target for the immune response, it also provides a target for siRNA therapeutics against HBV.31–35
siRNA or siRNA-producing plasmids must enter the cytoplasm to initiate RNA interference cascades, so it is important to assess whether the antibodies being used for targeting can be efficiently internalized into antigen-positive cells.36 In the current study, the scFv we used to target siRNA delivery can be effectively internalized into HBsAg-positive cells. As shown in Fig. 2D, fusion of a truncated protamine to the scFv and insertion of the Cκ sequence did not abrogate internalization. By using the sCκ-tP fusion protein, siRNA, SEC, and plasmid DNA were detected to be specifically delivered into HBsAg-positive HepG2.2.15 cells and effectively inhibited HBV gene expression and replication.
The targeted delivery of siRNA mediated by the sCκ-tP fusion protein was also evaluated in HBV transgenic mice, which are genetically modified by a HBV DNA fragment encoding the HBsAg, pre-S, and X antigens. FITC-siRNA was shown to be targeted delivered to the liver and kidneys where HBsAg-positive cells are located. Inhibition of HBV gene expression by siRNA–antibody or plasmid–antibody complexes was also detected in HBV transgenic mice. The inhibitory effect progressed for 11 days, and the inhibition induced by sCκ-tP-delivered siRNA or siRNA-producing plasmids was more effective at days 7 and 11 than the siRNA or plasmid only group, suggesting that they had a more persistent inhibitory effect.
Some reports caution that some siRNA has off-target effects, inducing Toll-like receptor (TLR) 7 or 8 signaling and IFN-α secretion.37, 38 In our experiment, no obvious increase in IFN-α level was detected after the treatment. Whereas the IFN response may be cell type specific, siRNA binding to the sCκ-tP fusion protein may prevent siRNA from being recognized by TLRs, thus avoiding the induction of an IFN response.
Our results show great potential for the use of siRNA in anti-HBV therapy, but some problems still need to be resolved before it can be clinically applied. First, high serum HBsAg levels in some chronic HBV patients may affect the therapeutic effect of siRNA as a result of partial neutralization of the sCκ-tP fusion protein by HBsAg. Pre-neutralization with anti-HBsAg monoclonal antibodies or scFv without the protamine fusion, or another method of reducing HBsAg levels before administration of the siRNA–antibody or plasmid–antibody mixture will help prevent this issue. Second, a slight increased ALT level was observed in mice treated with the prokaryotic purified sCκ-tP or csCκ-tP fusion proteins. Because the fusion protein used is a single-chain antibody, it seems unlikely to be caused by immune complex deposition. It might be caused by some content contaminated in the purified fusion protein, such as endotoxin. This problem may be resolved by using another purification system. Third, although siRNA or siRNA-producing plasmids were successfully delivered and remained functional, the mechanism of how the siRNA or siRNA-producing plasmids leave the endosomes and enter the cytoplasm remains unclear. Further experiments are required to define the exact trafficking pathway.
In conclusion, our results demonstrate that HBsAg is an ideal target for the specific delivery of siRNA or siRNA-producing plasmids against HBV. sCκ-tP-mediated delivery of siRNA and siRNA-producing plasmids effectively inhibited HBV gene expression, both in vitro and in vivo. Our findings should increase the potential of siRNA therapeutics for clinical application in anti-HBV therapies.
The authors thank Prof. Yun-Qing Li for confocal microscopy, Yun-Xin Cao for flow cytometry, Ying-Mei Wang and Fu-Cheng Ma for histological analyses, Dr. Nan-Chun Chen, Xin-Hai Zhang and Zhen-Yan for useful suggestions, Dr. Chang-Hong Shi for the support of animal experiment, and Dr. Xiao-Ying Lei and Yi Wan for assistance in the statistical analysis.