Engineering lentivirus envelope VSV-G for liver targeted delivery of IDOL-shRNA to ameliorate hypercholesterolemia and atherosclerosis

Lentiviral vectors (LVs) have been widely used as a tool for gene therapies. However, tissue-selective transduction after systemic delivery remains a challenge. Inducible degrader of low-density lipoprotein receptor is an attractive target for treating hypercholesterolemia. Here, a liver-targeted LV, CS8-LV-shIDOL, is developed by incorporating a hepatocyte-targeted peptide derived from circumsporozoite protein (CSP) into the lentivirus envelope for liver-targeted delivery of IDOL-shRNA (short hairpin RNA) to alleviate hypercholesterolemia. Tail-vein injection of CS8-LV-shIDOL results in extremely high accumulation in liver and nearly undetectable levels in other organs in mice. In addition, it shows superior therapeutic efficacy in lowering serum low-density lipoprotein cholesterol (LDL-C) and reducing atherosclerotic lesions over unmodified LV-shIDOL in hyperlipidemic mice. Mechanically, the envelope-engineered CS8-LV-shIDOL can enter liver cells via low-density lipoprotein receptor-related protein (LRP). Thus, this study provides a novel approach for liver-targeted delivery of IDOL-shRNA to treat hypercholesterolemia by using an envelope-engineered LV, and this delivery system has great potential for liver-targeted transgene therapy.


INTRODUCTION
Lentiviral vectors (LVs) have been widely used as a tool for gene and cell therapies, and the number of gene therapy clinical trials using LVs is increasing. 1,2To date, two lentivirus-based ex vivo gene therapies, Kymriah (Novartis) and Zynteglo (bluebird bio), have been approved. 3Besides, direct in vivo applications of lentivirus for treating patients with Parkinson's disease and age-related macular degeneration (AMD) have been reported. 4,5Compared with other retrovirus vectors, LVs possess several major advantages, which include efficient transduction of nondividing cells, low cytotoxicity and immunogenicity, relatively large transgene capacity of approximately 8-9 kb, and long-term transgene expression. 6,7Notably, LVs are commonly pseudotyped with the vesicular stomatitis virus envelope glycoprotein (VSV-G), which confers not only high particle stability and high viral titers but also broad cell tropism to the LVs, while the wide cell tropism can lead to the binding of VSV-G pseudotyped LVs to the surface of any cell encountered before reaching the target cells. 8Hence, improving the selective transduction of LVs to the target cells after systemic delivery remains a challenge.
As the crystal structure of VSV-G has been determined, 9,10 it is feasible to rationally design and engineer VSV-G to adapt the lentiviral tropism to the particular target cells.Previously, it has been identified that the circumsporozoite protein (CSP) of Plasmodium falciparum plays an important role in sporozoite entry into hepatocytes, 11,12 which is attributed to the specific binding of P. falciparum CSP peptides to hepatocytes. 13Similarly, two peptide segments derived from ApoB-100 have been shown to mediate the binding of low-density lipoprotein (LDL) to the LDL receptor (LDLR). 14Thus, we hypothesize that these peptides can be potentially exploited as the targeting ligands for engineering VSV-G to restrict LV entry to liver cells.
The inducible degrader of LDLR, also known as MYLIP, is an unique E3 ubiquitin ligase 15 that directly interacts with the cytoplasmic tail of LDLR and facilitates its ubiquitination and subsequent degradation in lysosome, 16,17 thus decreasing hepatic LDLR levels and reducing hepatic LDL uptake. 18,191][22] Recently, a cyclic peptide has been shown to dose-dependently increase LDLR levels via inhibiting the homodimerization of IDOL E3 ubiquitin ligase in hepatic cells. 235][26][27] Thus, developing RNAi therapeutics to suppress hepatic IDOL expression appears to be an alternative approach to treat hypercholesterolemia. Yet how to efficiently and selectively deliver therapeutic small interfering RNAs (siRNAs) against IDOL to hepatocytes after systemic delivery needs to be addressed.In the present study, we sought to develop a liver-targeted LV by engineering the envelope of lentivirus to facilitate targeted delivery of IDOL-shRNA (short hairpin RNA) to liver to more effectively ameliorate hypercholesterolemia and atherosclerosis (Figure 1), while also providing a potential lentivirus delivery system for liver-targeted transgene therapy.

Rational design and engineering of a liver-targeted LV system
Two hepatocyte-targeted peptides were selected as targeting ligands for engineering VSV-G protein.One was a 9 amino acid peptide (RLTRKRGLK) derived from ApoB-100 segments (AP), 14 and the other was a 20 amino acid peptide (HNMPNDPNRNV DENANANSAYC) derived from P. falciparum CSP. 13 Meanwhile, the prefusion form of VSV-G crystal structure was used to explore the feasible insertion sites. 10The amino acid sequences of glycoproteins G from various vesicular stomatitis virus serotypes were aligned and analyzed to find potential sites that might tolerate insertion of a foreign peptide.As shown in Figure 2, multiple potential insertion sites in VSV-G were identified on the basis of three criteria: flexible loop regions, unconserved regions, and regions rich in hydrophilic amino acids. 28,29Here, the amino acid was numbered referring to the literature elucidating VSV-G crystal structure. 9,10Thus, eight sites on VSV-G were chosen for incorporation of hepatocyte-targeting peptides: one site was within the N terminus of VSV-G protein, two sites were within the domain I region, three sites were within the domain III region, and the other two sites were located within the loop region (Figure 3).By site-directed mutagenesis, nine envelope variants with AP peptide incorporation at above eight sites were constructed, respectively (Table S1).Of them, three yielded viable vectors: two with AP peptide insertion at the N terminus of VSV-G, next to the signal peptide with or without a linker (GGGGS), and the other with AP peptide insertion between amino acids 201E and 202G.Similarly, nine envelope variants with CSP peptide insertion were constructed, respectively (Table S1).Among them, four yielded viable vectors: two with CSP insertion at the N terminus of VSV-G next to the signal peptide with or without a linker (GGGGS) and the other one with CSP peptide insertion between amino acids 174K and 175G, the 4th one with CSP peptide insertion between amino acids 351T and 352T.The representative titers of LVs pseudotyped with VSV-G variants are shown in Figures 3D and 3E.
The inhibitory potency of these selected siRNAs on IDOL expression was tested by western blot analysis in human HepG2 and mouse Hepa1-6 cells, respectively.After transfection with 30 nM siRNAs, it was found that siIDOL-619 and siIDOL-622 potently inhibited the expression of IDOL by $60%, compared with the scramble siRNA group, whereas the others appeared less effective in HepG2 cells (Figures S1A and S1B).Similar results were observed in mouse Hepa1-6 hepatic cells (Figures S2A and S2B).Also, we detected the regulatory effects of these siRNAs on the function of LDL cholesterol (LDL-C) uptake through suppressing IDOL expression in HepG2 and Hepa1-6 cells.After transfection with 30 nM siIDOL-485, siIDOL-619, siIDOL-622, siIDOL-1406, and siI-DOL-1508, DiI-LDL uptake assay was performed.It was observed The VSV-G full-length plasmid pMD2.g was used for the construction of variants with peptide insertions.(H and I) Lentiviral shRNA expression plasmids LV-shNC and LV-shIDOL with ZsGreen tag were constructed for lentivirus production with helper plasmid pMD2.g(F) and psPAX2 (G).The shRNA expression vectors were constructed using the U6 pol III promoter on the basis of a 19 nt siRNA sequence (scramble and siIDOL-622) (Table S3).All shRNA sequences have identical loop sequences.
that siIDOL-622 most potently increased LDL-C uptake by $30% compared with control in HepG2 cells (Figure S1C).Similar results were obtained in mouse Hepa1-6 cells (Figure S2C).Furthermore, we verified that siIDOL-622 treatment dramatically reduced IDOL levels, resulting in a significant increase in LDLR expression at the cell surface in HepG2 (Figures S1D and S1E) and Hepa1-6 cells (Figures S2D and S2E) by immunofluorescence and flow cytometry analyses.Thus, siIDOL-622 was selected for the subsequent construction of IDOL-shRNA LVs.
The siIDOL-622 sequence and a scrambled siRNA (catalog #A06001NC-RL; Genepharm) beginning with a G were embedded in shRNA scaffolds as previously described. 30The shRNA oligos contain restriction sites at both ends, a 19 nt sequence that is identical to siRNA sense strand, a hairpin loop sequence, and a 19 nt sequence that is the reverse complement of siRNA sense strand, as well as a poly T termination sequence.The above shRNA sequences are listed in Table S3.The lentiviral expression vectors pLVX-shIDOL and pLVX-shNC were constructed by subcloning the synthetic annealed shRNA duplexes at BamH I/EcoR I sites of pLVX-shRNA2 (catalog #632179; Clontech), allowing shRNA duplexes to be driven by U6 pol III promoter.

CS8-LV-shIDOL has extremely high accumulation in liver
To assess the biodistribution of hepatocyte-targeting peptide-incorporated lentiviral shRNA vectors in vivo, CS8-LV-shIDOL and LV-shIDOL were administered intravenously via tail vein to mice, respectively.One week after injection, mice were sacrificed, and the fluorescence intensity in five different organs (heart, liver, kidney, lung, and spleen) was detected.As shown in Figure 5, wild-type LV-shIDOL was accumulated mainly in liver and kidney and was less distributed in heart, spleen, and lung, while the liver-targeted LV CS8-LV-shIDOL had extremely high accumulation in liver and nearly undetectable levels in other four organs.The ex vivo images of dissected tissues taken by fluorescence microscopy showed similar results (Figure 5C).Thus, a novel liver-targeted lentivirus delivery system has been well established by incorporating the hepatocyte-targeting ligand CS8 peptide into VSV-G protein.

In vivo biosafety of CS8-LV-shIDOL
To test the in vivo biosafety of CS8-LV-shIDOL, we assessed routine blood parameters by measuring plasma alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (ALP), plasma urea, and creatinine (CREA).Biochemical assay results showed that the blood levels of ALT, AST, ALP, urea, and CREA in mice after injection with CS8-LV-shIDOL for 7 days were not significantly different altered compared with those of the saline-treated group (Figures S6A-S6E), demonstrating that CS8-LV-shIDOL did not show significant toxicity on the liver and kidney.Besides, no tissue or cell damage was observed in heart, liver, spleen, lung, and kidney in mice receiving CS8-LV-shIDOL after histology examination by H&E staining (Figure S6F).

CS8-LV-shIDOL shows enhanced therapeutic efficacy in lowering serum lipid and alleviating atherosclerotic lesions in mice
The therapeutic effects of liver-targeted LV CS8-LV-shIDOL on hyperlipidemia and atherosclerosis were evaluated in hyperlipidemic mice.][33] Meanwhile, mice received 1 Â 10 8 TU LV-shNC (negative control), LV-shIDOL, and CS8-LV-shIDOL per animal intravenously via the tail vein every eight weeks.Twenty-four hours after final injection of P407, mice were fasted for 8 h and then euthanized for blood and tissue collection.By analysis on the serum lipids, it was shown that the LV-shIDOL and CS8-LV-shIDOL groups had 29.74% and 45.82% decreases in serum LDL-C, as well as 14.98% and 30.05% decreases in serum total cholesterol (TC), respectively compared with the LV-shNC group (Figures 6B and 6C).The serum LDL-C and TC levels in the liver-targeted CS8-LV-shIDOL group were significantly lower than those in non-targeted LV-shIDOL group, indicating that incorporation of the hepatocyte-targeting ligand CS8 peptide into VSV-G protein facilitated liver-specific cellular uptake of LV-shIDOL and led to enhanced IDOL silencing efficiency.Furthermore, atherosclerotic lesion development of aortic root was measured using oil red O staining, it was found that atherosclerotic plaques in the CS8-LV-shIDOL-treated group were significantly milder than those in LV-shIDOL group and much less severe than those in the salineand LV-shNC-treated groups (Figure 6D).Although both LV-shIDOL and CS8-LV-shIDOL remarkably inhibited the expression of hepatic IDOL, CS8-LV-shIDOL showed an increase of $60% in IDOL silencing efficiency, leading to higher potency in up-regulating LDLR, compared with non-targeted LV-shIDOL (Figures 6E-6G).This was further confirmed by immunofluorescence analysis (Figure 6H).Taken together, liver-targeted CS8-LV-shIDOL had higher efficiency in hepatotropic delivery of IDOL-shRNA than non-targeted LV-shIDOL, leading to enhanced potency in inhibiting IDOL, up-regulation of LDLR, reduced serum LDL-C and TC, and improved therapeutic efficacy in reducing atherosclerotic lesions in hyperlipidemic mice.

CS8-LV-shIDOL exhibits improved therapeutic effects in reducing hepatic lipid accumulation and attenuating liver injury in mice
To assess the therapeutic efficacy of CS8-LV-shIDOL on reducing hepatic fat accumulation and attenuating liver injury in mice, histological analyses on liver sections using oil red O and H&E staining were performed.As shown in Figure 7A, CS8-LV-shIDOL treatment VSV-G variants, respectively.After 48 h, the levels of IDOL mRNA in HepG2 cells were quantified using qRT-PCR (A).After 72 h, the levels of IDOL and LDLR protein were measured using western blot (B and C), DiI-LDL uptake levels were measured using a multimode reader (D), then normalized to mock control.(E-G) Effects of CS8-LV-shIDOL on IDOL and LDLR expression in HepG2 cells.After transfection with LV-shNC, LV-shIDOL, and CS8-LV-shIDOL for 72h, IDOL protein levels were visualized using immunofluorescence assay (F).The cell surface LDLR protein levels were determined using flow cytometry (E) and evaluated using immunofluorescence assay (G).*p < 0.05, **p < 0.01, and ***p < 0.001 vs. mock control; ### p < 0.001 and #### p < 0.0001 vs. LV-shNC control (unpaired Student's t test).Results are given as mean ± SEM of three independent experiments.reduced hepatic lipid deposition more potently than LV-shIDOL treatment, although both notably decreased lipid deposition compared with the shNC treatment group.Additionally, CS8-LV-shIDOL treatment alleviated liver injury more effectively than in the LV-shIDOL group (Figure 7B).Furthermore, it was confirmed that administration of liver-targeted CS8-LV-shIDOL caused an average 42.3% decrease in hepatic triglyceride (TG) content, which appeared to be lower than the 28.9% reduction attained with non-targeted LV LV-shIDOL (Figure 7B).Taken together, these data demonstrated that liver-targeted CS8-LV-shIDOL was much more efficacious than non-targeted LV-shIDOL in reducing hepatic lipid deposition and ameliorating liver injury in mice.

LRP mediates the targeted entry of CS8-LV-shIDOL into hepatocytes
To identify the receptor that mediates the targeted entry of CS8-LV-shIDOL into hepatocytes, HepG2 cells were incubated with anti-LDLR monoclonal antibody at 37 C for 1 h to mask cell surface LDLR, a natural entry receptor for vesicular stomatitis virus, 34 which is widely expressed in various tissues, such as liver, spleen, intestine, adrenals, neurons, microglia, and white and brown adipose tissues. 21,22Then, LV-shNC, LV-shIDOL, and CS8-LV-shIDOL were added.After incubation at 37 C for 30 min, the supernatant containing LV was replaced with fresh MEM medium and incubated at 37 C for additional 16 h.Subsequently, the plates were rinsed with PBS and the infected HepG2 cells were analyzed for ZsGreen expression.It was observed that wild-type LV infection was efficiently blocked by anti-LDLR monoclonal antibody, while CS8-LV-shIDOL displaying CS8targeting ligand retained infectivity (Figures 8A-8C).However, when HepG2 cells were treated with monoclonal antibodies against both LDLR and low-density lipoprotein receptor-related protein (LRP) infection of HepG2 cells by CS8-LV-shIDOL was completely blocked (Figures 8A-8C).Thus, it was demonstrated that the targeted entry of CS8-LV-shIDOL into hepatocytes was mediated by the interaction of CS8 ligand with the LRP.

In silico docking of CS8-LV-shIDOL with LRP
The ligand-receptor interaction details for the affinity of CSP peptide to LRP were investigated using computer-based homology modeling and molecular docking analyses.First, the three-dimensional (3D) structure of CS8 peptide-incorporated VSV-G (GenBank: OQ561788) was built on the basis of the wild-type VSV-G crystal structure (PDB: 5I2S) (Figure 8D) using I-TASSER (http://zhanglab.ccmb.med.umich.edu/I-TASSER/).As shown in Figure 8E, it was observed that the targeting ligand CSP peptide was properly displayed without disturbing the native structure of VSV-G, implying the VSV-G engineered CS8-LV-shIDOL might retain a high level of vector titer, enabling it to favorably interact with the target receptor LRP.After-ward, the modeled structure of CS8 peptide-incorporated VSV-G was docked with the crystal structure of LRP1 CR.7 (PDB: 1J8E) using HawkDock web server (http://cadd.zju.edu.cn/hawkdock/), and the binding free energy (DG bind ) of the ten top-ranked binding poses was calculated in HawkDock web server using the molecular mechanics/GB surface area (MM/GBSA) method.On the basis of the results of docking, the docked pose with the lowest binding free energy was selected as the best conformation to visualize the key interacting residues in the docked complex using Molecular Operating Environment for Windows (Chemical Computing Group Inc., Montreal, QC, Canada).As indicated in Figure 8F, it was revealed that the interaction residues in CS8 peptide-incorporated VSV-G including Asn9 and Lys 11 (in VSV-G), Asn359, and Arg360 (in CS8 peptide) form four hydrogen bonds against Asn13, Ser14, Asp32, and Thr37 in LRP1 CR.7.

DISCUSSION
In the field of gene therapy, therapeutic viruses have provided treatment options for diseases that are beyond traditional approaches. 35ompared with other virus vector systems, LVs hold a unique set of advantages, including low immune response, high cargo capacity, and efficient transduction of nondividing cells. 7,36,37To date, two lentivirus-based ex vivo gene therapies, Kymriah (Novartis) and Zynteglo (bluebird bio), have been approved. 3In addition, direct in vivo clinical applications of lentivirus have been reported for the treatment of Parkinson's disease and AMD. 4,5However, the broad tropism of VSV-G pseudotyped lentivirus makes it a non-ideal gene therapy vector, 38 thus it is required to augment its infection specificity to the particular target cells in order to enhance the therapeutic efficacy and minimize the systemic toxicity.In this work, we developed a novel liver-targeted lentivirus-based system for targeted delivery of IDOL-shRNA into the liver with an improved cholesterol-lowering efficacy.
Since the crystal structure of the prefusion form of VSV-G has been previously delineated, 10 it appears to be feasible to identify the potential sites in VSV-G that are exposed on the molecule surface and tolerant of the insertion of targeting ligands via computational analysis.Theoretically, flexible loop regions with unconserved sequences and rich in hydrophilic amino acids are preferred because of their favorable accessibility to the protein surface. 28,29Accordingly, eight potential sites which could tolerate an insertion of 20-amino-acidlong peptide were identified in VSV-G.Subsequently, two liver-targeted peptides were incorporated into VSV-G at these potential insertion sites, respectively.By analysis on the titers of LVs pseudotyped with VSV-G variants, 7 VSV-G variants with titers comparable with the parent were selected for subsequent IDOL-shRNA LV construction.
Figure 5. Biodistribution of CS8-LV-shIDOL in C57/BL6 mice (A and B) Liver-targeted lentiviral vector CS8-LV-shIDOL was injected into the tail veins of C57/BL6 mice at a dose of 1 Â 10 8 TU per mouse.Control mice were injected with an equivalent dose of wild-type lentiviral vector LV-shIDOL or saline.One week after injection, mice were sacrificed, and the heart, kidney, liver, lung, and spleen were taken for imaging ex vivo.Fluorescence in collected organs was measured and quantified using the IVIS Imaging System.(C) Tissues were incubated at 4 C in 4% paraformaldehyde for 48 h.Then, sections were stained using DAPI prior to viewing with a fluorescence microscope.
1][22] To inhibit hepatic IDOL expression in this work, various IDOL-targeting siRNAs were designed on the basis of the homologous regions in IDOL mRNA sequence across human, rhesus monkey, mouse, rat, and golden hamster.By in vitro assay, siIDOL-622 was screened out with the most potent silencing effect on IDOL expression, the strongest regulatory function in up-regulating LDLR and increasing LDL-C uptake in HepG2 and Hepa1-6 cells.
In order to facilitate liver-targeted delivery of therapeutic siRNA against IDOL to hepatocytes after systemic delivery, on the basis of siIDOL-622 sequence, we constructed recombinant lentivirus IDOL-shRNAs pseudotyped with various liver-targeting ligandmodified VSV-G envelopes.After in vitro assay, it was found that CS8-LV-shIDOL had the highest potency in repressing IDOL expression, up-regulating LDLR, and elevating LDL-C uptake in HepG2 and Hepa1-6 cells, implying that incorporation of hepatocyte-targeting peptide CS8 into VSV-G did not disturb the envelope function.
Furthermore, we investigated the targeted delivery of CS8-LV-shIDOL to hepatocytes in vivo.It was observed that wild-type LV LV-shIDOL was distributed mainly in liver and kidney and less distributed in heart, spleen, and lung, while the liver-targeted LV CS8-LV-shIDOL had extremely high accumulation in liver and nearly undetectable levels in other four organs, indicating that incorporating the hepatocyte-targeting ligand CS8 peptide into VSV-G protein greatly improved the liver-targeted delivery of IDOL-shRNA to liver.Besides, we evaluated the in vivo therapeutic effects of CS8-LV-shIDOL on hypercholesterolemia and atherosclerosis in P407induced hyperlipidemic mice fed an atherogenic diet.The data indicated that CS8-LV-shIDOL had much enhanced therapeutic efficacy in reducing serum lipid and alleviating atherosclerotic lesions in mice compared with wild-type LV-shIDOL group.Beyond that, CS8-LV-shIDOL treatment showed superior therapeutic effects in attenuating hepatic lipid accumulation and alleviating liver injury over wild-type LV-shIDOL in mice.However, how IDOL inhibition is linked to the reduction of hepatic lipid and the attenuation of liver injury remains to be addressed.
Additionally, we explored the mechanism by which the hepatocytetargeting CS8 peptide promoted the liver-targeted delivery of CS8-LV-shIDOL to liver.It has been known that wide-type VSV-G pseudotyped lentivirus enters cells through LDLR, a highly ubiqui-tous receptor, which leads to its broad tropism. 34,39,40Interestingly, the CSP of P. falciparum is involved in sporozoite entry into hepatocytes, 11,12 and interacts with LRP in liver cells. 41As an LDLR family member, LRP is expressed at functionally significant levels in hepatocytes and LRP-mediated endocytosis of apoE-enriched chylomicron remnants occurs largely in liver. 42,43Also, a peptide of T7 phage protein has been revealed to mediate the targeting of hepatocytes via LRP. 44Thus, we speculated that CS8-LV-shIDOL can target liver cells via LRP.To test this speculation, an LDLR antibody was used to mask the LDLR in HepG2 cells prior to infection, thus allowing the wildtype LV-shIDOL infection was blocked.As a result, the envelope-engineered CS8-LV-shIDOL showed much higher infection levels relative to wild-type LV-shIDOL.Conversely, when the cells were treated with monoclonal antibodies against both LDLR and LRP, CS8-LV-shPCSK9 infection was completely blocked.Thus, we demonstrated that LRP mediates the targeted entry of CS8-LV-shIDOL into hepatocytes.Moreover, we uncovered the ligand-receptor interaction details for the affinity of CS8 peptide-incorporated VSV-G to LRP using computer-based homology modeling and in silico docking analysis.
In summary, we successfully developed a novel liver-targeted lentivirus vector system by incorporating a hepatocyte-targeted peptide derived from CSP into the lentivirus envelope for liver-targeted delivery of IDOL-shRNA to specifically inhibit hepatic IDOL expression, thus enabling more potent reduction of hypercholesterolemia and alleviating atherosclerosis.This hepatocyte-targeted lentivirus delivery system has great potential for liver-targeted transgene therapy.

Construction of lentivirus envelope mutant plasmids
The VSV-G expression plasmid pMD2.g was used as the template for all mutant constructions.Sites for mutagenesis were chosen on the basis of a VSV-G crystal structure, 9,10 which provided the means for the preliminary structural identification of surface loop regions that were envisioned possibly tolerating foreign peptide sequence insertions.Liver cell-specific peptides derived from AP (RLTRKRGLK) 14 and P. falciparum circumsporozoite (CS) protein (HNMPNDPNRNVDENANANSAYC) 13 were inserted into appropriate places in VSV-G protein by overlapping PCR.The peptide sequences and insertion sites on VSV-G are listed in Table S1.Insertion mutation of liver-targeting peptides in VSV-G encoded in pMD2.G (Addgene, Cambridge, MA) was carried out using two sets of primers for every insertion.The primers used are listed in Table S4.The resulting PCR products and the plasmid pMD2.G were digested with restriction enzymes and ligated to obtain the modified VSV-G constructs.

Design of siRNAs targeting IDOL and construction of lentiviral shRNA vectors
The mRNA sequence of IDOL can be retrieved using RefSeq in the National Center for Biotechnology Information (NCBI) Entrez Gene database, and DNAMAN (version 6) (Lynnon Biosoft) was used to search for mRNA conserved sequences.The IDOL-targeting siRNAs were designed using DSIR version 7 (http://biodev.extra.cea.Fr/DSIR/DSIR.html) 45and RNAstructure version 6.4 (http://rna.urmc.rochester.edu/RNAstructure.html). 46Moreover, the homology analysis of the candidate siRNAs was conducted by using BLAST pro-vided from NCBI GeneBank with the transcript reference sequence database (Transcript Reference Sequences) for excluding the possibility that siRNA non-specifically inhibits genes other than IDOL.All siRNAs (Table S2) were synthesized and purified by GenePharma (Shanghai, China) as well as dissolved to a concentration of 20 mM with sterilized and RNase-free water according to the manufacturer's protocol.The selected IDOL-targeting siRNA and a scrambled siRNA (catalog #A06001NC-RL; Genepharm) were embedded in shRNA scaffolds as previously described, 30 and the shRNA sequences used in this study are listed in Table S3.The lentiviral expression vectors pLVX-shIDOL and pLVX-shNC were constructed by subcloning the synthetic annealed shRNA duplexes at BamH I/EcoR I sites of pLVX-shRNA2 (catalog #632179; Clontech).

Production of recombinant LVs with modified envelopes
LVs were produced as previously described 47 with slight modification.Briefly, lentiviral plasmids encoding for VSV-G variants, matrix proteins (psPAX2), and ZsGreen (pLVX-shRNA2) were used to transfect 293T cells by 25 kDa LPEI.Forty-eight and 72 h post-transfection, lentivirus containing supernatant was harvested and concentrated using a precipitation-based method using PEG8000.Twentyfive percent polyethylene glycol (PEG) 8000 and 0.75 M NaCl were added to LV containing supernatant and incubated at 4 C for 5 h, with mixing every 20 min.Subsequently, the mixture was pelleted at 7,000 Â g for 15 min, and the pellet was resuspended in PBS.The final concentrated lentivirus was titered immediately and then aliquoted and stored at À80 C for further analysis.Viral titers were determined as described previously. 48,49The titer of viral preparations with different VSV-G variants was measured by performing transduction serial dilutions of concentrated lentivirus in 96-well plates with 293T cells seeded at 30,000 cells per well in presence of 8 mg/ mL polybrene.The fraction of transfected cells (ZsGreen) was determined by counting fluorescent versus nonfluorescent cells using a Zeiss AX10 fluorescence microscope.The titer was calculated on the basis of the following formula: TU ml À1 = (F Â N Â D Â 1,000)/V, where F is the percentage of fluorescent cells (ZsGreen), N is the number of cells at the time of transduction (corresponding to about 1 Â 10 À5 cells per well), D is the fold dilution of vector sample used for transduction, and V is the volume of diluted vector sample added into each well for transduction.

Cell culture and treatments
Human hepatoma (HepG2), mouse hepatoma (Hepa1-6), and HEK293T cells were obtained from China Infrastructure of Cell Line Resources (Beijing, China) and maintained in MEM (HepG2 cells) or DMEM (Hepa1-6 and HEK293T cells) supplemented with 10% FBS, 100 U/mL penicillin, and 100 mg/mL streptomycin in a humidified 5% CO 2 atmosphere at 37 C. Mouse normal cell line AML12 was obtained from the Cell Bank of the Chinese Academy of Science (Shanghai, China) and maintained in DMEM/F12 supplemented with 10% FBS, 10 mg/mL insulin, 5.5 mg/mL transferrin, 5 ng/mL selenium, 40 ng/mL dexamethasone, 100 U/mL penicillin, and 100 mg/mL streptomycin in a humidified 5% CO 2 atmosphere at 37 C.For LDLR protein quantification and DiI-LDL uptake measurement experiments, Opti-MEM medium was used instead of medium containing 10% FBS after transfection for 48 h.
For siRNA transfection, about 5 Â 10 5 HepG2 or Hepa1-6 cells were seeded per well into a 6-well plate and transfected with 30 nM negative control or 30 nM specific IDOL siRNAs (Table S2) using Lipofectamine 3000 (Invitrogen), respectively.After transfection for 48 h, cells were rinsed with PBS and incubated with serum-free Opti-MEM at 37 C for 24 h.IDOL, LDLR expression levels, and LDL uptake function were detected using western blot, immunofluorescence, flow cytometry analysis, and DiI-LDL staining, respectively.Quantification of LDLR protein levels was performed using western blot analysis.
For lentivirus shRNA transduction, about 5 Â 10 5 HepG2, Hepa1-6, or AML12 cells were seeded per well into a 6-well plate and transduced with 5 Â 10 6 TU lentivirus in the presence of 4 mg/mL polybrene (catalog #H9268; Sigma-Aldrich).After transduction for 24 h, the virus-containing medium was discarded and replaced with fresh growth medium.The cells were then cultured for additional 24 h and then rinsed with PBS and incubated with serum-free Opti-MEM at 37 C for 24 h.IDOL, LDLR expression levels, and LDL uptake function were detected using western blot, immunofluorescence, flow cytometry analysis, and DiI-LDL staining, respectively.Quantification of LDLR protein levels was performed using western blot analysis.

Western blot
Western blot was performed to measure the protein expression levels in HepG2 and Hepa1-6 cells or tissues, as previously described. 50estern blot analysis was performed to measure the protein expression levels in cells.HepG2 and Hepa1-6 cells were washed three times in ice-cold PBS and then lysed in ice-cold RIPA lysis buffer (50 mM Tris, 150 mM Na Cl, 1 mM ethylenediaminetetraacetic acid [EDTA], 1% Triton X-100, 0.5% sodium deoxycholate, and 0.1% SDS, [pH 7.4]) containing 1 mM PMSF.After centrifugation at 12,000 Â g for 15 min at 4 C, the supernatant was harvested.Total soluble protein was measured using BCA protein assay kit (Biouniquer, Beijing, China).An equal amount of protein from each sample was loaded in each lane of a 10% SDS-PAGE gel and separated.The separated proteins were electrophoretically transferred onto a 0.22 mm PVDF membrane (Merck Millipore, Darmstadt, Germany).The membrane was blocked with a solution of 0.1% (vol/vol) TBS-Tween 20 (TBST) containing 5% (w/v) nonfat milk for 1 h at room temperature, then incubated with corresponding primary antibodies against IDOL (1:1,000), LDLR (1:2,000), GAPDH (1:1,000), b-actin (1:1,000), and LRP1 (1:1,000) at 4 C overnight, followed by incubation with appropriate HRP-conjugated secondary antibodies (1:5,000, Fcmacs Biological) for 1-2 h at room temperature.After washing three times with TBST for 5 min, protein bands were developed using ECL (Thermo Fisher Scientific) and quantified using ImageJ software (National Institutes of Health, Bethesda, MD).

qRT-PCR analysis
Total RNA was extracted from HepG2 and AML12 cells and liver tissues by using RNAiso Plus reagent according to the manufacturer's instructions and quantified by measuring A260nm using Thermo NanoDrop 2000 (Thermo Fisher Scientific).For mRNA quantification, cDNA was generated using the BeyoRT II First Strand cDNA Synthesis Kit with gDNA Eraser.The qRT-PCR analyses were performed using BeyoFast SYBR Green qPCR Mix on a MX3000PTM qRT-PCR instrument (Agilent Technologies, Santa Clara, CA).To measure mRNA expression levels, the data were normalized to the housekeeping gene b-actin or GAPDH.Primers for each gene are listed in Table S5.The 2 ÀDDCt method 51 was used to calculate relative gene expression levels.

LDL uptake assay
The assay was performed as described previously 52 with slight modification.Briefly, HepG2 and Hepa1-6 cells were maintained in MEM and DMEM supplemented with 10% FBS, respectively.The cells were seeded in 96-well black plates at a density of 1 Â 10 4 cells per well and grown to 70%-80% confluence.Following transfection or transduction for 48 h, HepG2 and Hepa1-6 cells were incubated with serum-free Opti-MEM for another 24 h, and 20 mg/mL DiI-LDL was added and incubated in the dark for additional 4 h.Cells incubated with Opti-MEM without DiI-LDL and cells incubated with Opti-MEM in the presence of 20 mg/mL DiI-LDL were used as negative control and control for normalization, respectively.After rinsing 3 times with PBS, LDL uptake was measured using a fluorescence plate reader (Varioskan Fflash; Thermo Fisher Scientific) at an excitation wavelength of 520 nm and an emission wavelength of 580 nm.

Flow cytometric analysis
Flow cytometric analysis of the LDLR expressed on the cell surface was conducted as previously described 53 with slight modification.Following transfection or transduction for 48 h, the medium was changed to serum-free Opti-MEM and incubated at 37 C for 24 h, and then HepG2, Hepa1-6, and AML12 cells were digested with trypsin, detached by scraping, washed with PBS, and collected in a 1.5 mL tube, then fixed in 200 mL 4% (w/v) paraformaldehyde in PBS for 5 min at room temperature.Cells were incubated with 200 mL of 0.1% Tween in PBS (PBS-T) and blocked with 10% goat serum in 0.3 M glycine in PBS for 30 min, then incubated with rabbit anti-LDLR monoclonal antibody (catalog #ab52818; 1:100) for 30 min at room temperature, followed by incubation with Alexa Fluor 488-conjugated goat anti-rabbit IgG (catalog #D110061; 1:200) or Alexa Fluor 555-conjugated goat anti-rabbit IgG (catalog #D110070; 1:200) for 30 min at room temperature.After washing, detection for Alexa Fluor 488 was performed directly on a Guava Easy Cyte flow cytometer (Merck Millipore) at an excitation wavelength of 488 nm and an emission wavelength of 525 nm, while Alexa Fluor 555 was detected at an excitation wavelength of 555 nm and an emission wavelength of 580 nm.The levels of LDLR on the cell surface were analyzed using FlowJo version 7.6 with 10,000 cells.

Immunofluorescence assay
Detection of IDOL and LDLR in HepG2, Hepa1-6, and AML12 cells and liver tissues by immunofluorescence was performed as previously described 52,53 with minor modification.Briefly, after treatment, HepG2 and Hepa1-6 cells were rinsed 3 times with PBS for 5 min and fixed in 4% (w/v) paraformaldehyde in PBS for 30 min.Liver tissues were fixed in 4% (w/v) paraformaldehyde in PBS at 4 C for 48 h, embedded in paraffin, and sliced at 4 mm thickness.After deparaffinization and hydration, tissue sections were pretreated by heating for 20 min in sodium citrate solution (0.01 M, pH 6.0) in a 95 C water bath for antigen retrieval.Thereafter, the cells or tissue sections were blocked with 10% (v/v) goat serum in PBS-T for 1 h and incu-bated with rabbit anti-IDOL antibody (1:100; catalog #D126668) or rabbit anti-LDLR antibody (1:100; catalog #ab52818) overnight at 4 C, followed by incubation with Alexa Fluor 488-conjugated goat anti-rabbit IgG (1:200; catalog #D110061) or Alexa Fluor 555-conjugated goat anti-rabbit IgG (1:200; catalog #D110070) for 1 h at room temperature and counter-stained with DAPI (Key GEN Bio TECH, Nanjing, China) to show cell nucleus.Images were acquired using a Zeiss LSM700 confocal microscope (Zeiss, Oberkochen, Germany).

Homology modeling and molecular docking
The 3D structure of CS8 peptide-incorporated VSV-G (GenBank: OQ561788) was built on the basis of the wild-type VSV-G crystal structure (PDB: 5I2S) using I-TASSER (http://zhanglab.ccmb.med.5][56] The modeled structure of CS8 peptideincorporated VSV-G was docked with the crystal structure of LRP1 CR.7 (PDB: 1J8E) using HawkDock web server (http://cadd.zju.edu.cn/hawkdock/), 57,58 and the binding free energy (DG bind ) of the ten top-ranked binding poses was calculated in HawkDock web server using the MM/GBSA method.Then, the docked poses with the lowest binding free energies were selected as the best conformations, and the key interacting residues in those docked complexes were analyzed and mapped using Molecular Operating Environment for Windows.

Mice
All animal experiments were approved by the ethics committee of China Pharmaceutical University (#201601179, October 19, 2016) and conformed to the Guide for the Care and Use of Laboratory Animals published by the National Institutes of Health.Six-week-old male C57BL/6 mice were obtained from Qinglongshan Experimental Animal Breeding Farm (SCXK [Su] 2020-0001, Nanjing, China) and maintained on a 12 h light/dark cycle at 25 C.

In vivo distribution of liver-targeted lentiviral IDOL-shRNA
To enable assessment of biodistribution, liver-targeted CS8-LV-shIDOL was injected into the tail veins of C57/BL6 mice at a dose of 1 Â 10 8 TU per mouse.Control mice were injected with an equivalent dose of wild-type lentivirus LV-shIDOL or saline.One week after injection, mice were sacrificed, and the fluorescence intensity in five different organs (heart, liver, kidney, lung, and spleen) was detected using IVIS Kinetic Bioluminescence imager (PerkinElmer).Subsequently, tissues were fixed in 4% (w/v) paraformaldehyde and embedded in paraffin for histological examination.Tissue sections were prepared, and ZsGreen expression was analyzed immediately using a fluorescence microscope (AX10).
In vivo toxicity evaluation C57BL/6 mice were intravenously administered with the equivalent dose (1 Â 10 8 TU per mouse) of either LV-shIDOL or CS8-LV-shIDOL.Control mice were injected with saline.One week after injection, mice were euthanized, and blood and major organs were harvested.Assay kits (Beyotime) were used to measure serum levels of ALT, AST, ALP, plasma urea, and CREA.Organs were fixed in paraformaldehyde, sectioned, and stained with H&E for morphological analysis.
In vivo IDOL-shRNA treatment After acclimation for one week, mice were randomly divided into five groups (n = 8 each): (1) normal group: normal mice fed on a low fat diet (catalog #TP28602; TrophicDiet, Nantong, China) treated with saline; (2) model group: P407-induced hyperlipidemic mice fed an atherogenic diet (catalog #TP28600; TrophicDiet) treated with saline; (3) negative control shRNA group: P407-induced hyperlipidemic mice fed an atherogenic diet treated with LV-shNC; (4) LV-shIDOL group: P407-induced hyperlipidemic mice fed an atherogenic diet treated with LV-shIDOL; and (5) CS8-LV-shIDOL group: P407-induced hyperlipidemic mice fed an atherogenic diet treated with CS8-LV-shIDOL.In order to induce atherosclerotic lesions, all animals except the normal group were fed an atherogenic diet (containing 20% fat, 1.5% cholesterol, and 0.5% cholic acid; catalog #TP28600; TrophicDiet) were simultaneously treated every third day with an intraperitoneal injection of P407 (0.5 g kg À1 ) for a time period of 16 weeks.For negative control shRNA, LV-shIDOL, and CS8-LV-shIDOL groups, mice were received corresponding LVs (1 Â 10 8 TU per mouse) via tail-vein injection every eight weeks, while mice in the normal control group and model group were injected with saline (vehicle).Twenty-four hours after the final injection of P407, all mice were fasted for 8 h and then euthanized for collection of blood sample, aorta, lung, heart, liver, spleen, and kidney.Liver tissues were dissected and further analyzed using western blot.Heart, liver, spleen, lung, and kidney were fixed in 4% (w/v) paraformaldehyde and embedded in paraffin for histological examination.In addition, liver tissues were also embedded in optimal cutting temperature (OCT) freezing medium and stained with H&E for morphological analysis and oil red O to detect lipid accumulation.To determine lesions in aortic root, frozen sections of aortic root were prepared and then stained with oil red O solution.

Statistical analysis
All values are presented as mean ± SEM and were analyzed using Prism 6.0 software (GraphPad Software, La Jolla, CA).Normality of data was determined using the Kolmogorov-Smirnov test.Normally distributed variables between 2 independent groups were compared using unpaired Student's t tests, whereas non-normally distributed variables were analyzed using the Mann-Whitney U test.A two-tailed p value of <0.05 was considered to indicate statistical significance.

Figure 1 .
Figure1.Schematic diagram illustrating the mechanism of liver-targeted lentiviral vector CS8-LV-shIDOL treating atherosclerosis in C57/BL6 mice (A) P407-induced hyperlipidemic mice fed an atherogenic diet (AD) treated with saline.IDOL binds to and promotes ubiquitination of the intracellular tail of the LDLR, resulting in lysosomal degradation of the receptor.The aortas of P407-induced atherogenic mice show pathological lesions that are thickened.(B) P407-induced hyperlipidemic mice fed an AD treated with wild-type lentiviral vector LV-shIDOL.Lentiviral vector mediated shRNA knockdown of IDOL up-regulates LDLR and lowers circulating LDL levels.The lesions of the aorta in LV-shIDOL-treated atherogenic mice are milder than those in untreated atherogenic mice.(C) P407-induced hyperlipidemic mice fed an AD treated with liver-targeted CS8-LV-shIDOL.Compared with LV-shIDOL, liver-targeted CS8-LV-shIDOL has higher potency in inhibiting hepatic IDOL in vivo, resulting in higher LDLR levels and lower serum LDL levels.The lesions of the aorta in CS8-LV-shIDOL-treated atherogenic mice are milder than those in LV-shIDOL-treated atherogenic mice.

Figure 2 .
Figure 2. Analysis on viable insertion sites on VSV-G for displaying hepatocyte-targeting peptides (A) Elements of secondary structure are indicated in the upper of the amino acid sequence of VSV-G.The helices are shown in the pink rectangle, the strands in the yellow rectangle, and the coiled coils as a gray line.The secondary structure of VSV-G was analyzed using PSIPRED server (http://bioinf.cs.ucl.ac.uk/psipred/).(B) Hydrophobicity plot corresponding to sequences spanning the VSV-G protein of the crystal structure (PDB: 5I2S).The hydrophobicity of VSV-G was analyzed using ProtScale (https://web.expasy.org/protscale/).(C) Sequence alignment of envelope proteins in various vesicular stomatitis virus strains.Conserved residues between different vesiculovirus envelope proteins are highlighted in blue.IND, vesicular stomatitis Indiana virus (GenBank: AAA48370.1);NJE, vesicular stomatitis New Jersey virus (NCBI: YP_009047084.1);PIR, Piry virus (Swiss-Prot: Q85213.1);COC, Cocal virus (GenBank: ACB47437.1);CHA, Chandipura virus (NCBI: YP_007641380.1).Alignment of multiple sequences was conducted using DNAMAN version 6 software.

Figure 3 .
Figure 3. Liver-targeting peptide insertion sites on VSV-G and representative titers of lentiviral vectors pseudotyped with VSV-G variants (A) Scheme of the lentiviral retargeting strategy.Liver-targeting peptide was incorporated into the VSV-G, and the engineered envelope was used to pseudotype lentivirus.(B) Insertion sites of liver-targeting peptides were marked on the amino acid sequence of VSV-G.The domains of VSV-G are displayed in colors: red, lateral domain; blue, trimerization domain; orange, PH domain; green, fusion domain; and magenta, C-terminal part.(C) Crystal structure of the VSV-G protein.Insertion sites of liver-targeting peptides were indicated.The color code for different domains is as the same as in (B).The VSV-G protein crystal structure (PDB: 5I2S) was used and edited in Molecular Operating Environment for Windows.(D and E) VSV-G constructs with ligand insertion sites and representative titers of lentiviral vectors pseudotyped with different VSV-G variants.Error bars represent the standard error of the linear regression used to determine titers.(F and G) Lentiviral plasmids encoding for VSV-G variants (F) and matrix proteins (G).The VSV-G full-length plasmid pMD2.g was used for the construction of variants with peptide insertions.(H and I) Lentiviral shRNA expression plasmids LV-shNC and LV-shIDOL with ZsGreen tag were constructed for lentivirus production with helper plasmid pMD2.g(F) and psPAX2 (G).The shRNA expression vectors were constructed using the U6 pol III promoter on the basis of a 19 nt siRNA sequence (scramble and siIDOL-622) (TableS3).All shRNA sequences have identical loop sequences.

Figure 4 .
Figure 4. Functional analysis on the liver-targeted lentiviral shRNAs against IDOL in vitro (A-D) HepG2 cells were transfected with LV-shNC (negative control), LV-shIDOL (lentiviral IDOL-shRNA with wild-type VSV-G), and lentiviral IDOL-shRNAs with different VSV-G variants, respectively.After 48 h, the levels of IDOL mRNA in HepG2 cells were quantified using qRT-PCR (A).After 72 h, the levels of IDOL and LDLR protein were measured using western blot (B and C), DiI-LDL uptake levels were measured using a multimode reader (D), then normalized to mock control.(E-G) Effects of CS8-LV-shIDOL on IDOL and LDLR expression in HepG2 cells.After transfection with LV-shNC, LV-shIDOL, and CS8-LV-shIDOL for 72h, IDOL protein levels were visualized using immunofluorescence assay (F).The cell surface LDLR protein levels were determined using flow cytometry (E) and evaluated using immunofluorescence assay (G).*p < 0.05, **p < 0.01, and ***p < 0.001 vs. mock control; ### p < 0.001 and #### p < 0.0001 vs. LV-shNC control (unpaired Student's t test).Results are given as mean ± SEM of three independent experiments.

Figure 6 .
Figure 6.Therapeutic effects of CS8-LV-shIDOL in mice (A) Schematic diagram of the experimental procedure to develop an atherosclerosis model in C57/BL6 mice.Solid arrows indicate the time points for injection of lentiviral vectors.(B and C) P407-induced hyperlipidemic mice fed an atherogenic diet were injected with LV-shNC, LV-shIDOL, and CS8-LV-shIDOL (1 Â 10 8 TU per mouse at intervals of 8 weeks) for 16 weeks.At the end of experiment, serum, aorta, and liver samples were collected.CS8-LV-shIDOL was more effective in lowering serum LDL-C and TC compared with LV-shIDOL.Results are expressed as mean ± SEM (n = 8 per group).(D) Atherosclerotic plaques in aortic root were measured using oil red O staining.(E-H)The hepatic IDOL and LDLR levels in LV-shIDOL and CS8-LV-shIDOL-treated groups were detected using qRT-PCR (E), western blot (F and G) and immunofluorescence assay (H).*p < 0.05, **p < 0.01, and ****p < 0.0001 vs. LV-shNC group.## p < 0.01, ### p < 0.001, and #### p < 0.0001 vs. saline group.D p < 0.05 and DDD p < 0.001 vs. LV-shIDOL group.Scale bar, 50 mm.Data are representative of 3 independent experiments with similar results.

Figure 8 .
Figure 8. Binding of CS8 peptide-incorporated VSV-G to LRP mediates the targeted entry of CS8-LV-shIDOL into hepatocytes (A and B) HepG2 cells were treated with anti-LDLR and/or anti-LRP1 monoclonal antibody for 1 h before the addition of lentiviral vectors.After transfection with LV-shNC, LV-shIDOL, and CS8-LV-shIDOL for 1 h, cells were washed and incubated at 37 C for 16 h.Fluorescence was detected using microscopy (A) and flow cytometry (B and C).***p < 0.001; ### p < 0.001 and #### p < 0.0001; and DDD p < 0.001 and DDDD p < 0.0001 (unpaired Student's t test).Results are given as mean ± SEM of three independent experiments.(D) The 3D structure of VSV-G protein (PDB: 5I2S).(E) Modeling of CS8 peptide-incorporated VSV-G (CS8-VSV-G) was constructed using I-TASSER, and CS8 peptide is shown in black.All the structures were edited using Molecular Operating Environment for Windows.(F) In silico docking of CS8 peptide-incorporated VSV-G with LRP1 CR.7 using HawkDock.Key residues involved in the interactions are represented as stick models and indicated in black font for CS8-VSV-G and in red font for LRP1 CR.7.The hydrogen bonds are shown as black dotted lines.