Human CD64-targeted non-viral siRNA delivery system for blood monocyte gene modulation

A subset of phagocytes including inflammatory monocytes in blood migrate and give rise to macrophages in inflammatory tissues which generated the idea that blood monocytes are the therapeutic targets for drug delivery. Fc gamma receptor I (CD64) is a membrane receptor for the Fc region of immunoglobulin G, primarily expressed on monocyte-lineage, and H22 a monoclonal antibody for human CD64 had shown rapid blood monocyte binding and occupation in clinical studies. Small interfering RNA-mediated gene silencing as a therapeutic has been proposed and is a promising strategy in terms of its “knock-down” ability on the target gene prior to translation. However, its instability and off-targeting effect must be overcome for success in clinical studies. In this study, we developed a non-viral delivery system composed of oligo-nona-arginine (9R) and anti-human CD64 single chain antibodies (H22) for human monocyte-specific siRNA delivery. A targeted and efficient siRNA delivery mediated by anti-CD64 scFv-9R was observed in CD64 positive human leukemia cells, THP-1. With primary human blood cells, anti-CD64 scFv-9R mediated gene silencing was quantitatively confirmed representing blood monocyte selective gene delivery. These results demonstrate the potential of anti-CD64 scFv-9R mediated siRNA delivery for the treatment of human inflammatory diseases via blood monocytes gene delivery.

siRNA delivery. Recently, Her2-targeted scFv-protamine fusion protein-mediated siRNA delivery suppressed growth and metastasis of human primary breast cancer in mice 16 .
The Fc gamma receptor family encompasses membrane glycoprotein receptor for the Fc region of immunoglobulin G and distributed on diverse type of cells, from immune cells to endothelial, epithelial cells. Human Fc gamma receptor I (CD64) expression is primarily on monocytes and cells of macrophage lineages 17,18 . The H22, a monoclonal antibody targeting human CD64 have been applied for immunotoxins ablating CD64-positive monocytic leukemia 19 , and phase I clinical studies showed rapid monocyte binding of H22 in circulation 20,21 .
In this study, we investigate CD64-targeted siRNA delivery for human blood monocytes gene silencing. We conjugated nucleic acid binding nona-arginine peptide with an anti-human CD64 single chain antibody (H22). Anti-CD64 scFv-9R showed target-receptor mediated siRNA delivery and gene silencing in a human monocytic cell line (THP-1). In human peripheral blood cells (PBMCs), anti-CD64 scFv-9R showed monocytes-selective uptake and gene silencing without significant toxicity ex vivo. With a human monocyte engrafted mouse, scFv-9R/siRNA complex demonstrated human monocyte-targeted delivery in vivo, and collectively demonstrated the potential of CD64 targeted scFv-9R for blood monocyte siRNA delivery in human inflammatory disease.

Results
Schematic illustration of CD64-targeted siRNA delivery system. It has been reported that the monoclonal antibody for human CD64, H22 binds to different epitopes from IgG binding, representing target binding in the presence of serum IgG 22 , and H22 has shown rapid blood monocyte binding in phase I clinical trials 20,21 . Additionally toxins conjugated with H22 were efficiently internalized into CD64 positive leukemia cells and induced cancer cell apoptosis 23 . Based on these studies, we developed a CD64-targeted non-viral carrier consisting of a single chain antibody of H22 (anti-CD64 scFv) and 9-mer oligo-D-arginine (9R) as shown in Fig. 1a, having triple benefits of target specificity, gene stabilization and cell penetrating efficiency. For construction of the anti-CD64 scFv-expressing vector, the gene sequence encoding anti-CD64 scFv was cloned into pET21a. Anti-CD64 scFv contains a signal peptide (pelB) for periplasmic localization, which facilitates protein folding, linker sequence (GlySer 4 ) 3 and cysteine to create a disulfide bond with the terminal cysteine of oligo-nona-D-arginine (Fig. 1b). Oligo-arginine has been widely used for protein 24,25 and gene delivery 10,26,27 . In particular, D-form oligo-arginine has been reported to be more effective in improving gene transfection efficiency and in vivo stability than L-form. Reducible polymers based on the D-form oligo-arginine have shown significantly improved transfection efficiency compared with PEI and lipofectamine 28 . As described in reference 14 , oligo-D-arginine with an (Npys) cysteine residue was incubated with anti-CD64 scFv in a low pH phosphate buffer for conjugation and dialyzed to remove unconjugated oligo-arginine (MWCO; 12-14,000 Da) (Fig. 1c).
Characterization and functionality of purified anti-CD64 scFv and anti-CD64 scFv-9R conjugate. The anti-CD64 scFv expressed in a bacterial system (BL21 (DE3)) and purified using affinity chromatography provided a single band around 26 kDa in SDS-PAGE. The western blot result stained with an anti-histidine tag antibody showed the same band location as SDS-PAGE (Fig. 2a). After oligo arginine conjugation, anti-CD64 scFv-9R conjugate showed a single band at the expected position in SDS-PAGE and The conjugation efficiency was ~77% as quantified using a free thiol quantification kit (Fig. S1b). The functionality of anti-CD64 scFv was tested by measuring competitive binding with a commercial human CD64 antibody in CD64 positive THP-1 and CD64 negative MDA-MB-231 cells. As shown in Fig. 2b, THP-1 cells pre-incubated with anti-CD64 scFv (black thick line) demonstrated reduced commercial CD64 antibody (FITC) binding, representing that anti-CD64 scFv competes with commercial CD64 antibody. On the other hand, no competitive binding was observed between anti-CD64 scFv and other antibodies against CD14 and CCR2 in THP-1 cells. CD64 negative MDA-MB-231 cells did not show binding affinities for anti-CD64 scFv and commercial CD64 antibody. In the siRNA retardation study, anti-CD64 scFv-9R conjugate retarded the siRNA in the N/P ratio above 4 in contrast to anti-CD64 scFv (Fig. 2c). The size of anti-CD64 scFv-9R/siRNA was analyzed with transmission electron microscopy ( Fig. 2d) and DLS nano-size distribution (Fig. 2e).

Toxicity of anti-CD64 scFv-9R/siRNA complex on human PBMC. To test for cytotoxicity, anti-CD64
scFv-9R/siRNA complex treated PBMCs were stained with Annexin V (PE) and analyzed. In comparison with non-treated groups, and only siRNA treated groups, the anti-CD64 scFv-9R/siRNA complex treated group did not exhibit a significant change in Annexin V staining, indicating that anti-CD64 scFv-9R doesn't have significant toxicity. As a positive control, heat-shocked PBMCs showed substantial change in Annexin V staining (Fig. 5a). Additional cell viability test was performed with MTT assay. As shown in Fig. 5b, sorted CD14 (+ ) monocytes and lymphocytes were treated with scFv-9R/siRNA complex in the same concentration as the knock-down experiments. Compared with other groups, scFv-9R/siRNA complex showed no significant toxicity. To examine immune response, inflammatory gene mRNA levels were measured 30,31 . The naked siRNA treated groups showed slight TNF-alpha down regulation and IFN-gamma, IFN-beta, IL-6 up regulation, thought to be the TLR signaling mediated response as previously reported 32 . The scFv-9R complexed siRNAs compromised the siRNA mediated inflammatory gene induction and is due to the sequestering of the scFv-9R complexed siRNAs inside in higher N/P ratio (Fig. 5c). These results indicate that the anti-CD64 scFv-9R/siRNA complex has mild toxicity in the indicated amount, and masking the siRNA's natural immune response.

Discussion
Many papers proved that subsets of monocytes are the source of disease accumulating macrophages 2-5 , suggesting that blood circulating monocytes are appropriate targets for disease-targeted drug delivery. Phagocyte-targeted drug delivery, especially for monocyte and macrophage lineage have demonstrated its therapeutic effect in animal models 7,30 , and potential human clinical application. In our previous work, polymerization of D-form oligo-arginine with terminal cysteine residues (rPOA) improved its gene transfection ability for both siRNA & plasmid DNA. Despite high transfection efficiency, its non-target specificity limits local in vivo applications 35,36 . With the theory of human blood monocytes as drug delivery targets, we developed a non-viral siRNA delivery system based on a single chain antibody and oligo-arginine. Target receptor human CD64, a monocyte lineage-specific expression improves its target specificity. Some papers reported CD64 expression is elevated on circulating monocyte from patient with inflammatory disease and its expression is IFN-gamma signaling mediated 37,38 resulting in inducible inflammatory conditions, and also IFN-gamma is known for its role in M1 macrophage polarization. For the goal of targeting, we employed H22 (scFv), anti-human CD64 antibody as a targeting moiety. H22 is different from IgG for Fc receptor binding in terms of binding epitope and Fc gamma receptor I (CD64) specificity rather than other type of Fc receptors (CD32, CD16) 22 , and this has been applied for the development of immunotoxins targeting CD64-positive human myeloid leukemia in previous studies 19,23,39,40 . In this study, anti-CD64 scFv-9R/siRNA complexes were characterized for their target receptor-mediated cell binding and uptake, and lipofectamine comparable gene silencing efficacy in THP-1. Experiments with human PBMCs demonstrates blood monocyte specific uptake and gene silencing ex vivo and human monocyte-targeting effect in vivo. However we found mild silencing effect on protein levels which may be attributed to the non-proliferative or quiescent state of monocyte ex vivo and limited endosome escape of oligo-arginine. Generally for non-proliferative primary cell gene delivery, investigators add growth factors inducing proliferation such as IL-13, PHA in case of human T cells 14,41 . Meanwhile, siRNA mediated gene silencing in mouse blood monocytes was not significant compared with other phagocytic cells 7 . Furthermore, CpG ligand mediated siRNA delivery did not reduce the target gene in the human monocyte even with a high uptake 42 , suggesting that monocytes are cells relatively harder to transfect than other phagocytes and have their own endosomal degradation and clearance mechanism. In terms of clinical application, applying targeting moiety for human receptor improves its potency as a drug, jumping the cross-species reactivity problem, but it doesn't exclude the need for animal studies such as target specificity and efficacy in vivo. Humanized mice is an existing animal model for human immune cell study in vivo, and genetic engineering technologies have developed mouse strains more suitable for human immune system simulation. A recent paper reported tumor accumulation of human macrophages in a cancer cell grafted to TG mice as the same as in a human patient 43 . However those generations do not reflect whole human inflammatory disease pathology such as disease accumulation of phagocytes [43][44][45] in inflammation. Collectively, our data demonstrates the CD64-targeted siRNA delivery as a potent system for human blood monocyte gene delivery.

Materials and Methods
Vector construction. An 800 base pair sequence-encoding anti-human CD64 scFv with a C-terminal cysteine residue and a 6x His-tag was synthesized and cloned (Incorporation Bioneer) between xbaI and NotI restriction sites of the pET21a vector (Novagen, Madison, WI) for bacterial expression.
SiRNAs. Luciferase siRNA conjugated with FITC was synthesized and purchased from ST Pharm. Co. Ltd.
SiRNAs for cyclophilin B (siCypB), CD45 (siCD45), and luciferase (siLuci) were purchased from Dharmacon.  Retardation assay. The siLuci (40 pmol) was mixed with anti-CD64 scFv or anti-CD64 scFv-9R in diverse N/P ratios and incubated for 15-20 min at room temperature. Samples were loaded into a 2% agarose gel and electrophoresed, and detection was performed using a Kodak image station.
Anti-CD64 scFv-9R mediated cell binding of siRNA. For THP-1 binding, anti-CD64 scFv-9R and anti-CD64 scFv were mixed with 75 pmol siLuci (FITC) in an N/P ratio of 4 and incubated for 20 min at room temperature. The 2 × 10 5 of THP-1 cells were incubated with the anti-CD64 scFv-9R/siLuci (FITC) complex in either the presence or absence of free anti-CD64 scFv protein on ice for 30 min. The cells were washed twice with PBS and analyzed using a BD FACS Calibur.
In vivo human monocyte targeting effect. The 5-7 weeks aged NOD-scid (Jackson laboratory) mice were engrafted with FACS sorted human CD14 (+ ) monocytes (5-7.5 × 10 5 cells per mouse) by intravenous injection. After 1-2 hours, anti-CD64 scFv-9R/siRNA complexes (siRNA (FITC) 500 pmol, N/P ratio 4) were injected intravenously and incubated for additional 1-2 hours in dark. Total blood was harvested from mouse heart and lysed with red blood cell lysis buffer (Sigma Aldrich) and stained with mouse, human CD45 antibody (PerCP-Cy5.5) for flowcytometry analysis. The blood monocytes were indicated with forward scattering (FSC) and CD45 staining. All animal experimental procedures with NOD-scid mice were reviewed and approved by the Institutional Animal Care and Use Committee (IACUC) of the Hanyang University (HY-IACUC-16-0126) and was performed in accordance with the relevant guidelines. The mice were housed under specific pathogen free conditions. Blood profile. The NOD-scid (Jackson laboratory) mice were engrafted with FACS sorted human CD14 (+ ) monocytes (5-7.5 × 10 5 cells per mouse) and injected with scFv-9R/siRNA complex (siRNA (Cy5.5) 500 pmol, N/P ratio 4). At each time points, 30-50 ul of blood was harvested from orbital vein and incubated 30 min for serum isolation. Serum was 3 times diluted in PBS and detected for Cy5.5 (excitation: 695 nm, emission: 720 nm).

Statistical analysis.
Data are presented as the mean ± standard deviations. Statistical analyses were performed using the Student's t-test, One-way ANOVA and Tukey posttest and the GraphPad Prism 5 Project software.
Scientific RepoRts | 7:42171 | DOI: 10.1038/srep42171 Flowcytometry analysis. All treated replicates were assembled and prepared in one tube and performed with BD FACS Calibur (BD biosciences, San Diego, CA) and analyzed with CellQuest Software (BD Pharmingen, San Jose, CA).