Targeting lentiviral vector to specific cell types through surface displayed single chain antibody and fusogenic molecule
Viral delivery remains one of the most commonly used techniques today in the field of gene therapy. However, one of the remaining hurdles is the off-targeting effect of viral delivery. To overcome this obstacle, we recently developed a method to incorporate an antibody and a fusogenic molecule (FM) as two distinct molecules into the lentiviral surface. In this report, we expand this strategy to utilize a single chain antibody (SCAb) for targeted transduction.
Two versions of the SCAb were generated to pair with our various engineered FMs by linking the heavy chain and the light chain variable domains of the anti-CD20 antibody (αCD20) via a GS linker and fusing them to the hinge-CH2-CH3 region of human IgG. The resulting protein was fused to either a HLA-A2 transmembrane domain or a VSVG transmembrane domain for anchoring purpose. Lentiviral vectors generated with either version of the SCAb and a selected FM were then characterized for binding and fusion activities in CD20-expressing cells.
Certain combinations of the SCAb with various FMs could result in an increase in viral transduction. This two-molecule lentiviral vector system design allows for parallel optimization of the SCAb and FMs to improve targeted gene delivery.
KeywordsTransduction Efficiency Isotype Control Antibody Single Chain Antibody Fusion Activity Binding Avidity
Gene therapy is the introduction of a functional gene into a dysfunctional cell for a therapeutic benefit. To date, viral vectors remain the most commonly used gene delivery vehicles due to their high transduction efficiencies [1, 2]. In particular, lentiviral vectors represent one of the most effective gene delivery vehicles as they allow for stable long-term transgene expression in both dividing and non-dividing cells. In order to expand the targeted specificity of viral vectors beyond their natural tropism, numerous studies have been focused on pseudotyping lentiviral vectors with envelope glycoproteins derived from other viruses, such as the glycoprotein from vesicular stomatitis virus (VSVG) [3, 4]. However, since the VSVG is thought to recognize a ubiquitous membrane phospholipids instead of a unique cellular receptor, pseudotyping generates vectors with broad specificities [5, 6]. To mitigate this off-target effect, previous attempts have been devoted to engineer the viral glycoprotein to recognize a specific cellular target by insertion of ligands, peptides, or antibodies [7, 8, 9, 10, 11, 12, 13, 14, 15, 16]. Another approach involves bridging the viruses and the targeted cell with ligand proteins or antibodies [17, 18, 19, 20]. However, these modifications to the surface glycoprotein appear to perturb the natural fusion function of the glycoprotein, resulting in a reduction of transduction efficiency.
Recently, our lab has developed a strategy to target lentiviral vectors to specific cell types by incorporating a surface antibody specific to CD20 antigen and a fusogenic molecule (FM) as two distinct molecules . Kielian and co-workers reported several versions of the Sindbis virus glycoprotein that were less dependent on cholesterol for transduction . We applied these mutations (E1 226) to the binding defective Sindbis glycoprotein and observed that they were able to enhance transduction efficiency when paired with an anti-CD20 antibody (αCD20) . In this study, we report our attempt to utilize a single chain antibody (SCAb) to pair with a FM for targeting lentiviral vectors. Our SCAb is composed of variable domains of the heavy and light chains of αCD20, linked by a GS linker and fused to a hinge-CH2-CH3 region of human IgG. To anchor the SCAb onto the viral surface, we conjugated the SCAb with either the HLA-A2 transmembrane domain (SC2H7-A2) or the VSVG transmembrane domain (SC2H7-GS). We demonstrated that the lentiviral vector enveloped with either of these antibody configurations could achieve targeted transduction to CD20-expressing cells. We also compared the targeted transduction efficiency and the binding avidity of both versions of the SCAb and investigate the molecular roles of the displayed proteins in mediating lentiviral transduction.
Construction of SCAb for targeting
Production of lentiviral vectors
Incorporation of SCAb and FM onto lentiviral vectors
Targeted transduction of lentiviral vectors
We conducted transduction experiments to evaluate the efficiency of lentiviral vectors bearing both SCAb and FM to transduce the CD20-expressing cell line. The lentiviral vector bearing VSVG was used as a positive control, whereas the lentiviral vector co-displaying AB and FM was included as a negative control. Cell lines were transduced by indicated lentiviral vectors and analyzed by FACS five days post-transduction. The GFP expression level was detected to quantify the specificity and efficiency.
Among the various lentiviral vectors bearing the same SCAb but different FMs, different transduction efficiencies were observed. The lentiviral vector displaying SC2H7-A2 and SINmu exhibited 15% transduction efficiency. However, lentiviral vectors displaying other FMs (SGN, SGM, and AGM) resulted in specific transductions of 25% to 30%. A similar trend was observed in another independent study where the SCAb with VSVG transmembrane domain was used as the targeting antibody (SC2H7-GS) (Fig. 4A). In this case, the lentiviral vector bearing SINmu and SC2H7-GS was able to specifically transduce about 14% of the 293T/CD20 cells, whereas the specific transduction efficiency was increased to 25% when other FMs (SGN, SGM and AGM) were used in combination with the SC2H7-GS.
Assays for studying the entry mechanism
The second critical step of transduction pathway involved the pH-dependent fusion event leading to the release of the viral core. To verify the pH requirement, we incubated either FUGW/SC2H7-A2/SGN or FUGW/GP160 with 293T/CD20 or Ghost-CCR5 cells in the increased presence of bafilomycin, which can raise the pH of the endosomal compartment. We observed a dramatic decrease in transduction efficiency (FUGW/SC2H7-A2/SGN) in response to increasing amount of bafilomycin (Fig. 5A). In a control experiment where a pH-independent virus (FUGW/GP160) was used, an increase in transduction efficiency was observed, which was consistent with previously published data [29, 30]. Thus, the pH in the endosomal compartment is critical for viral membrane fusion.
pH dependency study on the FMs
Binding avidity of lentiviral vectors to target cells
The purpose of this study is to incorporate both membrane-bound SCAb and FM on the lentiviral surface to achieve targeted transduction to specific cell types. Previously, we reported a strategy of separating the binding and fusion functions of viral glycoprotein for cell specific targeting . By pairing the αCD20 with a more fusion active FM, the resulted lentiviral vectors showed enhanced transduction . In this study, we extended the targeting strategy to utilize a membrane-bound SCAb with the engineered FMs. Insertion of SCAb into the viral glycoprotein has shown to be able to redirect vector particles to specific cellular target [8, 9, 13]. However, these modifications usually resulted in reduced transduction efficiency. Our strategy of separating binding and fusion functions allows us to engineer a targeting lentiviral vector system by optimizing these two parameters in parallel without compromising their functions.
The lentiviral vectors bearing both SCAb and FM can specifically transduce CD20-expressing cells. The specific transduction occurs through a two-step process. First the virus must recognize and bind to CD20-expressing cells. Using flow cytometry and confocal microscopy, we verified that the SCAb was able to mediate the binding of the vector to the CD20 antigen on the cellular surface. Furthermore, the soluble αCD20 inhibition assay revealed that the targeting kinetics of the SCAb vector was inhibited in a dose-dependent fashion, confirming the binding requirement for the observed targeting. The second step for targeted transduction is the FM-mediated endosomal fusion to deliver the viral payload into the cell. A high titer and efficient transduction demonstrated that the FM was functional when combined with SCAb on the viral surface. Thus, the targeting lentiviral vector succeeds in these two steps to achieve efficient transduction.
As suggested from our previous studies, two different approaches can be applied to further optimize this two-molecule targeting strategy. By engineering the fusion loop of the SINmu, transduction can be enhanced . Lentiviral vectors incorporating SCAb and SINmu consistently yielded lower transduction efficiency as compared to viral vectors with other FMs (SGN, SGM, or AGM). The difference in transduction efficiency may have resulted from the endosomal fusion kinetics of the different FMs. Recent studies of alphavirus glycoproteins have indicated that mutation in the E1 fusion domain might favor an increase in endosomal fusion ability [31, 32, 33]. We suspected that our mutation in the E1 domain might have a similar role in lowering the activation energy for the fusion event. Our liposome-virus fusion assay revealed that SINmu was not fusion-active at pH = 6.2, while other FMs were active at this pH. This direct correlation between the pH of fusion and the transduction efficiency suggests that the FMs that are more active at a higher pH can have better capacity to mediate lentiviral transduction. Consequently, targeted transduction may be further improved by constructing a library of FMs and screening for a FM with higher pH fusion activity.
Another approach to optimize this two-molecule targeting strategy is to engineer the targeting antibody to be more efficiently incorporated onto the lentiviral vector surface. Having the targeting molecule more efficiently incorporated onto the vector surface could enhance the binding of the vector to the cognate receptor on the target cell surface, thereby increasing transduction efficiency. To enhance the display of SCAb onto the viral surface, we constructed two SCAbs, each fused to a different transmembrane domain: the HLA-A2 transmembrane domain or the VSVG transmembrane domain. The targeted transduction efficiency was consistently higher with the SC2H7-A2-bearing vector. The binding avidity from the scatchard analysis revealed that the FUGW/SC2H7-A2/SGN vector exhibited a slightly higher binding avidity as compared to FUGW/SC2H7-GS/SGN. The higher avidity of the SC2H7-A2-bearing vector may be due to more efficient incorporation of SC2H7-A2 onto the lentiviral vector surface. It has been proposed that lipid rafts can serve as assembly sites for the pseudotyped lentiviral vectors . Recent studies have demonstrated a correlation between transmembrane domain and raft association with efficient viral incorporation . Although these data indicate a role of the transmembrane interaction to facilitate more efficient incorporation onto the virus, further understanding is needed to identify the precise mechanism of the transmembrane to facilitate incorporation of both the SCAb and FMs.
Materials and methods
To generate the SCAb against the CD20 antigen, we first PCR-amplified the light chain variable region from an αCD20 hybridoma cell line (ATCC, Manassas, VA, HB-9803) with primers CD20Lvfw (5'-CTG ACC CAG ACC TGG GCG CAA ATT GTT CTC TCC CAG TCT CCA GCA ATC CTG TC-3') and CD20LvGSbw (5'-CAC CTC CTG AAC CAC CGC CGC TAC CGC CTC CGC CTT TCA GCT CCA GCT TGG TCC CAG CAC C-3'). The HLA-A2 leading peptide sequence was then added to the 5'-end of the light chain variable region with primers HLA-A2 (5'-GAA CAA TTT GCG CCC AGG TCT GGG TCA GGG CCA GAG CCC CCG AGA GTA GCA GGA CGA GGG TTC-3') and HLA-A2fw (5'-CTT AAG CTT ATG GCC GTC ATG GCG CCC CGA ACC CTC GTC CTG CTA CTC TCG GGG G-3'). We also PCR-amplified the heavy chain variable region with primers CD20hvGSfw (5'-GGT AGC GGC GGT GGT TCA GGA GGT GGC GGC AGT GGT GGA GGA TCT CAG GCT TAT CTA CAG CAG TCT GGG GCT GAG CTG-3') and CD20hvbw (5'-GTT TTG TCA CAA GAT TTG GGC TCA ACT GAA GAG ACG GTG ACC GTG GTC CCT GTG-3'). The PCR product was assembled with the light chain variable region using the primers HLA-A2fw and CD20hvbw. To fuse the hinge-CH2-CH3 domain to the HLA-A2 transmembrane domain, we PCR-amplified the hinge-CH2-CH3 domain and the HLA-A2 transmembrane domain using the primer pairs (CH2-CH3-Hingefw, 5'-GTC TCT TCA GTT GAG CCC AAA TCT TGT GAC AAA ACT CAC ACA TGC CCA CCG TGC CCA GCA CCT GAA CTC CTG GGG GGA CCG TC-3'; CH2-CH3bw, 5'-CTG GGA AGA CGG GGC CCC CTG TCC GAT CAT GTT CCT G-3') and (HLA-A2Tfw, 5'-GAC AGG GGG CCC CGT CTT CCC AGC CCA CCA TCC CC-3'; HLA-A2Tbw, 5'-CGA GCG GCC GCT CAC ACT TTA CAA GCT GTG AGA GAC ACA TCA GAG CCC-3'). The resulting two PCR fragments were assembled using primers CH2-CH3-Hingefw and HLA-A2Tbw. We then assembled the variable fragments with the CH2-CH3/transmembrane domain using the primers HLA-A2fw and HLA-A2Tbw. The assembled DNA was finally cloned into pcDNA3 (Invitrogen) via Hind3 and Not1 restriction sites. To construct a single chain antibody with the VSVG transmembrane domain, a forward primer (SC2H7fw, 5'-CCC CCA TCC CGG GAT GAG CTG ACC-3') and a backward primer (SC2H7bw, 5'-AGT ATC ACC GGC CCC CTG TCC GAT CAT GTT CCT GTA GTC-3') were used to amplify a portion of the CH2-CH3 domain of SC2H7-A2. In parallel, a forward primer (GSfw, 5'-ATG ATC GGA CAG GGG GCC GGT GAT ACT GGG CTA TCC AAA AAT CCA ATC GAG CTT-3') and a backward primer (GSbw, 5'-GAT CGA GCG GCC GCT TAC TTT CCA AGT CGG TTC ATC TCT ATG TCT GTA TAA ATC TGT CTT TTC-3') were used to amplify the transmembrane domain of VSVG. The DNA products from these two reactions were PCR-assembled using SC2H7fw and GSbw as the primer pair and the resulting product was cloned into pSC2H7-A2 to yield SC2H7-GS. The integrity of these constructs was confirmed by DNA sequencing.
Viral vector production
293T cells were seeded in a 6-cm culture dish in DMEM medium supplemented with fetal bovine serum (Sigma, St. Louis, MO, 10%), L-glutamine (10 mL/L), penicillin, and streptomycin (100 units/mL) the night prior to transfection. 293T cells were transfected at a confluence of 80~90% with 5 μg of lentiviral backbone vector (FUGW), 2.5 μg each of pMDLg/pRRE, pRSV-Rev, pFM, and a plasmid encoding an antibody (pSC2H7-A2, pSC2H7-GS or pAB) via the standard calcium phosphate precipitation technique . Cells were replenished with pre-warmed media 4 hours post-transfection. Vectors were harvested two days post-transfection and filtered through a 0.45-μm pore size filter (Nalgene, Rochester, NY). Lentiviral vectors were then further concentrated by ultracentrifugation (Optimal L-90K Ultracentrifuge, Beckman Coulter, Fullerton, CA) at 4°C, 25,000 rpm for 90 minutes and resuspended in appropriate volume of cold PBS.
Virus-cell binding assay
293T/CD20 or 293T cells were incubated with 2 mL of lentiviral vectors (FUGW/SC2H7-A2/FM, FUGW/SC2H7-GS/FM or FUGW/AB/FM) at 4°C for 1 hour. After extensive washing with cold PBS, cell-virus complexes were stained with anti-HA tag antibody (Miltenyi Biotec, Inc.) and analyzed by flow cytometry (FACSort, BD Bioscience).
Fluorescent images were acquired on a Zeiss LSM 510 META laser scanning confocal microscope equipped with Argon, red HeNe, and green HeNe lasers as well as a Coherent Chameleon Ti-Sapphire laser for multiphoton imaging. Images were acquired using a Plan-apochromat 63x/1.4 oil immersion objective. To image virus-cell binding, cells were seeded into a 35-mm glass-bottom culture dish and grown at 37°C overnight. The seeded cells were rinsed with cold PBS and incubated with concentrated viral particles for 1 hour at 4°C to allow for binding. The cells were washed with cold PBS to remove unbound particles, fixed with 4% formaldehyde on ice for 10 minutes, and then immunostained with monoclonal antibody specific for HIV capsid protein p24 and 4',6-diamidino-2-phenylindole (DAPI) antibody for nuclear staining. Monoclonal antibody against HIV-1 p24 (AG3.0) was obtained from the NIH AIDS Research and Reference Reagent Program (Division of AIDS, NIAID, NIH). Images were analyzed using the Zeiss LSM 510 software version 3.2 SP2.
Antibody Competition Assay
293T/CD20 cells were incubated with the lentiviral vector (FGUW/SC2H7-A2/SGN or FUGW/VSVG) and various amount of either the soluble αCD20 (BD Bioscience) or the isotype control antibody overnight. Cells were then replenished with fresh media and incubated for additional 72 hours before flow cytometry analysis.
293T/CD20 or Ghost-CCR5 (NIH AIDS Research and Reference Reagent Program) cells were pre-incubated with various amount of bafilomycin for 30 minutes, after which, the lentiviral vector (FUGW/SC2H7-A2/SGN or FUGW/GP160) was added. The vector and cell mixture was spun at 25°C, 2,500 rpm for 90 minutes using a RT legend centrifuge (Sorval). Cells were then incubated at 37°C and 5% CO2 and replenished with fresh media 3 hours later. Flow cytometry was then used to analyze the treated cells 3 days post-transduction.
Targeted transduction of 293T/CD20 cells
293T or 293T/CD20 cells were seeded on a 24-well cell culture plate and spin-transduced with 1.5 mL of indicated lentiviral vectors (FUGW/SC2H7-A2/FM, FUGW/SC2H7-GS/FM, FUGW/AB/FM, FUGW/SC2H7-A2, or FUGW/SC2H7-GS) at 25°C, 2,500 rpm for 90 minutes using a RT legend centrifuge. After replacing with fresh media, the treated cells were cultured for additional 5 days at 37°C and 5% CO2. Flow cytometry was then used to analyze transduction efficiency. The titer was determined by measuring GFP-positive cells in the dilution range that resulted in a linear relationship between the percentage of GFP-expressing cells and the amount of vectors added.
293T/CD20 cells were incubated with various amount of lentiviral vectors (FUGW/SC2H7-A2/SGN or FUGW/SC2H7-GS/SGN). Flow cytometry analysis was carried out to measure the geometric mean fluorescence (GMF) of the bound viruses stained by anti-HA antibody. The concentration of the lentiviral vectors was measured by a p24 antigen capture enzyme immunosorbent assay (ELISA) kit (ImmunoDiagnostics, Woburn, MA). Apparent Kd value was derived from the negative reciprocal of the slope of the linear fit to scatchard plots, which is the geometric mean fluorescence/concentration of lentiviral vector (GMF/concentration) against geometric mean fluorescence (GMF).
Virus-liposome fusion assay
1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) and 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine (DOPE) were purchased from Avanti Polar Lipids (Alabaster, AL, USA). Cholesterol (Chol) and sphingomyelin (SPM) from egg yolk were obtained from Sigma (St Louis, MO, USA). Liposomes were prepared by the extrusion procedure . Briefly, lipid mixtures (PC/PE/SPM/Chol molar ratio of 1:1:1:2) were dried from a chloroform solution under a stream of argon gas and further dried under vacuum for at least 3 hours. The lipid mixtures were hydrated in HNE buffer (5 mM HEPES, 150 mM NaCl, and 0.1 mM EDTA, pH 7.4). Subsequently, the lipid mixtures were extruded 20 times through 0.2 μm pore size polycarbonate filters (Avanti polar lipids). To monitor virus-liposome fusion, the concentrated viruses were incubated with 70 μM of octadecyl rhodamine B chloride (R18) (Molecular Probes, Carlsbad, CA, USA) in serum-free medium for 1 hour at room temperature. R18-labeled viruses were then mixed with liposomes (200 μM phospholipids) in a final volume of 0.4 mL. Fusion was triggered by adding the appropriate volume of 0.2 M acetic acid, pretitrated to achieve the desired pH. The dequenching signal of R18 fluorescence was measured 60 seconds after acidification with QuantaMaster QM-4SE spectrofluorometer (Photon Technology International, Lawrenceville, NJ, USA). The initial fluorescence of virus-liposome mixtures was set at 0% fusion, and the 100% fusion value was obtained by detergent lysis for each experiment using 0.1% of Triton X-100 .
- 3.Burns J, Friedmann T, Driever W, Burrascano M, Yee J: Vesicular stomatitis virus G glycoprotein pseudotyped retroviral vectors: concentration to very high titer and efficient gene transfer into mammalian and nonmammalian cells. Proc Nat Acad Sci USA 1993, 90: 8033-8037. 10.1073/pnas.90.17.8033PubMedPubMedCentralCrossRefGoogle Scholar
- 19.Roux P, Jeanteur P, Piechaczyk M: A versatile and potentially general approach to the targeting of specific cell types by retroviruses: application to the infection of human cells by means of major histocompatibility complex class I and class II antigens by mouse esctropic murine leukemia virus-derived viruses. Proc Nat Acad Sci USA 1989, 86: 9079-9083. 10.1073/pnas.86.23.9079PubMedPubMedCentralCrossRefGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.