Measles virus glycoprotein-based lentiviral targeting vectors that avoid neutralizing antibodies

Abstract

Lentiviral vectors (LVs) are potent gene transfer vehicles frequently applied in research and recently also in clinical trials. Retargeting LV entry to cell types of interest is a key issue to improve gene transfer safety and efficacy. Recently, we have developed a targeting method for LVs by incorporating engineered measles virus (MV) glycoproteins, the hemagglutinin (H), responsible for receptor recognition, and the fusion protein into their envelope. The H protein displays a single-chain antibody (scFv) specific for the target receptor and is ablated for recognition of the MV receptors CD46 and SLAM by point mutations in its ectodomain. A potential hindrance to systemic administration in humans is pre-existing MV-specific immunity due to vaccination or natural infection. We compared transduction of targeting vectors and non-targeting vectors pseudotyped with MV glycoproteins unmodified in their ectodomains (MV-LV) in presence of α-MV antibody-positive human plasma. At plasma dilution 1:160 MV-LV was almost completely neutralized, whereas targeting vectors showed relative transduction efficiencies from 60% to 90%. Furthermore, at plasma dilution 1:80 an at least 4-times higher multiplicity of infection (MOI) of MV-LV had to be applied to obtain similar transduction efficiencies as with targeting vectors. Also when the vectors were normalized to their p24 values, targeting vectors showed partial protection against α-MV antibodies in human plasma. Furthermore, the monoclonal neutralizing antibody K71 with a putative epitope close to the receptor binding sites of H, did not neutralize the targeting vectors, but did neutralize MV-LV. The observed escape from neutralization may be due to the point mutations in the H ectodomain that might have destroyed antibody binding sites. Furthermore, scFv mediated cell entry via the target receptor may proceed in presence of α-MV antibodies interfering with entry via the natural MV receptors. These results are promising for in vivo applications of targeting vectors in humans.

Figure 1. Schematic drawing of cytoplasmic tail-truncated hemagglutinin envelope proteins used for pseudotyping of lentiviral vectors.

In the mutated hemagglutinin protein (H_mut) that is derived from the NSe variant of the measles virus (MV) vaccine strain Edmonston B, mutations in the MV receptor recognition regions Y481A, R533A, S548L and F549S (ectodomain) are indicated by asterisks. Glycine-serine linker ((G₄S)₃) or the factor Xa cleavage site (IEGR) were used as linker region between H_mut and single-chain antibody (scFv). A histidine tag (H6) is present at the scFv C-terminus. The hemagglutinin protein derived from the NSe variant of the MV vaccine strain Edmonston B that is not mutated and does not display a scFv is labeled H_NSe. The hemagglutinin protein derived from the wild-type measles virus strain IC-B is labeled H_wt. All hemagglutinin proteins are truncated by 18 amino acids in their cytoplasmic tail (Δ18) to allow incorporation into the lentiviral envelope. The names of the respective vector particles pseudotyped with the depicted H variants are indicated on the left site. w/o: without.

Figure 2. Targeting vectors are protected against MV neutralizing antibodies.

The indicated vector particles were incubated in serial plasma dilutions of two different α-MV antibody-positive donors. (a) 3×10⁴ CD20-positive Raji cells (MOI 0.4) were added, or the dilutions were added to (b) CD105/CD20-positive HT1080-CD20 (MOI 0.3) or (c) HT1080-CD133 cells (MOI 0.3) that were seeded at a density of 1.7×10⁴ and 0.75×10⁴ cells per 96 well, respectively, 24 h before transduction. Forty-eight to 72 h later, the fraction of EGFP-positive cells was quantified by FACS analysis. The relative transduction efficiency compared to transduction in absence of plasma (medium control) of one representative donor is shown for each cell line.

Figure 3. Influence of vector dose on neutralization.

The indicated vector particles were incubated at varying doses in a 1∶80 plasma dilution containing (a) 1200 mU/ml or (b) 4900 mU/ml α-MV antibodies. Then, the dilutions were added to (a) CD105-positive HT1080-CD20 cells that were seeded at a density of 1.7×10⁴ cells per 96 well 24 h before transduction or (b) 3×10⁴ CD20-positive Raji cells were added to the dilutions. Forty-eight to 72 h later, the fraction of EGFP-positive cells was determined by FACS analysis. The relative transduction efficiency is shown (transduction efficiency in presence of a 1∶80 FCS dilution was set to 100%). Arrows indicate relative transduction efficiencies of <1%.

Figure 4. Influence of particle amount on neutralization.

Equal amounts of vector particles as determined by p24 ELISA were incubated in serial plasma dilutions of two different α-MV antibody-positive donors. The dilutions were added to (a) CD133-positive HuH7 cells that were seeded 24 h before transduction at a density of 1.0×10⁴ (CD133-specific vectors) and 5.0×10⁴ (MV_NSe-LV) cells per 96 well, respectively, to apply similar MOIs. Alternatively, the dilutions were added to (b) 1.0×10⁴ (CD20-LV) and 5.0×10⁴ (MV_NSe-LV) CD20-positive Raji cells, respectively. Seventy-two h later, the percentage of EGFP-positive cells was determined by FACS analysis. The relative transduction efficiency compared to transduction in absence of plasma of one representative donor is shown.

Figure 5. α-MV antibody-negative serum does not neutralize MV_NSe-LV or targeting vectors.

Equal amounts of physical particles of the indicated vector types were incubated in serial serum dilutions of an α-MV antibody-negative donor. Then, (a) 1×10⁴ (CD20-LV) and 2×10⁵ (MV_NSe-LV) CD20-positive Raji cells were added, respectively, or the dilutions were added to (b) CD133-positive HuH7 cells. These were seeded at a density of 1.0×10⁴ (targeting vectors) and 5.0×10⁴ (MV_NSe-LV) cells per 96 or 48 well, respectively, 24 h before transduction, to apply similar MOIs of vector particles. Forty-eight to 72 h later, the percentage of EGFP-positive cells was determined by FACS analysis. As control, medium without serum was used.

Figure 6. Targeting vectors are protected against the neutralizing antibody K71.

Equal amounts of physical particles of the indicated vector types were incubated in presence of increasing amounts of antibody K71 (K71-Ab; putative epitope near the mutation sites in H_mut-scFv constructs) and L77 (L77-Ab; putative epitope distant to the mutation sites in H_mut-scFv constructs), respectively, in a final volume of 100 µl. After incubation at room temperature for 1 h, (a) 3×10⁴ (CD20-LV) or 4.5×10⁵ (MV_NSe-LV) CD20-positive Raji cells were added, or the vectors were added to (b) 9.3×10⁴ (CD105-LV) or 3.4×10⁴ (MV_NSe-LV) CD105-positive HT1080 cells or (c) 3.4×10⁴ (CD133-LV) or 2.7×10⁴ (MV_NSe-LV) CD133-positive HuH7 cells, to apply an MOI of 0.25 for each vector type. Forty-eight hours later, the fraction of EGFP-positive cells was quantified by FACS analysis. Mean values (n = 3) and s.d. of the relative transduction efficiency compared to transduction in absence of antibody is shown for each cell line. (d) The indicated H proteins were expressed on the surface of HEK-293T cells. The control antibody K83 and the antibodies K71 and L77 were incubated with the cells, respectively, and a FITC-labeled secondary antibody was used to detect antibody binding to the different H proteins by FACS analysis. The percentage of FITC-positive cells subtracted by the staining of secondary antibody alone is shown. Arrows indicate 0% cell staining.