Combining T-cell-specific activation and in vivo gene delivery through CD3-targeted lentiviral vectors

Abstract

Genetic modification of T lymphocytes is a key issue in research and therapy. Conventional lentiviral vectors (LVs) are neither selective for T cells nor do they modify resting or minimally stimulated cells, which is crucial for applications, such as efficient in vivo modification of T lymphocytes. Here, we introduce novel CD3-targeted LVs (CD3-LVs) capable of genetically modifying human T lymphocytes without prior activation. For CD3 attachment, agonistic CD3-specific single-chain variable fragments were chosen. Activation, proliferation, and expansion mediated by CD3-LVs were less rapid compared with conventional antibody-mediated activation owing to lack of T-cell receptor costimulation. CD3-LVs delivered genes not only selectively into T cells but also under nonactivating conditions, clearly outperforming the benchmark vector vesicular stomatitis-LV glycoproteins under these conditions. Remarkably, CD3-LVs were properly active in gene delivery even when added to whole human blood in absence of any further stimuli. Upon administration of CD3-LV into NSG mice transplanted with human peripheral blood mononuclear cells, efficient and exclusive transduction of CD3+ T cells in all analyzed organs was achieved. Finally, the most promising CD3-LV successfully delivered a CD19-specific chimeric antigen receptor (CAR) into T lymphocytes in vivo in humanized NSG mice. Generation of CAR T cells was accompanied by elimination of human CD19+ cells from blood. Taken together, the data strongly support implementation of T-cell-activating properties within T-cell-targeted vector particles. These particles may be ideally suited for T-cell-specific in vivo gene delivery.

Graphical abstract

Figure 1.

CD3-LVs transduce T cells independently of activation or cytokine treatment. (A-B) Primary human PBMCs isolated from different blood donations were activated with αCD3 and αCD28 antibodies for 3 days in the presence of IL-2 before they were incubated with CD3-LVs or VSV-LV. Green fluorescent protein (GFP) expression was determined 6 days later by flow cytometry. (A) Representative dot plots show GFP expression in CD4⁺ and CD4⁻ populations gated for viable cells. (B) Scatter bar diagrams summarize the percentages of GFP⁺ cells as mean ± standard deviation (SD) of 3 independent experiments with 2-3 donors and 2-3 technical replicates. (C-D) Freshly isolated PBMCs were cultured overnight in presence of IL-2 or IL-7/IL-15, or in absence of cytokines, before they were transduced with CD3-LVs or VSV-LV. GFP expression was determined 6 days later by flow cytometry. (C) Scatter bar diagrams show the percentage of GFP⁺ cells in all viable single cells. (D) Representative dot plots show GFP expression in the CD8⁺ and CD8⁻ populations gated for viable cells. Data are mean ± SD from 1 experiment with n = 3 technical replicates; biological replicates from additional experiments with IL-2- and IL-7/IL-15-stimulated cells are shown in supplemental Figure 6C-F. **P < .01, ***P < .001, ****P < .0001 by 2-way analysis of variance (ANOVA) with Dunnett's correction (comparison with VSV-LV for each condition), or unpaired Student t test.

Figure 2.

CD3-LVs activate cytokine-cultured T cells and induce cytokine production and proliferation. PBMCs isolated from adult blood were cultured overnight in the presence of IL-2 only and then transduced with the indicated CD3-LVs or with VSV-LV. Controls were left untransduced (ut) or activated with αCD3/αCD28 antibodies (1 µg/mL αCD3, 3 µg/ml αCD28) until first medium exchange. (A-B) Expression of the activation markers CD69 (A) and CD25 (B) on all viable cells was followed for 6 days by flow cytometry. The number of activation marker positive cells are shown. N = 3 donors, mean ± standard error of the mean. *P < .05, **P < .01, ***P < .001, ****P < .0001 by 2-way ANOVA with Dunnett's correction. (C-E) One day postincubation with vector particles or recombinant antibodies, cytokines secreted into the cell culture supernatant were quantified. Scatter bar diagrams show the concentration of interferon-γ (C), tumor necrosis factor-α (D), and IL-2 (E) for each condition. Mean ± SD from 1 experiment with n = 3 triplicates each are shown. (F) PBMCs were stained with CellTrace Violet (CTV) before transduction ± costimulation to follow cell proliferation over time. Histograms show the fluorescence of CTV at day 5 posttransduction. Data are representative of 3 different donors. (G) Bar diagrams display T-cell expansion after 6 days compared with day 0 ± costimulation. N = 3 donors, mean ± SD. ns, nonsignificant; *P < .05 by 1-way ANOVA with Dunnett's correction. (H) Bar diagrams show percentages of GFP⁺ cells gated from all viable single cells ± costimulation at day 6 posttransduction. N = 3 donors. Mean ± SD. ns, nonsignificant by 2-way ANOVA with Sidak's correction.

Figure 3.

Endocytosis of the TCR/CD3 complex upon transduction. (A-B) Downmodulation of CD3 (A) and TCR (B) on PBMC stimulated only with IL-2 and transduced with CD3-LVs (in absence of Vectofusin-1), VSV-LV, or left untransduced but activated (αCD3/αCD28). The mean fluorescence intensities measured were normalized to those of untreated nonactivated cells. N = 3 donors, mean ± standard error of the mean. *P < .05, **P < .01, ***P < .001 by 2-way ANOVA with Dunnett's correction. (C-D) Inhibition of endocytosis with Bafilomycin A1 (Bfla1) (C) or NH₄Cl (D) increases transduction efficiency on Jurkat cells with CD3-LVs. Mean ± SD of 3-4 independent experiments (E), or 1 experiment with 3 technical replicates (D). *P < .05, **P < .01, ****P < .0001 by unpaired Student t test. (E) Transduction of primary PBMC cultured in IL-2 by CD3-LVs upon treatment with Bfla1. N = 2 technical replicates from 1 donor, mean ± SD.

Figure 4.

CD3-LVs transduce human T cells in whole blood. CD3-LVs and VSV-LV were tested for their ability to transduce T cells in whole blood of healthy donors in absence of additional stimuli and transduction enhancer. (A) Experimental procedure. (B) Representative dot plots show the percentages of GFP⁺ cells gated from all viable single cells at day 5 posttransduction in isolated PBMC. (C-D) Bar diagrams show the percentages of GFP⁺ cells of all viable single cells (C) and in CD4⁺ and CD8⁺ cell subsets (D) at day 5 posttranduction. *P < .05, **P < .01 by mixed-effect analysis with Sidak's correction. Mean ± SD of 3 independent experiments with 5 different donors. *P < .05, **P < .01 by mixed-effect analysis with Dunnett's correction.

Figure 5.

TR66-LV mediates efficient in vivo gene delivery into all T-cell subpopulations. (A) Experimental outline: NSG mice were transplanted with αCD3/αCD28-activated human PBMC by intraperitoneal (IP) injection followed by vector administration 1 day later and analyzed for transgene expression by flow cytometry 7 days after vector application. (B) Representative dot plots show transduced T cells gated from all viable single CD3⁺ cells harvested from the peritoneal cavity at 7 days after vector injection. (C-E) Scatter plots summarize the percentages of GFP⁺ cells gated from all viable single CD3⁺ cells for peritoneum (C), blood (D), and spleen (E). (F-H) Respective scatter plots of GFP⁺ cells gated from CD4⁺ and CD8⁺ T cells isolated from peritoneum (F), blood (G), and spleen (H) at the day of analysis. N = 3 mice per group, mean ± SD. **P < .01, ***P < .001, ****P < .0001 by 1-way analysis of variance with Sidak's or Dunnett's correction.

Figure 6.

Selectivity of TR66-LV in vivo. (A) Strategy for backgating of GFP⁺ cells generated in PBMC-NSG mice of Figure 5. All viable single GFP⁺ cells were assessed for expression of human CD45, CD3, and CD19. (B-D) Scatter plots for the percentages of GFP⁺ cells in the indicated fractions of peritoneum (B), blood (C), and spleen (D). Data are mean ± SD from n = 3 animals per group. ****P < .0001 by 1-way ANOVA with Sidak's correction.

Figure 7.

CD3-targeted LVs mediate functional CAR T-cell generation in vivo. For in vivo CAR delivery, huNSG mice were subcutaneously (SC) injected with human IL-7 or phosphate-buffered saline 4 days and 1 day before IV injection of TR66.opt-LV harboring the CD19-CAR or vehicle as control. (A) Experimental outline. (B) Representative plots show CAR⁺ cells in CD8⁺ and CD8⁻ T cells at day 26 postinjection of vector. (C) Line diagrams show CAR⁺ cells in CD3⁺ cells in the peripheral blood of individual mice in group 1 (IL-7), group 2 (TR66.opt-LV), and group 3 (TR66.opt-LV + IL-7). *^,#P < .05, **^,##P < .01, ***^,###P < .001 comparing group 3 to group 1 (*) and group 2 (^#) by 2-way ANOVA with Tukey's correction. Arrows point to tendencies for CAR T-cell detection in group 2, which were not statistically significant. (D) Scatter dot plots compare the percentages of CAR⁺ cells within the CD3⁺CD45⁺ T-cell compartment, CD45⁻ mouse cells and CD3⁻CD45⁺ non-T cells. (E-F) B-cell levels were determined in peripheral blood of the animals to assess CD19-CAR T-cell functionality. (E) Representative dot plots show CD19⁺CD20⁺ B cells at day 26 postinjection. (F) Line diagrams summarize relative levels of CD19⁺ cells compared with day 0 in the peripheral blood of individual mice in group 1 (IL-7), group 2 (TR66.opt-LV), and group 3 (TR66.opt-LV + IL-7). All data from n = 5-6 mice per group. *P < .05, **P < .01, ***P < .001 comparing groups 2 and 3 to group 1 by 2-way ANOVA with Turkey's correction.