CD62L as target receptor for specific gene delivery into less differentiated human T lymphocytes

Abstract

Chimeric antigen receptor (CAR)-expressing T cells are a complex and heterogeneous gene therapy product with variable phenotype compositions. A higher proportion of less differentiated CAR T cells is usually associated with improved antitumoral function and persistence. We describe in this study a novel receptor-targeted lentiviral vector (LV) named 62L-LV that preferentially transduces less differentiated T cells marked by the L-selectin receptor CD62L, with transduction rates of up to 70% of CD4+ and 50% of CD8+ primary T cells. Remarkably, higher amounts of less differentiated T cells are transduced and preserved upon long-term cultivation using 62L-LV compared to VSV-LV. Interestingly, shed CD62L neither altered the binding of 62L-LV particles to T cells nor impacted their transduction. The incubation of 2 days of activated T lymphocytes with 62L-LV or VSV-LV for only 24 hours was sufficient to generate CAR T cells that controlled tumor growth in a leukemia tumor mouse model. The data proved that potent CAR T cells can be generated by short-term ex vivo exposure of primary cells to LVs. As a first vector type that preferentially transduces less differentiated T lymphocytes, 62L-LV has the potential to circumvent cumbersome selections of T cell subtypes and offers substantial shortening of the CAR T cell manufacturing process.

Keywords: CAR T cells; L-selectin; LV; chimeric antigen receptor; naïve T lymphocytes; receptor-targeted viral vectors; ΔLNGFR.

Figure 1

Basic characterization of 62L-LV. (A) Physical properties of 62L-LV vector stocks. Three independently produced 62L-LV stocks were analyzed for size (left panel) and particle concentration (right panel) by nanoparticle tracking analysis (filled bar; technical triplicates) or p24-ELISA (open bar, biological replicates). Means and standard deviations (SD) are depicted. (B) 62L-LV stocks were titrated on HT1080_αHis cells or activated human PBMC. Individual results of biological replicates and means with standard error (SEM) are plotted. (C) The indicated panel of HT1080 cells was incubated with 2.5 µL 62L-LV stock or left untransduced. Four days later, antibody staining against ΔLNGFR allowed for the detection of transduced cells by flow cytometry. Left panel: Representative dot plots for one vector stock. Right panel: Percentages of ΔLNGFR positive cells after transduction with seven different vector stocks. Dashed lines indicate detection levels for each individual cell line. Individual results as well as means with SD are plotted. Statistical testing was calculated by using ordinary 1-way ANOVA. WT = parental HT1080 cells.

Figure 2

CAR gene delivery into primary lymphocytes by 62L-LV. Activated PBMCs were incubated with 62L-LV (blue dots) or VSV-LV (orange dots). (A) Transduction rates in the presence (+V1) or absence (-V1) of Vectofusin-1 as determined by ΔLNGFR expression. Left panel: Representative dot plots of 62L-LV transduced PBMC pre-gated for CD3⁺ cells. The percentage of ΔLNGFR expression presented as numbers in the individual gates refer to the CD4+ (upper gates) or CD4- cells (bottom gates), respectively. Right panel: results from seven different donors in four independent experiments analyzed 9 - 12 days post-transduction. For VSV-LV, V1 was not applied. Individual results of biological replicates and means with standard deviation (SD) are plotted. Statistical analysis was performed by using paired t-test for the comparison of 62L-LV +/-V1 and by using an unpaired t-test for the comparison of 62L-LV and VSV-LV. ns, not significant. (B) The total percentage of ΔLNGFR+ cells expressing CD62L is displayed for the CD4⁺ (left) and CD8⁺ (right) fractions. Transduction was performed in the absence of V1 for 62L-LV and VSV-LV, respectively. Cells from three different donors transduced with either vector in two individual experiments were tested for significant differences at each analysis time point individually by Fisher’s least significant difference (LSD) test. The mean with standard error (SEM) of eight biological replicates is plotted. The gating strategy is depicted in Supplementary Figure 15 .

Figure 3

Selectivity of 62L-LV binding to primary T lymphocytes. Activated PBMC incubated either with the CD62L-specific antibody (blue bars) or with the CD45-specific antibody (grey bars) at the indicated concentrations before PBS (open bars) or 62L-LV vector particles (filled bars) were added for 30 min at 4°C. (A) Fluorophore-labeled αCD62L and αCD45 antibodies were used to determine the staining intensity of CD62L and CD45 on activated PBMC at the indicated concentrations by flow cytometry. Mean fluorescent intensities (MFI) are shown. N=1. (B) PBMC was pre-incubated with biotin-labeled antibodies before the addition of 62L-LV particles. After vector incubation, cells were stained with fluorophore-coupled αCD3 and αLNGFR antibodies to allow for the detection of vector-bound T cells by flow cytometry. Background MFI (dashed line) was determined from samples incubated with PBS (w/o) for all antibody concentrations. Means of MFI and standard deviations (SD) of three technical replicas are depicted. Statistical testing was performed by using 2-way ANOVA. ns, not significant.

Figure 4

Shed CD62L does not influence vector binding. (A) Accumulation of sCD62L in the supernatant of PBMC. Frozen PBMCs were thawed, activated, and cultivated in the presence of IL-7 and IL-15. The complete supernatant of one well was collected on the indicated day and used for sCD62L quantification (left panel). The amounts of sCD62L present in three independent cultures on day 6 are shown in the right panel. Individual results and means and standard deviation (SD) are depicted. (B) 62L-LV (blue) or VSV-LV (orange) particles were incubated with fresh or frozen supernatant containing sCD62L (day 6 harvest) or cell medium (TCM) only. A mixture of vector stock and supernatant was incubated with activated PBMC for 30 min at 4°C. Flow cytometry was performed to analyze the content of vector-bound T cells by staining with fluorophore-coupled αCD3 and αLNGFR antibodies. Dot plots of vector-bound T cells are depicted in the right panel. The percentage of vector-bound cells is indicated. ΔLNGFR intensity [MFI] of vector-bound cells in three to 10 independent experiments is shown in the left panel. Individual results and means with SD are plotted. Statistical testing was performed by using 2-way ANOVA. ns, not significant.

Figure 5

Transduction on CD62L-enriched and -depleted T cells. (A) CD62L frequency on activated T cells before (T cells) and after magnetic-activated cell separation using CD62L-APC (clone REAL163) and anti-APC microbeads into CD62L-enriched (CD62L enr.) and CD62L-depleted (CD62L depl.) fractions. (B) Percentages of ΔLNGFR+ cells 3 days after transduction of the separated cell fractions shown in panel A with 62L-LV or VSV-LV. (C–E) CD62L expression after separation. In an independent experiment, the percentage of CD62L+ cells was determined directly after separation (day 0) and upon 7 days of cultivation. Data are shown as measured (C), left diagram), normalized to the values in the enriched fraction (C), right diagram), and as exemplary FACS plots on day 0 (D) and day 7 (E). Individual results as well as means with standard deviation are shown for three donors measured in technical triplicates. Statistical testing was performed by RM 2-way. ns, not significant.

Figure 6

Antitumoral activity of CAR T cells generated with 62L-LV. (A) Experimental setting. PBMC were activated for two days prior to 24h incubation with 62L-LV, VSV-LV, or PBS (control) and injected i.v. into NSG mice (n=3 per group). Nalm6 cells were injected on day 4 of post-adoptive cell transfer and their growth was monitored by bioluminescence imaging (BLI). (B) Monitoring for tumor load by BLI at the indicated days after adoptive cell transfer. Ventral images of each mouse are depicted. (C) Total body flux quantified at the indicated time points for the 62L-LV group (blue), the VSV-LV group (orange), and the control (grey). Individual results and mean with standard error (SEM) are plotted. The dotted line represents the background signal of mice without imaging substrate. Ordinary two-way ANOVA was used to determine statistics. P-values are indicated when below 0.05. (D) Cells isolated from the blood and organs of each mouse were analyzed by flow cytometry for viable, CD45 negative, CD19, and EBFP double-positive Nalm6 cells. The percentage of Nalm6 positive cells of all viable cells is depicted. Individual results and means with standard error (SD) are plotted. Ordinary one-way ANOVA was used to determine statistics. P-values are indicated when below 0.05.

Figure 7

Characterization of CAR T cells from the in vivo experiment (A) Monitoring of CAR T cells in the blood. The fraction ΔLNGFR+ cells within human CD4⁺ or CD8⁺ T cells was determined by flow cytometry on days 3, 10, and 17 of the in vivo experiment. Gating on viable human CD45⁺, CD3⁺, and respective lineage marker-positive cells was performed. Only samples with at least 20 events in the CD4⁺ or CD8⁺ gates were considered. Mean values with standard error (SEM) are depicted. N=3. (B-E) Cellular composition in the blood at the final analysis. Frequencies of human CD45⁺ cells (left), ΔLNGFR+/human CD4⁺ (middle), and ΔLNGFR+/human CD8⁺ (right) are shown in (B), those of Teff, Tem, Tcm, Tn within ΔLNGFR+/human CD4+ (left), and ΔLNGFR+/human CD8+ (right) in (C), and frequencies of non-exhausted ΔLNGFR+/human CD4+ (left) and ΔLNGFR+/human CD8+ (right) as determined by double-negative TIM-3 and LAG-3 expression in (D). Individual results of each mouse and mean values with standard deviation (SD) are depicted. Unpaired t-tests were performed to determine statistics. P-values are indicated when below 0.05.