A novel method to generate T-cell receptor-deficient chimeric antigen receptor T cells

Abstract

Practical methods are needed to increase the applicability and efficacy of chimeric antigen receptor (CAR) T-cell therapies. Using donor-derived CAR-T cells is attractive, but expression of endogenous T-cell receptors (TCRs) carries the risk for graft-versus-host-disease (GVHD). To remove surface TCRαβ, we combined an antibody-derived single-chain variable fragment specific for CD3ε with 21 different amino acid sequences predicted to retain it intracellularly. After transduction in T cells, several of these protein expression blockers (PEBLs) colocalized intracellularly with CD3ε, blocking surface CD3 and TCRαβ expression. In 25 experiments, median TCRαβ expression in T lymphocytes was reduced from 95.7% to 25.0%; CD3/TCRαβ cell depletion yielded virtually pure TCRαβ-negative T cells. Anti-CD3ε PEBLs abrogated TCRαβ-mediated signaling, without affecting immunophenotype or proliferation. In anti-CD3ε PEBL-T cells, expression of an anti-CD19-41BB-CD3ζ CAR induced cytokine secretion, long-term proliferation, and CD19⁺ leukemia cell killing, at rates meeting or exceeding those of CAR-T cells with normal CD3/TCRαβ expression. In immunodeficient mice, anti-CD3ε PEBL-T cells had markedly reduced GVHD potential; when transduced with anti-CD19 CAR, these T cells killed engrafted leukemic cells. PEBL blockade of surface CD3/TCRαβ expression is an effective tool to prepare allogeneic CAR-T cells. Combined PEBL and CAR expression can be achieved in a single-step procedure, is easily adaptable to current cell manufacturing protocols, and can be used to target other T-cell molecules to further enhance CAR-T-cell therapies.

Graphical abstract

Figure 1.

Anti-CD3ε PEBLs block surface CD3 expression. (A) Flow cytometric dot-plots illustrate surface CD3 downregulation in Jurkat by anti-CD3ε PEBLs compared with cells transduced with GFP alone (“Control”) or SEKDEL. (B) Surface CD3 expression in Jurkat transduced with the indicated constructs. Bars show a mean of 2 to 3 experiments for KEDL, SEKDEL, and PEBL 2,4,5,8,9,11, or individual results for the remainder. (C) Intracellular or surface expression of PEBL-derived anti-CD3ε scFv in Jurkat. Bars show a mean of 2 to 3 experiments for PEBL 2,4,8,9,11, or individual results for the remainder. (D) Flow cytometric histograms illustrate CD3 expression in peripheral blood T cells transduced with anti-CD3ε SEKDEL or PEBLs, relative to that of lymphocytes transduced with GFP alone, 8 to 13 days posttransduction.

Figure 2.

Anti-CD3ε PEBLs downregulate TCRαβ. (A) TCRαβ expression in GFP-positive T lymphocytes 5 to 9 days after PEBL transduction. Mean (± standard deviation [SD]) is shown for cells transduced with GFP only (“Control”; n = 25), PEBL2 (n = 4), and PEBL5 (n = 18); other data represent results of 1 or mean of 2 experiments. (B) Flow cytometric dot-plots illustrate TCRαβ downregulation in T lymphocytes compared with cells transduced with GFP only. (C) CD3/TCRαβ expression in Jurkat cells transduced with PEBL after long-term culture; Control, cells transduced with GFP alone. (D) Collective results of CD3/TCRαβ expression in long-term cultures of T lymphocytes (with 200 IU/mL IL-2) or Jurkat cells. Symbols indicate persistence of more than 90% reduction of surface CD3/TCRαβ in GFP⁺ transduced cells.

Figure 3.

CD3/TCRαβ downregulation by PEBL does not affect cell proliferation, but abrogates CD3/TCRαβ signaling. (A) Growth rate of Jurkat transduced with anti-CD3 PEBLs or GFP only (“Control”). Symbols indicate mean (± SD) of triplicate measurements. (B) Survival of PEBL-transduced or control T lymphocytes from 5 donors (7 experiments) cultured with IL-2 (200 IU/mL). Symbols indicate mean of triplicate measurements. (C) CD25 and CD69 mean fluorescence intensity (MFI) in Jurkat after 24 hours with OKT3 or nonreactive mouse IgG2a. Bars indicate mean (±SD) of triplicate measurements. (D) Viable PEBL or control T lymphocytes recovered from cultures with OKT3 compared with cultures without OKT3, all containing IL-2 (200 IU/mL). Symbols represent mean (±SD) of 9 measurements with cells from 3 donors. P values by Student t test are shown for significant differences (****P < .0001). (E) Jurkat cells transduced with either a TCR specific for HBV s183 or a vector containing neomycin-resistant gene only (“NeoR”) were transduced with anti-CD3 PEBL or mCherry only (“Control”) after neomycin selection. CD3, TCRαβ, and TCRVβ3 chain (part of the HBV s183 TCR) expression is shown; TCRVβ3 expression was tested on the cell surface, and intracellularly after cell permeabilization. (F) Transduced Jurkat cells shown in panel E were cocultured with T2 cells loaded with HBV s183 peptide for 24 hours. Shown are CD25 and CD69 MFI minus those measured after culture with T2 cells, but without peptide. Symbols represent mean of triplicate measurements.

Figure 4.

CAR expression and signaling in T cells with CD3/TCRαβ expression blockade. (A) Flow cytometric dot-plots illustrate CD3 downregulation and anti-CD19-41BB-CD3ζ CAR expression. Cells were transduced with the CAR construct followed by anti-CD3ε PEBL, or with GFP only followed by mCherry only (“Control”). (B) Percentage of T lymphocytes transduced with PEBL or GFP alone (“Control”) expressing anti-CD19-41BB-CD3ζ CAR 24 hours after CAR mRNA electroporation (n = 5), or 5 to 6 days after CAR viral transduction (n = 4); P = .207. (C) IFNγ production by PEBL or control T cells electroporated with CAR mRNA or no mRNA and cultured with CD19⁺ RS4;11 for 8 hours at E:T 1:2. Bars represent mean (±SD) of 9 measurements with cells from 3 donors; ****P < .0001. (D) T lymphocytes were first transduced with CAR and then transduced with either mCherry alone or anti-CD3 PEBL. Cells were then cultured with irradiated CD19⁺ OP-1 for 3 weeks. Results were compared with cells transduced with GFP only and then with mCherry only (“Control”). Symbols indicate mean (± SD) percentage cell recovery relative to number of input cells in triplicate cultures.

Figure 5.

Cytotoxicity of CAR⁺PEBL T lymphocytes. (A) Four-hour cytotoxicity assays of PEBL or control (mCherry-transduced) T cells from 3 donors electroporated either with anti-CD19-41BB-CD3ζ CAR mRNA or no mRNA against CD19⁺ ALL cell lines at 2:1 E:T (see also supplemental Figure 4). Symbols indicate mean of 3 measurements for each donor. (B) Cytotoxicity of CAR-transduced T lymphocytes from 2 donors, sequentially transduced with a retroviral vector containing either mCherry alone or anti-CD3 PEBL was tested against CD19⁺ cell lines. Control, cells transduced with GFP only followed by mCherry only. Shown are data for 4-hour assays against CD19⁺ ALL cell lines at 1:1 E:T (full set of data in supplemental Figure 4). Each symbol indicates mean of triplicate experiments for each donor. (C-D) T lymphocytes transduced as in panel B were tested for long-term cytotoxicity against Nalm6 transduced with mCherry. Leukemia cell growth was measured with IncuCyte Zoom System (Essen BioScience). Whole-well imaging of triplicate cultures at 80 hours; E:T 1:8, is shown in panel C; leukemia cell growth measurements at the indicated E:T ratios in panel D. ***P < .001; ****P < .0001.

Figure 6.

CD3/TCRαβ knock-down by PEBL prevents GVHD. (A) NSG mice were irradiated with 2.5 Gy and IV injected 1 day later with 1 × 10⁷ T lymphocytes transduced with either anti-CD3 PEBL or GFP only (“Control”; n = 8 per group). Body weight is expressed as change relative to weight on day 3 after irradiation. (B) Hemoglobin levels and (C) platelet counts in peripheral blood. (D) Kaplan-Meier overall survival curves and log-rank test. Mice were euthanized when weight reduction exceeded 20% in 2 consecutive measurements (additional data in supplemental Figure 6 and Table 3). (E) Human CD45⁺ cell counts in blood 18 days after T-cell injection. *P = .0148; ***P < .001.

Figure 7.

T cells with CD3/TCRαβ knock-down by PEBL and CAR expression kill leukemia cells in mice. (A) NSG mice were IV injected with 5 × 10⁵ Nalm6-luciferase cells. Three days later, mice received 2 × 10⁷ T-lymphocytes transduced with anti-CD19-41BB-CD3ζ CAR plus either PEBL or mCherry alone; other mice received tissue culture medium instead (“no T cells”). Bioluminescence images on day 3 are shown with enhanced sensitivity to illustrate Nalm6 engraftment. (B) Symbols correspond to the average bioluminescence signal in ventral and dorsal imaging. (C) Kaplan-Meier curves and log-rank test for overall survival. Mice were euthanized when the ventral and dorsal bioluminescence average signal reached 1 × 10¹⁰ photons per second. ****P < .0001. (D) NSG mice were IV injected with 5 × 10⁵ Nalm6-luciferase cells and with 2 × 10⁷ T lymphocytes on day 3 as described in panel A. Before T lymphocytes injection, mice received 2.5 Gy total body irradiation. Bioluminescence images on day 3 are shown with enhanced sensitivity to illustrate Nalm6 engraftment. (E) Symbols correspond to bioluminescence average by ventral and dorsal imaging. (F) Kaplan-Meier curves and log-rank test for overall survival. Mice were euthanized when the ventral and dorsal bioluminescence average signal reached 1 × 10¹⁰ photons per second, or when signs of GVHD (>20% weight reduction exceeded in 2 consecutive measurements, with reduced mobility and/or fur loss) were evident. GVHD occurred in 3 of the 5 CAR⁺mCherry mice and 0 of the 6 CAR⁺PEBL mice; relapse (“Rel.”) rates were 0 of 5 vs 2 of 6, respectively. **P = .0014; ***P = .0006.