Human T cells engineered with an HLA-A2-restricted murine T-cell receptor targeting glypican 3 effectively control human hepatocellular carcinoma in mice

Abstract

Background and aims: Glypican-3 (GPC3) is a promising target for T-cell therapy in HCC. While chimeric antigen receptor (CAR) T cells targeting GPC3 have demonstrated therapeutic efficacy, their effectiveness is limited by challenges such as low persistence and shedding of surface GPC3. Natural T-cell receptors (TCRs) may serve as an alternative, though identifying GPC3-specific TCRs within the endogenous repertoire is difficult.

Approach and results: We immunized human leucocyte antigen-A2 (HLA-A2) transgenic mice with an adenovirus expressing human GPC3, identifying a panel of TCRs that recognize the GPC3(522-530) epitope. We cloned 3 murine GPC3-TCRs (TCR-A, TCR-B, and TCR-C) and engineered primary human T cells (TCR-T). TCR-T cells effectively recognized GPC3 + HLA-A2 + human HCC cells, with recognition diminished by GPC3 silencing and HLA-A2 blockade. TCR-B-T and TCR-C-T cells showed the highest reactivity, with TCR-B-T cells exhibiting superior effector functions, proliferative capacity, and therapeutic efficacy in xenograft HCC models. Notable, TCR-B-T cells outperformed second-generation 41BB GPC3-specific CAR-T cells, attributed to lower exhaustion, enhanced proliferation, greater effector function, and improved resilience. Furthermore, mixed dosing of CAR-T and TCR-B-T cells was significantly more effective than staggered dosing of the same cell type, suggesting potential synergistic effects.

Conclusions: Transgenic TCRs join forces with CARs, expanding the arsenal of GPC3-targeting receptors for HCC T-cell therapy.

Keywords: T-cell therapy; chimeric antigen receptor.

Graphical abstract

FIGURE 1

Identification of TCR specific for human GPC3 in HHD-DR1 mice. (A) Schematic procedure to identify immunodominant GPC3 T-cell epitopes. HHD-DR1 mice were immunized with ADV-GPC3. Eight days later, splenocytes were isolated for epitope screening. Initially, mixtures of synthetic GPC3 peptides covering the entire GPC3 protein were tested, followed by individual peptides from the reactive mixture in the first screening. Splenocytes from a naive mouse served as a negative control. The response was evaluated by intracellular staining for IFNγ and TNFα in CD8⁺ and CD4⁺ splenocytes. (B) Dot plots showing TNFα and IFNγ production by CD8 T lymphocytes upon stimulation with Mix 7, peptide 511, or medium (no Ag). Data from a representative mouse in the ADV-GPC3 group and a naive mouse are shown. Cells are gated on CD8⁺. (C) Graphs displaying compiled data from the first and second screenings. The percentage of IFNγ⁺ cells in CD8 and CD4 T lymphocytes stimulated with peptide mixes and individual peptides from Mix 7 is shown. The dotted line indicates the threshold for a positive response (3 times the % of IFNγ⁺ cells observed in the naive mouse). Data represent one experiment out of 3. (D) Experimental design for identifying HLA-A2-restricted murine TCRs specific for pGPC3(522-530). Splenocytes from ADV-GPC3-immunized mice were stimulated with pGPC3(522-530), and specific CD8 T cells were single-cell sorted (as IFNγ^high cells) and expanded. The 24 expanded clones were processed for TCR library preparation. (E) Distribution of CDR3α and CDR3β region lengths (residue numbers) and usage of TRAV, TRBV, TRAJ, TRBD (when identified), and TRBJ segments in the GPC3-TCR repertoire. Abbreviations: HHD-DR1, double transgenic mouse strain carrying both chimeric HLA-A*0201 (α1-α2) and H-2Db (α3, transmembrane, and intracytoplasmic domains), as well as HLA-DR1 molecules; NGS, NOD scid gamma; NI, nonidentified; TCR, T-cell receptor; TRAV, TCR alpha variable gene; TRAJ, TCR alpha joining gene; TRBV, TCR beta variable gene; TRBD, TCR beta diversity gene; TRBJ, TCR beta joining gene.

FIGURE 2

Evaluation of CD8 dependency, avidity, and functional avidity of the GPC3-specific TCRs. (A) Transduction efficiency was assessed by measuring the surface expression of murine TCRβ (mTCRβ). (B) Genetically modified human T cells were stained with a saturating concentration of pGPC3(522-530)/HLA-A2 tetramer. (A and B) Dot plots are gated on CD4 (top) or CD8 (bottom) T cells. Numbers indicate the percentage of mTCRβ⁺ cells (A) or tetramer⁺ cells (B) within CD4 or CD8 T cells. (C) Genetically modified T cells were co-cultured with T2 cells pulsed with a saturating concentration of pGPC3(522-530) (1 µg/mL) or medium. The graphs display the response as % of CD137⁺ cells within transduced (mTCRβ⁺) CD4 or CD8 T cells, or total CD4 or CD8 T cells (for UTD). Data are shown as mean ± SD (2 replicates per condition). (A–C) UTD: untransduced T cells. Data are representative of 1 experiment out of 3. (D) TCR avidity was assessed by staining CD8 TCR-T cells with decreasing concentrations of the pGPC3(522-530)/HLA-A2 tetramer and analyzing the % of tetramer⁺ cells within the mTCRβ⁺CD8⁺ population. (E) Functional avidity was assessed by culturing CD8 TCR-T cells with T2 cells pulsed with serial dilutions of pGPC3(522-530) and measuring surface CD137 expression, as well as the production of IFNγ, IL-2, and TNFα (by ELISA). (D and E) Nonlinear fit curves are shown, with numbers indicating the EC50 values. Data represent the mean of 3 independent experiments with different donors. compare the different TCRs, the % of TCR⁺ CD8 T cells in each TCR-T cell line was normalized by adding UTD CD8 T cells. Abbreviations: IFN, interferon; TCR, T-cell receptor; UTD, untransduced.

FIGURE 3

TCR-engineered human CD8 T cells recognize GPC3⁺HLA-A2⁺ tumor cells. (A) Relative expression of GPC3 versus HLA-A2 molecules in different tumor cell lines was used in this study. GPC3 expression is shown as reverse transcription-quantitative PCR CT values and HLA-A2 expression is shown as median fluorescence intensity (MFI). (B) Heat map showing the average % of CD137⁺ cells within mTCRβ⁺ CD8 T cells stimulated with GPC3⁺HLA-A2⁺ and GPC3⁺HLA-A2^- tumor cells. For comparison, the average % of CD137⁺ cells in UTD CD8 T cells cultured under similar conditions is also shown. Data are compiled from 3 experiments for TCR-A and 7 experiments for TCR-B, TCR-C, and UTD. (C) CD8 TCR-T cells were co-cultured (6 h) with GPC3-silenced tumor cell lines and their respective shREN-treated control cells. (D) GPC3⁺HLA-A2⁺ and GPC3⁺HLA-A2⁻ tumor cells were incubated with anti-pan-HLA-I, anti-HLA-A2, or control IgG (mouse IgG1) mAbs before co-culture with CD8 TCR-T cells. (C and D) The response was evaluated by measuring CD137 expression in mTCRβ⁺ cells (upper panels) and IFNγ release (lower panels). PLCPRF5 cells were used as a negative control for TCR-A and TCR-B, and HEP3B-A2 cells for TCR-C. Data are shown as mean ± SD (2 replicates per condition) and are representative of two experiments (C) or 3 experiments (D). To compare different TCRs, % of TCR⁺ CD8 T cells in each TCR-T cell line were normalized by adding UTD CD8 T cells. Abbreviations: HLA, human leukocyte antigen; TCR, T-cell receptors; UTD, untransduced.

FIGURE 4

Effector and proliferative capacity of GPC3 TCR-T cells in response to GPC3⁺HLA-A2⁺ tumor cell recognition. (A and B) CD8 TCR-T (5 × 10⁴ mTCRb⁺/well) cells from 6 donors were co-cultured with serially diluted tumor cells for 24 hours. CD137 expression was detected by FACS (A), and IFNγ in the supernatant was measured by ELISA (B). (A) Percentages of CD137⁺ cells within TCRβ⁺ CD8 T cells are shown. (C) TCR-T cells (5×10⁴ mTCRβ⁺ T cells/well) from 6 donors were co-cultured with luciferase-expressing target cells at varying E:T ratios for 10 hours, and % cytotoxicity was calculated as described in Experimental Procedures. (A–C) Data shown are from donor 1, with data for donors 2–6 available in Supplemental Figure S7, http://links.lww.com/HEP/J637. (D) CTV-labeled CD8 TCR-T (5 × 10⁴ mTCRb⁺/well) cells were cultured alone (medium) or with irradiated tumor cells (1.25 × 10⁴ cells/well) with or without anti(α)-pan-HLA-I mAbs. IL-2 (10 U/mL) was added 48 hours poststimulation. After 96 hours, cells were collected, stained with anti-mTCRβ mAb, and analyzed by FACS. The graph shows the % of proliferating T cells (CTV^low) within mTCRβ⁺ CD8 T cells for donor 1. (E) Effector functions and proliferation of GPC3 TCR-T cells across donors in response to GPC3⁺HLA-A2⁺ tumor cells. For each donor, percentages of TCR+ CD8 T cells were normalized using autologous UTD CD8 T cells. (A-E) Data are shown as mean ± SD. (E) Statistical significance was calculated using Two-way ANOVA. ****p<0.0001. Abbreviations: CTV, CellTrace Violet; IFN, interferon; TCR, T-cell receptor.

FIGURE 5

TCR-B-T cells efficiently controlled tumor progression in ACT settings. (A–C) 6- to 8-week-old female NSG mice were subcutaneously (sc) implanted with PLCPRF5-A2 (B) or HEPG2 (C) cells into the right flank. Nine (B) or 12 (C) days later, mice were treated with human TCR-B- or TCR-C-T cells containing 8×10⁶ mTCRβ⁺ cells injected i.v. (N=7 mice/group). The % of mTCRβ⁺ CD8 T cells was equalized between samples by adding autologous UTD CD8 T cells. Control groups received a number of UTD CD8 T cells similar to the total number of CD8 T cells in the TCR-T-treated groups from the same donor (N=6–7 mice/group). Mice received human IL-2 i.p. on days 1, 2, 3, 5, 7, 9, 11, 13, and 15, and sc (near the tumor site) on days 3, 5, 7, 9, 11, 13, and 15 post-ACT. (B and C) Left: Average tumor size (mm²). Right: Overall survival. Data are represented as mean as mean ± SEM (left). Statistical significance was determined using nonlinear regression (curve fit) (left) and Mantel-Cox test (right). ***p<0.0005, *p<0.05. One experiment representative of 2 is shown. Abbreviations: ACT, adoptive T-cell therapy; NSG, NOD scid gamma; UTD, untransduced.

FIGURE 6

TCR-B-T cells exhibited enhanced engraftment capacity and ability to mount a vigorous response after ACT. (A) PLCPRF5-A2 tumor-bearing mice were treated as in Figure 5, and blood and tumor samples were analyzed. (B and C) Blood analysis. On days 2, 6, and 9, blood samples were collected (N=6 mice/group) and analyzed by FACS to assess the % of mTCRβ⁺ cells in the human (h) CD8⁺ population (B) and the absolute number of mTCRβ⁺ T cells and total hCD8 in 50 μL of blood (C). (D–F) Tumor analysis. On day 10, TCR-T (N=5 mice/group) or UTD (N=4 mice/group) treated mice were sacrificed, and the tumor cell suspension was analyzed by FACS. (D) Representative dot plots showing the % of human CD8 T cells within total living cells in the tumor on day 10 post-ACT. Dot plots are gated on living total cells. The graph shows the % relative to the UTD group. (E) Representative dot plots and graphs showing the % of mTCRβ⁺ cells within human CD8⁺ cells in the infusion product (input) and on day 10 in human CD8⁺ TILs. The dotted lines in the graph show the % of mTCRβ⁺ cells in the input of TCR-B- and TCR-C-T cells. (F) Representative dot plots and graphs showing the % of GzmB⁺ (left) and Ki-67 ⁺ (right) cells within mTCRβ⁺ CD8⁺ TILs. (E and F) Dot plots are gated on living CD8⁺ cells. (B–F) Data in graphs are represented as mean as mean ± SEM (B and C) or median (D–F). Symbols represent individual mice (D–F). Statistical significance was determined using one-way ANOVA (D) and paired t test (B, C, E, and F). ***p<0.0005, **p<0.005, *p<0.05. One experiment representative of 2 is shown. Abbreviations: NSG, NOD scid gamma; TCR, T-cell receptor; TIL, tumor-infiltrating T lymphocyte; UTD, untransduced.

FIGURE 7

TCR-B-T cells outperformed YP7-based CAR-T cells in ACT schedules. (A) Enrichment of CAR-GPC3⁺ and mTCRβ⁺ cells. Human CD8 T cells were transduced with retroviruses coding for CAR-GPC3 or TCR-B. On day 4, they were labeled with GPC3-Fc chimera protein plus APC-labeled anti-Fc mAb or APC-labeled anti-mouse TCRβ mAb, respectively, and CAR-GPC3⁺ and mTCRβ⁺ cells were isolated by FACS and expanded. (B) Surface GPC3 protein levels were assessed in PLCPRF5-A2 and HEPG2 cells by staining with anti-human GPC3 mAb (clone YP-7), or a control isotype mAb (Isot) and FACS analysis. (C) CAR-GPC3⁺ and mTCRβ⁺ T cells were co-cultured with serial dilutions of tumor cells. Cells were stained to detect CD137 on CD8 T cells by FACS. The % of CD137⁺ cells within CD8 T cells is shown. (D) T-cell-mediated HEPG2 killing was measured by a real-time impedance-based assay using CAR-GPC3⁺, mTCRβ⁺, or UTD cells. Normalized cell index data for each culture condition (left) and % of cytotoxicity at 4 and 16 hours from the addition of T cells (right) are shown. (E) 6- to 8-week-old female NSG mice were sc implanted with HEPG2. Twelve days later, mice were treated with 8×10⁶ CAR-GPC3⁺ or mTCRβ⁺ cells injected i.v. Mice received human IL-2 as described in Figure 5. Left: Experimental design. Middle: Graph of average tumor size (mm²) with a zoom region inset from the first 11 days of treatment. Right: Overall survival. (C–E) Data are represented as mean ± SD (C and D right), mean ±SEM (E, middle) or mean (D left). Two (C and D) or 7 (E) replicates per condition. Statistical significance was determined using nonlinear regression (curve fit) (E, middle) and Mantel-Cox test (E, right). ***p<0.0005, **p<0.005. One experiment representative of 2 (B–E) or 6 (A) is shown. Abbreviations: ACT, adoptive T-cell therapy; APC, Allophycocyanin; CAR, chimeric antigen receptor; CR, complete response; GPC3, Glypican-3; LT, liver transplantation; NSG, NOD scid gamma; TCR, T-cell receptor; UTD, untransduced.

FIGURE 8

TCR-B-T cells exhibited lower levels of exhaustion and enhanced response after transfer, compared to CAR-T cells. HEPG2 tumor-bearing mice were treated as in Figure 7E and blood and tumor samples were analyzed. (A) Blood analysis. On days 2, 6, and 12, blood samples were collected, and the absolute number of human CD8 T cells in 50 μL of blood was assessed by volumetric flow cytometry. (B–F) On day 13, mice were sacrificed, and the tumor cell suspension was stained and analyzed by FACS. (B) Representative dot plots and graphs showing the % of human CD8 T cells within total living cells in the tumor. Dot plots are gated on living total cells. The graph shows the % relative to the UTD group. (C–E) Representative dot plots, histograms, and graphs showing the percentage of GzmB⁺ (C) and Ki-67⁺ (D) cells, and PD-1 MFI (E) in CD8 TILs. (F) Representative histograms and graphs showing the MFI of surface GPC3 (left) and HLA-A2 (right) in CD45^−​FSC^hiSSC^hi cells from the tumor cell suspension. (A–F) Data are depicted as mean ± SEM in A (4–5 mice) and median (B–F). Symbols represent individual mice (B–F). Statistical significance was determined using two-way ANOVA (A–F); ****p<0.0001, ***p<0.0005, **p<0.005, *p<0.05. One experiment is representative of 2 experiments. Abbreviations: ACT, adoptive T-cell therapy; APC, Allophycocyanin; CAR, chimeric antigen receptor; GPC3, Glypican-3; MFI, median fluorescence intensity; TCR, T-cell receptor; TIL, tumor-infiltrating T lymphocyte; UTD, untransduced.