FVIII-specific human chimeric antigen receptor T-regulatory cells suppress T- and B-cell responses to FVIII

Jeongheon Yoon, Anja Schmidt, Ai-Hong Zhang, Christoph Königs, Yong Chan Kim and David W. Scott

Article Figures & Data


  • Figure 1.

    Generation of FVIII-specific CAR T cells. (A) Schematic view of transgenes used. FVIII-specific ANS8 CAR, containing the FVIII-specific scFv 1G10 in comparison with the empty Mock control. (B) Retrovirally transduced Mock or ANS8 CAR Tregs were incubated with biotinylated FVIII (1.8 μg/mL) or OVA (1.8 μg/mL). Cells were stained with PE-conjugated streptavidin and analyzed by FACS. GFP+ ANS8 CAR Tregs showed specific binding to biotinylated FVIII. (C) Transduced, cell proliferation dye-labeled ANS8 CAR T-effector cells were incubated with FVIII or FVIII in combination with irradiated autologous PBMCs for 6 days and analyzed for cell proliferation. ANS8 CAR–transduced cells strongly proliferated in the presence of FVIII and PBMCs. The histogram shows viable CD4+ cells. Similar results were obtained in independent experiments from at least 3 different healthy T-cell donors.

  • Figure 3.

    FVIII-specific and bystander immunosuppression by ANS8 CAR Tregs. Sorted human Treg and T-effector cells were transduced with Mock, ANS8 CAR, 17195 TCR, or Ob2F3 TCR, expanded, and rested. (A) Cell proliferation dye (CPD)-labeled 17195 T-effector cells were cocultured with different numbers of Mock or ANS8 CAR Tregs in the presence of FVIII protein (0.4 μg/mL) for 4 days. ANS8 CAR Tregs suppressed FVIII-specific effector T-cell proliferation. (B) CPD-labeled Ob2F3 T effectors were cocultured with different numbers of either ANS8 CAR or 17195 TCR Tregs in the presence of FVIII protein and pMBP for 4 days. ANS8 CAR and 19175 TCR Tregs suppressed MBP-specific effector T-cell proliferation but only when both pMBP and FVIII were present. This figure shows representative cell proliferation data of 17195 TCR (A) or Ob2F3 TCR (B) T-effector cells in the presence of the different Treg populations (left). Graphs show absolute counts of divided GFP+ T-effector cells after coculture with differently engineered Tregs (right, dotted line and closed circle, mock Tregs; closed square, ANS8 CAR8 Tregs; open square, 17195 TCR Tregs). Data indicate mean values for triplicates ± standard error of the mean (SEM). The experiment was repeated with separately transduced T cells from different donors with similar results.

  • Figure 2.

    CAR-mediated stimulation of FVIII-specific CAR Tregs. Sorted human Tregs were transduced with ANS8 CAR or Mock, expanded in the presence of IL-2, rested, and then stimulated with FVIII. Gating strategy for GFP+ and GFP cells is shown on the left. Intracellular staining for LAP and GARP in GFP+ cells was increased in the presence of FVIII but not in the presence of OVA. Foxp3 expression is increased in the ANS8 CAR GFP+ population in the presence of FVIII. Data are representative of 3 independent experiments.

  • Figure 4.

    In vitro suppression of antibody production by FVIII-specific CAR Tregs. Whole splenocytes isolated from E16×DR1 mice with high antibody titers (measured by ELISA) were cultured in the presence of FVIII for 6 days and antibody-secreting cells were enumerated by ELISPOT assay. Bars indicate mean ± SEM. (A) FVIII-specific antibody-secreting cells were detected in splenocyte cultures only in the presence of FVIII. (B) Coculture with either ANS8 CAR or 17195 TCR Tregs (but not Mock Tregs) inhibited formation of FVIII-specific antibody-secreting cells. The experiment was repeated with separately transduced Tregs from different donors with similar results.

  • Figure 5.

    Xenogeneic suppression of anti-FVIII antibody response in vivo by human ANS8 CAR Tregs. (A) Experimental schema. Female E16×DR1 mice (n = 5 per group) were subcutaneously immunized with FVIII in incomplete Freund’s adjuvant on day 0. Four hours after immunization, the mice were adoptively transferred with GFP+ Tregs expressing ANS8 CAR, 17195 TCR, or control Ob2F3 TCR. The anti-FVIII antibody levels were followed weekly after the immunization (dotted arrows). Mice received one additional challenge with FVIII, together with 10% TNP-SRBC to test antigen-specificity of tolerance on day 56 (solid arrows). (B) Time course of anti-FVIII antibody response. Data indicate mean ± SEM. Multiple Student t tests were performed to determine statistical significance. *P < .05, **P < .01.

  • Figure 6.

    Model of FVIII-specific human CAR Treg function. Cartoon depicting suppressive function of FVIII-specific CAR-engineered human Tregs to FVIII-specific TCR-engineered T-effector cells or antibody-producing B cells in response to FVIII. FVIII-loaded APCs might bring T-effector cells and Tregs into close proximity. Thus, activated Tregs can inhibit T-effector cell activation in this local milieu (eg, by IL-2 consumption). Inhibition of T-effector cells also leads to prevention of FVIII inhibitor formation as costimulatory signals of T-effector cells are essential for B-cell activation. In addition, activated Tregs might influence B cells directly by still unknown mechanisms.