A chimeric antigen receptor-based cellular safeguard mechanism for selective in vivo depletion of engineered T cells

Abstract

Adoptive immunotherapy based on chimeric antigen receptor (CAR)-engineered T cells has exhibited impressive clinical efficacy in treating B-cell malignancies. However, the potency of CAR-T cells carriethe potential for significant on-target/off-tumor toxicities when target antigens are shared with healthy cells, necessitating the development of complementary safety measures. In this context, there is a need to selectively eliminate therapeutically administered CAR-T cells, especially to revert long-term CAR-T cell-related side effects. To address this, we have developed an effective cellular-based safety mechanism to specifically target and eliminate the transferred CAR-T cells. As proof-of-principle, we have designed a secondary CAR (anti-CAR CAR) capable of recognizing a short peptide sequence (Strep-tag II) incorporated into the hinge domain of an anti-CD19 CAR. In in vitro experiments, these anti-CAR CAR-T cells have demonstrated antigen-specific cytokine release and cytotoxicity when co-cultured with anti-CD19 CAR-T cells. Moreover, in both immunocompromised and immunocompetent mice, we observed the successful depletion of anti-CD19 CAR-T cells when administered concurrently with anti-CAR CAR-T cells. We have also demonstrated the efficacy of this safeguard mechanism in a clinically relevant animal model of B-cell aplasia induced by CD19 CAR treatment, where this side effect was reversed upon anti-CAR CAR-T cells infusion. Notably, efficient B-cell recovery occurred even in the absence of any pre-conditioning regimens prior anti-CAR CAR-T cells transfer, thus enhancing its practical applicability. In summary, we developed a robust cellular safeguard system for selective in vivo elimination of engineered T cells, offering a promising solution to address CAR-T cell-related on-target/off-tumor toxicities.

Keywords: B cell aplasia 2; chimeric antigen receptor; on-target/off-tumor; safeguard mechanism; side effects.

Figure 1

Design and in vitro functionality of anti-CAR CAR-T cells. (A) Schematic of the CAR sequences for retroviral transduction. (B) Schematic of the mechanism of action of anti-CAR CAR-T cells. (C) Cell surface expression of CAR constructs after retroviral transduction measured by EGFRt co-expression. (D, E) 5x10⁴ anti-CAR CAR-T cells were co-cultured with B cell-depleted anti-CD19 CAR-T cells or untransduced cells at the indicated E:T ratios for 4 h. Representative flow cytometry plots (E:T 1:1) (D) and quantification (F) of intracellular staining of IFN-γ, TNF-α and IL-2. Pregated on CD3⁺ EGFRt⁺ living lymphocytes. (F, G) 4x10⁴ anti-CAR CAR-T cells (red: CD45.1⁺ EGFRt⁺) were co-cultured with B cell-depleted bulk anti-CD19 CAR-T cells (blue: CD45.1^- EGFRt⁺; gray: CD45.1^- EGFRt^-) at the indicated E:T ratios and analyzed at 0, 24 and 48 h. Representative flow cytometry plots (G) and quantification (H) of the co-culture killing assay. Pregated on CD3⁺ living lymphocytes. Data are shown as mean + SD in (C) and mean ± SD in (E, G). In (C), statistical analyses have been performed by Mann-Whitney nonparametric test. In (G) statistical analyses have been performed by two-way ANOVA with Dunnett’s multiple comparisons test using 0 h as reference. ns = p > 0.05, *p < 0.05, **p < 0.01.

Figure 2

anti-CAR CAR-T cells display in vivo killing in immunocompromised mice in a model of acute antigen encounter. (A) Schematic of the experimental setup. (B, C) Representative flow cytometry plots (B) and quantification (C) of anti-CD19 CAR-T cells (EGFRt⁺CD90.2⁺) in the peripheral blood at the indicated time points and in tissues in the presence or not of anti-CAR CAR-T cells. Data are shown as mean ± SD. In (C), statistical analyses were performed by unpaired t-test. ns = p > 0.05, **p < 0.01, ***p < 0.001.

Figure 3

anti-CAR CAR-T cells can restore B cells in CD19 CAR-treated immunocompetent mice. (A) Schematic of the experimental setup. (B) Representative flow cytometry plots of anti-CD19 CAR-T cells (EGFRt⁺CD90.1⁺) and anti-CAR CAR-T cells (EGFRt⁺CD45.1⁺) at day 6 post infusion according to the different treatment schemes. Cells are pregated on living CD3⁺ T cells. (C, D) Quantification of the frequencies of anti-CD19 CAR-T cells (C) and anti-CAR CAR-T cells (D) in blood over time. (E) Quantification of tissue-infiltrating anti-CD19 and anti-CAR CAR-T cells. (F) Representative flow cytometry plots of living CD19⁺ B cells according to the different treatment schemes at the indicated time points. (G) Kinetics of B cells in blood over time. Data are shown as mean + SD. In (C) and (G) statistical analyses were performed by two-way ANOVA with Sidak’s multiple comparisons test (C) and Dunnett’s multiple comparisons test using the anti-CAR group as reference (G). In (E) statistical analyses were performed with non-parametric Mann-Whitney t test. ns = p > 0.05, *p < 0.05, **p < 0.01, ***p < 0.001, ****p > 0.0001.

Figure 4

Reliable B-cell reconstitution after therapeutic anti-CAR application also in the absence of pre-conditioning treatment. (A) Schematic of the experimental setup (n=3). (B) Representative flow cytometry plots of B cells one week prior (upper) and six weeks after (lower) anti-CAR CAR-T cell transfer. (C, D) Quantification of B cells (C) and anti-CD19 CAR-T cells (D) in blood. (E, F) Absolute numbers of B cells (E) and anti-CD19 CAR-T cells (F) in secondary lymphoid organs at the endpoint. Data are shown as mean + SD. In (C, D), statistical analyses have been performed by two-way Anova with Dunnett’s multiple comparisons test using the mock group as reference. In (E, F), statistical analyses have been performed by Kruskal-Wallis test with Dunnett’s multiple comparisons test using the mock group as reference. ns = p > 0.05, *p < 0.05, ***p < 0.001.

Figure 5

anti-CAR CAR-T cells persist also in absence of pre-conditioning. (A) Representative flow cytometry plots of anti-CAR CAR-T cells one week (upper) and six weeks after (lower) anti-CAR treatment. (B, C) Quantification of anti-CAR CAR-T cells in blood (B) and lymphoid tissues at the endpoint (C). Data are shown as mean+SD. Grey area indicates the detection limit. In (B) statistical analyses have been performed by two-way Anova with Dunnett’s multiple comparisons test using the 4x10⁶ cells group as reference. In (C) statistical analyses have been performed by Kruskal-Wallis test with Dunnett’s multiple comparisons test using the ‘4x10⁶ cells’ group as reference. ns = p > 0.05, **p < 0.01.