Switch-mediated activation and retargeting of CAR-T cells for B-cell malignancies

Abstract

Chimeric antigen receptor T (CAR-T) cell therapy has produced impressive results in clinical trials for B-cell malignancies. However, safety concerns related to the inability to control CAR-T cells once infused into the patient remain a significant challenge. Here we report the engineering of recombinant antibody-based bifunctional switches that consist of a tumor antigen-specific Fab molecule engrafted with a peptide neo-epitope, which is bound exclusively by a peptide-specific switchable CAR-T cell (sCAR-T). The switch redirects the activity of the bio-orthogonal sCAR-T cells through the selective formation of immunological synapses, in which the sCAR-T cell, switch, and target cell interact in a structurally defined and temporally controlled manner. Optimized switches specific for CD19 controlled the activity, tissue-homing, cytokine release, and phenotype of sCAR-T cells in a dose-titratable manner in a Nalm-6 xenograft rodent model of B-cell leukemia. The sCAR-T-cell dosing regimen could be tuned to provide efficacy comparable to the corresponding conventional CART-19, but with lower cytokine levels, thereby offering a method of mitigating cytokine release syndrome in clinical translation. Furthermore, we demonstrate that this methodology is readily adaptable to targeting CD20 on cancer cells using the same sCAR-T cell, suggesting that this approach may be broadly applicable to heterogeneous and resistant tumor populations, as well as other liquid and solid tumor antigens.

Keywords: antibody engineering; autologous cell therapy; cancer; chimeric antigen receptor T cell; leukemia.

Fig. 1.

Relative orientations and in vitro activity of switches. (A) Schematic depicting the immunological synapse between the sCAR, switch, and antigen (Left); ribbon diagrams depicting the different switch designs (Right), where the light chain is shown in gold, heavy chain in green, and PNE in blue; the rigid linker on HCNT is shown in gray. The diagrams are oriented such that the antigen-binding interface is at the top. (B) Cytotoxicity of sCAR-T cells against CD19⁺ RS4;11 cells with titration of anti-CD19 Fab switch molecules. Corresponding IgG designs are shown in SI Appendix, Fig. S5A. (C) Table of EC₅₀ values of cytotoxicity from B. (D) Cytotoxicity of sCAR-T cells against CD19⁻ K562 cells with the anti-CD19 Fab switch molecules. Corresponding IgG switches are shown in SI Appendix, Fig. S5B. (E) IL-2 released from sCAR-T cells cultured with RS4;11 or K562 and 1-nM anti-CD19 Fab switches. Gray asterisks denote significance compared with wild-type switch and black asterisks denote significance between switches. (F) Cytotoxicity of sCAR-T cells against CD20⁺ Raji cells with the anti-CD20 Fab switch molecules. (G) IL-2 released from sCAR-T cells cultured with Raji cells and 1 nM anti-CD20 Fab switches. All assays were performed with effector to target (E:T) = 10:1 and 24 h incubation. Cytotoxicity was assessed by LDH release. Data are represented as mean ± SD. Where listed, values next to switches show EC₅₀ of cytotoxicity. Significance was calculated using the Student’s t test, where **P < 0.01 and ***P < 0.001.

Fig. 2.

Activity of sCAR hinge designs. (A) Cytotoxicity of sCAR-T cells with CD8, IgG4, or IgG4m hinges, or nontransduced T cells (mock) against CD19⁺ RS4;11 cells with titration of the anti-CD19 LCNT switch molecule. (B) IL-2 released from sCAR-T cells cultured with RS4;11 cells and 1-nM anti-CD19 Fab switches; gray asterisks denote significance between CART-19 and sCAR-T cells and black asterisks denote significance between sCAR–T-cell hinges; mean values for CART-19 is indicated by dotted horizontal line. (C) Quantification of the T-cell activation assay in the presence of 1 nM of each anti-CD19 Fab switch. (D) Cytotoxicity as described in A against CD19⁻ K562 cells. (E) Activation of sCAR-T cells with IgG4m hinge or mock cells in the presence of RS4;11 cells with or without 1-nM LCNT Fab switch measured by flow cytometry with staining for CD25 and CD69. (F) Cytotoxicity of sCAR-T cells against CD20⁺ Raji cells with titration of anti-CD20 Fab switch molecules. Values listed next to switches show EC₅₀ of cytotoxicity. All cytotoxicity and cytokine assays were performed with E:T = 10:1 and 24 h incubation. Cytotoxicity was assessed by LDH release. Data are represented as mean ± SD, and statistical significance was calculated using the one-way ANOVA with Tukey’s posttest (B and D), where *P < 0.05, **P < 0.01, and ***P < 0.001.

Fig. 3.

In vivo activity of switch and hinge designs in Nalm-6 xenograft model. NSG mice were inoculated with CD19⁺ Nalm-6 cells. After 6 d the mice were injected with luciferin and imaged on an IVIS before being randomly sorted into groups (n = 5) with representative tumor burden. Next, 40 × 10⁶ sCAR-T cells with a transduction efficiency of 50–75% were infused intravenously and switch-dosing commenced every other day at 0.5 mg/kg for 10 d. Tumor burden was followed by IVIS. (A) Representative IVIS images depicting three of the mice from each of the groups treated with the CD8, IgG4, or IgG4m sCAR-T cells and LCNT Fab switch are shown. (B) Quantified tumor burden (as average radiance from luciferase activity from each mouse) from A during switch-dosing period (n = 5). (C) Enumeration of T cells in peripheral blood from A at day 17 by flow cytometry using CountBright Beads (Thermo). (D) Representative IVIS images depicting three of the mice from each of the groups treated with the IgG4m sCAR-T cells and the anti-CD19 Fab switch designs indicated. (E) Quantified tumor burden from D during the switch-dosing period (n = 5). (F) Enumeration of T cells in peripheral blood from D at day 20 by flow cytometry as in C. Asterisks indicate significance from LCNT Fab group. Data are represented as mean ± SEM, and statistical significance was calculated using the two-way ANOVA with Bonferroni’s posttest (B and E) (shown for final time-point only) or by one-tailed Student’s t test (C and F), where *P < 0.05, **P < 0.01, and ***P < 0.001; ns, not significant.

Fig. 4.

sCAR–T-cell expansion and localization during tumor clearance. NSG mice were inoculated with Nalm-6 as described in Fig. 3. (A) After 6 d, mice were sorted into four groups. Three groups containing five cohorts of mice with representative tumor burden before being injected intravenously with 40 × 10⁶ CART-19, IgG4m sCAR-T cells with (+) LCNT Fab switch, or no T cells (PBS group). A fourth group consisted of two cohorts (n = 3) of mice injected intravenously with IgG4m sCAR-T cells without (−) LCNT Fab switch were analyzed at 8 h and 96 h only. LCNT Fab dosing (0.5 mg/kg) in the (+) group was started with initial T-cell infusion and continued daily for 5 d. Luminescence was measured at 8 h and subsequently every 24 h, as indicated. All cells were labeled with eFluor 450 cell proliferation dye before injection. (B) Quantified tumor burden (radiance, left axis, solid line) and quantified human IL-2 (pg/mL, right axis, dashed line) in peripheral blood from each cohort shown in A at each time point. Gray lines indicate tumor burden in PBS and IgG4m sCAR–T-cell −LCNT Fab controls. (C) The proportion of leukocytes that were human CD3⁺CD4/8⁺ in the blood, spleen, lungs and liver at 96 h in the +LCNT and −LCNT groups was quantified by flow cytometry. (D) Relative proliferation of T cells in the spleen, lung, and liver at 96 h in +LCNT and −LCNT Fab groups, determined by eFluor 450 dye dilution shown as 1/mean fluorescence intensity. Data are represented as means ± SEM, and statistical significance was calculated using the one-tailed Student’s t test (C and D), where *P < 0.05, **P < 0.01, and ***P < 0.001.

Fig. 5.

Influence of switch dosing regimen on sCAR–T-cell activation and efficacy. NSG mice (n = 3) were inoculated with Nalm-6 and 6 d later were engrafted with IgG4m sCAR-T cells (transduction efficiency and CD4:CD8 ratio of injected cells: 60%, 1:1.23) as described in Fig. 3. (A) Quantified tumor burden from groups dosed every day (q.d.), every other day (q.a.d.), or every fifth day (q.5.d.) with 0.5 mg/kg LCNT. Dosing was carried out for 15 d starting at day 6 (corresponding with IgG4m sCAR–T-cell infusion), as indicated by gray shading (day 6–21). The NS group received IgG4m sCAR-T cells and injections of PBS. Open circles denote the death of individual mice without evidence of tumor burden and likely reflect graft vs. host disease. Triangles in the q.5.d. group indicates mouse expiration because of increasing tumor burden (n = 3). (B) Levels of IL-2, IFN-γ, and TNF-α at 24 h after first dose, from groups dosed every day with 0, 0.05, 0.5, or 2.5 mg/kg of LCNT. CART-19 is labeled as “(19)” (n = 5). (C) Quantified tumor burden by IVIS from groups in B after 10 d. (D) Quantified tumor burden from mice dosed with 0.05 mg/kg LCNT for 10 d indicated by gray shading (day 6–16). Dosing was resumed at 0.5 mg/kg when the average tumor burden in group became significantly higher than CART-19 control at day 30, indicated by an asterisk and second gray shading (day 30–40). Triangles indicate the death of individual mice with significant tumor burden (n = 5). Transduction efficiency and CD4:CD8 ratio of injected cells: CART19 = 68%, 1:1.63 and sCAR-T = 75%, 1:1.89. (E) The phenotype of T cells in the peripheral blood from (B and C) after 21 d as the proportion of CD3⁺CD4/CD8⁺ by flow cytometry with staining for CD62-L and CD45RA (n = 5). Statistical significance was calculated using the one-tailed Student’s t test (B, C, and E), where *P < 0.05, **P < 0.01, and ***P < 0.001; NS, not significant.