Bispecific T Cell Engager with Trbc-Targeting Antibody Against Effector T Cells for the Treatment of T Cell Malignancy

Introduction: Immunotherapy for T-cell leukemias/lymphomas (TCLL) is not clinically available unlike B cell malignancies. In particular, bispecific T-cell engagers (BTCE) have been poorly developed because malignant cells express surface CD3, which makes anti-CD3 antibodies unavailable. Recently, Maciocia et al. (Nature Medicine 23; 1416: 2017) generated T cell receptor β-chain constant region (TRBC)1-targeted CAR-T cells to TRBC1-positive T-cell malignancies, which eradicate not only TRBC1(+) malignant clones but also approximately half of normal T lymphocytes. A phase I/II study on TRBC1-targeted CAR-T cells, unexpectedly, revealed the disappointing outcomes. In this study, we generated BTCE using an anti-TRBC2 antibody against normal effector T cells but not against malignant cell targets. We verified the hypothesis that BTCE with an anti-TRBC2 antibody would work well against TRBC1(+) malignant cells.

Methods: Anti-TRBC1 (human Jovi-1, Kd=2.6E-9 M) and anti-TRBC2 (KFN human Jovi-1, Kd=4.8E-7 M) antibodies, with agonistic activity, were produced as described by Ferrari et al. (Nat Commun 15; 1583: 2024). CD1a was selected as the tumor-associated antigen. BTCE was produced using Fabs of anti-CD1a antibody and anti-CD3 (or anti-TRBC1 or anti-TRBC2) antibodies and Fc (BTCE-CD3, BTCE-TRBC1, BTCE-TRBC2 respectively). CD1 expressions in T-ALL derived CD1(+) cell lines (HPB-ALL, JM and CCRF-CEM) were similar before and after CD3 (or TRBC1 or TRBC2) knock out (KO). Cytotoxicity was evaluated by co-culturing BTCE and calcein-labelled cell line with T-LAK for 1.5~2 h. For the in vivo study, human PBMCs and TRBC1(+) JM-luciferase cells were injected to immunodeficient NOD/Shi-scid, IL-2RγKO Jic（NOG）mice on day -10 and 0 respectively. BTCE-TRBC2 was injected twice weekly from days 1 to 25. Bioluminescence imaging was performed on days 14 and 28.

Results and Discussion: BTCE-CD3 induced a typical dose-dependent sigmoid curve in cytotoxicity against CD3 KO and CD3(+) HPB-ALL cells (E/T ratio: 10, n=3, EC50: 0.69 vs 66.0 pM; Emax: 78.5 vs 48.6 % respectively). EC50 and Emax against CD3(+) HPB-ALL cells were 96-fold higher and 1.6-fold lower respectively with significant difference (P=0.044, P=0.007 respectively) than those against CD3 KO HPB-ALL cells. BTCE-TRBC1 induced a similar sigmoid curve in cytotoxicity against TRBC1 KO and TRBC1(+) JM cells (E/T ratio: 10, n=3, EC50: 5.4 vs 20.0 pM; Emax: 107.8 vs 97.8 %, respectively). EC50 against TRBC1(+) JM cells was 4-fold higher with significant difference (P=0.012) than that against JM TRBC1 KO cells. These results indicate that a portion of BTCE-CD3 or BTCE-TRBC1 may be dysfunctional BTCEs that induce binding among cell lines via BTCEs, resulting in a significantly higher EC50 and lower Emax. Then, we evaluated the cytotoxicity of TRBC1(+) malignant cells by BTCE-TRBC2. BTCE-TRBC2 induced a typical dose-dependent sigmoid curve in cytotoxicity, with no significant difference, between TRBC1 KO and TRBC1(+) JM cells (E/T=10, n=3, EC50: 17.6 vs 14.2 pM; Emax: 87.0 vs 99.2 %, respectively). To assess the possibility of BTCE-TRBC1 against TRBC2(+) malignant cells, we evaluated the cytotoxicity of TRBC2(+) CCRF-CEM by BTCE-TRBC1. BTCE-TRBC1 exhibited similar and effective cytotoxicity, with no significant difference, between TRBC2 KO and TRBC2(+) CCRF-CEM cells (E/T=10, n=3, EC50: 20.7 vs 21.9 pM; Emax: 68.6 vs 69.3 %, respectively). In the in vivo study, bioluminescence imaging demonstrated marked reduction of tumor burden in mice treated with BTCE-TRBC2 on days 14 and 28.

Conclusion: These results strongly suggest that BTCE-TRBC2 functions effectively against TRBC1(+) tumor cells and vice versa. This concept is applicable to BTCE therapy for T cell malignancies.