Targeting refractory/recurrent neuroblastoma and osteosarcoma with anti-CD3×anti-GD2 bispecific antibody armed T cells

Abstract

Background: The survival benefit observed in children with neuroblastoma (NB) and minimal residual disease who received treatment with anti-GD2 monoclonal antibodies prompted our investigation into the safety and potential clinical benefits of anti-CD3×anti-GD2 bispecific antibody (GD2Bi) armed T cells (GD2BATs). Preclinical studies demonstrated the high cytotoxicity of GD2BATs against GD2+cell lines, leading to the initiation of a phase I/II study in recurrent/refractory patients.

Methods: The 3+3 dose escalation phase I study (NCT02173093) encompassed nine evaluable patients with NB (n=5), osteosarcoma (n=3), and desmoplastic small round cell tumors (n=1). Patients received twice-weekly infusions of GD2BATs at 40, 80, or 160×10⁶ GD2BATs/kg/infusion complemented by daily interleukin-2 (300,000 IU/m²) and twice-weekly granulocyte macrophage colony-stimulating factor (250 µg/m²). The phase II segment focused on patients with NB at the dose 3 level of 160×10⁶ GD2BATs/kg/infusion.

Results: Of the 12 patients enrolled, 9 completed therapy in phase I with no dose-limiting toxicities. Mild and manageable cytokine release syndrome occurred in all patients, presenting as grade 2-3 fevers/chills, headaches, and occasional hypotension up to 72 hours after GD2BAT infusions. GD2-antibody-associated pain was minimal. Median overall survival (OS) for phase I and the limited phase II was 18.0 and 31.2 months, respectively, with a combined OS of 21.1 months. A phase I NB patient had a complete bone marrow response with overall stable disease. In phase II, 10 of 12 patients were evaluable: 1 achieved partial response, and 3 showed clinical benefit with prolonged stable disease. Over 50% of evaluable patients exhibited augmented immune responses to GD2+targets post-GD2BATs, as indicated by interferon-gamma (IFN-γ) EliSpots, Th1 cytokines, and/or chemokines.

Conclusions: This study demonstrated the safety of GD2BATs up to 160×10⁶ cells/kg/infusion. Coupled with evidence of post-treatment endogenous immune responses, our findings support further investigation of GD2BATs in larger phase II clinical trials.

Keywords: Adoptive cell therapy - ACT; Neuroblastoma.

Figure 1

(A) Protocol schema. It shows the treatment schema for the phase I/II study. (B) Survival curves for phase I and phase II. It shows the Kaplan-Meier overall survival curves for phase I (blue, OS=18.0 months), phase II (red, OS=31.2 months), and all patients (green, OS=21.0 months). GM-CSF, granulocyte macrophage colony-stimulating factor; OS, overall survival.

Figure 2

(A) A bone marrow response with clearance of neuron-specific enolase (NSE) positive tumor cells in the bone marrow of an NB patient IT20111 after 8 infusions of GD2BATs. (B) The metabolic PET changes after five infusions of GD2 BAT in an OST patient IT20115. Arrow marks point the intracranial frontal and right maxillary sinus tumor extention before and after GD2BATs infusions. (C) Flow cytometry staining for CD4+ and CD8+ to evaluate presence of tumor infiltrating lymphocytes in the tumor for Pt # 20 115. Most intranasal tumor infiltrating lymphocytes were CD8⁺. Cells were gated on CD3⁺ cells and analyzed for FITC positive CD4 and PE positive CD8 T cells. (D) The proportion of CD3+A1 G4+ cells in the peripheral blood mononuclear cells of patient # 8 preinfusion and at various times during and after GD2BATs infusions, % of CD3⁺A1G4⁺ cells shows the persistence of GD2BATs in peripheral blood. CD3⁺A1G4⁺ cells were measured by using two color FACS gated on lymphocytes using anti-CD3 and A1 G4 anti-idiotypic antibody for hu3F8. PET, positron emission tomography; SUV, standardized uptake value.

Figure 3

(A) Patient’s IT00013 retroperitoneal soft tissue mass (white arrow mark) before (left panel) and after (right panel) eight infusions of GD2BATs. (B) Patient’s IT00031 intracranial dural mass (yellow arrow mark) before (left panel) and after (right panel) eight infusions of GD2BATs.

Figure 4

Anti-GD2 cytotoxicity (IFN-γ EliSpots). Each panel shows the pre-IT and post-IT IFN-γ EliSpots responses from PBMC after overnight stimulation. Phase II portion: (A–C) Responses of PBMC from the phase II NB patients to KCNR (p<0.02), MG63 (p<0.02), and K562 (NS). Phase I/II combined: (D–F) IFN-γ EliSpots responses of PBMC from all 18 of the phase I/II patients pre-IT and post-IT after overnight exposure to KCNR (p<0.03), MG83 (p<0.04), and K562 (not significant, NS), respectively. PBMC, peripheral blood mononuclear cells.

Figure 5

Cytokines/chemokines: (A) (IL-12) shows significant changes between pre-IT and post-IT levels of IL-12 in all 18 phase I/II patients (p<0.002). (B) (IL-12) shows a significant change between pre-IT and post-IT IL-12 levels for 10 NB in the phase II patients analyzed separately (p<0.02). (C) (MIP-1β) shows post-IT levels of MIP-1β increased significantly over pre-IT levels in 28 phase I/II patients (p<0.02). (D) (IL-10) shows the levels of IL-10 significantly increased from pre-IT to post-IT in all 18 phase I/II patients (p<0.02).