GD2-targeting CAR T-cell therapy for patients with GD2+ medulloblastoma.

PURPOSE:Medulloblastoma (MB), the most common childhood malignant brain tumor, has a poor prognosis in about 30% of patients. The current standard of care, which includes surgery, radiation and chemotherapy, is often responsible for cognitive, neurologic and endocrine side effects. We investigated whether chimeric antigen receptor (CAR) T-cells directed towards the disialoganglioside GD2 can represent a potentially more effective treatment with reduced long-term side effects. EXPERIMENTAL DESIGN:GD2 expression was evaluated on primary tumor biopsies of MB children by flow-cytometry. GD2 expression in MB cells was evaluated also in response to an EZH2 inhibitor (Tazemetostat). In in vitro, as well as in in vivo models, GD2+MB cells were targeted by a CAR-GD2.CD28.4-1BBζ (CAR.GD2)-T construct, including the suicide gene inducible-caspase-9. RESULTS:GD2 was expressed in 73.17% of MB tumors. The SHH and G4 subtypes expressed the highest levels of GD2, while the WNT subtype the lowest. In in-vitro co-culture assays, CAR.GD2 T-cells were able to kill GD2+MB cells. Pre-treatment with Tazemetostat upregulated GD2 expression, sensitizing GD2dimMB cells to CAR.GD2 T-cells cytotoxic activity. In orthotopic mouse models of MB, intravenously injected CAR.GD2 T-cells significantly controlled tumor growth, prolonging overall survival of treated mice. Moreover, the dimerizing drug AP1903 was able to cross the murine blood brain barrier and to eliminate both blood circulating and tumor infiltrating CAR.GD2 T-cells. CONCLUSIONS:Our experimental data indicate the feasibility of CAR.GD2 T-cell therapy. A phase I/II clinical trial will be conducted to evaluate the safety and therapeutic efficacy of CAR.GD2 therapy in high-risk MB patients.